From mwilson at the-wire.com  Mon Apr  7 08:26:24 2003
From: mwilson at the-wire.com (Mel Wilson)
Date: Mon, 07 Apr 2003 08:26:24 -0400
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com>
 <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com>
 <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com>
 <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com>
 <09is8votcf3116a94euite6vhdfffgefko@4ax.com>
 <slrnb8ti36.ldu.rmunn@localhost.localdomain> <x7smsvh7kr.fsf@guru.mired.org>
Message-ID: <w5Wk+ks/KT/A089yn@the-wire.com>

In article <x7smsvh7kr.fsf at guru.mired.org>, Mike Meyer <mwm at mired.org> wrote:
>Robin Munn <rmunn at pobox.com> writes:
>>     xover(start, end, open('foo.txt', 'w'))
>>
>> See? File-like object gets created, passed to your function, and then
>> automatically garbage-collected (and thus closed) when your function
>> returns.
>
>The garbage collection happening when your function returns is
>implementation specific. It works that way in CPython, but not
>Jython. You should always explicitly close files to make sure they get
>closed - and the output flushed, in this case - when you're done with
>them.

   That's why I prefer to think of Jython as a different
language.  The idea of a service that can clean up and quit
when the last client stops using it is just too appealing.

        Regards.        Mel.



From robin at jessikat.fsnet.co.uk  Thu Apr 17 19:06:47 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Fri, 18 Apr 2003 00:06:47 +0100
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org>
 <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk> <3E9EA707.4DB53582@engcorp.com>
 <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk>
 <mailman.1050587068.2733.python-list@python.org>
Message-ID: <c37WqKAHOzn+EwJz@jessikat.fsnet.co.uk>

In article <mailman.1050587068.2733.python-list at python.org>, Jp
Calderone <exarkun at intarweb.us> writes
......
>
>  Only if you direct the evolution in the wrong way >:)
>
>  Jp
.....
well if you know how to direct the solution the problem is solved (in
principle at least) :)
-- 
Robin Becker



From peter at engcorp.com  Tue Apr 22 05:22:25 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 22 Apr 2003 05:22:25 -0400
Subject: Python implementation for a web spider \ site mirroring
References: <662cede5.0304220114.54e4dd10@posting.google.com>
Message-ID: <3EA509D1.EDCA8733@engcorp.com>

Tzahi wrote:
> 
> I need to copy complete web sites form the web.
> 
> I think Python is a natural fit for this task, with the intensive HTML
> and HTTP support. Being lazy, I would rather have someone else's
> implementation, so that I do not start from scratch.

Perhaps you could use Tools/websucker.py from the standard distribution.

-Peter



From km-news1 at onlinehome.de  Tue Apr 22 09:31:40 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Tue, 22 Apr 2003 15:31:40 +0200
Subject: Pythonwin don't show the complete list in interactiv window!?
References: <oprnzwcechmu3s0q@news.cis.dfn.de> <d17pa.6563$8K2.53295@news-server.bigpond.net.au>
Message-ID: <oprn0862owmu3s0q@news.cis.dfn.de>

>> The interactive windows shows only a part of the list (last value 771).
>> (Win2000)
>
> It is a Scintilla bug or limitation. Newer versions of Scintilla don't
> stop at 771 but there are still limits.

Thanks.

For a Python beginner it's a little frustrating to search bugs in my own 
code but could not find and then later find out, that it was a tools bug.
Because Idle also has bugs (for example, after a little working time won't 
save my source file so i lost the new source-version!) i switched to 
ConTEXT (http://www.fixedsys.com/context/).

-- 
Gru? - regards Klaus :-)



From max at alcyone.com  Fri Apr 11 23:59:12 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 11 Apr 2003 20:59:12 -0700
Subject: Singletons
References: <roy-20E369.23282211042003@reader1.panix.com>
Message-ID: <3E978F10.6EA61594@alcyone.com>

Roy Smith wrote:

> It seems like a pretty reasonable implementation, but I don't
> understand
> one thing about it.  What is the point of:
> 
> # Store instance reference as the only member in the handle
> self.__dict__[ '_Singleton__instance' ] = Singleton.__instance
> 
> I don't see that it does anything useful.  Is there something subtle
> that I'm missing?

I presume you're confused about the '_Singleton__instance' key?  It's
because a double underscore for a class or instance attribute is
mangled.  For instance:

>>> class C:
...  def __init__(self):
...   self.__x = 123
... 
>>> c = C()
>>> dir(c)
['_C__x', '__doc__', '__init__', '__module__']
>>> c.x
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AttributeError: C instance has no attribute 'x'
>>> c._C__x
123

So this pattern is really just diddling with the new object and
assigning something to a "private" attribute (e.g., one prefixed with a
double underscore).

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ When you talk to her / Talk to her
\__/ India Arie
    WebVal / http://www.alcyone.com/pyos/webval/
 URL scanner, maintainer, and validator in Python.



From nikolai.kirsebom.NOJUNK at siemens.no  Thu Apr 24 07:29:44 2003
From: nikolai.kirsebom.NOJUNK at siemens.no (Nikolai Kirsebom)
Date: Thu, 24 Apr 2003 13:29:44 +0200
Subject: Question about accessing class-attributes.
Message-ID: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com>

Given the following class hierarchy:
	
class A(object):
	InstCount = 0
	def __init__(self):
		self.__class__.__dict__['InstCount'] += 1

class B(A):
	InstCount = 0

class C(A):
	InstCount = 0

Is the syntax used for incrementing the class-attribute for every
instanciation 'the way' it should be done ?

Nikolai Kirsebom






From andrew-pythonlist at puzzling.org  Sat Apr 19 13:10:22 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 20 Apr 2003 03:10:22 +1000
Subject: Parent references: hot or not?
In-Reply-To: <1050772061.14712.55.camel@ida.dylanreinhardt.com>
References: <1050763980.14721.41.camel@ida.dylanreinhardt.com> <20030419145756.GA25916@meson.dyndns.org> <1050771656.14717.48.camel@ida.dylanreinhardt.com> <1050772061.14712.55.camel@ida.dylanreinhardt.com>
Message-ID: <20030419171022.GC1316@frobozz.local>

On Sat, Apr 19, 2003 at 10:07:41AM -0700, Dylan Reinhardt wrote:
> On Sat, 2003-04-19 at 10:00, Dylan Reinhardt wrote:
> 
> > Sounds like I should be sure to write a __del__ method for Container
> > that ensures things work correctly.
> 
> And upon re-reading your post you say *not* to do that... :-)
> 
> I'd like to understand that a bit better.  What would be wrong with:
> 
> class Container:
> 
>    def __del__(self):
>       # ensure all Items are deleted first
>       for item in self.items.keys():
>          del self.items[item]

Calling self.items.clear() would be simpler than looping.

But it's redundant.  The garbage collector will collect cycles for you
automatically, so don't waste your time doing unnecessary house-keeping --
particularly because if you *do* define __del__, but somehow accidentally
end out with that object in a reference cycle anyway, it can't be collected.
See http://python.org/doc/current/lib/module-gc.html#l2h-312 for a brief
explanation.  

There are very few reasons to define __del__ in modern versions of Python.

-Andrew.





From martin at v.loewis.de  Sun Apr  6 17:40:25 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Sun, 06 Apr 2003 23:40:25 +0200
Subject: Debugging embedded python
In-Reply-To: <u4r5bgyfx.fsf@online.no>
References: <u4r5bgyfx.fsf@online.no>
Message-ID: <b6q6s1$1c9$00$1@news.t-online.com>

Syver Enstad wrote:
> I have the python interpreter embedded in a daemon/service (this is
> win2k so it's a service). I'd really like to be able to debug the
> python code that is running. I was thinking of popping up a console
> window to be able to run pdb, how does one go about it to accomplish
> this? Are there any code doing similar things or articles.

I recommend that you find a non-daemon mode of operation, and support
debugging there. This has tradition on Unix (atleast): many daemons
support a do-not-fork-into-background-do-not-fork-many-processes mode,
to simplify debugging.

If you find you must study the running service, use print statements.
There can be very advanced forms of print statements, like the cgitb
module for CGIs (which are also a form of embedding).

Regards,
Martin




From max at alcyone.com  Mon Apr 21 15:45:35 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 21 Apr 2003 12:45:35 -0700
Subject: coverting list back to string
References: <3cf72d52.0304211132.799d8afb@posting.google.com>
Message-ID: <3EA44A5F.6ADA09B8@alcyone.com>

scn wrote:

> my quest:  i know that output.write(remove) is not correct.  you
> cannot use the 'write' method on a list.

Presumably you want to join the list back into a string, with something
like:

	' '.join(remove)

or in general

	S.join(L)

where S is a delimiter string and L is a sequence.  Note that this may
have an surprising effect since L.split() with no argument will split on
_any_ block of whitespace, whereas ' '.join(L) will only join the parts
back with a single space:

>>> L = 'this has  a   lot    of    spaces'.split()
>>> ' '.join(L)
'this has a lot of spaces'

This may not bother you, but I thought I'd point it out as a clear
difference.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The revolution will not televised
\__/ Public Enemy
    Maths reference / http://www.alcyone.com/max/reference/maths/
 A mathematics reference.



From claird at lairds.com  Mon Apr 14 17:39:16 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 14 Apr 2003 21:39:16 -0000
Subject: inserting Python commands into HTML documents
References: <zpqdne0Kztcb0AejXTWcow@comcast.com> <6-CdnUrJoLKFkQajXTWcqA@comcast.com>
Message-ID: <v9mak43akfope4@corp.supernews.com>

In article <6-CdnUrJoLKFkQajXTWcqA at comcast.com>,
Em  <mcclaine at comcast.net> wrote:
>These options all seem to require me to install 3rd party software onto 
>the server of this company, and I believe that to be outside of the 
>scope of the test project.
>
>The only issue is that I want it to be invisible that I'm pulling from a 
>CGI script (I don't want the URL bar to say 
>"www.mywebsite.com/cgi-bin/myProj.py?page=projects", I want it to simply 
>say "www.mywebsite.com/projects.html" or "www.mywebsite.com/projects.php").
>
>Is there a way for me to send my output to /index.html, /projects.html, 
>etc.?
			.
			.
			.
We're using words in different ways.  Crudely, yes,
all the usual active page methods can pass data
either visibly (".../myProj.py?page=projects") or
not-so-visibly (".../myProj.py")--withOUT regard to
whether CGI is involved or not.

And equally without regard to language; PHP can do 
just as much and just as little, in this regard.

Given this, I conclude I have no idea what you mean
by "send my output to /index.html ..." and so on.
I'm reasonably confident Python can do it, whatever
it is.

The problem remains ill-specified.  One of the things
is that we'll need to know about the Web server avail-
able.  If someone has asked you, "Demonstrate how to
use a Web server that only supports PHP but without
using PHP", then my summary will be that he's hazing
you.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From sean at activeprime.com  Tue Apr 29 16:42:40 2003
From: sean at activeprime.com (Sean)
Date: 29 Apr 2003 13:42:40 -0700
Subject: Determining Unicode encoding.
Message-ID: <ef7a350b.0304291242.42c9b024@posting.google.com>

I'm really new to dealing with unicode, so please bear with me.  I'm
trying to add unicode support to a program I'm working on, and I'm
getting stuck a little when printing a unicode string to a file.  I
know I have to encode the string using an encoding (UTF-8, UTF-16,
latin-1, etc).  The problem is that I don't know how to determine what
the *right* encoding to use on a particular string is.  The way I
understand it, utf-8 will handle any unicode data, but it will
translate characters not in the standard ASCII set to fit within the
8-bit character table.  My problem is I'm handling data from a lot of
different encodings (latin, eastern, asian, etc) and I can't allow
data in the strings to be changed.  I also can't (at least I don't
know how to) determine what encodings the strings are using.  IE, I
don't know what strings are from what languages.  Is there any way to
determine, from the unicode string itself, what encoding I need to use
to prevent data loss?   Or do I need to find a way to determine
beforehand what encoding they are using when they are read in?

Am I even asking the right questions?  I'm really pretty lost and my
O'Reilly books arn't helping very much.

-Sean



From xxdigitalhellxx at hotmail.com  Mon Apr  7 15:18:43 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Mon, 07 Apr 2003 13:18:43 -0600
Subject: super question
In-Reply-To: <fd1c664e.0304071056.31a0d377@posting.google.com>
References: <fd1c664e.0304071056.31a0d377@posting.google.com>
Message-ID: <b6sj0l$m08$1@heckle.cc.uregina.ca>

Danra wrote:
> Hi,
> 
> I must be doing something wrong, but it seems that the typical use for
> the 'super' function, suggested in its help and in the python library
> refernce, does not work for me.
> 
> Here's my experience with 'super':
> 
> 
>>>>class b:
> 
> 	pass
> 
> 
>>>>class c(b):
> 
> 	pass
> 
> 
>>>>super(c)
> 
> Traceback (most recent call last):
>   File "<pyshell#6>", line 1, in ?
>     super(c)
> TypeError: super() argument 1 must be type, not classobj
> 
> 
>>>>super(c,c())
> 
> Traceback (most recent call last):
>   File "<pyshell#7>", line 1, in ?
>     super(c,c())
> TypeError: super() argument 1 must be type, not classobj
> 
> What am I doing wrong?
> 
> P.S. pardon the pun in the question's subject :)

I'm not sure if this is what you're looking for, but here it is...

 >>> class b:
...     pass
...
 >>> class c(b):
...     pass
...
 >>> super(type(c))
<super: <class 'class'>, NULL>
 >>> super(type(c), c)
<super: <class 'class'>, <class object>>

I'm not even familiar with the function and didn't even look it up, but 
maybe this is what you're looking for :)

-- 
Ryan




From b.maryniuk at forbis.lt  Thu Apr 24 08:06:20 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Thu, 24 Apr 2003 14:06:20 +0200
Subject: What means exactly "Memory error"?
Message-ID: <200304241406.20833.b.maryniuk@forbis.lt>

Hello, all.

I made an application which uses 4DOM. I have a big (~10M) autogenerated HTML 
file from an Oracle tool. The problem is to public it. Nobody want to open 
10MB HTML file in their browser. 

OK, so the main idea was to split it using frequently repeated <HR> tag. Then 
I got ~1500 parts. OK, now I parse each part with HtmlLib and keep it tree 
somewhere in the variable (in memory). Then after each part is parsed I want 
to do a lot of changes inside the document to adapt it to using with Zope in 
the cycle.

Using this idea with the files up to 2-3M -- no problem. But when I took 
biggest one, it crashes with strange weird stuff I can't exaplain. Somebody 
knows what "Memory error" might appear and how to avoid (if possible) it?

Used software:
	Python 2.2
	4Suite (from the CVS) 
	latest PyXML. 
	OS Linux SuSE 7.1 Pro with it originally compiled kernel 2.4.0-64GB-SMP

Hardware:
	Four CPU machine: Pentium III (Cascades) 700Mhz
	RAM (size): 4GB
	A lot of free space on hard disks...

Full traceback is here:
-----------------------8<-------------------------------------
Traceback (most recent call last):
  File "./mono2chunk.py", line 244, in ?
    mc.processDocument()
  File "./mono2chunk.py", line 138, in processDocument
    self.__doc_complete[_ref] = 
self._evaluateChunk(self.__parseHtmlDoc(partData), _ref)
  File "./mono2chunk.py", line 67, in __parseHtmlDoc
    doc = HtmlLib.FromHtml(''.join([Mono2Chunks.__html_header_,
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/HtmlLib.py", 
line 89, in FromHtml
    return Reader().fromString(text, ownerDoc, charset)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/HtmlLib.py", 
line 70, in fromString
    return self.fromStream(stream, ownerDoc, charset)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/HtmlLib.py", 
line 28, in fromStream
    self.parser.parse(stream)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/Sgmlop.py", 
line 57, in parse
    self._parser.parse(stream.read())
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/Sgmlop.py", 
line 166, in finish_starttag
    parent.appendChild(element)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/FtNode.py", line 256, in 
appendChild
    newChild._4dom_fireMutationEvent('DOMNodeInserted',relatedNode=self)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/FtNode.py", line 378, in 
_4dom_fireMutationEvent
    relatedNode,prevValue,newValue,attrName)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/Event.py", line 120, in 
initMutationEvent
    Event.initEvent(self,eventTypeArg, canBubbleArg, cancelableArg)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/Event.py", line 93, in 
initEvent
    self.bubbles = canBubbleArg
MemoryError
-----------------------8<-------------------------------------

-- 
Regards, Bogdan

A witty saying proves nothing.
                -- Voltaire





From rcook at llnl.gov  Tue Apr  1 23:05:17 2003
From: rcook at llnl.gov (Richard Cook)
Date: Tue, 1 Apr 2003 20:05:17 -0800
Subject: PyObject_CallFunction with keywords?
In-Reply-To: <p05200f08baaff8e7aee8@[134.9.17.165]>
References: <p05200f08baaff8e7aee8@[134.9.17.165]>
Message-ID: <p05200f09bab011e58a96@[134.9.17.165]>

Never mind, duh:
PyEval_CallObjectWithKeywords seems like it may do the trick.  :-)

At 6:39 PM -0800 4/1/03, Richard Cook wrote:
>Hi,
>Short question:
>Is it possible to pass keywords to a function with PyObject_CallObject()?
>
>
>Long question:
>I am using an embedded Python interpreter in our C program to 
>execute a method from a Python module (pexpect, to be exact).  At a 
>certain point, I wish to do the equivalent of the following Python 
>code in my C code:
>
>result =  obj.expect("hello", timeout=50);
>
>So I have a PyObject *f which points to obj.expect, and a PyObject 
>*s which is a list containing the string "hello".  I figure I have 
>to do:
>PyObject_CallObject(f, s, k)
>where k is the timeout parameter, but this does not match the param 
>list for PyObject_CallObject().
>
>My question is, how can I pass keyword parameters such as the 
>timeout parameter in my example to a function which accepts them? 
>The documentation for PyObject_CallObject says it is equivalent to 
>apply(o, args), but says nothing about the actual apply syntax:
>apply(o, args, keywords)
>where keywords is a dictionary.
>
>--
>Richard Cook
>Lawrence Livermore National Laboratory
>Bldg-451 Rm-2043, Mail Stop L-561
>7000 East Avenue,  Livermore, CA, 94550, USA
>phone (925) 423-9605 (work)    fax (925) 423-8704
>---
>Information Management & Graphics Grp., Services & Development Div., 
>Integrated Computing & Communications Dept.
>(opinions expressed herein are mine and not those of LLNL)


-- 
Richard Cook  
Lawrence Livermore National Laboratory
Bldg-451 Rm-2043, Mail Stop L-561       
7000 East Avenue,  Livermore, CA, 94550, USA
phone (925) 423-9605 (work)    fax (925) 423-8704
---
Information Management & Graphics Grp., Services & Development Div., 
Integrated Computing & Communications Dept.
(opinions expressed herein are mine and not those of LLNL)




From mcclaine at comcast.net  Mon Apr 14 01:56:37 2003
From: mcclaine at comcast.net (Em)
Date: Mon, 14 Apr 2003 01:56:37 -0400
Subject: Inserting Python commands into HTML documents
Message-ID: <zpqdne0Kztcb0AejXTWcow@comcast.com>

Hello,

I am an extremely new member of the Python family. I had a job interview 
  where they charged me with the task of creating a (fake) corporate 
website using Python in ten days--with two left now.

The problem is that what I have read seems to say Python is a CGI 
language... Am I remiss in believing that my website is going to be 
completely run from inside a couple .PY files in the CGI-BIN?

Is there a way for it to act more like PHP, as in run Python commands 
from inside the (index|contacts|about|etc).html or at least simulate 
that behavior? (This is on a linux server and I am not to use PHP, FYI.)


Thank you all for any assistance!




From gh at ghaering.de  Tue Apr  8 04:49:43 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: 8 Apr 2003 08:49:43 GMT
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com> <urnka.4658$Du.16154@newsc.telia.net> <i6059vgtvavjp99ilkfeo8rtc003mha57l@4ax.com>
Message-ID: <slrnb95307.1nc.gh@haering.opus-gmbh.net>

Rainer Klueting <rainer.klueting at epost.de> wrote:
> [...]
> I had already seen that there is another GUI toolkit called wxWindow,
> and I even had downloaded it. But my impression was, that Tkinter is
> kind of a GUI-standard for Python, and I tried to stick to the
> standard.
> 
> Is that true?

Tkinter is "standard" in the sense that it is shipped with Python. No
more, no less.

> Do you have any information about how widespread the use of
> wxWindows is? I don't want to steer into a dead-end-road.

For the usage of wxPython, the only reliable number you'll likely get
is the traffic on the wxPython mailing list. It'll show you that it is
widely used among Python programmers. [1]

-- Gerhard

[1] Those that do GUIs, anyway.



From rmunn at pobox.com  Wed Apr  2 12:26:23 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 17:26:23 GMT
Subject: Slicing in Python 2.3a2
References: <mailman.1049200983.30254.python-list@python.org>
Message-ID: <slrnb8lia7.ccp.rmunn@localhost.localdomain>

S?bastien Keim <s.keim at laposte.net> wrote:
> I have some troubles to understand how work slicing in python 2.3.
> In my examples, I will use the following list:
> >>> A = range(10)
> 
> Python 2.3 add a new argument to slice wich can for sample be used to
> get only the items at positive indexes:
> >>> A[::2]
> [0, 2, 4, 6, 8]

I think of the parts of the slice as "start, stop, step", and an empty
start *or* stop means "start/stop at the end of the sequence". *Which*
end depends on what direction you're going. For instance, the above
means "start at one end, stop at one end, step forward by twos". The
"start" end is the beginning, the "stop" end is the end.

Oh dear, I'm using the word "end" for two different meanings here. Let
me disambiguate by using the word "edge" instead.

   [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
   ^                            ^
   \- "front" edge              \- "back" edge

So A[::2] means "start at an edge, stop at an edge, step forward by 2".
Since the step is forward, the slice will start at the front edge and
stop at the back edge.

> This works also for subslices: A[x:y:z] is equal to A[x:y][::z]
> >>> A[1::2]
> [1, 3, 5, 7, 9]

Start at index 1, stop at an edge, step forward by 2. Since the step is
forward, the slice will stop at the back edge.

> With a negative value you can reverse a list:
> >>> A[::-1]
> [9, 8, 7, 6, 5, 4, 3, 2, 1, 0]

Start at an edge, stop at an edge, step backward by 1. Since the step is
backward, the slice will start at the back edge and stop at the front
edge.

> But I don't understand the result for a subslice with a negative third 
> argument:
> >>> A[1::-1]
> [1, 0]

Start at index 1, stop at an edge, step backward by 1. Since the step is
backward, the slice will stop at the front edge.

> Intuitively, I expected to get:
> >>> A[1:][::-1]
> [9, 8, 7, 6, 5, 4, 3, 2, 1]
> 
> The numarray package has the same behavior, so I guess it's intentional.
> >>> import numarray
> >>> numarray.array(range(10))[1::-1]
> array([1, 0])
> 
> Then what is the real semantic for this third argument?

The semantics I illustrated above appear to work for all the cases I
know of. Do you know of any counterexamples?

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From fperez528 at yahoo.com  Wed Apr  9 17:13:02 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 09 Apr 2003 15:13:02 -0600
Subject: Wraparound problems with time.clock() on Linux
Message-ID: <b722d0$bau$1@peabody.colorado.edu>

Hi all,

I wrote the following simple code to take some timing information along the
way as some very cpu-intensive code executes:

#cut-----------------------------------------------------------------------

import time

class Timer:
    """Stopwatch-like class to take timings during code execution.

    It is not meant to be an ultra-accurate tool, since it doesn't calibrate
    its own overhead.  But it will work well for timing reasonably
    long-running codes."""

    def __init__(self):
        """Return a new timer."""
        self.clock = time.clock
        self.start = self.clock()
        self.times = [self.start]
        self.splits = [0]
        self.tagged_splits = []

    def split(self,tag=None):
        """Take a split time reading."""
        self.times.append(self.clock())
        self.splits.append(self.times[-1]-self.times[-2])
        if tag:
            self.tagged_splits.append((tag,self.splits[-1]))

    def split_print(self,tag=None):
        """Take a split and print it."""

        self.split(tag)
        if tag is None:
            print "Last split timing: %s CPU seconds." % self.splits[-1]
        else:
            print "Split timing for <%s>: %s CPU seconds." % \
(tag,self.splits[-1])

    def print_tagged_splits(self):
        """Print a formatted table of all tagged splits."""
        print 'Summary of tagged splits (in CPU seconds)'
        print '-----------------------------------------'
        for tsplit in self.tagged_splits:
            print '%s: %s s' % tsplit
            
    def last_split(self):
        """Return the last split."""

        return self.splits[-1]

    def total(self):
        """Return total time elapsed since timer's creation."""
        return self.clock() - self.start

#/cut----------------------------------------------------------------------

The problem I am having is that in certain long runs, I get negative values
for the total() function.  I was wondering if there are wraparound issues I
should be aware of.  Here is an example of the output from two different
runs of my code:

This one was ok:
    CPU total: 831.26 s
    Splits   :
        build rho -> 683.73 s
        build sep Coul -> 0.24 s
        apply sep Coul -> 119.92 s
        build mra Coul -> 0.02 s
        apply mra Coul -> 19.21 s
        build plot arrays -> 8.11 s
        saving plots to disk -> 0.03 s

This one, on the other hand, generated negative times.  I know (from
timestamps on files made by the code as it ran), that the code took about
50 minutes of wall time to actually execute.  Since time.clock() is
supposed to return floats, I wasn't expecting wraparound problems.


    CPU total: -1418.007296 s
    Splits   :
        build rho -> -2211.787296 s
        build sep Coul -> 0.03 s
        apply sep Coul -> 683.05 s
        build mra Coul -> 0.18 s
        apply mra Coul -> 100.78 s
        build plot arrays -> 9.7 s
        saving plots to disk -> 0.0400000000001 s


Any ideas?  For reference, this is running on a RedHat 8.0 Intel box.

Thanks for any input.

cheers,

f.



From claird at lairds.com  Mon Apr 14 16:28:05 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 14 Apr 2003 20:28:05 -0000
Subject: Calling Python Code Snippets within Python Application
References: <236da4ed.0304140143.25abc577@posting.google.com> <slrnb9kumi.1q1.rmunn@localhost.localdomain>
Message-ID: <v9m6eljni1p58c@corp.supernews.com>

In article <slrnb9kumi.1q1.rmunn at localhost.localdomain>,
Robin Munn  <rmunn at pobox.com> wrote:
		.
		.
		.
>Look at the eval() or execfile() built-ins:
>
>    http://www.python.org/doc/current/lib/built-in-funcs.html
>
>Note that eval() evaluates an expression -- if you want to use eval() to
>run chunks of Python code, you might have to wrap those chunks in a
>function and eval() the result of calling that function. execfile() will
>work if the code you want is in a file; if it's in a database, you might
>have to write it to a temporary file and then execfile() that temporary
>file.
			.
			.
			.
?  When is it advantageous to write to a temporary file
and apply execfile() rather than exec directly?
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From sholden at holdenweb.com  Wed Apr 30 09:25:07 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 30 Apr 2003 13:25:07 GMT
Subject: importing modules with the same name but from different directories
References: <mailman.1050892833.31640.python-list@python.org> <b80pdp$81h$1@news.tu-darmstadt.de>
Message-ID: <T8Qra.32135$WQ4.29932@news2.central.cox.net>

"Mirko Koenig" <koenig at v-i-t.de> wrote in message
news:b80pdp$81h$1 at news.tu-darmstadt.de...
> On Mon, 21 Apr 2003 04:42:51 +0200, logistix wrote:
>
>
>
> >> -----Original Message-----
> >> From: python-list-admin at python.org
> >> [mailto:python-list-admin at python.org] On Behalf Of Mirko Koenig Sent:
> >> Sunday, April 20, 2003 9:26 PM
> >> To: python-list at python.org
> >> Subject: importing modules with the same name but from different
> >> directories
> >>
> >> I have a program consisting of many modules. Every modul has its own
> >> directory in a directory called modul, eg.
> >>
> >> .../modul/modul1
> >> .../modul/modul2
> >>
> >> In every modul directory is a diretory for the language files. These
> >> files are used to display the modules label in different languages. eg.
> >>
> >> .../modules/module1/language/german.py
> >> .../modules/module1/language/english.py
> >> .../modules/module2/language/german.py
> >> .../modules/module2/language/english.py
> >>
> >> My problem is:
> >> If is start modul1 from within its modul directory ( .../modules/modul1
> >> ) And modul1 calls modul2, then modul2 doesn't load its own language
> >> file. Instead it loads the langugae file of modul1.
>
> >>
> > Add an __init__.py file to each directory in your project.  This makes
> > the directories packages.  Then you can use a fully qualified path such
> > as "import modules.modul1.language.english as module1_english".  Note
> > that you can't use the same name for different modules at runtime, so
> > use "as xxx." to give them a more distinct name.
>
> Is it possible to to something like this:
> from modules.modul1 import *
> or
> from modules.modul1.language.english import text as modul1_text
> or
> import modules.modul1 as modul1 and than get the languag file
>

Mirko:

You seem to be able to want to do something like

if running_in_german:
    import module1.language.german as mod12lang
    import module2.language.german as mod2lang
elif running_in_English:
    import module1.language.english as mod12lang
    import module2.language.english as mod2lang
else:
    sys.exit("Don't know which language to use")

Note that dots in module names indicate "packaging": a package is a
directory, which can contain a __init__.py file to perform certain special
magic.

Another thing to try: the import of mod1lang could be handled in module1
rather than in the main program, and then the main program can simply "from
module1 import mod1lang", for example, to access variables and function and
class names bound in that module.

What you definitely *can't* do is this:

    m = "modulename_or_path"
    import m as something

To emulate this non-available solution and actually execute an import you
would need to use this:

    m = "modulename_or_path"
    exec "import %s as something" % m

I'd advise against any solution that involves exec'ing or eval'ing unless
there's really no alternative, which is not the case here. What you would
need for this is the __import__() function, built in to the language and
just waiting for you to take advantage of it.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From hungjunglu at yahoo.com  Sun Apr 27 12:35:08 2003
From: hungjunglu at yahoo.com (Hung Jung Lu)
Date: 27 Apr 2003 09:35:08 -0700
Subject: win32api registry methods and exception handling problem
Message-ID: <8ef9bea6.0304270835.96e779f@posting.google.com>

Hi,

The registry handling methods in win32api seem to have problems with
the Python exception handling. It took me some time, but I have
managed to isolate the problem to a simple program. But the problem is
more severe in more complicated situations, and there it seems to
happen with most (or all) registry methods (win32api.Reg...()).

The methods of _winreg module do not exhibit this problem. For now, I
am using _winreg. However, it would be nice if the problem could be
tracked down inside the win32api's source code and be fixed.

regards,

Hung Jung Lu

#----------------------------------------------------------------------
# OS used: Windows XP
#
# Python 2.1
#     program works correctly: correct exception message displayed
#
# Python 2.2.2 (win32all build 152)
#    (a) In Pythonwin intractive window, you get
#        SystemError: error return without exception set
#    (b) When running with python.exe, it crashes
#
# Python 2.3b1 (win32all build 153)
#    (a) In Pythonwin intractive window, you get
#        SystemError: error return without exception set
#    (b) When running with python.exe, program exits silently,
#        which is not the correct behavior
#
# Comments:
#     (a) equivalent _winreg module methods do not have problem.
#     (b) problem seems to happen with most, or all, win32api.Reg...()
#         methods, not in this particular example, but in more
#         complicated programs (involving classes)
#

import win32con
import win32api

def reg_operation():
    hkey = win32api.RegCreateKey(win32con.HKEY_CURRENT_USER,
r'Software\Spam')
    x = 3/0 # or a statement like: raise 'error'

try: reg_operation()
except: x = 5/0 # or a statement like: raise 'error'



From johnroth at ameritech.net  Wed Apr 16 20:07:49 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 16 Apr 2003 20:07:49 -0400
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net> <420ced91.0304152121.2af84a33@posting.google.com> <9t8na.27736$T34.825154@news2.tin.it>
Message-ID: <v9rrq9ercgto01@news.supernews.com>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:9t8na.27736$T34.825154 at news2.tin.it...
> sdhyok wrote:
>
> > Tyler Eaves <tyler at cg1.org> wrote in message
> > news:<bEGma.462$8c4.82 at tornadotest1.news.pas.earthlink.net>...
> >> Nicola Mingotti wrote:
> >>
> >> > import time
> >> > t = time.time()
> >> > # t is the number you asked for
> >> >
> >> > If you want to know more about this look at
> >> > the "time" module .
> >> >
> >> >
> >> > bye.
> >>
> >> The problem is that you can only represent dates from 1970 -
> >> 2030-someting that way.
> >
> > So, this is not the solution I am seeking.
> > Better module to handle wider range of time?
>
> In Python 2.3, module datetime may serve your needs:
>
> >>> datetime.date(2303,4,16).toordinal()
> 840893
> >>> datetime.date(1303,4,16).toordinal()
> 475651

Unfortunately, this appears to be based off of 1AD or
some such. I usually want 4712BC as the base - in other
words, the astronomical Julian date.

> >>>
>
> Of course, dates before the Gregorian calendar's establishment
> (which happened in different years in different nations...!)
> will give somewhat arbitrary results, but apart from that you're
> all right.

Technically, it's called a proleptic Gregorian date, that is, a
date created according to the Gregorian calendar rules before
it was used historically.

> The result is an integer, but it's not hard to turn
> it into a float if that's what you desire for some reason or other.
>
>
> Alex
>





From peter at engcorp.com  Mon Apr  7 23:04:56 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 07 Apr 2003 23:04:56 -0400
Subject: cgi form handler question
References: <3e90ff67$0$13167$5a62ac22@freenews.iinet.net.au> <mailman.1049691790.26318.python-list@python.org> <3E9158CA.C53791E5@engcorp.com> <3d06fae9.0304070522.2f9e7059@posting.google.com>
Message-ID: <3E923C58.4DAD8D8@engcorp.com>

jmdeschamps wrote:
> 
> Peter Hansen <peter at engcorp.com> wrote in message news:<3E9158CA.C53791E5 at engcorp.com>...
> > Ian Bicking wrote:
> > > for name in form.keys():
> > >     if not form[name].value:
> > >         print name
> >
> > The latter suggestion won't usually work, as the data sent
> > from the browser does not even contain the names of the fields
> > which are left empty.
> 
> ...left empty, -unless you use the form
> cgi.FieldStorage(keep_blank_values = 1) to get your form info as in
> http://www.python.org/doc/current/lib/node297.html
> 
> and as someone else might have said, 'Happy CGI-ing',

D'oh!  Did I get caught by assuming the behaviour I'd observed
using cgi.FieldStorage() recently was actually caused by the
browser instead of by cgi.py itself?  Dang... sorry to mislead.

-Peter



From regnivon at netscape.net  Mon Apr 21 15:32:11 2003
From: regnivon at netscape.net (scn)
Date: 21 Apr 2003 12:32:11 -0700
Subject: coverting list back to string
Message-ID: <3cf72d52.0304211132.799d8afb@posting.google.com>

my code:

import xreadlines

input = open('c:/python22/programs/linetest.txt', 'r') # 3 lines of
text

output = open('c:/python22/programs/output.txt', 'w')

for line in input.xreadlines():
    remove = line.split()
    remove[0:1] = []
    output.write(remove)

output.close()
input.close()

my quest:  i know that output.write(remove) is not correct.  you
cannot use the 'write' method on a list.  is there a better way to
convert the list to a string after the first element of the list has
been removed and then output the strings to a file?  i'm a newbie
programmer and need guru help. thank you.



From jerf at jerf.org  Tue Apr  8 00:02:07 2003
From: jerf at jerf.org (Jeremy Bowers)
Date: Tue, 08 Apr 2003 04:02:07 GMT
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <pan.2003.04.07.23.02.00.101458@jerf.org>

On Mon, 07 Apr 2003 01:06:06 +0000, Ben Finney wrote:
> These aren't in a state to be site packages; they're still undergoing
> rapid architectural changes.  Thus, using monthy.pth isn't an option
> (unless you can tell me otherwise).
> 
> Any guidance appreciated as I tread this unfamiliar ground.

Good luck; I asked this question a while ago and the consensus is,
depending on your point of view, "why the hell would you want to do that?"
or "can't be done".

It is interesting that all the Python programs I've seen that are broken
up into files at all (and not just 100+KB hunks of text) are plopped into
one directory.  

IMHO, this is the largest wart by far in Python when it comes to writing
"real" programs that really shouldn't have every file dumped into the same
directory. I also find the solution of putting references to the
application in the Python directory or the other standard solutions
completely unsatisfactory. It's one thing for a library to go live in "the
main namespace", it's quite another for my application to do so; this is
apparently the one namespace in Python that it is *acceptable* to pollute.

For instance, my application has a file in the top-level named "gui.py",
named that for the very good reason that it switches between the (currently)
2 GUI frameworks the program supports. Is it so inconceivable that
some other application might want to use that name too without conflict?

Applications aren't libraries and there should be somewhat better support
for this sort of thing. On the simplest level, one could rise up the
directories until one either hits root or some token file that says "Here,
I'm an application!" and use that as the first dir in the search path.

I'd PEP it but given that nobody else is uncomfortable polluting this part
of the namespace it seems the odds of it going anywhere are slim.

In the meantime, I've made do with the following boilerplate, dropped as
needed in files as I'm going:

import sys
import os
if __name__ != "__main__": # add parent dir to path
    f = __file__
    f = os.path.normpath(os.path.dirname(f) + os.sep + os.pardir)
    if f not in sys.path: sys.path.append(f)                     
else:
    sys.path.append("..")

Repeat the lines as appropriate to add the grandparent, etc. depending on
how deep the file is. 'import something_at_the_root_of_the_app' will then
work as you expect, provided there aren't any name conflicts with
libraries or internal to your app. Those do get hard to manage; I named a
package "commands" without realizing that is actually an obscure Python std
lib package. Note that as the Python library grows, along with the
popularity of certain almost-standard libraries like PIL, this sort of
thing is going to happen more often...

In particular, I have found this helpful in putting all the unit tests in
separate directories named "tests" under the things they are actually
testing, as this allows the test modules to directly import the things
they are testing, regardless of whether they've been packaged or are
available from the bare "import" command, without polluting the directory
itself with more files. (It's bad enough as it is with as much as a .py, a
.pyc, a #.py#, and a .py~ for each Python file when using emacs to write
Python.)




From whisper at oz.net  Tue Apr 15 14:14:02 2003
From: whisper at oz.net (David LeBlanc)
Date: Tue, 15 Apr 2003 11:14:02 -0700
Subject: syntax highlighting file python 2.1
In-Reply-To: <6gPma.6270$hF.51358@news-server.bigpond.net.au>
Message-ID: <GCEDKONBLEFPPADDJCOEEEIOJIAA.whisper@oz.net>

> -----Original Message-----
> From: python-list-admin at python.org
> [mailto:python-list-admin at python.org]On Behalf Of Neil Hodgson
> Sent: Tuesday, April 15, 2003 1:19
> To: python-list at python.org
> Subject: Re: syntax highlighting file python 2.1
> 
> 
> Luigi Monaco:
> 
> > Basically I require a full list of methods, modules or a way how to
> > extract this from a python release. Keywords etc. should
> > suffice from e.g.: /usr/share/vim/vim61/syntax/python.vim.
> 
>    This may help, although it can take a long time:
> http://www.scintilla.org/gen_python_api.zip
> 
>    Neil

PyXR will do this too (and also slowly)

http://members.bellatlantic.net/~olsongt/python/pyxr.html 

David LeBlanc
Seattle, WA USA 





From aleax at aleax.it  Mon Apr 28 10:07:27 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 14:07:27 GMT
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com>
Message-ID: <zAara.22169$K35.739509@news2.tin.it>

Aahz wrote:

> In article <b8hc64$6n2$1 at slb9.atl.mindspring.net>,
> Andrew Dalke <adalke at mindspring.com> wrote:
>>
>>        i = i + 1
> 
> I'm a little curious why you didn't write this as
> 
>     i += 1
> 
> Any reason (i.e. still need to support 1.5.2) or just habit?  If the
> latter, why haven't you changed your habit?

There is no special reason to "change that habit", e.g. not performance:

[alex at lancelot src]$ python Lib/timeit.py 'i=0; i+=1'
1000000 loops, best of 3: 0.284 usec per loop
[alex at lancelot src]$ python Lib/timeit.py 'i=0; i=i+1'
1000000 loops, best of 3: 0.281 usec per loop
[alex at lancelot src]$ python Lib/timeit.py 'i=0; i=i+1'
1000000 loops, best of 3: 0.281 usec per loop
[alex at lancelot src]$ python Lib/timeit.py 'i=0; i=i+1'
1000000 loops, best of 3: 0.284 usec per loop
[alex at lancelot src]$ python Lib/timeit.py 'i=0; i+=1'
1000000 loops, best of 3: 0.282 usec per loop
[alex at lancelot src]$ python Lib/timeit.py 'i=0; i+=1'
1000000 loops, best of 3: 0.278 usec per loop

No statistically significant / measurable difference.  The tiny difference
in conciseness is hardly compelling one way or another, either.


Alex




From jepler at unpythonic.net  Tue Apr 15 12:26:17 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Tue, 15 Apr 2003 11:26:17 -0500
Subject: file read : feature or bug ?
In-Reply-To: <3e9c2eea$0$6528$afc38c87@sisyphus.news.be.easynet.net>
References: <3e9c2eea$0$6528$afc38c87@sisyphus.news.be.easynet.net>
Message-ID: <20030415162616.GB30739@unpythonic.net>

When you're using text files, seek will not operate as you expect.  The
C standard defines that only offsets returned by ftell are legal
arguments to fseek when working with a text file, and Python doesn't
prevent you from shooting yourself in the foot here.

Jeff




From gh at ghaering.de  Tue Apr 15 08:10:45 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: 15 Apr 2003 12:10:45 GMT
Subject: Inserting HTML files as strings in database
References: <mailman.1050407951.26474.python-list@python.org>
Message-ID: <slrnb9ntbc.148.gh@haering.opus-gmbh.net>

vivek kumar <gupt_vive at hotmail.com> wrote:
> hi all,

> I am trying to make a search engine in python. I want to save the html
> files retrieved by the crawler into a Mysql database as strings
> (MEDIUMTEXT). But the characters like ' " etc are giving problems while
> inserting it as a string. [...]

It's generally a bad idea to insert the values into the SQL string
yourself. This, and quoting the values appropriately, is the job of the
DB-API module:

#v+
a_really_ugly_string = "".join([chr(x) for x in range(1,255)])

# cx is the connection object
cu = cx.cursor()
cursor.execute("""
    INSERT INTO MYTABLE(FOO) VALUES (%s)
    """, (a_really_ugly_string,))
cu.commit()
#v-

For inserting binary data (i. e. data containing the null byte) you'll need
a binary column type.

-- Gerhard



From tismer at tismer.com  Sat Apr  5 10:59:03 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sat, 05 Apr 2003 17:59:03 +0200
Subject: Stackless python
In-Reply-To: <221d8dbe.0304050531.107bc603@posting.google.com>
References: <221d8dbe.0304050531.107bc603@posting.google.com>
Message-ID: <3E8EFD47.7060804@tismer.com>

srijit at yahoo.com wrote:
> I am interested to try Stackless Python. I have downloaded python22.dll.zip. 
> How do I install this zip file? I already have standard python 2.2.2 installed.

You either install python22.dll in a local folder. When you start
Python from there, it will be stackless.
If you want to install it permanently, rename the dll in
windows/system32 and copy the new dll there.

In the version you have, there is a stacklessmodule.pyd which
should be in the local folder or be copied into /python22/DLLs.
This is temporary, the new version which I'm about to upload
tomorrow does not have that external file, since it is all moced into
the core dll, for simplicity.

> I also do not see any documentation like tutorials,
> example programs, demos etc.

There are a few test programs, if you download the CVS tree.
There is no public release, yet, this is still pre-alpha
stuff.
You can get all internal documentation by
 >>> help(stackless)

> Is it 100% compatible with standrad Python 2.2.2?

Yes, it should be. If not, let me know.

ciao - chris
-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From exarkun at intarweb.us  Wed Apr 30 11:36:24 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 30 Apr 2003 11:36:24 -0400
Subject: plugin system
In-Reply-To: <WHIra.27759$su3.3307778@twister.nyc.rr.com>
References: <WHIra.27759$su3.3307778@twister.nyc.rr.com>
Message-ID: <20030430153624.GA22939@meson.dyndns.org>

On Wed, Apr 30, 2003 at 04:56:22AM +0000, xam wrote:
> I am interested in writing a plugin system where the main app is python but
> it is capable of accepting helper "plugins" which can be either python or
> C++. I guess I want some sort of an interface spec for the plugins that can
> be implemented in either language. Which technologies should I research? Are
> there other projects that I can refer to as a model -
> 

  http://twistedmatrix.com/documents/TwistedDocs/current/api/

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 41 days, 11:04, 5 users, load average: 0.14, 0.11, 0.03




From oktaysafak at ixir.com  Wed Apr  2 02:43:38 2003
From: oktaysafak at ixir.com (Oktay Safak)
Date: Wed, 2 Apr 2003 10:43:38 +0300
Subject: unicode bug in turkish characters?
Message-ID: <008901c2f8eb$95a02a60$0200a8c0@b>

Hi everybody,

I think there is a problem with the Turkish encoding in Python's 
unicode support. I have Python 2.3a instaled on win98/ME. I came
across some strange behaviour while using the re module, and upon
investigating the issue I have narrowed it down to the following
(I set sys.setdefaultencoding("iso-8859-9") in sitecustomize py
and use IDLE) :

When I try to convert the character "i" to uppercase what comes
out is "I" where it should have a dot on top of it instead. Also,
when I try to convert the uppercase i with dot to lowercase, it
comes out as itself where "i" should be the character produced.
I'm unable to give IDLE window output because turkish characters
are involved which I cannot reproduce here. I did some investigation
with both iso-8859-9 and windows-1254 encodings and found that the
ordinal values before and after togglin the case of chars are as
follows:


                   |     original value     |   after case toggle  |
                   -------------------------------------------------
                   |    8859-9 | win-1254   |   8859-9 | win-1254  |
--------------------------------------------------------------------
small i w/o dot    |    253    |    253     |    73    |     253   |
--------------------------------------------------------------------
small i            |    105    |    105     |    73    |     73    |
--------------------------------------------------------------------
Capital I          |     73    |     73     |    105   |     105   |
--------------------------------------------------------------------
Capital I with dot |     221   |     221    |    221   |     221   |


THE CORRECT VALUES SHOULD INSTEAD BE:

                   |     original value     |   after case toggle  |
                   -------------------------------------------------
                   |    8859-9 | win-1254   |   8859-9 | win-1254  |
--------------------------------------------------------------------
small i w/o dot    |    253    |    253     |    73    |     73*   |
--------------------------------------------------------------------
small i            |    105    |    105     |    221*  |     221*  |
--------------------------------------------------------------------
Capital I          |     73    |     73     |    253*  |     253*  |
--------------------------------------------------------------------
Capital I with dot |     221   |     221    |    105*  |     105*  |


as you see only one of the 8 conversions is correct.

I have looked into the encodings directory and saw that the
iso-8859-9 file contains:

decoding_map.update({
        0x00d0: 0x011e, #  LATIN CAPITAL LETTER G WITH BREVE
        0x00dd: 0x0130, #  LATIN CAPITAL LETTER I WITH DOT ABOVE  ***1
        0x00de: 0x015e, #  LATIN CAPITAL LETTER S WITH CEDILLA
        0x00f0: 0x011f, #  LATIN SMALL LETTER G WITH BREVE
        0x00fd: 0x0131, #  LATIN SMALL LETTER DOTLESS I           ***2
        0x00fe: 0x015f, #  LATIN SMALL LETTER S WITH CEDILLA
})


the *** marked lines are relevant here. The mapping for these in
decimal seem to be:

 221:304

and

 253:305

respectively. So why aren't these the values I get?
I might be missing something but I have a feeling that this is
a bug since the case toggle works perfectly with turkish
characters that do not exist in ascii. With i and I though, which
do exist in ascii, it's all messed up.

Any ideas?





From salvatore at wanadoo.fr  Sat Apr  5 05:17:55 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Sat, 05 Apr 2003 12:17:55 +0200
Subject: In memory File
In-Reply-To: <3E8EABDB.9AA000A0@alcyone.com>
References: <b6m741$dqo$1@news-reader12.wanadoo.fr> <3E8EA197.3AE807EB@alcyone.com> <3E8EA9E4.60704@wanadoo.fr> <3E8EABDB.9AA000A0@alcyone.com>
Message-ID: <3E8EAD53.2070608@wanadoo.fr>

Erik Max Francis wrote:

> Salvatore wrote:
> 
> 
>>I wanted to redirect sys.stdout to StringIO object but il doesn't seem
>>to work
> 
> 
>>>>import StringIO
>>>>import sys
>>>>sys.stdout = StringIO.StringIO()
>>>>print "Told you"
>>>>print >> sys.stderr, sys.stdout.getvalue()
> 
> Told you
> 

I am a poor Python coder ;-)

Thanks Erik





From hwlgw at hotmail.com  Tue Apr 15 03:16:49 2003
From: hwlgw at hotmail.com (Will Stuyvesant)
Date: 15 Apr 2003 00:16:49 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <mailman.1049653811.11491.python-list@python.org> <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <qtom9vopignqfueuke26m2lbd1pn39mjcd@4ax.com>
Message-ID: <cb035744.0304142316.fc542d3@posting.google.com>

[Courageous]
> At the risk of boldly tossing incidental mud up in the face of
> the entirety of all Pythonistas, I really don't think our community
> has enough cohesion or clout to put together a uniform .NET/Web Services
> style package with standards-establishing authority. We're not Microsoft,
> after all, with a cool $billion to dispose of at a sneeze.
> 
> What you'll have to do instead is wait for the competitive feeding
> frenzy to end and allow the technically superior platform to present
> itself. This is both the bane and boon of open software in action,
> IMO, and at a guess it very much applies here.

Well said.  Or written.  Whatever.  I wish english was my native
language.  I would like to add something.  The major problem I face
when I want to use Python for web applications or services, is that
the host or internet provider just does not have Python installed, or
they have the old Python 1.5, like even sourceforge does.  Or no
``#!/usr/bin/env Python`` or something like that configured for CGI
scripts.  Running your own webserver is often just not allowed.

I feel we should concentrate, whoever of us can do such a thing, on
getting Python into *nix distros, especially the kind internet
providers use.  And talk to Apache so they change their default
configuration script or something (I am not sure about the
technicalities here, there is something like mod_python also but how
about support?) so it will not only allow #!/bin/sh but also Python. 
And for the rest of use mere programmers, we can kick ass producing
better tools for web environments.  For example: I have not seen a
good, easy to use, web application that works well with RDF.  How
about setting up a website with Python where you can input your RDF
file and it displays the tree with nice colors and URLs and so forth? 
There are plenty possibilities using Python for things like this, but
you have to show it, working examples, *on the web* or nobody will see
it.

I think CGI is the only standard that will last some time, maybe
better concentrate on that.

-- 
As in Protestant Europe, by contrast, where sects
divided endlessly into smaller competing sects and no
church dominated any other, all is different in the
fragmented world of IBM.  That realm is now a chaos of
conflicting norms and standards that not even IBM can
hope to control.  You can buy a computer that works
like an IBM machine but contains nothing made or sold
by IBM itself.  Renegades from IBM constantly set up
rival firms and establish standards of their own.  When
IBM recently abandoned some of its original standards
and decreed new ones, many of its rivals declared a
puritan allegiance to IBM's original faith, and
denounced the company as a divisive innovator.  Still,
the IBM world is united by its distrust of icons and
imagery.  IBM's screens are designed for language, not
pictures.  Graven images may be tolerated by the
luxurious cults, but the true IBM faith relies on the
austerity of the word.

                -- Edward Mendelson, "The New Republic", 
                   February 22, 1988



From mgerrans at mindspring.com  Sat Apr 12 22:49:08 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Sat, 12 Apr 2003 19:49:08 -0700
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
Message-ID: <b7aj86$q9i$1@slb6.atl.mindspring.net>

Learning Python is an excellent book, but unfortunately quite out of date.
It covers Python version 1.5, whereas 2.2.2 is the current release, with 2.3
in alpha.

Anyway, the syntax to read lines from a file without loading the whole thing
is now quite simple and clean:

for line in file( filename ):
   ... do your thing ...

or if you prefer open():

for line in open( filename ):
   ... do your thing ...





From staschuk at telusplanet.net  Mon Apr 28 00:29:53 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 27 Apr 2003 22:29:53 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <LNBBLJKPBEHFEDALKOLCEEFJEEAB.tim.one@comcast.net>; from tim.one@comcast.net on Sun, Apr 27, 2003 at 10:17:44PM -0400
References: <20030427183847.C28461@tibia.amotlpaa.bogus> <LNBBLJKPBEHFEDALKOLCEEFJEEAB.tim.one@comcast.net>
Message-ID: <20030427222953.C23486@tibia.amotlpaa.bogus>

Quoth Tim Peters:
  [...]
> I beg to differ, sir!  The people with the greatest power to change the docs
> are those who actually submit thoughtful doc patches.  On that basis, you're
> the most powerful fellow I've seen this month <wink>.

<megalomania>First, the Python docs... next, THE WORLD!  MWA HA
HA!  COWER BEFORE ME, YOU PUNY WORMS!</megalomania>

Ahem.  I mean, er, thanks.

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From aleax at aleax.it  Sat Apr 12 11:16:35 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 15:16:35 GMT
Subject: readline clear buffer
References: <slrnb9edf0.tds.luc@grizzly.dont-panic.info> <FWRla.4654$T34.116245@news2.tin.it> <mailman.1050154758.29973.python-list@python.org>
Message-ID: <n5Wla.4446$LB6.127749@news1.tin.it>

Skip Montanaro wrote:

>     Alex> Right: it seems that Python's readline.c does not expose
>     Alex> interfaces to rl_delete_text, rl_kill_text, or other GNU
>     Readline Alex> functions that would allow such behavior.
> 
> GNU readline has a huge API (I count over 340 extern declarations in
> readline/{readline.h,history.h}, the two programmer-visible #include
> files). I believe the approach all along has been to expose just those
> parts of the
> API necessary for the current task at hand.  I realize this means the
> interface will be incomplete, but it's not clear that most of it is
> necessary for common tasks.  So, yes, submit a patch which includes
> wrappers for the functions you're interested in.

I think that's a rather sensible pragmatical approach to finding the
right subset to expose out of a huge API (though proponents of the
"waterfall", AKA "big design up front", sect, are no doubt loading up
their flamethrowers -- if any lurk on c.l.py, that is;-).  And I
notice with pleasure that the OP is now testing out the small patch
to expose text-deletion and plans to submit it, so the incremental
approach to API exposure seems to be still working fine;-).


Alex




From beau at nyc-search.com  Tue Apr  1 16:53:39 2003
From: beau at nyc-search.com (beau at nyc-search.com)
Date: Tue, 01 Apr 2003 16:53:39 -0500
Subject: Python Programmers, NYC
Message-ID: <3E8A0A63.3F682C06@nyc-search.com>

Python Programmers, NYC
http://www.nyc-search.com/jobs/python.html





From bokr at oz.net  Thu Apr 17 17:17:34 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 21:17:34 GMT
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net> <3E9D35FC.FCAD17C9@alcyone.com> <b7mp2u$3k8$0@216.39.172.122> <3E9EEB54.682751BC@alcyone.com>
Message-ID: <b7n5le$srb$0@216.39.172.122>

On Thu, 17 Apr 2003 10:58:44 -0700, Erik Max Francis <max at alcyone.com> wrote:

>Bengt Richter wrote:
>
>> What other use is there (outside of strings and comments)?
>> Have we purposely been set up for gaucheries?
>> ;-)
>
>In shell-like languages, the backquotes are used to mean "Execute what's
>in between the backquotes as a shell command, and then substitute the
>(stdout) output of that command and continue, e.g.:
>
Yes, I'm aware. I thought we were talking about `...` in Python ;-)
(I'm also aware that they're on Guido's regrets list
http://www.python.org/doc/essays/ppt/regrets/PythonRegrets.pdf )

I guess my smiley needed more oomph 8-<;^D

Regards,
Bengt Richter



From erno-news at erno.iki.fi  Fri Apr 18 20:31:52 2003
From: erno-news at erno.iki.fi (Erno Kuusela)
Date: 19 Apr 2003 03:31:52 +0300
Subject: Python Thread Question
References: <bd993a2f.0304170534.245aa2d5@posting.google.com> <mailman.1050587609.6119.python-list@python.org>
Message-ID: <kur87z72fb.fsf@lasipalatsi.fi>

In article <mailman.1050587609.6119.python-list at python.org>, Jp
Calderone <exarkun at intarweb.us> writes:

|   Nope.  Threading costs you context switches.  A multi-threaded app using
| blocking IO will appear "faster" than a single-thread app doing the same,
| but both take longer to run than a single written app using non-blocking IO.

in addition, enabling threading (creating the first theread) turns on
all the locking machinery in the python interpreter. this causes
everything to slow down somewhat.  on my pc, running pystone in
a thread goes about 50% slower than normally.

  -- erno



From gh at ghaering.de  Tue Apr 22 18:09:13 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 23 Apr 2003 00:09:13 +0200
Subject: Can someone send me the latest Gadfly?
In-Reply-To: <0Ljpa.2547$%_3.2056417@newssrv26.news.prodigy.com>
References: <200304220934.38013.scott_list@mischko.com>
	<mailman.1051030174.3152.python-list@python.org>
	<0Ljpa.2547$%_3.2056417@newssrv26.news.prodigy.com>
Message-ID: <3EA5BD89.10503@ghaering.de>

drs wrote:
> "Gerhard H?ring" <gh at ghaering.de> wrote:
>>Scott Chapman wrote:
>> [where to find Gadfly download]
>>>Can someone help out here?
>>
>>http://sourceforge.net/projects/gadfly
>>http://sourceforge.net/project/showfiles.php?group_id=662
> 
> I am very confused ... You obviously did a reasonable amount of research to
> find this, and you are also clearly a sophisticated internet user ... a
> google search on "gadfly python" turned up the db as the second result, and
> an informational page linking to the db as the first.  what google searches
> did you try?

I just remembered seeing an announcement of Gadfly moving to Sourceforge 
with new maintainers here on this list. So I straightly went to 
http://sf.net/projects/gadfly/ :-)

-- Gerhard




From oktaysafak at ixir.com  Wed Apr 23 16:46:37 2003
From: oktaysafak at ixir.com (oktaysafak at ixir.com)
Date: Wed, 23 Apr 2003 22:46:37 +0200
Subject: In praise of PythonCard
Message-ID: <20030423204637.MQHZ239.nairobi@[212.252.122.82]>

> 
> Oktay Safak wrote:
> > Very rude of you, Laotseu ;)
> 
> Yes, I know, I shouldn't have mispelled 'smalltalk' !-)
                                 ^^^^  

You definitely need a good spellchecker :-P





From martin at v.loewis.de  Wed Apr 30 14:31:25 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 30 Apr 2003 20:31:25 +0200
Subject: question on the compile built in function.  Filename ignored?
References: <7rSra.706428$S_4.744795@rwcrnsc53>
Message-ID: <m31xzjhm6a.fsf@mira.informatik.hu-berlin.de>

Alex <delete.this.part.alex.news at attbi.com> writes:

> From the docs I would expect the line 
> 
>   File "<string>", line 1
> 
> to read
> 
>   File "abc", line 1
> 
> 
> What am I missing?

You are right: this is what should have been printed. You have found a
bug.

Actually, you are not the first one to find this bug; it will be fixed
in Python 2.3.

Regards,
Martin




From peter at engcorp.com  Sun Apr 27 20:04:58 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 27 Apr 2003 20:04:58 -0400
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <3EAC2962.2050609@soraia.com>
Message-ID: <3EAC702A.A65C22B9@engcorp.com>

Joe Francia wrote:
> 
> The Deitels have a textbook called _Python How to Program_ that is aimed
> at beginning programmers.  I'm not sure how good/bad it is.  Three out
> of four reviewers on Amazon seem to like it (the fourth guy gave it one
> star based on its lack of Perl coverage).

At a PyGTA meeting three or four months ago someone showed up with
a copy to give away.  I believe it went home with one person only
to come back the next month, to be given away again.  I'm pretty 
sure it wasn't because he'd read it cover to cover.

I'm not sure who got stuck with it last...

(To be more direct: I opened it to skim, and my first comment was
along the lines of "Gah!".  I believe "Yeesh!" and "Ugh" were heard
as well.  To be fair, these comments were directed at the quality of
the layout and the unreadable color scheme, not necessarily at the
content, which I was unable to get near.)

-Peter



From aleax at aleax.it  Sat Apr 19 18:11:35 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 22:11:35 GMT
Subject: Sorting without transitivity
References: <mailman.1050788132.16618.python-list@python.org>
Message-ID: <rQjoa.37575$LB6.855365@news1.tin.it>

Frank Niessink wrote:

> Hi all,
> 
> [It's been while since I've dealt with transitivity and stuff, so
> I hope I got the terminology right]
> 
> I have objects that are not transitivily ordered, e.g.:
> 
> A > B
> A == C
> B == C
> 
> Now I have a list [A, B, C] and I want to sort it.
> The resulting list should have B before A, I don't care where
> C is. However, the Python list.sort() method (Python 2.2.2 on cygwin)
> doesn't put A and B in the right order.

I think what you want is what's known as a *topological
sort* -- a sort obtaining any ordering that respect certain
given pairwise constraints that don't, however, define a
complete order.

> How can I get the desired behavior? Should I implement my own
> sort algorithm? Any pointers?

You can find topological sorts in Python on the net.  E.g.,

http://starship.python.net/crew/aaron_watters/kjbuckets/tsort.py

this one uses kjbuckets, but it's not hard to transliterate
to forms using other representation for the constraints and
other ways to store sets.


Alex




From sholden at holdenweb.com  Tue Apr 15 09:50:24 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 15 Apr 2003 13:50:24 GMT
Subject: Complexity of standard Python data structures
References: <mailman.1050354988.8671.python-list@python.org> <tdnk7dwlfda.fsf@shell01.TheWorld.com>
Message-ID: <A6Uma.99232$It5.2059@news2.central.cox.net>

"Chris Jones" <clj at acme.com> wrote in message
news:tdnk7dwlfda.fsf at shell01.TheWorld.com...
> Tim Peters <tim.one at comcast.net> writes:
>
> > [Tim]
> > > Note that it's not true that Python's list is really a vector/array --
> > > thelanguage doesn't define its implementation, and doesn't want to.
> >
> > [Marcus Alanen]
> > > Hmm.  So the Python language is more about specifying _what_ will
> > > happen, instead of _how_, with no concern for actual speed or memory?
> >
> > The Python language reference manual, and the Python standard library
> > reference manual, are like that, and that's par for the course for
language
> > specs:  they specify syntax ("how is it spelled?") and semantics ("what
does
> > it mean?").  Everything else falls under pragmatics, often called
"quality
> > of implementation" issues.  The STL is close to unique, outside the
> > real-time world, in making promises related to performance.  For the
other
> > side, I spent the first 15 years of my career working largely on
optimizing
> > Fortran compilers and runtime libraries:  no user base was more
concerned
> > about speed than Fortran's, yet the Fortran standard didn't breathe a
word
> > about how fast or slow any construct or standard function might be,
could
> > be, or should be.  Similarly, the IEEE-754 standard for floating-point
> > arithmetic doesn't say anthing about speed either.  Standards bodies are
> > generally loathe to constrain implementations.
> >
> > Of course the Python language itself is distinct from (albeit related
to)
> > its implementations.  People working on implementations are typically
quite
> > concerned about speed and memory.  The pragmatic tradeoffs can (and do)
vary
> > across implementations, though (same as for C, C++, Fortran, Java, etc).
>
> People wanting to program should nail a copy of this posting above their
> keyboard.

READERS: Please DO NOT FOLLOW THIS ADVICE.

Following these instructions, I printed the posting out and nailed it
directly above my keyboard. As a result my laptop no longer works -
apparently a TFT display doesn't take kindly to having nails driven through
it, and now Dell are saying that since this isn't accidental damage I'm not
covered by the extended warranty I purchased. Please therefore send me $650
so I can repair the damage I did by following your doubtless well-meant
advice.

just-having-fun-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From lozinski at jobmart.com  Mon Apr  7 15:30:43 2003
From: lozinski at jobmart.com (lozinski at jobmart.com)
Date: Mon, 7 Apr 2003 12:30:43 -0700 (PDT)
Subject: Plone Developer Needed
In-Reply-To: <20030407192803.29437.14104.Mailman@mail.python.org>
Message-ID: <1030407123043.232AAC/X.lozinski@maya>

A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 612 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030407/7d0fea0a/attachment.bin>

From hwlgw at hotmail.com  Fri Apr 25 12:32:32 2003
From: hwlgw at hotmail.com (Will Stuyvesant)
Date: 25 Apr 2003 09:32:32 -0700
Subject: Vim Newbie ? Python Edit/Compile/Run  cycle and integrating py help questions
References: <a791ab19.0304240532.5a025b81@posting.google.com> <3EA7EBD3.3010605@ghaering.de> <mailman.1051243975.12570.python-list@python.org>
Message-ID: <cb035744.0304250832.71aef2d9@posting.google.com>

[Hans Nowak]
> I have been trying to edit Python source code with vim on Mac OS X.  Most of 
> the settings in my .vimrc file (copied from Windows) work just fine, except for 
> one.  I use Tab to indent and Shift-Tab to dedent (a line, not a block).  Tab 
> behaves as usual, but Shift-Tab doesn't seem to work; it just behaves like I 
> pressed Tab.

In gvim on windows and linux (untested on Mac, since I have no access
to a Mac anywhere):

Start block selection with the v key, or with <SHIFT>v.  And then of
course the j key for line down and the l key for going right etc.

To indent a line, or a selected block, use <SHIFT>>>.  To dedent it do
<SHIFT><<.

-- 
This life is a test.  It is only a test.  Had this been an
actual life, you would have received further instructions as
to what to do and where to go.



From andrew.markebo at telia.com  Sat Apr  5 12:55:25 2003
From: andrew.markebo at telia.com (Andrew Markebo)
Date: Sat, 05 Apr 2003 17:55:25 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com>
 <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com>
 <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com>
Message-ID: <m3d6k0rfqp.fsf@telia.com>

/ carroll at tjc.com wrote:
|
| Looking at nntplib.py, most of the logic to do this is already there:
| xover calls longcmd to actually issue the command, and longcmd can
| take an optional File parameter to which getlongresp (which actually
| retrieves the response) will write its output, if present.

Which format should the xover output be in the file? I mean it is a
list of tuples or similar? How to squeeze that to a file? Have some
separator being escaped? 

Or have I missed something? You seem to have gotten it working?

   /Andy

-- 
 The eye of the linker rests on the libs!



From lists at ghaering.de  Sun Apr 13 09:33:20 2003
From: lists at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Sun, 13 Apr 2003 15:33:20 +0200
Subject: win32 COM test question
In-Reply-To: <cb035744.0304130224.6323b4a0@posting.google.com>
References: <mailman.1050209239.15961.python-list@python.org>
	<cb035744.0304130224.6323b4a0@posting.google.com>
Message-ID: <3E996720.5000608@ghaering.de>

Will Stuyvesant wrote:
> [Roman Yakovenko]
> 
>>http://www.hps1.demon.co.uk/users/andy/pyvb/spamsrv.htm
>>
>>And next time try to search for the answer with google ( python com
>>visual basic )
> 
> As a side note, quoted from above URL:
> 
> "Distribution is a problem - not everyone will want to know they have
> python installed, and a on-step installation that puts it there
> suitably hidden will be needed in time"

py2exe and Installer both solve this problem. At least py2exe can 
package COM clients and servers and it can register a Windows service 
written in Python for you.

-- Gerhard




From msg_2222 at yahoo.com  Mon Apr  7 15:44:14 2003
From: msg_2222 at yahoo.com (Rick Y)
Date: Mon, 7 Apr 2003 12:44:14 -0700 (PDT)
Subject: socket question(3) 
Message-ID: <20030407194414.52865.qmail@web20705.mail.yahoo.com>

installed openssl package from sunfreeware. error:

./viewcvs-install
Traceback (most recent call last):
  File "./viewcvs-install", line 35, in ?
    import compat
  File "./lib/compat.py", line 20, in ?
    import urllib
  File "/usr/local/lib/python2.1/urllib.py", line 26,
in ?
    import socket
  File "/usr/local/lib/python2.1/socket.py", line 41,
in ?
    from _socket import *
ImportError: ld.so.1: python: fatal: libgcc_s.so.1:
open failed: 
No such file or directory


followed these steps:

PATH=/usr/local/ssl/bin:${PATH}
export PATH

LD_LIBRARY_PATH=/usr/local/ssl/lib:${LD_LIBRARY_PATH}
export LD_LIBRARY_PATH

cd /usr/lib
ln -s ../local/ssl/lib/libcrypto.a libcrypto.a
ln -s ../local/ssl/lib/libcrypto.so.0.9.6
libcrypto.so.0.9.6
ln -s ../local/ssl/lib/libssl.a libssl.a
ln -s ../local/ssl/lib/libssl.so.0.9.6 libssl.so.0.9.6
ln -s libssl.so.0 libssl.so
ln -s libssl.so.0.9.6 libssl.so.0
cd /usr/bin
ln -s ../local/ssl/bin/openssl openssl
ln -s ../local/ssl/bin/c_rehash c_rehash
cd /usr/include
ln -s ../local/ssl/include/openssl openssl

__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com




From cpl.19.ghum at spamgourmet.com  Fri Apr 18 06:26:41 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Fri, 18 Apr 2003 12:26:41 +0200
Subject: python and macro excel
References: <5b8834c2.0304170138.12f1a393@posting.google.com>
Message-ID: <Xns93617E38BB105cpl19ghumspamgourmet@62.153.159.134>

Hello Georges,

> i want to integrate a macro in excel with python and execute it 
you allways have to get the win32all extenstion


What exactly do you want to do?

a) write a vba makro and execute it from Python?
o = win32com.client.Dispatch("Excel.Application")
o.Visible = 1
o.Workbooks.Add() # Double Check the actual code, changes sometimes

o.Run("NameOfYourMakro)

b) write a Python Program and execute it from VBA?
recommendation: build a COM-Server in Python, use it from VBA. More 
difficult. 

Harald



From pobrien at orbtech.com  Sun Apr  6 14:27:56 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 13:27:56 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <16016.27005.335657.809556@montanaro.dyndns.org>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	<7xn0j3fsdu.fsf@ruckus.brouhaha.com>
	<m21y0fa5kq.fsf@localhost.localdomain>
	<16016.27005.335657.809556@montanaro.dyndns.org>
Message-ID: <m2vfxr8or7.fsf@localhost.localdomain>

Skip Montanaro <skip at pobox.com> writes:

>     Patrick> The Scintilla text control (written in C and wrapped in
>     Patrick> wxPython by a control called wxStyledTextCtrl) handles the
>     Patrick> on-screen representation and user-manipulation of the text.  So
>     Patrick> that part is actually easy.  Just read in a file, feed it to
>     Patrick> Scintilla, and go.  The next step is wrapping all this in an
>     Patrick> Emacs-like framework.
> 
> If Scintilla doesn't distinguish between Emacs's notion of buffers (a chunk
> of text being edited) and windows (a view onto a buffer) you may find the
> task harder than it would first appear.

I don't think this will be an issue.  I'm implementing my own buffer
class that acts like an Emacs buffer, except that I delegate the
actual rendering and user interface to Scintilla.  Scintilla is used
like the Emacs window, and I can have multiple Scintilla views into
the same buffer and a change in one is reflected in the other.  And
like Emacs windows, each can have its own point into the same buffer.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------




From aahz at pythoncraft.com  Tue Apr 15 23:04:08 2003
From: aahz at pythoncraft.com (Aahz)
Date: 15 Apr 2003 23:04:08 -0400
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi>
Message-ID: <b7ih78$fb2$1@panix1.panix.com>

In article <slrnb9lsbe.4pa.marcus at infa.abo.fi>,
Marcus Alanen <marcus at infa.abo.fi> wrote:
>
>On the other hand, I've always wished for a document "language X for
>people know language Y / programming in general / think they know". We
>could skip all that "let's hold hands while I take you though the
>mysteries of a dictionary" nonsense, and just concentrate on the issue
>at hand: how does our knowledge from language X map to this new
>language? What pitfalls do we have in the language at the moment
>(e.g. python default arguments are in my humble opinion still
>counter-intuitive) ?

Right now, I'm trying to get some momentum moving on a book titled
_Python for VB6 Programmers_.  Anyone interested should subscribe to
marketing-python.

http://wingide.com/mailman/listinfo/marketing-python
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From selwyn at home.net.nz  Tue Apr 22 06:25:13 2003
From: selwyn at home.net.nz (selwyn)
Date: Tue, 22 Apr 2003 22:25:13 +1200
Subject: problem starting python interpreter in emacs (win98)
In-Reply-To: <b8301i$5nl2c$1@ID-114079.news.dfncis.de>
References: <V65pa.4289$AB5.573029@news02.tsnz.net> <b8301i$5nl2c$1@ID-114079.news.dfncis.de>
Message-ID: <%D8pa.4364$AB5.586778@news02.tsnz.net>

> Perhaps you need to add C:\Python22 to the PATH environment variable.

Doh! Works like a charm.

Thanks for your help :-)

Selwyn.




From opengeometry at yahoo.ca  Wed Apr  9 03:35:54 2003
From: opengeometry at yahoo.ca (William Park)
Date: 9 Apr 2003 07:35:54 GMT
Subject: sorting strings written to files
References: <2745895.1049870828@dbforums.com>
Message-ID: <b70igq$9n0ff$2@ID-99293.news.dfncis.de>

onefatman <member27857 at dbforums.com> wrote:
> 
> Can someone help me?
> 
> I have a small phone directory .dat file that has the entries
> written to it as strings with three fields in each, separated by \t
> and ended by \n.
> 
> e.g.
> phbook.write("Joe\tNobody\t62907243\n")
> phbook.write("Daniel\tStevens\t62962328\n")
> phbook.write("Ellen\tWay\t62910758\n")
> phbook.write("Jason\tGreig\t62314587\n")
> 
> it has a function that allows you to write new entries, but i need to
> know how to sort them by first name, and then by last name when the
> first names are the same.

If it's already written to file, then 
    man sort
If it's still in Python, then
    list.sort()

-- 
William Park, Open Geometry Consulting, <opengeometry at yahoo.ca>
Linux solution for data management and processing. 



From dan at grassi.org  Tue Apr  1 16:34:36 2003
From: dan at grassi.org (Dan Grassi)
Date: Tue, 1 Apr 2003 21:34:36 +0000 (UTC)
Subject: How to obtain python file path
Message-ID: <20030401163212631-0500@news.covad.net>

I need to obtain the full path of the python.py file that the program is 
executing from wether imported or run as main.

Sorry but I have looked and can't find the docs.

TIA,

Dan



From achrist at easystreet.com  Fri Apr  4 17:26:09 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Fri, 04 Apr 2003 14:26:09 -0800
Subject: Converting text to html
Message-ID: <3E8E0681.9005524D@easystreet.com>

Is there any recommended code available to convert plain text into
text that can go into an html document without breaking it?  I 
think that I just have to do these three translations:

	< =>  &lt;	> =>  &gt;   & 	=>  &amp;

But I'm not sure if there are any others.  I've got to do this 
quite a bit while the user is waiting, so it would be nice to
have something that did it in a way that was not a lot slower
than any of the alternatives.  

Surely someone must have cracked this nut sys.maxint times by now.

TIA


Al



From s.thuriez at laposte.net  Thu Apr 10 04:42:12 2003
From: s.thuriez at laposte.net (sebastien)
Date: 10 Apr 2003 01:42:12 -0700
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <yfs7ka3iom2.fsf@black132.ex.ac.uk>
Message-ID: <b1ab591.0304100042.1188e4bf@posting.google.com>

Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfs7ka3iom2.fsf at black132.ex.ac.uk>...
> s.thuriez at laposte.net (sebastien) writes:
> 
> > Hi,
> > 
> > Is there any PCA analysis tools for python ?
> 
> What is the analysis tool supposed to do?
> 
> Maybe this will do what you want, once you (downloaded and installed Numeric):
> 
> # Warning: hackish and not properly tested ripped out bit of code ahead
> # so no guarantees whatsoever
> # Anyway, it should at lesat sort of give you the idea
> # try pca(X); if that doesn't do what you want try pca(t(X))
> 
> from Numeric import take, dot, shape, argsort, where, sqrt, transpose as t
> from LinearAlgebra import eigenvectors
> 
> def pca(M):
>     "Perform PCA on M, return eigenvectors and eigenvalues, sorted."
>     T, N = shape(M)
>     # if there are less rows T than columns N, use
>     # snapshot method
>     if T < N:
>         C = dot(M, t(M))
>         evals, evecsC = eigenvectors(C)
>         # HACK: make sure evals are all positive
>         evals = where(evals < 0, 0, evals)
>         evecs = 1./sqrt(evals) * dot(t(M), t(evecsC))
>     else:
>         # calculate covariance matrix
>         K = 1./T * dot(t(M), M)
>         evals, evecs = eigenvectors(K)
>     # sort the eigenvalues and eigenvectors, decending order
>     order = (argsort(evals)[::-1])
>     evecs = take(evecs, order, 1)
>     evals = take(evals, order)
>     return evals, t(evecs)
> 
> 
> You can download Numeric and use it to compute the eigenvalues and
> eigenvectors of an array.
> 
> 
> > If it does, do you have any idea on how well it would scale ?
> 
> It should scale fine. If you experience speed problems, configure Numeric 23
> with ATLAS support (you have to install ATLAS and LAPACK first, of course).
> For large matrices, this should be *much* faster than handwritten C code that
> doesn't use ATLAS.
> 
> > 
> > I have already seen PyClimate (but it is not available for Windows
> > which will be one of the target). Is there some LAPACK like packages ?
> 
> Yes, Numeric and scipy. (www.numpy.org, www.scipy.org, I should think)
> 
> 'as


HI,

Thanks for the program. I have tested it with a test matrix that I
have and that I took from IDL help (Interactive data language).

The matrix is :

((2,1,3),
(4,2,3),
(4,1,0),
(2,3,3),
(5,1,9))


The principal component given by IDL is 

((0.87, -0.7, 0.69),
(0.01, -0.64, -0.66),
(0.49, 0.32, -0.30))

Runnnig the same matrix with the program that you have kindly provided
gives :

((-0.58, -0.67, 0.45),
(-0.23, -0.40, -0.88),
(-0.77, -0.62, -0.074))

Do you know why is there such a difference ??

I did not have the time yet to test it against other program.

Thanks.

Sebastien.



From tim.one at comcast.net  Tue Apr 29 11:43:23 2003
From: tim.one at comcast.net (Tim Peters)
Date: Tue, 29 Apr 2003 11:43:23 -0400
Subject: Tkinter exit problem in Win98
In-Reply-To: <20030429.095109.-177423.0.eltronic@juno.com>
Message-ID: <BIEJKCLHCIOIHAGOKOLHAEPEFHAA.tim.one@comcast.net>

{MvL]
>> That is the oldes Python-Bug, see www.python.org/sf/216289
>> Nobody has been able so far to study in detail what is causing
>> the problem and how it could be fixed.

[eltronic at juno.com]
> is it still a problem in python2.3 with Tk8.4, anyone?

It is in 2.3b1.  I haven't had time to try another Tcl/Tk release since the
last entry in the bug report, and-- alas --don't know whether I'll be able
to make time for it before 2.3 final.  As the bug report says, though, use
pythonw.exe instead of python.exe, and there's no problem.





From timr at probo.com  Fri Apr  4 01:22:29 2003
From: timr at probo.com (Tim Roberts)
Date: Thu, 03 Apr 2003 22:22:29 -0800
Subject: writing to windows memory
References: <3E8C01CE.6070601@netscape.net>
Message-ID: <h49q8v4aoduebq0n0o9h08ask5o3e8d2fl@4ax.com>

vector <Vector180W at netscape.net> wrote:
>
>newbie:
>pythonwin 2.2.2
>Id like to bind a button.widget that simply writes a one to windows 
>address 440h. any ideas on where to start.(the button stuf is trivial) i 
>need to know how to access the address space.

What, exactly, do you think that is going to do?
-- 
- Tim Roberts, timr at probo.com
  Providenza & Boekelheide, Inc.



From eppstein at ics.uci.edu  Tue Apr  1 11:50:42 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Tue, 01 Apr 2003 08:50:42 -0800
Subject: else: w/o except: - why not?
References: <mailman.1049211247.3694.python-list@python.org>
Message-ID: <eppstein-B27196.08504101042003@news.service.uci.edu>

In article <mailman.1049211247.3694.python-list at python.org>,
 Michael Chermside <mcherm at mcherm.com> wrote:

> I think that a "bare except" is useful in only two ways. The first
> is like a "finally" clause that's skipped if everything works.
[...snip...]
> The other situation is running "untrusted plug-in" code of some 
> sort.

I have a bare except wrapped around the main event loop of some python 
gui code.  It provides an opportunity to log the problem and give the 
user a chance to clean up instead of just immediately crashing.
Doesn't seem to fit into either of your two uses, but I think it's a 
legit use.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From dsada at asd.com  Wed Apr  2 11:34:37 2003
From: dsada at asd.com (John Does)
Date: Wed, 02 Apr 2003 16:34:37 GMT
Subject: that python regex viewer is kodos.sf.net [case closed!]
References: <pan.2003.04.02.11.03.07.614824@asd.com> <3E8AC763.C89B8B03@engcorp.com>
Message-ID: <pan.2003.04.02.16.35.51.467730@asd.com>

On Wed, 02 Apr 2003 06:20:03 +0000, Peter Hansen wrote:

> http://kodos.sourceforge.net/


thanks!



From peter at engcorp.com  Wed Apr 16 15:53:16 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 15:53:16 -0400
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com> <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk> <3E9D8B1B.24050D34@engcorp.com> <7h3brz6qmsc.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3E9DB4AC.E94829D9@engcorp.com>

Michael Hudson wrote:
> 
> And unless someone went to the bother of faking a poorly-scanned pdf
> of Knuth's paper, the quotation and attribution I gave is correct.

Oh, I quite believe that.  I've seen it too...

> It is entirely possible that Hoare said something like this *as well*,
> of course.  I don't have a reference, however, and have never seen
> one.

That is exactly what the pages I had found said.  The wording wasn't
quite the same, but the meaning was, and I vaguely recall that even
Knuth attributed the sentiment to Hoare (but I'll have to do another
search myself before I insist on that point at all).

-Peter



From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr 10 10:31:21 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 10 Apr 2003 14:31:21 GMT
Subject: Macros in Python?
References: <b71qde$q71$07$1@news.t-online.com>
Message-ID: <slrnb9b01m.108.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <b71qde$q71$07$1 at news.t-online.com>, Dominic wrote:
> 
> 
> (defmacro with-root-privs (() &body body)
>    (let ((oldidsym (gensym)))
>      `(let ((,oldidsym (geteuid)))
>         (seteuid 0)
>         (unwind-protect
> 	   (progn , at body)
> 	 (seteuid ,oldidsym)))))
> 
> Is there an easy way to produce similar code
> in Python which gurantees that certain things
> happen around some code? (Maybe by using
> generators?)
> 
> The normal approach would be to enclose everything
> with try: finally: but this would scatter the
> code all over.
> 
> Is something like this feasible?
> 
> def server(a,b):
>    if cond:
>        ...
>    else:
>      With_Root_Privs:
>         if x in user:
>            active.append(x)
>           (...)

Python has nothing like Lisp macros, but you can generically wrap
function calls by accepting and passing on arbitrary arguments with
the * and ** operators:

    def with_root_privs(func, *args, **kwargs):
        save_euid = os.geteuid()
        os.seteuid(0)
        try:
            func(*args, **kwargs)
        finally:
            os.seteuid(save_euid)

You can use the above like this:

    def activate_user(x):
        if x in user:
            active.append(x)
        ...
    with_root_privs(activate_user, x)

Alternatively, you can curry the above function:

    def with_root_privs(func):
        def call(func, *args, **kwargs):
            save_euid = os.geteuid()
            os.seteuid(0)
            try:
                func(*args, **kwargs)
            finally:
                os.seteuid(save_euid)
        return lambda *args, **kwargs: call(func, *args, **kwargs)

You can then write something like the following:

    def activate_user(x):
        if x in user:
            active.append(x)
        ...
    activate_user = with_root_privs(activate_user)
    ...
    activate_user(x)



From aahz at pythoncraft.com  Wed Apr  2 19:32:18 2003
From: aahz at pythoncraft.com (Aahz)
Date: 2 Apr 2003 19:32:18 -0500
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <nFKia.141$Sd.25@fed1read01> <mailman.1049328065.20513.python-list@python.org> <3E8B7D49.829034FF@engcorp.com>
Message-ID: <b6fvei$3bu$1@panix1.panix.com>

In article <3E8B7D49.829034FF at engcorp.com>,
Peter Hansen  <peter at engcorp.com> wrote:
>
>I hope Greg's comments about being able to disable this are on the
>mark, because otherwise this would be a clear case of a desire for
>(unnecessary, IMHO) optimization being given higher priority than
>maintaining one of Python's strongest advantages.  Speed will *never*
>be Python's strongest suit, but its dynamicism clearly is; let's not
>bugger that up!

Maybe I'm misunderstanding something, but I was under the impression
that disabling the ability to munge another module's namespace was at
least partly driven by the desire to enable secure code.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From gh at ghaering.de  Tue Apr 15 10:18:03 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Tue, 15 Apr 2003 16:18:03 +0200
Subject: any benchmark results for python?
In-Reply-To: <b7h40n$28s$1@skates.gsfc.nasa.gov>
References: <b7h40n$28s$1@skates.gsfc.nasa.gov>
Message-ID: <3E9C149B.7070706@ghaering.de>

J Li wrote:
> I am looking for benchmark test results for python, in comparison to other
> languages like Perl or Java or C++ [...]

"There are lies, damn lies, and benchmarks."

I'd recommend creating the benchmarks yourself for the particular use 
cases you have in mind. Be sure to exploit each language's features 
optimally lest you forge the results.

If you're just looking for the bare results of using the *same* 
algorithm (no libraries, not using any fancy language features) of 
various language implementations, you can look at:

The Great Computer Language Shootout
http://www.bagley.org/~doug/shootout/

HTH,

-- Gerhard





From anthony at interlink.com.au  Wed Apr 16 22:21:31 2003
From: anthony at interlink.com.au (Anthony Baxter)
Date: Thu, 17 Apr 2003 12:21:31 +1000
Subject: compatibility problem 
In-Reply-To: <1f4c702e.0304161116.5b0f6565@posting.google.com> 
References: <1f4c702e.0304161116.5b0f6565@posting.google.com>  
Message-ID: <200304170221.h3H2LWY02764@localhost.localdomain>

>>> Chad Scherrer wrote
> So I'm wondering what I should look for to make this work under 2.2.2.


The hashing algorithm and the ordering of heterogeneous types changed
between 2.1 and 2.2.

HTH,
Anthony
-- 
Anthony Baxter     <anthony at interlink.com.au>   
It's never too late to have a happy childhood.





From markscottwright at hotmail.com  Mon Apr  7 20:37:07 2003
From: markscottwright at hotmail.com (msw)
Date: Tue, 08 Apr 2003 00:37:07 GMT
Subject: How to protect python code ?
In-Reply-To: <mailman.1049722991.18682.python-list@python.org>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org>
Message-ID: <3E921A4C.1020908@hotmail.com>

Dave Brueck wrote:
> On Mon, 7 Apr 2003, msw wrote:
> 
> 
>>Jp Calderone wrote:
>>
>>>On Thu, Apr 03, 2003 at 09:35:49AM +0200, christophe grimault wrote:
>>>
>>>>Before trying my own solutions, I wonder if someone already has a solution
>>>>to this. We have been programming a quite large application in python and
>>>>it contains many original algorithm and other things. The application will
>>>>be sent to customers in a few month and since it is a commercial app, we
>>>>need some protection against copy (illegal copies) and we need to make
>>>>sure that critical parts of the code cannot be viewed and studied.
>>>>
>>
>>I'm really interested in this as well, and I think that it's an
>>important issue for python.  Unfortunately, as I'm sure you've
>>discovered by now, it's a question that comp.lang.python is
>>constitionally incapable of answering.  The only answers you will get in
>>this group are:
>>
>>1. intellectual property law is sufficient
>>2. no protection scheme is perfect
> 
> 
> Not true - check the archives. In the past people, including myself, have
> posted the general steps to do to "protect" the code. It's not worth the
> effort in most cases, IMO, but answers *have* been given.

If so, I was unable to find it, and I think this thread proves my point. 
  I count half a dozen responses to my post, not one of which provided 
any detail into protecting source code.  However, they all go into great 
detail, at varying levels of insult, about how I shouldn't be asking the 
question in the first place.  I have consulted the archives and the 
signal/noise ratio is similar.  No offense, but this post is a case in 
point (although more respectful than most) - a few vague sentences about 
how to accomplish this followed by several paragraphs implying that I 
don't understand the product requirements.

The software I write falls under the category of those that need to have 
the source code hidden from customers - for many reasons, not least of 
which is "because the marketing department says so".  Our customers are 
unlikely to go after "movl $0x0007, %ax" with a hex editor, but are 
almost certain to mess with code that looks like "licenseCount = 5".

My frustration with this question is that it seems to offend the c.l.p. 
readership to such a degress that they collectively decide to drown out 
any meaningful discussion on the matter.

>>I think it ought to be relatively easy to distribute a hacked version of
>>the python interpreter with your application that substitutes the
>>standard "import" code with code that first looks for .pys files (or,
>>what the hell, .tcl - let's not make this any easier) that are
>>des-encrypted .py files.
> 
> 
> It *is* relatively easy to do this - there are various ways to overload
> imports, import from libraries, add new extensions and import handlers,
> and you don't need a "hacked" interpreter - the simplest example of
> running Python from C code will do.

That was what I was suggesting - distributing myapp.exe along with 
python.dll, where myapp.exe installed a different __import__ that looked 
for files with a particular extention that were scrambled or encrypted. 
   Which then raises the question of how I determine dependencies for 
setup.py, etc.

msw




From amk at amk.ca  Wed Apr 23 13:09:37 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Wed, 23 Apr 2003 12:09:37 -0500
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com> <e9570f37.0304220625.11312687@posting.google.com>
Message-ID: <K2qdnXJ_ZeZMVTujXTWcqg@speakeasy.net>

On 22 Apr 2003 07:25:54 -0700, 
	Graham Fawcett <graham__fawcett at hotmail.com> wrote:
> I guess the moral of the story is, "contribute some time to
> PyPI/Distutils development."  Or, befriend a PyPI/Distutils developer
> and send him food and get-out-of-work-free tickets. ;-)

Contribute time, not money.  Money does not make interested developers
materialize, and money doesn't help figure out what to do.  The great
problem is that no one is particularly interested in improving Python's
packaging, not even to the extent of offering opinions on concrete proposals
or trying out code.

(Things that might help: join the Catalog-SIG and try a system that someone
else has built.  Comment on existing proposals like PEPs 262 and 314. 
Review the existing code and offer patches.)

--amk                                                    (www.amk.ca)
TAMORA: O cruel, irreligious piety!
      -- _Titus Andronicus_, I, i



From aahz at pythoncraft.com  Sat Apr 26 00:55:12 2003
From: aahz at pythoncraft.com (Aahz)
Date: 26 Apr 2003 00:55:12 -0400
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com>
Message-ID: <b8d3fg$i3b$1@panix3.panix.com>

In article <3EA97D61.3CC9FE74 at easystreet.com>,  <achrist at easystreet.com> wrote:
>
>Any suggestions of the shortest path from a Python program to RTF
>output?  RTF seems to be the best (reasonable combination of features
>and compatability) common denominator for exporting data so that
>(windows) users can bring it into their own word processing program,
>whatever that may be.  (Any problem with that?)

Another idea: how about PDF?  I don't know if the formatting gets
preserved at all if you cut'n'paste, though.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From jepler at unpythonic.net  Thu Apr 17 15:24:00 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Thu, 17 Apr 2003 14:24:00 -0500
Subject: "pythonic"
In-Reply-To: <qait9vsttagq0c8gk5f40ou991bitistfe@4ax.com>
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com> <qait9vsttagq0c8gk5f40ou991bitistfe@4ax.com>
Message-ID: <20030417192400.GE30789@unpythonic.net>

On Thu, Apr 17, 2003 at 03:39:44PM +0000, Courageous wrote:
> 
> >I do not know if this ignorance, laziness, or a meme, but I would like
> >to know if there a real definition of "pythonic". Anyone?
> 
> First you have to understand the word "unpythonic".
> 
> This is kind of like the word "communist". You know? During the
> 50's, your life was over if you were labeled a "communist"? 
> Guido, Tim, and team sit on a the Technical Committee for Un[ythonic
> Activities.
> 
> Be afraid, be very afraid.

uh oh.

Jeff Epler <jepler at unpythonic.net>




From sholden at holdenweb.com  Wed Apr 16 09:17:49 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 13:17:49 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <mailman.1050438685.19213.python-list@python.org> <sggp9v8lpk0045klmquc8gec1ju7q4kmpe@4ax.com> <slrnb9pt1o.iin.ramen@lackingtalent.com>
Message-ID: <1Kcna.110587$It5.3946@news2.central.cox.net>

"Dave Benjamin" <ramen at lackingtalent.com> wrote in message
news:slrnb9pt1o.iin.ramen at lackingtalent.com...
> In article <sggp9v8lpk0045klmquc8gec1ju7q4kmpe at 4ax.com>, Courageous wrote:
> > Anonymous functions are occasionally useful, as in Java's anonymous
> > class expressions, really used as a poor man's lambda of sorts. I've
> > seen this overused, though, that's for sure.
>
> How frequently a feature is useful depends on the problem domain. Java's
> anonymous classes are a poor substitute because they carry so much
baggage.
> For comparison, here's an implementation and usage of map in Java:
>
> static interface Functor {Object call(Object[] args);}
>
> static Object[] map(Functor func, Object[] array) {
>     Object[] result = new Object[array.length];
>     for (int i = 0; i < array.length; i++) {
>         result[i] = func.call(new Object[] {array[i]});
>     }
>     return result;
> }
>
> public static void main(String args[]) {
>     System.out.println(Arrays.asList(map(
>         new Functor() {
>             public Object call(Object[] args) {
>                 return args[0] + ", world!";
>             }
>         }, new String[] {"Hello", "Goodbye"}
>     )));
> }
>
> And the same usage in Python:
>
> >>> print map(lambda x: x + ', world!', ['Hello', 'Goodbye'])
> ['Hello, world!', 'Goodbye, world!']
>
> This is why Java is IMHO unfit for common functional tasks where Python
> shines with elegance. As to whether or not anonymous functions can be
> overused, I think any feature can be overused. Classes, for one. In fact,
in
> my day to day work, I find the two of similar usefulness, but maybe it's
> just because of the way I program. =)
>

I *still* say that Java is object-oriented COBOL. People tend to take
offence when I say it, but that example makes the point well. There's so
much crud in the typical Java application, most of which could probably be
macro-generated from a much more concise "super-source".

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From duncan-news at grisby.org  Mon Apr  7 15:32:37 2003
From: duncan-news at grisby.org (Duncan Grisby)
Date: Mon, 07 Apr 2003 19:32:37 +0000
Subject: Pictures from the Python UK conference
Message-ID: <5f0b6$3e91d255$516049d2$2705@nf1.news-service.com>

Hi,

I took lots of digital pictures at the Python UK conference. In case
you're interested, you can see them here:

  http://www.grisby.org/Photos/128/

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --



From srijit at yahoo.com  Sun Apr 13 14:58:36 2003
From: srijit at yahoo.com (srijit at yahoo.com)
Date: 13 Apr 2003 11:58:36 -0700
Subject: Protected Methods and Python
Message-ID: <221d8dbe.0304131058.1ecd1c3c@posting.google.com>

I do not see a  concept of Protected Methods in Python 2.2 like C++ and Ruby.
Is this a serious limitation for Object Oriented Programing with Python? 
Or may be it is possible to emulated Protected Methods in Python.

I look forward to some discussion by OOP experts.

Regards,
Srijit



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Sun Apr 27 09:05:29 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Sun, 27 Apr 2003 15:05:29 +0200
Subject: problems using magic "<broadcast>" socket address on different platforms
Message-ID: <3eabd599$0$49102$e4fe514c@news.xs4all.nl>

Hello
I'm trying to use UDP broadcasting but there are some weird things going
on regarding Python's magic "<broadcast>" address (which is
the INADDR_BROADCAST address).


Using the following piece of Python code to send UDP broadcast messages:

----
from socket import *

sock = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP)
sock.setsockopt(SOL_SOCKET, SO_BROADCAST, 1)
sock.sendto("Het Bericht", ("<broadcast>", 2000) )
print "done"
----

This works fine. The problems are in the server code:

----
from socket import *
import select
import sys

sock = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP)
sock.setsockopt(SOL_SOCKET, SO_BROADCAST, 1)
sock.bind( ("<broadcast>", 2000) )

while 1:
	(rs,ws,es)=select.select([sock],[],[],1)
	if sock in rs:
		(data, addr) = sock.recvfrom(9999)
		print "Got data: {%s}" % data
		print "from:",addr
	else:
		print ".",
		sys.stdout.flush()
----		

When running this on Linux, everything works as expected.
When running the server code on Windows, Python fails to bind
the socket on the magic "<broadcast>" address:
socket.error: (10049, "Can't assign requested address")

I've had reports of users that the same problem is present on Mac OS X.

Because I don't have access to other platforms than Windows (NT) and Linux,
I am wondering what the results are on other systems such as Solaris, *BSD...

--Irmen de Jong

PS 
The really weird part is that a piece of C code that directly uses the
BSD socket API *also* fails on windows with the same error
(bind: Cannot assign requested address, when trying to bind the socket
on the INADDR_BROADCAST address). The exact same piece of C code runs fine
on Linux. Used GCC in both cases. If you're interested in this code let me know,
I thought it was too big to post here.

PPS
To get my code to work on windows, I had to use the magic "" address
(INADDR_ANY). But I don't think that's right, "<broadcast>" should work,
shouldn't it?




From abpillai at lycos.com  Mon Apr 14 05:48:20 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 14 Apr 2003 02:48:20 -0700
Subject: Python Jobs
Message-ID: <bd993a2f.0304140148.36298ef3@posting.google.com>

Can somebody give me a python job ? I am quite bored of my current job
and I love python programming. I need a change desperately but I am
not
getting it. I am ready to relocate out of my country to U.S/U.K for
the
job. 

If somebody thinks he has a job, my profile is avaialable to them for
review. Mail me at this email address. Or visit my webpage
http://members.fortunecity.com/anandpillai.

I dont want to write further, but I am quite desperate for a shift
from C/C++.

Thank You.

Anand



From cswiger at mac.com  Thu Apr 10 20:32:56 2003
From: cswiger at mac.com (Chuck Swiger)
Date: Thu, 10 Apr 2003 20:32:56 -0400
Subject: popen(), sendmail: Success?
In-Reply-To: <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de>
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de>
Message-ID: <3E960D38.6020601@mac.com>

In article <3e95d63f.2635484 at news.muenster.de>, Martin Bless wrote:
> To send mails from Python here's the example my provider provides:
> This http://faq.puretec.de/skripte/python/3.html
>
> The relevant lines are these:
>
> sendmail = os.popen("/usr/lib/sendmail -t", "w")
> sendmail.write(msg)
> sendmail.close()
>
> This works just fine. But is it enough? How do I know the operation
> succeeded?

If you plan on sending any significant amount of mail, it is recommended 
  that your code talk SMTP to port 25 and pipeline all of the messages 
you need to send in one continuous stream.  If you fork a sendmail 
subprocess for each message you send, that will create a wastefully huge 
amount of system load.

This has a nice side-effect: your code becomes more portable since you 
can point to a SMTP server anywhere that is willing to relay your 
messages, not just one on localhost.  Nor does the local host have to 
have /usr/lib/sendmail, meaning your code will run under Windows or 
elsewhere non-Unixlike.

With regard to error handling, sendmail will generate an error return 
value; adding the "-v" flag to your sendmail invocation will show the 
details of the SMTP or LMTP transaction if you want to see what's going 
on for debugging purposes.  I'd assume there is an smtp module that 
would let you see the success or error code and messages from doing 
SMTP, as well.

Good luck,
-Chuck

PS: Note that handing the mail off does not guarantee that the email 
will actually be read by the intended receipient: if that is part of 
your meaning of "the operation succeeded", you might want to generate 
messages which request the user to mail a password/token back or click 
an URL to confirm that the email went through.  See the Mailman mailing 
list program (www.list.org) for examples of "good email handling" practices.






From godoy at metalab.unc.edu  Tue Apr  8 18:01:13 2003
From: godoy at metalab.unc.edu (Jorge Godoy)
Date: Tue, 08 Apr 2003 19:01:13 -0300
Subject: Best way to make a weighted sum
References: <867ka46awj.fsf@ieee.org> <mailman.1049834294.20008.python-list@python.org>
Message-ID: <86r88c3aza.fsf@ieee.org>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> writes:

> Jorge Godoy <godoy at metalab.unc.edu> wrote previously:
> |    def CalculaSoma(self):
> |        ie = self.ie
> |        iter = len(ie)
> |        for i in ie:
> |            produto = i * self.pesos[iter]
> |            self.soma = self.soma + produto
> |            iter = iter - 1
> |I don't know if I got it in the most pythonic way in my code.
>
> Looks fine.  You could be somewhat more concise with:
>
>     from operator import mul, add
>     amts, weights = pesos.keys(), pesos.items()
>     sum = reduce(add, map(mul, amts, weights))/reduce(add, weights)
>
> I'll leave it to a native speaker to translate my var names to
> Portuguese.
>
> Yours, Lulu...

Thanks.


This is also the kind of thing I was looking for. The reduce() and
zip() functions are very interesting. And the operator module was
already being used in other parts of the code, so why not here, in
this class? :-) 

-- 
Godoy.      <godoy at metalab.unc.edu>

	"... The name of the song is called 'Haddocks' Eyes'!"
	"Oh, that's the name of the song, is it?" Alice said, trying to
feel interested.
	"No, you don't understand," the Knight said, looking a little
vexed.  "That's what the name is called.  The name really is, 'The Aged
Aged Man.'"
	"Then I ought to have said "That's what the song is called'?"
Alice corrected herself.
	"No, you oughtn't:  that's quite another thing!  The song is
called 'Ways and Means':  but that's only what it is called you know!"
	"Well, what is the song then?" said Alice, who was by this time
completely bewildered.
	"I was coming to that," the Knight said.  "The song really is
"A-sitting on a Gate":  and the tune's my own invention."
		-- Lewis Carroll, "Through the Looking Glass"



From aleax at aleax.it  Tue Apr 29 08:11:21 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 12:11:21 GMT
Subject: Determining the volume ID for a file/drive?
References: <uv7savkms6lsgchcv6jh8ahtccu9lon3oi@4ax.com>
Message-ID: <JZtra.19431$3M4.563286@news1.tin.it>

carroll at tjc.com wrote:

> Is there any way to determine the volume serial number for a disk,
> given the filename (or drive, since I can determine the drive from the
> filename with os.path.splitdrive(os.path.abspath(file))?
> 
> My environment is Windows & Python 2.2.2, and I don't need for my code
> to be portable.

Quick and dirty...:

import os
x = os.popen('C:\\')
x.readline(); x.readline()
vsn = x.readline().strip()[-9:]
x.close()


Perhaps too dirty.  If you want to do it via a Win32 API call, you
need to determine what call is it (e.g. via MSDN) and then use
win32all or ctypes to go and call it.  But parsing dir's output
is quite a bit quicker to program;-).


Alex





From vjb at vuit.com  Tue Apr 29 11:47:47 2003
From: vjb at vuit.com (Vince Buonassisi)
Date: Tue, 29 Apr 2003 08:47:47 -0700
Subject: import problem
Message-ID: <ED1CFDBD-7A59-11D7-B1FA-000A95682CD2@vuit.com>

hi,

i am having problems importing a module in my python script.  i have 
the following two import statements in the script:

from WVWService.util import XMLUtil
from WVWService.test import HTTPClient

the first import works just fine.  but, the second one i get the error 
"ImportError: No module named test".  i have checked for mispellings;  
i have checked to see if the directory test exists; i have checked the 
system path by printing out sys.path from w/i the script to see if the 
directory path is in sys.path (/home/vjb/workspace/wvwBase3/src/wvw).  
i don't understand why the first import works and the second one does 
not since the directories util and test reside in the same directory 
path.

i am using python 2.2.2 on redhat 8.0.

any help would be appreciated.

thanks,

vince





From hllee678 at hotmail.com  Mon Apr 21 08:03:08 2003
From: hllee678 at hotmail.com (=?big5?B?p/UgvuWswg==?=)
Date: Mon, 21 Apr 2003 20:03:08 +0800
Subject: (no subject)
Message-ID: <Sea2-F35u96FxpJaqTT0000c778@hotmail.com>

Hi,
I have a problem about M2Crypto.
I am using M2Crypto to write the SSL server and client.
My server was based on Medusa.
The client built a SSL connection to server,
and then created a new thread to receive the response from server.

The code of client:
rfile=sock.makefile('rb',-1) #sock is the ssl connection between server and 
client
try: 
    data = rfile.read(100)
    if not data:
        print "no data, maybe try to close connection"   
except:
    print "exception, maybe try to close connection"   

When the server shutdown and close the SSL connection between server and 
client,
the client receiving thread raised a windows exception on rfile.read 
procedure:
Unhandled exception in python.exe (PYTHON21.DLL): 0xC0000005: Access 
Violation

The code of server:
sock.shutdown(2)         # socket is the ssl connection between server and 
client
sock.close() 

Please help me if you know how to solve this problem.
Thank you very much!


_________________________________________________________________
??????? MSN Messenger????????????????????????
???? http://messenger.msn.com.tw/ 





From spam at magnetic-ink.dk  Tue Apr  1 03:39:43 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Tue, 1 Apr 2003 08:39:43 +0000 (UTC)
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com>
Message-ID: <d2c6a353-612a-4f45-8ddd-bddc04d59e6a@news.szn.dk>

Chris wrote:

> I would like to request that Python support multi-line
> comments (e.g. /* */ in C).

You could always enclose comments in triple-quotes:

#v+

comment = """
Sed ut perspiciatis, unde omnis iste natus error sit voluptatem
accusantium doloremque laudantium, totam rem aperiam eaque ipsa,
qu? ab illo inventore veritatis et quasi architecto beat? vit?
dicta sunt, explicabo.  Nemo enim ipsam voluptatem, quia voluptas
sit, aspernatur aut odit aut fugit, sed quia consequuntur magni
dolores eos, qui ratione voluptatem sequi nesciunt, neque porro
quisquam est, qui dolorem ipsum, quia dolor sit, amet, consectetur,
adipisci velit, sed quia non numquam eius modi tempora incidunt,
ut labore et dolore magnam aliquam qua?rat voluptatem.  Ut enim
ad minima veniam, quis nostrum exercitationem ullam corporis
suscipit laboriosam, nisi ut aliquid ex ea commodi consequatur?
Quis autem vel eum iure reprehenderit, qui in ea voluptate velit
esse, quam nihil molesti? consequatur, vel illum, qui dolorem
eum fugiat, quo voluptas nulla pariatur?
"""

#v-


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From yaipa at yahoo.com  Sat Apr 19 14:48:35 2003
From: yaipa at yahoo.com (yaipa h.)
Date: 19 Apr 2003 11:48:35 -0700
Subject: Parent references: hot or not?
References: <mailman.1050762692.5175.python-list@python.org>
Message-ID: <6e07b825.0304191048.3fe5976d@posting.google.com>

Dylan,

Might be worth the look...

 Multiple dispatch  
  . Generalizing polymorphism with multimethods

 by David Mertz, Ph.D. (mertz at gnosis.cx)

 link,
    http://www-106.ibm.com/developerworks/linux/library/l-pydisp.html

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dylan Reinhardt <python at dylanreinhardt.com> wrote in message news:<mailman.1050762692.5175.python-list at python.org>...
> Hi all,
> 
> I'm working on a project where I have multiple instances of a container
> class that each hold a large number of another class.
> 
> I need the "contained" object to be able to make use of methods and
> attributes in its container. Inheritance won't do the trick since the
> things I need to refer to will be pretty dynamic.
> 
> A very minimal example of what I'm doing:
> 
> ---------------
> 
> class Container:
>    def __init__(self):
>       self.items = {}
> 
>    def add_item(self, item_name):
>       self.items[item_name] = Item(self)
> 
>    some_attr = 'spam'
> 
> 
> class Item:
>    def __init__(self, parent):
>       self.parent = parent
> 
>    def some_method(self):
>       return self.parent.some_attr
> 
> ---------------
> 
> This seems to be working well, but I want to check my assumptions.
> 
> If I'm understanding correctly:
> 
> 1. Any Item's self.parent is only a *reference* to its parent.
> 2. Holding a reference to a parent isn't some kind of recursive paradox.
> 3. As I add Items, Container size will grow linearly, not exponentially
> 4. It will be impossible to fully delete a non-empty Container.
> 
> I think those are valid assumptions, but would appreciate any
> confirmation, contradiction, or insight anyone cares to offer.
> 
> Also, if there's some other consideration or reason why holding a
> reference to one's parent is a Bad Idea, I'd love to hear about it.
> 
> Thanks in advance,
> 
> Dylan



From Bill.Scherer at verizonwireless.com  Tue Apr  8 13:18:02 2003
From: Bill.Scherer at verizonwireless.com (Scherer, Bill)
Date: Tue, 8 Apr 2003 13:18:02 -0400 (EDT)
Subject: ANN: SQLObject 0.3
In-Reply-To: <1049822101.734.40.camel@lothlorien>
Message-ID: <Pine.LNX.4.44.0304081317440.23210-100000@bagend.odc.vzwcorp.com>

On 8 Apr 2003, Ian Bicking wrote:

> On Tue, 2003-04-08 at 09:13, Scherer, Bill wrote:
> > On 7 Apr 2003, Ian Bicking wrote:
> > 
> > > SQLObject 0.3 has been release:
> > >   http://sqlobject.org
> > 
> > Is support for Oracle planned?
> 
> Sure, someone just has to add it (I don't have Oracle).  It should be
> quite easy.

Cool. I'll give it a go and let you know.

> 
>   Ian
> 

-- 
Bill.Scherer at Verizon Wireless
RHCE 807101044903581





From marcus at infa.abo.fi  Mon Apr 14 16:27:48 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 14 Apr 2003 22:27:48 +0200
Subject: Complexity of standard Python data structures
References: <mailman.1050343651.17799.python-list@python.org>
Message-ID: <slrnb9m2qj.9r2.marcus@infa.abo.fi>

On Mon, 14 Apr 2003 14:04:57 -0400, Tim Peters <tim.one at comcast.net> wrote:
>If you want to clear a list,
>
>   del list[:]
>
>is the usual way to spell it;
>
>   list[:] = []

OK, that has slipped past me, thanks for clarifying. And it'll save
the bandwidth of one mail :)

>Note that it's not true that Python's list is really a vector/array -- the
>language doesn't define its implementation, and doesn't want to.  It is true

Hmm. So the Python language is more about specifying _what_ will
happen, instead of _how_, with no concern for actual speed or memory?

Marcus



From peter at engcorp.com  Mon Apr 21 07:52:11 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 21 Apr 2003 07:52:11 -0400
Subject: Mail and text extraction
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com> <3EA34DAE.EC67C1E3@engcorp.com> <pan.2003.04.21.13.28.15.303279.2294@mindlessSPAM.com>
Message-ID: <3EA3DB6B.DAC620BD@engcorp.com>

Dino Levy wrote:
> 
> On Mon, 21 Apr 2003 03:47:26 +0200, Peter Hansen wrote:
> 
> 
> > Why not download the whole message, then scan it after downloading? You
> > provide no good reason, so far, for why you wouldn't just do this much
> > simpler step.  Keep the design simple, modular, clean, and you can do
> > most of what you want with simple, modular, existing solutions in
> > standard modules.  Why wouldn't you do that?
> >
> > -Peter
> 
> Well yes, that's antoher way of doing it, but I wanted to make it done in
> one breath, like connecting and immediately parsing the text, but it's
> obviously not achievable:)

Oh, it's quite achievable that way.  It's just the wrong way to do it.

Mixing in the various operations like that is a form of optimization,
and since nothing's actually working yet it's clearly premature, and
therefore not good. :-)

> I don't want MIME, I really need POP3 connections...
> Ok, thanks anyway

MIME and POP3 are orthogonal.  MIME is an encoding, POP3 is a
transport protocol (or something).  You can have one, or the other,
or both, or neither...

By the way, the "quite achievable" part above simply means you could
easily build this capability into custom software.  The "wrong way"
just means that would likely be a large waste of time, given that 
the individual pieces already exist in standard modules, quite well
tested and debugged.  Just combine them in appropriate ways and keep
life simple.

-Peter



From fredrik at pythonware.com  Fri Apr 25 13:51:41 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 25 Apr 2003 19:51:41 +0200
Subject: What means exactly "Memory error"?
References: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net> <mailman.1051198958.16669.python-list@python.org> <slrnbai52u.8cv.hat@se-46.wpa.wtb.tue.nl>
Message-ID: <b8bshn$d00$1@main.gmane.org>

Albert Hofkamp wrote:

> Do you need all data DOMmed at once?
> You may be able to have one DOM tree at a time, dropping and reloading
> everytime you switch file.

an alternative is to use an incremental tree builder, and process the
subtrees as they arrive.

here's an example, using my elementtree module:

    from elementtree import ElementTree

    class MyBuilder(ElementTree.TreeBuilder):

        def end(self, tag):
            elem = ElementTree.TreeBuilder.end(self, tag)
            if elem.tag == "SCENE":
                # process(elem) in some way, and write it out
                # ElementTree.ElementTree(elem).write(sys.output)
                elem.clear() # nuke it

    parser = ElementTree.XMLTreeBuilder()
    parser._target = MyBuilder() # plug in a custom builder!

    tree = ElementTree.parse(filename, parser)

I've tested this with a 10 megabyte XML file created by concatenating
Jon Bosak's Hamlet XML file over and over again, and wrapping it all in a
single document element.

the resulting file contains 720 scenes (about 15k each, in average).

the above script requires about 4.5 megabytes to run to completion, and
about 2 minutes processing time (on a really slow machine).

if I comment out the elem.clear() call, the script requires about 75 mega-
bytes, in about 15 minutes (13 of which were spent on swapping; I ran the
test on a machine with 96 megabytes RAM and really slow disks... ;-)

for more information on element trees, see:

    http://effbot.org/zone/element-index.htm
    http://www.xml.com/pub/a/2003/02/12/py-xml.html

</F>







From cpitaar at yahoo.com.ar  Wed Apr 30 08:23:27 2003
From: cpitaar at yahoo.com.ar (Carlos P.)
Date: 30 Apr 2003 05:23:27 -0700
Subject: Raising objects
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it>
Message-ID: <453d10b2.0304300423.224abe55@posting.google.com>

> restriction will be removed -- in particular, because I do not know
> of "use cases" where the ability to raise new-style classes or instances
> thereof would be a major boon to an application.  Maybe you have some...?

In fact, I have no one. I've just written a nice metaclass Metaclass
which automatically generates instance/static properties/methods from
getters/setter/methods according to naming conventions, and naive
stuff like that. I intended to use class Object (a new style class
with metaclass Metaclass) as a base class for my application. The
point of using it with exceptions is only to be consistent and, before
your answer, it was natural to try to derive an exception from
ObjectException (a subclass of Exception and Object). Nothing sooo
important, just would be nice if possible.
Thank you.



From jidanni at dman.ddts.net  Tue Apr 15 01:59:54 2003
From: jidanni at dman.ddts.net (Dan Jacobson)
Date: Tue, 15 Apr 2003 13:59:54 +0800
Subject: do python's nifty indentation rules spell the death of one-liners?
Message-ID: <87he90b8rp.fsf@jidanni.org>

I need to use one liners every day, especially in Makefiles.  I thus
need to know how to make any kind of python program into a big long one
liner. However, right out of the starting gates I run into
$ python -c 'print 2;for i in (1,4):print i'
  File "<string>", line 1
    print 2;for i in (1,4):print i
              ^
SyntaxError: invalid syntax

Help. BTW tell me what to do in all the other cases when ";" can't be
a replacement for newline.  I.e. does python's nifty indentation rules
spell the death of one-liners?

BTW, how do I import a module from the command line when using -c? Is
there an option switch for importing modules?

This makes no output and returns no error value to the shell:
$ python -c 'import math' -c 'for i in range(200,500,50): print
  i,360/2/math.pi*math.asin(i/1238.4)'

If I found a bug, report it for me.
-- 
http://jidanni.org/ Taiwan(04)25854780



From dsavitsk at e-coli.net  Fri Apr 25 18:18:03 2003
From: dsavitsk at e-coli.net (dsavitsk)
Date: Fri, 25 Apr 2003 22:18:03 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com>
Message-ID: <vuiqa.3522$%_3.2810696@newssrv26.news.prodigy.com>

<achrist at easystreet.com> wrote in message
news:3EA97D61.3CC9FE74 at easystreet.com...
> Any suggestions of the shortest path from a Python program to RTF
> output?  RTF seems to be the best (reasonable combination of
> features and compatability) common denominator for exporting
> data so that (windows) users can bring it into their own word
> processing program, whatever that may be.  (Any problem with
> that?)
>
> Requirements are about like this:
>
> 1. Don't need an edit control, just want to write a file, but if
> I have to open an edit window, I can.
>
> 2. Don't want to do too much fancy.  Text, headings, bulleted
> lists,  maybe some variations in font size.
>
> 3. Development platform is WinNT, target platform is any reasonably
> recent Windows machine (eg either anything newer than Win95 release 1
> or Win98 or better).

Your requirements suggest to me that html is the best common denominator,
and it's easier to write.

-d





From andrew-pythonlist at puzzling.org  Sun Apr 27 04:36:16 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 18:36:16 +1000
Subject: Forgetting "()" when calling methods
In-Reply-To: <U0Mqa.13860$K35.457281@news2.tin.it>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <U0Mqa.13860$K35.457281@news2.tin.it>
Message-ID: <20030427083616.GP26530@frobozz.local>

On Sun, Apr 27, 2003 at 07:54:28AM +0000, Alex Martelli wrote:
> Erik Max Francis wrote:
> 
> > Jeremy Fincher wrote:
> > 
> >> Perhaps the answer is to remove the __nonzero__ method on
> >> functions/methods.  Logically, what purpose does it serve?
> > 
> > # f holds a callable object or None
> > if f:
> > f(...)
> > 
> 
> I think "if f is None:" and even more "if callable(f):" are
> better ways to express this test -- explicit is better than
> implicit.  However, if this (inferior) expression is indeed

Taken to extremes, "explicit is better than implicit" suggests we should be
using a statically typed language.... ;)

Fortunately, The Zen of Python has plenty of mildly contradictory
statements, which helps in defending one's viewpoint :)

-Andrew.





From max at alcyone.com  Fri Apr 25 05:57:16 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 25 Apr 2003 02:57:16 -0700
Subject: Char to ASCII and back ?
References: <roy-CF32FA.11400525042003@malibu.unice.fr>
Message-ID: <3EA9067C.199071E0@alcyone.com>

Jean-Paul Roy wrote:

> How do you transform a char (or a one-char string) into the integer
> ascii code and back ? Say to code a message.

You want ord and chr, respectively:

>>> ord('A')
65
>>> chr(65)
'A'

> The string type is not mutable in Python. Is there something like the
> StringBuffer class in Python whose instances would be mutable ?

You almost certainly want a variant of the StringIO class, either in the
StringIO (Python-only) or cStringIO (optimized C):

>>> import StringIO
>>> output = StringIO.StringIO()
>>> output.write("Hello, world.\n")
>>> output.getvalue()
'Hello, world.\n'
>>> input = StringIO.StringIO("Goodbye, world.\n")
>>> input.read()
'Goodbye, world.\n'

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The actor is not quite a human being -- but then, who is?
\__/ George Sanders
    Bosskey.net: Unreal Tournament 2003 / http://www.bosskey.net/ut2k3/
 A personal guide to Unreal Tournament 2003.



From jkraska at san.rr.com  Tue Apr 15 22:55:46 2003
From: jkraska at san.rr.com (Courageous)
Date: Wed, 16 Apr 2003 02:55:46 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <roy-06B012.20145112042003@reader1.panix.com> <bJ2ma.5447$883.4511@nwrdny02.gnilink.net> <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net> <b7ig7d$ale$1@panix1.panix.com>
Message-ID: <h8hp9v8ojhhn85uc407b4flo708789qluv@4ax.com>

>>But is that /guaranteed/ (as it is with STL), or does it just happen to
>>be that way with the current implementation (however unlikely it is to
>>change)?

>Both, sort of.  Python uses dicts internally; you can rely on any
>changes to achieve roughly equivalent or better performance guarantees.

Yeah, but dictionaries are only /guaranteed/ to scale O(N). 'Course,
I recall that a fairly good mathematician or two seems to think that
amortized times are fine; these /tend/ to scale O(1). It doesn't hurt
at all that Python's dictionary implementation has been tuned and
tested more than just about any dictionary implementation in any
production environment anywhere*. :)

My interpretation of his question makes me wonder how he came to believe
that Guido and Tim are on a standards committee, dictating how the host
of various and sundry Python compiler developers will design Python
implementations. Or perhaps he fears that someone will steal Python from
us and do something evil, like replace dict with a red-black tree. :-)

performed-an-extensive-analysis-twenty-years-from-now-and-mailed-
it-backwards-through-time-ly yrs*

C//




From gangli at msn.com  Tue Apr  8 11:54:05 2003
From: gangli at msn.com (gang li)
Date: 8 Apr 2003 08:54:05 -0700
Subject: [ANN?] Boston PIGs
References: <mailman.1049783173.32214.python-list@python.org>
Message-ID: <9dfc16ab.0304080754.459eace3@posting.google.com>

It is greate idea.  I am in Boston area, and defintly like to attend a
 meeting once a month.

Jack Diederich <jack at performancedrivers.com> wrote in message news:<mailman.1049783173.32214.python-list at python.org>...
> After six months of subscribing to c.l.py I haven't seen any activity that
> specifically targets Boston.  As implied, that is where I am.   If there was a 
> caucus I would attend a monthly Python Interest Group.  Heck, if arstechnica 
> can pull fifty people a couple times a year it can't be all that hard.
> 
> Would anyone be interested in attending a regular Boston Python meeting?
> 
> Possible names abound,
>   BAPIG (Boston Area Python Interest Group)
>   BPIGs (Boston Python Interest Group)s
>   BPUG (feh)
>   BeePee (phonetical)
> 
> -jack



From peter at engcorp.com  Tue Apr  1 08:47:40 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 08:47:40 -0500
Subject: spawning a separate process
References: <b6c2c6$lr7$1@aspen.sucs.soton.ac.uk>
Message-ID: <3E89987C.5B133AED@engcorp.com>

Mark Light wrote:
> 
>        I have a python program from which I wish to spawn an os.system call
> (a bash script) that may take several minutes to run - whilst the rest of
> the script carries on. I have tried using bits of code from books - but it
> never seem to work quite right

In addition to P?draig's response, here's another possibility:

import threading
class RunInBackground(threading.Thread):
    def __init__(self, command):
        threading.Thread.__init__(self)
        self.command = command
        self.start()
        self.done = False

    def run(self):
        import os
        os.system(self.command)
        self.done = True

Then do something like:

bgThread = RunInBackground('/bin/bash mycommand')

Using the popen approach is no doubt better if you don't have
"special needs", but this might make it easier for you to 
extend the functionality if you decide you need more than what
you originally asked for.

-Peter



From staschuk at telusplanet.net  Wed Apr 16 16:50:11 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 14:50:11 -0600
Subject: A better unittest
In-Reply-To: <1EadnQb6ZY_RDwCjXTWcog@comcast.com>; from tjreedy@udel.edu on Wed, Apr 16, 2003 at 01:36:03PM -0400
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com>
Message-ID: <20030416145011.E4266@tibia.amotlpaa.bogus>

Quoth Terry Reedy:
  [...]
> I like having the actual values too.  It should make debugging an
> unexpected failure easier.  But what if each value is more that a few
> bytes?  Example (from my future plans): two lists of all permutations
> of range(8), one produced by an obvious recursion and the other by a
> much less obvious iteration.

Now and then I replace
    self.assertEquals(longlist1, longlist2)
with
    map(self.assertEquals, longlist1, longlist2)

It's not quite ideal; in some cases it would be nice to see the
index of the mismatch too, for example.

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From aleax at aleax.it  Wed Apr 23 18:07:26 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 22:07:26 GMT
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org> <GXupa.83382$T34.2079161@news2.tin.it> <5jadeg285s.fsf@xoxy.net>
Message-ID: <y8Epa.51043$LB6.1153097@news1.tin.it>

Charlie Reiman wrote:

> Alex Martelli <aleax at aleax.it> writes:
> 
>> <posted & mailed>
>> > 
>> For your algorithm, why not just, for example:
>> 
>> class Infimum:
>>     def __cmp__(self, other): return -1
>> infimum = Infimum()
>> 
>> class Supremum:
>>     def __cmp__(self, other): return 1
>> supremum = Supremum()
>> 
>> as coded these have obvious anomalies, e.g. infimum < infimum -- but
>> those are easy to remove if they give problems to your algorithm.
> 
> 
> Are there any dangers in something like this:
> 
> a = [1, infimum, infimum]
> a.sort()

Quite possibly -- so IF your algorithms sorts lists, then the "easy
to remove" anomalies should be removed.

> Should the __cmp__ methods return 0 for equality (identity, in this
> case)?

If you need sorting, yes -- "return (other is not self) and -1" will
suffice.  But, should you never need to compare infimum to itself,
then the precaution, cheap as it may be, won't pay for itself.  You
pays your money and you takes your choices.


Alex




From syver-en+usenet at online.no  Sun Apr  6 08:35:22 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 06 Apr 2003 14:35:22 +0200
Subject: I'm new to programming
References: <oprm4ogjb9acxezr@news.datafast.net.au>
Message-ID: <u1y0f7qid.fsf@online.no>

Casey Jones <cajones at datafast.net.au> writes:

> Inspiration would be appreciated.

video games, sound or graphics? Try pygame
http://www.pygame.org



-- 

Vennlig hilsen 

Syver Enstad



From exarkun at intarweb.us  Tue Apr  8 01:12:55 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 8 Apr 2003 01:12:55 -0400
Subject: Cross-reference 'import' in a class hierarchy
In-Reply-To: <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <20030408051255.GA1934@meson.dyndns.org>

On Tue, Apr 08, 2003 at 04:55:09AM +0000, Ben Finney wrote:
> On Tue, 08 Apr 2003 15:44:10 +1200, Greg Ewing (using news.cis.dfn.de) wrote:
> > It's been suggested before (and I happen to agree) that Python should
> > provide a general way of making *relative* references to modules or
> > packages.
> 
> I continue to be amazed that there's simply no way to say
> 
>   import "../foo/ook.py"
> 
> as should be possible to implement any well-architected class hierarchy.

  It's simple.  Code should not depend directly on code higher than it in
the package hierarchy.  If it does (and it isn't one of a few extremely rare
cases), it is misdesigned.

  Code shallow in the hierarchy should manage required references to objects
required by code deep in the hierarchy by passing them around explicitly. 
Any other solution is just bunk.

  Jp

-- 
http://catandgirl.com/view.cgi?90
-- 
 up 19 days, 1:01, 3 users, load average: 0.00, 0.02, 0.00




From tim at bladerman.com  Fri Apr  4 07:53:57 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Fri, 04 Apr 2003 12:53:57 GMT
Subject: Interactive loop with InteractiveInterpreter
Message-ID: <Fffja.3972$Du.12917@newsc.telia.net>

I am trying to make an interactive interpreter for python source code.

It is all fun and games with Interactiveinterpreter.runcode() as long as it
is simple statements like

print "hello"
print "world"

Then I can feed it one line at the time and probably highlight the actuall
line in the editor but when it comes to a  loop I must feed runcode the
whole statement, the whole loop.

for i in list:
    print i

The above can be feed to runcode but I want to step trough the list and se
each printing operation not all at once.

Is it possible to step through a loop at all? With interactive interpreter?
I can't seem to find an easy solution, I guess InteractiveConsole won't do
it either. Is it maybe better to do it in a completley different way?


Any ideas or suggestions are welcome.

Thanks

Tim





From marcus at infa.abo.fi  Thu Apr 17 11:45:45 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 17 Apr 2003 17:45:45 +0200
Subject: Complexity of standard Python data structures (Re: Speed of string += string)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi> <Oqwma.11524$LB6.319489@news1.tin.it> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <2DFma.19063$T34.605833@news2.tin.it>
Message-ID: <slrnb9tfdn.lgn.marcus@infa.abo.fi>

On Mon, 14 Apr 2003 21:21:02 GMT, Alex Martelli <aleax at aleax.it> wrote:
>It basically seems to me that you're basing your expectations
>of Python docs, and those of C docs, on very different, unfair
>standards -- i.e., perhaps subconsciously, singling Python and
>its docs out for criticism by standards you don't apply to C...

ok, you are right, I was being unfair.

On the other hand, given the amount of questions regarding big-O
notation & string += string at the moment on c.l.p, perhaps there
should be a note -- a "warning" sounds a bit extreme?  Also in the docs
of C functions, of course!  Sometimes, unix man pages have usage
examples. Perhaps there should be "anti-usage" examples as well?  :)

>Wrong, actually.  E.g., what happens when you loop doing alist.append
>N times depends NOT just on alist being "a vector/array", but also,
>and crucially so!!!, on the overallocation strategy (size vs capacity 
>in std:: vector terms).  If you think that strategy is "obvious
>from that one single sentence" you're sorely mistaken -- as is shown
>by the fact that Python's strategy in that regard has historically
>changed reasonably recently (almost nobody noticed, except a few
>big-O purists who sleep easier at night knowing append IS now
>amortized O(1) in theory, too, not just in practice;-0).  And what

ok, that's correct as well, I generalized far too much. Curiously,
people do seem to start explaining the python list as e.g. "a C++
vector", or "a contiguous array of heterogeneous elements". I do
understand that Python people don't want to make any guarantees,
although describing some implementation might make it easier
for people to learn (or migrate to) Python.

>> - programming is about meeting requirements
>> - those requirements usually involve correctness to a certain
>> degree, _but_ also memory usage, speed, user interface functionality,
>> error handling, modularity, documentation, extendability.
>order in which you list the "also" requirements is rather weird
>with respect to most code a typical application programmers
>develops in the course of his or her career;-).

Just scramble the order to suit your or the typical application
programmer's tastes. ;)

>Still, this discussion helped me -- thanks.  Basically, you're
>saying you found the Python docs you used to be no better than
>those for other languages, or, at least, not better by enough
>to satisfy your specific interests -- an early focus on HOW
>things are being done rather than on WHAT EXACTLY is being done,
>implementation rather than interface, so to speak.  So, from
>my current, incomplete understanding, I think this is a rather
>"niche", marginal issue.

I didn't mean to start any crusade to change Python docs. If
you consider it a niche, that's just fine with me.

Marcus



From kp at kyborg.dk  Tue Apr  8 10:43:23 2003
From: kp at kyborg.dk (Kim Petersen)
Date: Tue, 08 Apr 2003 16:43:23 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk>
Message-ID: <3E92E00B.6000708@kyborg.dk>

> (defun foo()
>   (message "when foo is called folding is: %s" case-fold-search))
> (defun bar()
>   (message "when bar is called folding is: %s" case-fold-search)
>   (let ((case-fold-search t))
>     (foo))
>   (message "now folding is: %s again" case-fold-search))
> (bar)


hmmm - occured to me that you might mean that case-fold-search was a 
global variable first - then local.... but that wouldn't
matter to this since we would need to do lookup in locals first _then_ 
in globals (as expected)



-- 
Med Venlig Hilsen / Regards

Kim Petersen - Kyborg A/S
IT - Innovationshuset
Havneparken 2
7100 Vejle
Tlf. +4576408183 || Fax. +4576408188




From news at snakefarm.org  Tue Apr  8 17:23:30 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 8 Apr 2003 21:23:30 GMT
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com>
Message-ID: <slrnb96fei.ul2.news@cipw08.physik.uni-bonn.de>

Heiko Henkelmann wrote:
> I need to split a string into an array of equally long substrings, like
> e.g.:
> 
> '1234567890'   -> ['12', '34','56','78','90']
> 
> What is the most efficient way to do this?

I don't know how efficient it is but this is what came to my mind:

import re
filter(None, re.split('(.{2})', '1234567890'))

cg.




From marklumsden at comcast.net  Wed Apr 30 10:53:29 2003
From: marklumsden at comcast.net (Mark Lumsden)
Date: 30 Apr 2003 07:53:29 -0700
Subject: win32com and makepy problem with LabVIEW
Message-ID: <8653a280.0304300653.72ce712@posting.google.com>

Hi,

I'm having some problems using makepy and win32com to talk to LabVIEW
with python.  I have a number of methods in LabVIEW that require "By
Ref" variable types and therefore I need to use the early binding
provided from makepy.  So I run makpy on the LabVIEW server and it
generates the code correctly.  However, whenever I try to initiate a
session (using win32com.client.Dispatch), it doesn't seem to enable
early binding.  Forcing early binding from gencache.EnsureModule
generates a 'Library not registered' error.  In addition, when I run
'makepy -i' I get the following output:

>>> Warning - could not load registered typelib
'{9A872073-0A06-11D1-90B7-00A024CE2744}'
LabVIEW 6.1 Type Library
 {9A872073-0A06-11D1-90B7-00A024CE2744}, lcid=0, major=5, minor=2
 >>> # Use these commands in Python code to auto generate .py support
 >>> from win32com.client import gencache
 >>> gencache.EnsureModule('{9A872073-0A06-11D1-90B7-00A024CE2744}',
0, 5, 2)

I believe the Warning in the first line above is at the root of the
problem.

I can work around this problem - if I rename the file
9A872073-0A06-11D1-90B7-00A024CE2744x0x5x2.py generated from makepy to
labview.py and then load this file directly. i.e.:

import labview
test=labview.Application()

Then the early-binding works and everything is fine.  I can live with
this solution but I'd be happier knowing what the problem is.

Thanks,
Mark



From hal_wine at yahoo.com  Fri Apr  4 19:40:31 2003
From: hal_wine at yahoo.com (Hal Wine)
Date: Fri, 04 Apr 2003 16:40:31 -0800
Subject: Invalid format character in string
In-Reply-To: <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com> <mailman.1049493668.26508.python-list@python.org> <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>
Message-ID: <3E8E25FF.5060600@yahoo.com>

Nikolai Kirsebom wrote:
> Hi,
> The problems arise where the HTML editor used (to produce the file in
> the first place) has put various expressions of %; and %>.  These are
> HTML codings the creator of the HTML file does not really know about.
> As an example, I've just told them (the users creating the HTML file)
> that they should write the text "%(name)s" where they want the name
> generated and "%(age)s" where they want the age generated.  Then my
> script is run for a set of person objects producing a set of HTML
> files with the correct names and corresponding ages. 

Try replacing those sequences with entries that you add to your 
dictionary to map it back to the original:
   html_in = html_in.replace('%;', '%(_SEMI)s' )  \
                    .replace( '%>;', '%(_SEMI_GT)s')
   dict.update({'_SEMI':'%;', '_SEMI_GT':'%>;'})
   html_out = html_in % dict

--Hal




From fidtz at clara.co.uk  Fri Apr 25 10:31:49 2003
From: fidtz at clara.co.uk (Fidtz)
Date: 25 Apr 2003 07:31:49 -0700
Subject: Reading a non-standard floating point format
Message-ID: <24944f78.0304250631.1778654e@posting.google.com>

I'm trying to do a database extraction program. Part of this is
reading a large number of files with a 48bit floating point format
that I think from the definition and searching on google, is unlike
any on any other system (at least current ones anyway).

While I think I can see the way to do this with various bit-grubbing
techniques, the potential for error is massive, esp considering my non
comp-sci background.

Does anyone know of a library in python or even (fairly) easily
wrappable C that might let me specify the format in a more high level
way?

Dominic



From Bill.Scherer at verizonwireless.com  Tue Apr  8 10:13:30 2003
From: Bill.Scherer at verizonwireless.com (Scherer, Bill)
Date: Tue, 8 Apr 2003 10:13:30 -0400 (EDT)
Subject: ANN: SQLObject 0.3
In-Reply-To: <1049776366.759.96.camel@lothlorien>
Message-ID: <Pine.LNX.4.44.0304081013030.16340-100000@bagend.odc.vzwcorp.com>

On 7 Apr 2003, Ian Bicking wrote:

> SQLObject 0.3 has been release:
>   http://sqlobject.org

Is support for Oracle planned?

> 
> Changes
> -------
> 
> * Automatic table creation based off class.
> * Create classes from database.
> * Dynamically add and remove columns to live classes and databases.
> * anydbm-based backend (queryable database without SQL).
> * Better caching.
> * Some bugs fixed.
> 
> What is it?
> -----------
> 
> SQLObject is an object-relational mapper.  It allows you to translate
> RDBMS table rows into Python objects, and manipulate those objects to
> transparently manipulate the database.
> 
> SQLObject emphasizes convenience.  It's easy to learn and easy to use,
> and doesn't require too much typing or use of any tools outside of your
> editor and your Python source.
> 
> 
> 

-- 
Bill.Scherer at Verizon Wireless
RHCE 807101044903581





From grante at visi.com  Mon Apr 21 14:14:22 2003
From: grante at visi.com (Grant Edwards)
Date: 21 Apr 2003 18:14:22 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org>
Message-ID: <3ea434fd$0$180$a1866201@newsreader.visi.com>

In article <mailman.1050946774.25112.python-list at python.org>, Jp Calderone wrote:

>> I'm blabbering! Anyway, my problem is this. I tried to compile
>> a simple Hello World program into something runnable on
>> non-python computers, and it came out being 7MB in size. Is
>> there something I'm doing wrong? Just so you know, I do know
>> somewhat what I'm doing. I know that it's incorporating all of
>> the GUI. So is there a way to not incoporate all of the gui,
>> and still have a functional program? I'm thinkin in the 800k
>> range for a simple program? The reason I ask is because I'm
>> thinking of porting something I made in VB to Python before I
>> develop it any further.
> 
> Though 7MB is slightly larger than the numbers I've seen, it is
> in the right ballpark.
> 
> It's easy to see why.  You're distributing an entire runtime
> environment, instead of just your application.  When you write
> VB code, you're only distributing your application.

Has anybody tried distributing their Python app w/o bundled
Python but with a wrapper/isntaller that will download and
install Python if it isn't installed?  I'm currently
distributing a very simple wxWindows app bundled with Python
and wxWindows (so the download is pretty huge).  For customers
that already have Python and wxWindows, it seems like a waste.

I've now idea how many customers (if any) already do have
Python. I suspect very few do, so it's sort of a moot point.

-- 
Grant Edwards                   grante             Yow!  CHUBBY CHECKER just
                                  at               had a CHICKEN SANDWICH in
                               visi.com            downtown DULUTH!



From grante at visi.com  Fri Apr 25 17:44:44 2003
From: grante at visi.com (Grant Edwards)
Date: 25 Apr 2003 21:44:44 GMT
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com> <3ea947e1$0$198$a1866201@newsreader.visi.com> <mailman.1051301620.25438.python-list@python.org>
Message-ID: <3ea9ac4c$0$179$a1866201@newsreader.visi.com>

In article <mailman.1051301620.25438.python-list at python.org>, Fredrik Lundh wrote:
> Grant Edwards wrote:
> 
>> Tkinter is indeed a pretty good place to start -- just try not
>> to think about the fact that you're pulling in a Tcl interpreter
>> also.  ;)
> 
> I do find it a bit worrying that every time I start a Python
> program, it pulls in a C runtime library.  Guess I have to do
> something about that...

But those C runtime library routines aren't redundant. they're
providing functionality that Python doesn't implement. OTOH,
having both Tcl and Python interpreters seems a bit redundant.

I have used Scheme with Tk bindings, and that didn't use Tcl.
The Scheme interpreter was bound directly to the Tk widgets.
That results in several advantages to the application
programmer, but disadvantages to the binding maintainer. I
completely understand why the designers of Tkinter chose to do
it the way they did -- and I probably would have done the same
thing.

-- 
Grant Edwards                   grante             Yow!  While I'm in
                                  at               LEVITTOWN I thought I'd
                               visi.com            like to see the NUCLEAR
                                                   FAMILY!!



From afriere at yahoo.co.uk  Tue Apr 15 03:32:26 2003
From: afriere at yahoo.co.uk (Asun Friere)
Date: 15 Apr 2003 00:32:26 -0700
Subject: A python server
References: <15f445fb.0304140751.170d410e@posting.google.com>
Message-ID: <38ec68a6.0304142332.4a901e14@posting.google.com>

santacruz at southern.edu (dansan) wrote in message news:<15f445fb.0304140751.170d410e at posting.google.com>...
> Hello all!
> 
> I have read and re-read the PY Sockets HOWTO, and I guess I'm still in
> the dark about how to do some stuff.
> 
> I'm trying to write a little server that will allow connections via
> telnet-like clients to play a board game. 

Rather than using low-level sockets, you might want to have a look at 'telnetlib'
in the standard library.



From ramen at lackingtalent.com  Tue Apr  1 10:54:50 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 15:54:50 -0000
Subject: Slicing in Python 2.3a2
References: <mailman.1049200983.30254.python-list@python.org>
Message-ID: <slrnb8jamj.421.ramen@lackingtalent.com>

In article <mailman.1049200983.30254.python-list at python.org>,
S?bastien Keim wrote:
> I have some troubles to understand how work slicing in python 2.3.
> In my examples, I will use the following list:
> >>> A = range(10)
> 
> ...
> 
> But I don't understand the result for a subslice with a negative third 
> argument:
> >>> A[1::-1]
> [1, 0]
> Intuitively, I expected to get:
> >>> A[1:][::-1]
> [9, 8, 7, 6, 5, 4, 3, 2, 1]

You can get this with A[:0:-1].

> The numarray package has the same behavior, so I guess it's intentional.
> >>> import numarray
> >>> numarray.array(range(10))[1::-1]
> array([1, 0])
> 
> Then what is the real semantic for this third argument?

I think it makes sense if you think in terms of "start", "stop", and "step".
The step amount represents the value that is added to the implicit counter
with each iteration. I think that the confusion actually comes from the
defaults for start and stop if they are not supplied. It seems from your
example (and the [::-1] form) that the defaults for start and stop are
swapped if the step is negative.

HTH,
Dave



From rmunn at pobox.com  Mon Apr 21 13:47:03 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 21 Apr 2003 17:47:03 GMT
Subject: "pythonic"
References: <mailman.1050921572.30147.python-list@python.org>
Message-ID: <slrnba7mt3.shc.rmunn@localhost.localdomain>

soporte_cali <soporte_cali at terra.com.co> wrote:
> I am quite new to python so please forgive me if my question is too 
> newby. I am running a phyton exe file, linux red hat, it is an indexing
> system that takes a flat file and index it. the py program once runned
> is giving  me this:
> 
>  ./index.py:57: SyntaxWarning: import * only allowed at module level
>   def createIndex(dbname,idtokenNum=1,keySize=10,idtokenChar=' '):
> error: unknown library " <nameoflibrary.py> " does not exit 
> Library  nameoflibrary.py  Does not exist
> 
> so where should I place the name of this library? I mean is there any
> special place to put it? libnames? well thanks in advance.. 

That doesn't look like any Python error message *I'm* familiar with...

Is this the actual error message, or your own paraphrase? If it's the
latter, plase copy and paste the actual error message. It would also be
helpful if you could show the code that is causing the error. Not the
entire file, of course, but maybe the line that's causing the error plus
the surrounding code (say, the function that line appears in).

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From staschuk at telusplanet.net  Tue Apr 29 12:18:42 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 10:18:42 -0600
Subject: Question about accessing class-attributes.
In-Reply-To: <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk>; from mwh@python.net on Tue, Apr 29, 2003 at 08:54:54AM +0000
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304251030.30fd1f94@posting.google.com> <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk>
Message-ID: <20030429101842.A338@tibia.amotlpaa.bogus>

Quoth Michael Hudson:
> mis6 at pitt.edu (Michele Simionato) writes:
  [...]
> > 2) why you say this makes classes unhashable ?
> 
> Well, I thought it would.  I was wrong:
  [...]
> Then you get this kind of weirdness:
> 
> />> class C(type):
> |..     def __eq__(self, other):
> |..         return 1
> \__ 

Can't the instances of this metaclass be made sensible for use in
dicts simply with
    def __hash__(self):
        return hash(C)
for example?  (To ensure that if x == y then hash(x) == hash(y),
as desired.)  Or does this do weird things in a metaclass context?

(Ignoring issues arising from the brokenness of the __eq__:
    class foo(object):
        __metaclass__ = C
    foo() == 3               # true!
    hash(foo()) == hash(3)   # probably false
and like cases.)

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From andy at reportlab.com  Thu Apr 17 16:16:26 2003
From: andy at reportlab.com (Andy Robinson)
Date: Thu, 17 Apr 2003 21:16:26 +0100
Subject: Python and Schools
References: <mailman.1050499405.10336.python-list@python.org> <eGdna.110601$It5.106431@news2.central.cox.net> <mailman.1050503666.19709.python-list@python.org> <uXhna.110755$It5.95103@news2.central.cox.net>
Message-ID: <vo2u9v0vh8b5o06mdjj6plpj5f97df4idi@4ax.com>

>> Or, if my memory isn't failing me:
>>
>>     sarr = Array("the", "quick", ... "dog")
>>
>
>Indeed, though that's new to me (never used VB, just the scripting version).
>Now let's create a dictionary-of-objects equivalent using the
>DictionaryOfObjects() function ;-)

I think VB actually started to borrow Python ideas and
features a few years back.  The Collection classes 
got a lot friendlier, foreach loops over collections..

Years ago someone here said that Python won't take over 
the world but its good ideas would end up in other languages...

- Andy Robinson



From max at alcyone.com  Mon Apr 14 01:59:14 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sun, 13 Apr 2003 22:59:14 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de>
Message-ID: <3E9A4E32.F179FDDC@alcyone.com>

"Greg Ewing (using news.cis.dfn.de)" wrote:

> I don't find those guidelines particularly useful except
> in fairly special cases. More generally, the rule of thumb
> I use is that __str__ is for the "normal" output of the
> program, whatever that might be, whereas __repr__ is for
> debugging output.

Same here.  Usually I define __str__ to be something short and
meaningful, and __repr__ to be something like

	def __repr__(self):
	    return "<%s @ 0x%x (%s)>" % \
	           (self.__class__.__name__, id(self), str(self))

That way it helps to see things like:

>>> Course(lowEarthOrbit, lowMarsOrbit).transfers()
[<ExtractionTransfer @ 0x82893ac (out of `low orbit around `Earth'')>,
<HohmannTransfer @ 0x812cef4 (Hohmann from `Earth' to `Mars')>,
<InsertionTransfer @ 0x82adb64 (into `low orbit around `Mars'')>]

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Freedom is a road seldom travelled by the multitudes
\__/ Public Enemy
    ZOE / http://www.alcyone.com/pyos/zoe/
 A simple Python OpenGL rendering engine.



From dino_levy at mindlessSPAM.com  Sun Apr 20 19:19:57 2003
From: dino_levy at mindlessSPAM.com (Dino Levy)
Date: Mon, 21 Apr 2003 01:19:57 +0200
Subject: Mail and text extraction
Message-ID: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com>

I have the following problem:
I'm trying to make a script which will connect to POP3 server, print
number of mails, print their size and subject, and then will ask the user
which mail should be downloaded...
So far so good, but I have problem: I also want to implement the option
for my script to search for a previously determined pattern and then
download the text with the pattern. I know I sound weird and am not
clearly explaining the problem, but this is a short example...
You get a mail with a subject Monthly payment, and it's fairly large but
all you want to do is to download ( or see ) whether you have been given
a raise. The
mail looks like this:
Agnell company

-Worker		Payment
Joe Doe		500$
Mary Smith	550$
Karen Bay	495$

-New Employee: Peter Maley

some text...

-Raises:
some text

And all you want to do is to connect to your POP3 server, and download
the section called Raises ( for example, two or three lines after the
title ), so all you want to do is to download those few lines containing your
text... ( ONLY those lines, not the whole e-mail )
I know my explanation is lame and clumsy and my english is poor, but
please help me if you understand what I'm talking about and if you have
the solution...
Thanks



From exarkun at intarweb.us  Sat Apr  5 13:21:22 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sat, 5 Apr 2003 13:21:22 -0500
Subject: Why use a GUI framework?
In-Reply-To: <1g5u8vc1k8t97jg8vsm2gi81sfnu7cm1hf@4ax.com>
References: <1g5u8vc1k8t97jg8vsm2gi81sfnu7cm1hf@4ax.com>
Message-ID: <20030405182122.GA23069@meson.dyndns.org>

On Sat, Apr 05, 2003 at 07:47:05PM +0200, Jane Doe wrote:
> Hi,
> 
> 	I know a bit more about the multiple options available to
> build a GUI-based application (Windows, in our case), but I haven't
> found information on the point of using another tool since the
> win32all extension is available to use the native Win32 widgets. 
> 
> Do wxWindows et al. just make it easier, or is it a must-use for
> people who wish to write OS-independant Python applications?
> 
> If win32all is good enough, does anybody have a good tutorial on using
> this add-on to build GUIs?

  If portability to non-Windows platforms is not important to you, win32all
is probably good enough.

  The other GUI toolkits make it -possible-, not easier, to write
OS-independant Python applications.

  Jp

-- 
A sad spectacle.  If they be inhabited, what a scope for misery 
and folly.  If they be not inhabited, what a waste of space.
                -- Thomas Carlyle, looking at the stars
-- 
 up 16 days, 14:00, 7 users, load average: 1.25, 1.31, 1.27




From robin at alldunn.com  Tue Apr 15 20:00:35 2003
From: robin at alldunn.com (Robin Dunn)
Date: Wed, 16 Apr 2003 00:00:35 GMT
Subject: Using property in classic class may not work
In-Reply-To: <3E9C99E9.9000201@eb2.net.au>
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it> <a8b7f07a.0304150527.5b0b1187@posting.google.com> <hRUma.23848$T34.709378@news2.tin.it> <3E9C99E9.9000201@eb2.net.au>
Message-ID: <3E9C9D1B.2000303@alldunn.com>

I have it on my ToDo list for figuring out how what exactly the problem 
is (I have a general idea) and fixing it.  Not sure when I'll get to it 
though, hopefully soon...

--Robin


Neil Hodgson wrote:
> [With some editing to isolate the points]
> 
> [Neil Hodgson]
>  The class in question is derived from a wxWindows class, so I can't
> make it derive just from object. The "What's New in Python 2.2"
> document only mentions property wrt new-style clases.
> 
> [A. Lloyd Flanagan]
> However, the last time I tried to subclass from both wxFrame and
> object, wxPython freaked out (that's a technical term for "I don't
> remember what happened, but boy, it sure didn't work").
> Not sure about the latest version of wxPython, though -- there's
> binaries for python 2.3 out by now.  I'll have to try it again..
> 
> [Alex Martelli]
> If wxPython keeps freaking, you may have to give up on properties
> if you HAVE to inherit its classes (can't just wrap with automatic
> delegation) and go back to "good old" (yeah right) __setattr__.  Or,
> gently pressure wxPython's maintainers to have them move to
> new-style classes... whatever aspects of old-style classes they
> may be taking for granted at present, migrating to new-style can
> hardly be a major job, in my experience.
> 
> [Me again]
> Robin may feel constrained to keep wxPython compatible with multiple 
> Python major versions.
> Here is the "freaking out" with wxPython+wxGTK 2.4.0.7 on Python 2.2 
> caused by adding 'object' as a second base class to the most derived 
> frame class, that is, the class that has the property:
> 
> Traceback (most recent call last):
>   File "DemonTerm.py", line 1343, in ?
>     app = DemonTermApp(sys.argv)
>   File "DemonTerm.py", line 1290, in __init__
>     wxApp.__init__(self, 0)
>   File 
> "/usr/local/ActivePython-2.2/lib/python2.2/site-packages/wxPython/wx.py", 
> line 1802, in __init__
>     _wxStart(self.OnInit)
>   File "DemonTerm.py", line 1327, in OnInit
>     self.frame = DemonFrame(BasicScreenDefs)
>   File "DemonTerm.py", line 577, in __init__
>     TerminalGUI.TerminalFrame.__init__(self, screenDefs)
>   File "TerminalGUI.py", line 762, in __init__
>     style=wxDEFAULT_FRAME_STYLE | wxNO_FULL_REPAINT_ON_RESIZE)
>   File 
> "/usr/local/ActivePython-2.2/lib/python2.2/site-packages/wxPython/frames.py", 
> line 150, in __init__
>     self._setOORInfo(self)
>   File 
> "/usr/local/ActivePython-2.2/lib/python2.2/site-packages/wxPython/windows.py", 
> line 60, in _setOORInfo
>     val = apply(windowsc.wxEvtHandler__setOORInfo,(self,) + _args, _kwargs)
> TypeError: Type error in argument 1 of wxEvtHandler__setOORInfo. 
> Expected _wxEvtHandler_p.
> 
>    OOR is wxPythonese for 'Original Object Return' but I don't know what 
> that really means.
> 
>    Neil
> 
> 




From sholden at holdenweb.com  Sat Apr 19 09:52:57 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sat, 19 Apr 2003 13:52:57 GMT
Subject: Pair of prescription sunglasses left at PyCon sprint
References: <mailman.1050709532.4784.python-list@python.org>
Message-ID: <Zwcoa.134503$It5.83738@news2.central.cox.net>

"Christopher Armstrong" <radix at twistedmatrix.com> wrote in message
news:mailman.1050709532.4784.python-list at python.org...
> I found a pair of prescription sunglasses on the table where the
> Twisted gang was sprinting. If anyone can claim them, send me an
> address to ship them to. Sorry for not posting about this earlier, I
> had forgotten about them and they just turned up.
>

There's a lost property bit in the Wiki somewhere, maybe you should also add
to that?

--
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From unix at unix.com  Fri Apr 11 02:26:14 2003
From: unix at unix.com (Gilles)
Date: 11 Apr 2003 01:26:14 -0500
Subject: pyserial newbie questions
References: <gilles.celli-943DC4.10083810042003@webmail.restena.lu> <8tfla.154$N5.27502@mencken.net.nih.gov>
Message-ID: <110420030823304299%unix@unix.com>

In article <8tfla.154$N5.27502 at mencken.net.nih.gov>, Mike
<uptonothing at yahoo.com> wrote:

Hello all,

Fank Zheng sent me this sample via email, and it works great.


import serial
import time
    s = serial.Serial(300) #here u can set timeout
    command = '\x00'*20
    s.write(command) #the return value of write(..) is the actual size
written.
    time.sleep(0.01)
    s.write('HELLO') 
recv = s.read(5) #if u don't set timeout, here will stop until get 5
bytes from the serial port
    if recv == 'READY':
        #do something 

Thanks again Frank :)

Gilles




> "Gilles" <gilles.celli at ecgs.lu> wrote in message
> news:gilles.celli-943DC4.10083810042003 at webmail.restena.lu...
> > Hello folks,
> >
> > Since I am very new to Python I have the following question:
> >
> > I try to write a program in Python (on RHat 8.0 Linux) with the
> > "pyserial" module which communicates to a device via RS232 serial cable.
> >
> > The program should do:
> >
> > 1. Set Com port  to 300,n,8,1
> > 2. Send 20 NULL characters Chr(0)
> > 3. Wait 10 ms ( may vary in your case)
> > 4. Send a string "HELLO"
> > 5. Wait to receive ?READY? string  from device
> > 6. If above string OK SEND   ?START?
> >
> > I figured out how to set the port, send the 20 NULL characters :)
> >
> > But I have no idea how to "wait for 10ms", send the HELLO string and
> > read the READY string...
> >
> > Anyone has a clue or source code sample ?
> >
> > Thanks a lot.
> >
> > Gilles
> 
> I don't know if this is the answer you're looking for but you can use the
> sleep(secs) function in the time module.
> 
> <code>
> import time
> 
> time.sleep(.01)  # sleep 10 ms
> ...
> </code>
> 
>



From mertz at gnosis.cx  Mon Apr 28 12:13:40 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Mon, 28 Apr 2003 12:13:40 -0400
Subject: More rants on a bad sort
In-Reply-To: <Xns936B6B217EB86duncanrcpcouk@127.0.0.1>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
 <698f09f8.0304262115.386c2571@posting.google.com>
 <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com>
 <7ULqa.9804$3M4.264757@news1.tin.it> <g3qdncyq1Zx2FzGjXTWcrg@comcast.com>
 <mailman.1051497463.22158.python-list@python.org>
 <Xns936B6B217EB86duncanrcpcouk@127.0.0.1>
Message-ID: <0MVr+kKkXUES092yn@gnosis.cx>         

|>     [u'x', 'x', chr(255)].sort()
|>     ['x', chr(255), u'x'].sort()
|>     [1j, '1', 1].sort()
|>     [1, 1j, '1'].sort()

|I must admit I was surprised that 3 of these failed to sort, since I'd
|guessed that you would have selected 2 that worked and 2 that failed. So my
|question is, did one of the last two actually work on your system?

It was a trick question in that respect too.

But the problem is that I honestly do not even UNDERSTAND why .sort()
does what it does here.  If we start from the premise that some things
are comparable, and other things are not, then we can observe that:

    >>> u'x' < 'x', 'x' < chr(255)
    (0, 1)

But

    >>> u'x' < chr(255)
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    UnicodeError: ASCII decoding error: ordinal not in range(128)

The sorting algorithm winds up only performing allowed comparisons in:

    [u'x', 'x', chr(255)].sort()

Well... OK.  We have a similar story with complex numbers:

    >>> '1' < 1, 1j < '1'
    (0, 1)
    >>> 1 < 1j
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    TypeError: cannot compare complex numbers using <, <=, >, >=

Given all that, is it POSSIBLE to construct a list containing a string,
int, and complex, that will not crash on a sort?  I don't know!  I
couldn't find one--and I THINK the relative order of complex and strings
precludes it... but actually proving this one way or another is
decidedly non-obvious.

Yours, Lulu...

--
---[ to our friends at TLAs (spread the word) ]--------------------------
Echelon North Korea Nazi cracking spy smuggle Columbia fissionable Stego
White Water strategic Clinton Delta Force militia TEMPEST Libya Mossad
---[ Postmodern Enterprises <mertz at gnosis.cx> ]--------------------------





From pobrien at orbtech.com  Sun Apr  6 21:53:41 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 20:53:41 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com>
Message-ID: <m2k7e75aze.fsf@localhost.localdomain>

llothar at web.de (Lothar Scholz) writes:

> > "comparable in power."  The reason is that there is a ton of elisp
> > code out there, including most of Emacs itself, that could be
> > translated to work with PyAlaMode.  My goal isn't to be able to
> > convert in a completely automated fashion, but close enough to be
> > attractive to both an elisp programmer and a Python programmer.
> 
> Sorry, have you ever written a elisp program, something that is larger
> then
> 100 lines. I can't think about a way to provide this in python without
> writting a lisp like interpreter. Remember that the scope of elisp
> variables for example are not lexical. This is something very hard to
> emulate in python.
> I would really suggest to look at the structure of keyhandling,
> keymaps and buffers. And then forget about lisp and design a good
> python editor.

No, the only elisp I've dealt with has been trivial changes to .emacs
and similar files.  I don't want to write a lisp-like interpreter.

All I *really* want to support are the Emacs concepts, not the actual
code, especially if there is no simple way to automatically translate
elisp into Python.  I've been assuming there is an advantage to making
the PyAlaMode API conceptually similar to the Emacs API.  Maybe there
isn't and I'm giving the Emacs API more credit than it deserves.

Is it enough to support the following:

* default Emacs keybindings
* ability to remap keybindings
* multiple buffers and multiple windows
* multiple modes
* keyboard friendliness (no need for a mouse)
* minibuffer-style control of environment

If not, what's missing and what's the definition of a "good editor"
written in Python?  Should it allow customizations and extensions,
like Emacs, but be a completely new and Pythonic API?

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From jacek.generowicz at cern.ch  Thu Apr 17 11:16:05 2003
From: jacek.generowicz at cern.ch (Jacek Generowicz)
Date: 17 Apr 2003 17:16:05 +0200
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <87k7dt5m3j.fsf@interim.henrik-motakef.de> <1azna.115402$It5.22896@news2.central.cox.net>
Message-ID: <tyfhe8xqhmy.fsf@pcepsft001.cern.ch>

"Steve Holden" <sholden at holdenweb.com> writes:

> While Guido is known to regret the inclusion of lambda in Python

But he regrets the inclusion of a crippled lambda, rather than the
inclusion of support for anonymous functions (IIRC).



From Patrick.Carabin at sciencesnaturelles.be  Tue Apr 15 05:21:28 2003
From: Patrick.Carabin at sciencesnaturelles.be (Patrick Carabin)
Date: Tue, 15 Apr 2003 11:21:28 +0200
Subject: files: direct access - NOT sequential
Message-ID: <03041511212800.01079@pc20_118>

i search a means to get " direct access " inside files, eg mixing 
read & write operaions for the same file
example:
given a file with contents : "qwerty" , positionning on the "w", and 
writing a "N" , so the file becomes "qNerty"
How can i achieve this in Python ?

TIA,
Patrick Carabin

Patrick.Carabin at sciencesnaturelles.be 
Patrick.Carabin at NatuurWetenschappen.be .
Institut Royal des Sciences Naturelles de Belgique 
http://www.SciencesNaturelles.be/
Koninklijk Belgisch Instituut voor Natuurwetenschappen 
http://www.NatuurWetenschappen.be/ 

?Le bonheur n'est pas au bout du chemin, le bonheur est le chemin.?
?Het geluk is niet op het einde van de weg, het geluk is de weg.?
						Dala? Lama.




From max at alcyone.com  Sat Apr 12 04:55:47 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sat, 12 Apr 2003 01:55:47 -0700
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk> <3E97D324.8000009@v.loewis.de>
Message-ID: <3E97D493.659A567B@alcyone.com>

"Martin v. L?wis" wrote:

> When Eric Max said "its identity is unimportant", he really
> meant "it's unimportant whether they are interned, and their
> identity is preserved in deepcopy" :-)

See, a little money under the table never hurts.  Now I don't even have
to defend myself :-).

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Who's not sat tense before his own heart's curtain?
\__/ Rainer Maria Rilke
    Bosskey.net: Unreal Tournament 2003 / http://www.bosskey.net/ut2k3/
 A personal guide to Unreal Tournament 2003.



From drew at astro.pas.rochester.edu  Thu Apr  3 22:17:19 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 3 Apr 2003 19:17:19 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <mailman.1049388384.29614.python-list@python.org>
Message-ID: <9383481e.0304031917.1d0100bf@posting.google.com>

Jp Calderone <exarkun at intarweb.us> wrote in message news:<mailman.1049388384.29614.python-list at python.org>...

> 
>     class Conductor(object):
>         def set_voltage(self, value):
>             # Twiddle whatever needs to be twiddled
>         def get_voltage(self):
>             # Lookup whatever the voltage is
>             return it
>         voltage = property(get_voltage, set_voltage)
> 
>     c = Conductor()
>     c.voltage = 3
> 
>   This is more along the lines of what you'd really do, anyway, right? 
> After all, if "voltage" is simply a global, then your program or library
> will be forever tied to controlling a single item across which an emf can
> be put.  Not very scalable, right?
> 
>   Jp
> 
> -- 
> Lowery's Law:
>         If it jams -- force it.  If it breaks, it needed replacing anyway.

Well, not exactly what I had in mind.. how bout this?

class dacvoltage() :
    def __init__(self,dacnum,volt=0.0,dacfunc=maindac) :
        self.dacnum = dacnum
        self.volt = volt
        self.writedac=dacfunc
        self.writedac(self.dacnum,self.volt)
    def __assign__(self,other) :
        # todo: add code to check type of other..
        # if it is another dacvoltage object,
        # return other, not self...
        # this will rebind the name!
        try:
            self.writedac(self.dacnum,other) # might barf.
            self.volt=other  # didn't barf? remember what was assigned.
        except:
            # complain bitterly here if writedac barfs.
        return self # retain identity..

# instantiate some dac voltage objects for the first time..
# (or re-instantiate old ones, if above todo gets done..)

vdduc = dacvoltage(1, -4.0)
vreset = dacvoltage(2, -3.0)

vbias = dacvoltage(0,dacfunc=auxdac) # bias dac uses special converter..

vbias = 2.5

# etc etc...
# do some stuff at these voltages..


# modify the voltages, in a very simple, readable syntax.
vdduc = -3.8 # try a reduced supply
vreset = -3.2 # and increased reset drive...
vbias = 2.3 # and decreas the bias a little..

# do some more stuff with new voltages, etc etc...


Drew



From exarkun at intarweb.us  Thu Apr 17 09:33:16 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 17 Apr 2003 09:33:16 -0400
Subject: inheriting variables
In-Reply-To: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de>
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de>
Message-ID: <20030417133316.GB25373@meson.dyndns.org>

On Thu, Apr 17, 2003 at 10:34:08AM +0000, Haran Shivanan wrote:
> This seems like a fairly straight forward thing but I'm not sure about how
> to implement it.
> I have a base class from which I'm deriving several new classes:
> class A:
> 	mem_a=0
> 	mem_b=1
> class B(A):pass
> 
> I want B to inherit mem_a and mem_b from A without my having to explicit do
> something like:
> 
> self.mem_a = 0
> 
> in the constructor for class B.

    class A:
        a = 0
        b = 1

    class B(A):
        pass

    b = B()
    print b.a, b.b

  Isn't this the behavior you desire?  If not, could you clarify your
question?

  Jp

-- 
It is practically impossible to teach good programming style to
students that have had prior exposure to BASIC: as potential
programmers they are mentally mutilated beyond hope of
regeneration.        -- Dijkstra
-- 
 up 28 days, 9:02, 5 users, load average: 0.04, 0.03, 0.00




From paul at boddie.net  Mon Apr 14 06:21:46 2003
From: paul at boddie.net (Paul Boddie)
Date: 14 Apr 2003 03:21:46 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>	 <mailman.1049530929.9225.python-list@python.org>	 <a3b5b608.0304051049.7048d76@posting.google.com>	 <mailman.1049571010.25625.python-list@python.org>	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>	 <mailman.1049653811.11491.python-list@python.org>	 <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org>
Message-ID: <23891c90.0304140221.20e9e7fb@posting.google.com>

Ian Bicking <ianb at colorstudy.com> wrote in message news:<mailman.1050210449.24337.python-list at python.org>...
>

[WebKit on Twisted]

> Using the Twisted CGI code, it would be a quick change to make it
> dispatch to a thread instead of a process (as with CGI).

One of the main concerns about making Twisted the "Pythonic standard"
is that people who demand a non-threaded, non-CGI handler model for
scalability reasons (and this did come up recently on
comp.lang.python, so it is a valid concern) are very likely to fall
outside the resulting audience. Unfortunately, especially given the
success of mod_python and other process-forking frameworks, that could
amount to a lot of people.

> I'm still thinking about the strategy around it, which is why I haven't
> actually tried implementing it.  As an experiment it's not that
> interesting to me, but maybe it could be more than that.

Well, I've been experimenting with implementing a small framework on
top of WebKit and mod_python (with Zope being another likely
candidate), and it has to be said that some frameworks share enough
similarity and do provide enough flexibility to make
"super-frameworks" possible. Moreover, with developments like POSH...

  http://poshmodule.sourceforge.net

...it might even be possible for such "super-frameworks" to paper over
the fundamental but desirable differences between the underlying
frameworks. For example, some people might want a threaded model,
whilst others might prefer a process-forking model together with
similarly convenient data sharing semantics.

As for the issue of standard APIs, I believe that such APIs really
only serve a useful purpose as being something to build on that is
widely agreed upon, recognised and stable; they don't necessarily
serve the purpose of being something developers really want to write
to directly. For example, Java Servlets have arguably served as a
foundation for more high-level toolkits, whereas writing directly to
the Servlet API is something that the most productive developers
stopped doing years ago.

Paul

P.S. The Web frameworks shootout paper is most informative and serves
as an interesting guide as to the most accessible frameworks that
could be subverted (or "leveraged" depending on your mindset) by such
"super-frameworks".



From dave at pythonapocrypha.com  Fri Apr 18 00:37:44 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Thu, 17 Apr 2003 21:37:44 -0700 (PDT)
Subject: Mem "leak" w/ long-running network apps?
Message-ID: <Pine.LNX.4.51.0304171941001.22199@ns0.tbye.com>

I've spent the last few weeks trying to track down a memory leak in a
long-running network app (an HTTP proxy for large-ish objects) and am
trying to find out if anyone else has encountered something similar.

I narrowed the problem down to the receiving side of the proxy, and today
in socketmodule.c I noticed that when you call sock.recv(n) a string of
size n is created and then resized after the recv to the actual size of
the data received. There's nothing wrong with that in and of itself, but
when I replaced it with the function below to do receives the ever-growing
memory problem went away (so now instead of calling sock.recv(amnt) I do a
specialmodule.recv(sock, amnt)). The function merely uses a static buffer
and then creates a Python string of just the size needed - the resize is
what got eliminated. I initially encountered the problem on Python 2.1.3
but then moved to 2.2.2 to use the gc module enhancements.

My questions are:

(1) has anybody else run into a similar problem before? This may not be
very common because the app is a little unusual in that it's long-running,
handles hundreds of concurrent connections, each connection is usually for
a large (tens to hundreds of megabytes) object, the data is all proxied
rather than being served off disk or generated by the app, and both the
upstream and downstream connections are generally fast (aggregate
throughput for the server is usually in the 100-200 Mbps range for a P3
900.

(2) Any ideas on why using the normal socket.recv resulted in ever-growing
memory use? I've spent ages using the gc module and other tools to track
down objects that should have been freed, cyclic references, etc., and
don't see any problems there and I don't think the memory is really being
leaked (in the C sense). Could it be a heavily fragmented heap or
something like that? I'd notice that after my tests ran for a long time
I'd stop them and memory usage would drop down after awhile by a few
megabytes, and upon starting my tests again (without restarting my app)
mem usage would drop down some more but not all the way down and then it
would gradually grow again to a new high-water mark, so that overnight my
process had hundreds of MB of RAM.  With my recv-replacement I'm holding
steady at about 30 MB, which is normal.

(3) Does anyone see any glaring errors in my function below? Seems to work
well enough. :)

Anyway, I don't think there's a bug in Python, and my function is
certainly not patch-worthy because it really works just for my use, and
now that my problem is gone it's mostly out of curiosity that I'm trying
to better understand why my problem is gone, but I'd appreciate any
insight or hints.

Thanks,
Dave

Here's the function. I can get away with using a single large buffer for
all my receives because while the server is a mixture of threading and
poll-based, all the I/O happens sequentially in one thread against sockets
that are known to be ready for I/O.

#define MAX_RECV_SIZE 1048576
static char RECV_BUFF[MAX_RECV_SIZE];

static PyObject *
recv_wrapper(PyObject *self, PyObject *args)
{
    PyObject *py_sock;
    PyObject *py_str;
    int sock;              /* Output socket */
    int len, n;

    if (!PyArg_ParseTuple(args, "Oi:recv", &py_sock, &len))
            return NULL;

    if (len < 0)
    {
      PyErr_SetString(PyExc_ValueError, "negative buffersize in recv");
      return NULL;
    }

    if (len > MAX_RECV_SIZE-1)
    {
      PyErr_SetString(PyExc_ValueError, "buffersize too large in recv");
      return NULL;
    }

    sock = PyObject_AsFileDescriptor(py_sock);
    Py_BEGIN_ALLOW_THREADS;
    n = recv(sock, RECV_BUFF, len, 0);
    Py_END_ALLOW_THREADS;
    if (n == -1)
      return PyErr_SetFromErrno(RecvError);

    py_str = PyString_FromStringAndSize(RECV_BUFF, n);
    if (py_str == NULL)
      return NULL;

    return py_str;
}




From blunck at gst.com  Wed Apr 16 10:54:06 2003
From: blunck at gst.com (Christopher Blunck)
Date: Wed, 16 Apr 2003 10:54:06 -0400
Subject: How many of you are Extreme Programmers?
In-Reply-To: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com>
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com>
Message-ID: <20030416145406.GA19919@homer.gst.com>

On Wed, Apr 16, 2003 at 09:36:14AM -0500, sismex01 at hebmex.com wrote:
> Many of us have used XP techniques without actually knowing
> that our methodology was labelled as "XP", it comes natural
> in the scope of our tools and goals.

Absolutely!  Which is why I believe that many of us in the Python
community follow many of the principals of "XP" without knowing it.


> I don't know; "OOP", "XP", "Structured Programming", etc.. are
> all monickers, buzzwords, fads, which come and go; they come
> into existance when some mass-market-oriented technical book
> writer says "Hey, this programmer is successful, let's see
> what he's doing"; once done, the writer creates some catchy
> name for this guy's techniques, and sells his book.
> 
> I really dislike market speak, as you can tell. :-)

Me too.  What's discouraging is that it often takes books, buzzwords,
and fads to get the attention of higher-ups.  Developing software
is not all that difficult - it's always so confusing why people make
software development more complex than it actually is.


-c

-- 
 10:50am  up 20:02,  1 user,  load average: 0.87, 0.91, 0.91




From jkn at nicorp.f9.co.uk  Wed Apr  2 09:17:23 2003
From: jkn at nicorp.f9.co.uk (Jon Nicoll)
Date: 2 Apr 2003 06:17:23 -0800
Subject: slurping yahoogroups postings with Python, like yahoo2mbox?
Message-ID: <8351bb33.0304020617.61007d5f@posting.google.com>

hi there
    Does anyone know of a python equivalent (or near-equivalent, that
I can fiddle with) of the Perl module yahoo2mbox? This 'slurps'
Yahoogroups postings into an mbox format.

Thanks for any pointers

    Jon N



From grante at visi.com  Wed Apr 30 10:54:24 2003
From: grante at visi.com (Grant Edwards)
Date: 30 Apr 2003 14:54:24 GMT
Subject: How to call built in function 'open' from inside object that has 'open' method?
References: <db81c492.0304300526.7b8d3c86@posting.google.com> <mailman.1051711257.436.python-list@python.org> <b8onf8$b7d$1@panix2.panix.com>
Message-ID: <3eafe3a0$0$184$a1866201@newsreader.visi.com>

In article <b8onf8$b7d$1 at panix2.panix.com>, Aahz wrote:

>>> More details on reasons. I have met this problem trying to write a
>>> simple script use ActiveScripting inside IE.  I am trying "file =
>>> open(...) but because context window object has 'open' method that
>>> just opens window it is called first.
>>
>>Unless you're typing "file = self.open(...)", the class you're in has
>>nothing to do with it.
> 
> Not quite true; what happens with
> 
>     class C:
>         def open(self, name):
>             f = open(name)

When I do it, it calls the builtin open().  What happens when
you do it?

----------------------testit.py----------------------
class C:
    def __init__(self):
        print "C.init()"
        print open
        print self.open
    def open(self,fname):
        print "C.open(%s)" % fname
        print open
        print self.open
        return open(fname)

c = C()
f = c.open("foo")
print f
---------------------------------------------

$ python2 testit.py
C.init()
<type 'file'>
<bound method C.open of <__main__.C instance at 0x81667c4>>
C.open(foo)
<type 'file'>
<bound method C.open of <__main__.C instance at 0x81667c4>>
<open file 'foo', mode 'r' at 0x8166178>


-- 
Grant Edwards                   grante             Yow!  What GOOD is a
                                  at               CARDBOARD suitcase ANYWAY?
                               visi.com            



From elechak at bigfoot.com  Mon Apr 21 22:05:38 2003
From: elechak at bigfoot.com (Erik Lechak)
Date: 21 Apr 2003 19:05:38 -0700
Subject: methods and vars in zope objects
Message-ID: <c42e062e.0304211805.4ced1175@posting.google.com>

Hello all,

I have been beating my head against zope for a few days now.  I am
writing python scripts, and I can't seem to find what methods and
member variables go along with the various zope objects.  I would like
to be able to use dir(obj), obj.__class__ or type(obj) like I do when
I am using python everywhere else.  But It won't let me.  The
documentation is lengthy; I read throught it, and I must have missed
it.    I am looking at the code now ... but it's just raising my blood
pressure.

Is there a simple script, document, manual, or button that will list
the methods and member vars available to a certain object?  Something
like dir().

Thanks,
Erik Lechak



From max at alcyone.com  Fri Apr 11 20:05:57 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 11 Apr 2003 17:05:57 -0700
Subject: module for doing unix "find"-like things?
References: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com> <a8b7f07a.0304111559.6707284b@posting.google.com>
Message-ID: <3E975865.3D85A4E7@alcyone.com>

"A. Lloyd Flanagan" wrote:

> If such a module doesn't exist I might write one.  A friendlier (more
> pythonic?) interface to os.stat() would be an obvious starting point.

The added bonus is that it wouldn't be that hard.  Definitely sounds
like something worth doing.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ It is human nature to think wisely and act foolishly.
\__/ Anatole France
    Bosskey.net: Quake III Arena / http://www.bosskey.net/q3a/
 A personal guide to Quake III Arena.



From beliavsky at aol.com  Tue Apr 15 08:58:30 2003
From: beliavsky at aol.com (beliavsky at aol.com)
Date: 15 Apr 2003 05:58:30 -0700
Subject: book on Python for Fortran programmers
Message-ID: <3064b51d.0304150458.595afc30@posting.google.com>

Python has been suggested by several posters to comp.lang.fortran as a
scripting language worth learning. What are some good book(s) and web
site(s) to get an experienced Fortran 95 programmer started in Python?



From mads at troest.NEVERMORE.dk  Tue Apr  8 14:08:58 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Tue, 8 Apr 2003 20:08:58 +0200
Subject: Loading a class the name of which is only known runtime
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net> <3BDka.4759$Du.16741@newsc.telia.net>
Message-ID: <1ruvtw54py90d.1sob8ay5wclmb.dlg@40tude.net>

On Tue, 08 Apr 2003 17:24:15 GMT, Tim Gahnstr?m /Bladerman wrote:

Hi Tim, and thanks for your response.

> If this is what ou are asking for it is certanly possible.
> 
> if test:
>     import modul1
>     a=class1()

Actually, what I am looking for is more alone the lines of:

	moduleToImport = "SomeModule" # <- Retrieved as CGI argument
	classToInstantiate = "SomeClass" # <- Retrieved as CGI argument

	import moduleToImport
	someInstance = classToInstantiate()

Of course, the above does not work. But is there any way to achieve what I 
outline here?

Regards,
   /\/\\ads Orbesen Troest



From peter at engcorp.com  Mon Apr 28 23:17:49 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 28 Apr 2003 23:17:49 -0400
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk>
Message-ID: <3EADEEDD.7B740C6C@engcorp.com>

Garrison Tsang wrote:
> 
> I just signed up and this is my first post.
> 
> Knowing nothing about Python at the moment, I need to convert a python-code
> into compatible c-code.  I recall that were some program/utility that does
> the transformation for you.

I doubt it.  Maybe it would be helpful if you explained why you think you
need to convert the Python program into C?  Does it not run fast enough?
Does it have to interface with something else that is in C?  Does it have
to run on a platform on which Python is not supported?

If you haven't got a really good reason, you might be better off leaving
the code in Python.  If you think it's an issue that you don't know how
to program in Python yet, believe me, you'll become more proficient,
faster, with Python than you are with C in a very short time.

-Peter



From ark at research.att.com  Fri Apr  4 16:33:34 2003
From: ark at research.att.com (Andrew Koenig)
Date: Fri, 4 Apr 2003 16:33:34 -0500 (EST)
Subject: output buffering--am I missing something obvious?
In-Reply-To: <16013.63756.223973.484125@montanaro.dyndns.org> (message from
	Skip Montanaro on Fri, 4 Apr 2003 15:28:44 -0600)
References: <yu99istu9ds9.fsf@europa.research.att.com> <16013.63756.223973.484125@montanaro.dyndns.org>
Message-ID: <200304042133.h34LXYt02986@europa.research.att.com>

Skip> I tried using 1.5.2 (!) on a Solaris 8 machine (it's what comes on the
Skip> Software Companion CD for Sol8) I get the same results:

>>>> import sys
>>>> sys.stdout.write("foo")
foo> sys.stdout.flush()
>>>> sys.stdout.write("bar\n")
Skip> bar

Skip> Bit rot somewhere?  It certainly shouldn't do that.

In a follow-up message, I suggested that interactive input tossed
unwritten output.  I'm now not so sure it's that simple, because if I
do this:

>>> sys.stdout.write("foooooooooooooooooo"); sys.stdout.flush()

it appears to write the argument string *and then to backspace over it*,
erasing the characters from view.

Very odd.






From mads at troest.NEVERMORE.dk  Tue Apr  8 15:11:31 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Tue, 8 Apr 2003 21:11:31 +0200
Subject: Loading a class the name of which is only known runtime
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net> <3BDka.4759$Du.16741@newsc.telia.net> <1ruvtw54py90d.1sob8ay5wclmb.dlg@40tude.net> <d1Fka.7728$y3.941559@news010.worldonline.dk>
Message-ID: <on9yn6snibwu.1opr6c4g1aaj0.dlg@40tude.net>

On Tue, 8 Apr 2003 21:01:56 +0200, Bjarke Dahl Ebert wrote:

Hi Bjarke;

> Have a look at the builtin function __import__.
> themodule = __import__(themodulename)
> theclass = getattr(themodule, theclassname)
> a = theclass()

Thanks for your suggestion, this seems like just what I am after!

> Of course, there are all kinds of security issues with doing __import__ and
> getattr like this in a CGI script...

Well, you're right of course, but I actually don't plan to allow arbitrary 
imports from the CGI parameter (although I know my post to Tim suggested 
that). What I am doing is having a central python script registered as 
handler for some file extensions in Apache, and I have pages implemented as 
python classes. What I want to do is for the central python handler to 
"dispatch" the page (class) given as argument. Now, if I restrict all these 
page objects to reside in a fixed package, and only allow for import of 
that package, I should think that the security issue (arbitraty module 
import) goes away, as only page-objects inside the page-package can be 
specified.

At any rate, thank you very much for your feedback!

Regards,
   /\/\\ads Orbesen Troest



From news at snakefarm.org  Tue Apr  8 12:30:37 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 8 Apr 2003 16:30:37 GMT
Subject: fdopen() question
References: <slrnb95s3m.u9b.news@cipw08.physik.uni-bonn.de>
Message-ID: <slrnb95u9d.r1m.news@cipw08.physik.uni-bonn.de>

Carsten Gaebler wrote:
> 
> Traceback (most recent call last):
>   File "./listener.py", line 27, in ?
>     f = os.fdopen(fd, 'r')
> OSError: [Errno 9] Bad file descriptor
> 
> What am I missing?
> 

Seems I have to open the file in 'w+' mode to make it work.

cg.



From exarkun at intarweb.us  Thu Apr 10 17:58:29 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 10 Apr 2003 17:58:29 -0400
Subject: Python compared toJava
In-Reply-To: <20030410172428.GA4764@isis.gerrietts.net>
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com> <20030410172428.GA4764@isis.gerrietts.net>
Message-ID: <20030410215829.GA24402@meson.dyndns.org>

On Thu, Apr 10, 2003 at 10:24:28AM -0700, Geoff Gerrietts wrote:
> Quoting Blaktyger (blaktyger at hotmail.com):
> > Hi!
> > 
> > I recently read that the Python interpreter is like the Java Virtual Machine.
> > Is that true?
> 
> It's hard to tell exactly what you mean by this, but since you seem to
> have received an unusually high proportion of flip or catty responses,
> I'll try to reply.
> 
> A Java program runs inside the context of the JVM. A Python program
> runs inside the context of the Python interpreter. In this way they
> are similar: the actual process observed in the operating system's
> process table will be the JVM or interpreter, not the program itself.

  Runtime context is a property of exactly that - the runtime.  You're
describing CPython, it seems, which, while the canonical Python runtime, is
not the only Python runtime (given reasonable bounds for what one considers
"Python").  See Jython, for example, where the answered to the OP's question
would be "No, it is not *like* the Java Virtual Machine, it *is* the Java
Virtual Machine".  Consider also a future in which a smart person has
written a native compiler.

  The language can remain the same while the runtime environment changes. 
Don't confuse properties of one with properties of the other.

> [snip - more good information about CPython]

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 21 days, 17:01, 8 users, load average: 2.05, 2.10, 2.07




From gh at ghaering.de  Mon Apr 28 22:40:06 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 29 Apr 2003 04:40:06 +0200
Subject: POD vs. reST for standalone writing?
In-Reply-To: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
Message-ID: <3EADE606.5080209@ghaering.de>

eichin at metacarta.com wrote:
> So, after a recent bad experience with LyX and version skew, I was
> looking at alternatives for software end-user and administrator
> documentation.  The constraints are pretty simple:
> 
> * produce good html and PDF
>   ** via linux tools, and highly preferably free ones
> * be editable on multiple platforms
>   ** linux first, but also, osx, windows
> * support "literal computer output" and screen shot
> [...]

I'd give OpenOffice a try. I haven't tried the HTML export much. PDF 
export exists in the 1.1 beta.

-- Gerhard




From aleax at aleax.it  Wed Apr 16 06:24:13 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 10:24:13 GMT
Subject: encapsulation checking features
References: <3e9b1298@cs.colorado.edu> <U8Gma.14362$LB6.388431@news1.tin.it> <3e9ca7a5@cs.colorado.edu>
Message-ID: <hbana.29034$T34.835762@news2.tin.it>

Eric Scharff wrote:

> Thanks for the reference to PyChecker, I'll have to check it out.
> 
> I've not read enough python code to know how people traditionally treat
> "instance variables that ought to be public" and "instance variables that
> ought to be private" (other than prepending underscores, which I never
> liked).

Whether you like it or not, it *IS* exactly the Pythonic approach --
if you don't prepend an underscore to the name, you're conventionally
saying that name IS part of the public interface to the module or
class -- how it's IMPLEMENTED in the future is up to you, but client
code is entitled to rely on the existence of SOME implementation.


>> But I really don't see the problem here.  If you want myVar.x to
>> call an accessor, you make x into a property of myVar's class,
>> period.  Absolutely no need to design trivial do-nothing accessors
>> either -- you can perfectly well have all client code access all
>> instance variables you want, and if and then such "variables"
>> need to be "computed on the fly" or have special effects when
>> set, then you make them into properties, hey presto.
> 
> I come from a Smalltalk background, which implicitly treats all direct
> variable references as private to a class, requiring do-nothing accessors.
> The advantage of these in Smalltalk is that I find it makes refactoring
> easier - you can ask Smalltalk to tell you everywhere where the "myVar:"
> setter is called.  I could grep all of my python code for ".myVar =" but
> that technique is slightly less accurate.

Python is full of features that will make it nearly impossible (not
just "slightly less accurate") to statically identify "all places
where X is performed".  For example, consider identifying "all
places where function F is called" (no matter whether it's a method,
etc).  Grepping for "F(" isn't just "slightly less accurate", it's
simply broken, because it will miss all the occurrences of e.g.
    mylist.append(F)
in some place in the code with some other places then having e.g
    mylist[chooser()]()

If your main purpose in life is easily and statically finding "all
places where X is performed", then Python is one of the worst
languages you could possibly choose -- most aspects of its power,
such as "everything is first-class", dynamism, etc, etc militate
_against_ allowing easy and accurate static analysis.

Even if one was willing to bear the burden of zillion of silly
empty deuced boilerplate accessors everywhere (which I wouldn't be:
the fact that Python lets me do without boilerplate is one of the
crucial ways it enhances my productivity), Python STILL wouldn't
let you satisfy your heart's desire in this.  E.g. even if you
did use bedraggled setMyVar and getMyVar semantically-redundant
accessors everywhere, STILL in order to easily find all calls to
them you'd have to impose LOTS more limitations and discipline
on client-code -- no sticking such methods in lists and dicts, no
passing boundmethods as callback arguments to other components,
and so on, and so forth.

I don't think that trying to code Smalltalk in Python is a winning
idea, in general -- any more than trying to code Fortran in C, or
any other such combination.  "If you want Smalltalk, you know
where to find it", to paraphrase Ritchie [some claim it's a legend
that Ritchie ever said this, but the concept is surely germane;-)].

If you find that your overall productivity gets _damaged_ by
using Python Pythonically (for example because you so strongly
dislike some of Python's rules and conventions -- such as the
leading-underscore idea -- that you refuse to use them, or spend
your time and energy loathing what you're coding rather than
coding it;-), then it's quite possible that you should instead
try Ruby, or Smalltalk variants (Squeak?), or other languages
yet.  Personally, I think MOST programmers will easily be able
to adapt, but I've seen some counterexamples where somebody was
*SO* utterly attached to idioms and approaches of some other
language (Perl being the one I observed most often) that, no
matter what, they kept "trying to code Perl [or other X] in
Python" and ended up with a net loss in terms of productivity
*AND FUN* (I think it's important to have FUN programming --
if you spend your time and energy *fighting* your tools because
you hate the way they work, it's unlikely that you're having
fun, unless, perhaps you have a peculiarly pugnacious
personality...;-).


>>> These are great features to have, but it'd be nice to have warnings to
>>> discourage their use at times.
> 
>> I can see that for the "ensure all instance variables exist at
>> the end of __init__".  I can't see that, given the existence of
>> properties, for "no access to instance attributes from outside
>> of methods".  Perhaps you can clarify what advantage the latter
>> would provide, that properties just can't...?
> 
> Here's an example:  I implemented a simulation of something using python
> objects that represent objects in the simulation.  In my rapid
> prototyping,
> the objects had two "aspects", the simulation and the visualization. 
> Every object had a tkItem instance variable that the object used to
> configure
> using Tkinter.  Now, I want to create an easy way to enable and disable
> the
> views, or have more flexible views.  Turning off the visualization means
> finding all the references to tkItems and changing their structure.
> 
> Yes, it was badly written to start, but it made sense at the time. :)  Now
> I want to get as much built-in help for refactoring as possible.

I'm not sure I entirely follow -- you're talking about restructuring
the CLIENT code of your simulation?  Where said client code is entirely
under your control?  You just can't do it by static analysis, not
reliably for all usage patterns that client-code may have used -- but
if the client code is entirely under your control, why wouldn't just
grepping for the attributename you need to refactor around be entirely
satisfactory?  If you've used peculiar ways of access, such as
getattr with runtime-computer string expressions, you'll need a bit
more work, of course, but how do you think any static analysis tool
might help you with those...?


Alex




From ckea25d02 at sneakemail.com  Wed Apr  2 18:12:10 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 03 Apr 2003 11:12:10 +1200
Subject: Comments
In-Reply-To: <b6ff2r$g1a$1@news.schlund.de>
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <435d40e0.0304011457.56e32bd0@posting.google.com> <sKSdnfwAUb0mrxejXTWcog@comcast.com> <435d40e0.0304021052.736293b3@posting.google.com> <b6fc6e$1fe$1@panix2.panix.com> <b6ff2r$g1a$1@news.schlund.de>
Message-ID: <b6fq8g$4unng$1@ID-169208.news.dfncis.de>

Carsten Gaebler wrote:
> But multi-line strings are evil, too. At least when they occur in an 
> indented block and leading whitespace is not allowed in the string 
> your're building

Last time I pondered this issue, I decided that the
correct way to fix this would be to have a statement
like

   string my_string:
     |  The string can contain any characters and
     |have multiple lines and leading whitespace,
     |and its indentation level is made clear by
     |the leading | characters.

Teaching the tokenizer to understand this could
be an interesting exercise, however...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From staschuk at telusplanet.net  Mon Apr  7 23:05:31 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 7 Apr 2003 21:05:31 -0600
Subject: how to test when correct output is "random"?
In-Reply-To: <b6qoto$7v6a1$1@ID-169208.news.dfncis.de>; from ckea25d02@sneakemail.com on Mon, Apr 07, 2003 at 02:57:40PM +1200
References: <20030404215312.GE7895@unpythonic.net> <mailman.1049564865.5908.python-list@python.org> <b6qoto$7v6a1$1@ID-169208.news.dfncis.de>
Message-ID: <20030407210531.A4539@tibia.amotlpaa.bogus>

Quoth Greg Ewing (using news.cis.dfn.de):
> Steven Taschuk wrote:
> > Consider two implementations of this function:
> > 	def permute(seq):
> > 		"""Returns a copy of seq, randomly permuted."""
> > 		# ...
> 
> In that kind of situation, what you should probably
> be checking is not that a particular permutation is
> produced, but just that the output is *some* permutation
> of the input.

Indeed.  I think I mentioned this idea under the heading "sanity
checks" later in my post.  I introduced permute() (which is not
the OP's example, note) to illustrate why I don't like the idea of
mocking the PRNG and testing for specific output.

> You may also want to call it a few times and check
> that it returns a different permutation each time.

Though note that such a test can produce false negatives.

  [...]
> If you're worried that the output might pass those
> tests but still not be random enough, then you're
> into statistical testing, as Mr. Tasch suggested.

Am I Mr. Tasch?  If so, whose post do you think you're quoting
above and, I infer, responding to?  I am puzzled.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"[T]rue greatness is when your name is like ampere, watt, and fourier
 -- when it's spelled with a lower case letter."      -- R.W. Hamming




From max at alcyone.com  Tue Apr 22 18:27:34 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 15:27:34 -0700
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org>
	        <3EA5B173.9420324C@alcyone.com>
	        <eppstein-D065BC.14323022042003@news.service.uci.edu> <mailman.1051048660.32673.python-list@python.org>
Message-ID: <3EA5C1D6.A0F30D9A@alcyone.com>

Skip Montanaro wrote:

> This is why you should avoid version dependencies in your code
> wherever
> possible. ;-) (Not casting aspersions toward Erik or David.  It's just
> that
> you're bound to make mistakes with this sort of minutiae.)

Indeed.  This is why I, too, use and recommended the old try: ... except
NameError: ... mechanism.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From jkraska at san.rr.com  Sat Apr 12 22:30:20 2003
From: jkraska at san.rr.com (Courageous)
Date: Sun, 13 Apr 2003 02:30:20 GMT
Subject: exceptions == errors?
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com> <roy-AB9C56.20201307042003@reader2.panix.com> <og0ma.8129$T34.232220@news2.tin.it>
Message-ID: <2tih9vgrb1qcmaum842qjmc2ocgkfq87ri@4ax.com>

>> And, of course, "catch" is spelled "except" in Python :-)
>
>One of the very, VERY few places where I'd prefer the C++ choice

I would have, too.  Water under the bridge. Perhaps sometime
between now and the next Hundred Years Language, someone will
"fix" it. :-)

C//




From puccio_13 at yahoo.it  Wed Apr 16 10:37:02 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Wed, 16 Apr 2003 16:37:02 +0200
Subject: reading Unix mailbox
References: <hZbna.17434$DN.480144@tornado.fastwebnet.it> <mailman.1050499476.11070.python-list@python.org>
Message-ID: <K1ena.17601$DN.484833@tornado.fastwebnet.it>

Skip Montanaro wrote:

>     Alessio> I am using Python2.3 alpha and I thought there was something
>     Alessio> updated to iterate over the messages of a mailbox in mbox
>     Alessio> format, but the next() method still return a deprecated
>     RFC822 Alessio> message object reference, am I wrong? Is there a way
>     to read Alessio> mailboxes in mbox format but iterating over
>     MIMEMessages?
> 
> Try:
> 
>     from spambayes import mboxutils
>     mbox = mboxutils.getmbox("somefile")
>     for msg in mbox:
>         do_stuff(msg)
> 
> Skip

But that requires to have spambayes, doesn't it?

-- 
bye
Alessio Pace



From evan at 4-am.com  Wed Apr 16 23:06:58 2003
From: evan at 4-am.com (Evan Simpson)
Date: Wed, 16 Apr 2003 22:06:58 -0500
Subject: do python's nifty indentation rules spell the death of one-liners?
In-Reply-To: <1050544015.99640.1@iris.uk.clara.net>
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org> <1050544015.99640.1@iris.uk.clara.net>
Message-ID: <b7l5n5$b9n$1@main.gmane.org>

Garry Knight wrote:
> Maybe I didn't completely understand what the OP meant, but this isn't a
> one-liner. It's a five-liner.

python -c """exec('''print 2$$for i in (1,4):$$  print 
i$$'''.replace('$$','\n'))"""

Yech,

evan @ 4-am






From staschuk at telusplanet.net  Thu Apr 17 15:31:10 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 13:31:10 -0600
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <3E9E3040.5040409@tismer.com>; from tismer@tismer.com on Thu, Apr 17, 2003 at 06:40:32AM +0200
References: <3E9E3040.5040409@tismer.com>
Message-ID: <20030417133110.F7968@tibia.amotlpaa.bogus>

Quoth Christian Tismer:
  [...]
> At the same time, I dismissed all of my Stackless 1.0 code, which was
> continuation-based, an absolute no-no in Guido's eyes. [...]

Is there a short way to say why Guido didn't want continuations?
Or should I go digging in the archives?

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From andrew at acooke.org  Mon Apr 28 17:15:06 2003
From: andrew at acooke.org (andrew cooke)
Date: Mon, 28 Apr 2003 17:15:06 -0400 (CLT)
Subject: Slice confusion : a[n:p] is a list exclude the last  element p 
     (thus spoke Dijkstra)
In-Reply-To: <yu99bryqjqwq.fsf@europa.research.att.com>
References: <mailman.1051532982.25352.python-list@python.org> 
     <b8k0fl$dsa$1@news.hananet.net> 
     <yu99bryqjqwq.fsf@europa.research.att.com>
Message-ID: <3925.200.27.214.82.1051564506.squirrel@127.0.0.1>

Andrew Koenig said:
> Incidentally, Dijkstra's argument about ``natural numbers'' translates
> directly into an argument about pointers in C, which permits a pointer
> to any element of an array, or to a position immediately past the last
> element of an array, but does not allow a pointer to a position before
> the beginning of an array.

otoh, pointers are annoyingly different to iterators (at least in java -
don't know about c++) which start before (you need to call next() to get
the first item).  this has the advantage of allowing iterators over empty
sets (when hasNext() is false immediately), but it's fiddly converting
between the two approaches.

sorry, this is not really python related (well, i guess it also applies to
the new iterator/coroutine stuff if you use them explicitly)

andrew

-- 
http://www.acooke.org/andrew




From news at datatailors.xs4all.nl  Tue Apr 15 11:21:45 2003
From: news at datatailors.xs4all.nl (Peter van Kampen)
Date: Tue, 15 Apr 2003 17:21:45 +0200
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <bDSma.22753$T34.687158@news2.tin.it>
Message-ID: <slrnb9o8s8.iea.news@datatailors.com>

In article <bDSma.22753$T34.687158 at news2.tin.it>, Alex Martelli wrote:
> Aahz wrote:
> 
>> In article <v9e21mhemm8c2 at corp.supernews.com>,
>> Paul Watson <pwatson at redlinec.com> wrote:
>>>
>>>We need to teach students correct design priciples to build something
>>>greater than what already exists.  We will never get very far if we
>>>require everyone to start with a quark or atom.  Yes, of course we need
>>>some people who design silicon and create microcode.  They will learn
>>>the low-level details what they need to know as they need it.  Knowing
>>>great design and organization principles will enable them to make the
>>>most of it.
>> 
>> While there's some truth to that, try explaining why the following code
>> is a Bad Idea to someone who has no basic understanding of CS:
>> 
>>     s = ''
>>     for i in range(1000000):
>>         s += str(i)
> 
> I will gladly accept the challenge, if we can agree on the details to
> determine whether I've won or lost the bet.  What I require is somebody
> with real, earnest *interest* in *understanding* this stuff -- I'll
> gladly undertake to supply all necessary *background*, but I have no
> idea on how to awake real interest in somebody whose motivation for
> learning is just "serving time" or some vague hope of getting a good
> paying job (and that would be just as much of a problem, whether the
> subject was "CS", "programming", or "artichoke farming techniques").

I would gladly volunteer to be that somebody, although I realize it
won't happen (Practicallity beats geograpy ;-)) But I think you'll be
surprised about the number of people who *would* want to understand
the basics without necessarily (re-)entering university for a
CS-course. Let me tell you my background, perhaps it can serve as an
example.

I am a programmer by accident. I love the pleasure of 'creating' when
writing programs/scripts/whatever. I studied English literature but I
have toyed with computers over the years, mostly as a means to
store/retrieve stuff my brain (sadly) keeps forgetting (that's why I
have always been interested in databases).

Thanks to BASIC I learned on a Commodore 64 and dBase I learned on my
parents computer, back in the 1980's I found myself faced with the
opportunity/neccesity (as I said, totally by accident) to develop an
MS-Access-application in 1999 (a long story I won't bother you with).
After that was finished I was rather hooked (again after almost 10
years of hardly touching a computer, couldn't afford on either). Once
I learned to connect databases to the internet I could make a living
out of it (my re-discovery of the computer coincided with the
.com-hype), but I could not afford to keep up with Microsofts
relentless 'upgrading' so I have gradually moved to linux over the
past 2 years. Now I do mostly php/mysql-stuff and try to learn python.

I do not understand discussions about O(N) en O(N^2) and would be very
interested to learn such things. I have bought 'Python in a Nutshell'
and I love it how you explain about tokens, identifiers, keywords and
so on. Very few (if at all) 'general purpose' textbooks take the
trouble to explain stuff like that.

> If the interest is there, and enough Python background to understand
> exactly what this program snippet DOES, then in about one or two
> hours of face to face instruction I can give a model that will allow
> the student to predict this snippet's disastrous performance and
> compare it to the expected performance for, e.g.,
> 
>     s = ''.join([str(i) for in in range(1000000)])
> 
> It's not rocket science (that's _another_ thread...;-) -- at this
> level, it's really simple arithmetic, a decent mental model of "an
> abstract underlying machine", AND an interest in understanding this.

So +1 for an article/book. I think you could write another best-seller
explaining this and similar basic things most CS-students take for
granted while there are many more users of python (and many other
high-level languages) who use it just to get their jobs done...

PterK

-- 
Peter van Kampen
pterk -- at -- datatailors.com



From greg at brondo.com  Mon Apr  7 11:00:01 2003
From: greg at brondo.com (Greg Brondo)
Date: Mon, 07 Apr 2003 15:00:01 GMT
Subject: Zope large PDF question
In-Reply-To: <e09988ac.0304070214.4212e2e9@posting.google.com>
References: <e09988ac.0304070214.4212e2e9@posting.google.com>
Message-ID: <Rngka.4414$2x2.2043725@dca1-nnrp1.news.algx.net>

David Currie wrote:

> Hi everyone.
> 
> I am having some problems serving large(ish) PDF files from Zope. 
> Basically, Acrobat Reader seems to hang trying to open a PDF directly
> from a Zope server when used as an IE plugin.  Downloading the file
> with right-click/save-as works fine, opening the file in the plugin
> when it is stored on an Apache server on the same machine and opening
> the file from the same place also works without any problems.  So it
> looks like the Reader plugin does not play well with Zope.  Does
> anyone have any ideas on why this could be happening?  Does Zope have
> problems transferring large files under some circumstances?
> 
> Any ideas/fixes would be much appreciated. :)
> 
> Thanks.
> 
> David Currie
Actually, you're probably hitting the same bug I did with IE.  It's 
documented in MS's knowledge base.  Try this:  In acrobat under General 
Prefs you should be able to set acrobat to NOT open the document in IE. 
  Instead, it will open in a standalone acrobat.  Try that and see if it 
fixes the problem.

Greg B.




From rstephens at vectron.com  Tue Apr 29 20:03:40 2003
From: rstephens at vectron.com (Ron Stephens)
Date: 29 Apr 2003 17:03:40 -0700
Subject: Python Tutorials
References: <15451.8030105@earthlink.net> <yyyznm9thov.fsf@adams.patriot.net>
Message-ID: <a5415304.0304291603.4171ec39@posting.google.com>

Nick Vargish <nav at adams.patriot.net> wrote in message news:<yyyznm9thov.fsf at adams.patriot.net>...
> That's a great resource, and points me to several tutorials for topics
> I keep meaning to learn more about, wxPython for example.
> 
> However, all the <font size = 6> tags are... umm... just not conducive
> to readability. It makes your list text bigger than the headlines on
> most other web pages.
> 
> Thanks for assembling the list,
> 
> Nick

OK, good criticism, Nick. I have gone back and changed the page (and
the rest of the web site) to use smaller size fonts, specifically 4
where I had used 6 on the tutorials page. I know I should use a good
CSS style sheet, but I am hopelessly inept at 'style' , 'fashion', or
'art' of any kind.


Does it look better now?

If anyone wants to donate me a decent CSS sheet, I'm open to
suggestions ;-)))



From aussie2010 at yahoo.com  Tue Apr  1 17:45:44 2003
From: aussie2010 at yahoo.com (Chris)
Date: 1 Apr 2003 14:45:44 -0800
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <3E8974E5.88B861D7@engcorp.com>
Message-ID: <435d40e0.0304011445.3abf38ff@posting.google.com>

Peter Hansen <peter at engcorp.com> wrote in message news:<3E8974E5.88B861D7 at engcorp.com>...
> Chris wrote:
> > 
> > I would like to request that Python support multi-line 
> > comments (e.g. /* */ in C).
> 
> There are at least two reasons why someone would want to do
> that, and of course it's been suggested before, but what's
> your rationale behind this?  There are generally simple 
> answers that address each possible use case, but perhaps you've
> actually got an original idea about this...
> 
> -Peter

I wanted to comment out code in a large if statement that was getting an error.

Thanks to all who suggested """~""".



From reply.in at the.newsgroup  Sun Apr  6 16:02:10 2003
From: reply.in at the.newsgroup (Rene Pijlman)
Date: Sun, 06 Apr 2003 22:02:10 +0200
Subject: Is there swap function in Python?
References: <b6q0n6$1c55$1@mail.cn99.com>
Message-ID: <3r119vosrdukh6til7bdmgredlhhehqg64@4ax.com>

Xin Liu:
>I know it is a too sample question. I can't find anything in my python
>documents about the swap funtion. 

What do you mean by the swap function? What should it do
precisely?

-- 
Ren? Pijlman



From aleax at aleax.it  Sun Apr 13 12:34:59 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 16:34:59 GMT
Subject: How to subclass datetime?
References: <ba1e306f.0304130339.3bece039@posting.google.com>
Message-ID: <Tkgma.8679$LB6.235147@news1.tin.it>

<posted & mailed>

Jane Austine wrote:

> Running Python 2.3a2 on win XP.
> 
> I'm interested in the new datetime module and I want to subclass
> date/datetime class to modify the behaviour. However, simply
> inheritting the classes(like "class mydate(date):" ) didn't work.

Please define "didn't work"...:

>>> import datetime
>>> class mydate(datetime.date):
...   pass
...
>>> xx=mydate(2003,4,14)
>>> print xx
2003-04-14
>>>

Seems to be working just fine.  What exactly isn't working for you?


> Should I resort to delegation?

Maybe, but it's hard to say unless you can be more specific!


Alex




From pan-newsreader at thomas-guettler.de  Thu Apr 10 10:52:31 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Thu, 10 Apr 2003 16:52:31 +0200
Subject: Cookbook Was: Macros in Python?
References: <b71qde$q71$07$1@news.t-online.com> <l_0la.1989$d8.141@nwrdny02.gnilink.net> <b73f31$r95$02$1@news.t-online.com>
Message-ID: <pan.2003.04.10.16.52.31.251249.815@thomas-guettler.de>

On Thu, 10 Apr 2003 11:57:04 +0200, Dominic wrote:
> P.S. It would be nice to have a database for python usage patterns.

Search for python and cookbook. It's great and available as book, too.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From nNOSPAMhaldimann at gmx.ch  Wed Apr 30 18:30:33 2003
From: nNOSPAMhaldimann at gmx.ch (Nik H.)
Date: Thu, 01 May 2003 00:30:33 +0200
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com>
Message-ID: <3EB04E89.8060509@gmx.ch>

Christopher Blunck wrote:
> What advice can you offer on spreading the testing bug around?  

Teach it at universities.

This is my first year as a comp sci major. I've been doing unit testing 
before, and I was surprised that we've actually been taught how to write 
unit tests right from the start in programming classes, along with other 
agile concepts like iterative development and pair programming (alas 
using Java not Python). It's interesting to see what effect this has on 
people that have done little or none real software development before - 
unit testing comes absolutely natural to them.

I read some accounts on universities still teaching the waterfall model 
and stuff. So this post is meant as a heads up. There might be a new 
generation of programmers with mad unit testing skills on the rise. <wink>

Nik





From pinard at iro.umontreal.ca  Mon Apr 28 08:55:10 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 28 Apr 2003 08:55:10 -0400
Subject: Pyrex Hard Case
In-Reply-To: <607d8200.0304272112.2f02bcf5@posting.google.com>
References: <607d8200.0304272112.2f02bcf5@posting.google.com>
Message-ID: <oqy91uye6p.fsf@alcyon.progiciels-bpi.ca>

[Rodrigo Benenson]

> Basically there is a Macro function that I can't map to Pyrex, so I
> create an equivalent dummy function and hand edit the .c file erasing
> the definition.

I'm not fully sure I'm replying to your question, but this might work:

---------------------------------------------------------------------->
cdef extern from *:
    MACRO_NAME(EXPECTED_ARGUMENTS)
----------------------------------------------------------------------<

If you do that, every time you use MACRO_NAME as a Python function, the
net effect will be to call the macro, and Pyrex will make no attempt to
declare it in the generated C code.  Of course, you may replace `*' above
by the header file defining the macro, if it needs an explicit #include.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From nirina at mail.blueline.mg  Sun Apr  6 18:25:09 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Mon, 7 Apr 2003 02:25:09 +0400
Subject: need help with game
Message-ID: <007501c2fc8c$9573a400$f9a2383e@raseliar>

"BBFoto" <bbfoto2002 at netscape.net> wrote in news:
3e8b1b0f.8380222 at news-server...
> I did a search on the PySol directory.  Nowhere is there a file called
> "PySol.py".  There is a series of files in sequence named
> "PySol_15.pyc"  "PySol_16.pyc"  "PySol_20..pyc"  "PySol_21.pyc" and
> "PySol_22.pyc".

Hmmm.... usually you do not have to deal with compiled files, but you can
try the following suggestion. Do this from the Python interpreter (the
prompt begins with >>> ) and not from dos prompt.

>>> yourdir='c:/program files/pysol/pysol-4.80/data'

to short cut the name directory, then

>>> import os
>>> os.chdir(yourdir)

This changes directory to yourdir. To confirm that, you can get the current
working directory by typing:

>>> os.getcwd()

You have also to set the path to yourdir, so Python finds correctly his way.
You do that with sys module:

>>> import sys
>>> sys.path

This prints a list of all paths Python needs to run properly, you append the
path to yourdir to this list with:

>>> sys.path.append(yourdir)

and yourdir appears in the list when you type again:

>>> sys.path

Now, your are able to load those pyc files with, for example:

>>> import PySol_22

If this succeds, your game should start or the prompt should return to

>>>

without complain. Typing:

>>> PySol_22

should return something like:

<module 'PySol_22' from 'c:/program
files/pysol/pysol-4.80/data\PySol_22.pyc'>

To abstract what you've done, one can say that you have loaded into the
interpreter an OBJECT called 'PySol_22' with all its ATTRIBUTES (which
attributes are originally defined in the file 'PySol_22.py' ). Get the list
of attributes with the built-in function dir:

>>> dir(PySol_22)

this returns a list like:

['__builtins__', '__doc__', '__file__', '__name__', 'afunction', 'aclass',
'another', 'main']

To use them, remember how to use os and sys modules. In this hypothetical
list, you can output some doc and request help with:

>>> print PySol_22.__doc__

>>> help(PySol_22)

and

>>> help(PySol_22.main)

i speculate 'main' is the function to execute your game, so:

>>> PySol_22.main()

should start it. That is one way, to run your game.
You may wonder why compiled Python files are unusual?
Ordinary, when creating Python program, one writes Python scripts in a text
editor (a new window with idle, edit with dos, notepad with w98 ...) then
saves it in a file with .py extension. Let's try:

save as 'hello.py' the following line:

print "hello, Python world!"

and as 'list.py' in the same directory the next 4 lines :

from os import listdir, getcwd
for file in listdir(getcwd()):
    if file.endswith('py') or file.endswith('pyc'):
        print file

To run a Python program, start a dos session, change to the directory where
you have saved the files above and type at the prompt (you will type exactly
the same thing even if you are in a bash or csh shell on Lunix box), for
example:

python list.py

to run 'list.py'. This will display at the screen all files which end with
'py' and 'pyc' and among them 'hello.py' and 'list.py'. Note that there is
no compiled file, no .pyc file, associated with them (for the moment!). I
let you run 'hello.py' and see what it does.
If you just type

python

at the dos prompt, a Python shell, an interpreter, starts. On my machine,
the prompt turns to this:

Python 2.3a1 (#38, Dec 31 2002, 17:53:59) [MSC v.1200 32 bit (Intel)] on
win32
Type "copyright", "credits" or "license" for more information.
>>>

Now, to execute a Python program, you can use the built-in function
execfile. Below, i copy some code and their predicted behavior:

>>> execfile('list.py')
list.py
hello.py
>>> execfile('hello.py')
hello, Python world!
>>> execfile('list.py')
list.py
hello.py

Note again, there is no file with .pyc extension associated with 'list.py'
or 'hello.py'.
Another way to run Python program is to import them, e.g.:

>>> import hello
hello, Python world!

But now, if you try first:

>>> execfile('list.py')

then:

>>> import list

you will see where compiled Python files come from. Typing the module name
tells you from where the module is loaded.

>>> list
<module 'list' from 'list.py'>

Exit now the Python shell and return to dos session:

>>> import sys; sys.exit()

So, when importing module the machine creates compiled Python files. Can
they be used without 'pure' Python file?
Create another directory, copy there those .pyc files and change into this
directory; from your dos prompt, type:

mkdir pyc
copy *.pyc pyc
cd pyc

If you try to execute a compiled file with

python hello.pyc

the program may run but return error like:

run_pyc_file: nested_scopes: 0
or
RuntimeError : Bad magic number in .pyc file

In this new directory, start again a Python shell. Of course, built-in
function execfile does not work now as it expects a human readable code, and
a .pyc file is machine readable code, hence the syntax error you
experienced. But import works fine:

>>> import hello
hello, Python world!
>>> import list
list.pyc
hello.pyc
>>> hello
<module 'hello' from 'hello.pyc'>
>>> list
<module 'list' from 'list.pyc'>

ouf, that becomes too long.
I hope this helps you to clarify your first contact with Python.
Distributing Python program in compiled package is not (in my own and humble
opinion) the spirit of Python, Python is open source and free. To encourage
you to go on, below you find a program you can use, modify, improve... Cut
and paste it and save as 'sfkb.py' or whatever you want. To run, type as
usual at a dos prompt:

python sfkb.py

or double-click on the icon.
The program uses winsound module, so it works only on win32 platorm.

Python is very well documented, but do not hesitate to ask the list, you are
in the Python world!

Don't forget, Python is free!
I 'll be glad if blind people can hear the sound from keyboard

i do not pretend to be a Python expert, i'm just a Python user
all other people in the Python-list are more authoritative than i.

Sincerly yours
nirinA

### -------cut here --------------
try:
    import winsound
except ImportError:
    print "this script requires winsound module"

import os
import math
import wave
from Tkinter import *

print "This may take a while...(about 6 minutes on my machine)\n\
When the 25 wave files are build, two windows will appear\n\
activate the one with title Keyboard\n\
and you can play sounds by pressing alphabetical keys.\n\
\n\
Next time you use this program, comment out wave generator code\n\
to avoid rebuilding 25 wave files.\n\
\n\
Create a directory named /images and copy on it some GIF images,\n\
uncomment out canvas example code to demonstrate some Tkinter features\n\
Have fun!"

wavename = []
parameters = (1,2,44100,44100,'NONE','not compressed')

for frequence in range(100,725,25):
    wn = 'f' + str(frequence) + '.wav'
    wavename.append(wn)

###comment out wave generator code when wave files are created
###--wave generator--
    wo = wave.open(wn, 'w')
    wo.setparams(parameters)

    for i in range(1, 44101):
        amplitude = int(30000*abs(math.sin(math.pi*i/frequence)*
                                  math.sin(math.pi*i/(frequence-5.))))
        wp = wave.struct.pack('h', amplitude)
        wo.writeframes(wp)

    wo.close()
###--wave generator--

root = Tk()
root.title("Monty Python's Flying Circus")
Label(root, text='  Python  '*10, fg='blue').pack()

def call(event):
    print "ni here, this is a sound generator"
def press_a(event):
    winsound.PlaySound(wavename[0],winsound.SND_ASYNC)
    return
def press_b(event):
    winsound.PlaySound(wavename[1],winsound.SND_ASYNC)
    return
def press_c(event):
    winsound.PlaySound(wavename[2],winsound.SND_ASYNC)
    return
def press_d(event):
    winsound.PlaySound(wavename[3],winsound.SND_ASYNC)
    return
def press_e(event):
    winsound.PlaySound(wavename[4],winsound.SND_ASYNC)
    return
def press_f(event):
    winsound.PlaySound(wavename[5],winsound.SND_ASYNC)
    return
def press_g(event):
    winsound.PlaySound(wavename[6],winsound.SND_ASYNC)
    return
def press_h(event):
    winsound.PlaySound(wavename[7],winsound.SND_ASYNC)
    return
def press_i(event):
    winsound.PlaySound(wavename[8],winsound.SND_ASYNC)
    return
def press_j(event):
    winsound.PlaySound(wavename[9],winsound.SND_ASYNC)
    return
def press_k(event):
    winsound.PlaySound(wavename[10],winsound.SND_ASYNC)
    return
def press_l(event):
    winsound.PlaySound(wavename[11],winsound.SND_ASYNC)
    return
def press_m(event):
    winsound.PlaySound(wavename[12],winsound.SND_ASYNC)
    return
def press_n(event):
    winsound.PlaySound(wavename[13],winsound.SND_ASYNC)
    return
def press_o(event):
    winsound.PlaySound(wavename[14],winsound.SND_ASYNC)
    return
def press_p(event):
    winsound.PlaySound(wavename[15],winsound.SND_ASYNC)
    return
def press_q(event):
    winsound.PlaySound(wavename[16],winsound.SND_ASYNC)
    return
def press_r(event):
    winsound.PlaySound(wavename[17],winsound.SND_ASYNC)
    return
def press_s(event):
    winsound.PlaySound(wavename[18],winsound.SND_ASYNC)
    return
def press_t(event):
    winsound.PlaySound(wavename[19],winsound.SND_ASYNC)
    return
def press_u(event):
    winsound.PlaySound(wavename[20],winsound.SND_ASYNC)
    return
def press_v(event):
    winsound.PlaySound(wavename[21],winsound.SND_ASYNC)
    return
def press_w(event):
    winsound.PlaySound(wavename[22],winsound.SND_ASYNC)
    return
def press_x(event):
    winsound.PlaySound(wavename[23],winsound.SND_ASYNC)
    return
def press_y(event):
    winsound.PlaySound(wavename[24],winsound.SND_ASYNC)
    return

### z key plays windows sound

def press_z(event):
    winsound.PlaySound("SystemExit", winsound.SND_ALIAS)
    return

###--canvas example
##list_image = os.listdir('./images/')
##from random import choice
##image_name = choice(list_image)
##img = PhotoImage(file='./images/'+image_name)
##canvas = Canvas(root, width=img.width(), height=img.height())
##canvas.bind("<Button-1>",call)
##canvas.create_image(0,0,anchor=NW,image=img)
##canvas.pack()
###--canvas example

top = Toplevel()
Label(top, text='Sound From KeyBoard',
      font=('Helvetica',20), fg='cyan').pack()
Label(top, text='Python is Cooooool!', fg='green').pack()
Label(top, text='Nirvana of the Snake', fg='yellow').pack()
top.title('Keyboard')
top.bind("<a>",press_a)
top.bind("<b>",press_b)
top.bind("<c>",press_c)
top.bind("<d>",press_d)
top.bind("<e>",press_e)
top.bind("<f>",press_f)
top.bind("<g>",press_g)
top.bind("<h>",press_h)
top.bind("<i>",press_i)
top.bind("<j>",press_j)
top.bind("<k>",press_k)
top.bind("<l>",press_l)
top.bind("<m>",press_m)
top.bind("<n>",press_n)
top.bind("<o>",press_o)
top.bind("<p>",press_p)
top.bind("<q>",press_q)
top.bind("<r>",press_r)
top.bind("<s>",press_s)
top.bind("<t>",press_t)
top.bind("<u>",press_u)
top.bind("<v>",press_v)
top.bind("<w>",press_w)
top.bind("<x>",press_x)
top.bind("<y>",press_y)
top.bind("<z>",press_z)

Button(top, text='destroy',
       fg ='red', command=top.destroy).pack()
Button(root, text='quit',
       fg ='red', command=root.quit).pack()
root.mainloop()

###----end here ----------------------------









From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr  3 14:13:33 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 3 Apr 2003 19:13:33 GMT
Subject: newbie os.stat question
References: <AV_ia.7382$4P1.570905@newsread2.prod.itd.earthlink.net>
Message-ID: <slrnb8p1ut.174.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <AV_ia.7382$4P1.570905 at newsread2.prod.itd.earthlink.net>,
Tipton Bandy wrote:
> 
> I've got my first python utility working pretty good, but
> I just need help on one point.
> 
> I need to test a filename and return true or false based on 
> whether the file is readable and writable by the user running
> my utility. Oh, yes, this is on Linux:)
> I'm used to writing things like, 'if [ -w $filename ]' in bash.
> 
> I think I should use os.stat() somehow, but it seems like a lot
> of code would be required to check the appropriate field (user,
> group, or others) based on the ownership of the file and the
> uid of the user.
> 
> This seems like such a common thing.  Is there a pre-written function
> to handle it?

os.access

(But this is fairly useless, because the situation could change by
the time you actually try to use the file.  You should just attempt
to use the file and catch OSError exceptions:

    try:
        f = open(filename)
    except OSError, e:
        if e.errno == errno.EPERM:
           print 'Operation is not permitted'

)



From graham__fawcett at hotmail.com  Tue Apr 22 10:25:54 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 22 Apr 2003 07:25:54 -0700
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com>
Message-ID: <e9570f37.0304220625.11312687@posting.google.com>

Grant Edwards wrote:
> In article <mailman.1050946774.25112.python-list at python.org>, Jp Calderone wrote:
> 
>>It's easy to see why.  You're distributing an entire runtime
>>environment, instead of just your application.  When you write
>>VB code, you're only distributing your application.
> 
> 
> Has anybody tried distributing their Python app w/o bundled
> Python but with a wrapper/isntaller that will download and
> install Python if it isn't installed?  I'm currently
> distributing a very simple wxWindows app bundled with Python
> and wxWindows (so the download is pretty huge).  For customers
> that already have Python and wxWindows, it seems like a waste.
> 
> I've now idea how many customers (if any) already do have
> Python. I suspect very few do, so it's sort of a moot point.
> 

Excellent point. I took a peek at the Chandler 0.2 package that was
made available recently. It was ~13Mb, and it seemed an awful shame to
have so much "bloat" when a pre-installed interpreter (and wxPython
lib) could have cut the package down to almost nothing.

It would be great to see this integrated with the Python Package Index
as well, for a CPAN-style (grab what I need when I need it)
installation. Well, I assume that's the point of PyPI in the first
place.

Presumably a small C program could be written and made available as a
bootstrap to be placed in binary executable installers (Inno Setup,
etc) that could detect the absence of a Python environment and
download/install it prior to final installation of the application.

The Java Web Start model might be worth consideration as well. It
would cordon off the interpreter into a separate, single download,
take care of the PyPI/Distutils stuff, etc. and still leave the client
with a configuration-free environment. Just fire up the Python Web
Start, download the application stub, and let PWS do the rest. This
would also isolate and take care of differences between build
environments (*nix vs. Windows, for example, where, in the latter, the
presence of a compiler is an abberation).

I guess the moral of the story is, "contribute some time to
PyPI/Distutils development."  Or, befriend a PyPI/Distutils developer
and send him food and get-out-of-work-free tickets. ;-)

-- Graham



From tweedgeezer at hotmail.com  Wed Apr 30 00:32:45 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 29 Apr 2003 21:32:45 -0700
Subject: Python 2.3b1: RuntimeError using rexec
References: <b8mh4l$2oo0$1@nntp6.u.washington.edu> <m31xzl6pac.fsf@mira.informatik.hu-berlin.de>
Message-ID: <698f09f8.0304292032.23b0c6a7@posting.google.com>

martin at v.loewis.de (Martin v. L?wis) wrote in message news:<m31xzl6pac.fsf at mira.informatik.hu-berlin.de>...
> eval should work, but it won't be safe if you cannot trust the string.

I'm curious, if the string was eval'ed in an environment that included
nothing except an empty __builtins__, would there be any non-DoS
security hole?  Obviously the attack could DoS by making some value
10**10**10**10 or something, but is there any actual *security* breach
possible?

Jeremy



From warlock at eskimo.com  Fri Apr 25 03:21:47 2003
From: warlock at eskimo.com (Jim Richardson)
Date: Fri, 25 Apr 2003 00:21:47 -0700
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth> <mailman.1051241558.31268.python-list@python.org>
Message-ID: <b94nn-i04.ln1@grendel.myth>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 24 Apr 2003 23:30:43 -0400,
 Tim Peters <tim_one at email.msn.com> wrote:
> [Jim Richardson]
>> I have come across a weirdness. I have a number of floats, (lat long
>> pairs) which I stuff into a dict, but when I take them out, my nice %2f4
>> number, (say, 47.5894) is something long and *almost* right, like
>> 47.589333333333333339. What gives? I am not doing anything special as
>> far as I can tell, just stuffing the numbers into a dict, and recalling
>> it later. This is with Python 2.2 on Linux, and no, it's not an original
>> Pentium :)
> 
>> ...
>> Linux, because eventually, you grow up enough to be trusted with a fork()
> 
> OTOH, I'm not sure anyone grows up enough to be trusted with a float <wink>:
> 
>     http://www.python.org/doc/current/tut/node14.html
> 
> 


I figured it was something like that, It doesn't make any difference in
my application, I just found it kind of odd.

I have no real schooling in programming, all I learn is as I go,
sometimes, stuff like this, trips me up. 

<Jim, who *still* doesn't grokk self.>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+qOILd90bcYOAWPYRAtLvAJ90amgaw1KvVFr0MSsAJluZ0DJN5QCgpI7n
EYLMHLtMIdNY/vnHQ9nJLAQ=
=8fIH
-----END PGP SIGNATURE-----

-- 
Jim Richardson         http://www.eskimo.com/~warlock

Lintel, because you don't support heavy loads with windows.



From evan at 4-am.com  Fri Apr 18 16:39:02 2003
From: evan at 4-am.com (Evan Simpson)
Date: Fri, 18 Apr 2003 15:39:02 -0500
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
Message-ID: <3EA06266.4060901@4-am.com>

Guy Middleton wrote:
> 	>>> reduce(lambda a, b: a+b, [y[1] for y in x])
> 
> But I don't really like lambda, is there a better way to do this?

What's wrong with:

t = 0
for y in x:
     t += y[1]

The extra variable?  The number of lines?

Cheers,

Evan @ 4-am






From aahz at pythoncraft.com  Sun Apr 13 11:24:45 2003
From: aahz at pythoncraft.com (Aahz)
Date: 13 Apr 2003 11:24:45 -0400
Subject: Cryptography (was Re: How to protect python code ?)
References: <3E8BE455.4000904@novagrid.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com> <OIqdnV1p3aljSw-jXTWcpg@comcast.com>
Message-ID: <b7bvft$os4$1@panix1.panix.com>

In article <OIqdnV1p3aljSw-jXTWcpg at comcast.com>,
Terry Reedy <tjreedy at udel.edu> wrote:
>
>Consider cryptography in general.  Try asking any security/spy agency
>for details on their cryptography methods.  Now notice that
>publications in the field mostly (all?) come from academics who do not
>actually do cryptography!  The science of keeping secrets secret is
>generally keep secret, as much as possible, by its practitioners.

Not that I'm disagreeing with your main point (about protecting Python
code), but cryptographic systems that are developed privately are much
easier to hack (historically, at least) than ones developed in public.
Anyone who cares about security should read comp.risks.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From jeethur at sancharnet.in  Mon Apr  7 14:07:23 2003
From: jeethur at sancharnet.in (Jeethu Rao)
Date: Mon, 07 Apr 2003 23:37:23 +0530
Subject: Comms Protocol design dilemma
Message-ID: <000001c2fd30$8c021740$7f88013d@JeethuRao>

Hi,
I've been working on a encrypted instant messenger server for
quite some time now. I've had to dump the code about 3 times
(I changed the protocols). And quite frankly I don't feel that
bad while throwing out code written in python, since its very
easy to rewrite the same functionality (Throwing away *hard written*
C code really breaks my heart.)

Now all the past three implementations are based on a packet
Oriented scheme <TCP/IP, Blowfish Encryption,16 byte packet hdr>.
My friends in here are suggesting me to use a line oriented protocol
(Like Microsoft Messenger), some have even suggested using XML
(Which I vehemently hate) like in Jabber. I'm of the opinion that
XML tends to waste a lot of bandwidth.

Since most of the data is binary, If I go in for line oriented protocol,
I'll have to base64 (or something similar) all the encrypted data,
Which will be an added overhead.

The primary dilemma is, What protocol is better suited, 'text based' or
'packet based' ?

The other problem is in the implementation.
The old implementations used a simple scheme, every incoming
Connection would be handled by a thread from a thread pool.
But now, the client is going to be connected to the server
All the time its online. Thus, a lot of clients means a lot
Of threads running in parallel, doing nothing. I'm not quite sure
If the Python threads waiting on sockets would yield as soon
as it gets its share of the VM (They're all serialized, right ?).
And since there's a lot of encryption/ decryption , hashing, random
number generation involved, I can pretty much rule out asynchronous
sockets. Is there any better strategy to tackle this problem?


Thanks,

Jeethu Rao





From ark at research.att.com  Wed Apr  9 18:52:31 2003
From: ark at research.att.com (Andrew Koenig)
Date: Wed, 9 Apr 2003 22:52:31 GMT
Subject: string() - Is this some kind of sick joke?
References: <36dbf824.0304091142.527af834@posting.google.com>
Message-ID: <yu99fzor9tcg.fsf@europa.research.att.com>

Daniel> Observe from the 2.2.2 interpreter (and 2.2.1 in HPUX):

     >>> y="'hello'"
     >>> import string
     >>> string.strip(y,"'")
     Traceback (most recent call last):
     File "<interactive input>", line 1, in ?
     TypeError: strip() takes exactly 1 argument (2 given)

On the other hand, the following works just fine:

     >>> y = "'hello'"
     >>> str.strip(y,"'")
     'hello'

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From aleax at aleax.it  Fri Apr 18 02:19:44 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 06:19:44 GMT
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de> <mailman.1050586468.29594.python-list@python.org> <rozna.115488$It5.70451@news2.central.cox.net> <b7mn1i$2icfr$2@ID-148859.news.dfncis.de> <LBBna.43583$T34.1210957@news2.tin.it> <b7nttn$2p743$3@ID-148859.news.dfncis.de>
Message-ID: <4OMna.30867$LB6.707286@news1.tin.it>

Haran Shivanan wrote:

> module "Alex Martelli" core dumped with the following output:
> [snipped informative stuff]
>> 
>> This prints 11, as apparently required by your specs.  I hope
>> this can help you make the custom metaclass you need to
>> implement your actual, exact specifications, of course.
> Thanks *a lot* for your help.

You're welcome!

> Do you know of any on-line resources where I can get info
> on metaclasses and using them?

In your shoes, I'd use Advanced Groups Search in Google,
looking for 'metaclass' in comp.lang.python -- I've posted
a few examples &c, and so have others.  Mertz and Simionato
have an article online, at
http://www-106.ibm.com/developerworks/linux/library/l-pymeta.html

Regular Google for:
   python custom metaclass 
shows the M & S article as well as archival copies of some 
of my posts all within the first 10 of 450 hits.  In general, 
whenever you are looking for information on a subject, a try
with google is worth it -- asking around is fine too, of
course, but google is truly helpful more often than not.


> I was only dimly aware of their presence until your post.

That's a widespread issue -- mostly because very few people
really NEED to build classes with peculiar semantics, and
that's what custom metaclasses are for, after all.  But when
one does have such a need, a custom metaclass should be the
first possibility considered -- and clearly such a need is
exactly what you believe you're facing here.


Alex




From gerrit.muller at embeddedsystems.nl  Wed Apr  2 01:39:07 2003
From: gerrit.muller at embeddedsystems.nl (Gerrit Muller)
Date: Wed, 02 Apr 2003 08:39:07 +0200
Subject: Problem with Idle
References: <435d40e0.0304011438.5b7b0595@posting.google.com>
Message-ID: <3E8A858B.5070606@embeddedsystems.nl>

Chris wrote:
> I have a module and a main program that imports it. When I make
> changes to the module in one IDLE window, and run the program in the
> main window, the changes are not picked up (I did do ctrl-s in the
> module window to save it). But if I go to the MS-DOS prompt and run
> the main program it does run the module with the latest changes to
> it??????

This behavior is a well known trap for new Python users. There are 2 
solutions:
1) call reload() for the changed module
2) dowload idlefork idle 0.9.2a (?), which will do all the reloads if 
you press F5 "Restart". This idle version will hopefully be part of the 
Python 2.3 distribution.

kind regards Gerrit

-- 
Gaudi systems architecting:
http://www.extra.research.philips.com/natlab/sysarch/




From aahz at pythoncraft.com  Thu Apr 17 10:48:27 2003
From: aahz at pythoncraft.com (Aahz)
Date: 17 Apr 2003 10:48:27 -0400
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <b7kd74$jqm$1@panix2.panix.com> <hr9s9v8pgqr154h15lsjbk1pjs5u5n6n65@4ax.com> <Inyna.27603$LB6.635831@news1.tin.it>
Message-ID: <b7merr$mmn$1@panix2.panix.com>

In article <Inyna.27603$LB6.635831 at news1.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>
>let me try to channel GvR (even though I'm quite a fan of one-line
>blocks myself, particularly for such idioms as "if <xxx>: break")...:
>
>Having MTOWTDI ain't Pythonic.  The ability to write either:
>
>    if finished: break
>
>OR quite equivalently
>
>    if finished:
>        break
>
>is therefore a wart.  If only one of the two must survive, it's
>clearly the second one (more general).  This will also ease the
>task of semi-bright editors (right now, my GVIM tries to help
>me by autoindenting right after each such "one-line block"...).

That's not the main reason, though.  The main reason is that Guido
believes that control structures should stand out clearly.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From a.schmolck at gmx.net  Mon Apr  7 09:22:07 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 07 Apr 2003 14:22:07 +0100
Subject: isearch, LEAP and patents  was: Re: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org> <yfsr88fkppr.fsf@black132.ex.ac.uk> <eaeka.1333$1h6.8933@news-server.bigpond.net.au>
Message-ID: <yfs1y0el9xc.fsf_-_@black132.ex.ac.uk>

[crossposted to comp.emacs; followups set]

[THE and LEAP(tm) [sic!] (http://humane.sourceforge.net/home/) seem rather
similar to emacs and isearch]

"Neil Hodgson" <nhodgson at bigpond.net.au> writes:

> Alexander Schmolck:
> 
> > He certainly knows about Emacs (have a look at the index of his book), but
> I
> > guess obtaining patents and attracting attention as some sort of visionary
> is
> > unnecessarily complicated by pointing out the prior art yourself. And
> although
> > the most crucial (and brilliant) idea, incremental search, is nicked (and
> > uncredited on his webpage),
> 
>    Priority for incremental search is uncertain - Jef Raskin has been
> pushing LEAP since at least the mid 80s with the Swyftcard and Canon Cat
> products. The earliest publication reference found in a quick google is
> November 1985. When did EMACS get incremental search?
> http://archaic-apples.shauny.de/files/cat/canon.html
> 
>    Neil

Emacs has been around since the late 70ies and the last time I looked around
on the web I came to the conclusion that swyftcard and canon cat came into
existence at a time when emacs already had incremental search. Given that I'm
not positively certain, however, I should have phrased this more carefully.

Maybe someone on comp.emacs can fill us in on isearch's origins and whether
raskin holds any patents that are relevant to emacs (he certainly seems to
either hold or have applied for patents concerning THE[1] and from what I gather
from his webpage pretty much everything that would seem original enough to get
patented is present in emacs)?

[1] Raskin holds a number of patents and has several relating to UIs pending.

 the legalese for THE states: 

 "Distribution of this in-progress software is for testing and illustrative
 purposes only and does not confer any rights, including rights to reproduce
 copyrighted material and patented inventions."

 (BTW, any idea why this clearly not open source project is hosted by SF?)

alex



From alloydflanagan at attbi.com  Fri Apr 25 17:10:54 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 25 Apr 2003 14:10:54 -0700
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>	<Kq9qa.9083$h%2.957726@read1.cgocable.net>	<mailman.1051273778.5621.python-list@python.org> <fzo6emyy.fsf@python.net> <mailman.1051277805.18035.python-list@python.org>
Message-ID: <a8b7f07a.0304251310.63a7b4b1@posting.google.com>

Gerhard Haering <gh at ghaering.de> wrote in message news:<mailman.1051277805.18035.python-list at python.org>...
<big snip>
> On Windows, try this:
> 
> C:\>python
> Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
> Type "help", "copyright", "credits" or "license" for more information.
>  >>> from wxPython.wx import *
>  >>> f = wxFrame(NULL, 1, "foo")
>  >>> f.Show()
> 
> This will segfault Python. IIRC, this was what I did in my very first 
> exploration of wxPython.
> 
> I believe the reason of this crash is that the wxWindows main loop isn't 
> initialized, yet. What I've done may be not very clever, but in my 

As a follow-up to my earlier reply, this seems to work well:

>>> from wxPython.wx import *
>>> from threading import *
>>> class wxThread(Thread):
...    def run(self, my_app):
...        my_app.MainLoop()
...
>>> x = wxThread()
>>> app = wxPySimpleApp()
>>> x.run(app)
>>> f = wxFrame(None, 1, "foo")
>>> f.Show()
1
>>> f.SetTitle("foobar")

etc.
You've got your application main loop running in a separate thread,
but it seems to 'see' f just fine, and window events get processed.



From a.schmolck at gmx.net  Mon Apr  7 20:16:06 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 08 Apr 2003 01:16:06 +0100
Subject: exceptions == errors?
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
Message-ID: <yfsistpkfnd.fsf@black132.ex.ac.uk>

Mark Harrison <mh at pixar.com> writes:

> Are exceptions always considered errors, or there the same feeling
> as in C++ that exceptions can be used for non-error situations?

No. Yes: iterators raise an StopIteration exception to signal that there are
no more elements left. Thus

  for x in container: print x
  print "done"

will print xs until a StopIteration is raised by the iterator returned by
container's __iter__ method, at which point execution quietly continues with
the ``print "done"`` statement.

> 
> I am asking because I am writing a network server that has the
> general outline
> 
> 	while 1:
> 		read request
> 		process request
> 		send response
> 
> So, I have the option to
> 
> 1.  embed the "send response" code in the request processing code

Sounds somehow bad.

> 
> 2.  Have the request processing code generate formatted responses
>     that are returned to the main loop
> 
> As is usually the case, adding robust error checking and reporting
> throws a kink into the works.   I have to do a lot of exception
> wrapping in the request processing code in order to ensure that
> there are no code paths where an exception short-circuits a
> response generations.
> 
> Therefore, I'm thinking of structuring the code like this:
> 
> 	while 1:
> 		try:
> 			read request
> 			process request
> 		catch MyException,e:
> 			send response contained in e
> 		catch Exception,e:
> 			send "internal error" response
> 
> Am I setting myself up for trouble if I follow this route?
> All advice gratefully received.


I don't fully understand your problem from your description, but maybe you
want some variation on this?


try: 
    # you'll likely want try: except: statements nested in here for error
    # handling
    read request
    process request
# irrespective of whether an error occured send a response
finally:
    send response



'as



From staschuk at telusplanet.net  Sat Apr 19 13:26:32 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 19 Apr 2003 11:26:32 -0600
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <Fy7oa.53694$T34.1483263@news2.tin.it>; from aleax@aleax.it on Sat, Apr 19, 2003 at 08:13:25AM +0000
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it>
Message-ID: <20030419112632.C587@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
  [...]
> "sum([y[1] for y in x])" would be very simple and direct, IMHO; the
> "same" thing expressed with "reduce(operator.add,[y[1] for y in x])"
> may only feel "simple and direct" to PhD's in maths -- it feels
> _abstruse_ to most people, in my opinion.

What if operator.add were changed to take any number of arguments?
Then you could write
    operator.add(*[y[1] for y in x])
Clear?  Abstruse?

Btw, I don't know about most people, but I don't have a PhD in
math, and reduce(operator.add, L) seems simple and clear to me.
(I do agree that there is a point beyond which abstraction brings
diminishing returns; but reduce is well inside that limit, imho.)

-- 
Steven Taschuk                            staschuk at telusplanet.net
"Our analysis begins with two outrageous benchmarks."
  -- Clinger et al., "Implementation strategies for continuations"




From mis6 at pitt.edu  Wed Apr  9 14:52:47 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 9 Apr 2003 11:52:47 -0700
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com> <b6soav$7p2$1@panix1.panix.com> <Qdpka.2889$Zu2.2627@news01.roc.ny.frontiernet.net> <a8b7f07a.0304080603.473a8d90@posting.google.com> <WRBka.134$jS3.16@news02.roc.ny.frontiernet.net>
Message-ID: <2259b0e2.0304091052.52d4c900@posting.google.com>

Lee Harr <missive at frontiernet.net> wrote in message news:<WRBka.134$jS3.16 at news02.roc.ny.frontiernet.net>...
> I guess I do not get it.
> 
> 
> >>> class A(object):
> ...   def f(self):
> ...     print 'Af'
> ... 
> >>> class B(object):
> ...   def f(self):
> ...     print 'Bf'
> ... 
> >>> class C(A, B):
> ...   def f(self):
> ...     super(C, self).f()
> ... 
> >>> c=C()
> >>> c.f()
> Af
> 
> 
> Should this be showing
> Af
> Bf
> ?
> 
> How would I even get that output?  I know I can do:
> 
> class C(A, B):
>   def f(self):
>     A.f(self)
>     B.f(self)
> 
> 
> but I can't seem to get that out of super()

If metaclasses don't scare you, here there is the solution:

class MakeCooperative(type):
    "A custom metaclass making cooperative a given list of methods"
    def __init__(cls,name,bases,dic):
        "Makes cooperative the methods in the method list"
        for meth in cls.methods:   
            setattr(cls,meth,cls.coop_method(meth,dic.get(meth)))
    def coop_method(cls,name,method):
        "Given a regular method, returns a cooperative method"
        def _(self,*args,**kw):
            try: supermethod=getattr(super(cls,self),name)
            except: pass # do nothing
            else: supermethod(*args,**kw) # calls the supermethod
            if method: method(self,*args,**kw) # calls the method
        return _
    def method(meta,methods=''):
        """Generates the list of cooperative methods from an input string
        and returns the metaclass"""
        meta.methods=methods.split()
        return meta
    method=classmethod(method) # this is a metaclass-level classmethod
    
class A(object):
    __metaclass__ = MakeCooperative.method('f')
    def f(self):
        print 'Af'

class B(object):
    def f(self):
       print 'Bf'
       
class C(A, B):
    def f(self):
        print 'Cf'

c=C()
c.f()

# Output:
# Bf
# Af
# Cf



From rmunn at pobox.com  Sat Apr  5 12:56:13 2003
From: rmunn at pobox.com (Robin Munn)
Date: Sat, 05 Apr 2003 17:56:13 GMT
Subject: Input from a file as a command line argument
References: <3E8EDBDA.7020701@pbezone.net>
Message-ID: <slrnb8th5u.ldu.rmunn@localhost.localdomain>

Teemu Luojola <teemu.luojola at pbezone.net> wrote:
> I have read through Python documentation and tried several searches, but 
> I can't find a simple answer to this question.
> 
> If I make an executable script in linux environment, how can I pass 
> input to that script from a file on command line? Like the following:
> 
> $ script.py input.txt

In this case, the filename would be found as sys.argv[1]. (You'd have to
import the sys module in your code, of course). You would want to do
something like:

    try:
        filename = sys.argv[1]
    except IndexError:
        # No parameters at all: print usage information and exit
        usage()
        sys.exit(1)
    input_file = file(filename, 'r')

A couple words of explanation of the above code. First, sys.argv is a
list that works very much like the argv[] array in C code: sys.argv[0]
will be the name of the current process (usually the filename of your
"main" Python source file, but that's not 100% guaranteed), sys.argv[1]
will be the first command-line parameter, etc. There is no sys.argc
because you can get similar results by doing len(sys.argv). Beware --
len(sys.argv) will always be equal to (number_of_parameters + 1) because
sys.argv[0] is guaranteed to contain something.

Now: the "try: ... except IndexError:..." block is an example of Python
exception-handling. IndexError gets thrown if you try to access an index
that's off the end of an array. In this script, that will only happen if
the script was run with no arguments at all, in which case we should
print out some usage information and exit with an error code.

Next I create a file object with a call to file() (you can also call
open() -- the names "file" and "open" both refer to the same function).
I prefer to call file(), though, because it reminds me of a very
important point: do *not* use the name "file" for your own use! It will
cause subtle and unexpected bugs later in your code. Use names like
"input_file" or "output_file", which are more descriptive anyway.

By the way, if any of this seems insultingly obvious to you, I
apologize. I figure it's better to explain too much and be redundant
than run the risk of explaining too little and leaving the other person
just as confused as before.

> or perhaps
> 
> $ script.py < input.txt

In this case, the file would be available as sys.stdin, and it would
already be open, so you wouldn't need to call file().

> I make now a comparison to a Perl script. In Perl I would write the 
> following to handle the input (line by line) from a file:
> 
> while (<>) {
> 	## (do something to $_)
> }

In Python that would be done as:

    for line in input_file:
        # Do something with line

or, if you are using stdin:

    for line in sys.stdin:
        # Do something with line

> and I can execute the script simply with
> 
> $ script.perl < input.txt
> 
> and it would process each line in input.txt.
> 
> How would a similar Python code look like?

There you go. One last caveat: it's been over a year since I did any
Perl programming, so I don't remember whether the "while (<>)" idiom in
Perl strips newlines off the end of lines or not. In Python, newlines
are *not* stripped of the end of lines in the for loop. So take that
into account in your processing: if you want to strip the newlines off
the end, you might want to do something like this as the first line of
your loop:

    for line in sys.stdin:
        line = line.rstrip('\n')
        # Further processing

Hope this helps.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From peter at engcorp.com  Wed Apr  2 10:22:24 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 10:22:24 -0500
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com> <b6esse$pge$1@heckle.cc.uregina.ca>
Message-ID: <3E8B0030.E3B0DE00@engcorp.com>

sik0fewl wrote:
> 
> Ben Hutchings wrote:
> > In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer wrote:
> > <snip>
> >
> >>    3.  At most three Is, Xs and Cs may appear in any given literal.
> >
> > <snip>
> >
> > OK cleverclogs, how do we write 39?
> 
> Although I honestly don't know how, I'm sure it's possible. Here's how I
> would do it, although it's probably not the right way. My roman numerals
> are a bit rusty. :)

This was a frequent problem with the early generations of Roman numerals,
but newer technology now allows production of stainless Roman numerals,
which merely acquire a very light layer of surface oxidation which then
prevents further degradation and flaking as with rust.

I believe Google could provide more detail with a search using 
"stainless roman numerals" or something like that.

-Peter



From aleax at aleax.it  Tue Apr 15 11:15:57 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 15:15:57 GMT
Subject: files: direct access - NOT sequential
References: <mailman.1050416124.6482.python-list@python.org>
Message-ID: <NmVma.17291$LB6.442904@news1.tin.it>

<posted & mailed>

Patrick Carabin wrote:
   ...
> how do i have to manage it to keep the contents that are already in
> the file ? Do You have some example?

As I already posted, mode 'r+' is what you need if the file already
exists, in order to keep its previous contents except when you
specifically overwrite them ('r+b' if it's a binary-file).


Alex
 



From aahz at pythoncraft.com  Thu Apr 17 02:25:52 2003
From: aahz at pythoncraft.com (Aahz)
Date: 17 Apr 2003 02:25:52 -0400
Subject: PythonCard vs. Boa Constructor (was Re: Python for VB6 programmers)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com> <6e07b825.0304162200.34884d36@posting.google.com>
Message-ID: <b7lhdg$q0f$1@panix3.panix.com>

In article <6e07b825.0304162200.34884d36 at posting.google.com>,
yaipa h. <yaipa at yahoo.com> wrote:
>
>Here in the SF Bay Area I have only met one VB programmer willing to say
>that he uses VB.net. The other VB programmers I find are talking about
>either sticking with VB6 (most), taking Java courses (some) or playing 
>around with the C# idea (few).  Most share my feelings about the .net
>thing. That it will not be stable until release +3.x and there are 
>little to no existing libraries. You know, batteries. ;-) So yes there 
>is a window and it is wide open.
>
>With that said, without a slick, stable GUI builder with all the 
>look and feel of the VB GUI builder it's a long way from here
>to there. I mean what do you expect them to do, pass in command-line 
>arguments or hand assemble TK widgets. :^) 

Good point.  Anyone here willing to take a whack at claiming that
PythonCard, Boa Constructor, or something else would the best path for a
VB6 programmer?
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From cben at techunix.technion.ac.il  Tue Apr 15 07:21:40 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Tue, 15 Apr 2003 14:21:40 +0300 (IDT)
Subject: Catching Save - newbie
In-Reply-To: <3e9bd039_2@news.estpak.ee>
References: <3e9bd039_2@news.estpak.ee>
Message-ID: <Pine.GSO.4.51.0304151405060.2075@techunix.technion.ac.il>

Vahur Lokk wrote on 2003-04-15:

> Hi!
> What I want to do is write a small daemon that can catch Save File type of
> events in the system and manipulate files being saved.
> Having mostly sorted out the rest of the program I cannot think of a way to
> catch these saves.
>
Do you mean saves from inside your python program or any saves done by
other programs running on your computer?

In the first case, you can try to replace the builtin `file` function
(and its `open` alias) to return a wrapper that does what you need
when one calls the `.write()` and `.close()` methods of it.
Subclassing `file` sounds like a good way for this.

In the later case, there are two basic approaches:

1. Poll the file system for changes.  libfam_ will do it better than
   you can probably bother to write.  In any case you must realize
   that some updates will be missed and notifications might be
   delayed.  There is no way to get reliable update notifications on
   unix (perhaps there is on windows).

.. _libfam: http://oss.sgi.com/projects/fam/

2. Create a new filesystem driver (or modify existing one) that will
   communicate notifications to your user-land program and replace
   your filesystem with it.  This can be made reliable but is very
   much work and a huge hassle to install.

In any case, you'll have to define what you mean by "Save File"
events.  Under the hood programs can modify files continuosly; at best
you can aim for the close event (not in approach 1 of case 2).  Also
note that you probably care also for rename actions - most
self-respecting editors save files by writing a new file and then
atomically renaming over the old version (or some variant thereof).

Perhaps you should rethink your goals.  Why would you want to catch
file modifications and how do you want to manipulate them?  Perhaps
there are better ways?  What is your real goal?

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From sholden at holdenweb.com  Wed Apr 16 09:19:18 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 13:19:18 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <mailman.1050453385.17091.python-list@python.org> <b7iigt$1d3jt$1@ID-169208.news.dfncis.de> <slrnb9pt59.iin.ramen@lackingtalent.com>
Message-ID: <qLcna.110588$It5.96231@news2.central.cox.net>

"Dave Benjamin" <ramen at lackingtalent.com> wrote in message
news:slrnb9pt59.iin.ramen at lackingtalent.com...
> In article <b7iigt$1d3jt$1 at ID-169208.news.dfncis.de>,
> Greg Ewing (using news.cis.dfn.de) wrote:
> > There isn't any widely understood name for what Python used
> > to do, because no other language has ever been quirky
> > enough to do it that way. It was lexical, but with only
> > three levels visible at any one time.
>
> Three? I thought it was two! Locals, globals, and...?
>

... builtin

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From a.schmolck at gmx.net  Wed Apr  9 16:04:29 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 09 Apr 2003 21:04:29 +0100
Subject: Macros in Python?
References: <b71qde$q71$07$1@news.t-online.com>
Message-ID: <yfsy92jh1yq.fsf@black132.ex.ac.uk>

Dominic <oblivious at web.de> writes:

> 
> Is something like this feasible?
> 
> def server(a,b):
>    if cond:
>        ...
>    else:
>      With_Root_Privs:
>         if x in user:
>            active.append(x)
>           (...)
> 
> Ciao,
>   Dominic
> 

No. you can only do something like this:

def withRootPrivs(fun):
    uid = geteuid
    setuid(0)
    try:     return fun()
    finally: setuid(uid)

def server(a,b):
    if cond:
       ...
    else:
       def useRootPrivs():
           if x in user:
              active.append(x)
           (...)
       withRootPrivs(useRootPrivs)

(well, technicall you could screw around with the source code of the function
but that way lies insanity)

'as



From duncan at NOSPAMrcp.co.uk  Thu Apr 24 04:34:38 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 08:34:38 +0000 (UTC)
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <3EA5B173.9420324C@alcyone.com> <mailman.1051056037.8077.python-list@python.org> <3EA5DE1E.22704437@alcyone.com> <m3r87tdbo3.fsf@mira.informatik.hu-berlin.de> <mailman.1051081416.21801.python-list@python.org> <m34r4pgh33.fsf@mira.informatik.hu-berlin.de>
Message-ID: <Xns93676070D39E9duncanrcpcouk@127.0.0.1>

martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=) wrote in 
news:m34r4pgh33.fsf at mira.informatik.hu-berlin.de:

> Ian Bicking <ianb at colorstudy.com> writes:
> 
>> Note that the idiom I've been using (and which I've seen elsewhere) is:
>> 
>> True, False = 1==1, 0==1
>> 
>> Which retains proper boolean objects.
> 
> For Python 2.x, this does not really matter much. Yes, you do get
> references to Py_True/Py_False, but they are still just integer
> objects with the value 1/0.

Not entirely accurate. In Python 2.3, True and False have type bool which 
is a subclass of int, so this could make a difference.

Even in Python 2.1 & 2.2 there can be a subtle difference. If, for example, 
you use the win32 extensions to access COM objects, then calling a COM 
method that accepts a variant type with 1 as a parameter will pass an 
integer, but 1==1 will make it pass a boolean. The win32 com support did 
some strange tricks with object identity to tell integers and booleans 
apart.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From nav at adams.patriot.net  Tue Apr 29 12:02:56 2003
From: nav at adams.patriot.net (Nick Vargish)
Date: 29 Apr 2003 12:02:56 -0400
Subject: Python Tutorials
References: <15451.8030105@earthlink.net>
Message-ID: <yyyznm9thov.fsf@adams.patriot.net>

That's a great resource, and points me to several tutorials for topics
I keep meaning to learn more about, wxPython for example.

However, all the <font size = 6> tags are... umm... just not conducive
to readability. It makes your list text bigger than the headlines on
most other web pages.

Thanks for assembling the list,

Nick

-- 
#include<stdio.h> /* SigMask 0.3 (sig.c) 19990429 PUBLIC DOMAIN "Compile Me" */
int main(c,v)char *v;{return !c?putchar(*v-1)&&main(0,v+ /* Tweaks welcomed. */
1):main(0,"Ojdl!Wbshjti!=obwAqbusjpu/ofu?\v\1");}  /* build: cc -o sig sig.c */



From dirk at gerrits.homeip.net  Wed Apr  9 15:54:19 2003
From: dirk at gerrits.homeip.net (Dirk Gerrits)
Date: Wed, 09 Apr 2003 21:54:19 +0200
Subject: Odd behaviour of dir(object) with metaclasses.
Message-ID: <b71toc$k7m$1@main.gmane.org>

Hi,

I am a bit puzzled by the behaviour of the dir builtin. I'll illustrate 
this with the following Python shell session:

 >>> class MetaBase(type):
	def mb(self):
		print self
 >>> class MetaDerived(MetaBase):
	def md(self):
		print self
 >>> class Base(object):
	__metaclass__ = MetaBase
	def b(self):
		print self
 >>> class Derived(Base):
	__metaclass__ = MetaDerived
	def d(self):
		print self
 >>> b = Base()
 >>> d = Derived()
 >>> def not_magic(str):
	return not (str.startswith("_") or str.endswith("_"))
 >>> filter(not_magic, dir(b))
['b']
 >>> filter(not_magic, dir(d))
['b', 'd']

Nothing out of the ordinary here, just as I would expect. But lets take 
a look at the classes now:

 >>> filter(not_magic, dir(Base))
['b']
 >>> filter(not_magic, dir(Derived))
['b', 'd']

Perhaps its just me, but I think this is really odd. The way I would 
expect it, dir should list the accessable attributes of the object. And 
that would mean:
(1) its own attributes
(2) its class's attributes and recursively of its class's bases
(3) (if it's a class object) recursively the attributes of its bases

For b and d, the dir listings are clearly correct. But Base's and 
Derived's the dir listings are incomplete, they don't seem to take (2) 
into account:

 >>> Base.mb() # No AttributeError
<class '__main__.Base'>
 >>> Derived.mb() # Idem
<class '__main__.Derived'>
 >>> Derived.md() # Dito
<class '__main__.Derived'>

Perhaps I just don't fully understand what dir does. The first bit in 
the docs says:

     Return an alphabetized list of names comprising (some of) the
     attributes of the given object, and of attributes reachable from it:

This would seem to correspond with my feeling of what its supposed to 
do, but doesn't. However, reading further, we see that the current 
behaviour is indeed as it is documented:

     No argument:  the names in the current scope.
     Module object:  the module attributes.
     Type or class object:  its attributes, and recursively the
         attributes of its bases.
     Otherwise:  its attributes, its class's attributes, and recursively
         the attributes of its class's base classes.

Why is there an unnatural (IMHO) distinction between type/class objects 
and other non-module objects? I assume that a class's class is not 
considered become dir comes from the pre-metaclass era? Shouldn't the 
docs really read

     Return an alphabetized list of names comprising (some of) the
     attributes of the given object, and of attributes reachable from it:

     No argument:  the names in the current scope.
     Module object:  the module attributes.
     Otherwise:  its attributes, its class's attributes, and recursively
         the attributes of its class's base classes. And if it's a class
         object, also the attributes of its own base classes recursively.

?
(Or something with the same meaning but in a less confusing wording. ;))

I hope you'll be able to clarify these things for me a bit.

Cheers,
Dirk Gerrits






From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Sat Apr  5 14:35:19 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Sat, 05 Apr 2003 21:35:19 +0200
Subject: distributed computing implementations
In-Reply-To: <mailman.1049306344.1908.python-list@python.org>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <b6f66v$28k$1@slb3.atl.mindspring.net> <mailman.1049306344.1908.python-list@python.org>
Message-ID: <3e8f2ff5$0$49103$e4fe514c@news.xs4all.nl>

Jp Calderone wrote:
> On Wed, Apr 02, 2003 at 10:25:20AM -0700, Andrew Dalke wrote:
> 
>>robin:
>>
>>>If you want the biggest, boldest approach and don't care about
>>>overhead, use CORBA.
>>
>>What overhead would this be?  From what I see of omniORB,
>>there isn't really that much.  Also, CORBA is the most complete,
>>eg, it allows callbacks and passing around object references
>>the others don't have.
>>
> 
> 
>   I'm not sure about XML-RPC, SOAP, or Dopy, but I know Twisted Spread can
> pass references around like this (and you can tell it just -how- you want
> them passed), and I think Pyro can too.

Certainly. You can pass around the actual Pyro proxy objects, without 
bothering about object location, ID, etc, or pass around the object's UID.

Pyro also allows callbacks, although you have to do a little bit of extra 
coding to enable them.


>>I've wanted to do something in CORBA for years.  I've
>>never gotten there.  One problem is that I'm used to Python,
>>where I don't need to describe the interface beforhand.
>>CORBA wants that IDL, and a change in the object's interface
>> must be reflected in the IDL.  That just seems tedious to
>>me now.
>>
> 
> 
>   Right.  I think this is at least part of the overhead Robin was talking
> about.  Several of the other schemes don't require this.

> 
> 
>>I also do nearly everything in Python, so don't need the
>>ability for different langauges to interoperate.  I just pass
>>around Python objects.

I'd say: use Pyro (but I'm biased ofcourse ;=)
You won't have to specify an interface other than your regular
Python class, and it's designed for a Python-only environment.

--Irmen de Jong.




From aleax at aleax.it  Sun Apr 20 19:07:19 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 20 Apr 2003 23:07:19 GMT
Subject: Good Python generators example?
References: <paDoa.1611$hT2.1228371@news2.news.adelphia.net>
Message-ID: <HKFoa.40060$LB6.906470@news1.tin.it>

Robert Oschler wrote:

> Can someone point me to a nice concise Python 2.2 compatible example of
> the use of generators?

You might like my tutorial on iterators and generators at:

http://www.strakt.com/docs/eup02_itgen_alex.pdf

the example I'm thinking of is an external mergesort based on generators
(just in case you must ever sort files larger than memory...;-).


Alex




From pobrien at orbtech.com  Sun Apr  6 10:59:10 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 09:59:10 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <m3r88g9iqt.fsf@mira.informatik.hu-berlin.de>
Message-ID: <m2fzovaczl.fsf@localhost.localdomain>

martin at v.loewis.de (Martin v. L?wis) writes:

> pobrien at orbtech.com (Patrick K. O'Brien) writes:
> 
> > If you were crazy enough to think that you could create a program
> > along the lines of Emacs, but written in Python, how would you go
> > about it?  How would you design the domain model for files, buffers,
> > windows, and frames?  How would you allow the same level of
> > customizability?  How would you map functions (or methods or whatever)
> > to keys?  Any thoughts?
> 
> This problem is underspecified, atleast for the actual implementation
> strategy that follows: Do you need to support both terminal and GUI
> mode? Multiple GUI libraries? Do you need API compatibility with Emacs?

I only want to support the wxPython GUI and use Scintilla.  But I want
to wrap Scintilla so that no matter what methods Scintilla provides
the primitives used to customize PyAlaMode look like Emacs:

  forward_char
  forward_line
  forward_paragraph
  ...

I want the API and behavior to be similar enough to Emacs that an
elisp programmer would feel comfortable with the PyAlaMode API.  But I
also want Python programmers to be comfortable with it as well.  So
strict compatibility with Emacs is not the goal, simply close enough
to benefit from the lessons learned by Emacs and to attract Emacs
users and developers to PyAlaMode.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From s.thuriez at laposte.net  Mon Apr  7 11:01:30 2003
From: s.thuriez at laposte.net (sebastien)
Date: 7 Apr 2003 08:01:30 -0700
Subject: FTP : get  dir name with space caracters embedded  + is ther something quicker than LIST
References: <b1ab591.0304030143.52422a3e@posting.google.com> <459f9b68.0304041432.3451f57b@posting.google.com>
Message-ID: <b1ab591.0304070701.11ce900f@posting.google.com>

matt at teaphoo.fsnet.co.uk (Matt Russell) wrote in message news:<459f9b68.0304041432.3451f57b at posting.google.com>...

Hi to you all,


* For getting the listing of a directory, I do not feel any difference
in speed in between ftp.nlst ("/") and ftp.retrlines("LIST",
list.append). My connexion is extremly slow. This may explain why.
Also using nlst, I do not know which is a directory or a file. I have
trouble then to walk the directory structure in search for my file.

* I do not want to use the os.path.walk since this program send
several thread to each of the directory. I feel this recursive walk is
simple enough not to use os.path.thread.

* For names wich includes spaces caracters, I think that the advices
that you gave me will do the trick. I was just hoping for something
more general. But since I will work on a the same couple of servers it
will do the trick.

Thanks for yours quick responses.

Sebastien.



> I think FTP.nlst([ directory ])
> solves your problem.
> you may be able to use this method in conjunctio n with os.path.walk
> to achieve your goal.
> Matt
> 
> s.thuriez at laposte.net (sebastien) wrote in message news:<b1ab591.0304030143.52422a3e at posting.google.com>...
> > Hi ,
> > 
> > I have 2 problems regarding FTP methods.
> > 
> > 
> > My GOAL :
> > ---------
> > 
> > I am looking for a specific file on a huge ftp server. The structure
> > of the ftp server is not known. I have to determine which are
> > directories and which are files so that it can recursively go into the
> > subdirectories to scan the files (it is multi threaded).
> > 
> > It is working well but I have 2 problems.
> > 
> > 
> >  My trouble :
> > -----------------
> > 
> > It am using the ftp.retrlines('LIST',liste.append)
> > and analyse which one are directories since they have the letter d at
> > the beginning of the line.
> > 
> > 1) The LIST METHOD IS EXTREMLY SLOW !!!
> > 
> > In certain directories there is thousand of file and the LIST method
> > is extremely slow.
> > It works but is there other method to get the directories without
> > having to do a LIST ?
> > 
> > 2) DIRECTORY NAME WITH SPACES  
> > 
> > LIST is slow but it works. One other problem is the directory name in
> > which there is spaces : like
> > "data file today".
> >  I can do a :
> > elements= line.split(" "), 
> > 
> > but the data before my directory name has not the same format : for
> > example :
> > 
> > .....
> > may 22   2001     name of directory
> > mar 15    02:00   other directory name
> > .....
> > 
> > I do not know how to get the directory name simply.
> > 
> > One method would be to get a len(elements) and if len(elements) > 9
> > for example on my server. All the last elements would be part of
> > directory name. It is not so clean since it may depend on the FTP
> > server configuration.
> > 
> > Do you have any idea ?
> > 
> > Thanks for having taken the time to read my mail.
> > 
> > Sebastien.



From martin at v.loewis.de  Wed Apr 30 14:28:53 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 30 Apr 2003 20:28:53 +0200
Subject: Building Python 2.2.2 - configure error
References: <65a7284e.0304291557.6b3e437f@posting.google.com>
Message-ID: <m365ovhmai.fsf@mira.informatik.hu-berlin.de>

nraavi_news at yahoo.com writes:

> ALL these symbols exist in the library. 

Why do you think this? It appears that NONE of the symbols exist in
the library, or else there would be no error message.

> cc -o conftest conftest.c -I/opt/tcl/include -L/opt/tcl/lib -lm
> -ltcl8.4 -ltk8.4 -lpthread

Perhaps you need to link with -lX11 to get the X functions?

Regards,
Martin



From aleax at aleax.it  Thu Apr 17 12:58:58 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 16:58:58 GMT
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com> <hr9s9v8pgqr154h15lsjbk1pjs5u5n6n65@4ax.com> <Inyna.27603$LB6.635831@news1.tin.it> <5qit9v4pcrn4haj40gfi61egrg1pi0cg7d@4ax.com>
Message-ID: <m3Bna.43429$T34.1204317@news2.tin.it>

Courageous wrote:

> 
>>Having MTOWTDI ain't Pythonic.  The ability to write either:
>>
>>    if finished: break
>>
>>OR quite equivalently
>>
>>    if finished:
>>        break
>>
>>is therefore a wart.
> 
> Hrm. Conceputalization? I consider these two equivalent. I suppose

They ARE -- I DID say "quite equivalently" in the snippet you
just quoted.  But, ones can be seen as conceptualizing the
"guarded break" as a single unit, the other twists the knife
in the wound of the fact that there IS no such single unit, so
that the if and the break are, alas, separate statements.

> it's bias. Years of using languages where WSDM (white spaces don't
> matter), allows me to elide away some under some circumstances. :)

Sure -- that's just like saying that:

    if x > 23:

and

    if 23 < x:

are fully equivalent at a reasonable level of abstraction.  Of
course they are.  But most people are more used to the first form
(variable-operator-constant appears to be a more ofter used
sequence than constant-operator-variable) so they may read it
more smoothly, while the second form requires a fraction of a
second's worth of attention -- just because it's very slightly
less usual, not because of any intrinsic difference.


> BTW, there's a thing I do in some programming environments, that
> I personally really like to do, but I see few other people do. I'll
> show you:
> 
> if (alpha)       { DoSomething();        return x; }
>     if (beta)        { DoSomethingElse();    return y; }
>     if (delta)       { DoOtherThing();       return z; }
> 
> I like to use *columns*, in other words. I find this to be neat and
> readible, and considerably superior to what would otherwise pass as
> the defacto alternative.
> 
> I think you'll see how the example applies to the thread's subject...

Of course (though it definitely doesn't apply to Python!-).


Alex




From carsten at gehling.dk  Wed Apr 30 05:31:38 2003
From: carsten at gehling.dk (Carsten Gehling)
Date: Wed, 30 Apr 2003 11:31:38 +0200
Subject: SV: set cookie problem (slightly OT)
In-Reply-To: <wiGra.5646$%_3.3813575@newssrv26.news.prodigy.com>
Message-ID: <BHEFKIOIMEDGILGDACKDEEDGDFAA.carsten@gehling.dk>

> Fra: python-list-admin at python.org
> [mailto:python-list-admin at python.org]P? vegne af drs
> Sendt: 30. april 2003 04:13
> Emne: set cookie problem (slightly OT)
>
>
> I have a web directory (apache2/FreeBSD) with ssi enabled but not cgi.  in
> order to run simple python scripts i use something like
>
> <!--#exec cmd="/usr/local/bin/python /path/to/script.py" -->
>
> which works surprisingly well.  I cannot figure a way to set a cookie,
> however, as i am not able to add to the header. Anyone have any
> suggestions?

Normally, Apache (and other webservers like IIS) pre-process headers UNLESS
you prefix the script name with "nph-". Try to do this:

<!--#exec cmd="/usr/local/bin/python /path/to/nph-script.py" -->

And - of course - rename your script accordingly. I haven't tried it, but it
should give you greater control with the headers.

- Carsten





From skip at pobox.com  Sun Apr  6 15:22:47 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sun, 6 Apr 2003 14:22:47 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <b6pri6$drf$1@panix2.panix.com>
References: <m2u1dcw22s.fsf@localhost.localdomain>
        <m2k7e7addl.fsf@localhost.localdomain>
        <b6pgcm$1t3$1@panix3.panix.com>
        <m27ka7aae0.fsf@localhost.localdomain>
        <b6pri6$drf$1@panix2.panix.com>
Message-ID: <16016.32391.485749.650299@montanaro.dyndns.org>

    aahz> Bare emacs is unusable.

I'm not sure what you're referring to.  No Emacs installation I'm aware of
is "bare".  (I presume you're not referring to the predump "temacs"
executable.)  I find "bare" vi pretty darn inscrutable as well.  I know how
to 'dd', 'x', 'i', 'a' and ':wq[!]', and can navigate with the arrow keys.
Just enough to tweak the occasional config file.  Anything more substantial
requires Emacs in some fashion (even if it's just X11 copy/paste from an
Emacs buffer running on another machine).  It all boils down to what you're
productive with.  I began using Emacs more than 20 years ago, so at this
point any editor in which typing 7-bit ASCII doesn't result in those
characters winding up in the file I'm editing seems weird.

I can actually understand the vi/emacs standoff.  They represent two
similarly powerful ways of editing files.  What bugs the hell out of me is
watching someone stubbornly refuse to switch from /bin/sh.  They type long
shell commands with no command/file completion or command recall of any
kind, even if bash or tcsh is installed and available.  You get stuck
standing there watching them type and make mistakes, then have to type the
whole damn thing over again with corrections.  Makes you want to scream,
"Just move over and let me drive you fool!  I have better things to do with
my time!"

Skip




From pobrien at orbtech.com  Mon Apr  7 10:14:35 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 07 Apr 2003 09:14:35 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <7xn0j3fsdu.fsf@ruckus.brouhaha.com> <Yseka.1370$1h6.8783@news-server.bigpond.net.au>
Message-ID: <m2fzou5r90.fsf@localhost.localdomain>

"Neil Hodgson" <nhodgson at bigpond.net.au> writes:

> Paul Rubin:
> 
> > I think you're going to need some C extensions to represent buffers
> > similar to how Emacs does it.  It's hard to move blocks of characters
> > around in Python with the array module without constantly re-allocating
> > the arrays.
> 
>    The only addition to array that is really needed to provide all the
> building blocks used in Scintilla is a method for moving blocks around in an
> array. Others that would be nice to have but are not as important are fast
> ways to add space to the end of an array and to copy between arrays.
> However, the real time sink in a GUI editor is most commonly in measuring
> and drawing the text, not in the speed of the underlying text storage data
> structures.

And I'm more than happy to delegate all the display issues to
Scintilla.  ;-)

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From olczyk at interaccess.com  Fri Apr 18 18:05:31 2003
From: olczyk at interaccess.com (Thaddeus L Olczyk)
Date: Fri, 18 Apr 2003 22:05:31 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <b7m546$2a83$1@news.peterlink.ru> <lyptnlxq1n.fsf@cartan.de> <m24r4xdyd6.fsf@mycroft.actrix.gen.nz> <lyd6jlxkpi.fsf@cartan.de> <usft9v01l8fcdg8erdl2kishg0cnr1t9dp@4ax.com> <lyadepp1ck.fsf@cartan.de>
Message-ID: <1lt0avslergps3bdv1lpvr48kef6t03640@4ax.com>

On 17 Apr 2003 17:53:15 +0200, Nils Goesche <cartan at cartan.de> wrote:

>Thaddeus L Olczyk <olczyk at interaccess.com> writes:
>
>> On 17 Apr 2003 16:27:53 +0200, Nils Goesche <cartan at cartan.de> wrote:
>> 
>> >A definition that seems sensible to me would be something like: A
>> >language is functional if it somehow contains the lambda calculus.
>> >If you can take lambda calculus formulae and readily translate them
>> >into working code.
>
>> That makes nearly every modern language functional.
>
>Certainly not Java or C# ;-) 
Yes. Both contain lambda calculus. As does any language
which supports objects.

In ( a sort hand wavy way ),
class Lambda
{
public:
        retval exec(arglist){...}
}

Java ( can't say I know enough about C# ) may lack some syntactic
sugar to make it pretty, but it's there.

>But even if that's true -- then so be it.
>If it is true that modern languages tend to be functional, too, then
>you could call that progress.  Why should we change the definition of
>``functional?? every year in order to keep the ratio of functional to
>non-functional languages constant over time?
>
While it's true that in math and CS, many definitions evolve over
time ( ones that come to mind right now are function/procedure
[which were distinct, now considered the same thing], fiber bundles,
sheaves, topological space ... ), that's all just bullshit thrown out
to distract people from *your stupidity*.

That's because the definition that you have suggested is not some
formal definition appearing in seminal "Introduction to Functional
Programming" type books or papers. It's *your* definition, and it
sucks. Because it's your definition and not a formal one, there is
no suggestion that I am changing the definition over time. As I said,
it's simply a distraction to hide the fact that your definition
*sucks*.

As for your definition, it sucks because it fails to make any
distinction. Almost everyone would agree that languages like
Haskell, ML, Clean, Scheme and  ( sometimes ) Lisp are functional.
And your definition includes those. But your definition would say that
languages like C, Eiffel and Smalltalk ( not to mention C++, Java and
C# ) are functional, which almost no one would agree with.

>> >And you have to be a bit relaxed when using this requirement; for
>> >instance, writing an Y-combinator in ML is very hard because of the
>> >stupid type checker.
>> 
>> No it is easy.
>> 
>> In OCaml, 
>>     let rec fix f x=f (fix f) x;;
>
>That's cheating and you know it :-)
>
No. It's not.
It's a standard definition of Y in a combinatorial approach
to computing.

>> For a more traditional approach.
>> type 'a rf = RF of ( 'a rf -> 'a );;
>> 
>> let y f x= 
>>   let g (RF h) z=
>>       f( h (RF h) )  z in
>>   (g (RF g)) x;;
>> 
>> Of course the example above looks like:
>>    let fact x= y ( fun q->
>>                              (fun n-> 
>>                                 if n<2 then 1 
>>                                 else n* (q (n-1)))) x;;
>> 
>> When I did it this way the first time it was a bit difficult getting
>> the type right.
>
>I know: Been there done that ;-\  Took me some time.  
What can I say. Some people are slower than others.

>Certainly /much/
>longer than simply translating the lambda calculus formula to Lisp, so
>I refuse calling it easy, even if the ML code that finally works is
>short.
And know we see why some people are slower.
I never even thought of "translating" the code.
I did what every one is taught in every math class.
I rederived it in ML. In the process I found it easy.
The only sticking point was that I did not understand
user defined types well enough to write HOFs. Once
I understood that it was trivial.

>
>> But the reason was that I was new to the OCaml typing system. It
>> wasn't what the type should look like, it was just getting the
>> declaration right.
>
>If it's so easy, why did they even bother putting it into the ML FAQ?
Because there are lazy people who don't write code from first
principles, or bother to learn enough about user defined types
to write HOFs ( the true FP practitioner doesn't just use HOFs
he writes them too ).

>
>> While not an expert, I should say this type of infinite cascading
>> type and data is typical of a class of ML programs. I believe the
>> little MLer has an example early on.
>> 
>> This should also be easy for a Lisper/Schemer to grasp as it just
>> extends on some of their ideas.
>
>Sure, once you know how to do it, it seems easy and obvious.  But that
>is true about almost all problems.
No. It's easy and obvious because I learned enough OCaml to make it
easy and obvious. These are aspects of the language which are
fundamental, and anyone taking a professional approach to programming
in ML should be familiar with it enough to understand how to write
HOF's. It's only hackers who have a hard time implementing Y in ML.


--------------------------------------------------
Thaddeus L. Olczyk, PhD
Think twice, code once.



From opengeometry at yahoo.ca  Wed Apr 16 17:13:49 2003
From: opengeometry at yahoo.ca (William Park)
Date: 16 Apr 2003 21:13:49 GMT
Subject: Problem Cron-Job using Python-Script
References: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de>
Message-ID: <b7kh2d$1k2h6$2@ID-99293.news.dfncis.de>

Andreas Penzel <mail at andreas-penzel.de> wrote:
> Hello!
> 
> I have a Python-Script to create Log-Files und work with them.
> Very well.
> But if I start the Script with Cron it works not correct.
> I often use os.system() to send commands to the shell including paths from
> the os.
> Could this be the problem?
> And if, how to solve?

- Use full path
- Post the error message.  We are not mind-readers.

-- 
William Park, Open Geometry Consulting, <opengeometry at yahoo.ca>
Linux solution for data management and processing. 



From bokr at oz.net  Sat Apr 26 12:57:01 2003
From: bokr at oz.net (Bengt Richter)
Date: 26 Apr 2003 16:57:01 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <b8edot$jua$0@216.39.172.122>

On Fri, 25 Apr 2003 23:46:50 +0200, Frantisek Fuka <fuka at fuxoft.cz> wrote:

>When I try to call methods, I sometimes forget to include the 
>parentheses. Instead of:
>
>if object.isGreen():
>	do something...
>
>i sometimes write:
>
>if object.isGreen:
>	do something...
>
>If I understand it correctly, the if statement in this case tests if 
>pointer to hasParent method is non-zero, which is of course always True 
>so no error is reported and "do something..." always gets executed, so 
>the application behaves in quite different way than I expected.
>
>You can say to me "Don't forget to always include the parenteses" but 
>I'm still curious if this cannot be somehow configured, so that I get 
>error when I try to just access the method pointer (".isGreen") instead 
>of calling the method (".isGreen()"). I understand that there has to be 
>the possibility of accessing the method pointers but I usually don't 
>need it and for a beginner like me this makes the applications very hard 
>to debug.
>
I think it's easier to do it the other way around. I.e., always leave off
the parenthesis, and make your isGreen a property if it has to be
evaluated each time it's accessed. The get-function of the property would
then return a bool (1 or 0 or True of False) and if you put () on that you
would get an exception. Type
    help(property)
for more info. Don't forget to derive your class from object if you want
to define properties for it.

Some people get nervous about the ambiguity of an attribute name that
doesn't just dereference to an object, but calls a method instead.
You might want to call special attention to property usage in the docs
for your class if there's anything tricky at all. But it is nice
to be able to write stuff like widget.bgcolor='lime' and have a visual
side effect if desired, etc. Some prefer widget.setBgColor('lime') though.
OTOH, properties let you add side effects to ordinary attributes without
changing the accessing code.

Regards,
Bengt Richter



From oblivious at web.de  Sun Apr 13 03:05:41 2003
From: oblivious at web.de (Dominic)
Date: Sun, 13 Apr 2003 09:05:41 +0200
Subject: Macros in Python? (PEP 310)
In-Reply-To: <z2Zla.5338$LB6.149344@news1.tin.it>
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org> <b770t8$2b8$00$1@news.t-online.com> <b771jb$ac6$03$1@news.t-online.com> <a8b7f07a.0304111546.3e5a5636@posting.google.com> <wJRla.3135$LB6.88227@news1.tin.it> <b79jii$njr$06$1@news.t-online.com> <z2Zla.5338$LB6.149344@news1.tin.it>
Message-ID: <b7b25d$raj$06$1@news.t-online.com>

> Not sure what "generator hack" you're considering.
 From Beni (a previous posting):

 >>To complete the list of solutions, there is an evil way to abstract
 >>try..finally that is not recommended due to it's unreadability to
 >>anybody that doesn't know the trick.  It also only works in CPython
 >>with reference counting that invokes __del__ immediately.

 >>The basic idea was to use a for loop and a generator:

 >>def with_root_privs():
 >>    oldid = geteuid()
 >>    seteuid(0)
 >>    yield None # dummy
 >>    seteuid(oldid)

 >>for _ in with_root_privs:  # `_` is dummy
 >>    # Dohere whatever you wanted.  The generator will cause this to be
 >>    # executed exactly once, with pre- and post- code excuted...

^^^^^^^^ The code needs not to be written into a function,
that's what I tried to say with "separate".

 >>The problem, of course is that the for loop's body can raise an
 >>exception, or exit the loop in other evil ways.  Luckily in CPython,
 >>exiting the loop immediately drops the only reference to the
 >>generator, so hooking it's __del__ would do the trick.  Writing it as
 >>an iterator class instead of a generator would be be ugly but we can
 >>only do it once, by a universal wrapper taking a naive generator and
 >>returning a safe one.  Google for "abstracting try..finally with
 >>generators" in c.l.py if you are interested.  Again, it's hard to
 >>recommend this hack.  Instead, consider pushing PEP 310 .

Thank you very much for your detailed answers Alex :-)

Ciao,
  Dominic




From exarkun at intarweb.us  Sat Apr  5 14:17:53 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sat, 5 Apr 2003 14:17:53 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <a3b5b608.0304051049.7048d76@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com>
Message-ID: <20030405191753.GA23326@meson.dyndns.org>

On Sat, Apr 05, 2003 at 10:49:46AM -0800, Giorgi wrote:
> > 
> > Right now, no, there isn't any Right Way to do Python on the web.  Zope
> > might be the closest thing, but there is no consensus.  I can't really
> > figure out what a consensus would be, but I imagine there's at least
> > some things we can come to consensus on... maybe underlying
> > infrastructure, or ways we can share code (like making templating
> > separate from the frameworks).  There's still a lot of work to be done
> > on that path, though.
> > 
> 
> The question is that the only way to establish Python as a language
> for corporate level web-programming is to have the standard approach
> if not introduced, then at least supported by the Python consortium...

  Python consortium?

  Heee.

  Jp

-- 
http://catandgirl.com/view.cgi?44
-- 
 up 16 days, 16:00, 5 users, load average: 1.13, 0.93, 0.85




From donn at u.washington.edu  Wed Apr 16 12:06:55 2003
From: donn at u.washington.edu (Donn Cave)
Date: 16 Apr 2003 16:06:55 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org> <b7di0n$dd6gj$1@ID-169208.news.dfncis.de> <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net> <1050390062.742825@yasure> <b7i1mm$17cpu$1@ID-169208.news.dfncis.de>
Message-ID: <b7jv2v$1vd4$1@nntp6.u.washington.edu>

Quoth "Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com>:
| Donn Cave wrote:
|> The idea that repr() must marshal the object is a sorely misguided
|> extension of this.  It is not consistently supported enough to be
|> clearly useful, and it has done some harm, e.g., it apparently
|> contributed to a move to make the repr for floats to display full
|> precision.
|
| No, I think the reason for that is to provide something
| which can be relied upon to tell you the "whole truth"
| about the value of a float; str(aFloat) lies to you slightly,
| exacerbating people's confusion about the behaviour of
| float arithmetic.
|
| Of course, it's still not *quite* the whole truth --
| to get that you'd have to display it in binary.
| But it's better than it used to be.

But the idea that repr should be the function that conveys that
"whole truth" is suspiciously consistent with the marshalling idea.
Cf. following Tim Peters articles along those lines -
http://groups.google.com/groups?q=g:thl323609740d&dq=&hl=en&lr=&ie=UTF-8&selm=000001bf275c%2402370d20%241d2d153f%40tim
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&selm=mailman.988673761.22305.python-list%40python.org

	Donn Cave, donn at u.washington.edu



From hal_wine at yahoo.com  Tue Apr  1 12:32:13 2003
From: hal_wine at yahoo.com (Hal Wine)
Date: Tue, 01 Apr 2003 09:32:13 -0800
Subject: converting from shell script to python
In-Reply-To: <b6bc47$3enbd$1@ID-99293.news.dfncis.de>
References: <kfn0jaoig1.fsf@shaoz.activesky.com.au> <b6bc47$3enbd$1@ID-99293.news.dfncis.de>
Message-ID: <3E89CD1D.8010503@yahoo.com>

William Park wrote:
> Shao Zhang <shao at cia.com.au> wrote:
> 
>>I need to convert a very simple shell script to python in order get it
>>run on both windows and unix. However I cannot quite work out a simple
NB:           ^^^^^^^

Which means that the class path needs to look like a normal file 
path on each platform. Which, afaik, _does_ mean using 
os.path.join and os.pathsep

As 'elegant' as I can get is
import os
ROOTDIR = r'c:\lib'
JARS = ( ('a', 'a.jar'),
          ('b', 'b.jar'),
          ('c', 'c.jar'))
CLASSPATH = os.pathsep.join(
               map(lambda x: os.path.join( ROOTDIR, *x ), JARS) )
os.system( 'java -classpath %s foo' % (CLASSPATH) )

> 
>>Will perl be a better
>>choice for things like this?

Not IMO -- it's harder to find the equivalent of os.pathsep...




From rmunn at pobox.com  Wed Apr  2 19:36:01 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 03 Apr 2003 00:36:01 GMT
Subject: Some newbie questions the FAQs didn't answer!
References: <6ffd97db.0304021132.62f86c7b@posting.google.com> <mailman.1049313785.3664.python-list@python.org>
Message-ID: <slrnb8mbfp.dph.rmunn@localhost.localdomain>

Verlee <verlee at web.de> wrote:
> Hi,
> 
>> [...] But then I read that in order to run a python program, the
>> machine had to have the python VM installed, and > the user had to go
>> to a command line and type "python <program name>>.py".
> 
>> What I had planned to do was write Win32 GUIs, but I would guess that
>> most common users don't have Python installed.  Is there some way to
>> write the program in C++ and then have it bring up the GUI written in
>> Python without Python being installed?  Wouldn't that require the code
>> being compiled into machine code?

A better way would be to write the whole thing in Python and then use a
tool like py2exe to bundle up the Python interpreter with your code in a
single .EXE file. (See below.)

> There is a tool called py2exe that converts .py-files to .exe-files
> which can be executed without having the Python interpreter installed.
> I figure this is what you want to have?
> (I?ve been worrying a lot about this, too...)
> It can be downloaded at http://sourceforge.net/projects/py2exe/
> Haven?t had time to try it yet, though. There may also be other tools
> that do the same, but this is the (only) one I have heard of.

I've done this; not with any large application, but with a small
wxPython test program that brought up a GUI window. It produced a single
.EXE, which I was able to copy to a machine that didn't have Python
installed, and it worked perfectly.

BTW, if you're planning on writing GUIs in Python, I would recommend
staying *far away* from the Windows API, and using one of the
cross-platform Python GUI toolkits instead. I once spent three months
trying to write something in the Windows API -- it was *painful*. Using
one of the cross-platform toolkits, on the other hand, will not only be
*much* more pleasant, it will also allow you to produce programs that
can run on Unix, Windows, and Macintosh systems, all with no extra
effort on your part! There are several toolkits available, but I'm
partial to wxPython myself: I've found it quite easy to understand and
use, and (unlike some other toolkits) it looks native on each platform.
That is, on Windows it looks just like a normal Windows app; on
Macintosh it looks just like a normal OS X app (or OS 9 if that's what
the user is running); and on Unix it looks just like a normal GTK app.

Anyway, just some food for thought.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From reply.in at the.newsgroup  Thu Apr 10 19:17:44 2003
From: reply.in at the.newsgroup (Rene Pijlman)
Date: Fri, 11 Apr 2003 01:17:44 +0200
Subject: Can I define nested classes in Python?
References: <169a7283.0304101432.13f5786c@posting.google.com> <1JGdnc952fcaaAijXTWcrg@comcast.com>
Message-ID: <6dub9v0rjv0s7r87hhjsjqb0oqjuu9m8ko@4ax.com>

Terry Reedy:
>sdieselil:
>> Can I define nested classes in Python
>
>Try it and see what happens.

This may have been how Thomas Edison constructed the light bulb,
but this is not the Right Way to learn a programming language.

A better approach is to read the manual.
http://www.python.org/doc/current/ref/nested-scopes.html

Although in this case, I must admit, the manual primarily serves
to steer Python newbies towards a career in light bulb
development :-/

But wait a minute, why would a newbie want to use nested classes
anyway?

-- 
Ren? Pijlman



From artyprog at wanadoo.fr  Thu Apr 17 19:23:58 2003
From: artyprog at wanadoo.fr (Salvatore)
Date: Fri, 18 Apr 2003 01:23:58 +0200
Subject: variable in Python
In-Reply-To: <mailman.1050617370.30482.python-list@python.org>
References: <b7lvbh$an3$1@news-reader13.wanadoo.fr> <ad052e5c.0304171017.45b8884f@posting.google.com> <mailman.1050617370.30482.python-list@python.org>
Message-ID: <b7nd4k$19f$1@news-reader11.wanadoo.fr>

Steven Taschuk wrote:
> Quoth Dan Bishop:
>   [...]
> 
>>The latter: All variables in Python are pointers in disguise.
> 
>   [...example...]
> 
>>Note that I used a list instead of an int.  With ints, the distinction
>>between value and reference semantics doesn't matter as much because
>>ints are immutable.
> 
> 
> And this is an extremely important point: C programmers who are
> familiar with pointer semantics would expect that after
>     a = 3
>     b = a
>     a = 4
> we would find b == 4.  This is why, as Aahz pointed out, thinking
> of Python variables as pointers can lead one astray, and why you
> often see Python variables described as simply "names for
> objects", and = as not a value-copying operation (as it is in C)
> but a name-binding operation.
> 

Thanks Steven




From donnal at donnal.net  Tue Apr 15 13:28:29 2003
From: donnal at donnal.net (Donnal Walter)
Date: Tue, 15 Apr 2003 12:28:29 -0500
Subject: Python and Schools
In-Reply-To: <slrnb9o8s8.iea.news@datatailors.com>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <bDSma.22753$T34.687158@news2.tin.it> <slrnb9o8s8.iea.news@datatailors.com>
Message-ID: <b7hfd4$4oe$1@main.gmane.org>

Peter van Kampen wrote:
> In article <bDSma.22753$T34.687158 at news2.tin.it>, Alex Martelli wrote:
> 
>>Aahz wrote:
>>
>>
>>>In article <v9e21mhemm8c2 at corp.supernews.com>,
>>>Paul Watson <pwatson at redlinec.com> wrote:
>>>
>>>>We need to teach students correct design priciples to build something
>>>>greater than what already exists.  We will never get very far if we
>>>>require everyone to start with a quark or atom.  Yes, of course we need
>>>>some people who design silicon and create microcode.  They will learn
>>>>the low-level details what they need to know as they need it.  Knowing
>>>>great design and organization principles will enable them to make the
>>>>most of it.
>>>
>>>While there's some truth to that, try explaining why the following code
>>>is a Bad Idea to someone who has no basic understanding of CS:
>>>
>>>    s = ''
>>>    for i in range(1000000):
>>>        s += str(i)
>>
>>I will gladly accept the challenge, if we can agree on the details to
>>determine whether I've won or lost the bet.  What I require is somebody
>>with real, earnest *interest* in *understanding* this stuff -- I'll
>>gladly undertake to supply all necessary *background*, but I have no
>>idea on how to awake real interest in somebody whose motivation for
>>learning is just "serving time" or some vague hope of getting a good
>>paying job (and that would be just as much of a problem, whether the
>>subject was "CS", "programming", or "artichoke farming techniques").
> 
> 
> I would gladly volunteer to be that somebody, although I realize it
> won't happen (Practicallity beats geograpy ;-)) But I think you'll be
> surprised about the number of people who *would* want to understand
> the basics without necessarily (re-)entering university for a
> CS-course. Let me tell you my background, perhaps it can serve as an
> example.
> 
> I am a programmer by accident. I love the pleasure of 'creating' when
> writing programs/scripts/whatever. I studied English literature but I
> have toyed with computers over the years, mostly as a means to
> store/retrieve stuff my brain (sadly) keeps forgetting (that's why I
> have always been interested in databases).
> 
> Thanks to BASIC I learned on a Commodore 64 and dBase I learned on my
> parents computer, back in the 1980's I found myself faced with the
> opportunity/neccesity (as I said, totally by accident) to develop an
> MS-Access-application in 1999 (a long story I won't bother you with).
> After that was finished I was rather hooked (again after almost 10
> years of hardly touching a computer, couldn't afford on either). Once
> I learned to connect databases to the internet I could make a living
> out of it (my re-discovery of the computer coincided with the
> .com-hype), but I could not afford to keep up with Microsofts
> relentless 'upgrading' so I have gradually moved to linux over the
> past 2 years. Now I do mostly php/mysql-stuff and try to learn python.
> 
> I do not understand discussions about O(N) en O(N^2) and would be very
> interested to learn such things. I have bought 'Python in a Nutshell'
> and I love it how you explain about tokens, identifiers, keywords and
> so on. Very few (if at all) 'general purpose' textbooks take the
> trouble to explain stuff like that.
> 
> 
>>If the interest is there, and enough Python background to understand
>>exactly what this program snippet DOES, then in about one or two
>>hours of face to face instruction I can give a model that will allow
>>the student to predict this snippet's disastrous performance and
>>compare it to the expected performance for, e.g.,
>>
>>    s = ''.join([str(i) for in in range(1000000)])
>>
>>It's not rocket science (that's _another_ thread...;-) -- at this
>>level, it's really simple arithmetic, a decent mental model of "an
>>abstract underlying machine", AND an interest in understanding this.
> 
> 
> So +1 for an article/book. I think you could write another best-seller
> explaining this and similar basic things most CS-students take for
> granted while there are many more users of python (and many other
> high-level languages) who use it just to get their jobs done...
> 
> PterK
>

Me, too, Peter. +1 on the article/book as well.

I am a physican first and programmer second. I've never taken a course 
in computer science, and I do not know the meaning of O(N). OTOH, just 
from following comp.lang.python I think I have a pretty good idea why

     s += str(i)

would be a bad idea. Nevertheless, I certainly would not be opposed to 
deepening my understanding. (PythonIAN is great, BTW.)

Donnal Walter
Arkansas Children's Hospital
mindwrapper.org






From bbull at pathfire.com  Fri Apr 11 12:13:18 2003
From: bbull at pathfire.com (Brian Bull)
Date: 11 Apr 2003 09:13:18 -0700
Subject: Dynamically loading python classes
Message-ID: <9f363ad9.0304110813.3778cc8a@posting.google.com>

I am writing an application that has to dynamically load classes from
other modules based on certain criteria.  I load the class initially
like this:

 self.dict["handleUnknown"] =
new.classobj("handleUnknown",(handleUnknown.handleUnknown,),{})

I later initialize this class by calling

runningClass = self.dict[clas](self.pluginLog)

where clas is the name of the class I want to load such as
"handleUnknown" in this example

I then execute methods against the class object and delete it when I
am finished.

What I am looking for is a way to load the class "fresh" each time I
instantiate it.  For instance, if I make a change to the loaded class,
I want that change to be activated the next time the class is loaded
without having to restart the main program that loads the classes.

I also need to be able to load the module and class from strings
instead of actually using the module/class name.  ie.

runningClass = handleUnknown.handleUnknown()

This is what I want to be able to do (something similar to this)
className = "handleUnknown"
runningClass = className.className()

Any help would be appreciated.  If you need a better explanation,
please email me (bbull at pathfire.com).  I can also send my code if
needed.



From nessus at mit.edu  Fri Apr 18 19:22:34 2003
From: nessus at mit.edu (Douglas Alan)
Date: Fri, 18 Apr 2003 19:22:34 -0400
Subject: A new and very robust method for doing file locking over NFS?
References: <lcu1cw525a.fsf@gaffa.mit.edu> <5j65pbai03.fsf@xoxy.net>
 <mailman.1050695250.12755.python-list@python.org>
Message-ID: <lclly7ie6d.fsf@gaffa.mit.edu>

Jeff Epler <jepler at unpythonic.net> writes:

> If the actions between the dashed lines all took place within the same
> wall second, client 1 may read its own cached data instead of the data
> written by client 2.  Our eventual solution was a kernel patch which let
> us evict any cached data for a particular file from memory, and we did
> it every time we performed a locking operation.  You can bet that
> performance suffered, but at least it was correct.

Hmmm, why not just fetch the time in between locking and unlocking and
if a second has not passed, sleep for the difference?  I suppose this
wouldn't work if you routinely need to lock a file for writing at a
rate of more than once a second, but there probably aren't that many
applications that write to files over NFS that need to be able to to
do that.

Thanks for the info -- it is very helpful.

|>oug



From a.schmolck at gmx.net  Tue Apr 15 21:19:38 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 16 Apr 2003 02:19:38 +0100
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com>
Message-ID: <yfsistfb5n9.fsf@black132.ex.ac.uk>

Dave Benjamin <ramen at lackingtalent.com> writes:

> In article <yfs7k9vcu2p.fsf at black132.ex.ac.uk>, Alexander Schmolck wrote:
> > I think your life might be much happier if you just went to
> > http://www.drscheme.org/, dowloaded Dr Scheme and used *that* for
> > webdevelopment. Then you can use a real functional programming language and
> > read books by people who actually have a real clue about functional
> > programming (there is lots of *excellent* material available on-line and for
> > free, like SICP and HTDP, the latter uses Dr Scheme -- also have a look at
> > http://www.cs.brown.edu/courses/cs173/2001/Lectures/). Why settle for some
> > second-rate immitation? [1]
> 
> I may just take your advice and investigate that. Thanks for the reference.
> But keep in mind that I'm not saying "I don't like Python, it's not
> functional enough". I'm saying "I love Python, it lets me work in both an FP
> and OOP style, and this makes me happy. I just don't want to see FP support
> drop out of the picture."

Well it seems to me that you're currently *really* keen on functional
programming. You can do that (and learn about it) much better in scheme than
in python, so even if you want to program in python for the rest of your life
afterwards it might well be worth it. I mean you wouldn't want to learn OO in
C, would you -- there a fewer good ressources and it's just much more of a
pain (this comparison is of course overly pessimistic -- C is just horrible
and python really isn't).

Plus, Dr Scheme is a rather nice learning environment, has plenty of cool
stuff (and quite a few libraries) and let's you do OO, too (if you use Emacs
check out quack.el, BTW).


> That's true, but then I have to name "bar". And part of the joy of FP is not
> having to name *everything*. Naming things is important if you need to call
> them up later, but it's a hassle to name intermediate values

No argument here. I tend to think python would have been a better language if
it had included something like smalltalk-like blocks from day 1 (I still
really like python).

> >> keyword-argument trick wasn't really that bad, once you got used to it. In
> >> any case, with dynamic scoping, this problem has completely disappeared.
> >                  ^^^^^^^^^^^^^^^
> > not quite.
> 
> Am I using the wrong terminology here? I'm specifically talking about the
> named-argument hack to get around Python's previous lack of nested scopes.
> This specific problem, as far as I know, is gone now. Am I missing something?

Yes but certainly not due to dynamic scoping (which python doesn't even have,
as you can see from the raging Emacs thread). It now has non-crippled lexical
scope :)

(have a look at http://www.supelec.fr/docs/cltl/clm/node43.html for a good
explanation)

> Well, I don't think people usually write functions for list comprehensions
> at all.

Never mind, I just shouldn't write postings when I'm tired (like now:).

'as



From llothar at web.de  Mon Apr  7 09:30:47 2003
From: llothar at web.de (Lothar Scholz)
Date: 7 Apr 2003 06:30:47 -0700
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain>
Message-ID: <6ee58e07.0304070530.79d9c994@posting.google.com>

pobrien at orbtech.com (Patrick K. O'Brien) wrote in message news:<m2k7e78j5s.fsf at localhost.localdomain>...

> Actually, we don't have to clone everything about the interface.  I'm
> using wxPython, so the interface is pretty and more easily configured.
> But some people actually like the minibuffer, so we'll need to support
> something like it and something less ugly.  Heck, I'm actually running
> three different versions of the interface now, and I only wrote the
> code this week (minus the almost two year's worth of code in PyCrust).
> 
> > There are only a handful of features that are normally the reason to
> > keep this as your primary editor. For me its the window handling C-X
> > 1, the way to prefix commands with a repeat command, the macro
> > recording and a the autoindentation with TAB.
> 
> Except everyone will likely disagree about what the handful of
> features are.  ;-) In any case, I'm targeting the bare minimum to
> start.
> 
> > To make clear, i'm using XEMACS 8 hours a day for years now. But
> > it's absolutely losing ground compared to modern editors. And i
> > don't think that extensibility in the emacs sense is really
> > necessary today or a plus, because people want out of the box
> > solutions that are specialized to their domain.
> 
> What would you like to see in a modern version?
> 

Look at: 

http://www.eagle-software.com/

This people had a lot of good ideas. Also the codeguide people at
www.codeguide.com (????) - a java ide - have implemented codetips/code 
insight in the way i like them.



From chlump at yahoo.com  Wed Apr  9 15:42:44 2003
From: chlump at yahoo.com (Daniel Gowans)
Date: 9 Apr 2003 12:42:44 -0700
Subject: string() - Is this some kind of sick joke?
Message-ID: <36dbf824.0304091142.527af834@posting.google.com>

There is some sick joke in the python 2.2.2 documentation I have. 
This is located in the ActivePython documentation:

strip(s[, chars]) 
  Return a copy of the string with leading and trailing characters
removed. If chars is omitted or None, whitespace characters are
removed. If given and not None, chars must be a string; the characters
in the string will be stripped from the both ends of the string this
method is called on.

The same goes (in the docs) for rstrip and lstrip.

But does it actually work that way?  No!

Observe from the 2.2.2 interpreter (and 2.2.1 in HPUX):

>>> y="'hello'"
>>> import string
>>> string.strip(y,"'")
Traceback (most recent call last):
  File "<interactive input>", line 1, in ?
TypeError: strip() takes exactly 1 argument (2 given)
>>> 

What gives?  Did I get the bogus version?  Is this wishful thinking
documentation?  I also tried  y.strip("'")  with the same results.

If this can't be done this way (oh, how I love chomp()) then is there
a simple, elegant way to do it?



From sholden at holdenweb.com  Thu Apr 17 11:04:55 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 15:04:55 GMT
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de> <mailman.1050586468.29594.python-list@python.org>
Message-ID: <rozna.115488$It5.70451@news2.central.cox.net>

"Jp Calderone" <exarkun at intarweb.us> wrote ...
> On Thu, Apr 17, 2003 at 10:34:08AM +0000, Haran Shivanan wrote:
> > This seems like a fairly straight forward thing but I'm not sure about
how
> > to implement it.
> > I have a base class from which I'm deriving several new classes:
> > class A:
> > mem_a=0
> > mem_b=1
> > class B(A):pass
> >
> > I want B to inherit mem_a and mem_b from A without my having to explicit
do
> > something like:
> >
> > self.mem_a = 0
> >
> > in the constructor for class B.
>
>     class A:
>         a = 0
>         b = 1
>
>     class B(A):
>         pass
>
>     b = B()
>     print b.a, b.b
>
>   Isn't this the behavior you desire?  If not, could you clarify your
> question?

Probably not.

>>> class A:
...   a = 0
...   b = 1
...
>>> class B(A):
...   pass
...
>>> b = B()
>>> A.a = 23
>>> b.a
23
>>>

I suspect the OP wants a separate class variable for each subclass. Right,
Haran?

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From theller at python.net  Thu Apr 17 10:01:55 2003
From: theller at python.net (Thomas Heller)
Date: 17 Apr 2003 16:01:55 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <d6jmqr2q.fsf@python.net> <Xns936072232AA84duncanrcpcouk@127.0.0.1>
Message-ID: <he8xgr3g.fsf@python.net>

Duncan Booth <duncan at NOSPAMrcp.co.uk> writes:

> Anyone fancy a patch that gives the output below? (I do hope it doesn't word wrap).
> 
> FFFF
> ======================================================================
> FAIL: test_failUnlessEqual (__main__.Test)
> ----------------------------------------------------------------------
> TestFailed: failUnlessEqual
> 'xxxxxxxxxxxxxxxxxx...xxxxyzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz...
>                    ...    ^
> 'xxxxxxxxxxxxxxxxxx...xxxxazzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz...
>                    ...    ^
>   File "J:\temp\test.py", line 7, in test_failUnlessEqual
>     self.failUnlessEqual(s, t)
> 
[and so on]

This looks great, but has problems. It is slow. For some input, at
least.

I didn't have enough patience to wait for the output of your script,
instead I changed the 10000 character strings to 100 characters and then
it only took some seconds.

Maybe you should create a new method (failUnlessEqual_ShowDiff or so)
which will do the fancy formatting.

Maybe the diff formatting should be put into the TestReport, so that
at least the
    FFFF
line is printed more quickly, and the formatting doen afterwards.

Maybe a time limit should be introduced.

Thomas



From pinard at iro.umontreal.ca  Tue Apr 29 08:00:54 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 29 Apr 2003 08:00:54 -0400
Subject: Pyrex Hard Case
In-Reply-To: <3EAD7A15.6060508@elo.utfsm.cl>
References: <20030428171412.ZMGU12530.mta2@mta2.real.mail.entelchile.net>
	<mailman.1051552208.15532.python-list@python.org>
	<3EAD7A15.6060508@elo.utfsm.cl>
Message-ID: <oqof2pzf61.fsf@titan.progiciels-bpi.ca>

[Rodrigo Benenson]
> [Francois Pinard]
> > [rdrb at 123.cl]

> >>>cdef extern from *:
> >>>    MACRO_NAME(EXPECTED_ARGUMENTS)

> >>I had already done that.
> >>extern MACRO_NAME(EXPECTED_ARGUMENTS)

> > This is not the same thing.

> [...] I did test your suggestion and it creates the same output.  It's
> obvious it should do, because we are defining and External Function
> Prototype.

This is strange that you get unchanged output.  A first reason is how the
the `cdef extern from *:' is documented to work.  Another is that it works
as documented for me (not on your problem, of course, but for similar ones).
A last reason, admittedly vaporous, is that Pyrex sources finality is to
control how C code is generated, and so, it is likely that changing Pyrex
sources will change generated C.  I'm sorry to not have helped you better,
and I wish you will succeed in finding an elegant solution to your problem.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From martin.vonloewis at hpi.uni-potsdam.de  Tue Apr 22 15:20:20 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Tue, 22 Apr 2003 21:20:20 +0200
Subject: Problems compiling _tkinter
References: <mailman.1051002635.15634.python-list@python.org>
Message-ID: <b844lk$94i$1@zeppelin.rz.uni-potsdam.de>

> build/temp.linux-i686-2.2/tkappinit.o -L/usr/X11R6/lib -L/usr/local/lib
> -ltk8.4 -ltcl8.4 -lX11 -o build/lib.linux-i686-2.2/_tkinter.so
> WARNING: removing "_tkinter" since importing it failed
>
> The only other errors in the make are some warnings about tempnam_r in
> posixmodule.c.
>
> Any ideas?

You can ignore the compiler warnings: they are about constness problems.

To find out why linking the module fails, you should try to import the
module manually, and see what error message that produces. A copy
of the module should still exist in build/lib*, under a different name.

Regards,
Martin





From staschuk at telusplanet.net  Tue Apr 15 17:32:05 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 15 Apr 2003 15:32:05 -0600
Subject: Printing Columns
In-Reply-To: <20030415172106.IOMK15908.priv-edtnes16-hme0.telusplanet.net@mail.python.org>; from Greg.Lindstrom@acxiom.com on Tue, Apr 15, 2003 at 12:04:55PM -0500
References: <20030415172106.IOMK15908.priv-edtnes16-hme0.telusplanet.net@mail.python.org>
Message-ID: <20030415153205.A3151@tibia.amotlpaa.bogus>

Quoth Lindstrom Greg - glinds:
  [...]
> myList = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H']
  [...]
> 	A	D	G
> 	B	E	H
> 	C	F
> 
> problem is, the code (all 12 lines) is a bit convoluted and would not be
> easy for the next person using it to figure out (and in my mind,
> maintainability is priority one; that's why I use Python).  I would like to
> know if there is a "pythonic" way to do this. So, here are the specs....
> 
> Given a list of length N with 0<=N<=infinity and a number of columns, C (C>0
> and may be less than N), how do we print out the elements of the list in the
> above (column major?) form?

N might be infinite?  That introduces some difficulties.  I'll
assume you meant 0 <= n < infinity.

You don't specify which of these arrangements is preferred:
    0  3  6  8            0  3  6  9
    1  4  7  9            1  4  7
    2  5                  2  5  8
I'll assume the former.

The most straightforward method I can think of is this:

    def chop(lst, n):
        """Chop lst into sublists of n items."""
        return [lst[i:i+n] for i in range(0, len(lst), n)]

    def columnize(lst, n):
        """Arrange the elements of lst into n columns."""
        # assemble long & short columns
        q, r = divmod(len(lst), n)
        longcount = r # number of long columns
        longlen = q + 1 # length of long columns
        shortlen = q # length of short columns
        longcols = chop(lst[:longcount*longlen], longlen)
        shortcols = chop(lst[longcount*longlen:], shortlen)
        # pad short columns
        shortcols = [c + [None] for c in shortcols]
        # transpose
        rows = map(list, zip(*(longcols + shortcols)))
        # unpad short columns
        del rows[-1][longcols:]
        return rows

Example:

    for row in columnize(range(37), 5):
        print '\t'.join(map(str, row))

    0       8       16      23      30
    1       9       17      24      31
    2       10      18      25      32
    3       11      19      26      33
    4       12      20      27      34
    5       13      21      28      35
    6       14      22      29      36
    7       15

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From max at alcyone.com  Tue Apr  1 17:20:05 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 01 Apr 2003 14:20:05 -0800
Subject: else: w/o except: - why not?
References: <mailman.1049156943.12744.python-list@python.org> <v8j9pb3v2cnra6@corp.supernews.com> <698f09f8.0304011201.1a4939a1@posting.google.com>
Message-ID: <3E8A1095.A7414F8@alcyone.com>

Jeremy Fincher wrote:

> Wouldn't that be better written like thiS?
> 
> try:
>   f1()
> finally:
>   f2()

No, that means something different.  The finally clause _always_ gets
executed; the except clause only gets executed if an exception occurs.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ A good indignation brings out all one's powers.
\__/ Ralph Waldo Emerson
    Esperanto reference / http://www.alcyone.com/max/lang/esperanto/
 An Esperanto reference for English speakers.



From mcfletch at rogers.com  Tue Apr  8 17:48:59 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 08 Apr 2003 17:48:59 -0400
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <BAB8B22D.2022%dogen@mindspring.com>
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org> <BAB8B22D.2022%dogen@mindspring.com>
Message-ID: <3E9343CB.4090107@rogers.com>

Here's an algorithm that should give you a decent speedup (it's based on 
the suggestions that I _think_ Steven was giving you)...

words = [...]
backwardMap = {...} # last-char to reversed word-list, construct from words

for word in words:
    for back in backwardMap[ word[-1]]:
       subset = min(len(word),len(back))
       if word[:subset] == back[:subset]:
          good_chance_you_have_one_so_explore_further()

For even faster operation, pre-calculate and store the word-lengths in 
the backwardMap so that you don't have to calculate them each time. 

Basically, this should reduce the search space to 1/64th of total (or 
so) for each word with a simple O(1) hash-table lookup, with each check 
considerably faster in exchange for an O(n) setup fee (reversing and 
taking the length of each word, then adding to the dictionary).  Key is 
setting up the correct data-structures so that when you're doing the 
actual run-time searches you can rapidly whittle down the search space 
to a reaidly-processed sub-set.  You see the same thing in fuzzy-text 
searching, BTW.

You can also do neat things such as using the tails beyond subset to 
lookup third/fourth words in the middle (you need a forward map to make 
that work efficiently, of course).  Keep in mind that if you find a 
palindrome set which exactly matches on both sides you get the 
possibility of having an inifinite sequence of those nesting in, so 
you'll want a mechanism to stop that at some point :) .

Enjoy,
Mike


Jon Brown wrote:

>Steven,
>
>Actually the program does exactly what you describe, except from the inside
>out because it starts in the middle, and it cuts out of the loop if the
>characters don't match.
>
>To look for pals it spins off big, big numbers of combinations, true enough.
>That's what behind the need for speed. Since Python is inherently on the
>pokey side, I expected the slowness but wonder if I'm missing a cool way to
>pick up the pace. But I suspect the solution will involve C++, and I've
>never linked C/C++ with Python.
>
>Jonedary 
>
>  
>
>>A smarter search would do better.  One approach would be to build
>>the palindrome from both ends instead of from just one end.  For
>>example, suppose your code understood that a palindrome starting
>>with 'orange' must end with a word 'o', 'ro', 'aro', 'naro',
>>'gnaro', 'egnaro', or a word ending with 'egnaro'.  If there are
>>no such words, it can skip 1/n of the search space, namely all the
>>combinations which start with 'orange'.
>>    
>>
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From a.schmolck at gmx.net  Fri Apr 11 19:40:09 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 12 Apr 2003 00:40:09 +0100
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com>
Message-ID: <yfsel48fvs6.fsf@black132.ex.ac.uk>

Erik Max Francis <max at alcyone.com> writes:

> Fortepianissimo wrote:
> 
> > I can't seem to find the exact description between these two in the
> > Library Reference - so am I correct in assuming that deepcopy() will
> > not actually duplicate an intern'ed string?
> 
> Strings are immutable, so what good would a copy (via copy or deepcopy)
> do you?  You cannot possibly change an existing string object, so its
> identity is unimportant.

No it isn't. The idea of symbols (viz interned strings) is their uniqueness
(amongst other things, establishing identity is faster then establishing
string equality). So I don't think the OP's question is unreasonable: you
might well want to deepcopy some container that contains symbols (and you'd
like to preserve their identity).

I'm pretty sure that deepcopy will not duplicate objects it knows to be
immutable, but explicitly documenting this behavior (even as undefined) might
be useful.

'as



From ramen at lackingtalent.com  Tue Apr 29 02:28:06 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 29 Apr 2003 06:28:06 -0000
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com> <7ULqa.9804$3M4.264757@news1.tin.it> <g3qdncyq1Zx2FzGjXTWcrg@comcast.com> <mailman.1051497463.22158.python-list@python.org>
Message-ID: <slrnbas6t1.j6l.ramen@lackingtalent.com>

In article <mailman.1051497463.22158.python-list at python.org>,
Lulu of the Lotus-Eaters wrote:
> And to be even trickier, answer the question for the following lists:
> 
>     [u'x', 'x', chr(255)].sort()
>     ['x', chr(255), u'x'].sort()
>     [1j, '1', 1].sort()
>     [1, 1j, '1'].sort()
> 
> Then try to explain the answer to ANYONE you know.

I don't know what you're talking about. They all return "None" for me.
<ducking>

Dave



From imbosol-1049687461 at aerojockey.com  Mon Apr  7 00:03:18 2003
From: imbosol-1049687461 at aerojockey.com (Carl Banks)
Date: Mon, 07 Apr 2003 04:03:18 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net> <yfsd6jzkney.fsf@black132.ex.ac.uk>
Message-ID: <aM6ka.40992$wM6.11447@nwrdny01.gnilink.net>

Alexander Schmolck wrote:
> Carl Banks <imbosol-1049649076 at aerojockey.com> writes:
>> Next three functions are "save-current-buffer", "with-current-buffer",
>> and "with-temp-buffer".  These are actually not functions, but special
>> forms and macros.  Python has no analogue to these (barring some
>> obscure metaclass tricks). 
> 
> Care to share the obscure metaclass trick to implement with-* style macros in
> python?

I doubt it'll work for with-style macros, but this shows how to do a
conditional operator:

http://tinyurl.com/8yne


>> There is one potentially disasterous pitfall you should be aware of.
>> Emacs does all kinds of strange things with namespaces.  It uses
>> dynamic binding, unlike Python, which uses lexical binding.  Emacs has
>> one monolithic global namespace, as opposed to Python, which has an
>> organized infrastructure of "global" namespaces, that is, modules.
>> Also, Emacs has weird namespaces like buffer local variables.  One
>> thing you'll have to do is think about how to sort all these namespace
>> issues out.
> 
> Yes, these are minor issues (BTW you forgot to mention that emacs also has
> separate namespaces for functions/macros and variables (and, in a slightly
> different way, lots of other things)).

Good point.  I think namespace is a major issue, especially if you're
going to follow the Lisp code.


-- 
CARL BANKS



From nas-usenet at arctrix.com  Sat Apr  5 23:32:59 2003
From: nas-usenet at arctrix.com (Neil Schemenauer)
Date: Sun, 06 Apr 2003 04:32:59 GMT
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org>
Message-ID: <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>

Ian Bicking <ianb at colorstudy.com> wrote:
> The next step I can envision would be for Twisted to become the
> canonical infrastructure, where higher-level web frameworks are all
> built ontop of it.

That's not feasible.  Some applications would require extensive
modification to work on top of an event driven framework.

  Neil



From jdhunter at ace.bsd.uchicago.edu  Mon Apr 14 14:10:53 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Mon, 14 Apr 2003 13:10:53 -0500
Subject: debug build of python2.2 for win
Message-ID: <m2adetospe.fsf@mother.paradise.lost>

Any one know where I can get a debug build of python 2.2.2 for win32?

Thanks,
John Hunter




From jhs_bkk at nospam.yahoo.com  Mon Apr  7 12:22:38 2003
From: jhs_bkk at nospam.yahoo.com (Jason Smith)
Date: Mon, 07 Apr 2003 23:22:38 +0700
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no> <3E915BA0.A4B5286@engcorp.com> <pan.2003.04.07.11.36.03.78556@idi.ntnu.no>
Message-ID: <3e91a7e1_2@127.0.0.1>

Nils Grimsmo wrote:

> run() is called when you push the button. run() sets the label for the
> button every second for four seconds, but the button label is not updated
> before you return from run().

Disclaimer:  I'm a week-old Python programmer from a long Perl background.

Maybe you should forego time.sleep() in an interactive GUI program? 
Sleeping is asking the kernel to knock your program completely out.  You
don't want to be sleeping when a user tries to click a button, after all! 
Although I'm new to GUI programming.

I had a similar problem (waiting for blocking I/O), and I used a wxPyTimer
to get control back to my program so at least the wxButton would pop back
up while the network talk ensued.  So that's what I used for your program. 
I have no idea if this is the right solution, however.  YMMV.

(Tested on Gentoo)

from wxPython.wx import *

class MainWindow(wxFrame):
    label  = ['foo', 'bar']
    button = None

    def __init__(self, parent, id, title):
        self.togcount = 0      # Counts the SetLabel calls

        wxFrame.__init__(self, parent, -4, title, size=(100,100))
        self.button = wxButton(self, 13, self.label[0])
        EVT_BUTTON(self, 13, self.run)
        self.Show(true)

    # Controls the toggle function
    def run(self, event):
        self.timer = wxPyTimer(self.toggleButton)
        self.timer.Start(1000)
        self.toggleButton()     # Gets the toggling going right now

    # Toggles the button
    def toggleButton(self):
        self.togcount += 1
        if self.togcount == 4:  # Done
            self.timer.Stop()
            self.togcount = 0

        self.button.SetLabel(self.label[self.togcount % 2])

app = wxPyApp()
frame = MainWindow(None, -1, 'Hello, World')
app.MainLoop()



----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---



From peter at engcorp.com  Thu Apr 24 11:44:39 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 24 Apr 2003 11:44:39 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <X8Mpa.30197$y3.2398527@news010.worldonline.dk>
Message-ID: <3EA80667.BA87E94@engcorp.com>

Max M wrote:
> 
> Peter Hansen wrote:
> 
> > For throwaways, I don't write quite so many tests, and in many cases if
> > it's only a "small" throwaway script (e.g. one or two subroutines only,
> > plus a little mainline) I will write a test or two for the subroutines
> > (because I'm doing TDD, not because I want to have unit tests to make
> > sure my script works properly... if you see the distinction) but never
> > for the mainline.
> 
> While I do understand the principles of tdd, testing first and all that,
> I have a hard seing why it should be an advantage to do it in the early
> stages of a project.

I think you're leading us in a new direction.  Aahz asked about "throwaway
scripts", and to me that has the meaning of "a script with a very clear
and specific but short-lived purpose".  In other words, about the farthest
thing from "early stages of a project" that you could imagine.  There is
no project: there is a little script, then nothing.  Maybe calling them
nihilistic scripts would be better. :-)

> Normaly if a am entering a new domain, where I don't have a clear
> picture of the requirements, I try out different approaches. Each
> approach using a different combination of ideology, objects and functions.

Agreed.  As others have pointed out, XP calls this a Spike Solution
(picture driving a spike all the way through the problem in the most 
expedient, ugliest manner possible, to explore it, but with no intention
of using that as a the real solution).  Maybe Exploratory Surgery would
be another good view of it.

> This is all to get a feel of the best approach in this particular
> domain. Some times I try out a *lot* of approaches before I end up with
> a particular elegant design.
> 
> At which point I am ready to "solve the problem". It is at this point I
> can see the point of unit testing.
> 
> In tdd do you really use testing for the first "creative" fase also?

Nope, other than, as I said elsewhere in this thread, when I have very
clear ideas about a specific "unit" of code that I want to include even
in my spike, in which case I may well write it test-first in order to
have some confidence that it will work at all.  Failing to do that would
mean a higher likelihood of going down the wrong path, or spending way
too long on a fruitless candidate solution, because my code isn't doing
what I think it's doing.

The XP/TDD view of all this would be like this: "Initially, write tests 
for *everything*.  When you learn when you don't need to write tests
for something, don't do it.  When you realize you made a mistake in not
writing a test, back up a little and slow down, writing tests again 
until you're back on track."

-Peter



From a.schmolck at gmx.net  Sun Apr 13 20:11:19 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 14 Apr 2003 01:11:19 +0100
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk> <3E977928.B3A6DC7E@engcorp.com> <yfsznmueec1.fsf@black132.ex.ac.uk> <m3he92pef3.fsf@mira.informatik.hu-berlin.de> <yfsr886e3gt.fsf@black132.ex.ac.uk> <b7c85g$slh$03$1@news.t-online.com>
Message-ID: <yfs4r52djko.fsf@black132.ex.ac.uk>

"Martin v. L?wis" <martin at v.loewis.de> writes:
> I meant to explain all this in the paragraph on mortality of interned
> strings. 

And so you did, I was just being daft.

'as



From dave at boost-consulting.com  Tue Apr 22 22:35:43 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 22 Apr 2003 22:35:43 -0400
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com>
 <a3b5b608.0304080302.c26ad68@posting.google.com>
 <af6b5bdd.0304181001.558da6f@posting.google.com>
 <84of317cd6.fsf@boost-consulting.com> <b7vlfb$27c$0@216.39.172.122>
 <84of30m4cp.fsf@boost-consulting.com>
 <p0p8avsd7nlqu52p4rnjobuemctd3foil8@4ax.com>
 <848yu3o5o6.fsf@boost-consulting.com>
 <6a36e729.0304221350.5d30c3bb@posting.google.com>
Message-ID: <847k9mgcu8.fsf@boost-consulting.com>

bob at redivi.com (Bob Ippolito) writes:

> David Abrahams <dave at boost-consulting.com> wrote in message news:<848yu3o5o6.fsf at boost-consulting.com>...

>> Sure, you have lots of options. Also:
>> 
>>     xml.sometag(class_("value2"), attr1 = "value1")["contents"]
>> 
>>     xml.sometag({'class':"value2", 'while':"value3"}, attr1 = "value1")["contents"]
>> 
>>     xml.sometag(attr1 = "value1", _ = {'class':"value2", 'while':"value3"})["contents"]
>> 
>> it's all in the fun of designing domain-specific sublanguages.
>
> well, in our implementation (off the top of my head) these are valid:
>
> xml.sometag(_class = 'value2', attr1='value1')['contents']
> xml.sometag({'class':'value2'}, attr1='value1')['contents']
> xml.sometag('class="value2"', attr1='value1')['contents']
> xml.sometag(('class', 'value2'),), attr1='value1')['contents']
>
> The preferred is the first, it's easiest.  It's assumed that you will
> never want an attribute that actually starts with an underscore.  If
> you do (probably never), you'll have to use one of the other three
> options.

Hey, thanks for piping up, Bob!  If you've got an implementation of
this, you must be the same Bob who sat on the couch at PyCon and coded
this stuff up while Donovan Preston and I were designing the
syntax... you sure were a Python-coding machine!  A pleasure to meet
you; I never caught your last name.  I guess if you still have an
implementation, it's still working out for you... (?)

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From ckea25d02 at sneakemail.com  Tue Apr  8 22:56:59 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 09 Apr 2003 14:56:59 +1200
Subject: Surprise: del weakref.WeakKeyDictionary[ item ] raises RuntimeError
In-Reply-To: <mailman.1049836994.31575.python-list@python.org>
References: <mailman.1049836994.31575.python-list@python.org>
Message-ID: <b701jn$9ls3d$1@ID-169208.news.dfncis.de>

Mike C. Fletcher wrote:
>        for ref in self.data.iterkeys():
>            o = ref()
>            if o == key:
>                del self.data[ref]
>                return
> 
> Using keys seems to avoid the 
> problem, but with the obvious disadvantages of keys vs. iterkeys.

Maybe a separate list of keys could be maintained
alongside the dict?

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From amk at amk.ca  Thu Apr 17 14:58:58 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Thu, 17 Apr 2003 13:58:58 -0500
Subject: Pythonic way of web-programming
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <B8Pma.20554$T34.656139@news2.tin.it> <mailman.1050471334.22000.python-list@python.org> <0gqdnfw_XuYm3gCjXTWcow@speakeasy.net> <11p2n-ci3.ln1@nb2.stroeder.com>
Message-ID: <bgCdnemerfrvZAOjXTWcoA@speakeasy.net>

On Thu, 17 Apr 2003 16:06:57 +0200, 
	Michael Str?der <michael at stroeder.com> wrote:
> I'm using a multi-thread FastCGI web application with mod_fastcgi. Or maybe 
> I did not get your point.

My point is that you're very rare; most people are not using multithreaded
FastCGI applications.

--amk



From zhitenev at cs.vsu.ru  Thu Apr 10 12:18:04 2003
From: zhitenev at cs.vsu.ru (Lexy Zhitenev)
Date: Thu, 10 Apr 2003 20:18:04 +0400
Subject: wxPython - Can you explain why I crash
References: <33803989.0304090652.174ac9de@posting.google.com>
Message-ID: <b745fs$c1k$1@serv.vrn.ru>

"Miki Tebeka" <tebeka at cs.bgu.ac.il> wrote in message:
news:33803989.0304090652.174ac9de at posting.google.com...
> I've written this little client. It used to work fine but now no
> matter which button I press the application crash.
> I'm running Python2.2 and wxPython 2.4.0.7
> Any ideas?
>

Nothing crashed on Python2.2.1, wxPython 2.3.3.1

I think there are some instability in your chain. I had many problems when I
used Stackless python DLL. This is not your program, this is, I think,
wxPython. Try to install a different version.]

Regards, Lexy.






From gh at ghaering.de  Tue Apr 15 06:07:10 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Tue, 15 Apr 2003 12:07:10 +0200
Subject: Width of tabs -- was: Re: vim configuration help
In-Reply-To: <b7gj7r$qh9$1@news.web.de>
References: <P4Oma.78974$rd4.2975295@twister.austin.rr.com> <b7gc3g$lb54$1@ID-141451.news.dfncis.de> <mailman.1050395849.18888.python-list@python.org> <b7gj7r$qh9$1@news.web.de>
Message-ID: <3E9BD9CE.7010006@ghaering.de>

Mirko Zeibig wrote:
> Gerhard Haering wrote:
>>Karl Scalet wrote:
>>>looking in my vimrc file: [...]
>>>set tabstop=4
>>>[...]
>>
>>No, no, no! That's evil!
>>
>>Tabs are 8 characters wide. Period.
> 
> says who?

If the Python interpreter encounters a tab character in Python source 
code, it treats it as an equivalent of an indentation of 8 spaces.

Not knowing about this fact will only bite you if you mix tabs and 
spaces in Python sourcecode, which is a really silly thing to do in the 
first place (and python -t will warn you about it).

There's plenty of other software and hardware that behaves the same, 
though I'm not aware of any standard that would mandate this behaviour.

-- Gerhard





From tlavoie at acm.org  Tue Apr 15 11:19:02 2003
From: tlavoie at acm.org (Tim Lavoie)
Date: 15 Apr 2003 10:19:02 -0500
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1049901964.20840.clpa-moderators@python.org> <b72fo5$bq6$1@panix1.panix.com> <87ptnuxryk.fsf@theasylum.dyndns.org> <b7g2fk$g1u$1@panix3.panix.com> <87smsj4z4t.fsf@theasylum.dyndns.org> <h2Vma.17871$y3.1620573@news010.worldonline.dk>
Message-ID: <87of374wm1.fsf@theasylum.dyndns.org>

>>>>> "Anders" == Anders J Munch <andersjm at inbound.dk> writes:

    Anders> "Tim Lavoie" <tlavoie at acm.org> wrote:
    aahz> So how does Capon differ from SCons?
    >> >> Well, the name has funnier connotations, for one.  :)
    >> 
    aahz> You deserve to be roasted for that comment.
    >>  Well, sometimes you stick your neck out...  "Capon: Batteries
    >> included, nuts aren't."  :)

    Anders> <g> actually the name comes from the boring "Cascading
    Anders> property notation".  Gives me an idea for a slogan though:
    Anders> "Less is more".

    Anders> nah-maybe-not-ly y'rs, Anders

 *SNORT* You know, I'm really glad I wasn't slurping my coffee while
 reading this one. It's bad enough that I'll have to explain to people
 why I'm giggling.

  Cheers,
  Tim

-- 
"I went to the museum where they had all the heads and arms from the
statues that are in all the other museums."
                -- Steven Wright



From aleax at aleax.it  Sat Apr 26 06:26:15 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 10:26:15 GMT
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.29.48.204375@soul.net> <RZlqa.4182$lf3.712485@cletus.bright.net>
Message-ID: <b9tqa.9443$K35.308314@news2.tin.it>

George Kinney wrote:

> "lost" <lost at soul.net> wrote in message
> news:pan.2003.04.25.23.29.48.204375 at soul.net...
>> On Fri, 25 Apr 2003 22:51:12 +0000, Alex Martelli wrote:
>> Hi Alex,
>> sorry i wasn't a bit more specific in what i needed, but what you posted
>> was exactly what i was looking for!  Thanks a lot.
>>
> 
> FWIW, theres also:
> 
> from glob import glob
> file_list = glob.glob(SomePath)
> outfile = open(OutputFilename, 'w')
> 
> for file in file_list :
>     print file
>     outfile.write(file + '\n')
> 
> outfile.close()

Yes, but - I suggest you AVOID using names of builtin types as names
for your own variables; it's a habit that will eventually "bite" you,
wasting some of your time in the chase for some obscure bug.  Just
code the for loop more readably as:

for filename in file_list :
    print filename
    outfile.write(filename + '\n')

and I think the quality of your code gets substantially enhanced.


Alex




From aleax at aleax.it  Thu Apr 17 13:35:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 17:35:39 GMT
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de> <mailman.1050586468.29594.python-list@python.org> <rozna.115488$It5.70451@news2.central.cox.net> <b7mn1i$2icfr$2@ID-148859.news.dfncis.de>
Message-ID: <LBBna.43583$T34.1210957@news2.tin.it>

Haran Shivanan wrote:

> module "Steve Holden" core dumped with the following output:
>> 
>>>>> class A:
>> ...   a = 0
>> ...   b = 1
>> ...
>>>>> class B(A):
>> ...   pass
>> ...
>>>>> b = B()
>>>>> A.a = 23
>>>>> b.a
>> 23
>>>>>
>> 
>> I suspect the OP wants a separate class variable for each subclass.
>> Right, Haran?
> Right.
> The variables in the subclass should not be bound to the parent.
> Any ideas?

Then you need to explicitly request copies.  What runs when B is
*created* (as opposed to, when B is *instantiated*) is the
initializer of B's metaclass, so to get non-standard behavior at
class-creation time you need a custom metaclass (which in practice
means your B class will be close to a new-style class, because
subclassing type is far and away the sensible way to get custom
metaclasses) -- might as well stick it in A, as long as you're
at it, so all descendants of A will automatically inherit it.

Basically, you want to re-bind (to the same name, and to a shallow
copy of the value, but in B itself) all the attributes of (direct?) 
base classes that are neither special (magical names) nor private 
(name starts with underscore) nor callable (if the exact condition 
you want to apply is different, adjust the following code, of course).

So, for example, and assuming you don't care about supporting 
multiple inheritance or only want the "copy behavior" from the 
one leftmost direct bast class if any:

import copy

class CopyingMeta(type):
    def __new__(metaclas, clasname, bases, clasdict):
        if bases:
            for n, v in vars(bases[0]).iteritems():
                if n in clasdict or n.startswith('_'
                    ) or callable(v): continue
                clasdict[n] = copy.copy(v)
        return type.__new__(metaclas, clasname, bases, clasdict)

class A:
    __metaclass__ = CopyingMeta
    a = 11
    b = 22

class B(A): pass

b = B()
A.a = 23
print b.a


This prints 11, as apparently required by your specs.  I hope
this can help you make the custom metaclass you need to
implement your actual, exact specifications, of course.


Alex




From member27604 at dbforums.com  Tue Apr  8 21:32:03 2003
From: member27604 at dbforums.com (dolephi9080)
Date: Wed, 09 Apr 2003 01:32:03 +0000
Subject: [help]Is this a bug or my code's fault
References: <2741546.1049794193@dbforums.com>
Message-ID: <2745274.1049851923@dbforums.com>

Thank you, your explanation is enough for me.

--
Posted via http://dbforums.com



From riemer at lincvs.org  Mon Apr  7 17:18:58 2003
From: riemer at lincvs.org (Tilo Riemer)
Date: Mon, 7 Apr 2003 23:18:58 +0200
Subject: Announce: apm.py 0.1.0
Message-ID: <20030407211858.GA19329@ppprs1.phy.tu-dresden.de>

Hello,

apm.py 0.1.0 is released.

The goal of apm.py is providing of an uniform and platform independent
interface to APM. At the moment it implements only battery relevant
functions. The interface is not stable yet.

There are implementations for NetBSD 1.6 and Linux 2.4. An implementation
for FreeBSD 4 coming soon.

Because I don't have dozens of laptops and operating systems I need help! It
would be great to have a Python modul which covers APM more or less
completely for many platforms one day.



Best regards, Tilo




From python at dohao.org  Tue Apr 22 01:36:46 2003
From: python at dohao.org (Wenshan Du)
Date: 21 Apr 2003 22:36:46 -0700
Subject: Zope and FreeBSD
References: <mailman.1050931234.15277.python-list@python.org>
Message-ID: <e786d63d.0304212136.42e01fa5@posting.google.com>

thanks, Andrew



From jxlu at fudan.edu.cn  Tue Apr  8 21:26:26 2003
From: jxlu at fudan.edu.cn (Lu Jian Xiong)
Date: Wed, 09 Apr 2003 09:26:26 +0800
Subject: It seems that some *.o file is not installed, but how can I get them?
Message-ID: <0HD100EE2YK4UG@mail.fudan.edu.cn>

Hello,
I am a newbie of python.
When I was trying the example read from the following URL,
http://www.python.org/doc/current/ext/simpleExample.html,
it showed the following error message:

#It seems that some *.o file is not installed, but how can I get them?
#What should I do to find which RPM or tgz package they are included with?

[user0 at localhost user0]# cat spammodule.c
#include <Python.h>
static PyObject *
spam_system(self, args)
    PyObject *self;
    PyObject *args;
{
    char *command;
    int sts;

    if (!PyArg_ParseTuple(args, "s", &command))
        return NULL;
    sts = system(command);
    return Py_BuildValue("i", sts);
}
int main()
{
    return (0);
}
[user0 at localhost user0]$ gcc -I /usr/include/python2.2 spammodule.c -lpython2.2
/usr/bin/ld: cannot find -lpython2.2
collect2: ld returned 1 exit status
[user0 at localhost user0]$ gcc -I /usr/include/python2.2 -L /usr/lib/python2.2/config spammodule.c -lpython2.2
/usr/lib/python2.2/config/libpython2.2.a(posixmodule.o): In function `posix_tmpnam':
posixmodule.o(.text+0x3977): the use of `tmpnam_r' is dangerous, better use `mkstemp'
/usr/lib/python2.2/config/libpython2.2.a(posixmodule.o): In function `posix_tempnam':
posixmodule.o(.text+0x385e): the use of `tempnam' is dangerous, better use `mkstemp'
/usr/lib/python2.2/config/libpython2.2.a(complexobject.o): In function `_Py_c_pow':
complexobject.o(.text+0x37c): undefined reference to `pow'
complexobject.o(.text+0x3bc): undefined reference to `cos'
complexobject.o(.text+0x3cd): undefined reference to `sin'
complexobject.o(.text+0x3e6): undefined reference to `exp'
complexobject.o(.text+0x3f7): undefined reference to `log'
/usr/lib/python2.2/config/libpython2.2.a(floatobject.o): In function `float_pow':
floatobject.o(.text+0x138a): undefined reference to `pow'
/usr/lib/python2.2/config/libpython2.2.a(dynload_shlib.o): In function `_PyImport_GetDynLoadFunc':
dynload_shlib.o(.text+0x1ca): undefined reference to `dlopen'
dynload_shlib.o(.text+0x207): undefined reference to `dlsym'
dynload_shlib.o(.text+0x217): undefined reference to `dlerror'
/usr/lib/python2.2/config/libpython2.2.a(thread.o): In function `PyThread_start_new_thread':
thread.o(.text+0xb3): undefined reference to `pthread_sigmask'
thread.o(.text+0xd9): undefined reference to `pthread_create'
thread.o(.text+0xf3): undefined reference to `pthread_sigmask'
thread.o(.text+0x12a): undefined reference to `pthread_detach'
/usr/lib/python2.2/config/libpython2.2.a(posixmodule.o): In function `posix_openpty':
posixmodule.o(.text+0x17fe): undefined reference to `openpty'
/usr/lib/python2.2/config/libpython2.2.a(posixmodule.o): In function `posix_forkpty':
posixmodule.o(.text+0x188f): undefined reference to `forkpty'
collect2: ld returned 1 exit status
[user0 at localhost user0]$ slocate thread.o
[user0 at localhost user0]$ slocate python.o
/usr/lib/python2.2/config/python.o
[user0 at localhost user0]$ rpm -qa| grep ^python
python-docs-2.2.2-7
python-optik-1.3-2
python-tools-2.2.2-7
python-devel-2.2.2-7
python-2.2.2-7
[user0 at localhost user0]$ rpm -ql python-devel| grep [.]o\$
/usr/lib/python2.2/config/python.o







From eppstein at ics.uci.edu  Mon Apr 21 20:57:13 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Mon, 21 Apr 2003 17:57:13 -0700
Subject: Regarding Dictionaries in python
References: <mailman.1050957134.26691.python-list@python.org>
Message-ID: <eppstein-D6EFEA.17571321042003@news.service.uci.edu>

In article <mailman.1050957134.26691.python-list at python.org>,
 "Mehta, Anish" <Anish.Mehta at enst-bretagne.fr> wrote:

> why it is not showing the items in order in which i have inserted. 
> Please tell me why it is happing like this.

Because dictionaries store things out-of-order, in hash tables, so that 
they can perform lookups quickly.  If maintaining the order of some 
things is important, that's what lists are for.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From patrick.bussi at free.fr  Sun Apr  6 11:11:59 2003
From: patrick.bussi at free.fr (Patrick Bussi)
Date: Sun, 6 Apr 2003 17:11:59 +0200
Subject: ClientCookie.read() failing on some servers
Message-ID: <20030406152829.BEE11C131@postfix3-1.free.fr>

Hello,

I'm calling for help to diagnose this fail of ClientCookie.py as an URL 
opener (just in in place of urllib2.ulropen()).

Here is my simple program :

-------snip-------
#! /usr/bin/env python
'''Simple test of urllib2 request with ClientCookie opener
   usage:
   $ python test.py www.python.org
'''
import sys
import ClientCookie, urllib2, urllib
from urllib2 import Request

def timeStamp():
    import time
    t=time.time()
    return t, time.ctime(t)

def test(h):
    '''h is the host name. Caution : no protection against mistakes
    '''
    req = urllib2.Request('http://'+h)
    req.add_header('Host',h)
    req.add_header('User-agent', 
'Mozilla/5.5.(X11;.U;.Linux.2.4;.en-US;.0.8).Gecko/20010409')
    req.add_header('Accept','*/*')
    req.add_header('Accept-Language','en')
    req.add_header('Accept-Encoding','gzip,deflate,compress,identity')
    req.add_header('Keep-Alive','300')
    req.add_header('Connection','keep-alive')

    print '\n'.join(['%s' %k for k in (req.get_full_url(), 
req.get_selector(), req.headers)])
    response=ClientCookie.urlopen(req)
    print '\n2\n', response.info()
    if response: print response.read()
    print '\n3\n', '%s -  Data for server %s was :\n%s' %(timeStamp()[1], 
req.get_host(), req.get_data())


if __name__ == '__main__':
    import os, sys
    print 'Linux', ''.join(['%s'*3 %os.uname()[2:]])
    print 'Python', sys.version
    try:
        print sys.argv[1]
    except:
        print __doc__, '\nPython interpreter is :', sys.version
        sys.exit('Bad news')
    test(sys.argv[1])

-------snap-------


-------snip-------


-------snap-------


-------snip-------


-------snap-------





From cben at techunix.technion.ac.il  Sun Apr 13 13:31:45 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Sun, 13 Apr 2003 20:31:45 +0300 (IDT)
Subject: a size-intelligent progress bar?
In-Reply-To: <mailman.1049910978.11449.python-list@python.org>
References: <1PYka.1140$RX1.361@newsfep3-gui.server.ntli.net>
 <mailman.1049910978.11449.python-list@python.org>
Message-ID: <Pine.GSO.4.51.0304132026220.26349@techunix.technion.ac.il>

Michael Stenner wrote on 2003-04-09:

> On Wed, Apr 09, 2003 at 06:20:02PM +0100, Tom Chance wrote:
> > Hullo,
> >
> > I'm making a shell-based program that includes a progress bar (like wget's
> > one). At the moment, I've got it printing an equals sign every 2% so I get
> > a progress bar 50 "="s wide, which fits my Konsole perfectly. This isn't
> > very handy for other shell sizes though :)
> >
> > Does anyone have, or can anyone point me in the direction of a way of making
> > a decent progress bar that will automagically detect the width of the shell
> > open, even resizing it mid-"progress"?
>
> I don't know how portable it is, but the environment variable COLUMNS
> appears to tell me my terminal width and follows resizes.  If you
> check that on each update, it should do what you want.
>
It's there on any normal unix.  It is set up correctly in 95% of the
cases and relying on it is OK (many command-line programs do) - if
it's not the user will eventually get annoyed and fix it by 'eval
`resize`' ;-).  However it's *not* going to be updated when the
terminal resizes - it's impossible for an environment variable your
program sees to change after your program has started (unless the
program itself changes it).  The best way in a shell script would be
to take the output of `tput cols` every time; if you don't like to
launch a process every time, turn to ioctl/curses...

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>

Never spend several weeks on a single homework - or you will miss a
nice Linux conference ;-(.




From ckea25d02 at sneakemail.com  Tue Apr  8 20:51:43 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 09 Apr 2003 12:51:43 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1049814375.24609.python-list@python.org>
References: <m2u1dcw22s.fsf@localhost.localdomain>        <3E9040CF.2040606@removethis.free.fr>        <m2k7e7addl.fsf@localhost.localdomain>        <b6pgcm$1t3$1@panix3.panix.com>        <m27ka7aae0.fsf@localhost.localdomain>        <6ee58e07.0304061722.2472b3d0@posting.google.com>        <3E9175A5.1060607@kyborg.dk>        <yfsof3ijmua.fsf@black132.ex.ac.uk>        <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org>
Message-ID: <b6vq8r$9qev7$1@ID-169208.news.dfncis.de>

Skip Montanaro wrote:
> Look again.  In Elisp, when foo() looks up the value of case-fold-search it
> will find the value defined in the let statement inside bar() because it's
> lexically scoped  (looks up the call chain for values).

Your terminology is confused -- that's *dynamic* scoping.

> If this was Python,
> the reference to case-fold-search in foo() would be resolved at the global
> scope because Python is statically scoped (looks in locals, globals, then
> builtins only

*This* is called lexical scoping (since the scope is determined
by where the code is written). (Although calling it "static"
is correct too, since it's not dynamic.)

 > ignoring nested functions which this case doesn't involve).

In recent Pythons, this is no longer true -- intermediate
lexical scopes are searched, too (although only for _reading_,
just to keep things Pythonically perverse.:-)

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From wedjlok at attbi.com  Mon Apr 28 19:37:39 2003
From: wedjlok at attbi.com (Eric Anechiarico)
Date: Mon, 28 Apr 2003 23:37:39 GMT
Subject: saving output to file
References: <MUera.658292$L1.187296@sccrnsc02> <3ead7e89$0$49116$e4fe514c@news.xs4all.nl>
Message-ID: <7Xira.661399$F1.86830@sccrnsc04>

Correct, I am just having the results print out directly within the IDE
window.

That looks as if it what I was looking for and I thank you for the response.
I will give it a shot.

One more thing, though.  Do I need to specify a path to the resultfile that
I will be writing to, or will it just default to the directory that Python
resides in when I execute it?

Eric


"Irmen de Jong" <irmen at -NOSPAM-REMOVETHIS-xs4all.nl> wrote in message
news:3ead7e89$0$49116$e4fe514c at news.xs4all.nl...

>
> I'm guessing that you are now "print"-ing the results of your calculations
> to the screen, as in:
>
> def fib(n):
>     a,b = 0,1
>     while b<n:
>         print b,
>         a,b = b,a+b
>
> Instead of printing them to the screen, print them to a file:
>
> def fib(n, resultfile):
>     a,b = 0,1
>     while b<n:
> print >>resultfile, b,
>         a,b = b,a+b
>
> fib(2000, open("fibonacci_to_2000.txt",'w'))
>
>
> You might also want to check the following parts of the manual:
>
http://www.python.org/doc/current/tut/node9.html#SECTION00920000000000000000
0
> http://www.python.org/doc/current/lib/module-StringIO.html
>
> If you want you can even replace the 'default' standard output (your
> screen) with your own file object:
>
> sys.stdout=open("my_stdout.txt",'w')
>
> Does this make sense?
>
> --Irmen
>





From antispam.rvercoutere at ge-serveur.insa-lyon.fr  Thu Apr  3 10:50:26 2003
From: antispam.rvercoutere at ge-serveur.insa-lyon.fr (Roro)
Date: Thu, 3 Apr 2003 17:50:26 +0200
Subject: [wxPython] wxMDIChildFrame
Message-ID: <b6hl4f$hfj$1@demo2.univ-lyon1.fr>


Hello everybody,

I want to create a window with others windows inside : My goal is to open
several Images and apply filters on it.

In concrete terms, my main window is with a bar menu : I click on  'Open
File' Item and select images witch I wanna see.
And I open  Images in a window by image, and I use wxMDIParentFrame and
wxMDIChildFrame to create my windows.


To create Windows, I use these lines :

 def OnNewWindow(self, evt, ):
        / Open Dialog window witch ask witch files to open
        /
        self.winCount = self.winCount + 1
        self.win = wxMDIChildFrame(self, Self.winCount+100, "Child Window:
%d" % self.winCount)
        / ...
        /
        self.win.Show(true)

The problem comes now, I want to select a window and apply a filter on it
(filters are in my menu bar). So, I want to select image's window, click in
the menu bar the filter and see the result in the  selected window.
I can only apply my filter on the last opened window. the others windows
become ghost-windows (I can see them but i can't kill them). I think, it's
because 'win' points on the last window and run over the previous created
window.
I want to manage my windows, but I didn't find any solution.

Is it possible to create an indirection for 'win', by passing an argument
with the new name for the window  ?
Can I use 'Self.winCount' to refer to my ghost-windows ? Using their ID
number ?


Thanks in advance for any help.





From jjl at pobox.com  Thu Apr 24 14:08:47 2003
From: jjl at pobox.com (John J. Lee)
Date: 24 Apr 2003 19:08:47 +0100
Subject: Python 2.2.2 and 2.3a2 compile problems on Mingw32
References: <mailman.1051001914.4729.python-list@python.org> <87he8q4993.fsf@pobox.com> <mailman.1051042335.19989.python-list@python.org>
Message-ID: <87fzo7kbtc.fsf@pobox.com>

Gerhard H?ring <gh at ghaering.de> writes:

> John J. Lee wrote:
> > "Andreas KLostermann" <bluetigger at web.de> writes:
> >
> >>I wanted to compile Python on Mingw32. That's a GNU C++ Compiler for
[...]
> > Sorry if this is a stupid question, but are you using the cygwin
> > port?
> 
> Apparently Andreas was trying to compile Python using the *native*
> win32 GNU toolset, aka mingw (minimalist GNU for Windows). This isn't
[...]

D'Oh.  Didn't read the question properly...


John



From syver-en+usenet at online.no  Sun Apr  6 16:31:30 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 06 Apr 2003 22:31:30 +0200
Subject: Debugging embedded python
Message-ID: <u4r5bgyfx.fsf@online.no>

I have the python interpreter embedded in a daemon/service (this is
win2k so it's a service). I'd really like to be able to debug the
python code that is running. I was thinking of popping up a console
window to be able to run pdb, how does one go about it to accomplish
this? Are there any code doing similar things or articles.


-- 

Vennlig hilsen 

Syver Enstad



From tjreedy at udel.edu  Fri Apr 11 08:48:50 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Fri, 11 Apr 2003 08:48:50 -0400
Subject: string() - Is this some kind of sick joke?
References: <36dbf824.0304091142.527af834@posting.google.com> <1dpc9v4beon9upq2sd3clfb2plh7nlqm7p@4ax.com>
Message-ID: <J0GdnakOdOO8JQujXTWcqQ@comcast.com>

> >What gives?  Did I get the bogus version?  Is this wishful thinking
> >documentation?  I also tried  y.strip("'")  with the same results.
>
> I get that, too, on string.strip(y,"'"), but y.strip("'") works
fine:

I believe the inconsistencies have been fixed for the future 2.2.3 and
2.3.b1.  In the meanwhile, use the new string methods.

tjr





From mis6 at pitt.edu  Fri Apr 25 17:31:42 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 25 Apr 2003 14:31:42 -0700
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com> <SXcqa.6797$K35.186991@news2.tin.it> <b8bmaq$mhl$1@panix3.panix.com>
Message-ID: <2259b0e2.0304251331.620356e5@posting.google.com>

aahz at pythoncraft.com (Aahz) wrote in message news:<b8bmaq$mhl$1 at panix3.panix.com>...
> I've got one throwaway script, called x.py.  Any time I need to write a
> throwaway script, I edit that file, deleting the current contents first.
> (I'll admit that I do usually have multiple instances of that file in
> different directories, but the ones in directories get used strictly for
> the projects in those directories -- my standard is the one in my home
> directory.)

I have also x.txt, x.html, x.tex, ... ;)



From news03 at zurgy.org  Tue Apr 22 06:19:00 2003
From: news03 at zurgy.org (Ben Lamb)
Date: Tue, 22 Apr 2003 11:19:00 +0100
Subject: Compiling a program the embeds python
Message-ID: <3ea51711$0$4862$cc9e4d1f@news.dial.pipex.com>

I'm trying to compile an example program written in C that embeds Python.
The porgram is:

#include <Python.h>

int main(int argc, char *argv[])
{
        Py_Initialize();
        PyRun_SimpleString("from time import time,ctime\n"
                        "print 'Today is ',ctime(time())\n");
        Py_Finalize();
        return 0;
}


I use the following to compile it:

ben at localhost:~/prog/pyembed$ gcc -I/usr/local/include/python2.2
-L/usr/local/lib/python2.2/config -Xlinker -export-dynamic -lm -lpython2.2
-ldl -lpthread -lutil pyembed.c
/tmp/ccLcdPKW.o: In function `main':
/tmp/ccLcdPKW.o(.text+0x7): undefined reference to `Py_Initialize'
/tmp/ccLcdPKW.o(.text+0x14): undefined reference to `PyRun_SimpleString'
/tmp/ccLcdPKW.o(.text+0x1c): undefined reference to `Py_Finalize'
collect2: ld returned 1 exit status


I installed Python 2.2 on my Linux system by compiling it from sources with
the default options. I've tried using GNU C 2.95 and 3.2, ld 2.12.90 but
get the same errors.

Python runs fine on the system and I have successfully built other programs,
such as modpython, that embed it. modpython seems to build itself using
exactly the same options as I am trying above.

There are no libpython.so files on my system, just libpython2.2.a that lives
in /usr/local/lib/python2.2/config

Please can someone point out what I am doing wrong.

Thanks,

Ben Lamb.



From aleax at aleax.it  Wed Apr 23 04:33:19 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 08:33:19 GMT
Subject: i18n in a library
References: <b85hjt$6l2na$1@ID-15245.news.dfncis.de>
Message-ID: <jdspa.47509$LB6.1087058@news1.tin.it>

<posted & mailed>

Mathias Waack wrote:

> I'm using some different applications which all use the same package.
> Lets assume there are app "A" and "B" and a package "L". Now I've got
> a question regarding the translation of these apps:
> 
> It's easy to prepare translation using gettext.py. But I don't want
> to include all strings from the library in each app. So I get 3
> files: A.po, B.po and L.po. Is it possible to setup the gettext
> module to use both A.po and L.po to lookup strings for app A?
> 
> I think there must be a simple solution because this should be a
> common problem for each library writer. But I can't find a solution
> in the python docs:(

I think what you want is gettext.translation, which offers you
finer-grained control than gettext.install.  I.e., within the
modules of library L, you could use

    import gettext
    _ = gettext.translation('L').ugettext

so that in those modules function _ will look things up in domain
'L', rather than in whatever domain is "installed" globally.


Alex




From exarkun at intarweb.us  Tue Apr 15 14:10:56 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 14:10:56 -0400
Subject: Twisted docs [was: Re: A python server]
In-Reply-To: <15f445fb.0304150911.385c7202@posting.google.com>
References: <15f445fb.0304140751.170d410e@posting.google.com> <3E9ADED2.280A671F@engcorp.com> <15f445fb.0304150911.385c7202@posting.google.com>
Message-ID: <20030415181056.GD32515@meson.dyndns.org>

On Tue, Apr 15, 2003 at 10:11:18AM -0700, dansan wrote:
> [snip]
> 
> That is what I would like.  If twisted would allow me to concentrate
> on the game-logic, and tournament logic, that would be great... but I
> have no clue where to start.

  I'm not sure what you've looked at yet, but these might be of use:

    http://twistedmatrix.com/documents/howto/servers
    http://twistedmatrix.com/documents/howto/clients
    
  or taking the Perspective-Broker approach:

    http://twistedmatrix.com/documents/howto/pb-usage

  For either one, looking at

    http://twistedmatrix.com/documents/howto/async
    http://twistedmatrix.com/documents/howto/reactor-basics

  first is probably a good idea.  Of course, the Howtos are arranged in what
we think is the most reasonable order to read them, and I've just pulled
these five out of the main list, so if something doesn't make sense, going
back and reading some of the earlier ones will probably help.  Specific
questions are always welcome on the Twisted mailing list, so if you get
stumped, by all means - ask.

  Looking at examples is often helpful, too.  So you might want to check out
the doc/examples/ directory in the source tarball or in CVS.

  Jp

-- 
http://catandgirl.com/view.cgi?90
-- 
 up 26 days, 15:02, 4 users, load average: 0.29, 0.10, 0.02




From haris at email.hinet.hr  Mon Apr 28 08:17:38 2003
From: haris at email.hinet.hr (Haris Bogdanovic)
Date: Mon, 28 Apr 2003 14:17:38 +0200
Subject: pyrex exceptions
Message-ID: <b8j67p$aqpk$1@as201.hinet.hr>

How do I handle exceptions in Pyrex ?
I have a C++ wrapper around some C library and I'm doing a Python wrapper
around that library (copying C++ wrapper object model).
In C++ wrapper there is a class Exception which is derived from C++ standard
exception class.
In Pyrex it is not yet possible to derive an extension type from Python
class so I would have to create extension type for handling exceptions from
scratch.
If I want to raise an exception like this :
raise ExceptionClass(error)
how should then ExceptionClass be defined ?
I'm using wxPython GUI so I want to show the message box (window) which
displays error text when raising exception and then let user to choose from
options of ending or continuing the program from some previous place in code
or correcting error which caused the exception.

Can you please give me just the bare bones of how should that exception
class look like.(in Pyrex, as extension type)
I'm also little bit unfamiliar with what exactly goes on when program raises
an error with class instance as attribute. Could you explain that just in a
few words ?

Thank you
Haris





From BPettersen at NAREX.com  Wed Apr 16 15:37:55 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 16 Apr 2003 13:37:55 -0600
Subject: Big-O notation
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDF8E@admin56.narex.com>

> From: Ken R. [mailto:kriley1 at hotmail.com] 
> 
[...]
> It's been nearly forever since my compsci algorithms class so be
> gentle if I'm glaringly wrong :). 

ditto.

[...]
> No one decides what order an algorithm is, it is an inherant property 
> of the algorithm.

... which implies that it must have been _proven_ (it's a theoretical
system :-)

> If you execute quicksort on a wide variety of sample sets of different
> sizes then measure and plot the processor time vs sample size, it will
> approximate the graph of NlogN. Therefore the algorithm has order
> NLogN. The same will apply to bubble sort (order N^2), etc.
[...]

A commonly held belief, but as I said, it has to be proven, not
measured. Importantly, the statement entirely depends on the innocent
phrase "wide variety of sample [data]" (which would be valid if you were
trying to approximate the statistical mean of the algorithm). When
measuring, most people (should be?) are interested in _expected_
performance however, which means you need a "representable sample" (of
the domain), but even when well defined and acknowledged
programmers/scientists/... have historically proven to be highly
subjective when determining 'representable' <wink>. [Hint: you're
assuming you're only sorting random data].

-- bjorn




From staschuk at telusplanet.net  Wed Apr 30 00:45:46 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 22:45:46 -0600
Subject: Raising objects
In-Reply-To: <ltxra.31737$K35.1001968@news2.tin.it>; from aleax@aleax.it on Tue, Apr 29, 2003 at 04:09:53PM +0000
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it>
Message-ID: <20030429224546.B3128@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
  [...]
> No, in 2.2 (and 2.3 beta 1) exceptions must still be classic-classes
> or instances thereof [or strings for backwards compatibility with the
> dark and distant past], and should be derived from Exception (but this
> is a recommendation, not enforced).  I do not know when (or if) this
> restriction will be removed -- in particular, because I do not know
> of "use cases" where the ability to raise new-style classes or instances
> thereof would be a major boon to an application.  Maybe you have some...?

I don't have a real use case either, but it does seem rather
arbitrary.  Is there an implementation reason for this restriction?

A somewhat silly possible use:  A depth-first search to find a
single object satisfying a given property.  An implementation
(ab)using exceptions for control flow:

    def _search(node):
        if condition(node):
            raise node
        else:
            for child in childrenof(node):
                _search(child)

    def findone():
        try:
            _search(root)
        except Node, found:
            return found
        else:
            return None

If the objects in the search tree are instances of new-style
classes, the restriction would be inconvenient.  It would be
simple enough -- though ugly -- to wrap them in an old-style
class for raising, of course.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"[T]rue greatness is when your name is like ampere, watt, and fourier
 -- when it's spelled with a lower case letter."      -- R.W. Hamming




From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Mon Apr 21 06:13:18 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Mon, 21 Apr 2003 12:13:18 +0200
Subject: Multiple assignment?
In-Reply-To: <mailman.1050918753.29031.python-list@python.org>
References: <mailman.1050918753.29031.python-list@python.org>
Message-ID: <3ea3c43d$0$49102$e4fe514c@news.xs4all.nl>

Mehta, Anish wrote:
> Can someone tell me the what is being done here in this line of code.
> 
> dir, file = os.path.split(input)
> 

Your message title was already on the right track.
This is a Tuple Assignment (also called Tuple Unpacking).

os.path.split returns a tuple with 2 elements; (directoryname, filename).
Using tuple unpacking these two elements of the tuple
are assigned to "dir" and "file", respectively.


--Irmen

PS: bad idea to use "dir" and "file" because
you'll hide the builtin "dir" and "file" identifiers.




From aleax at aleax.it  Mon Apr 14 13:04:32 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 17:04:32 GMT
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it> <a8b7f07a.0304140646.71675ca1@posting.google.com>
Message-ID: <ASBma.13064$LB6.352323@news1.tin.it>

A. Lloyd Flanagan wrote:

> Alex Martelli <aleax at aleax.it> wrote in message
> news:<6EWla.4594$LB6.132415 at news1.tin.it>...
>> A. Lloyd Flanagan wrote:
>> 
>> > Unless you absolutely have to have it, don't use it.
>> 
>> "Absolutely have to have it", just like "screws up big time", are
>> silly overbids IMHO -- over-reactions that I don't see as justified
>> at all!  Why can't we all just keep SOME perspective, please...?!
>> 
> 
> Perspective?  Moi??
> 
> Seriously, I was going to clarify my statement, then realized that
> you'd stated the trade-offs better than I could.  And thanks very much
> for the benchmarks, that kind of data is really helpful.

Always glad to be of help!


> Perhaps I should add a corollary to Knuth's law:  "Premature brevity
> is the root of all misunderstandings"?

Nice observation;-).  Would need expanding, though -- communication
needs some shared starting point, and while assuming more of a start
than is present may induce misunderstanding, the reverse assumption
inevitably means high verbosity...


> I think the root of the issue here is that the documentation could be
> clearer about when to use __slots__ and when not to.  We could cut and
> paste much of your last message into the docs, and it would help a
> lot.  How do changes and updates make it into the official docs,
> anyway?

Changes are made to the official docs just like to any other part of
Python - you prepare a proposed patch (to the LaTeX sources, but, with
some imitation of surrounding context, you don't really need to "get"
LaTeX for that purpose...) and submit it to sourceforge -- the doc
czar, Fred Drake, eventually accepts or rejects the patch, or maybe
suggests some rework before it can be accepted.  You're welcome to
use my previous post, in whole or in part, to prepare such a patch,
if you wish (can't do it myself right now...).


Alex




From mathias-usenet at valpo.de  Wed Apr  2 05:13:31 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 2 Apr 2003 10:13:31 GMT
Subject: Python equivalent of _tempnam?
References: <3e8ab45b$1@mail.hmgcc.gov.uk>
Message-ID: <b6ed4b$4daat$1@ID-15245.news.dfncis.de>

"Richard" <richardd at hmgcc.gov.uk> wrote:

> I am writing an application which produces file output. Files
> are outputted to a directory and I want to ensure that if an
> output filename is given that already exists in that directory a
> number is affixed to the filename to ensure that the existing
> file is not appended to or overwritten. i.e. output_0.txt, then
> output_1.txt etc... 
> 
> I have been told that a way of doing this in C is by using
> _tempnam( ). 

Where is this function defined? There exists a function tmpnam(), 
which use is depracted and dangerous. Maybe you're talking about 
mkstmp(), which _can_ work in nearly the same manner as described 
above. 

> However I want to write this in Python, so is there
> an equivalent function? 

Maybe you should look into the tempfile module. 

> My application will only be executed
> occasionally so I am unable to keep an internal record of file
> numbering. I assume that I need to find the highest number file
> (if exists) in the directory and then increment that by one 
> before creating the new file. 

Which offers a great starting point for nearly all kinds of race 
conditions;)

Mathias



From skip at pobox.com  Tue Apr 22 12:37:01 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 22 Apr 2003 11:37:01 -0500
Subject: xmlrpc introspection
In-Reply-To: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com>
References: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com>
Message-ID: <16037.28589.172440.897944@montanaro.dyndns.org>

    Bryan> i'm interested in implementing the xmlrpc introspecton methods
    Bryan> ...

Google is your friend.  Searching for "xmlrpc introspection" yields this
promising item.

    http://xmlrpc-c.sourceforge.net/xmlrpc-howto/xmlrpc-howto-api-introspection.html

Skip




From max at alcyone.com  Mon Apr 28 17:26:20 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 28 Apr 2003 14:26:20 -0700
Subject: Coding style and Python versions (was Re: Another itertool 
 function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com> <zAara.22169$K35.739509@news2.tin.it>
Message-ID: <3EAD9C7C.B4C9B8F1@alcyone.com>

Alex Martelli wrote:

> There is no special reason to "change that habit", e.g. not
> performance:

Brevity sure can be:

    veryLongIdentifierToIncrement = veryLongIdentifierToIncrement + 1

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Give me chastity, but not yet.
\__/ St. Augustine
    The laws list / http://www.alcyone.com/max/physics/laws/
 Laws, rules, principles, effects, paradoxes, etc. in physics.



From nagylzs at freemail.hu  Fri Apr 11 08:24:33 2003
From: nagylzs at freemail.hu (=?ISO-8859-1?Q?Nagy_L=E1szl=F3_Zsolt?=)
Date: Fri, 11 Apr 2003 14:24:33 +0200
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com> <b766fi$p3b$1@newstree.wise.edt.ericsson.se> <16411664.REEpLcBOk5@lothar.cengija.com>
Message-ID: <3E96B401.5070006@freemail.hu>

>
>
>other things -- so it isn't a real program'. Now I'm looking for a free 
>install-shield like software.
>
Try NSIS (installer of winamp, made by NullSoft). It is free, makes a 
compact installer and it has support for displaying eulas (and many more).

I met some people in the US who saw my perfect NSIS installer and payed 
me to write a much bigger, much slower, much worse setup with 
InstallShield for the same program. Crazy people. They will never be 
satisfied with a good program. They have false dreams and they want to 
see them in real. :-(

Laci 1.0






From jjl at pobox.com  Tue Apr 22 15:34:00 2003
From: jjl at pobox.com (John J. Lee)
Date: 22 Apr 2003 20:34:00 +0100
Subject: Python 2.2.2 and 2.3a2 compile problems on Mingw32
References: <mailman.1051001914.4729.python-list@python.org>
Message-ID: <87he8q4993.fsf@pobox.com>

"Andreas KLostermann" <bluetigger at web.de> writes:

> I wanted to compile Python on Mingw32. That's a GNU C++ Compiler for
> Windows and it can configure...  The "make" process begins
> promissing, but fails at "posixmodule.c"...
> 
> This Problem is the same for 2.2.2 and 2.3a2.
> Can someone help me?
[...]

Sorry if this is a stupid question, but are you using the cygwin port?

Or is the cygwin port merged into Python proper, now?


John



From max at alcyone.com  Tue Apr 22 23:06:29 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 20:06:29 -0700
Subject: SGMLParser questions
References: <igmbavg5ughls6ksiden2rgnael1s0ub8q@4ax.com> <mailman.1051061136.16221.python-list@python.org>
Message-ID: <3EA60335.E3A047C6@alcyone.com>

Mike Rovner wrote:

> dict(attributes).get['src']

<nitpick value="I think you meant

	dict(attributes).get('src')

?"/>

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Life is painting a picture, not doing a sum.
\__/ Oliver Wendell Holmes, Jr.
    Max Pandaemonium / http://www.maxpandaemonium.com/
 A sampling of Max Pandameonium's music.



From tim at bladerman.com  Fri Apr 18 17:31:14 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Fri, 18 Apr 2003 21:31:14 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org>
Message-ID: <C8_na.646$dP1.2863@newsc.telia.net>

"Evan Simpson" <evan at 4-am.com> wrote in
> Guy Middleton wrote:
> > >>> reduce(lambda a, b: a+b, [y[1] for y in x])
> >
> > But I don't really like lambda, is there a better way to do this?
>
> What's wrong with:
>
> t = 0
> for y in x:
>      t += y[1]
>
> The extra variable?  The number of lines?

I am really curious about that to, I would most definitley say that this is
the most pythonic way. It is simple and easily readabel by anyone and I am
sure it is just as fast any of the other way.

I am not complaining now, I am sure there is a good reason and I am just
curious.

Thanx,

Tiim






From vze4rx4y at verizon.net  Tue Apr 29 02:37:52 2003
From: vze4rx4y at verizon.net (Raymond Hettinger)
Date: Tue, 29 Apr 2003 06:37:52 GMT
Subject: Another itertool function?
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no>
Message-ID: <45pra.20060$xw4.19534@nwrdny01.gnilink.net>

[Magnus Lie Hetland]
> The need for the following function just cropped up in a parser I'm
> writing; it seems sufficiently general to perhaps be a candidate for
> itertools -- and I can't see any way of implementing it with the
> current itertools functions:

Okay, I tried it and the result was horrible:

def weave(*iterables):
    "Works perfectly except fails to terminate"
    return ifilter(None, imap(type(chain([])).next,
                cycle(imap(chain, iterables, repeat(repeat(None))))))
>
> def weave(*iterables):
>     iterables = map(iter, iterables)
>     while iterables:
>         for it in iterables[:]:
>             try:
>                 yield it.next()
>             except StopIteration:
>                 iterables.remove(it)
>
> Another possible name (which is a bit more unwieldy) would be
> 'roundrobin'.

I find this name to be the most descriptive.


> I expect there would have to be more use-cases to justify it, but I
> just thought I'd air the idea...

Probably there should be more use cases.
It does seem to be a general purpose tool
for combining iterators.


Here's another idea that was submitted:

def windowed(iterable, group=2):
    "returns (i[0], i[1], ... i[group-1]), (i[1], i[2], ... i[group]), ..."
    it = iter(iterable)
    result = []
    while True:
        result.pop(0)
        result.append(it.next())
        if len(result) == group:
            yield group



Raymond Hettinger





From richardd at hmgcc.gov.uk  Mon Apr 14 03:14:13 2003
From: richardd at hmgcc.gov.uk (Richard)
Date: Mon, 14 Apr 2003 08:14:13 +0100
Subject: Decoding / Encoding ASN1 BER?
Message-ID: <3e9a5fc6@mail.hmgcc.gov.uk>

Hi,

Can anyone tell me the best way of  encoding and decoding BER ASN1 using
Python? I'm looking for the most popular solution and if possible any code
examples.

Thanks

Richard






From JoeyTaj at netzero.com  Tue Apr 29 01:38:30 2003
From: JoeyTaj at netzero.com (Paradox)
Date: 28 Apr 2003 22:38:30 -0700
Subject: Calling dotNet components from Python.
Message-ID: <924a9f9c.0304282138.15aa7635@posting.google.com>

I was wondering if it is possible to call a dotNet dll written in C#
from within a Python script. If anyone knows of any documentation or
example code that accomplishes something like this. It seems that you
can call a com component but all my searching on dotNet has not found
anything about using dotNet components in Python. Thanks



From dan65536 at hotmail.com  Tue Apr  1 21:48:31 2003
From: dan65536 at hotmail.com (daniel w. moore)
Date: 1 Apr 2003 18:48:31 -0800
Subject: ternary operator
Message-ID: <33b491f.0304011848.d87dcc6@posting.google.com>

anyone heard about the results ofPEP 308 voting?  What do people think
of this, particularly Pythonic Internals and Syntax Parsing experts (i am not one!).

-dan



From grahamd at dscpl.com.au  Fri Apr 25 04:05:00 2003
From: grahamd at dscpl.com.au (Graham Dumpleton)
Date: 25 Apr 2003 01:05:00 -0700
Subject: xmlrpc introspection
References: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com> <af6b5bdd.0304240434.fd7e6ad@posting.google.com> <UR0qa.77049$gK.181217@rwcrnsc52.ops.asp.att.net>
Message-ID: <dc6f5c99.0304250005.43ffbfcc@posting.google.com>

"Bryan" <belred1 at yahoo.com> wrote in message news:<UR0qa.77049$gK.181217 at rwcrnsc52.ops.asp.att.net>...
> i see a scalability problem with listMethods.  we have a server that
> dynamically loads many providers.  each provider registers one or more
> services.  each service handles one or more methods.  so a typical method
> name would be service.method or for example calculator.add.  when the server
> receives calculator.add, it knows to send it to all providers that
> registered for calculator service.  now when our server receives a
> listMethods request, it has to return all the method names in all providers.
> this could be hundreds or even thousands.  it would be nice if it was
> standard that there was a listServices request and listMethods took an
> optional service parameter.  then you can scale this up.  i noticed that
> it's typical to have method names in the format service.method so i'm too
> far off with this.  i'm probably going to have to implement it this way, i
> just wish it was the "standard" way.

That this presents a problem is not so much a deficiency of the the introspection
API but that many XML-RPC server side implementations only really easily allow
you to have XML-RPC services hosted out of a single URL. To my mind, different
web services should not all be at the same URL if they do different things. If you
can oragnise your server side code so that distinct services are on different URLs
then the introspection API will only be returning methods for that specific service
when called against its URL and you don't have the problem, unless you have a
bizare number of methods which are all to do with that service.

For example, calculator service could be accessed as:

  http://somesite/service/calculator

If you had a database access service, it could be accessed as:

  http://somesite/service/database

Have a look at the python manual for OSE (http://ose.sourceforge.net) to see how
it manages to have all registered services access using different URLs.



From bokr at oz.net  Fri Apr 25 21:54:06 2003
From: bokr at oz.net (Bengt Richter)
Date: 26 Apr 2003 01:54:06 GMT
Subject: Exception and finally question
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com> <3EA86EDA.EFD023C0@alcyone.com>
Message-ID: <b8coru$u8m$0@216.39.172.122>

On Thu, 24 Apr 2003 16:10:18 -0700, Erik Max Francis <max at alcyone.com> wrote:

>Tung Wai Yip wrote:
>
>> How to I access the exception object in a handler?
>>
>> try:
>>     raise NameError, 'HiThere'
>> except NameError:
>>     #how to print the message 'Hi There' here?
>
>	try:
>	    raise NameError, "Hi, there."
>	except NameError, e:
>	    print e
>
>Note that e is actually the exception _object_, not just the string (or
>whatever other argument) it was constructed with:
>
>>>> try:
>...  raise NameError, "What's up, dog?"
>... except NameError, e:
>...  print repr(e)
>...  print str(e)
>... 
><exceptions.NameError instance at 0x815b3b4>
>What's up, dog?
>
Sometimes it's useful to print the specific exception name along with the message,
if you're catching more than one kind of instance, e.g.,

 >>> def tex(ex):
 ...     try:
 ...         raise ex, 'a message'
 ...     except Exception, e:
 ...         print 'Exception "%s" message: "%s"' % (e.__class__.__name__, e)
 ...
 >>> tex(NameError)
 Exception "NameError" message: "a message"
 >>> tex(NotImplementedError)
 Exception "NotImplementedError" message: "a message"
 >>> tex(IOError)
 Exception "IOError" message: "a message"

Hm, how would you catch all exceptions raised as strings (e.g., raise "ho", "hum")
and print both string and 2nd arg? Do I have to dig with inspect?
It might be convenient if the system would convert such exceptions to a standard one,
so we could write e.g.,
    except  OldStringException, e:
        print e, e.args
and see the particular string and also a message arg.

Regards,
Bengt Richter



From aleaxit at yahoo.com  Thu Apr 10 03:25:46 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: Thu, 10 Apr 2003 09:25:46 +0200
Subject: Adding item in front of a list
References: <mailman.1049902101.26281.python-list@python.org>
Message-ID: <b7369r030oa@enews4.newsguy.com>

Tim Peters wrote:

> [Andrew Koenig, starting with l=[2, 3, 4]]
>> ...
>> I would have thought that after l.insert(-1, 1), l would be
>> [2, 3, 1, 4], but it doesn't work that way.
> 
> Alas, list.insert() existed before sequence indices were generalized to
> give
> a "count from the right end" meaning to negative index values.  When the
> generalization happened, it appears that list.insert() was just
> overlooked.
> 
> I'd like to change this.  If I did, how loudly would people scream?

If you presented it as a *bug fix* I wouldn't scream -- I'd PURR...!!!

Consider:

>>> def ins1(al, where, what):
...     al.insert(where, what)
...
>>> def ins2(al, where, what):
...     al[where:where] = [what]
...

are these two functions equivalent?  It sure seems so...:


>>> a=list('lopkan'); ins1(a, 2, 'i'); print a
['l', 'o', 'i', 'p', 'k', 'a', 'n']
>>> a=list('lopkan'); ins2(a, 2, 'i'); print a
['l', 'o', 'i', 'p', 'k', 'a', 'n']

BUT the bug breaks the equivalence when where<0...!

>>> a=list('lopkan'); ins2(a, -2, 'i'); print a
['l', 'o', 'p', 'k', 'i', 'a', 'n']
>>> a=list('lopkan'); ins1(a, -2, 'i'); print a
['i', 'l', 'o', 'p', 'k', 'a', 'n']
>>>

Yeah, I know, the docs now carefully guard this assertion of equivalence
with an "if i >= 0" and just as carefully specify that if i < 0 then the
item being "inserted" is actually prepended to the sequence.  So I'd say
there's a bug in the DOCS, too...;-)


> Guido says he also wishes list.insert() had been defined with the
> arguments in the opposite order, so that list.insert(object) could have a
> natural
> default index argument of 0.  I'd like to change that too, but it's
> clearly too late for that one.

Yep.  This has to go into a list of "tiny warts that will never be
fixed because we DO need backwards compatibility, sigh".


Alex




From jepler at unpythonic.net  Sat Apr  5 11:15:19 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Sat, 5 Apr 2003 10:15:19 -0600
Subject: building a .py files with distutils
Message-ID: <20030405161517.GA13698@unpythonic.net>

I'm writing a program which uses a generated .py file (it's generated based
on a yapps grammar file).  I know how to generate the .py file either as a 
shell command (python yapps.py gram.g) or as a python expression
(yapps.generate("gram.g"))

I'd like to use distutils to build the program, but I don't see an easy
way to cause a .py file to be built at the right time.  

I tried overriding build_ext and handling the .g -> .py in the subclass.
But when I did that, gram.py/gram.pyc aren't written in INSTALLED_FILES,
so bdist_rpm (at least) fails.  I tried overriding get_outputs in addition
to build_extensions, but that function doesn't even seem to be called.

(I suspect that using ext_modules would also make my built rpm be .i386
instead of .noarch, but this is a comparatively minor concern)

The next thing I tried was subclassing build_py and replacing
find_package_modules and build_modules.  This seems to have worked, at
least for the case where the .g file is inside a package.

Is there a better method I should have used?

Jeff




From andrew at acooke.org  Tue Apr 29 13:49:15 2003
From: andrew at acooke.org (andrew cooke)
Date: Tue, 29 Apr 2003 13:49:15 -0400 (CLT)
Subject: Visual Basic - Python compare
In-Reply-To: <uptn56xxi.fsf@fitlinxx.com>
References: <mailman.1051603275.12037.python-list@python.org> 
     <1bf4033f.0304290627.40f7d831@posting.google.com> 
     <uptn56xxi.fsf@fitlinxx.com>
Message-ID: <3339.10.2.0.2.1051638555.squirrel@127.0.0.1>

komodo has a gui builder these days (don't know how good it is).
http://www.activestate.com/Products/Komodo/more_information.plex

andrew

David Bolen said:

> fcslka at yahoo.com (Kent) writes:
>
>> 7.  In the interest of being honest, Python support in development
>> tools is weak.  There are many IDEs for Python, most of which are
>> immature, or are Python-only.  What needs to happen is for full
>> support of Python to be built into existing IDEs.  There's Visual
>> Studio .NET support, but boy, it's expensive.  NetBeans and Eclipse
>> could really help us out here if we could get more than just cursory
>> support for Python in those otherwise great environments.
>
> At the risk of over-emphasizing this aspect, I do think that the OP
> needs to consider this when it comes to replacing a VB environment
> with Python, particularly when it sounds like the application is
> already fairly well separated into back-end logic (the C++ COM
> objects) and front end UI (the VB code).
>
> I'd completely agree that in a pure language comparison, there's very
> little to recommend VB over Python.  However, to the extent that you
> are really comparing the RAD for VB development and/or its ability to
> design and implement the UI (and only the UI) for a system, the
> comparison is a lot closer.
[...]

-- 
http://www.acooke.org/andrew




From revanna at mn.rr.com  Sat Apr 19 17:12:57 2003
From: revanna at mn.rr.com (Anna)
Date: Sat, 19 Apr 2003 16:12:57 -0500
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it>
Message-ID: <pan.2003.04.19.21.12.54.280480@mn.rr.com>

On Sat, 19 Apr 2003 20:55:06 +0000, Alex Martelli wrote:

> So, I tried implementing sum in the simplest way, as a separate patch to
> operator.c (leaving add alone) -- THAT gives me 124 microseconds, a 50%
> speedup with respect to the simple for loop and clearly sufficient to
> justify sum's existence (together with its simplicity, clarity, and likely
> frequency of use).  As I was at it, I also ensured that operator.sum(x)
> would immediately delegate to ''.join(x) if x's first item is a string --
> this way it gets the same top performance as ''.join(x), avoiding the trap
> into which the "for xx in x: t+=xx" falls here (for x=map(str,range(999)),
> sum(x) clocks in at 79 microseconds, vs 78 for ''.join(x) and 1260 for the
> simple loop with += ...).
> 
> Interesting... I have my doubts on whether Guido would accept such a patch
> just a week before a beta, but maybe I should try submitting it anyway...
> 
> 
> Alex

+1

Anna



From martin at v.loewis.de  Fri Apr 25 17:56:32 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 23:56:32 +0200
Subject: zipfile with umlaut in filename
In-Reply-To: <3ea9a4bf$1_1@news.vo.lu>
References: <3ea9a4bf$1_1@news.vo.lu>
Message-ID: <b8caue$g65$05$3@news.t-online.com>

Patrick Useldinger wrote:

> Hello everybody,
> I have a problem when creating an archive using zipfile. If the name of 
> the file to be added contains a special character like an umlaut, this 
> character gets replaced by another character.
> What can I do against this?

I'm not sure you can do anything. Zipfiles only support ASCII file 
names, to my knowledge.

Regards,
Martin




From dave at boost-consulting.com  Tue Apr 29 21:58:26 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 29 Apr 2003 21:58:26 -0400
Subject: POD vs. reST for standalone writing?
In-Reply-To: <u3ck0696n.fsf@boost-consulting.com> (David Abrahams's message
 of "Tue, 29 Apr 2003 21:55:12 -0400")
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <u3ck0696n.fsf@boost-consulting.com>
Message-ID: <uu1cg4ugt.fsf@boost-consulting.com>

David Abrahams <dave at boost-consulting.com> writes:

> ReST rocks.  I'm using it to write a book, and think that it comes
> out very readably and is miles easier to write than any tweaky **ML
> or TeX.
>
> An example of something recent (not from the book):
> http://www.boost-consulting.com/writing/new-iter-concepts.html

For balance:

The biggest problem with ReST at the moment, I think, is that the only
output formats (that I care about) receiving any attention are HTML
and TeX, which means that other formats like PDF end up getting
generated from TeX via other tools.  The TeX generation is pretty
decent, but not really professional since it'ss really misusing TeX as
page layup format (c.f. PDF) instead of letting TeX do the formatting.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From roy at panix.com  Tue Apr 22 08:48:34 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 22 Apr 2003 08:48:34 -0400
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net> <nCUoa.41676$LB6.941312@news1.tin.it> <pan.2003.04.22.03.08.58.526973@mn.rr.com> <3japa.44301$LB6.997057@news1.tin.it>
Message-ID: <roy-6ECAE1.08483422042003@reader1.panix.com>

Alex Martelli <aleax at aleax.it> wrote:
> You're welcome, but for "sum([]) returns 0" you should thank Guido
> and Tim (it would raise a ValueError if I had implemented it my way;-).

It's hard to imagine that anybody who would put complex number support 
in the core of the language, and use the One True Symbol for the 
imaginary unit would think sum([]) should be anything but zero.  I'm 
with Guido j^2 on this one.



From aleax at aleax.it  Sun Apr 13 17:53:02 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 21:53:02 GMT
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <%djma.11763$T34.381354@news2.tin.it> <roy-031327.16584913042003@reader1.panix.com>
Message-ID: <2%kma.9988$LB6.267062@news1.tin.it>

Roy Smith wrote:

> Alex Martelli <aleax at aleax.it> wrote:
>> #define private public
>> #define protected public
>> at the top of your .cpp file, before any #include, in most C++
>> implementations
> 
> I'm confused.  You started out to show that C++'s "protected" keyword is
> evil,

I did no such thing -- I just mentioned Stroustrup's regrets about it (and
my agreement with him on this point).  It's more of a redundant complexity
than any substantial "evil".

> but what you've really shown is that C++'s preprocessor is evil.

Nope.  If your preprocessor didn't implement #define, you could do
just the same trick, even more messily, with a copy and paste from
the relevant .h files and any text editor of your choice.  The
preprocessor just makes it handier, avoiding the evils of copy and
paste coding for this, at least;-).

> Can they both be true? :-)

Possibly, but that's not what we're discussing, anyway.


Alex




From gradha at titanium.sabren.com  Tue Apr  1 14:46:43 2003
From: gradha at titanium.sabren.com (Grzegorz Adam Hankiewicz)
Date: Tue, 1 Apr 2003 21:46:43 +0200
Subject: Mimic a web surfer ?
In-Reply-To: <3e894b7d$0$29158$626a54ce@news.free.fr>
References: <3e894b7d$0$29158$626a54ce@news.free.fr>
Message-ID: <20030401194643.GB31829@pedos.es>

On 2003-04-01, shagshag13 <shagshag13 at yahoo.fr> wrote:
> i would like to develop a script which will mimic a web surfer,
> that is :
> 
> - follow links as a human and thus always come from previous page
> (i think there is something to do with http 'referer' and cookies)
> - be able to log if necessary
> 
> is there anything already existing ? i'm looking for any tips
> which will help me to start...

Take a look at http://sourceforge.net/projects/webunit/. It looks
much more complex than http://tclwebtest.sourceforge.net/. The
latter, however, is written in TCL. You could wrap it around under
python, but it's style would feel obscene.

-- 
 Please don't send me private copies of your public answers. Thanks.




From pobrien at orbtech.com  Sun Apr  6 19:19:43 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 18:19:43 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049669290.9639.python-list@python.org>
Message-ID: <m2u1db5i40.fsf@localhost.localdomain>

Sean 'Shaleh' Perry <shalehperry at attbi.com> writes:

> On Saturday 05 April 2003 22:51, Patrick K. O'Brien wrote:
> > Here is the background.  I'm working on a file editor (called
> > PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> > PyCrust code base.  I've got a basic version working, and it supports
> > editing multiple files and all the usual, basic features.  But I want
> > it to be as customizable and extensible as Emacs.  In fact, I want to
> > model it after Emacs, so that elisp code could be rewritten in Python
> > and work with PyAlaMode.  To do that, I've got to support the same
> > primitives as Emacs; expose files, buffers, regions, and windows the
> > same way Emacs does; handle the Emacs keybindings and ability to remap
> > keys, the same way Emacs does, etc.
> >
> 
> I was actually pondering this earlier in the week as I read more of
> the elisp and emacs docs.
> 
> One thing about emacs is that even though it is based on elisp most
> people can blissfully ignore that and just use it.  Please, if you
> make a newer, better, GUI editor with a Python backend do not limit
> it or even aim it at the Python community.  Make it the great new
> thing for ALL text editing.  This is why I use emacs over vi -- I
> can teach the editor about whatever it is I am editing and decrease
> the amount of time I spend typing in text. 6 years of vi and now a
> year and a half with emacs i can't imagine going back.  To get me to
> leave emacs will require something with all of emacs' might and more
> modern features.

I am aiming it for the Python community, but don't intend to criple it
in any way.  The underlying Scintilla text control can provide
coloring, indenting, folding, brace matching and such for a wide
variety of programming languages.  So the plan is to support whatever
Scintilla supports.  But the early focus is on Python features.

> real pop up hints, suggestive programming, etc as seen in commercial
> IDEs would be nice.  As much as the IDEs have improved they are
> still one trick ponies that can only be used to edit source so a
> real editor like emacs still wins overall.  Support for refactoring
> regardless of the language.  World peace.  you name it.

PyAlaMode already supports pop up hints and autocompletion for Python
source files.  Perhaps the Bicycle Repair Man folks can suggest ways
to integrate BRM for the refactoring features.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From mwh at python.net  Wed Apr  2 10:54:01 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 2 Apr 2003 15:54:01 GMT
Subject: stop exeuction and give my a prompt
References: <b6dara$5mv$1@news.asu.edu>
Message-ID: <7h3k7eclwtj.fsf@pc150.maths.bris.ac.uk>

David Bear <david.bear at asu.edu> writes:

> I would like to have the ability to insert a statement in my code that would 
> stop program execution at that point and give me a python prompt.  I don't 
> know what to call this. 
> 
> What I would like to do is to stop program execution and inspect the state 
> of certain dictionaries --- I don't want to insert a bunch of print 
> statements and watch output fly of the screen.  I'd also like to be able to 
> fiddle with these dictionaries when my program stops.  
> 
> Any pointers?

Look at the documentation for the standard library module 'code'.

Cheers,
M.

-- 
  You have run into the classic Dmachine problem: your machine has
  become occupied by a malevolent spirit.  Replacing hardware or
  software will not fix this - you need an exorcist. 
                                       -- Tim Bradshaw, comp.lang.lisp



From zhaojs at yahoo.com  Thu Apr 17 07:22:01 2003
From: zhaojs at yahoo.com (Jinsong Zhao)
Date: 17 Apr 2003 04:22:01 -0700
Subject: check to see if a symbol in a string has a binding
Message-ID: <kfed6jlv06e.fsf@alumni.cse.ucsc.edu>

Sorry if the question is too simple. I want to see if the name inside
the string has a binding and callable. For example,

1. func_string="func"
2. func_string="func_string"
3. func_string="dir"

So the in case 1, "func" does not have a binding; case 2 "func_string"
has a binding but not callable; in case 3 "dir" has a binding and is
callable. How do I proceed based on the content in func_string? This
is rather baffling.

Any suggestion? Thanks in advance.

Jinsong





From nicodemus at globalite.com.br  Sun Apr  6 09:27:30 2003
From: nicodemus at globalite.com.br (Nicodemus)
Date: Sun, 06 Apr 2003 10:27:30 -0300
Subject: Python & Music: Performance question
In-Reply-To: <1759581e.0304060218.164d4cd4@posting.google.com>
References: <1759581e.0304060218.164d4cd4@posting.google.com>
Message-ID: <3E902B42.1090702@globalite.com.br>

David Asorey ?lvarez wrote:

>Hello, group.
>
>I'm developing a small music program, and I have some doubts.
>
>   I want define a class, Note, which represents a note. It has two
>attributes, pitch and duration (quarter note, half note, etc). There
>will be a lot of instances of this object, maybe hundreds of notes,
>and the program needs access to all of this objects and read its
>attributes.
>   Which implementation will be more efficient?
>
>Implementation A:
>
>class Note:
>   def __init__(self, pitch, duration):
>      self.pitch = pitch
>      self.duration = duration
>   def change_pitch(factor): pass
>   def change_duration(factor): pass
>   # more methods ...
>  
>
Try using __slots__... this will save memory.

class Note(object):
    __slots__ = ['pitch',  'duration' ]
    def __init__(self, pitch, duration):
        self.pitch = pitch
        self.duration = duration
    # more methods

>Implementation B:
>
>class Note:
>   def __init__(self, pitch, duration):
>      self["pitch"] = pitch
>      self["duration"] = duration
>   def change_pitch(factor): pass
>   def change_duration(factor): pass
>   # more methods ...
>  
>
This won't even work, unless you define a __setitem__ method. And if 
you're looking for efficiency, that's not the way to go.

Hope that helps,
Nicodemus.






From bdesth at removethis.free.fr  Wed Apr  2 08:57:24 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 02 Apr 2003 15:57:24 +0200
Subject: Best practice for object attributes?
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk>
Message-ID: <3E8AEC44.2090909@removethis.free.fr>

Michael Sparks wrote:
> Hello,
> 
> 
> If I want to model objects with 3 attributes based on the behaviour of
> python 2.2/2.3 I'ved noticed I can use either of the following idioms:
> 
> class shrubbery1(object):
>    name = None
>    knightowner = "ni"
>    flowering = False
> 
> or:
> class shrubbery2(object):
>    def __init__(self):
>       self.name = None
>       self.knightowner = "ni"
>       self.flowering = False
> 
> In both cases I gain objects which have the same 3 attributes.
> Differences between
> them I find are:
>    * shrubbery1 has 3 class attributes as well, with the initial values
>      indicated. This isn't the case with shrubbery2.
> 
>      >>> [x for x in dir(shrubbery1) if not x[0] == "_"]
>      ['flowering', 'knightowner', 'name']
> 
>      >>> [x for x in dir(shrubbery2) if not x[0] == "_"]
>      []
> 
>    * shrubbery2 has 3 entries added to self.__dict__ which act as the
>      attribute stores. This is not the case with shrubbery1.
> 
>      >>> shrubbery1().__dict__
>      {}
>      >>> shrubbery2().__dict__
>      {'flowering': False, 'name': None, 'knightowner': 'ni'}
> 
> 
> Personally I find the former (shrubbery1) clearer, but I'm posting to
> find out what people think is generally the better of the two. Do people
> find any inherent advantages to one versus the other? 

One outstanding advantage of the second 'idiom' is that it's the right 
way to define *instance* attributes !-) (which is probably what your 
looking for)

With the first form you get *class* attributes. Make 2 instances of 
Shrubbery1, change the name of first one to 'Brian', print the name of 
the second one, and then try to understand what is the purpose of the 
'self' argument in methods.

> Thanks for any comments,

You're welcome.

Laotseu





From aleax at aleax.it  Sun Apr 13 05:24:05 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 09:24:05 GMT
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com> <ugham-t2f.ln1@grendel.myth> <fi%la.7870$T34.222571@news2.tin.it> <lhvmm-7c7.ln1@grendel.myth>
Message-ID: <V0ama.9033$T34.282930@news2.tin.it>

Jim Richardson wrote:
   ...
> I don't know how fast/slow the dict would be to tell the truth, it just
> doesn't seem that "elegant" and elegance, is often (wrongly I know)
> associated with speed.

It's quite usual that in Python the simplest and most elegant solution
can also be very fast.  But I dispute that dicts aren't elegant!!!


> Having said that, I have found that usually, the simpler the script, and
> the closer to the pythonic "metal" it is, the faster it is. For some
> reason, the folks who wrote python, are a lot better at programming that
> I am :)

Sure.  But dicts are VERY "close to the Pythonic metal"...!


Alex




From inforcati at yahoo.com.br  Fri Apr 25 16:00:46 2003
From: inforcati at yahoo.com.br (Mauricio)
Date: 25 Apr 2003 13:00:46 -0700
Subject: Permanent Clock
Message-ID: <d613c14.0304251200.33f1b741@posting.google.com>

Is there some way of accessing a permanent clock, i.e., a clock that
works even if the system is rebooted or poewered off?
I would like to increment a variable each day, independent of the
system date (once this one users can change!), but not leaving my
script running all the time. Better explaining: every time tha script
is run, it must now if the current day is the same of tha last time
the script was started.If not, increment tha variable!

  thanks for helping,



From aleax at aleax.it  Mon Apr 28 04:08:44 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 08:08:44 GMT
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <ql_qa.18612$K35.626741@news2.tin.it>
Message-ID: <gk5ra.13911$3M4.379232@news1.tin.it>

Lorenzo Bolognini wrote:

> "Steven Hodgen" <steven at twitch.net> ha scritto nel messaggio
> news:oproar1sjtx3zh5c at news.mindspring.com...
> 
>> However, Java is not a great language for beginners.  Python, on the
> other
>> hand, is ideal.  If I can find a textbook that is even ok, I may
> consider
>> changing, since the language is so much better for students at this
> level.
> 
> I'm a beginner and I liked very much
> 
> Practical Python
> By Magnus Lie Hetland

It's an excellent book indeed.  However, Steven is looking for
a *TEXT*-book -- and Magnus's book isn't one (no exercises)...
it may be one of the best tools to LEARN Python, but textbooks
typically have certain requirements, which this book doesn't meet.


> And you could find the tut's at this page useful for the very
> beginning
> http://hetland.org/python/

Excellent tutorials (translating a few of them into Italian was
one of the first things I did as I was learning Python), but
irrelevant wrt the need to select *TEXTBOOKS*, alas.


Alex




From staschuk at telusplanet.net  Mon Apr 28 00:13:52 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 27 Apr 2003 22:13:52 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <1051075309.87191@yasure>; from donn@drizzle.com on Wed, Apr 23, 2003 at 05:21:50AM -0000
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <f6a31cb8.0304220810.547f5acd@posting.google.com> <oRKdnYDxiYyb5TijXTWcpA@comcast.com> <1051075309.87191@yasure>
Message-ID: <20030427221352.A23486@tibia.amotlpaa.bogus>

Quoth Donn Cave:
  [...problem with an implementation of __repr__ in Mailman...]
> So here's this object.  You can't repr() it, including implicitly
> from the interpreter prompt - you get a traceback instead.  Of course
> that also goes for the list object that contains this object.  To
> me, this is not right, and maybe it ought to be another point to
> consider if we're reconsidering the documentation for __repr__.
> Wish I could specify this better, but I believe there's a minimalist
> way to think about implementing __repr__ that will tend to make it
> less interesting but more reliable.

That's an excellent example of how to do __repr__ wrong, and I
agree that the docs should reflect the underlying principle.  I
think it can be expressed as a consequence of the debugging focus
of __repr__; it's hard to inspect an object while debugging if the
act of inspection itself crashes.

(Related point: a repr should be reliably printable in diverse
environments -- which probably requires sticking to printable
ASCII -- again to ensure debuggability.)

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From mike at bindkey.com  Tue Apr  8 17:11:14 2003
From: mike at bindkey.com (Mike Rovner)
Date: Tue, 8 Apr 2003 14:11:14 -0700
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com>
Message-ID: <b6vdsg$65b$1@main.gmane.org>

def split(S,N):
  """S - source string
     N - number of parts
  """
  lens=len(s)
  l=lens/N
  return [S[i:i+l] for i in xrange(0,lens,l)]

Mike

PS. Task description is too vague.

"Heiko Henkelmann" <heiko at hhenkelmann.de> wrote in message
news:b6vd0e$i1b$01$1 at news.t-online.com...
> I need to split a string into an array of equally long substrings, like
> e.g.:
>
> '1234567890'   -> ['12', '34','56','78','90']
>
> What is the most efficient way to do this?
>
> Heiko







From rshaw2 at midsouth.rr.com  Wed Apr 30 14:21:27 2003
From: rshaw2 at midsouth.rr.com (Richard)
Date: 30 Apr 2003 11:21:27 -0700
Subject: Python 2.3b1 and PIL 1.1.4
Message-ID: <84e0f331.0304301021.3a367d05@posting.google.com>

I recently installed Python 2.3b1 and the appropriate PIL 1.1.4b1 from
effbot download (both for win32). I then tried to run a simple image
editing program I wrote with Tkinter and ran into a little problem.
The PIL install for Python 2.3 does not include the _imagingtk.pyd
file while the 2.2 version does.

Is this file not longer needed? I am not importing it directly. The
exception is as follows:

Exception in Tkinter callback
Traceback (most recent call last):
  File "C:\Python23\lib\lib-tk\Tkinter.py", line 1337, in __call__
    return self.func(*args)
  File "C:\Tools\ImageTweaker\imagetweak.pyw", line 61, in open
    self.show(self.image)
  File "C:\Tools\ImageTweaker\imagetweak.pyw", line 71, in show
    self.image_label = viewer.UI(self, image)
  File "C:\Tools\ImageTweaker\viewer.py", line 23, in __init__
    self.image = ImageTk.PhotoImage(im)
  File "C:\Python23\lib\site-packages\PIL\ImageTk.py", line 114, in
__init__
    self.paste(image)
  File "C:\Python23\lib\site-packages\PIL\ImageTk.py", line 179, in
paste
    import _imagingtk
ImportError: No module named _imagingtk

Thanks



From a.schmolck at gmx.net  Thu Apr 24 20:02:52 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 25 Apr 2003 01:02:52 +0100
Subject: string to variable name
References: <b89qi8$2oij$1@agate.berkeley.edu>
Message-ID: <yfsist3wij7.fsf@black132.ex.ac.uk>

>   FILE = open("<mytextfile");
>   while (<FILE>){
>     /^(\w+)\s+(\w+)\s+(\w+)$/;
>     inst.$2 = $1;
>   }
> }


> Is there any way to (easily) achieve what I want to do in python?

Sure: you can do it even shorter and almost as unreadably with (UNTESTED):

for line in file("mytextfile"):
    inst.__dict__.update(dict([line.split()[:2]]))

a better way would be to use ``setattr(inst, key, value)``

> Hey! Could I do it by overloading the special function that's called when you
> treat something like a dictionary?  i.e. could I make my class respond to 
>   key = "bar"
>   inst[key] = "baz"
> as though it got
>   inst.bar = "baz"

Sure.

> 
> How?

if you really feel the need, have a look at __setitem__ in the python language
reference

'as



From afriere at yahoo.co.uk  Mon Apr 28 00:54:12 2003
From: afriere at yahoo.co.uk (Asun Friere)
Date: 27 Apr 2003 21:54:12 -0700
Subject: Date validation
References: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au> <mailman.1051257039.22404.python-list@python.org>
Message-ID: <38ec68a6.0304272054.6e4d2eb2@posting.google.com>

"Fredrik Lundh" <fredrik at pythonware.com> wrote in message news:<mailman.1051257039.22404.python-list at python.org>...
> if you want your program to check how many days there are in
> a given month, use the calendar module:
> 
> >>> import calendar
> >>> firstday, ndays = calendar.monthrange(2003, 4)
> >>> print ndays
>  30
> >>> print calendar.day_abbr[firstday]
> 'Tue'
> 
> </F>
Unfortunately calendar.monthrange does not work for years < 1970 in
the latest stable release, as I just discovered :(   (To be more
preceise, it doesn't seem to work for any year in the from the 2nd
Century till 1970, but it does work for all but the last year of the
1st century).  I see this has been fixed in 2.3a2.

While checking this out, I noticed that calender.py in 2.3a2 imports
'datetime,'
for which there is no python file.  Is this now the wrapper to the
various clib implementations?



From staschuk at telusplanet.net  Thu Apr  3 23:28:22 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 3 Apr 2003 21:28:22 -0700
Subject: split encloser
In-Reply-To: <435d40e0.0304031530.1c197742@posting.google.com>; from aussie2010@yahoo.com on Thu, Apr 03, 2003 at 03:30:10PM -0800
References: <435d40e0.0304031530.1c197742@posting.google.com>
Message-ID: <20030403212822.A2938@tibia.amotlpaa.bogus>

Quoth Chris:
> string.split() takes a delimiter and works fine as long as the
> delimiter isn't part of the data fields. But frequently they are.
> e.g. 'John Doe,135 South Main St.,#122, Springfield, Iowa' or
>       ' so long goodbye see ya'
> 
> Because the fields can contain the delimiter in some cases, an
> encloser is usually used (typically "") to handle those fields.
> 
> The above strings would be written:
> 'John Doe,"135 South Main St., #122", Springfield, Iowa'
> and 
> '"so long" goodbye "see ya"'

What if the field data contains double quotes?

> I don't understand regular expressions but I was wondering if anyone
> that did knew of a way to get re.split() to handle "enclosers" as used
> above.

Why use a regular expression?  string.split can do the trick:

	in_out = line.split('"')
	fields = []
	for i in range(len(in_out)):
		if i % 2:
			results.append(in_out[i])
		else:
			results.extend(in_out[i].split(','))

Or, more clearly:

	fields = []
	while line:
		if line.startswith('"'):
			endquote = line.index('"', 1)
			field = line[1:endquote]
			# assumed " is followed by ,
			line = line[endquote+1:]
		else:
			field, line = line.split(',', 1)
		fields.append(field)

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From pw at panix.com  Thu Apr 17 16:06:42 2003
From: pw at panix.com (Paul Wallich)
Date: Thu, 17 Apr 2003 16:06:42 -0400
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <871y01f2hw.fsf@deneb.enyo.de> <3E9F1FE8.5030409@removethis.free.fr>
Message-ID: <pw-A79DD1.16064217042003@reader1.panix.com>

In article <3E9F1FE8.5030409 at removethis.free.fr>,
 laotseu <bdesth at removethis.free.fr> wrote:

> Florian Weimer wrote:

>  > In the strictest sense, the language has to be purely
>  > applicative (or free of side effects).  Common Lisp satisfies the
>  > first definition, but not the second.
> 
> Do you mean that it is *absolutely not possible* to write a whole 
> program in CL whithout any side effect ? Or that many features in the 
> language, that you can choose to use or not, are not side effect free ?

It's the second. In a strictest-sense functional language, it is 
impossible to write a whole program that *does* have side effects. 
This leads to the usual amusements such as passing the universe to a 
function that then returns a similar-but-not-identical universe with one 
property differing...

paul



From imbosol-1050250211 at aerojockey.com  Sun Apr 13 12:18:49 2003
From: imbosol-1050250211 at aerojockey.com (Carl Banks)
Date: Sun, 13 Apr 2003 16:18:49 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <bJ2ma.5447$883.4511@nwrdny02.gnilink.net> <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net>
Message-ID: <J5gma.9328$Gb1.6409@nwrdny01.gnilink.net>

Mads Orbesen Troest wrote:
> On Sun, 13 Apr 2003 01:05:11 GMT, Carl Banks wrote:
> 
>> Constant.
>> Constant.
>> Constant.
> 
> But is that /guaranteed/ (as it is with STL), or does it just happen to be 
> that way with the current implementation (however unlikely it is to 
> change)?

It's not guaranteed, but if it ever changes, I'll eat my hat.


-- 
CARL BANKS



From phil at river-bank.demon.co.uk  Tue Apr 22 15:11:40 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Tue, 22 Apr 2003 20:11:40 +0100
Subject: Problem with PyQt on SuSE 8.2
In-Reply-To: <b842hl$dsa$1@online.de>
References: <b842hl$dsa$1@online.de>
Message-ID: <200304222011.40917.phil@riverbankcomputing.co.uk>

On Tuesday 22 April 2003 7:44 pm, Bernd Lachner wrote:
> Hello,
>
> we try ro run several Python programs which are using PyQt on SuSE 8.2.
> Every time the same error message occurs:
>
> Traceback (most recent call last):
>   File "./kdevpy.py", line 16, in ?
>     from qt import *
>   File "/usr/lib/python2.2/site-packages/qt.py", line 39, in ?
>     import libqtc
> ImportError: /usr/lib/python2.2/site-packages/libqtcmodule.so: undefined
> symbol: _ZNK16QAssistantClient9classNameEv
>
> We tried the original PyQt packages which comes with SuSE 8.2. We also have
> compiled PyQt from the source RPMs, but the same error occurs.
>
> Is ther e anybody who can help us? Thanks in advance.

You don't say which version of Qt.

If you are using Qt v3.1.1 or earlier, and you do a "make install" (rather 
than leaving the build in situ) the libqassistantclient library gets 
corrupted because of over-enthusiastic stripping.

Phil




From anton at vredegoor.doge.nl  Mon Apr 21 17:31:09 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 21 Apr 2003 23:31:09 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <mailman.1050953314.21065.python-list@python.org>
Message-ID: <b81od0$gcc$1@news.hccnet.nl>

Steven Taschuk <staschuk at telusplanet.net> wrote:

>Quoth Anton Vredegoor:
>> "A.M. Kuchling" <amk at amk.ca> wrote:
>> 
>> >http://mail.python.org/pipermail/python-dev/1999-July/000467.html
>> 
>> Thanks. It seems Tim is making a heroic effort to explain the new in
>> terms of the old. This necessarily must fail, but what other option is
>> there instead of rebuilding all terminology out of gotos?
>
>I am not sure it needs to fail, nor that it is a bad idea to
>"rebuild" everything out of gotos.

Please let's see what our forefather said about this, centuries ago
when the world was young and fresh:

<Tim>

CONTINUATIONS

Given the pedagogical structure so far, you're primed to view
continuations as an enhancement of coroutines. And that's exactly what
will get you nowhere <wink>. Continuations aren't more elaborate than
coroutines,they're simpler. Indeed, they're simpler than generators,
and even simpler than "a regular call"!  That's what makes them so
confusing at first: they're a different *basis* for *all* call-like
behavior.  Generators and coroutines are variations on what you
already know; continuations challenge your fundamental view of the
universe.

</Tim>

>From this I get the impression that continuations are not explainable
in terms of the other structures because they're simpler. Application
of common sense logic is sufficient to conclude that it's inefficent
to explain amoebae in terms of elephants. Maybe if two twin elephants
could be found that differ only one amoeba from each other there would
be a basis for such explorations but it would still demand the utmost
disciminative powers from the observers. Furthermore I politely refuse
the implication of generalizing my remarks beyond the scope of the
link under our scrutinization and also confirm that I have no
objections whatsoever to rebuilding terminology out of gotos save
perhaps that in modern Python it's possible to pop stacks anyway one
wants to, which might provide a better analogy for the low level
operations.

>One possible method for understanding continuations, for those
>with C (or similar language) experience:  Assume you have only one
>control flow construct, a conditional goto.  Show how if/else,
>loops, and function call/return can be implemented using
>conditional goto.  Then show how coroutines, generators, and
>backtracking can be implemented using it.  Notice especially your
>use of constructions which consist of "save state, goto there" and
>"goto there, restore state" combined in useful ways, for various
>values of "state".
>
>Having thus prepared the conceptual ground, *then* go read
>material on continuations.

Seems like sound advice. However, I would like to make one observation
to save this discussion from the morasses of technological epicycles.
These continuations might seem difficult and hard to understand for
most people (including me, being on the wrong side of enlightenment in
this matter), however in years even longer ago when there were no such
sophisticated counting mechanisms as today are provided for even young
children, it was deemed a sign of superior intelligence to even be
able to divide two large integer numbers. The differance is the
conceptual framework and computer equipment that were absent then and
that are present now to fit the problem space in. I believe similar
advances in the concepts of continuations brought about by our
conceptual forerunners will enable the masses in the end to follow and
come to grips with the matter at hand. Therefore it is advisable to
refrain from esoteric behavior around these -expected to become so
RSN- probably relatively simple concepts.

Anton




From oren-py-l at hishome.net  Sat Apr 19 04:15:55 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Sat, 19 Apr 2003 04:15:55 -0400
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <OT0oa.690$l75.32@tornadotest1.news.pas.earthlink.net>
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <Gu_na.34381$LB6.779929@news1.tin.it> <OT0oa.690$l75.32@tornadotest1.news.pas.earthlink.net>
Message-ID: <20030419081555.GA84526@hishome.net>

On Sat, Apr 19, 2003 at 12:38:06AM +0000, Tyler Eaves wrote:
...
> > Using lists 10 or 100 times as long seems to show O(N) behavior for
> > each case, as could be expected.  So, the simplest (and I would agree
> > thereby most Pythonic) approach is over twice as fast as the fancy ones
> > (measurements on 2.2.2 seem to show similar results, all approaches
> > being about 20% slower than in 2.3, uniformly).
> > 
> > 
> > Alex
> 
> Wow, I feel vindicated now for never trying to wrap my brain around stuff
> like list comprehensions. *g*  

You are welcome to feel great about never having learned the meaning 
of the reduce builtin. But list comprehensions have nothing to do with 
it - they are just a shorthand notation for a loop that appends to a 
list. The performance of list comprehensions is exactly the same as the 
equivalent loop construct. In this case, though, the version of the
code using the list comprehension was not equivalent - it created a 
temporary list where none was necessary and was therefore slower than 
the trivial Pythonic implementation.

    Oren




From rmunn at pobox.com  Mon Apr 14 11:05:44 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 14 Apr 2003 15:05:44 GMT
Subject: Confusion: assignments create object references.
References: <0C642BE4EE085149B52839125D4FC9D422BD9F@idc1exc0002.corp.global.level3.com> <mailman.1050079403.16409.python-list@python.org>
Message-ID: <slrnb9kurf.1q1.rmunn@localhost.localdomain>

Jp Calderone <exarkun at intarweb.us> wrote:
> On Fri, Apr 11, 2003 at 10:28:42AM -0600, Havens, Peter wrote:
>> 
>> >>> L = 0
>> >>> M = L
>> >>> L
>> 0
>> >>> M = 1
>> >>> L
>> 1
>> >>>
>> 
> 
>   The above never happened, I assure you. >:)    Python doesn't work this
> way.

Indeed. I cannot reproduce this behavior either:

[rmunn at localhost ~]$ python
Python 2.2.2 (#1, Jan 12 2003, 12:07:20)
[GCC 3.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> L = 0
>>> M = L
>>> L
0
>>> M = 1
>>> L
0
>>>

Try it again. Do you still get the same result?

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From max at alcyone.com  Mon Apr  7 02:18:48 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sun, 06 Apr 2003 23:18:48 -0700
Subject: Emacs python-mode problem
References: <m2of3ktq67.fsf@db.hit.edu.cn> <bd993a2f.0304062214.2027bed8@posting.google.com>
Message-ID: <3E911848.D8AFE286@alcyone.com>

Anand B Pillai wrote:

> (setq auto-mode-alist
>        (append '(("\\.C$"  . c-mode)
>                 ("\\.cc$" . c++-mode)
>                 ...

Note that .C (on case sensitive filesystems, anyway) is usually a _C++_
file, not a C one.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Things are as they are because they were as they were.
\__/ Thomas Gold
    CSBuddy / http://www.alcyone.com/pyos/csbuddy/
 A Counter-Strike server log file monitor in Python.



From roy at panix.com  Sun Apr 13 11:04:58 2003
From: roy at panix.com (Roy Smith)
Date: Sun, 13 Apr 2003 11:04:58 -0400
Subject: How to do per-line profiling?
Message-ID: <roy-7A3D8E.11045713042003@reader1.panix.com>

Are there any tools to do profiling on a line-by-line basis?

I've got a program that does a lot of low-level string manipulation, 
including building up strings character by character using "s1 += s2" 
constructs.  I knew this was evil, but it was quick-and-dirty and worked.  
Run time is just slow enough to be really annoying, but not slow enough to 
have ever convinced me to rewrite it :-)

All this recent talk about the quadratic complexity of string addition has 
made me finally go back and fix it.  I tore out all the string additions 
and replaced them with appending to a list, an join()ing the list at the 
end.  To my astonishment, this made it run about 10% slower!

The Python library profiler shows that, all my time is still spent 
groveling in string creation, and most of that is due to one routine 
(reformatted to fit the screen):

ncalls  tottime  percall  cumtime  percall filename:lineno(function)
     1    0.010    0.010   21.550   21.550 profile:0(app.main())
     1    0.000    0.000   21.540   21.540 <string>:1(?)
     1    0.000    0.000   21.540   21.540 ...(main)
     2    0.640    0.320   21.070   10.535 ...(parse)
 10382   14.270    0.001   17.210    0.002 ...(getNextStatement)
 10374    2.100    0.000    3.210    0.000 ...(processInsert)
 84882    2.120    0.000    2.120    0.000 ...(getLine)

The problem is that getNextStatement() is 165 lines of messy state machine 
code and it's not at all clear where the bad spots are.  At this point, 
what I really need is a profiler that can show me how much time is spent in 
each line of code, not just in each function.  Does such a thing exist?



From bignose-hates-spam at and-zip-does-too.com.au  Sun Apr  6 21:06:06 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Mon, 07 Apr 2003 01:06:06 GMT
Subject: Cross-reference 'import' in a class hierarchy
Message-ID: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake>

Howdy all,

So, I'm fairly experienced with Python for shell programming, fairly
experienced with OO programming, but am just now using the one for the
other.

I'm designing a class hierarchy, using Python package directories for
semi-discrete sets of modules; all Good Things if I understand
correctly.

However, some of the modules need to refer to modules elsewhere in the
hierarchy, and I can't find how to manage filesystem paths elegantly in
Python 'import' statements.

  Project monty
    |--- Package foo
    |     |--- Module ook
    |     |--- Module eek
    |--- Package bar
          |--- Module ork
          |--- module urk

In ork.py I want:

    import eek

How do I set up my paths the Right Way so that this will just work?  How
will it be maintained when modules move around, get renamed, split into
sub-packages, et cetera?

These aren't in a state to be site packages; they're still undergoing
rapid architectural changes.  Thus, using monthy.pth isn't an option
(unless you can tell me otherwise).

Any guidance appreciated as I tread this unfamiliar ground.

-- 
 \        "I installed a skylight in my apartment. The people who live |
  `\                          above me are furious!"  -- Steven Wright |
_o__)                                                                  |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From thoaan at stud.cs.uit.no  Wed Apr  9 07:58:18 2003
From: thoaan at stud.cs.uit.no (Thomas Aanensen)
Date: Wed, 9 Apr 2003 13:58:18 +0200
Subject: Sockets on PocketPC
Message-ID: <b711rb$15rb$1@news.uit.no>

Hi

Have anybody used sockets on a PocketPC?

I get an "ImportError: No module named _socket"

Isn't sockets supportet in the PPCPython lib?


Thomas





From rmunn at pobox.com  Tue Apr 22 14:37:21 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 22 Apr 2003 18:37:21 GMT
Subject: coverting list back to string
References: <3cf72d52.0304211132.799d8afb@posting.google.com> <878yu33a16.fsf@vercingetorix.vercingetorix> <3cf72d52.0304220541.144f537b@posting.google.com>
Message-ID: <slrnbaae7b.uvq.rmunn@localhost.localdomain>

I rearranged your post so the question-and-answer order makes sense.
(Please don't top-post, see http://www.caliburn.nl/topposting.html for
reasons why).

scn <regnivon at netscape.net> wrote:
> Dialtone <dialtone at despammed.com> wrote in message news:<878yu33a16.fsf at vercingetorix.vercingetorix>...
>> regnivon at netscape.net (scn) writes:
>> 
>> > my quest:  i know that output.write(remove) is not correct.  you
>> > cannot use the 'write' method on a list.  is there a better way to
>> > convert the list to a string after the first element of the list has
>> > been removed and then output the strings to a file?  i'm a newbie
>> > programmer and need guru help. thank you.
>> 
>> Sure:
>> 
>> input = file('c:/python22/programs/linetest.txt', 'r')
>> output = file('c:/python22/programs/output.txt', 'w')
>> 
>> for line in input:
>>     remove = line.split()
>>     remove = remove[1:]
>>     output.write(''.join(remove))
>> 
>> input.close()
>> output.close()
> 
> Thank you for your help.  This worked.  I incorporated your suggestion
> of 'remove = remove[1:]' instead of my original code. I'd like to
> figure out how to get each string to output line-by-line, instead of
> one continuous string. If you have any suggestion, I'd appreciate it.
> Thanks again!

Try adding "output.write('\n')" at the end of your for loop:

    for line in input:
        remove = line.split()
        remove = remove[1:]
        output.write(' '.join(remove))
        output.write('\n')

That will get you a newline after each string. Notice also that I put a
space as the delimiter for the join operation. If you use the empty
string ('') as the delimiter, you'll get the input fields all squashed
together, which is probably not what you wanted.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From aleax at aleax.it  Sat Apr 12 11:53:38 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 15:53:38 GMT
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com>
Message-ID: <6EWla.4594$LB6.132415@news1.tin.it>

A. Lloyd Flanagan wrote:

> Alexander Schmolck <a.schmolck at gmx.net> wrote in message
> news:<yfsistmh4ec.fsf at black132.ex.ac.uk>...
>  
>> __slots__ is the one addition in python2.2 that really annoys me. It
>> seems an ill-conceived jumble of conceptually unrelated things (bug
>> avoidance, optimization, safety, encapsulation) that screws up reflection
>> big time for little apparent gain. As Knuth said ...
>> 
>> 'as
> 
> I agree.  The more I looked at it, the more it became clear that
> __slots__ turns a class into something very non-pythonic, with
> significantly different behavior.  A lot of utilities, debuggers, and
> advanced programming tricks will be coming with this disclaimer:
> "does not work with any class which defines __slots__"

I think it's an important memory optimization, PERIOD.  The other
conceptually unrelated things (bug avoidance -- I can't even SEE the
others!) aren't in __slots__' own design (even though a commentator
who's usually quite perceptive seems to have seen the "bug avoidance"
feature that wasn't really there).

As Knuth said, *premature* optimization is the root of all evil in
programming.  But __slots__ may allow very substantial optimization
that need not be premature.  I think it should just have been more
clearly documented as a pure optimization issue right from the start,
that's all.


> Unless you absolutely have to have it, don't use it.

"Absolutely have to have it", just like "screws up big time", are
silly overbids IMHO -- over-reactions that I don't see as justified
at all!  Why can't we all just keep SOME perspective, please...?!

Consider a typical use case, a class whose instances have and need
a single attribute 'x', and of which we need to generate 10,000:

[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s' def 
__init__(self): self.x=23' 'z=[X() for i in xrange(9999)]'
10 loops, best of 3: 6.51e+04 usec per loop
[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s' def 
__init__(self): self.x=23' 'z=[X() for i in xrange(9999)]'
10 loops, best of 3: 6.56e+04 usec per loop
[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s ' 
__slots__=["x"]' -s' def __init__(self): self.x=23' 'z=[X() for i in 
xrange(9999)]'
10 loops, best of 3: 3.36e+04 usec per loop
[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s ' 
__slots__=["x"]' -s' def __init__(self): self.x=23' 'z=[X() for i in 
xrange(9999)]'
10 loops, best of 3: 3.34e+04 usec per loop
[alex at lancelot python2.3]$

Do we "*absolutely* have to have" the roughly factor-of-2 speedup given
by __slots__?  Clearly not -- we COULD live with a program that's
roughly twice as slow, and here we're measuring only initialization
anyway.  But, consider also the memory-churning, particularly when
MORE instances of class X need to be around, say 100,000...:

[alex at lancelot python2.3]$ time python -O timeit.py -s 'class X(object):' -s 
' __slots__=["x"]' -s' def __init__(self): self.x=23' 'z=[X() for i in 
xrange(99999)]'
10 loops, best of 3: 4.28e+05 usec per loop
16.49user 0.74system 0:17.58elapsed 97%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (507major+5528minor)pagefaults 0swaps

[alex at lancelot python2.3]$ time python -O timeit.py -s 'class X(object):' 
-s' def __init__(self): self.x=23' 'z=[X() for i in xrange(99999)]'
10 loops, best of 3: 1.19e+06 usec per loop
47.46user 0.65system 0:49.05elapsed 98%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (507major+13183minor)pagefaults 0swaps
[alex at lancelot python2.3]$

the issue isn't so much that the elapsed time, per-loop and overall, is
now roughly optimized by a factor of THREE, rather than just TWO as when
we were creating just 10,000 instances -- more significant may be the
fact that VM consumption (as measured by page faults, since Linux doesn't
deign tell us the actual numbers;-) decreases by over 2 times too.  Do
we "*absolutely* have to have* that?  By no means, but since it can be
VERY nice indeed to have, what's the big deal with having it *for those
few classes where __slots__'s restrictions are acceptable, and tens or
hundreds of thousands of instances can well be necessary*?  What is it
that you think makes __slots__ so terrible, compared with many other
possible optimizations?!


So my advice remains: consider __slots__ for classes which may need to
be instantiated a HUGE number of times, if the restrictions that __slots__
imposes can be lived with.  Like for all optimizations, you'll probably
want to measure things, at some level, BEFORE you optimize.  Still,
profiling isn't necessarily all that useful here, since __slots__'s main 
use is in reducing your memory footprint (your working-set) and profilers
aren't very good at showing you where you're spending memory and what the
overall effects of your memory consumption on system performance are (maybe
your program's running fine, but if it hogs more memory than it really needs
it's slowing down _other_ simultaneous activities on the same computer -- 
and whether that matters or not may be dependent on considerations not
directly related to your program, that no profiler, per se, can ever 
capture...;-).  [I really wish I had a good way to find out how much
memory a given container -- together with all of its contents -- is using...
unfortunately, I don't know of any good way to answer that question!!!].


Alex




From rmunn at pobox.com  Wed Apr  2 11:59:32 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 16:59:32 GMT
Subject: Pycon DC was terrific
References: <5b4785ee.0303310948.26afe16b@posting.google.com>
Message-ID: <slrnb8lgns.ccp.rmunn@localhost.localdomain>

Wade Leftwich <wade at lightlink.com> wrote:
[snip words of praise about PyCon DC 2003]
> Many thanks to the PSF, the sponsors, the organizers, Guido, and
> conference chairman / master of ceremonies Steve Holden.
> 
> Wade Leftwich
> Ithaca, NY

Hear, hear! *applauds*

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From member27604 at dbforums.com  Thu Apr  3 22:43:19 2003
From: member27604 at dbforums.com (dolephi9080)
Date: Fri, 04 Apr 2003 03:43:19 +0000
Subject: [Que]Do Boa Constructor have any IntelliSense?
Message-ID: <2728009.1049427799@dbforums.com>

Do Boa Constructor have IntelliSense input. Without this function it's
hard to write some a little big applications. And I cannot find some IDE
support IntelliSense very well include the python-win. Do anyone have
any suggesion? Thx

--
Posted via http://dbforums.com



From duncan at NOSPAMrcp.co.uk  Thu Apr 17 07:14:07 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 17 Apr 2003 11:14:07 +0000 (UTC)
Subject: Parsing
References: <9afd3f36.0304170247.2a082573@posting.google.com>
Message-ID: <Xns93607C5492A1Eduncanrcpcouk@127.0.0.1>

iamshady at rediffmail.com (Jim Shady) wrote in 
news:9afd3f36.0304170247.2a082573 at posting.google.com:

> What I'd like to do is run through each entry with x.readlines() and
> put all 4 elements in an array for further processing. Right now I'm
> trying with:
> 
> for chunk in psfile.readlines():
>     if i <= 4:
>         cur_chunk.append(chunk[:-2])
>         i = i+1
>         if i > 4:
>             i = 0
>             cur_chunk = []
> 
> for x in bigchunk:
>     print x
> 
> This somehow doesn't work. I am a newbie to Python and would
> appreciate any help.

What is the relationship between cur_chunk and bigchunk supposed to be? The 
code you posted doesn't do anything with cur_chunk, perhaps you want to 
append it to bigchunk before clearing it.

Other hints:
Forget using readlines, just iterate over the file directly.
Dont try to strip newlines from the end of the line using chunk[:-2], much 
better just to use the strip method (or the rstrip method if you don't want 
to lose leading spaces).
>From the description, it sounds as though the blank lines are separating 
the chunks, so you might want to collect lines up to but not including the 
blank line.

Putting that altogether you might get something like this:

bigchunk = [[]]
for line in psfile:
    line = line.rstrip()
    if line:
        bigchunk[-1].append(line)
    else:
        bigchunk.append([])

# Remove last chunk if it wasn't used.
if not bigchunk[-1]:
    bigchunk.pop(-1)
    
import pprint
pprint.pprint(bigchunk)


-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From martin at v.loewis.de  Wed Apr  2 16:42:14 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 02 Apr 2003 23:42:14 +0200
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
Message-ID: <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>

escalation746 at yahoo.com (robin) writes:

> If you want the biggest, boldest approach and don't care about
> overhead, use CORBA.

I can't agree with that analysis: CORBA has, of all your alternatives,
the least network bandwidth requirements.

Regards,
Martin



From jack at performancedrivers.com  Tue Apr 15 21:54:07 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Tue, 15 Apr 2003 21:54:07 -0400
Subject: Why functional Python matters
In-Reply-To: <slrnb9p3vo.vg8.ramen@lackingtalent.com>; from ramen@lackingtalent.com on Tue, Apr 15, 2003 at 11:04:06PM -0000
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <slrnb9p3vo.vg8.ramen@lackingtalent.com>
Message-ID: <20030415215406.J1039@localhost.localdomain>

On Tue, Apr 15, 2003 at 11:04:06PM -0000, Dave Benjamin wrote:
> But, there is of course the risk of setting precedent that there are more
> than one ways to do it, and this is why I brought up Perl. I think that
> there is a subtle fear in the Python community of letting Python become too
> much like Perl, especially since both languages have similar application
> domains. There are other options besides TMTOWTDI and military rigor. This
> is precisely why I like Python, because it seems to be a good compromise
> between the two.

No problem, lets just yank out list comprehensions :)

list comps versus map/filter comes down to personal preference and coding 
style.  I like map & filter.

fearing TMTOWTDI is good, but python hasn't fallen down because people can
write abominations like

i = 0
sum = 0
while (i < len(mylist)):
  sum = sum + mylist[i]
  i = i + 1

at a guess most people do

sum = 0
for (guy) in mylist:
  sum += guy

or my favorite

sum = reduce(int.__add__, mylist, 0) # forget that zero at your peril

TOOWTDI when overdone makes for a bondage and discipline language.
The BDFL gets to decide the balance, sometimes with people whispering in his
ear, and sometimes with people screaming.

The ternary vote thing was more like a room full of crazy homeless people 
talking to themselves.
  
-jack




From mcfletch at rogers.com  Tue Apr 15 21:20:17 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 15 Apr 2003 21:20:17 -0400
Subject: =?ISO-8859-1?Q?R=E9ception_et_d=E9codage_d=27une__?=
 =?ISO-8859-1?Q?trame_UDP?=
In-Reply-To: <3E9CA7C2.2C9730EA@engcorp.com>
References: <3e9c9b1e$0$1509$7a628cd7@news.club-internet.fr> <3E9CA7C2.2C9730EA@engcorp.com>
Message-ID: <3E9CAFD1.5030402@rogers.com>

Peter Hansen wrote:

>Benoit BESSE wrote:
>  
>
>>Bonsoir tout le monde,
>>
>>qui a d?j? ?crit un module qui ?coute, re?oit, d?code et insert dans une BD
>>une trame UDP.
>>Je voudrais d?coder la trame UDP ainsi et l'ins?rer dans une table :
>>    - Port Source
>>    - Port Destination
>>    - Longueur
>>    - Somme de contr?le
>>    - Donn?es
>>
>>Merci par avance.
>>    
>>
>
>Hmm... almost looks like another "Twisted has that" but I'm 
>not sure what "BD" is so I can't be sure...  ;-)
>
>-Peter
>  
>
I think "BD" is refering to "database", or more generally, a data 
storage mechanism (table is mentioned later), but my french is very 
rusty.  Took me a while to figure out Somme de contr?l (checksum) and 
Donn?es (data AFAICT from context, though it might mean "arguments" 
("given stuff")).  I think the OP is looking for the socket module (i.e. 
a low-level module), though as Peter mentions, Twisted would likely be 
able to handle listening for UDP packets as well.

Original poster, I believe there is a French Python list if you'd like, 
though honestly it's kinda fun piecing together what French messages 
mean :) .

Oh, and to answer the actual question (who's done it already), dozens, 
maybe hundreds or thousands of people, depending on whether you care 
about it being in Python or not, and whether you're going to sweat the 
details :) .

Enjoy,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From dick at charts.org  Mon Apr 21 21:36:15 2003
From: dick at charts.org (Rich)
Date: 21 Apr 2003 18:36:15 -0700
Subject: Keyboard Commands
References: <Pine.SOL.4.51.0304201825001.6411@herald.cc.purdue.edu> <81l7avg7r77ko9p80mggjlfdmb2bd98b51@4ax.com>
Message-ID: <d62f103e.0304211736.3330b672@posting.google.com>

OK that helps some but it still isn't working perfectly.
is this right:

    self.Game.bind('<Control-h>',self.HighlightKey())

    def HighlightKey(self):
        self.FileOptionsHelp.Highlight()
     	if self.Yote.HighlightEnable == 1:
	    self.FileOptionsHelp.Highset.set(1)
        else:
	    self.FileOptionsHelp.Highset.set(0)
    def Highlight(self):
        if self.GameScreen.Yote.HighlightEnable == 1:
            self.GameScreen.Yote.HighlightEnable = 0
        else:
            self.GameScreen.Yote.HighlightEnable = 1


Shouldn't hitting ctrl-h cause HighlitghtEnable to switch values?



QWE <leomasi at libero.it> wrote in message news:<81l7avg7r77ko9p80mggjlfdmb2bd98b51 at 4ax.com>...
> On Sun, 20 Apr 2003 18:27:21 -0500, "richard.a.charts.1"
> <charts at purdue.edu> wrote:
> 
> >Hey,
> >I'm working on a program and I want to use keyboard shortcuts for some of
> >my menu items.  I can't seem to find the way to implement this.  Can
> >anyone point me in the right direction?  ex.  Ctrl-x, ctrl-F1, etc.
> >
> >Thanks for any help.
> >And please feel free to email ur answers to me directly.
> >
> >Charts
> from Tkinter import *
> 
> root = Tk()
> 
> def hello():
>     print "Hello"
> 
> frame = Frame(root)
> root.bind("<Control-p>", lambda e: hello())
> Label(text="Press Controll-p").pack()
> frame.pack()
> 
> root.mainloop()



From dberlin at dberlin.org  Sun Apr 27 11:49:02 2003
From: dberlin at dberlin.org (Daniel Berlin)
Date: Sun, 27 Apr 2003 11:49:02 -0400
Subject: Why doesn't the uu module give you the filename?
Message-ID: <C4C4782E-78C7-11D7-B62F-000A95A34564@dberlin.org>

While it's simple enough to get the uu module to uudecode a string 
(using StringIO), it's impossible to get it to handle you the filename 
the uuencoded thing specifies.

IE given

begin 644 a.ii.gz
<whatever>
end

Their is no way to get the decode function to tell you the thing is 
named a.ii.gz.
Of course, it uses this filename itself in creating an output file if 
you don't specify one.  It just won't tell *you* what the filename is.
I could just give it no output file, and let it create it, then 
determine the name of the file it created, but this seems like a very 
large kludge.
Besides, I am decoding from/to a string, in memory. I don't want to 
start have it write things to the disk for no reason.

The context of all of this is that I have a program that is converting 
text that possibly contains uuencoded attachments into a bunch of SQL 
statements to insert into a database (It's converting a GNATS bug 
database to a Bugzilla one. It's a rewrite of an incredibly ugly, slow, 
barely functional perl script that spews errors at random and leaks 
memory for no reason :P).
I had to cut/paste the decode function from the uu module into a new 
module and make it return the filename, just so that i could get access 
to it.
This seems a bit silly.
The decode function has no return value right now, so giving it one 
shouldn't break existing applications (since none of them should be 
expecting it to return anything).
I believe it should return the filename specified in the begin line.

As an added bonus, it would be even nicer if it also returned the start 
and end position of the decoded portion inside the input text.  that 
way if one wants to replace the entire uuencoded text with something 
like, say, "See bug attachments for <filename>", you can do it easily. 
:P

As i said, i've got a version of uu.decode that does all of this, i'll 
happily submit it as a patch if people agree i'm right.
--Dan





From jerf at jerf.org  Tue Apr  1 17:06:49 2003
From: jerf at jerf.org (Jeremy Bowers)
Date: Tue, 01 Apr 2003 22:06:49 GMT
Subject: I need a lot of help!
References: <31ef0f43.0304011336.1d2777ac@posting.google.com> <f_2cnTUxwtvYlRejXTWcqQ@comcast.com>
Message-ID: <pan.2003.04.01.17.06.39.649300@jerf.org>

On Tue, 01 Apr 2003 16:49:14 -0500, Terry Reedy wrote:
> You appear to be programming in some version of Basic.  This is the
> newsgroup for Python, which is quite different.  You need to ask your
> questions elsewhere.

It also sounds an awful lot lot like homework...



From logistix at cathoderaymission.net  Sun Apr 20 22:42:51 2003
From: logistix at cathoderaymission.net (logistix)
Date: Sun, 20 Apr 2003 22:42:51 -0400
Subject: importing modules with the same name but from different directories
In-Reply-To: <b7vhad$1t6$1@news.tu-darmstadt.de>
Message-ID: <000001c307af$b4cd23c0$20bba8c0@XP>


> -----Original Message-----
> From: python-list-admin at python.org 
> [mailto:python-list-admin at python.org] On Behalf Of Mirko Koenig
> Sent: Sunday, April 20, 2003 9:26 PM
> To: python-list at python.org
> Subject: importing modules with the same name but from 
> different directories
> 
> 
> Hi
> 
> The subject is a little bit long but i didn't how how to 
> describe it shortly.
> 
> I hope my english is good enough to explain my problem.
> 
> I have a program consisting of many modules. Every modul has 
> its own directory in a directory called modul, eg.
> 
> .../modul/modul1
> .../modul/modul2
> 
> In every modul directory is a diretory for the language 
> files. These files are used to display the modules label in 
> different languages. eg.
> 
> .../modules/module1/language/german.py
> .../modules/module1/language/english.py
> .../modules/module2/language/german.py
> .../modules/module2/language/english.py
> 
> My problem is:
> If is start modul1 from within its modul directory ( 
> .../modules/modul1 ) And modul1 calls modul2, then modul2 
> doesn't load its own language file. Instead it loads the 
> langugae file of modul1.
> 
> In modul1 i wrote sys.path.append( "./languages" ) then 
> import german. In modul2 i wrote the same. I know in that way 
> it coulnd't work.
> 
> What can i do, that every modul knows its 'homedir', no 
> matter from where i start it ?
> 
> I don't want to set environment variables, because the 
> programm is for beginners, that are not used to pc's and want 
> an easy going programm and not to set env vars.
> 
> I'm developing under linux if that help solving my problem
> 
> thanx for comments
> 
> Mirko Koenig
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 

Add an __init__.py file to each directory in your project.  This makes
the directories packages.  Then you can use a fully qualified path such
as "import modules.modul1.language.english as module1_english".  Note
that you can't use the same name for different modules at runtime, so
use "as xxx" to give them a more distinct name.





From tismer at tismer.com  Tue Apr 22 22:28:15 2003
From: tismer at tismer.com (Christian Tismer)
Date: Wed, 23 Apr 2003 04:28:15 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <20030422202417.A1462@tibia.amotlpaa.bogus>
References: <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <3EA2C43F.2080802@tismer.com> <20030421121843.A596@tibia.amotlpaa.bogus> <3EA4A843.2050507@tismer.com> <20030422202417.A1462@tibia.amotlpaa.bogus>
Message-ID: <3EA5FA3F.90305@tismer.com>

Steven Taschuk wrote:
> Quoth Christian Tismer:
> 
>>Steven Taschuk wrote:
> 
>   [...]
> 
>>>Before studying continuations, function calls are function calls
>>>and loops are loops.  While studying continuations, function calls
>>>are not function calls and loops are not loops.  After studying
>>>continuations, once again function calls are function calls and
>>>loops are loops.
>>
>>But then, you should have understood how to build function
>>calls and loops from continuations, and that in fact the is no
>>"call" or "return" in the first place, but these are artificial,
>>from a POV where continuations are the basic building block.
> 
> 
> Indeed; after studying continuations you are enlightened.
> 
> (My blurb above is a well-known quote about Zen, after the
> substitutions
>     s/Zen/continuations/g
>     s/men/function calls/g
>     s/mountains/loops/g

Continuations aren't perlish.
I have advertized them to them, but they didn't want 'em,
although the Perl engine is sa(i)d to be stackless.
Maybe, this is for the better. Perl is already slightly
loaded with features... (psssst, did I say anything? No, I didn't)

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From usenet_spam at janc.invalid  Tue Apr  8 12:14:56 2003
From: usenet_spam at janc.invalid (JanC)
Date: Tue, 08 Apr 2003 16:14:56 GMT
Subject: What did happen to PyXPCOM?
References: <mailman.1049616578.27671.python-list@python.org>
Message-ID: <Xns9357B9A342JanC@213.119.4.37>

Alessandro Bottoni <alessandro.bottoni at infinito.it> schreef:

> I recently visited the ASPN page devoted to PyXPCOM and discovered
> that the PyXPCOM tar.gz/zip files are not on the web site any more.
> The most obvious explanation is that PyXPCOM is now hosted by the cvs
> of Mozilla, as promised. 

It's in CVS but not in the standard binaries:
<http://lxr.mozilla.org/seamonkey/source/extensions/python/>
<http://lxr.mozilla.org/seamonkey/source/extensions/python/xpcom/doc/configure.html>

-- 
JanC

"Be strict when sending and tolerant when receiving."
RFC 1958 - Architectural Principles of the Internet - section 3.9



From heiland at ncsa.uiuc.edu  Fri Apr 25 07:40:24 2003
From: heiland at ncsa.uiuc.edu (Randy Heiland)
Date: Fri, 25 Apr 2003 06:40:24 -0500
Subject: PyXML - setContentHandler??
Message-ID: <3EA91EA8.2050505@ncsa.uiuc.edu>

I'm using a chunk of code that does:

 >>> parser = xml.sax.make_parser()
<define a handler>
 >>> parser.setContentHandler(handler)

however, using PyXML, I get:
AttributeError: SAX_expat instance has no attribute 'setContentHandler'

Any advice welcome.  Thanks.
--Randy




From sandskyfly at hotmail.com  Mon Apr 28 12:18:25 2003
From: sandskyfly at hotmail.com (Sandy Norton)
Date: 28 Apr 2003 09:18:25 -0700
Subject: Pyrex Hard Case
References: <607d8200.0304272112.2f02bcf5@posting.google.com>
Message-ID: <b03e80d.0304280818.5d36a998@posting.google.com>

You may want to also email the Pyrex list: pyrex at lists.copyleft.no 

ciao,

Sandy



From edsZZZ_2 at yahoo.com  Mon Apr 14 15:57:12 2003
From: edsZZZ_2 at yahoo.com (Eric Scharff)
Date: 14 Apr 2003 13:57:12 -0600
Subject: encapsulation checking features
Message-ID: <3e9b1298@cs.colorado.edu>

I love python's semantics for rapid prototyping, but I was wondering if
there was a way to turn on warnings so that for other kinds of code, I
can avoid certain common uglinesses. :)

Specifically, is there a "lint-like" tool to:

 * warn if instance variables are created outside of a constructor?

  class K:
    def __init__(self):
      self.a = 6  #okay
    def example(self):
      self.b = 66 # not as good

  myVar = K()
  myVar.c = 666  # really bad

 * warn if instance variables are used outside of instance methods
   (encouraging the use of getters and setters)
  
These are great features to have, but it'd be nice to have warnings to
discourage their use at times.

-Eric
 (genuine email replies, remove ZZZ from email address)



From nospam at somedomain.com  Thu Apr 10 08:27:06 2003
From: nospam at somedomain.com (maxx)
Date: Thu, 10 Apr 2003 14:27:06 +0200
Subject: newbie: strange python problem
Message-ID: <pan.2003.04.10.12.27.06.600319@somedomain.com>

Simple code (I think):

---
#!/usr/bin/python

import os

for line in os.popen('ls -la /').readlines():
    direntry = line.split()
    print direntry
    print direntry[0]
    print direntry[1]
    print direntry[2]
---

No list-items with an index above 1 are printed. They produce a
'IndexError: list index out of range' error... which is strange, because
the list 'direntry' is 9 items large, so listitems up to index 8 should be
printed... right?

maxx



From tim_one at email.msn.com  Sat Apr 26 10:35:46 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sat, 26 Apr 2003 10:35:46 -0400
Subject: RuntimeError at end of test.autotest in Python 2.3b on windows, tmpfile also fails
In-Reply-To: <3eaa75a0$0$49106$e4fe514c@news.xs4all.nl>
Message-ID: <LNBBLJKPBEHFEDALKOLCCEDIEIAB.tim_one@email.msn.com>

[Irmen de Jong]
> when running "import test.autotest" in Python2.3b on windows xp,
> I get a RuntimeError at the end:

As the download page says, please report bugs on SourceForge.

It so happens that the RuntimeError is a consequence of test_imp failing,
which in turn is a consequence of doing "import test.autotest" instead of
"python lib/test/regrtest.py", so you can forget that one (I'm about to
check in a fix, as the cause happened to be obvious to me).  I don't know
about your other failure, and it's impossible to track clues via a newsgroup
(that's what the bug tracker is for).  Please report what you know about
that on SourceForge; the tests all passed on Win2K and Win98SE using Pythons
installed from the installer we shipped.





From rainer.klueting at epost.de  Tue Apr  8 08:12:39 2003
From: rainer.klueting at epost.de (Rainer Klueting)
Date: Tue, 08 Apr 2003 14:12:39 +0200
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com> <urnka.4658$Du.16154@newsc.telia.net> <b68f71a7.0304080323.75b5ed1a@posting.google.com>
Message-ID: <2me59vcigrdkoeapbtb267i46jf07kr9gr@4ax.com>

Hi Alan,

On 8 Apr 2003 04:23:24 -0700, alan_salmoni at yahoo.com (Alan James
Salmoni) wrote:

>As Tim said, wxPython has an excellent wxGrid widget capable of doing
>what you want (see http://salstat.sunsite.dk/screenshots.html for some
>examples), but I also understand that there is a similar kind of thing
>which can be used with the Tkinter toolkit known as the tkTable. I
>haven't used it myself and therefore cannot comment on it, but there
>is a spreadsheet written in it called Novagrid which you might like to
>try:
>
>http://sourceforge.net/projects/novagrid
>
>Hope this helps!

There is a folder on my hard disc called "download", and in there is a
file named novagrid-1.0.tgz.gz. ;-) I downloaded it about a week ago
but didn't try it.

On my lengthy search for Python stuff I had already found some
information on tkTable. May be I misunderstood something --I dropped
that, because I thought, tkTable is written in TCL, and there is no
Python-interface.

Okay, I will find out: I only have to install novagrid!

Thank you

Rainer Klueting, Stuttgart, Germany




From phil at river-bank.demon.co.uk  Sun Apr 13 19:10:10 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Mon, 14 Apr 2003 00:10:10 +0100
Subject: Implementing Metatypes in C
Message-ID: <200304140010.10884.phil@riverbankcomputing.co.uk>

I want to create a type (in C) that has some additional, type specific, 
information associated with it.

That type will then have many sub-types created, also in C. Instances of those 
sub-types will then be created in Python.

As I understand it, I have to create a metatype for my new types so that they 
get created with the correct size to contain the additional I want to store 
in the type. To do this I am simply sub-typing from PyType_Type but 
specifying an "appropriate" value for the tp_basicsize field. The sub-types 
are then created on the heap by calling the metatype.

The problem I had was understanding what the "appropriate" value was. At first 
my metatype data structure was...

struct metatype {
	PyTypeObject supertype;
	ExtraTypeData extra;
};

Eventually I worked out that Python allocates additional fields (as well as 
PyTypeObject) when allocating new types on the heap. This is from 
typeobject.c in the Python 2.3a2 source...

/* The *real* layout of a type object when allocated on the heap */
/* XXX Should we publish this in a header file? */

If I add the extra fields in my struct metatype, everything works.

Two questions...

1. Is my understanding correct, or is there a more correct way to do what I 
want that avoids the problem?

2. If my solution is correct, can the structure defined in typeobject.c be 
published in a header file as the comment says? Please?

Thanks,
Phil




From neil.padgen at mon.bbc.co.uk  Tue Apr 29 05:15:39 2003
From: neil.padgen at mon.bbc.co.uk (Neil Padgen)
Date: Tue, 29 Apr 2003 09:15:39 +0000
Subject: Function to convert fetchone() result from a list to a dictionary
References: <c31eab9b.0304282332.32799f80@posting.google.com>
Message-ID: <b8lfrr$21c$1@nntp0.reith.bbc.co.uk>

<posted & mailed>

On Tuesday 29 April 2003 7:32 am, Carsten Gehling wrote:

> I'm fairly new to Python, and have yet to discover alle the little
> details, that may make my programs look leaner and perhaps become
> faster...
> 
> Using the DB API to access databases, I've stumbled upon what I
> would call a shortcoming in the API (IMHO). All the fetch commands
> return the results as lists instead of dictionaries.

Take a look at the cursorclass argument:

Python 2.1.3 (#1, Jan 22 2003, 15:10:34) 
[GCC 2.95.3 20010315 (release) [FreeBSD]] on freebsd4
Type "copyright", "credits" or "license" for more information.
>>> import MySQLdb
>>> db = MySQLdb.connect(db='fbis_test')
>>> c = db.cursor(cursorclass=MySQLdb.cursors.DictCursor)
>>> c.execute("""SELECT * FROM city""")        
613L
>>> c.fetchone()
{'cou_id': 'SUI', 'id': 'AARAU', 'city_name': 'AARAU'}
>>> 

-- Neil




From zhaojs at yahoo.com  Wed Apr 30 19:30:26 2003
From: zhaojs at yahoo.com (Jinsong Zhao)
Date: 30 Apr 2003 16:30:26 -0700
Subject: Ctrl-C to stop C++ code in python
Message-ID: <kfellxr5zsd.fsf@alumni.cse.ucsc.edu>

Some of my problems arise from extension in C++. Here is one of them. 

I have C++ codes that do a lot of computations in seconds to
minutes. The user may get bored and want to stop it by pressing
Ctrl-C, when it is running inside the C++ code. Unfortunately,
currently python does not detect that until the control is back to
python, which may be one minute later. This delay is very
frustrating.

I haven't found any solution in the Python Cookbook. Anyone to give a
hint as to how to achieve the "immediate kill" effect?

Jinsong



From dcengija_remove_ at inet.hr  Wed Apr 16 03:07:40 2003
From: dcengija_remove_ at inet.hr (Davor Cengija)
Date: Wed, 16 Apr 2003 09:07:40 +0200
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net> <slrnb9ooms.sho.ramen@lackingtalent.com> <pan.2003.04.15.22.22.18.262135@removethis.rosecott.ukfsn.org>
Message-ID: <1216579.Cm5jTlKPyA@lothar.cengija.com>

Ken Parkes wrote:

> On Tue, 15 Apr 2003 19:51:37 +0000, Dave Benjamin wrote:
> 
> 
>> Also, don't forget, sometimes the best things in life are free:
>> http://www.ibiblio.org/obp/thinkCSpy/
> 
> 
> Have you tried printing this?  When I try to print it on an HP Deskjet
> 840C (CUPS on Mandrake 9) I lose the last letter on every line and the
> last two lines on every page completely.  Have fiddled endlessly to no
> avail....and it looks such a good intro.

It seems that the PDF file is somehow broken.

Download the latex sources and compile them, the instructions are included. 
Now you should have the correct PDF.
-- 
Davor Cengija, dcengija_remove_ at inet.hr



From wtung at cisco.com  Tue Apr 22 14:29:33 2003
From: wtung at cisco.com (Wai-Yip Tung (wtung))
Date: Tue, 22 Apr 2003 11:29:33 -0700
Subject: Off topic: What usenet news reader to use?
Message-ID: <003901c308fd$1f2b5ef0$9c8e6b80@amer.cisco.com>

I haven't used Usenet news for a long time. Now that I have subscribed
to python-list I find that email does not scale well to the high volumne
of discussion. Instead of deleting emails all day I think I should go
back to Usenet news. Is there any good Windows news reader you would
recommend. I'm happily using Opera as the browser. But it's
revolutionary news and email reader eludes me. What are you using?

By the way I'm also looking for a free or low cost email client. My
company is using Outlook but I want to find one for personal use. Thank
you for you suggestions.

Wai Yip Tung





From duncan-news at grisby.org  Thu Apr 10 20:03:33 2003
From: duncan-news at grisby.org (Duncan Grisby)
Date: Fri, 11 Apr 2003 00:03:33 +0000
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <d116fbae.0304100704.68ef4833@posting.google.com>
Message-ID: <c48$3e960655$516049d2$3498@nf1.news-service.com>

In article <d116fbae.0304100704.68ef4833 at posting.google.com>,
Uche Ogbuji <uche at ogbuji.net> wrote:

>True, and Mike Olson did some pretty thorough analysis on this.  See:
>
>http://www-106.ibm.com/developerworks/library/ws-pyth9/

Interesting article. Unfortunately, the first two timings of the CORBA
client aren't timing what Mike thinks they are. The first time, for
"connecting to server" doesn't actually connect to the server at all.
It just creates an object reference for it.

The second time, for "send string" _is_ timing sending a string, but
that is also when the TCP connection to the server is made. A
significant portion of the time is spent setting up the connection,
not transferring the string. If I modify the client to do the first
call twice, I get

Connecting to server
Time to connect to server, 0.000386

Sending a long string to the server
Time to send a string of 21000 chars, 0.002099

Sending a long string to the server
Time to send a string of 21000 chars, 0.001034

Recieving a long stirng from the server
Time to receive a string of 22000 chars, 0.001088

Sending lots of ints to the server
Time to send 5000 ints, 0.921309 (0.000184 per call)


So you see that the first call takes about twice as long as the second
one. Another interesting thing if you care about raw speed is that
CORBA strings are not allowed to have embedded nulls in them, and
undergo code set conversion, and checking these things slows them
down. Using a sequence of octets, which doesn't have any checks or
conversions brings the time down to 0.000827 and 0.000839 for sending
and receiving respectively.

For comparison, the raw socket client has these times on my machine:

Connecting to server
Time to connect to server, 0.001273

Sending a long string to the server
Time to send a string of 21000 chars, 0.000988

Recieving a long stirng from the server
Time to receive a string of 22000 chars, 0.000873

Sending lots of ints to the server
Time to send 5000 ints, 10.290595 (0.002058 per call)


Lies, damn lies, and statistics...

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --



From aleax at aleax.it  Mon Apr 14 10:30:38 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 14:30:38 GMT
Subject: accessing a functions var from a subfunction.
References: <a3cae241.0304140616.1833e5c4@posting.google.com>
Message-ID: <iCzma.16405$T34.518484@news2.tin.it>

Sebastian Wilhelmi wrote:

> Hi,
> 
> I would like to do the following:
> 
> -------8<-------8<-------8<-------8<-------
> def test ():
>     count = 0
>     def inc_count ():
>         count += 1
>     inc_count ()
>     inc_count ()
>     print count
> 
> test ()
> -------8<-------8<-------8<-------8<-------
> 
> This doesn't work (and I even understand, why ;-)

Specifically: a nested function cannot *RE-BIND* a variable of
an outer function.

> One solution is the following, which I however do not see as very
> clean or nice.
> 
> -------8<-------8<-------8<-------8<-------
> def test ():
>     count = [0]
>     def inc_count ():
>         count[0] += 1
>     inc_count ()
>     inc_count ()
>     print count[0]
> 
> test ()
> -------8<-------8<-------8<-------8<-------
> 
> Now my question: Is there some way to achieve this with a nicer
> syntax?

Depends on your tastes in syntax, e.g.:

def test():
    class Bunch: pass
    loc = Bunch()
    loc.count = 0
    def inc_count():
        loc.count += 1
    inc_count()
    inc_count()
    print loc.count

or:

def test():
    test.count = 0
    def inc_count():
        test.count += 1
    inc_count()
    inc_count()
    print test.count

and no doubt quite a few others.

> Using 'global' would not count, as that would make a variable
> unnecessarily known to the outside.

The two solutions above outlined differ in this respect -- variable
'loc' in the former is local to function test, while function attribute
test.count in the latter is "known to the outside" (it survives each
execution of test and can be examined "from the outside").


Class Bunch (or some highly refined version thereof) IS typically
around whenever I program, see for example:

http://mail.python.org/pipermail/python-list/2002-July/112007.html

for a typical presentation.  Having available the MetaBunch there
described, I'd start the function as follows:

def test():
    class Bunch(MetaBunch): count=0
    loc = MetaBunch()
    def inc_count():
        loc.count += 1

etc.  I do find such constructs often handy...


Alex




From aleax at aleax.it  Sun Apr 27 03:45:07 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 07:45:07 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com>
Message-ID: <7ULqa.9804$3M4.264757@news1.tin.it>

Terry Reedy wrote:

> "Erik Max Francis" <max at alcyone.com> wrote in message
> news:3EAB6A14.707AD826 at alcyone.com...
>> Jeremy Fincher wrote:
>>
>> > Perhaps the answer is to remove the __nonzero__ method on
>> > functions/methods.  Logically, what purpose does it serve?
>>
>> # f holds a callable object or None
>> if f:
>>     f(...)
> 
> or
> print f and f() or 'cannot do that' # assumes boo(f()) always True
> 
> I consider it a feature that all objects have truth value (except for
> perversely written user classes), so that conditional expressions and
> statements always work without exception.

It's a feature, just like it WAS a feature that you USED TO be able
to always test "if a>b:" without exception.  But features can and do
get removed when it's deemed too likely that they "let errors pass
silently" -- the ability to compare a>b unrestrictedly was removed
in the move 2.0 -> 2.1 (when either comparand is a complex and the
other any kind of number) and it may get restricted even further.  By
the same token, the ability to "use the truth value" of a function or
method object COULD be removed for similar reasons.

It's sad that this enhancement cannot be used, to support such tests
as "if f:" rather than "if f is not None:".  But -- I would HOPE
somebody tried to make to the BDFL that "the use case" for keeping
functions and methods 'true' would be that horrid single-line print,
so unreadable, fragile, and unPythonic... if you managed to convince
him that THIS is the usage that such truth values empower, then I
think we'd see the truth value go away by 2.3 beta 2!-)


Alex




From duncan at NOSPAMrcp.co.uk  Wed Apr 16 09:39:41 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 16 Apr 2003 13:39:41 +0000 (UTC)
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <roy-92A049.08504216042003@reader1.panix.com>
Message-ID: <Xns935F94F66A99Eduncanrcpcouk@127.0.0.1>

Roy Smith <roy at panix.com> wrote in news:roy-
92A049.08504216042003 at reader1.panix.com:

>  Imagine you've got a program which runs in O(N^2) 
> time.  If you could find a way to reduce it to O(N), for an input set of 
> just 10 items, you would have speeded it up by a factor of 10!  Compared 
> to upgrading your Python interpreter, you did 50-100 times better tuning 
> the algorithm.  Now try to consider the implications of going from 
> O(N^2) to O(N) with 25,000 input items!

I believe you are making some unreasonable assumptions here. Remember that 
if I have an algorithm that is O(N^2) that is really just a shorthand for 
saying that it will have a running time a*N^2 + b*N * c where a, b, and c 
are constants but for sufficiently large N only the N^2 term matters.

Now imagine that you have a program that runs in O(N^2) time and you have 
10 items. If a is 1 and b is 1000, the N^2 term doesn't start to dominate 
until you have at least 1000 items. If these are times in milliseconds, 
then your 10 item program would take ~10 seconds to run, but improving the 
algorithm from O(N^2) to O(N) (with the same value for b) would have no 
noticeable effect.

Yes, if you have 25,000 items then for my chosen values of a and b the O(N) 
algorithm is probably going to result in a significant speedup (unless it 
increases b by a factor of more than 26), but you can't just simplify this 
to say that O(N) is going to be better in all cases.

If you try to apply this to the real world you may find there are all 
manner of compromises to be made. The most obvious one is in sorting where 
a real implementation of an O(n log n) sort will adapt to using an O(n^2) 
sort on subsets of the data because it turns out that the dumb sorting 
techniques are faster until n gets quite large.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From uselpa at myrealbox.com  Tue Apr 29 15:03:41 2003
From: uselpa at myrealbox.com (Patrick Useldinger)
Date: Tue, 29 Apr 2003 21:03:41 +0200
Subject: zipfile with umlaut in filename
In-Reply-To: <m3fzo28ip9.fsf@mira.informatik.hu-berlin.de>
References: <3ea9a4bf$1_1@news.vo.lu> <b8caue$g65$05$3@news.t-online.com> <3eaa52b1_1@news.vo.lu> <m3fzo28ip9.fsf@mira.informatik.hu-berlin.de>
Message-ID: <3eaecb38_2@news.vo.lu>

Martin v. L?wis wrote:

> And why would you think that \x94 is "o-umlaut"?

I don't, WinZip does ;-)
And being compatible to WinZip doesn't seem so bad to me.

> It appears that Python puts the non-ASCII characters into the zipfile
> as the system reports them (i.e. in the ANSI code page on your
> system); winzip apparently converts them to some other encoding first,
> probably code page 437 or code page 850.
> 
> To get the effect of Winzip in Python, when adding a file with name N,
> do
>   zfile.write(N, unicode(N, "cp1252").encode("cp437"))
> and then try to open the zipfile with Winzip.

You are right, this works (both code pages, actually). I must definitely 
revise the encoding and codepage stuff... but you've saved me from 
writing a decompression program myself.

Thank you.





From syver-en+usenet at online.no  Wed Apr  2 08:10:07 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 02 Apr 2003 15:10:07 +0200
Subject: (no subject)
References: <mailman.1049246464.32312.python-list@python.org>
Message-ID: <uk7edjb9s.fsf@online.no>

"Santoshi Reddy" <santoshi_reddy at hotmail.com> writes:

> hi I am new to python.can someone tell me what _int_ is?

You mean integer? That is a whole number, it is a numeric datatype
that supports only numbers without fractions. (1 not 1.6)

-- 

Vennlig hilsen 

Syver Enstad



From lost at soul.net  Mon Apr 14 19:33:01 2003
From: lost at soul.net (lost)
Date: Tue, 15 Apr 2003 00:33:01 +0100
Subject: splitting strings into variables
Message-ID: <pan.2003.04.14.23.33.00.456110@soul.net>

hi there,
i've read about splitting strings, and can do
strname.split(delimeter)
but what i need is something like this:
strname = hello:there
strname.split(":")
print strname
hello

so :there is taken away, and i just have hello?  how can i do that with
split?

thanks in advance!




From jolderdissen at ixiacom.com  Wed Apr 30 16:44:41 2003
From: jolderdissen at ixiacom.com (Jan Olderdissen)
Date: Wed, 30 Apr 2003 13:44:41 -0700
Subject: win32 adding a tcpip alias to an interface
References: <m3el3k2avv.fsf@moroni.pp.asu.edu>
Message-ID: <b5d0bv86tr41122td0ss0tuooi818htcts@4ax.com>

David,

the netsh command line utility should be very helpful here - it is
Microsoft's command line interface to network configuration. In
particular, 'netsh interface ip add address <arguments>' comes to
mind.

Start up netsh and play with help a bit. The interface is a bit
retarded, but you can figure it out after a bit of twiddling.

Jan

On 30 Apr 2003 09:44:20 -0700, David Bear <david.bear at asu.edu> wrote:

> But, I'm worried about how one might add an ip address to
> a windows tcpip stack.  Anyone have any ideas how to do this?




From mwm at mired.org  Thu Apr 10 12:29:01 2003
From: mwm at mired.org (Mike Meyer)
Date: Thu, 10 Apr 2003 16:29:01 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org> <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk> <x7el4jibzl.fsf@guru.mired.org> <Kp2ja.2589$y3.426276@news010.worldonline.dk> <x7znn6gkfn.fsf@guru.mired.org> <3e912546$0$10390$edfadb0f@dread11.news.tele.dk> <swYka.972$Jl.399908@read2.cgocable.net>
Message-ID: <x7u1d69v01.fsf@guru.mired.org>

[Context recovered from top posting.]

myk <mykbourassa at cogeco.ca> writes:
> Anders J. Munch wrote:
> >"Mike Meyer" <mwm at mired.org> wrote:
> >>>No limit on programmer time?
> >>You just have to hire more programmers.
> >That only works if your project isn't late <0.7 wink>
> 
> Hire more programmers?!?!?!  Mr. Brooks ("The Mythical Man-month" -
> Addison-Wesley) would have you for breakfast :)

You're not hiring more programmers to speed up the project, you're
hiring more programmers to replace the ones who died of old age.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From dirk at gerrits.homeip.net  Mon Apr 14 14:19:49 2003
From: dirk at gerrits.homeip.net (Dirk Gerrits)
Date: Mon, 14 Apr 2003 20:19:49 +0200
Subject: Protected Methods and Python
In-Reply-To: <3E99B586.5040107@skynet.be>
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <3E99B586.5040107@skynet.be>
Message-ID: <b7eu2s$gg1$1@main.gmane.org>

Jan Dries wrote:
> Don't worry too much about protected methods. Bjarne Stroustrup, creator 
> of the C++ language, considers them to be a design mistake, and now 
> regrets they were ever added to C++. See "The Design and evolution of 
> C++" by Bjarne Stroustrup, page 302.
> Java and I assume Ruby as well simply copied the concept from C++.

I don't have the book with me, but are you sure it wasn't protected DATA 
members that Bjarne regretted?

Regards,
Dirk Gerrits






From fuka at fuxoft.cz  Thu Apr 24 05:35:41 2003
From: fuka at fuxoft.cz (Frantisek Fuka)
Date: Thu, 24 Apr 2003 11:35:41 +0200
Subject: Problem with IDLE
Message-ID: <b88b6g$5u$1@ns.felk.cvut.cz>

I tried using BaseHTTPServer in my application which I develop using 
IDLE on Windows 98 SE. The problem is  that while the app is running, 
the server log is going to stdout but IDLE is not displaying it! In 
fact, IDLE looks totally dead, not even allowing me to switch to its 
window and issue "Ctrl C" to stop my app. Only when my app stops or 
crashes, all the stdout output appears in IDLE window and I regain 
control. Does this mean I cannot debug this sort of app using IDLE? Or 
am I supposed to add something to the main application loop 
(serve_forever()) so that IDLE "has time" to update and check for Ctrl C?

In the end, I debugged the app by editing in IDLE and running it from 
Cygwin Python opened in another window. :)

By the way, the app (my first) can be used to get URL addresses of 
QuickTime movie files at www.apple.com/trailers - useful if you have 
slow connection and have to use downloader programs to get the files 
(because they normally doesn't allow Resume). You can get it from 
http://www.fuxoft.cz/slowtime/slowtime.py




From nav at adams.patriot.net  Fri Apr 25 13:12:25 2003
From: nav at adams.patriot.net (Nick Vargish)
Date: 25 Apr 2003 13:12:25 -0400
Subject: OO + Python Developers Wanted
References: <mailman.1051183297.29110.python-list@python.org>
Message-ID: <yyyznmezekm.fsf@adams.patriot.net>

"Neeta Ramchandani" <nramchandani at harveynash.com> writes:

> This is a role for an Investment Bank in the City, for duration of
> 3-6 months. Rates are OPEN!!

Which "City" would that be?

Nick

-- 
#include<stdio.h> /* SigMask 0.3 (sig.c) 19990429 PUBLIC DOMAIN "Compile Me" */
int main(c,v)char *v;{return !c?putchar(*v-1)&&main(0,v+ /* Tweaks welcomed. */
1):main(0,"Ojdl!Wbshjti!=obwAqbusjpu/ofu?\v\1");}  /* build: cc -o sig sig.c */



From bokr at oz.net  Wed Apr 23 16:16:14 2003
From: bokr at oz.net (Bengt Richter)
Date: 23 Apr 2003 20:16:14 GMT
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <mailman.1051111176.26457.python-list@python.org>
Message-ID: <b86sae$as7$1@216.39.172.122>

On Wed, 23 Apr 2003 11:18:45 -0400, Oren Tirosh <oren-py-l at hishome.net> wrote:

>On Wed, Apr 23, 2003 at 03:44:57PM +0200, Frantisek Fuka wrote:
>> I want a simple test to see if str contains any occurences of substr. I 
>> am currently using this:
>> 
>> import string
>> if string.find(str,substr) + 1:
>>     ...do something...
>> 
>> But it somehow looks "ugly" to me. Is there more "standard" way to test 
>> this?
>
>if haystack.count(needle):
>   ...
>
>"If the count of needles in haystack is nonzero then..."
>
Except it's unnecessary to count past one, so it's sometimes
going to be inefficient.

Regards,
Bengt Richter



From delete.this.part.joelj at attbi.com  Wed Apr 30 13:57:19 2003
From: delete.this.part.joelj at attbi.com (Joel Jensen)
Date: Wed, 30 Apr 2003 17:57:19 GMT
Subject: Python / Zope Flash Remoting
References: <MPG.1919beb84d09f671989680@netnews.attbi.com>
Message-ID: <MPG.1919ca7a9399cd85989682@netnews.attbi.com>

Sorry about the repost-- my email address was wrong.


I am trying to convert some perl/php code to python for use in zope.
I have never done anything like this. 

If anyone would like to help, please post a response.


Flash Remoting is a way to share datastructures between Cold Fusion and 
flash,make a dataset, varible, or array in cold fusion/java, retreve it 
as a map,variable,array in flash via flashs' OO language actionscript.
http://www.macromedia.com/software/flashremoting/


This protocol is referred to by macromedia as AMF
These guys:
http://sourceforge.net/projects/amfphp
http://www.simonf.com/flap/

have hacked the protocol and written scripts to send datastructures from 
perl and php.
Make an array in PHP, call it from flash..

It would be nice if there was a version of this code available in 
python.

Zope is a Python based Object Oriented web application platform.
www.zope.org 
If flash could share datasets with ZOPE, a developer could make a Zope 
site to contain the data structure, searches, and navigation structure, 
and flash designers could access these datastructures to easily build 
dynamic flash sites.

Zope and flash already support xml-rpc, however AMF is far better.
http://www.zope.org/Members/teyc/howxmlrpc

thanks

/joel




From skip at pobox.com  Wed Apr  2 11:38:43 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 2 Apr 2003 10:38:43 -0600
Subject: distributed computing implementations
In-Reply-To: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
Message-ID: <16011.4627.603644.908973@montanaro.dyndns.org>

    robin> I am attempting to summarise the distributed computing
    robin> implementations available to Python programmers. My conclusions
    robin> so far are as follows:

    [CORBA
     SOAP, XML-RPC
     Dopy, Pyro, Twisted Spread
    ]

    robin> Comments?

If you're willing to bend multi-language interoperability a bit you can
speed up XML-RPC quite a bit by using marshal or pickle to prepackage your
arguments and return values.  Saves wasting a helluva lot of little angle
brackets.  I have two versions of frequently called methods on my server,
one that is Python-only (does the marshal thing) and one that is
cross-platform (doesn't marshal args and return values).

Skip




From staschuk at telusplanet.net  Sun Apr 20 02:31:53 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 20 Apr 2003 00:31:53 -0600
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <Iuhoa.58588$T34.1568946@news2.tin.it>; from aleax@aleax.it on Sat, Apr 19, 2003 at 07:31:52PM +0000
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it>
Message-ID: <20030420003153.C8441@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
> Steven Taschuk wrote:
  [...]
> > What if operator.add were changed to take any number of arguments?
> > Then you could write
> >     operator.add(*[y[1] for y in x])
> > Clear?  Abstruse?
> 
> That would be quite acceptable.  [...]
> What would you have operator.add return when called with no
> arguments, though?  Either 0 or 0.0 or '' or ... have problems.

That is a problem.  Of course, reduce solves it tidily.  :)

  [...]
> Guess we'll have to agree to disagree.  [...]  I now
> think that reduce is more trouble than it's worth and Python
> would be better off without it, if it was being designed from
> scratch today -- it would not substantially reduce (:-) Python's 
> power and WOULD substantially ease the teaching/&c task.  [...]

I'll defer to your experience teaching Python.  (I have none.)

On reflection, I realize that I have prior experience (from math
reading and play) with many concrete examples of what reduce
abstracts: big sigma notation for sums, big pi notation for
products, similar notations for set unions and intersections,
etc..  Familiarity with these concrete examples makes the
abstraction of reduce sit very comfortably.  "I see, reduce turns
+ into big sigma."  But perhaps such familiarity is not typical in
Python's, er, target market.

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From tim at bladerman.com  Fri Apr  4 14:14:46 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Fri, 04 Apr 2003 19:14:46 GMT
Subject: Interactive loop with InteractiveInterpreter
References: <mailman.1049481848.13031.python-list@python.org>
Message-ID: <GQkja.4019$Du.13206@newsc.telia.net>

"Michael Chermside" <mcherm at mcherm.com> wrote

> > If I am about to send a for statement or a while statement to the
> > interpreter I skip that line and set up my own loop with the parameters
from
> > the the loopline and then feed each line to the interpreter several
times,
> > handling the looping myself.
>
> If ALL you have to handle is the code:
>
>     for i in list:
>         print i
>
> Then I think you can probably "set up your own loop". But I hope you
> realize that this doesn't generalize very well. If you had this:

Hmm, yes you are right ofcourse, it actually scales lousy.

> Fortunately, are correct that looking at debugging hooks is the
> way to go for what I _think_ you want. (I think you want an interactive
> display that stops after each "line" of code, right?) Whether or
> not that's done via debuggermodule, I'm not sure, but if you explore
> the docs a little and still have questions, I'm sure others here can
> answer them.

I will have a good look at them, they seem to be much more what I want. I
will probalby post a few questions here :-)

Thanks a lot for your effort!

Tim





From martin at v.loewis.de  Fri Apr 18 06:04:21 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 18 Apr 2003 12:04:21 +0200
Subject: question on buildinh /demo/embed
References: <mailman.1050657062.4933.python-list@python.org>
Message-ID: <m3znmo86l6.fsf@mira.informatik.hu-berlin.de>

"Bjorn Pettersen" <BPettersen at NAREX.com> writes:

> .. which Martin knows (but Jing might not) would also include the
> compiler the OS and potentially the hardware (is there any OSen left
> that run on two hw configs?)

Certainly: Solaris, Linux, OpenBSD, and, recently (again) WinXP/2k3.

> ps: yes, my first job was in tech support, so we'll need your
> registration number too <wink>.

Share the pain, brother :-)

Martin



From DennisR at dair.com  Fri Apr 11 09:41:25 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Fri, 11 Apr 2003 13:41:25 GMT
Subject: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com>
Message-ID: <9Czla.46$yn7.5370091@newssvr13.news.prodigy.com>

> Is there a way to deploy the Python "runtime" with your Python
> scripts?  ... Windows ... smaller than 7 MB download
Check out my 800 KB download at http://www.spamai.com/eval.php  By
installing it, you will see the directory structure and files included.  The
key statements to have Python use the installed structure are:

    import sys
    sys.path = [".", ".\\bin", ".\\bin\\libstd", ".\\bin\\libxtra"]

Look at the /bin directory and its subdirectories to see how these
statements map to files distributed.

I do not have a GUI package.  I include an custom http server in above 800
KB and use dynamic web pages for  user interface.  Including a GUI package
will surely push the download byte count higher.

As far as tools to decide what to include, Python itself will write the .pyc
files it needs.  You simply include those files in what you distribute.  I
use WinRAR (http://www.rarlabs.com/ - $29) for self-extracting compression
and TinySetup (http://www.tinysetup.com - no cost).

> In other words, it would be nice
> to be able to simply unzip the files needed to run my scripts without
> the user having to download and install the full Python windows setup,
> and then download and install my Python scripts seperately.
Too error prone, IMO.  At 800 KB combined, it is preferable to simply
combine them.  In the rare case where someone already has Python installed,
I still want to install an embedded version so that I know exactly what the
Python operating environment is.
--

Dennis Reinhardt

DennisR at dair.com
http://www.spamai.com





From imbosol-1050195048 at aerojockey.com  Sat Apr 12 21:05:11 2003
From: imbosol-1050195048 at aerojockey.com (Carl Banks)
Date: Sun, 13 Apr 2003 01:05:11 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com>
Message-ID: <bJ2ma.5447$883.4511@nwrdny02.gnilink.net>

Roy Smith wrote:
> As another example, I can guess that Python dictionaries work like STL 
> maps on the inside, but that would be just a guess.

And, it turns out, a poor one.


>  Is inserting an 
> element logN?  

Constant.

> Is keys() linear?

Yes, but unlike STL map, the keys aren't sorted.  If you want sorted
keys, you'd have to also perform an O(N log N) sort.

> Is del on an element logN?

Constant.

>  Is 
> has_key() logN?

Constant.


> Those would be my guesses.  But they're just guesses.

STL map is a binary tree.  Python dict's are hash tables.  I think
getting a sorted list of keys is the only order-of-magnitude advantage
a binary tree has over a hash.


-- 
CARL BANKS



From tjreedy at udel.edu  Fri Apr 11 13:35:09 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Fri, 11 Apr 2003 13:35:09 -0400
Subject: Confusion: assignments create object references.
References: <mailman.1050080058.23175.python-list@python.org>
Message-ID: <VLicnQuBj-mmZgujXTWcrg@comcast.com>

"Havens, Peter" <Peter.Havens at Level3.com> wrote in message
news:mailman.1050080058.23175.python-list at python.org...
>I'm a silly rabbit. I hand copied the test incorrectly.

Many of us have learned to always cut&paste -- carefully.

The real test was:

>>> L = 1
>>> M = L

This binds M to an (immutable) int object with value 0

>>> M = 0

The (re)binds M to an (immutable) int object with value 1.  This has
nothing to do with L.

>>> L
1

>Those are the real results (I double checked), but now I'm really
confused. Why does the >former test (the one I lied about) above give
expected results, and the later doesn't?

Because your expectation is still wrong.  Modifying a (mutable) object
(by rebinding one of its slots) is different from rebinding a name to
another object.

Terry J. Reedy





From max at alcyone.com  Wed Apr  2 18:43:59 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 02 Apr 2003 15:43:59 -0800
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com> <b6esse$pge$1@heckle.cc.uregina.ca> <pan.2003.04.02.15.14.54.788223@mn.rr.com> <slrnb8mgum.f98.philh@cabalamat.uklinux.net>
Message-ID: <3E8B75BF.62C6D33C@alcyone.com>

phil hunt wrote:

> The Romans generally preferred not to use the subtractive principle,
> hence: XXXVIIII

According to _Universal History of Numbers_, the subtractive principle
was a relatively recent invention; throughout most of the direct use of
Roman numerals, it wasn't used.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Said it? yep / Regret it? nope
\__/ Ice Cube
    Blackgirl International / http://www.blackgirl.org/
 The Internet resource for black women.



From joshway_w/o_spam at myway.com  Fri Apr 25 14:47:26 2003
From: joshway_w/o_spam at myway.com (Joshua Marshall)
Date: Fri, 25 Apr 2003 18:47:26 +0000 (UTC)
Subject: pythong referencing/aliasing
References: <ylcqa.9116$h%2.963225@read1.cgocable.net> <b8bpf5$hmj$1@ginger.mathworks.com> <mailman.1051291480.8457.python-list@python.org>
Message-ID: <b8bvru$1tk$1@ginger.mathworks.com>

Gerhard H?ring <gh at ghaering.de> wrote:
> Joshua Marshall wrote:
...
>> 
>> You're basically asking two questions, and both answers are "no".
>> There is no way to associate two variables the same address, and there
>> is no REF/address-of operator in Python.

> Such an operator isn't necessary because creation of references is the 
> default behaviour in Python.

<snip>

Oh I agree it's not necessary; it's not even necessary in C.  You can
always leave your data on the heap--malloc a chunk of memory and pass
that address around instead of taking the address of a variable.



From ckea25d02 at sneakemail.com  Sun Apr  6 22:33:54 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 14:33:54 +1200
Subject: What are the differences between SWIG and Boost.Python?
In-Reply-To: <XSXja.346441$3D1.186570@sccrnsc01>
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org> <XSXja.346441$3D1.186570@sccrnsc01>
Message-ID: <b6qnh6$832vf$1@ID-169208.news.dfncis.de>

Bryan wrote:
> in pyrex, don't you have to redefine all the structures?

Currently, yes, although some people have been experimenting
with tools for semi-automatic translation of C headers. I'll
probably look into this issue myself at some point.

 > i haven't dived into pyrex all that much and i know it's
 > an incredible project, but redefining this header stuff
 > just keeps me away from it.

You need to weigh up the effort needed to rewrite the declarations
in Pyrex against the effort needed to write all the C code for
conversions, refcounting, error handling, etc. (and get it right!)
by hand.

I'd encourage you to at least try using Pyrex for a small
project, so you'll have some experience from which to decide
where the balance lies for you.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From fredrik at pythonware.com  Fri Apr 25 12:24:51 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 25 Apr 2003 18:24:51 +0200
Subject: Date validation
References: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au> <b8ap7f$4go$1@main.gmane.org> <20030425080658.GG15695@frobozz.local>
Message-ID: <b8bnf3$i8m$1@main.gmane.org>

Andrew Bennetts wrote:

> > > I mean I *could* write this myself, but I had to ask the computer how
> > > many days in April - someone with a better grasp of the date must have
> > > written this stuff before :)
> >
> > don't you have any knuckles on your left hand?
>
> And which knuckles do you use to remember which years are leap years?  ;)

in what part of the world does the number of days in April
vary depending on the year?

> > if you want your program to check how many days there are in
> > a given month, use the calendar module:
>
> He does.  In his words: "how can I check whether or not such strings
> represent a valid date?"
>
> His problem is verifying that the entire date is correct.  By the time you
> factor in the varying days of the months and leap years, I can understand
> wanting to use a prewritten and tested module rather than an ad-hoc solution
> for a problem that many people must have encountered before.

"calendar" is a prewritten and tested module, and has been part
of Python's standard library since around 1066 AD.

by using calender, he can add 3-4 lines of obviously correct code
to his program, and move on to the next problem...

in contrast, you're suggesting that he

1) update to a newer Python version, and end up with a program
that may or may not work if run on an older version, or

2) backport changes to a newer Python version to his own Python
install, and end up with a nonstandard Python environment and a
program that may or may not work if run on an older version, or

3) add 550 lines of recently written, rather complex, and yet to be
widely used code to his application, and end up with a program that
may or may not work...

but I can of course not stop him from doing things in the hardest
possible way; I'm just pointing out that there's another way to solve
his problem using tools that's already installed on his machine... if
you have a problem with that, that's your problem.

(it could be worse, of course: I'm still waiting for the let's-use-a-
metaclass crowd to discover this thread...)

</F>







From oren-py-l at hishome.net  Thu Apr 24 11:54:13 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Thu, 24 Apr 2003 11:54:13 -0400
Subject: What means exactly "Memory error"?
In-Reply-To: <200304241836.38456.b.maryniuk@forbis.lt>
References: <200304241406.20833.b.maryniuk@forbis.lt> <16040.122.803504.236859@montanaro.dyndns.org> <200304241836.38456.b.maryniuk@forbis.lt>
Message-ID: <20030424155413.GA88893@hishome.net>

On Thu, Apr 24, 2003 at 06:36:38PM +0200, Bo M. Maryniuck wrote:
> On Thursday 24 April 2003 17:19, Skip Montanaro wrote:
> >     Bo> Somebody knows what "Memory error" might appear and how to avoid
> > (if Bo> possible) it?
> > It means a memory allocation using malloc or realloc failed.  In short,
> > virtual memory was exhausted.
> 
> Mmm... OK, but what else I can to do? Yes, Python process grows up to ~1.2G in 
> the memory. The main code where it crashes it's like:
> 
> parsedBranches = {}
> for onePart in aLotOfParts:
> 	parsedBranches[generateId()] = HtmlLib.FromHtml(onePart)
> 
> ...and after 600-700 parts HtmlLib (4DOM) just crashes.
> 
> Any ideas?

Try to use a streaming XML API like SAX instead of DOM. For the splitting
operation you describe there shouldn't be any need to store the entire file
in memory. You may not even need and XML parser at all - looking for the 
HR markers could probably be done by a few regular expressions without full 
parsing of the XML stream.

    Oren




From tjreedy at udel.edu  Tue Apr  1 19:57:16 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 1 Apr 2003 19:57:16 -0500
Subject: I need help
References: <31ef0f43.0304011338.7ae3d45b@posting.google.com>
Message-ID: <exqdnZ43INeIqRejXTWc3g@comcast.com>

"Jewlz408" <Jewlz408 at aol.com> wrote in message
news:31ef0f43.0304011338.7ae3d45b at posting.google.com...
> I need a lot of help
> What do you know about Arrays
> How many storage Locations will be allocated for each of the given
arrays?
> 1) Dim A(1 to 10)
> 2) Dim B(-3 to 5)
> 3) Dim C$( 12)
> 4) Dim D( 1 to 4, -2 to 2)
> 5) Dim E(4,6)

So little that you should not worry about it.

> And What does the following programs do with its output?

Nothing, it does not have any.

> Dim Table( 1 to 5, 1 to 3)
> For I = 1 to 5
> For J = 1 to 3
> Let Table(I,J) = 1
> Next J
> Next I
> End
>
> What does the program do?

Initialize a 2d array to 1's.

You would be much better off learning Python than Basic.  Visit
www.python.org and you can find a bieginner's tutorial.

TJR





From iamshady at rediffmail.com  Wed Apr  9 02:45:42 2003
From: iamshady at rediffmail.com (Jim Shady)
Date: 8 Apr 2003 23:45:42 -0700
Subject: String splitting question
Message-ID: <9afd3f36.0304082245.33670341@posting.google.com>

Hello,

I am a beginner at Python and I just joined this group.

I have a string:

abcd/df/a/iiwk/abcdefghijkl/b/c

I need to get the longest string between the /s of the string. For
example, longest_str() for the above line should return
'abcdefghijkl'. How do I go about doing this?

Thanks in advance.

Regards,
Jim



From guy at obstruction-no-spam.com  Tue Apr 22 13:22:39 2003
From: guy at obstruction-no-spam.com (Guy Middleton)
Date: Tue, 22 Apr 2003 17:22:39 GMT
Subject: another python style question -- copying a data file
Message-ID: <zTepa.127177$Vzu.30273@news02.bloor.is.net.cable.rogers.com>

Ok, here's another style question.

I want to copy a data file, but if the file doesn't already exist, I don't
care.  I do care about any other errors, such as wrong permissions.

I have the following code, is this a typical Python way to do this?

    try:
        os.stat(DATAFILE)
        os.system("cp %s %s.new" % (DATAFILE, DATAFILE))
    except OSError, arg:
        # ENOENT is ok, we didn't do the copy
        # anything else is a problem
        if arg.errno != errno.ENOENT:
            raise




From alloydflanagan at attbi.com  Fri Apr 25 17:00:47 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 25 Apr 2003 14:00:47 -0700
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>	<Kq9qa.9083$h%2.957726@read1.cgocable.net>	<mailman.1051273778.5621.python-list@python.org> <fzo6emyy.fsf@python.net> <mailman.1051277805.18035.python-list@python.org>
Message-ID: <a8b7f07a.0304251300.e9d10d7@posting.google.com>

Gerhard Haering <gh at ghaering.de> wrote in message news:<mailman.1051277805.18035.python-list at python.org>...
<big snip>
> 
> C:\>python
> Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
> Type "help", "copyright", "credits" or "license" for more information.
>  >>> from wxPython.wx import *
>  >>> f = wxFrame(NULL, 1, "foo")
>  >>> f.Show()
> 
> This will segfault Python. IIRC, this was what I did in my very first 
> exploration of wxPython.
> 
> I believe the reason of this crash is that the wxWindows main loop isn't 
> initialized, yet. What I've done may be not very clever, but in my 
> opinion, a Python app should *never* crash. The whole point of me using 
> a highly dynamic language like Python is that it will give me exceptions.

That's a good point.  You can avoid a crash by doing 
app = wxPySimpleApp()
before creating the wxFrame.  But then your frame won't do anything
until you execute
app.MainLoop()
at which point you can't get back to the interpreter until the
application has exited.  Maybe we could start the loop in a separate
thread...



From ajsiegel at optonline.net  Tue Apr 22 09:40:13 2003
From: ajsiegel at optonline.net (Arthur)
Date: Tue, 22 Apr 2003 09:40:13 -0400
Subject: instance creatiion as an "event"
Message-ID: <000801c308d4$b4381890$0c02a8c0@Arts>

Alex:

>I don't understand this. When you call:

>p = TestPoint(args)
>p.redraw()

>it's EXACTLY AS IF TestPoint.__init__'s very last statement was

>self.redraw()

Standalone, yes.  But, unless I'm missing something, the effects in an
inheritance chain is different. If I inherit from TestPoint, and need to
call TestPoint.__init__ in the Point_Child.__init__ to pick up initilization
routines common to the chain, I will now unavoidably be calling TestPoint's
redraw method before I call Point_Child's.  When the main reason I am
inheriting from TestPoint is to override its redraw method, and in fact to
avoid calling it.  If TestPoint has attributes that Point_Child does not
that are referenced in its redraw method, I will have an attribute error on
the calling of TestPoint's redraw form Point_Child.

This is the crux of the problem as I see it, and I am prepared to be
embarrassed if I am misinterpreting the behavior under these circumsntances

Perhaps a "try:.... except AttributeError: pass" for all redraw methods
would be a straightforward solution to this issue. But does not seem fully
elegant.  There are cases where the attributes are in fact the same, but the
routines very different - a Point determined as the circumcenter of a
circle, versus a Point determined as an incenter of a circle.  True, these
classes would more likely be siblings, than parent/child.  So perhaps I am
suffereing from trying to be overly theorectical and pure.  On the other
hand, I am hoping for a design that is "naively" extensible.  Someone coming
in to extend and doing so by inheriting from an existing class would,
ideally, not need to be in touch with concerns at this level.

Next step is to try to digest the implications for my "real world" in your
suggestion as to working with the condition

if hasattr(sys, 'ps1')

Thanks .

Art








From teemu.luojola at pbezone.net  Sat Apr  5 08:36:26 2003
From: teemu.luojola at pbezone.net (Teemu Luojola)
Date: Sat, 05 Apr 2003 16:36:26 +0300
Subject: Input from a file as a command line argument
Message-ID: <3E8EDBDA.7020701@pbezone.net>

I have read through Python documentation and tried several searches, but 
I can't find a simple answer to this question.

If I make an executable script in linux environment, how can I pass 
input to that script from a file on command line? Like the following:

$ script.py input.txt

or perhaps

$ script.py < input.txt

I make now a comparison to a Perl script. In Perl I would write the 
following to handle the input (line by line) from a file:

while (<>) {
	## (do something to $_)
}

and I can execute the script simply with

$ script.perl < input.txt

and it would process each line in input.txt.

How would a similar Python code look like?

Teemu Luojola




From viszneki at telerama.com  Wed Apr  9 15:15:46 2003
From: viszneki at telerama.com (viszneki at telerama.com)
Date: Wed,  9 Apr 2003 15:15:46 -0400
Subject: atexit functionality
In-Reply-To: <10074955979500@aluminium.rcp.co.uk>
References: <mailman.1049821814.20582.python-list@python.org> <10074955979500@aluminium.rcp.co.uk>
Message-ID: <1049915746.3e947162ca34b@webmail.telerama.com>

Quoting Duncan Booth <duncan at rcp.co.uk>:

> In what way is asking for atexit.register to throw an IndexError exception 
> every time it is called NOT introducing a negative side effect?

He's right you know, below is quite an index error resulting from deleting 
items from atexit._exithandlers. (sarcasm)

root at crackbox:~# python
Python 2.2.1 (#1, Apr 11 2002, 15:27:15)
[GCC 2.95.3 20010315 (release)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import atexit
>>> def test(s):
...     print s
...
>>> for each in ['one','two','three','four','five']:
...     atexit.register(test,each)
...
>>> atexit._exithandlers
[(<function test at 0x8144ecc>, ('one',), {}), (<function test at 0x8144ecc>, 
('two',), {}), (<function test at 0x8144ecc>, ('three',), {}), (<function test 
at 0x8144ecc>, ('four',), {}), (<function test at 0x8144ecc>, ('five',), {})]
>>> del atexit._exithandlers[1:4]
>>> atexit._exithandlers
[(<function test at 0x8144ecc>, ('one',), {}), (<function test at 0x8144ecc>, 
('five',), {})]
>>> # i'm exiting now
...
five
one
root at crackbox:~# WOW! No index errors! What a suprise!
bash: WOW!: command not found

Thanks for playing Duncan. You know I was in good spirits and happy to 
contribute something until you and ol Stevey Holden decided to lecture me on 
Internet Etiquette, and tell me how much of a jerk I was to not have time to 
sign up for every mailing list I'd like to contribute something to. Gee I'm 
such a jerk.

If anyone has anything positive to say about my idea, please say so. In my last 
email to this mailing list, I posted some hypothetical code to do the job:

return _exithandlers[len(_exithandlers)]

Shortly after having sent it, I realized it wouldn't work because of the way 
Python does things (note: that does not mean index error, or ANY OTHER KIND OF 
ERROR for that matter.) Assigning the returned value, storing it, and then 
deleting it later, does not result in deletion of an item from _exithandlers, 
because it is only deleting the symbolic reference. The item in _exithandlers 
remains. I tried this myself, by adding that line of code to the end of 
atexit.register, just to make sure, because I am admittedly no master at 
pointer dynamics in Python (something I wish I knew more about.)

Thank you everybody for your time,
Donny Viszneki.





From ckea25d02 at sneakemail.com  Thu Apr  3 20:07:30 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 04 Apr 2003 13:07:30 +1200
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <slrnb8orsl.174.do-not-spam-ben.hutchings@tin.bwsint.com>
References: <mailman.1049296805.12278.python-list@python.org> <b6fvj9$518n3$1@ID-169208.news.dfncis.de> <slrnb8orsl.174.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <b6ilcb$5q3lq$1@ID-169208.news.dfncis.de>

Ben Hutchings wrote:
> But since the use of subtraction was apparently a late addition
> to the system, we should allow the purists to do without it:

This raises the question of whether subtractive and
non-subtractive forms should be allowed together in
the same literal, or even different literals of the
same module. If not, it would be prudent to have
two different feature domains, e.g.

   from __ancient_history__ import __roman_numerals__

for non-subtractive forms, and

   from __middle_ages__ import __roman_numerals__

for subtractive forms.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From tungwaiyip at yahoo.com  Tue Apr 22 20:17:36 2003
From: tungwaiyip at yahoo.com (Tung Wai Yip)
Date: Tue, 22 Apr 2003 17:17:36 -0700
Subject: SGMLParser questions
Message-ID: <igmbavg5ughls6ksiden2rgnael1s0ub8q@4ax.com>

I'm using SGMLParser to parse HTML. It works great in handling the lax
HTML format. I have a few questions however:

1. unknown_starttag(tag, attributes) - why is 'attributes' a list of
(name, value) pairs instead of a dictionary? I incline to use
something like

  attributes.get['src']

but this doesn't work for the list. Is there some short cut to
retrieve value in a list of (name, value) pairs?

2. Does it supports namespace?

3. handle_starttag() turn tags into lower case. This work perfectly
for HTML. But can it works case sensitively?

Wai Yip Tung




From jkraska at san.rr.com  Fri Apr 18 21:20:32 2003
From: jkraska at san.rr.com (Courageous)
Date: Sat, 19 Apr 2003 01:20:32 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <b7pchr$ig0$1@main.gmane.org> <mailman.1050690871.4874.python-list@python.org>
Message-ID: <0191avo4391fumu3rg6k0ok6lrbde52hro@4ax.com>

>AFAIK, generator functions have turned out to be quite difficult in Java
>too, but that didn't seem to stop their acceptance.

I'll chime in here.

I am emphatically _against_ targeting subsets of features that cater
to a virtual machine. The Java guys can get it together, or fall into
a hell of conformity, for all I care. Same with .NET.

C//




From bokr at oz.net  Wed Apr 23 23:42:27 2003
From: bokr at oz.net (Bengt Richter)
Date: 24 Apr 2003 03:42:27 GMT
Subject: Creating dynamically named lists/dictionaries
References: <ydFpa.15$IO5.10324@newsfep2-gui.server.ntli.net>
Message-ID: <b87mf3$shh$0@216.39.172.122>

On Thu, 24 Apr 2003 00:22:18 +0100, "Rogue 9" <rogue9 at ntlworld.com> wrote:

>Hi all,
>	I'm new to both programming and Python in particular and have run in to a
>snag while trying to write my own program.I have looked through books and
>the web but haven't as yet found an answer so I though I might try a post
>here to see if anyone could point me in the right direction.
>	Okay my program is for analysing a Lottery Draw results text file I have
>pre-formatted so that each line contains 7 comma separated values (Draw
>number and six values representing the numbers drawn in numerically
>ascending order.
>	I've sussed out how to open the file read in the lines and I have split
>the strings first by the '\n' newline symbol and then I have split it
>again by ','.I've also turned the values back into integers as they were
>strings.Now what I've ended up with is a list containing one draw result
>for each week (768 at the last count)as individual items in that list e.g
>list =[[767,11,22,33,44,45,46]......[1,20,33,41,45,46,47]]
>	What I really want is a list or dictionary for each individual weeks
>result but I can't seem to work out how to dynamically create a list or
>dict with a unique name for each week.The results file will grow and I
>don't fancy creating 768+ lines of code just to instantiate a list or
>dict.At first I tried looping a variable equal to the length of the file
>as in:
>	x=len(filelength)
>	'draw'+str(x)=[drawresult]
>I thought this would dynamically make a whole bunch of lists such as
>draw1,draw2 etc. but all I got was errors.
>	I hope I have explained my 'problem' sufficiently for someone to help or
>to at least point me in the right direction to work towards a solution.
Maybe this will give you some ideas...

====< lottery.txt >=================
1,11,12,13,14,15,16
2,21,22,23,24,25,26
3,31,32,33,34,35,36
4, -- intentionally bad line --
5,xx,52,53,54,55,56 -- also bad
767,11,22,33,44,45,46
====================================
====< lottery.py >==================
def test(filepath):
    drawdict = {}
    for drawline in file(filepath):
        nums = map(str.strip, drawline.split(','))
        if nums == ['']: continue # skip blank lines silently
        try:
            nums = map(int, nums)
            if len(nums) != 7: raise ValueError
        except ValueError:
            print 'Bad data line: %s' % `drawline`
        else:
            drawdict['draw%s' % nums[0]] = nums[1:]
    return drawdict

if __name__ == '__main__':
    import sys
    if len(sys.argv)>1:
        dd = test(sys.argv[1])
        for k,v in dd.items():
            print '%10s: %s' % (k, v)
        print 'Just press Enter to exit ;-)'
        while 1:
            key = raw_input('Enter draw id in form "drawN": ')
            if not key: break
            val = dd.get(key, '(None stored in dict)')
            print '    Result for drawing "%s" is %s' % (key, val)
    else:
        print 'Usage: [python] lottery.py filepath'
 ====================================

 [20:37] C:\pywk\clp>python lottery.py
 Usage: [python] lottery.py filepath

 [20:37] C:\pywk\clp>python lottery.py lottery.txt
 Bad data line: '4, -- intentionally bad line --\n'
 Bad data line: '5,xx,52,53,54,55,56 -- also bad\n'
    draw767: [11, 22, 33, 44, 45, 46]
      draw3: [31, 32, 33, 34, 35, 36]
      draw2: [21, 22, 23, 24, 25, 26]
      draw1: [11, 12, 13, 14, 15, 16]
 Just press Enter to exit ;-)
 Enter draw id in form "drawN": xxx
     Result for drawing "xxx" is (None stored in dict)
 Enter draw id in form "drawN": draw2
     Result for drawing "draw2" is [21, 22, 23, 24, 25, 26]
 Enter draw id in form "drawN": draw767
     Result for drawing "draw767" is [11, 22, 33, 44, 45, 46]
 Enter draw id in form "drawN":

 [20:37] C:\pywk\clp>python
 Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
 Type "help", "copyright", "credits" or "license" for more information.
 >>> import lottery
 >>> d = lottery.test('lottery.txt')
 Bad data line: '4, -- intentionally bad line --\n'
 Bad data line: '5,xx,52,53,54,55,56 -- also bad\n'
 >>> d
 {'draw767': [11, 22, 33, 44, 45, 46], 'draw3': [31, 32, 33, 34, 35, 36], 'draw2': [21, 22, 23, 2
 4, 25, 26], 'draw1': [11, 12, 13, 14, 15, 16]}
 >>> d.keys()
 ['draw767', 'draw3', 'draw2', 'draw1']

Hm, that bad data output might not be the best way to handle the errors. And you might
want to move stuff into test to do more of interest there, or write another routine that
uses test (and give it a better name ;-) Whatever, have fun.

Regards,
Bengt Richter



From tim_one at email.msn.com  Sun Apr 20 12:51:41 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sun, 20 Apr 2003 12:51:41 -0400
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <b7ufg4$o1d$1@news.hccnet.nl>
Message-ID: <LNBBLJKPBEHFEDALKOLCKEIFEHAB.tim_one@email.msn.com>

[A.M. Kuchling, points to
 http://mail.python.org/pipermail/python-dev/1999-July/000467.html
]

[Anton Vredegoor]
> Thanks. It seems Tim is making a heroic effort to explain the new in
> terms of the old. This necessarily must fail, but what other option is
> there instead of rebuilding all terminology out of gotos?

You're missing a decade of history <wink>:  as that post says at the start,

    Here's my biased short course.

I argued in favor of adding generators to Python starting in about 1991.
Every time I tried, the effort got hijacked by people going on about
continuations.  The post's primary point was that generators *can* be
explained "in terms of the old", and easily.  I think it succeeded at that.
Coroutines are harder, but can still be seen as a natural extension.
Continuations require a different mindset entirely, and I believe that post
was instrumental in (finally) getting generators taken seriously as a
feature that didn't require buying into continuations first.

Continuations are a different *foundation* for implementing control-flow
features, more than they're a feature in their own right.  While they're an
elegant foundation, they're not the foundation Guido (or 99.9+% of the
world's other language designers) chose at the start, and it remains
unlikely that any Python Guido works on will adopt them.





From max at alcyone.com  Wed Apr 30 18:20:14 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 30 Apr 2003 15:20:14 -0700
Subject: Self modifying systems question
References: <mailman.1051708787.23125.python-list@python.org>
Message-ID: <3EB04C1E.4CE0092B@alcyone.com>

Sten Kvamme wrote:

> Thanks to all of you who answered my question.
> Is it also possible to save the changes made to an instance in a
> persistant way?

I think it depends on what you mean by persistent.  If you want to have
objects that stick around from run to run, your best bet is probably the
pickle/cPickle module.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The conviction of wisdom is the plague of man.
\__/ Montaigne
    Esperanto reference / http://www.alcyone.com/max/lang/esperanto/
 An Esperanto reference for English speakers.



From xantor at linux.be  Sat Apr 19 04:22:35 2003
From: xantor at linux.be (Michael Anckaert)
Date: 19 Apr 2003 10:22:35 +0200
Subject: error in string module documentation?
Message-ID: <1050740555.624.4.camel@carpathia>

Hello all,
In the string module documentation I read the following:


lstrip(s[, chars])
        Return a copy of the string with leading characters removed. If
        chars is omitted or None, whitespace characters are removed. If
        given and not None, chars must be a string; the characters in
        the string will be stripped from the beginning of the string
        this method is called on.
        
But I get the following error from my program:

  File "./game.py", line 35, in run_game
    item = string.lstrip(cmd, "look")
TypeError: lstrip() takes exactly 1 argument (2 given)

Is there an error in the documentation or is it my error?

-- 
Greetings,
Michael Anckaert, aka The XanTor
OpenPGP key: F7A6C3AB
Fingerprint: A329 43FC 3953 A944 5DDC  2E05 8E5D AD60 F7A6 C3AB
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 196 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/python-list/attachments/20030419/298c6b3f/attachment.sig>

From spam at mattbergin.co.uk  Mon Apr 14 13:53:21 2003
From: spam at mattbergin.co.uk (Matt Bergin)
Date: Mon, 14 Apr 2003 18:53:21 +0100
Subject: Reading/writing files
In-Reply-To: <pan.2003.04.14.17.32.01.209427@soul.net>
References: <pan.2003.04.14.17.32.01.209427@soul.net>
Message-ID: <b7esij$p7f$1@wisteria.csv.warwick.ac.uk>

lost wrote:
...
> I need something like this
> 
> for line in a file
> 	print line
> print done

for line in file('afile.txt', 'rb')
     print line
print 'done'

> and for writing, I've tried doing file.write(blah) (blah is a list), but
> it wont write.
> I need to write lists into a file, but i cant.  are there other methods
> for printing to a file, other than file.write?

You need to convert whatever you want to a file to a string first. Take 
a look at the pickle module.

 From the Python tutorial:
----------------
If you have an object x, and a file object f that's been opened for 
writing, the simplest way to pickle the object takes only one line of code:

pickle.dump(x, f)

To unpickle the object again, if f is a file object which has been 
opened for reading:

x = pickle.load(f)

(There are other variants of this, used when pickling many objects or 
when you don't want to write the pickled data to a file; consult the 
complete documentation for pickle in the Library Reference.)
----------------

Or you could do

file.write(':'.join(list))

To join the list with colons (if it is a list of strings), and to read 
it back you could do

list = line.split(':')

> Also, do i need to take into account the \n at the end of a line?

line = line.rstrip() will remove the \n.

-- 
Matt Bergin




From jason at xxspamxx.methodize.org  Sat Apr  5 18:57:15 2003
From: jason at xxspamxx.methodize.org (Jason Brome)
Date: Sat, 05 Apr 2003 23:57:15 GMT
Subject: [Survey] Yet Another Open Source RSS News Reader?
References: <khudnWYcCKiwghKjXTWcpg@comcast.com>
Message-ID: <v3Kja.1404$py2.100@twister.nyc.rr.com>

Gary,

Just wanted to clear up a couple of points -

> 3.  nntp/rss (http://www.sourceforge.net/projects/nntprss).  An
> interesting approach that parses RSS feeds and makes them available to any
> NNTP (Usenet) news reader application.  Uses Mark Pilgrim's Ultra-liberal
> RSS Parser.

While inspired by Mark Pilgrim's discussions, nntp//rss's parse-at-all cost
implementation is quite different to Mark's Python parser.

> Does not seem to support syndication lists.

As of version 0.3, nntp//rss supports the import of OPML-based subscription
lists.  Future releases will have additional subscription list features,
along with tighter integration with public RSS feed directories.

Good luck with your project - at the end of the day, the more choice a user
has, the better.  I've spoken to a number of people about their requirements
for a newsreader / aggregator, and their requirements do vary quite
considerably.  As you point out, it's certainly not a one-fits-all scenario.
One point I would make - if you can, make sure you support If-Modified-Since
/ ETag within your HTTP client.  As these applications become more popular,
we, as developers, should ensure that we build products that make
considerate use of RSS publishers' bandwidth.

Regards,

Jason

--
Jason Brome
nntp//rss Project
jason at methodize.org





From sholden at holdenweb.com  Mon Apr 14 10:21:32 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Mon, 14 Apr 2003 14:21:32 GMT
Subject: Newbie: Example of GUI front-end for SQL
References: <pan.2003.04.12.02.09.40.159546@nospam.net>
Message-ID: <Mtzma.83640$It5.49749@news2.central.cox.net>

"c42" <nospam at nospam.net> wrote in message
news:pan.2003.04.12.02.09.40.159546 at nospam.net...
> I'm researching a new project that will include a GUI front-end
> (several screens) to an SQL server (probably Postgre or MySQL).
>
> Can someone recommend an open source program that I can review that will
> give me an idea if this is possible with python?
>
> It is also important that the program be able to run on Windows and Linux.

The PythonCard project has a simple database GUI as one of its sample
applications, so you might want to take a look at that. PythonCard is based
on wxPython which is in turn based on the highly-portable wxWindows, so your
portability goals need not be compromised. See

    http://pythoncard.sourceforge.net/

Or you can just take my workd for it, it's perfectly possible for almost any
modern relational database :-)

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From donn at drizzle.com  Fri Apr 18 00:29:28 2003
From: donn at drizzle.com (Donn Cave)
Date: Fri, 18 Apr 2003 04:29:28 -0000
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <mailman.1050634170.24420.python-list@python.org>
Message-ID: <1050640161.271511@yasure>

Quoth Steven Taschuk <staschuk at telusplanet.net>:
| Quoth Donn Cave:
| > Quoth Steven Taschuk <staschuk at telusplanet.net>:
|   [...]
|>| Second, I have some sense that you feel these terms are not merely
|>| unclear but misleading or wrong.  Is this true?  If so, do you
|>| have alternative fuzzy terms in mind (the "data" vs the "object"),
|>| or would you prefer that the documentation avoid fuzzy terms
|>| entirely, and only give examples?
|> 
|> The data vs. object distinction is the best I can do so far.  It
|> isn't all that fuzzy, either.  It's vague, in that it applies to
|> a very high level of generality, but it does not seem ambiguous
|> to me.
|
| Calling for __str__ to return the data of the object might imply,
| undesirably imho, that, for example, str(file('foo')) should
| return a string containing the contents of the file.  Likewise
| for, e.g., database connection objects.
|
| So it seems a little fuzzy to me, in that it needs qualification
| to avoid such implications.

That's interesting, that example sort of occurred to me as one
that demonstrates the principle.  I wasn't thinking of "file",
actually, rather some hypothetical object that represents a file
("file" represents an I/O stream.)  But we're definitely on the
same page here.  Your objection is probably reasonable, though
arguable, but the thing is, we're on ground here that we can
reason about.  It may be underspecified, but it isn't the kind
of thing I call fuzzy.  I agree that it's worth a brief sentence
or two, but at some point one has to leave it to the programmer's
judgement what's actually going to be desirable.

| Reading this thread has made me wonder whether current practice is
| uniform.  It would not be desirable imho for the language
| reference to say, "people do different things with __repr__,
| including <list of three things>", and not endow one of them with
| some normative weight.
|
| ... How about this?  (Draft.)
|
| __repr__(self)
|
|     Called by the repr() built-in function and by string conversions
|     (reverse quotes) to compute a string representation of the object.
|     The return value must be a string object.
|
|     The primary audience of this representation is programmers seeking
|     detailed information about the object for debugging or inspection
|     purposes; thus it is important that the representation be
|     information-rich and unambiguous.  This representation should not
|     usually be parsed or manipulated by other code.
|
|     In some cases the best information-rich and unambiguous
|     representation is a Python expression which evaluates to the
|     object, that is, for eval(repr(x)) to return an object which is
|     equal to x (given an appropriate environment).  For example, the
|     built-in numeric, sequence, and mapping types all have such
|     __repr__ implementations.  Where such a representation is
|     infeasible or undesirable, this method should return a string of
|     the form "<...some useful description...>".
|
|     (... examples ...)
|
| __str__(self)
|
|     Called by the str() built-in function and by the print statement
|     to compute a string representation of an object.  The return value
|     must be a string object.
|
|     This differs from __repr__() in that its primary audience is not a
|     programmer inspecting the object during debugging.  What the
|     primary audience is depends on the object: the __str__
|     representation might, for example, be designed for display to a
|     user of the application, or for processing by other code which
|     requires a string.  Thus this representation might be less
|     detailed, less precise, or in some way more convenient than that
|     computed by __repr__.
|
|     (... examples ...)

That's a big improvement over the current documentation.  The __str__
section could use more work.  I really think the last sentence is as
wrong as it is right - sure it might be, or it might be more detailed
etc.  Cf. rfc822.Message.__str__ - complete contents of the header.
In what way is that less detailed or precise, or more convenient?
It is however exactly the ``data''.

Your __repr__ draft works fine for me.

	Donn Cave, donn at drizzle.com



From noway at sorry.com  Wed Apr 30 19:56:15 2003
From: noway at sorry.com (Giovanni Bajo)
Date: Wed, 30 Apr 2003 23:56:15 GMT
Subject: About "Safe Asynchronous Exceptions for Python"
Message-ID: <zoZra.81868$DT4.2490014@twister1.libero.it>

Hello,

I was reading http://www.cs.williams.edu/~freund/papers/02-lwl2.ps. I
understand the problem, but my first dumb question is: when a signal
happens, could the interpreter check if the next statement is a try
instruction (for a try/finally block) and, in that case, delay the execution
of the handler until the try instruction has been executed? In other words,
given the suggested resolution (in the syntactic sugar form), Python could
always consider the statement before a try/finally block as within an
implicit "initially:" block. Wouldn't this solve automatically the resource
acquisitition/usage/release idiom? In fact, to be exception safe one must
already put a try block IMMEDIATLY after the resource acquisition statement.

I can't even see how such a resolution would break existing code, since
signals are asynchronous so no code can be affected if their handling is
postponed a little bit later.

What am I missing?
-- 
Giovanni Bajo






From BPettersen at NAREX.com  Tue Apr 29 20:15:43 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Tue, 29 Apr 2003 18:15:43 -0600
Subject: Question about accessing class-attributes.
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE169@admin56.narex.com>

> From: Alex Martelli [mailto:aleax at aleax.it] 
> 
> Bjorn Pettersen wrote:
> 
> >> From: Alex Martelli [mailto:aleax at aleax.it]
> > [...]
> 
> [special-methods problem snipped]
> 
> Ah yes, I see now!  Yes, functions such as len() rely on slots in
> the type object, e.g. as you've noticed:
> 
[...]
> zone of the code.  Maybe you can post this as a bug in 2.3 beta 1
> on sourgeforge (ideally showing where in the docs it defines the
> semantics that it then doesn't respect) so we can get this looked
> at by the few people who really DO grasp these parts...;-).  There
> is probably some sound implementation reason for the current
> behavior, but if so it should be better documented, I think.

SF 729913. I have not been able to find an exact description of the
lookup rules for new-style classes anywhere.

> > The relevance this has to super, is that you couldn't do a 
> > (contrived) example like:
> > 
> >   for item in super(MyClass, self):
> >       ...
> > 
> > (in general, you can't pass super(MyClass, self) to a 
> > function expecting a MyClass instance, even if MyClass 
> > defines no new behavior...)
> 
> Yes, of course not -- super(MyClass, self) is NOT an instance 
> of MyClass (isn't THAT why you're calling super...?).  

I'm also missing an exact description of what super() returns. It looks
like it's an object that when you call method 'm', invokes the proper
super class' method 'm', and if that 'm' invokes any otherm method,
'm2', the original objects definiton 'm2' is used. So it looks like it
should be substitutable, or at least a first class object, but as you
said: "of course not". Am I missing some documentation, or is it
unreasonable of me to think it should work this way (what way should it
work... exactly)?


> > Perhaps a more realistic exmaple?:
> > 
> >   class Secure(object):
> >     def __init__(self, name):
> >       self.name = name
> >   
> >     def __getattr__(self, attr):
> >       return getSecureMethod(self.name, attr, getUserPrivileges())
> >         
> >   class LogFoo(Secure):
> >     def __init__(self):
> >       Secure.__init__(self, 'SalaryDB')
> >       
> >     def foo(self):
> >       print 'calling foo'
> >       return super(LogFoo, self).foo()
> >       
> >   lf = LogFoo()
> >   print lf.foo()
> 
> Sorry, I don't get it.  How is this related to metaclasses...?

Not terribly obvious was it... :-)  And it's a bad example too. The
thinking was that if Super's metaclass could intercept special methods
through it's __getattr__, the object returned from super() could look
more like a regular object.  That Super.__getattr__ doesn't check for a
__getattr__ in the __mro__ chain is a problem, but a separate one (sorry
for the confustion).

-- bjorn




From martin at v.loewis.de  Tue Apr 22 01:47:08 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 22 Apr 2003 07:47:08 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <erk5avkh423eh8bc13ctvtbjqt9f6kt4kv@4ax.com> <84smscm5df.fsf@boost-consulting.com> <pf58avcm8u8r84183tia6alfo8gctc4fdu@4ax.com> <84el3vptmf.fsf@boost-consulting.com>
Message-ID: <m3k7dndqxv.fsf@mira.informatik.hu-berlin.de>

David Abrahams <dave at boost-consulting.com> writes:

> > One *can* on some systems, although this isn't portable, and not
> > supported on many, and guaranteed on none. Hence in the part you
> > deleted: "although certain patterns of setjmp/longjmp aren't
> > always safe." :)
> 
> References, please?

Hansen, The C++ Answer book, describes a coroutine library
build on top of setjmp/longjmp (in response to an exercise
set by Stroustrup).

Regards,
Martin



From bourassa-m at rmc.ca  Tue Apr  8 15:18:02 2003
From: bourassa-m at rmc.ca (bourassa)
Date: Tue, 08 Apr 2003 15:18:02 -0400
Subject: Pathname problem
Message-ID: <3e9321b5$1@rmcnt03a.rmc.ca>

Hi there.  I am just trying to load files using PIL.  I get the following:

im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
Traceback (most recent call last):
   File "<pyshell#13>", line 1, in ?
     im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
   File "C:\Python22\Lib\site-packages\PIL\Image.py", line 1479, in open
     fp = __builtin__.open(fp, "rb")
IOError: [Errno 2] No such file or directory: 
'C:\\MATLAB6p1\toolbox\\pout.tif'

I have stepped through and if the file in C:\ or C:\MATLAB6p1 it loads 
with no problem.  The minute I try to get it out of "toolbox" (which 
exists) I get the above error.  I'm using  PIL-1.1.4a3.win32 on a Win2K 
PIV.  TIA.

majb




From darin at REMOVEblueyonder.co.uk  Tue Apr 22 14:49:30 2003
From: darin at REMOVEblueyonder.co.uk (Darin Willits)
Date: Tue, 22 Apr 2003 18:49:30 GMT
Subject: Embedding Python in an MFC application
References: <a7gpa.456$OK2.4518100@news-text.cableinet.net>
Message-ID: <_8gpa.460$OK2.4518100@news-text.cableinet.net>

Darin Willits wrote:

> Hi,
> 
> I know that the first response to this will be "Why are you embedding
> when you could just extend?" but please believe that I have my reasons
> and for now this is what I have to do.  However I am something of a
> newbie with this stuff so bear with me.
> 
> Now for the question:  I have created a simple MFC application (with
> AppWizzard) which mearly calls Py_Initialize() and Py_Finalize() in
> the
> OnInit function.  When I build and run this app it starts the python
> interpreter fine, and shuts it down without a noticable problem.
> However when I close the application I get a huge memory leak.  Now
> for
> the wieird part.  If I switch the above application from linking to
> MFC as a shared dll, to statically linking to MFC the memory dump is
> gone.
> Why is this?  This is driving me nuts as I really need to use the MFC
> as a shared dll.
> 
> Please tell me I am missing something simple.  I am learning this
> stuff as I go and this one has me stumped.
> 
> Thanks,
> 
> Darin


Sorry forgot to mention that I am using Python 2.2.2 and Visual Studio 6
with Service Pack 5.

Thanks,

Darin



From nilsgri at idi.ntnu.no  Mon Apr  7 06:05:56 2003
From: nilsgri at idi.ntnu.no (Nils Grimsmo)
Date: Mon, 07 Apr 2003 12:05:56 +0200
Subject: wxPython update problem
Message-ID: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no>

hi!

i have a problem getting wxPython to update the gui while the thread is
running user procedures. i have tried calling wxWindow.Refresh() and
wxWindow.Update() in various orders both on the wxFrame and the wxButton, 
but the gui doesn't update.

i have posted source for a test program below. what i would like it to do,
is to update the button with the new label before control is returned from
run().


from wxPython.wx import *
import time

class MainWindow(wxFrame):
    label = ["foo", "bar"] 
    button = None
    def __init__(self, parent, id, title):
        wxFrame.__init__(self, parent, -4, title, size = (100, 100))
        self.button = wxButton(self, 13, self.label[0])
        EVT_BUTTON(self, 13, self.run)
        self.Show(true)

    def run(self, event):
        for c in range(2):
            for x in [0, 1]:
                self.button.SetLabel(self.label[x])
          	# please update. pretty, please!
		time.sleep(1.0)

app = wxPyApp()
frame = MainWindow(None, -1, "Hello World")
app.MainLoop()


another thing i am wondering about is to disable the button
temporarily. must i use wxEvtHandler::Disconnect?

-- 
nils grimsmo
nilsgri at idi.ntnu.no




From durdn at yahoo.it.oops!.invalid  Thu Apr  3 11:52:45 2003
From: durdn at yahoo.it.oops!.invalid (Nicola Paolucci)
Date: Thu, 03 Apr 2003 16:52:45 GMT
Subject: jython
In-Reply-To: <XoDia.3747$2x2.1694570@dca1-nnrp1.news.algx.net>
References: <XoDia.3747$2x2.1694570@dca1-nnrp1.news.algx.net>
Message-ID: <xFZia.1471565$Wr.54990795@Flipper>

Hi Greg,

Greg Brondo wrote:
> is Jython still unde development?

Personally I'm not involved with Jython development (though I use it in 
some of the projects I work on).
I had the same question few days ago and looked at Sourceforge here:

http://sourceforge.net/project/stats/index.php?report=months&group_id=12867

It seems that there is still recent CVS activity ...

Bye,
	Nick

-- 
#Remove .oops!.invalid to email or feed to Python:
'Tmljb2xhIFBhb2x1Y2NpIDxuaWNrQG5vdGp1c3RjYy5jb20+'.decode('base64')




From nomad*** at ***absamail.co.za  Fri Apr 11 15:46:49 2003
From: nomad*** at ***absamail.co.za (Nomad)
Date: Fri, 11 Apr 2003 21:46:49 +0200
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com> <b766fi$p3b$1@newstree.wise.edt.ericsson.se> <16411664.REEpLcBOk5@lothar.cengija.com>
Message-ID: <amfd9vsugg1hd0cd8qrjdm033opugi5i6e@4ax.com>

On Fri, 11 Apr 2003 13:26:32 +0200, Davor Cengija
<dcengija_remove_ at inet.hr> wrote:

>Frithiof Andreas Jensen wrote:
>
>> 
>> "kkennedy" <titanrebel at comcast.net> wrote in message
>> news:14bd1399.0304091426.1b08e005 at posting.google.com...
>>> I would like it to be smaller than
>>> the 7 MB download available from python.org, and be able to be
>>> installed transparent to the user.
>> 
>> On windows, Biig == Good!!
>> 
>> Users are so used to huge installs & downloads that they expect them and
>> they will be suspicious of anything that takes less than 30 minutes to
>> install - especially if it costs money. The "no-brain-install" is the
>> right idea, thoug.
>
>(really off-topic, but worth of mentioning)
>
>I wrote a simple python program with gui for a friend of mine and his 
>projects at local university. His co-workers didn't want to use it since 
>'it's too small, obviously it doesn't do what we need'. It's about 45k of 
>python source. Then I created self-standing exe with mcmillan's installer. 
>It's now about 7mb in size. Now it looks 'respectable' and 'worth of trying 
>but not for serious work since it doesn't have setup.exe with EULA and 
>other things -- so it isn't a real program'. Now I'm looking for a free 
>install-shield like software.

Have a look at "Inno Setup" from "http://www.innosetup.com".  Very
nice, and looks like an install-shield/wise install to the passing
eye.

>So, it isn't the quality of a software what only counts. Actually, I'm not 
>sure that the quality is the most important thing these days.

Of course not -- it's got to be bloated, but pretty...


-- 
Nomad

Wondering of the vast emptyness of the 'net
in search of something cool.



From dripton at ripton.net  Tue Apr 29 12:30:03 2003
From: dripton at ripton.net (David Ripton)
Date: Tue, 29 Apr 2003 12:30:03 -0400
Subject: The behavior of List.insert
In-Reply-To: <b8m8dn$b03$1@pegasus.csx.cam.ac.uk>
References: <b8m8dn$b03$1@pegasus.csx.cam.ac.uk>
Message-ID: <3EAEA88B.6000803@ripton.net>

York wrote:

> given
> 
> mylist = [1,2,3,4,5], 
> 
> mylist.insert(-1,99) will make mylist to be [99,1,2,3,4,5]
> 
> just wonder why not let it to be [1,2,3,4,99,5]

This is fixed in Python 2.3b1.

-- 
David Ripton    dripton at ripton.net





From marcus at infa.abo.fi  Thu Apr 17 11:54:37 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 17 Apr 2003 17:54:37 +0200
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org> <roy-31F2E6.07562115042003@reader1.panix.com>
Message-ID: <slrnb9tfub.lgn.marcus@infa.abo.fi>

On Tue, 15 Apr 2003 07:56:21 -0400, Roy Smith <roy at panix.com> wrote:
>I certainly agree that mutable strings would be useful in some 
>situations.  For example, in the quadratic run-time program I mentioned 
>in a parallel thread, I could easily see doing:
>
>s = ""
>s.reserve (len (line))
>for word in line.split():
>   blah blah blah
>   s += stuff
>
>and being very happy that I was now getting linear run time with only a 
>trivial change in logic.
>
>The problem is, strings are currently documented as being immutable.  
>Making them mutable would break existing code.  I can, for example, find 
>the length of a string and use that length someplace later in my 
>program, knowing that it hasn't changed.  Not to mention using strings 
>as dictionary keys.

It would be nice if a Python optimizer of sorts could be developed to
create the above code (or, rather, behaviour) given the following
code:

s = ""
for word in line.split():
	s += stuff

I'm not a Python developer, but I _think_ that this problem generally
could be hard/impossible to fix, given the extremely dynamic nature of
Python. So any solution would need to "nail down" some things in
Python, and remove some of its dynamic nature. The question would then
be if this is considered "premature optimization".

I'd love to be told otherwise, though. If a feasible solution exists,
more power to you :)

-- 
Marcus Alanen
maalanen at abo.fi



From christophe.grimault at novagrid.com  Thu Apr  3 02:35:49 2003
From: christophe.grimault at novagrid.com (christophe grimault)
Date: Thu, 03 Apr 2003 09:35:49 +0200
Subject: How to protect python code ?
Message-ID: <3E8BE455.4000904@novagrid.com>

Hello,

Before trying my own solutions, I wonder if someone already has a 
solution to this.
We have been programming a quite large application in python and it 
contains many
original algorithm and other things. The application will be sent to 
customers in a few
month and since it is a commercial app, we need some protection against 
copy (illegal
copies) and we need to make sure that critical parts of the code cannot 
be viewed and studied.

One way is to deliver only .pyc files, but I guess this is quite a poor 
protection, uh ?

An other way is to move some parts to a C API ext module and deliver 
only the .so,
but this is quite time consuming ...

Does someone has a mean to do this, knowing that we don't seek absolute 
protection, but
something that discourages people that will just "try a little"

Thanks in advance




From jkraska at san.rr.com  Sun Apr 20 13:56:00 2003
From: jkraska at san.rr.com (Courageous)
Date: Sun, 20 Apr 2003 17:56:00 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <mailman.1050854561.19910.python-list@python.org>
Message-ID: <g8n5avgooim13l90bv5t9gu6ilg1oar101@4ax.com>

>They also seem to have the property that before you understand them, no
>explanation will improve your understanding; and after you understand
>them you lose the ability to explain it <wink>.

Stacklessness is actually far harder to explain than a continuation.

In environments that I work on where continuations aren't available,
but the same features must exist, the hand-rolled equivalent is a set
of "parked" threads whereby some scheduler is written to selectively
activate them. That's how they work, conceptually. A continuations is
a marker for a "place one can return to whenever one might want". No
big deal, really. Doesn't have to be complicated.

'Course the implementation details may be a bit harder to explain,
depending. :-)

C//




From yaipa at yahoo.com  Fri Apr 11 13:46:35 2003
From: yaipa at yahoo.com (yaipa h.)
Date: 11 Apr 2003 10:46:35 -0700
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <6e07b825.0304110946.50aeb8c7@posting.google.com>

> - 50% of all exercises "to teach programming" are already solved in the 
> standard-library or are a method of a buildin object / are allready build 
> in the core language

Funny when my US Army .45 cal pistol class champion uncle taught me to
shoot he told me it was better to learn the right way first so not to
develop the bad habits in the first place.  I find that today's "program
in a box" college courses bread more bad habits then they actually teach 
good programming practices.  Learning from the Python modules is a fine 
way of learning the right way first.

Also, programming at the entry level is a bit like your basic hand tools. Life 
becomes greatly more interesting/easier if you know how to use them CORRECTLY.
All be it most of us have no desire to become carpenters, auto mechanics, &c.
Python is an excellent language to teach early on to find those that love
the discipline and educate the rest on how to use today's version of a hand 
tool CORRECTLY.

> - 45% of these exercises are solved within "the python cookbook", 
> available also online
> 
Ditto


> - 50% of all exercises "to teach programming" are already solved in the 
> standard-library or are a method of a buildin object / are allready build 
> in the core language

Harald Massa <cpl.19.ghum at spamgourmet.com> wrote in message news:<Xns935A6D886AC5cpl19ghumspamgourmet at 62.153.159.134>...
> > [cut]
> > I think python is a very good language for education. You could get in
> > contact with local schools.
> 
> Thomas, I am not so sure about this.
> 
> Yes, I love Python, and it is the most elegant programming language I 
> ever met. 
> 
> Python gives you objectoriented, functionorientet and imperative 
> programming paradigm in one fitting package.
> 
> Python does not bother beginners with a million typecasts. Very great.
> 
> For education there are small drawbacks:
> 
> - 50% of all exercises "to teach programming" are already solved in the 
> standard-library or are a method of a buildin object / are allready build 
> in the core language
> 
> - 45% of these exercises are solved within "the python cookbook", 
> available also online
> 
> - if you learn programming with Python and later have to use Java or 
> Visual Basic, it will be the most frustrating experience for the young 
> fellows.
> 
> 
> Anyway, I regret not having met Python earlier in my programming career.



From koenig at v-i-t.de  Mon Apr 21 08:50:01 2003
From: koenig at v-i-t.de (Mirko Koenig)
Date: Mon, 21 Apr 2003 14:50:01 +0200
Subject: importing modules with the same name but from different directories
References: <mailman.1050892833.31640.python-list@python.org>
Message-ID: <b80pdp$81h$1@news.tu-darmstadt.de>

On Mon, 21 Apr 2003 04:42:51 +0200, logistix wrote:



>> -----Original Message-----
>> From: python-list-admin at python.org
>> [mailto:python-list-admin at python.org] On Behalf Of Mirko Koenig Sent:
>> Sunday, April 20, 2003 9:26 PM
>> To: python-list at python.org
>> Subject: importing modules with the same name but from different
>> directories
>>
>> I have a program consisting of many modules. Every modul has its own
>> directory in a directory called modul, eg.
>> 
>> .../modul/modul1
>> .../modul/modul2
>> 
>> In every modul directory is a diretory for the language files. These
>> files are used to display the modules label in different languages. eg.
>> 
>> .../modules/module1/language/german.py
>> .../modules/module1/language/english.py
>> .../modules/module2/language/german.py
>> .../modules/module2/language/english.py
>> 
>> My problem is:
>> If is start modul1 from within its modul directory ( .../modules/modul1
>> ) And modul1 calls modul2, then modul2 doesn't load its own language
>> file. Instead it loads the langugae file of modul1.

>> 
> Add an __init__.py file to each directory in your project.  This makes
> the directories packages.  Then you can use a fully qualified path such
> as "import modules.modul1.language.english as module1_english".  Note
> that you can't use the same name for different modules at runtime, so
> use "as xxx." to give them a more distinct name.

Is it possible to to something like this:
from modules.modul1 import * 
or
from modules.modul1.language.english import text as modul1_text
or
import modules.modul1 as modul1 and than get the languag file 

Mirko Koenig



From aleax at aleax.it  Tue Apr 22 07:54:53 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 11:54:53 GMT
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com> <mailman.1050087439.11600.python-list@python.org> <435d40e0.0304130212.5659bd4d@posting.google.com> <7Bgma.8749$LB6.237526@news1.tin.it> <435d40e0.0304152211.5a9511d4@posting.google.com> <ZZ7na.27561$T34.822409@news2.tin.it> <14848047.0304212004.5aff060d@posting.google.com>
Message-ID: <h4apa.71820$T34.1889768@news2.tin.it>

<posted & mailed>

Arm wrote:
   ...
>>     [ <expression> for <target> in <iterable> ]
>> 
>> is known in Python as a "List Comprehension".  There is no issue of
>> "with no loop" -- the "for" keyword inside the list comprehension
>> is indicating exactly the fact that the loop is taking place.  (List
   ...
>>     for mo in are.finditer(instring):
>>         ...
>> (whether written out like this, or inside a list comprehension,
>> makes no difference) makes mo assume, one after the other, the
>> values of RE matchobjects for each of those non-overlapping
   ...
> Alex,
> Thanks for the great explanation. 

You're welcome!

> It seems like both List
> Comprehension and Iterators could use more documentation on
> python.org. 

What, and leave us poor book-writers unemployed?-)  Seriously,
though, the slides of my tutorial on iterators and generators
are on http://www.strakt.com/docs/eup02_itgen_alex.pdf and I'd
have no problem fleshing this out into a text article if there
were demand.  And I'm not sure what more there IS to say about
list comprehensions at a tutorial level -- I cover them in a
page in Python in a Nutshell, and, since the page is part of
Chapter 4, which happens to be the sample chapter for the
Nutshell on O'Reilly's site, you can just download
http://www.oreilly.com/catalog/pythonian/chapter/ch04.pdf
and get it all (and much more:-) for free.

> I think it would have been a lot more straightforward and
> easier to comprehend (at least for beginners), if the object of a
> "list comprehension" had been limited to a "sequence" or an
> "iterable". (I would even prefer a different keyword when using
> iterables).

I'm not sure what you mean here by "a different keyword" or even
by "the object" being "limited".  Any sequence is iterable --
just call iter(S) on any sequence S and you'll get an iterator
on it, so, S is iterable.  Some iterables aren't sequences --
i.e. there are objects X such that iter(X) gets you an iterator,
but you cannot perform other sequence operations on X, such
as indexing X[n] for suitable integers n, slicing, etc, etc.

But in a construct such as "for x in X", whether as a loop
statement or inside a list comprehension, you're only using
the fact that X is iterable -- you're not slicing X, etc etc,
so why would you possibly care in this context whether X is
sliceable etc (i.e. a sequence) or not (a non-sequence iterable)?

And the object you're iterating on MUST be iterable -- no ifs
and buts -- if it isn't you get a suitable error message, e.g.:

>>> [x for x in 42]
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: iteration over non-sequence

So, I must be completely missing your point, because I really
do not understand WHAT changes (in the language? in the docs?)
you think "would have been a lot more straightforward and
easier to comprehend (at least for beginners)".  Clarify pls?


> It doesn't make sense to be iterating through an iterator.

Since iterators are iterable, why, sure it does!  given any
iterator x, iter(x) is x.  E.g.:

>>> x=iter(range(999))
>>> iter(x) is x
True

You can iterate on any iterable, and iterators are iterable,
so _of course_ you can iterate on iterators -- it would be
utterly weird if you *couldn't*.


> What is the gain from allowing the iterator to be placed in the
> position a container should be - given that iterable
> containers/objects must return the iterator anyway? 

Are you asking for usage examples where it's particularly
helpful to use an iterator object directly in loops, rather
than being forced to have a full-fledged sequence in order
to perform iteration?  Please read the above-mentioned sample
chapter (half a page on iterators, one-plus on generators --
it's a Nutshell, so of course it's quite terse;-) and the
above-mentioned tutorial slides (well they're just slides
of course, but...).  Then consider for example a task such
as: I have a function process_body(body) that takes a 'body'
(an iterable whose items are lines) and does wonderful things
by looping on it, e.g.:

def process_body(body):
    for line in body:
        process_line(line)

However what I have is a "message" -- an iterable whose
items are lines, but, before the "body", there's an unknown
amount of lines I don't care about (call them "headers")
terminated by a blank line (a line that's whitespace).  So,
how do I extract the body in order to pass it to the above
function process_body?

Thanks to the fact that I can use a for loop on an iterator
just like on any other iterable, it's dead easy:

def process_message(message):
    it = iter(message)
    for line in it:
        if line.isspace(): break
    process_body(it)

Easy, isn't it?  And we could care less whether 'message'
is a full-fledged sequence (e.g. a list of lines) or an
iterable that's not a sequence (e.g. an open text file) --
since all we do is get its iterator, as long as it's an
iterable, we're in clover.  So why ever would one want to
have to have sequences, distinguish (e.g. with different
keywords) between iterating on sequences and iterating on
other iterables, forbid iterating on iterators (why?!),
and all the other complications that you appear to be
advocating?  Guess I must be missing something here...

> And it seems like
> in order to implement the "iterating through an iterator" syntax, an
> equally confusing call was added to iterators - "return self".

If and when you choose to implement an iterator class, you
will need it to have (i.e., define or inherit) a special
method named __iter__ that does indeed use "return self".  So
what's confusing about it (quite apart the fact that there is
absolutely no "call" in this return statement, so, I DO find
it confusing indeed that you choose to call it a "call"!-)?

Just the same "return self" idiom is widespread in many other
special methods (e.g., __iadd__, __imul__, ... just to name
a few), and in a quite widespread programming style (supporting
"method chaining") also in normal methods.  So what's your problem?


Alex




From imbosol-1049773207 at aerojockey.com  Tue Apr  8 01:25:12 2003
From: imbosol-1049773207 at aerojockey.com (Carl Banks)
Date: Tue, 08 Apr 2003 05:25:12 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net> <b6qs04$7eqi1$1@ID-169208.news.dfncis.de> <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net> <b6tedc$8rgif$1@ID-169208.news.dfncis.de>
Message-ID: <Y2tka.1182$d8.143@nwrdny02.gnilink.net>

"Greg Ewing (using news.cis.dfn.de)" wrote:
> Carl Banks wrote:
>> Second, given that the keyboard focus can only be active in one buffer
>> at a time, the idea of a current buffer is absolutely necessary.
> 
> There certainly needs to be a notion of which buffer
> has the keyboard focus. But I was thinking of the way
> that lots of elisp code uses current-buffer as a
> kind of implicit parameter, when it really has
> nothing to do with keyboard focus. The fact that
> there exist things like with-buffer is a testament
> to this.

Yes, I saw that that's what you meant later.  Sorry.


-- 
CARL BANKS



From mcfletch at rogers.com  Fri Apr 11 11:26:18 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Fri, 11 Apr 2003 11:26:18 -0400
Subject: Surprise: del weakref.WeakKeyDictionary[ item ] raises RuntimeError
In-Reply-To: <b701jn$9ls3d$1@ID-169208.news.dfncis.de>
References: <mailman.1049836994.31575.python-list@python.org> <b701jn$9ls3d$1@ID-169208.news.dfncis.de>
Message-ID: <3E96DE9A.9080002@rogers.com>

Greg Ewing (using news.cis.dfn.de) wrote:

> Mike C. Fletcher wrote:
>
>>        for ref in self.data.iterkeys():
>>            o = ref()
>>            if o == key:
>>                del self.data[ref]
>>                return
>>
>> Using keys seems to avoid the problem, but with the obvious 
>> disadvantages of keys vs. iterkeys.
>
>
> Maybe a separate list of keys could be maintained
> alongside the dict? 


It would have a pretty significant performance penalty if you always 
maintained a list.  Although I guess that doing two dictionaries:

    ref : object
    id : ref

And then doing all the logic to make sure they keep in synch would give 
you acceptable performance (and much faster than the current version for 
del (at the expense of lots of memory used and slower performance 
elsewhere)).

An "atomic" threading mechanism applied to the iterator-based del method 
might work just as well, I suppose.

In the meantime, using keys() *is* actually "maintaining" the seperate 
key-list for as long as it's needed.

Oh well, has to be some downside to iterators or we wouldn't have any 
fun :) ,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From staschuk at telusplanet.net  Fri Apr 25 11:21:52 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 09:21:52 -0600
Subject: What means exactly "Memory error"?
In-Reply-To: <200304251618.14933.b.maryniuk@forbis.lt>; from b.maryniuk@forbis.lt on Fri, Apr 25, 2003 at 04:18:14PM +0200
References: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net> <mailman.1051198958.16669.python-list@python.org> <slrnbai52u.8cv.hat@se-46.wpa.wtb.tue.nl> <200304251618.14933.b.maryniuk@forbis.lt>
Message-ID: <20030425092152.F261@tibia.amotlpaa.bogus>

Quoth Bo M. Maryniuck:
> On Friday 25 April 2003 12:56, Albert Hofkamp wrote:
> > Do you need all data DOMmed at once?
> 
> That what I doing right now. Well, it not so easy to make an app with high 
> perfomance where to parse it *twice* and re-collect data from the disk 
> storage. But I see there is no other way...

Others have suggested using a streaming API (such as SAX), and
that seems like the obvious way to reduce your program's memory
consumption.

Certainly breaking the document up at <hr>s is easy with SAX.
What are you doing to the parts afterwards?  For example, when
you're transforming one part, do you need access to all the other
parts?

If you do need access to all the other parts, one alternative
approach would be to replace your dict parsedBranches with a
shelf.  This would move the unneeded parts out to disk, saving
memory (at the cost of time, of course), providing a crude kind of
virtual memory.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From saveenr at online.microsoft.com  Sun Apr 13 20:30:58 2003
From: saveenr at online.microsoft.com (Saveen Reddy [MSFT])
Date: Sun, 13 Apr 2003 17:30:58 -0700
Subject: properties & exceptions
Message-ID: <3e9a0144$1@news.microsoft.com>

I was suprised by a python behavior I saw concerning exceptions thrown when
accessing a class through a property. Essentially, I am seeing an unexpected
exception, one that implies the property doesn't exist when in fact it does.
I haven't found any docs that explain this behavior and so my questions are
(1) is this behavior "by-design"? and (2) Besides not using properties, is
there a workaround?

(I see this behavior in Python 2.2.1 and 2.2.2)

Example code:

class FOO :

 def __init__( self ) :
  self.__bar__ = "Hello World"

 def _getbar1( self ) :
  return self.__bar__

 def _getbar2( self ) :
  raise "TROUBLE!"
  return self.__bar__

 def bar3( self ) :
  raise "TROUBLE!"
  return self.__bar__

 bar1 = property ( _getbar1 )
 bar2 = property ( _getbar2 )

foo = FOO()

#------------------------
#The line below works as expected ...
#-----------------------

>>> foo.bar1
'Hello World'


#------------------------
# When the line below runs ... Hmm... An exception was
# raised, yes. But I didn't
# expect AttributeError. I
# expected to receive the exception I threw, not a different one
#-----------------------

>>> foo.bar2
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AttributeError: FOO instance has no attribute 'bar2'

#------------------------
# Below is call to a normal method which has the
# behavior I expected, i.e. I received the proper exception
#-----------------------

>>> foo.bar3()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "proptest.py", line 16, in bar3
    raise "TROUBLE!"
TROUBLE!

Any wisdom is appreciated.
-s





From scott.daniels at acm.org  Mon Apr 14 21:36:28 2003
From: scott.daniels at acm.org (Scott David Daniels)
Date: Mon, 14 Apr 2003 18:36:28 -0700
Subject: splitting strings into variables
In-Reply-To: <pan.2003.04.14.23.33.00.456110@soul.net>
References: <pan.2003.04.14.23.33.00.456110@soul.net>
Message-ID: <3e9b6211$1@nntp0.pdx.net>

lost wrote:
>...
> strname.split(":")
> print strname
> hello
You are close.  Strings are values, not variables, so you
don't modify strname directly.  Try:
     strname = "hello:there"
     first,last = strname.split(":")
     print first

Of course, this is often better as:
     first,last = strname.split(":", 1)
if you know there is a colon, or as:
     first = strname.split(":")[0]

-Scott David Daniels





From maney at pobox.com  Wed Apr  2 00:13:17 2003
From: maney at pobox.com (Martin Maney)
Date: Wed, 2 Apr 2003 05:13:17 +0000 (UTC)
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <3E89EC01.F0CF6D31@engcorp.com> <b6d0jm$n7j$1@mailgate2.lexis-nexis.com>
Message-ID: <b6drhd$4or$2@wheel2.two14.net>

rzed <Dick.Zantow at lexisnexis.com> wrote:
> As an enhancement, I'd like to suggest two-way conversion to the
> similar Pig-Latinate numeric system of the form XIA (9), CIXA (91), or
> IIVA (7).

+1 ROFLASTC



From creedy at mitretek.org  Wed Apr  9 10:05:28 2003
From: creedy at mitretek.org (Chris Reedy)
Date: Wed, 09 Apr 2003 10:05:28 -0400
Subject: Import 'filename'? (was Re: Cross-reference 'import' in a classhierarchy)
In-Reply-To: <b6vri8$9ob33$1@ID-169208.news.dfncis.de>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <3e932e9f$1_2@corp-news.newsgroups.com> <b6vri8$9ob33$1@ID-169208.news.dfncis.de>
Message-ID: <3E9428A8.9090005@mitretek.org>

Greg Ewing (using news.cis.dfn.de) wrote:
> Chris Reedy wrote:
> 
>> I had to think about this one for a while. Here's the issue I see:
>>
>> What exactly does import '../foo/ook.py' actual do?
> 
> 
> You're being distracted by a red herring, I think. What
> I took the original post to mean was doing the equivalent
> thing in the Python module namespace, not the file
> system namespace.

You may be right. Thinking about this in the module namespace makes this 
seem better.

> Some suitable syntax would need to be devised,
> perhaps
> 
>   import ..foo.ook
> 
> although I'm not sure how you would refer to the
> imported module after that... maybe an "as" clause
> should be mandatory in this case?

There was a suggestion made in an old e-mail for something like __root__ 
or maybe __parent__ meaning the parent module of this one. So you would 
say something like:

   import __root__.foo.ook

I think that would have to be backed up by globals, something like 
__root__ or __parent__ referring to the parent of this module and maybe 
__self__ (or __this__ or __module__ or ...) which would refer to the 
current module. This would allow you do to things like:

   import __root__.foo.ook
   __root__.foo.ook.f()

Otherwise, I think you're right, an "as" clause would be mandatory.

> If you want to work in the file system namespace, you
> can already do that:
> 
>   ook = __import__("../foo/ook.py")
> 
>> Based on some reading in the email archives, there is clearly a desire 
>> (maybe a mandate by the BDFL) that some sort of correspondence be 
>> guaranteed to exist between the above four definitions (especially 1, 
>> 2, and 3).
> 
> 
> I sometimes wonder whether the Python module namespace and
> the file system should be one and the same thing. I certainly
> believe that PYTHONPATH is a mistake -- the environment required
> for a program to function correctly is a property of the program
> itself, not of the user's personal settings.

I don't like the idea of working in the file system namespace. I 
observed a whole discussion on the GCC mailing list of when (in an os 
independent fashion) two files are "the same". I remember a discussion 
of looking at things like inodes. I don't remember any resolution. I 
don't find myself that unhappy with PYTHONPATH.

> Something better is needed, I'm just not sure exactly what
> yet.

Agreed.

-- 
This is informal and not an official Mitretek Systems position.
Dr. Christopher L. Reedy, Senior Principal Software Engineer
Mitretek Systems, 3150 Fairview Park Drive South, Falls Church, VA 
22042-4519
Email: creedy at mitretek.org  Phone: (703) 610-1615  FAX: (703) 610-2203



-----------== Posted via Newsfeed.Com - Uncensored Usenet News ==----------
   http://www.newsfeed.com       The #1 Newsgroup Service in the World!
-----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =-----



From h_schneider at marketmix.com  Thu Apr 17 09:54:53 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Thu, 17 Apr 2003 15:54:53 +0200
Subject: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net>
Message-ID: <b7mbnb$pp2$06$1@news.t-online.com>

No permission probs, logged in as Admin. Platform is Win2K, not updated from
NT.
There are no exceptions. Reading a registry key just claims it would not
exist. When I run the same code outside the service wrapper, I can read the
key.

--Harald


"Thomas Heller" <theller at python.net> schrieb im Newsbeitrag
news:y929gt00.fsf at python.net...
> "Harald Schneider" <h_schneider at marketmix.com> writes:
>
> > Hi,
> >
> > I created a NT service under with py2exe under Python 2.1
> > The service ran without any problems.
> >
> > Then I upgraded to Python 2.2 and py2exe for 2.2 and now I have a big
> > problem:
> > All functions that access registry keys from within the service fail.
> > When I run my application outside the Service class, everything works
well.
> >
> > Any idea?
>
> Whithout any further information, one can only guess.
> Permission problems? Did you not only move from Python 2.1 to 2.2,
> but also from NT to 2000?
>
> What error do you get?
>
> Thomas





From aleax at aleax.it  Tue Apr 29 08:16:55 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 12:16:55 GMT
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <slrnbas791.j6l.ramen@lackingtalent.com>
Message-ID: <X2ura.19473$3M4.564080@news1.tin.it>

Dave Benjamin wrote:

> In article <UXara.22428$K35.745468 at news2.tin.it>, Alex Martelli wrote:
>> "How many items does XX[a:b] have?"
>> 
>> With Python's design (following Koenig's principle), the normal
>> answer (when a and b are both within XX'x boundaries) is:
>>         b - a
>> 
>> If the last-bound had been deemed to be included, it would have been:
>>         b - a + 1
>> 
>> Now, that innocuous-looking "+ 1" there is the cause of untold misery
>> in programming -- the root of most "off-by-one" errors.  By removing

Hmmm, I had let my enthusiasm take over here a bit -- make that "many"
rather than "most"!-)

>> the need for it, Koenig's principle makes you more productive.
> 
> Interesting! I've never heard that before. Thanks for the explanation.

You're welcome!


> I've definitely been bitten by this when trying to treat a 1D array as a
> 2D array by multiplying rowcounts. It's even more confusing when it's not
> + 1 anymore, it's + something else... or is it minus? Ahhh!

Oh yes -- and languages where array indexing starts at ONE, rather than
at zero, are even more HORRID that way!


Alex




From aleax at aleax.it  Thu Apr 17 06:39:53 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 10:39:53 GMT
Subject: PythonCard vs. Boa Constructor (was Re: Python for VB6 programmers)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com> <6e07b825.0304162200.34884d36@posting.google.com> <b7lhdg$q0f$1@panix3.panix.com>
Message-ID: <Zvvna.38978$T34.1112133@news2.tin.it>

Aahz wrote:

> In article <6e07b825.0304162200.34884d36 at posting.google.com>,
> yaipa h. <yaipa at yahoo.com> wrote:
>>
>>Here in the SF Bay Area I have only met one VB programmer willing to say
>>that he uses VB.net. The other VB programmers I find are talking about
>>either sticking with VB6 (most), taking Java courses (some) or playing
>>around with the C# idea (few).  Most share my feelings about the .net
>>thing. That it will not be stable until release +3.x and there are
>>little to no existing libraries. You know, batteries. ;-) So yes there
>>is a window and it is wide open.
>>
>>With that said, without a slick, stable GUI builder with all the
>>look and feel of the VB GUI builder it's a long way from here
>>to there. I mean what do you expect them to do, pass in command-line
>>arguments or hand assemble TK widgets. :^)
> 
> Good point.  Anyone here willing to take a whack at claiming that
> PythonCard, Boa Constructor, or something else would the best path for a
> VB6 programmer?

Personally I would suggest BlackAdder -- it's a commercial product,
but not too costly, and while still Beta it's IMHO more stable than
Boa (haven't tried PythonCard, natch).  The look and feel of each
of these IDE's is obviously hugely different from that of VB, though.

Not sure if ActiveState's "Visual Python" add-on for Microsoft Visual
Studio .NET can provide sufficient functionality, as I haven't tried
it yet -- at least the MS look and feel should be there, though,
since it DOES work on top of MS's IDE...!-)


Alex




From pobrien at orbtech.com  Fri Apr 18 10:54:39 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 18 Apr 2003 09:54:39 -0500
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050394215.20671.python-list@python.org> <b7i26g$17t5m$1@ID-169208.news.dfncis.de> <m24r4yh1l8.fsf@orbtech.com> <b7l1ov$22o4q$1@ID-169208.news.dfncis.de> <m2brz514kb.fsf@orbtech.com> <b7nr2r$2t3nf$1@ID-169208.news.dfncis.de>
Message-ID: <m2of33svo0.fsf@orbtech.com>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> writes:

> Patrick K. O'Brien wrote:
> > Assuming I had the
> > mechanism in place to bind keystrokes to a user-created command, how
> > would I know what parameters that command expected?
> 
> I think there's a mistake here in thinking of commands
> as being bound to keystrokes and nothing else. It
> would be better to think in terms of handlers for
> keystrokes being bound to particular objects, or
> classes of objects.
> 
> For instance, a command that operated on the text
> in a buffer would be attached to the Buffer class,
> and one that operated on a window would be
> attached to the Window class.
> 
> So the command for binding a command wouldn't
> just say "bind this function to this key", it
> would say "bind this function to an occurrence
> of this key being received by this class of
> object".

Hmm.  Sounds a bit like the typical event handling in a GUI.  That
could be good.  So, off the top of my head, let's say each object had
a dictionary used to register keychord/event handler pairs, and that
the event handler functions were bound to the object as methods.  That
would allow the functions access to attributes and other methods of
the object.  If the function/method were bound to the object
containing most of the state needed by the method, there would be no
need to pass parameters to the method.  Other objects could be
accessed by navigating the object graph.

Did I understand you correctly?

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From robin at jessikat.fsnet.co.uk  Thu Apr 17 09:25:35 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Thu, 17 Apr 2003 14:25:35 +0100
Subject: fixedpoint.py vs Python 2.1
Message-ID: <$uJOMKAPtqn+Ew9c@jessikat.demon.co.uk>

I'm having a bit of difficulty with the fixedpoint module in Python 2.1.

I get the following error

Python 2.1.3 (#35, Apr  8 2002, 17:47:50) [MSC 32 bit (Intel)] on win32
Type "copyright", "credits" or "license" for more information.
>>> from fixedpoint import FixedPoint as fp
>>> fp(fp(.175,3)*fp(22,2),2)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "fixedpoint.py", line 368, in __mul__
    return _mkFP(self._roundquotient(n1 * n2, _tento(p)), p, type(self))
  File "fixedpoint.py", line 497, in _mkFP
    f = FixedPoint()
TypeError: object of type 'type' is not callable
>>>

is there some reason for passing type(self) around rather than
self.__class__ etc. 
-- 
Robin Becker



From aleax at aleax.it  Wed Apr 30 04:18:54 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 08:18:54 GMT
Subject: Raising objects
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it> <mailman.1051678851.21661.python-list@python.org>
Message-ID: <OFLra.36816$K35.1143840@news2.tin.it>

Steven Taschuk wrote:

> Quoth Alex Martelli:
>   [...]
>> No, in 2.2 (and 2.3 beta 1) exceptions must still be classic-classes
>> or instances thereof [or strings for backwards compatibility with the
>> dark and distant past], and should be derived from Exception (but this
>> is a recommendation, not enforced).  I do not know when (or if) this
>> restriction will be removed -- in particular, because I do not know
>> of "use cases" where the ability to raise new-style classes or instances
>> thereof would be a major boon to an application.  Maybe you have some...?
> 
> I don't have a real use case either, but it does seem rather
> arbitrary.  Is there an implementation reason for this restriction?

In a sense, yes: there is no way to check if something "is a new-style
class" as opposed to "any type whatsoever" -- so one could end up
raising lists, tuples, strings (shades of the past...), whatever.

What I hope to see happen eventually is a spec that raise can use:

  -- any subclass of Exception (by that time a new-style class), OR
  -- for backwards compatibility any old-style class, OR
  -- for VERY backwards compatibility, any string

[and also INSTANCES as well as classes for the first two cases].  And
maybe we can dispense with the third subcase, since string exceptions
HAVE been deprecated for MANY, MANY years by now, after all.

But that wouldn't make your intended use-case work, anyway:


> A somewhat silly possible use:  A depth-first search to find a
> single object satisfying a given property.  An implementation
> (ab)using exceptions for control flow:
> 
>     def _search(node):
>         if condition(node):
>             raise node
>         else:
>             for child in childrenof(node):
>                 _search(child)
> 
>     def findone():
>         try:
>             _search(root)
>         except Node, found:
>             return found
>         else:
>             return None
> 
> If the objects in the search tree are instances of new-style
> classes, the restriction would be inconvenient.  It would be
> simple enough -- though ugly -- to wrap them in an old-style
> class for raising, of course.

Exactly: even in my hoped-for new spec, you'd have to wrap this
in a subclass of Exception.  Facilitating the [ab]use of the
exception mechanism for generalized control flow is, I suspect,
NOT on the list of priorities in the BDFL's mind;-).  Feel free
to open a PEP to settle that beyond any doubt, though...;-)


Alex




From aahz at pythoncraft.com  Sat Apr 26 00:55:49 2003
From: aahz at pythoncraft.com (Aahz)
Date: 26 Apr 2003 00:55:49 -0400
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com>
Message-ID: <b8d3gl$i6a$1@panix3.panix.com>

In article <pan.2003.04.26.04.22.16.562041 at mn.rr.com>,
Anna <revanna at mn.rr.com> wrote:
>
>My daughter just asked me about this "python stuph". So I pulled up
>livewires.org.uk and showed her some basics.
>
>She just coded her first while loop and for loop... she thought it was
>kewl! In fact, we had to kick her off the computer and send her to bed cuz
>she's got so much to do tomorrow...

Yay!
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From news at snakefarm.org  Thu Apr  3 08:30:09 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: Thu, 03 Apr 2003 15:30:09 +0200
Subject: FTP : get  dir name with space caracters embedded  + is ther
   something quicker than LIST
References: <b1ab591.0304030143.52422a3e@posting.google.com> <m37kab27pw.fsf@telia.com>
Message-ID: <b6hd11$soj$1@news.schlund.de>

Andrew Markebo wrote:

> | .....
> | may 22   2001     name of directory
> | mar 15    02:00   other directory name
> | .....
> 
> Implementation error ;-) Maybe should try line.split(" ", 3), telling
> it only to split the 3 first spaces, the rest should probably stick
> together..

Since the columns seem to be separated by more than one space he would 
have to say

line.split(None, 3)

to actually split off the directory name. 'None' tells split() to treat 
any amount of whitespace as one single separator and to ignore leading 
and trailing whitespace.

cg.




From hygt2 at cam.ac.uk  Tue Apr 29 08:07:02 2003
From: hygt2 at cam.ac.uk (Garrison Tsang)
Date: Tue, 29 Apr 2003 13:07:02 +0100
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com>
Message-ID: <b8lpof$q0f$1@pegasus.csx.cam.ac.uk>

> I doubt it.  Maybe it would be helpful if you explained why you think you
> need to convert the Python program into C?  Does it not run fast enough?
> Does it have to interface with something else that is in C?  Does it have
> to run on a platform on which Python is not supported?
>
> If you haven't got a really good reason, you might be better off leaving
> the code in Python.  If you think it's an issue that you don't know how
> to program in Python yet, believe me, you'll become more proficient,
> faster, with Python than you are with C in a very short time.
>
> -Peter

Thanks Peter.

I know nothing about Python but know a fair bit about C.

There is a program written in Python by a friend.  It takes in a file and
decrypts it , which I would like to integrate into a C program I am writing.
I want to see how the python code was implemented and whether I can
implement it is C.

Any solution for this?





From aleax at aleax.it  Sun Apr 27 03:52:37 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 07:52:37 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <mailman.1051422522.15507.python-list@python.org>
Message-ID: <9%Lqa.13853$K35.457009@news2.tin.it>

Andrew Bennetts wrote:
   ...
> lack of a better word.  It's usually not any clearer to say
>     
>     if len(myList) == 0:
>         print "I have a list of stuff"

The "== 0" part here is making it to the reverse of what you probably mean.


> than
> 
>     if myList:
>         print "I have a list of stuff"

Absolute agreement here.


> I think the same applies to testing to see if a variable holds a callable,

Not at all.  "if callable(thevariable):" is the one obvious way to check
THAT.  "if thevariable:" is going to have you try to call the number 23,
the string 'loolapalooza', and non-empty dicts and lists.

> just as much as it applies to checking the length of something.  Removing

The two cases are very different.  "if len(myList):" is equivalent to
"if myList:" *for a known given type of myList* -- a container.  In
the callable case, you _don't know_ the type -- and are checking the
VALUE to try and infer the TYPE.  Checking callable(f) is more obvious
and direct -- it clearly and unambiguously shows what you're checking.

> this feature will lead to unnecessarily longer code for minimal benefit.

I dispute the "unnecessarily" and the "minimal".  Tests for "is None"
add clarity (when they're applicable), and tests for callable(f) even
more so.  And giving newbies a little help by diagnosing a very common
newbie error more promptly would help (or at least it might, were it
not for some newbies' insistence that the diagnosis should be performed
by static analysis instead -- it just can't, and thus said insistence
convinces me that this is not worth putting to a PEP -- however, I find
the arguments against it quite unconvincing, except for the "we can't
break this idiom because even though it's inferior it's widely used").


Alex





From mh at pixar.com  Fri Apr 11 13:40:44 2003
From: mh at pixar.com (Mark Harrison)
Date: 11 Apr 2003 10:40:44 -0700
Subject: exceptions == errors?
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
Message-ID: <c703043e.0304110940.1aa40822@posting.google.com>

To wrap this up, I will be changing my code so that
all server responses are generated with exceptions.
This has the nice side effect that all unexpected
exceptions (which roughly translate to either "should
not occur" conditions or bugs) will generate a nice
internal error response with enough diagnostics to
figure out what went wrong.

Performance is not that much of an issue because the
server actions are extremely heavyweight.

Thanks to all,
Mark



From peter at engcorp.com  Fri Apr 11 10:34:24 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 11 Apr 2003 10:34:24 -0400
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net>
Message-ID: <3E96D270.B8E99B27@engcorp.com>

Mads Orbesen Troest wrote:
> 
> On Fri, 11 Apr 2003 10:48:49 +0200, Harald Massa wrote:
> 
> > Yes, I love Python, and it is the most elegant programming language I
> > ever met.
> 
> I agree it's mostly elegant, but the OO features are clumsy in my oppinion
> (coming as a professional developer from other languages). Not as bad as
> Perl's bless-a-dictionary-into-being-a-class retro-fitting hack, of course,
> but really; all the self me here and self me there is - to me at last -
> exceedingly clumsy and a constant source of frustration.

>>> import this


-Peter



From anton at vredegoor.doge.nl  Sun Apr 20 11:42:24 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Sun, 20 Apr 2003 17:42:24 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net>
Message-ID: <b7ufg4$o1d$1@news.hccnet.nl>

"A.M. Kuchling" <amk at amk.ca> wrote:

>http://mail.python.org/pipermail/python-dev/1999-July/000467.html

Thanks. It seems Tim is making a heroic effort to explain the new in
terms of the old. This necessarily must fail, but what other option is
there instead of rebuilding all terminology out of gotos?

By the way, I have a hunch the anagram script below is somehow related
to the operations described, since the script "changes directions"
depending on the order the lists items are popped. A simple
permutation could just do a series of divmods. 

If not, please excuse me, it's just a natural defensive reaction to
try to hold of the unknown with the shield of the known. In fact
knowledgeable people are welcome to hide behind *my* ignorance :-)

Anton

#anagram.py by Anton Vredegoor anton at vredegoor.doge.nl Apr 20, 2003
"""
    Starters(L) returns a 2-tuple for each *different* item
    in the list L:
    
    a) the index of the first occurrence of the item,
    b) the number of anagrams starting with this item.

    The function starters is used by the function anagram to 
    construct indexed anagrams of a list.
"""

def starters(L):
    #return a list of 2-tuples containing info about list L
    n,np,R = len(L),1,range(len(L))
    bf = [L[:i].count(L[i]) for i in R]
    for i in R: np = np*(n-i)/(bf[i]+1)
    return [(i,np*L[i:].count(L[i])/n) for i in R if not bf[i]]
    
def anagram(L,index=0):
    #return an anagram of list L
    remain, res, T = index, [], L[:]
    while T:
        for j,k in starters(T):
            if remain-k < 0:
                res.append(T.pop(j))
                break
            remain -= k
    return res

def nperm(L):
    #return the number of possible anagrams of list L
    return reduce(lambda a,b:a+b,[k for j,k in starters(L)])

def test():
    L = [[1],[1],[2],[3]]
    np = nperm(L)
    for i in range(np):
        print '%2i %s' %(i,anagram(L,i))

if __name__=='__main__':
    test()




From rmunn at pobox.com  Thu Apr 10 13:34:31 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 10 Apr 2003 17:34:31 GMT
Subject: atexit functionality
References: <mailman.1049821814.20582.python-list@python.org>
Message-ID: <slrnb9am2o.bom.rmunn@localhost.localdomain>

viszneki at telerama.com <viszneki at telerama.com> wrote:
> I am not a member of the Python mailing list, I dont' have the time to read any 
> mailing lists at the moment. So if someone could please respond to me 
> personally, that would be great.
> 
> Generally speaking I have a lot of objects that clean themselves up before 
> python actually exits, so I was wondering if a deregister function could be 
> provided.
> 
> Rather than a deregister function, I thought about making the following 
> modification to atexit.py after _exithandlers.append((func,targs,kargs))
> 
> return _exithandlers[len(_exithandlers)]
> 
> My intention is to be able to store the returned value as a member variable in 
> a class that registers its own cleanup functions. But if the class cleans 
> itself up, the class could "clean up" its cleanup registration with atexit by 
> saying something like del self.cleanup

You shouldn't need to return the just-inserted tuple. The caller already
has a reference to the function that was just registered -- it has to
have that reference, since it just passed that reference in as a
parameter! So you could write an "unregister" function like so:

    def unregister(func):
    """Unregister a function from the exit-handlers list.
    
    func - function to be unregistered.
    """
    for i in range(len(_exithandlers)):
        if _exithandlers[i][0] is func:
            del _exithandlers[i]
            break

Notice that I used the "is" operator to check for object identity.

If you think this is useful, feel free to prepare a patch and submit it
to the python-dev list for consideration. Be prepared to back it up with
real use cases demonstrating why this should be added, though. One of
the reasons Python is still so clean is that Guido won't make changes
without good, compelling reasons. This has (for the most part) avoided a
lot of cruft getting into the language.

Carefully-avoiding-mentioning-ternary-operators-ly y'rs, (oops, too late)

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From garryknight at gmx.net  Thu Apr 10 19:20:58 2003
From: garryknight at gmx.net (Garry Knight)
Date: Fri, 11 Apr 2003 00:20:58 +0100
Subject: Can I define nested classes in Python?
References: <169a7283.0304101432.13f5786c@posting.google.com> <1JGdnc952fcaaAijXTWcrg@comcast.com> <6dub9v0rjv0s7r87hhjsjqb0oqjuu9m8ko@4ax.com>
Message-ID: <1050016856.33115.1@doris.uk.clara.net>

Rene Pijlman wrote:

> But wait a minute, why would a newbie want to use nested classes
> anyway?

A newbie to Python isn't necessarily a newbie to programming. As I'm sure
you know.  :o)

-- 
Garry Knight
garryknight at gmx.net   ICQ: 126351135
Linux registered user 182025



From andrew at acooke.org  Tue Apr 29 14:11:02 2003
From: andrew at acooke.org (andrew cooke)
Date: Tue, 29 Apr 2003 14:11:02 -0400 (CLT)
Subject: how to parse numeric data files
In-Reply-To: <3EAEB951.5DC25694@engcorp.com>
References: <20030429132519.4e92271e.gry@ll.mit.edu> 
     <3EAEB951.5DC25694@engcorp.com>
Message-ID: <3390.10.2.0.2.1051639862.squirrel@127.0.0.1>

if you do parse with regexps, you may want to consider enfocing rules that
detect when files aren't formatted as you expect.  for example, specify
evey line completely (don't throw away the end if you don't expecet
anything else there) and specify what information should always be
present.  otherwise you can miss data (eg when a file format is "improved"
slightly during an upgrade to one of your testing machines) for a long
time.

personally, i'd try to go via a complete formal spec using a grammar to
avoid this kind of problem.  the example you gave should be pretty easy to
parse since every field is prefixed with a name (looks like LL(1), in
which case yapps should work (know nothing about this, it's just first hit
on googling for "python parser")
http://theory.stanford.edu/~amitp/Yapps/).

another disadvantage of regular expressions is that you need to add error
handling yourself.  a good parser generator should give you pretty useful
error handling from the start.  maybe this doesn't matter so much if this
is for internal use, though.

the big disadvantage of a formal grammar + parser is (of course) that as
soon as the language isn't directly specified by a grammar that the parser
generator can handle, knowing "that boring stuff about LALR" can be
critical.  the good news with LL(1) is that it is easy to understand (your
engineers will find it intuitive), the bad news is that LL(1) is pretty
weak, so it's more likely to fail.  however, a quick scan of yapps
suggests it has extensions that make it considerably more powerful
(actually, from 30 seconds first impression, i'm quite impressed - otoh it
may be quite slow).

if i were you i'd try yapps in my free time and see how it goes.  if it's
too nasty, try regexps and cross your fingers...

andrew

Peter Hansen said:

> george young wrote:
>>
>> We have several electronic testing machines(of various ages and
>> manufacturers) that spew out testing data files in various ascii
>> formats.
>> Currently we have a nasty mess of awk/shell/C/fortran programs that
>> extract and process some data from these files.  I have a dream of
>> a suite of simple, clear, maintainable python programs to do these
>> tasks.
>>
>> The trick is I hope to come up with something that our hardware
>> engineers can understand and maintain easily without studying
>> things like BNF, LALR etc. (they won't).
>>
>> [Below is a sample of one of the worst formats, shortened from a 40MB
>> file!]
>
> The format is quite amenable to parsing with re, if rather large...
> the real question is how much of that data do you need, and what do
> you need to do with it?  What do your current scripts actually do?
> Also, how much of the content you showed is *fixed* format, and how
> much of the format can vary?  Is anything optional?
>
> If you want the hardware engineers to be able to maintain it, you might
> want to support a kind of "template" specification, where you provide the
> names of various tags which are recognized (e.g. "PH_lot_id:") and you
> automatically extract the appropriate value found thereater.
>
> What you're trying to do is not really that complex, I think, so I
> do think you should be able to find a good simple solution for it.
> As you said, this _is_ Python...
>
> -Peter
> --
> http://mail.python.org/mailman/listinfo/python-list
>
>


-- 
http://www.acooke.org/andrew




From cnetzer at mail.arc.nasa.gov  Wed Apr 23 03:35:39 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 23 Apr 2003 00:35:39 -0700
Subject: why not say just what wants an integer?
In-Reply-To: <87d6jeeyo3.fsf@jidanni.org>
References: <87d6jeeyo3.fsf@jidanni.org>
Message-ID: <1051083339.976.19.camel@sayge.arc.nasa.gov>

On Tue, 2003-04-22 at 19:27, Dan Jacobson wrote:
> Say, does lack of ability to figure out how to make this work 
> $ python -c 'print range(map(lambda x:x+1,(11,131,11)))'
> Traceback (most recent call last):
>   File "<string>", line 1, in ?
> TypeError: an integer is required
> 
> indicate 1. newbie, 2. lazy newbie, 3. can't really blame me?

Probably 1.

Note the following:

$ python -c 'print map(lambda x:x+1,(11,131,11))'
[12, 132, 12]


So the returned value of the map() is a list.  In older python, you
could use the apply() builtin function to unpack the list and provide
its elements as the arguments to a function.  Recent python (2.0 and
newer?) has syntax to do this:

python -c 'print range(*map(lambda x:x+1,(11,131,11)))'
[12, 24, 36, 48, 60, 72, 84, 96, 108, 120]

Note the asterisk.  In most cases, one must be specific about wanting to
treat a sequence as multiple arguments, and the asterisk notation makes
this quite convenient.

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From rob at zoism.org  Wed Apr  2 22:33:37 2003
From: rob at zoism.org (Rob Brown-Bayliss)
Date: 03 Apr 2003 15:33:37 +1200
Subject: plugin system?
In-Reply-To: <20030331014306.GB21551@unpythonic.net>
References: <mailman.1048836529.7090.python-list@python.org>
	 <KVVga.31670$i26.818965@news2.tin.it> <1049072379.21707.34.camel@musicbox>
	 <20030331014306.GB21551@unpythonic.net>
Message-ID: <1049340834.8760.22.camel@musicbox>

Whats with the following exec statment?

> In the code, you would instantiate the class with a helper function:
>     def get_plugin_factory(name):
>         # If you have packages (eg monty_python.spam.VikingChorus
>         # or spam.spam.spam.spam.baked-beans.spam) this won't work
>         # as written
>         m, f = name.split(".")
>         d = {}
>         exec "import m" in d
>         return eval(name, d)

I admit I havn't tried it yet, but what does the exec "import m" in d
do?

I had done this in the past:

def change_gear(self,newcog):
  if sys.modules.has_key(newcog):
### CUT SOME GUI SPECIFIC STUFF ###	
  else:
    try:
	setcog = __import__(newcog)
	cog = setcog._setup(self)
	self.caseing.append_page(cog,gtk.Label(newcog))
	self.caseing.set_current_page(self.caseing.page_num(cog))
    except ImportError, msg:

self.caseing is a tabed notbook gui element, and each module imported
(setcog statment) contains code for a new page in the notebook...

How does the __import__(somemodule) differ from the exec line int he
previous example?

Also what happens what I really want to do is create multiple instances
of some "plugin" code using Jeff's example rather than one instance of
it?

The more I think about my problem the more I realise that it's a
combination of changing attributes in the parent object and introducing
new code to modify the behaviour of the parent object... 


-- 

*  Rob Brown-Bayliss
*  =================	
*      zoism.org






From aleax at aleax.it  Mon Apr 14 17:39:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 21:39:52 GMT
Subject: Variable inside module
References: <mailman.1050351342.5655.python-list@python.org>
Message-ID: <IUFma.19201$T34.609834@news2.tin.it>

Marcelo A. Camelo wrote:

> 
> Hi!
> 
> Can someone please explain-me what is
> happening here:
> 
> Given the followin bar.py file (placed
> inside a foo/ directory):
> 
> var = 'foo'
> 
> def SetVar():
> global var
> var = 'bar'
> 
> def PrintVar():
> global var
> print 'var:', var
>       
> In the same foo/ directory, I've placed
> the following __init__.py:
> 
> from bar import *

"from X import" binds names in the current module (or
current local namespace) to the values those names
hold _at this time_ in module X.  And this is all it
ever does and all it will ever do.


> Then, in the interactive python interpreter:
> 
> >>> import foo
> >>> print foo.var
> foo
> >>> foo.PrintVar()
> var: foo
> >>> foo.SetVar()
> >>> foo.PrintVar()
> var: bar
> >>> print foo.var
> foo
> >>>
>       
> What I don't understand is that foo.var seems
> to be two different variables: one accessed
> inside the bar.py file and one accessed from
> within the python interpreter. What is going on
> here?

Essentially the same thing as having, in foo/__init__.py
(which is the module body for "module foo" -- that's how
packages work):

import bar
var = bar.var
SetVar = bar.SetVar
PrintVar = bar.PrintVar
del bar

any future rebinding of bar.var will have no effect
on what foo.var is bound to now.

Yes, there ARE two different names (and variables are
just that -- names): foo.var and bar.var.


I've often recommended forgetting the existence of
statement from -- the confusion it can cause is not
justified by the modest advantages it can provide
when compared to using import (plus potentially one
or more assignments).


Alex




From exarkun at intarweb.us  Fri Apr 11 12:41:51 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 11 Apr 2003 12:41:51 -0400
Subject: Confusion: assignments create object references.
In-Reply-To: <0C642BE4EE085149B52839125D4FC9D422BD9F@idc1exc0002.corp.global.level3.com>
References: <0C642BE4EE085149B52839125D4FC9D422BD9F@idc1exc0002.corp.global.level3.com>
Message-ID: <20030411164151.GC31416@meson.dyndns.org>

On Fri, Apr 11, 2003 at 10:28:42AM -0600, Havens, Peter wrote:
> I am confused about when assignments create references to objects as
> opposed to copying the value of those objects.

  That's easy.  It -always- creates references and -never- copies the value
of an object.

  Jp

> I realize this is probably a very common bit of confusion, but I haven't
> had much luck finding anything in the Python FAQ, from googling the
> comp.lang.python news group (voluminous thread tangents), or from the
> regular documentation I've read which has only fueled my confusion. To
> wit, I read the following "rule" in the O'Reilly book "Learning Python":
> 
> "...Python assignment stores references to objects in names or data
> structure slots. It always creates references to objects, instead of
> copying objects."
> 
> That seems fairly straight forward. So, I ran this test:
> 
> >>> L = [0,1,2]
> >>> M = L
> >>> L
> [0, 1, 2]
> >>> M[0] = 9
> >>> L[0]
> 9
> >>>
> 
> So far so good. This works as I would expect given the rule above.
> 
> >>> L = 0
> >>> M = L
> >>> L
> 0
> >>> M = 1
> >>> L
> 1
> >>>
> 

  The above never happened, I assure you. >:)    Python doesn't work this
way.

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 22 days, 13:01, 5 users, load average: 0.99, 0.98, 0.99




From abelikov72 at hotmail.com  Wed Apr 30 17:57:40 2003
From: abelikov72 at hotmail.com (Afanasiy)
Date: Wed, 30 Apr 2003 21:57:40 GMT
Subject: fcntl in windows
References: <k2rvav06nhjgtr578fa5ostftfl69rethp@4ax.com> <mailman.1051717847.12855.python-list@python.org> <HSSra.42028$K35.1214886@news2.tin.it> <7h33cjz51y0.fsf@pc150.maths.bris.ac.uk>
Message-ID: <gsh0bvkcan4tsr4kc3m0s4tcreh6tuq780@4ax.com>

On Wed, 30 Apr 2003 17:29:46 GMT, Michael Hudson <mwh at python.net> wrote:

>Alex Martelli <aleax at aleax.it> writes:
>
>> Jp Calderone wrote:
>> 
>> > On Wed, Apr 30, 2003 at 03:26:29PM +0000, Afanasiy wrote:
>> >> Is there an fcntl in Windows? A stub, a wrapper, etc?

>Indeed, it would seem to me that implementing a version of fcntl on
>Windows would be flat out insane -- surely the sensible thing to do
>would be provide two implementations of whatever it is you need, the
>unix one of which calls fcntl.

Insane? I strongly disagree.

Sensible thing? I strongly disagree.

The sensible thing would provide the functionality
in a well designed platform independent module.

It is no matter though, the answer to my question is no.

-AB



From ramen at lackingtalent.com  Tue Apr 15 19:04:06 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 15 Apr 2003 23:04:06 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr>
Message-ID: <slrnb9p3vo.vg8.ramen@lackingtalent.com>

In article <3E9C9A91.8090209 at removethis.free.fr>, laotseu wrote:
> Even for OO programmers, functionnal features in Python are IMHO a great 
> plus, and BTW functionnal and OO paradigm does not have to conflict 
> (that would be functionnal vs imperative and object vs procedural). CLOS 
> is one of the great system objects out here, and it's been implemented 
> on top of a functionnal language, so...

This is a good point. O'Caml comes to mind as well--a functional language
with object support right out of the box. There is no reason the two
methodologies need to be mutually exclusive, or step on each other's toes.

But, there is of course the risk of setting precedent that there are more
than one ways to do it, and this is why I brought up Perl. I think that
there is a subtle fear in the Python community of letting Python become too
much like Perl, especially since both languages have similar application
domains. There are other options besides TMTOWTDI and military rigor. This
is precisely why I like Python, because it seems to be a good compromise
between the two.

I think that list comprehensions bridge the gap somewhat between FP and OO
methods, and this is one thing I really like them for. For instance:

  [obj.method(x) for x in data]

looks nicer to me than:

  map(lambda x: obj.method(x), data)

because it puts the emphasis on the method call, not the variable.

> In fact, I'd rather regret that Python is not more functional *and* 
> object oriented !-)

In what ways? =)

Thanks for your input,
Dave



From theller at python.net  Tue Apr 15 14:50:32 2003
From: theller at python.net (Thomas Heller)
Date: 15 Apr 2003 20:50:32 +0200
Subject: create a new instance in C of a Python subclass
References: <36ko9vknj9mm6m1ad18r3dm51q3t02pe03@4ax.com>
Message-ID: <he8zegsn.fsf@python.net>

Paul Miller <paul at fxtech.com> writes:

> I've got a C extension type that is subclassed in Python. In C, I want to
> create a new instace of the subclass, which I have a pointer to. What
> magical incantation will allow this? I'm assuming some combination of
> PyObject_New and PyObject_Init are required, using the ob->ob_type pointer?

You do the same as in Python: call the type.

Does this help?

Thomas



From aahz at pythoncraft.com  Thu Apr 24 19:37:19 2003
From: aahz at pythoncraft.com (Aahz)
Date: 24 Apr 2003 19:37:19 -0400
Subject: string to variable name
References: <b89qi8$2oij$1@agate.berkeley.edu>
Message-ID: <b89sff$s4e$1@panix2.panix.com>

In article <b89qi8$2oij$1 at agate.berkeley.edu>,
Benjamin Hartshorne  <ben at hal.rescomp.berkeley.edu> wrote:
>
>I have a function that I want to change the variables in my class:
>
>def myassigner():
>  file = open("mytextfile", 'r')
>  for line in file.readline():
>    m = re.match("^(\w+)\s+(\w+)\s+(\w+)$", line)
>    (newval, var, oldval) = m.groups()
>    inst.var = newval

setattr(inst, 'var', newval)
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From peter at engcorp.com  Tue Apr  1 14:44:01 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 14:44:01 -0500
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <3E89EC01.F0CF6D31@engcorp.com>

Mike Meyer wrote:
> 
> Abstract
> 
>     This PEP proposes adding Roman numerals as a literal type.  It
>     also proposes the new built-in function "roman", which converts an
>     object to an integer, then converts the integer to a string that
>     is the Roman numeral literal equivalent to the integer.
> 
> Rationale
> 
>     Roman numerals are used in a number of areas, and adding them to
>     Python as literals would make computations in those areas easier.
>     For instance, Superbowls are counted with Roman numerals, and many
>     older movies have copyright dates in Roman numerals.  

In addition, any calculations involving the Roman calendar 
(cf http://www.clubs.psu.edu/aegsa/rome/romec.html) would clearly benefit
from such a module.

My kalend.py module, currently running at 87,543 lines of code, could
be shrunk substantially if it could do its calculations using the
new roman type.  *Thank you*, Mike, for starting the process of getting 
this valuable feature into Python.

(I've also considered whether we should be including the roman 
alphabet in the source encodings (PEP 263) so that, for example, 
names like "MAIOR" and "IVSTINIANUS" would be properly interpreted
as "major" and "Justinians".  Unfortunately, this requires coupling
that PEP with PEP 235 (Migration to Case-Insensitive Python), and 
coupling is always a bad idea in computing.)

Oh, and a Festive Veneralia to you all. :-)

-Peter



From mt at open2web.com  Mon Apr 28 15:22:58 2003
From: mt at open2web.com (Don Low)
Date: 28 Apr 2003 19:22:58 GMT
Subject: problem with string.lower function
References: <b8jrgf$ashp7$1@ID-145503.news.dfncis.de> <3EAD7281.D1BB6D90@engcorp.com>
Message-ID: <b8jv2i$anjik$1@ID-145503.news.dfncis.de>

Thanks, you're right. The line I added used spaces for indentation while
the other lines used tabs. 

>>         for eachLine in allLines:
>>         test = string.lower(eachLine) // change is here
>>         if string.find(test, word) > -1:
>>                 count = count + 1
>>                 print eachLine,
>> 

Mark



From mwh at python.net  Sun Apr  6 13:33:11 2003
From: mwh at python.net (Michael Hudson)
Date: Sun, 6 Apr 2003 17:33:11 GMT
Subject: negative values in profile output
References: <mailman.1049479748.24160.python-list@python.org> <7h38yunk17w.fsf@pc150.maths.bris.ac.uk> <qfZja.18781$It5.10774@news2.central.cox.net>
Message-ID: <7h3u1dbil9k.fsf@pc150.maths.bris.ac.uk>

"Steve Holden" <sholden at holdenweb.com> writes:

> "Michael Hudson" <mwh at python.net> wrote in message
> news:7h38yunk17w.fsf at pc150.maths.bris.ac.uk...
> > "Matthew M. Boedicker" <mboedick at mboedick.org> writes:
> >
> > > When using the profile module to profile a long-running program,
> > > I am getting some negative values:
> > >
> > > 6753628 -4036.027   -0.001 -4036.027   -0.001
> gobject_sql.py:36(<lambda>)
> > >
> > > Am I correct to assume these don't have special meaning and are the
> > > result of something wrapping?
> >
> > They may be a result of mis-calibration of the profiler -- see the
> > docs.
> >
> > Last time i used the profiler, I more-or-less ignored the values
> > output and just assumed I was doing the right thing when they got
> > smaller...
> >
> 
> But when a number comes up negative, does "smaller" mean more negative or
> closer to zero?

I was consciously avoiding thinking too hard about that :-)

Cheers,
M.

-- 
  This is an off-the-top-of-the-head-and-not-quite-sober suggestion,
  so is probably technically laughable.  I'll see how embarassed I
  feel tomorrow morning.            -- Patrick Gosling, ucam.comp.misc



From nirina at mail.blueline.mg  Wed Apr  2 04:02:52 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Wed, 2 Apr 2003 13:02:52 +0400
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server>
Message-ID: <01dd01c2f8f7$07840ba0$f7a2383e@raseliar>

> >"BBFoto" <bbfoto2002 at netscape.net> wrote:
> >
> >> I run the Python executable and type in "PySol" at the prompt on the
> >> black screen and get "name not defined."
> >
> >What happens if you type:
> >import PySol

> "module not found"

I suppose 'PySol.py' is the full name of your game and it is located at
'c:/your/game/directory', so you can try:

import os
os.chdir('c:/your/game/directory')
execfile('PySol.py')

or in one line:

execfile('c:/your/game/directory/PySol.py')

nirinA








From syver-en+usenet at online.no  Wed Apr  9 22:34:47 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 10 Apr 2003 04:34:47 +0200
Subject: HTML -> HTMLGen (or similiar)
References: <aT1la.8930$8Z5.56436@sea-read.news.verio.net>
Message-ID: <ur88bozaw.fsf@online.no>

VanL <vlindberg at verio.net> writes:

> Hello,
> 
> I have a bunch of existing HTML that I need to integrate into a new
> web app.  I would ideally like to use one of the existing HTML
> generation systems (HTMLgen, HyperText, any others) but I don't want
> to have to build up an equivalent representation for the many pages I
> need to work with.
> 
> The pages validate as HTML 4.01, so no strange measures are
> necessary. However, does anyone know of a package that can take
> existing (valid) HTML and make it into an object that I can
> manipulate?

I seem to remember that the python xml package has modules to read
HTML into a DOM tree. I think you have to download the xml package to
get it (it's not in the version included with python as far as I remember).
 

-- 

Vennlig hilsen 

Syver Enstad



From ianb at colorstudy.com  Wed Apr 30 20:30:45 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 30 Apr 2003 19:30:45 -0500
Subject: Web programming and a different "type" problem
In-Reply-To: <slrnbb0eoo.fb6.jon+usenet@snowy.squish.net>
References: <mailman.1051532982.25352.python-list@python.org>
	 <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com>
	 <slrnbarhfb.67t.jon+usenet@snowy.squish.net>
	 <b8lj2q$l14$1@newshost.accu.uu.nl>
	 <m3fzo1v561.fsf@ferret.object-craft.com.au>
	 <mailman.1051654786.4587.python-list@python.org>
	 <m365owlu6v.fsf@ferret.object-craft.com.au>
	 <mailman.1051673506.20716.python-list@python.org>
	 <slrnbb0eoo.fb6.jon+usenet@snowy.squish.net>
Message-ID: <1051749044.690.30.camel@lothlorien>

On Wed, 2003-04-30 at 16:07, Jon Ribbens wrote:
> In article <mailman.1051673506.20716.python-list at python.org>, Ian Bicking wrote:
> > It might be better if the error is signaled earlier and in a more
> > pleasant way, but IMHO throwing away duplicate values when they
> > aren't expected just covers up bugs.  But I don't think that's what
> > you're doing in this case (though it sounds like jonpy does this,
> > for instance)...?
> 
> You think it would be better to throw an exception if two form
> variables have the same name unexpectedly? I... disagree ;-)

Of course it is better, why not?  Unexpected behavior is probably a bug,
exceptions are a proper response to exceptional behavior.  So long as
you have a way to indicate whether you expect multiple form values for a
name, you should use that information.  Using it to throw away
information seems like the worst option.

  Ian






From aleax at aleax.it  Sat Apr 19 11:36:04 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 15:36:04 GMT
Subject: getting an item's list index
References: <mailman.1050757111.12171.python-list@python.org>
Message-ID: <E1eoa.57848$T34.1541880@news2.tin.it>

<posted & mailed>

Michael Anckaert wrote:

> Hello all,
> 
> I have a list called contents. If one of the elements has the value
> "item1" for example, how can I get that elements index?

This is one question, and I see it has been answered repeatedly and
correctly -- contents.index(item) will return the index of the FIRST
occurrence of value item in list contents (or raise an exception if
the value never occurs).


> for item in contents:
> # here I need to get that items index

This is a very, VERY different question, as I see some respondents
have mentioned, but it's probably worth clarifying.  There is NO
way you can retrieve the exact index of an item's value once you
choose to loop with "for item in contents:" -- you'll only get the
index of the FIRST occurrence of each item, each time you meet
that item.  For example:

>>> contents=list('paraply')
>>> contents
['p', 'a', 'r', 'a', 'p', 'l', 'y']
>>> for item in contents:
...     idx = contents.index(item)
...     print (idx, item),
...
(0, 'p') (1, 'a') (2, 'r') (1, 'a') (0, 'p') (5, 'l') (6, 'y')
>>>

See?  Each 'p' is always associated with index 0, that of the
first occurrence of 'p', and similarly each 'a' with index 1.

Even if you KNOW that the list contents has no duplicates, it's
still UN-advisable to use a loop such as this -- each call to
index takes time proportional to the resulting index, so the
overall time of the loop should end up taking time proportional
to the SQUARE of the list length's, so-called "O(N squared)".
For example:

[alex at lancelot python2.3]$ python timeit.py -s'c=range(1000)' 'for it in c: 
i = c.index(it)'
10 loops, best of 3: 4.02e+04 usec per loop

[alex at lancelot python2.3]$ python timeit.py -s'c=range(2000)' 'for it in c: 
i = c.index(it)'
10 loops, best of 3: 1.6e+05 usec per loop
[alex at lancelot python2.3]$

See?  A list twice as long takes up FOUR times as long for
the loop -- and:

[alex at lancelot python2.3]$ python timeit.py -s'c=range(3000)' 'for it in c: 
i = c.index(it)'
10 loops, best of 3: 3.6e+05 usec per loop
[alex at lancelot python2.3]$

list three times as long, takes NINE times as long for the loop -- etc.
Time for the loop grows like the square of the list's length, in fact.


You can do MUCH better, in such a loop, by looping on indices
directly -- easiest in the coming Python 2.3:

for idx, item in enumerate(contents):
    print (idx, item),

but not TOO terrible even in good old Python 2.2:

for idx in xrange(len(contents)):
    item = contents[idx]
    print (idx, item),


See what happens to the resulting time...:

[alex at lancelot python2.3]$ python timeit.py -s'c=range(3000)' 'for i, it in 
enumerate(c): pass'
1000 loops, best of 3: 1.12e+03 usec per loop
[alex at lancelot python2.3]$

these 1.12 milliseconds per loop correspond to the 360 milliseconds
per loop withe the "for it in c: i = c.index(it)" approach -- a
speedup of over THREE HUNDRED TIMES even for a list of reasonably
moderate length (3000 items), growing without bounds as the list gets
longer and longer.


An excessive focus on performance is a widespread disease in programmers,
but SOME modicum of attention to performance, when you can get speedups
of hundreds and hundreds of times without actually complicating your
code (in fact making it simpler and more elegant), is probably worthwhile.

So, summarizing: if you're looping over a whole list and may need to
know, within the body of the loop, up to what index in the list you've
reached -- forget the list's method .index, and ensure your LOOP itself
tells you the current index -- with enumerate if you can upgrade to
Python 2.3 (a bit too early for that, it's still in alpha -- but the
first beta will be out VERY soon!!!) -- or in Python 2.2 or earlier by
looping directly with the index, in xrange(len(contents)), and using
indexing, contents[i], to get the current item, rather than doing things
the other way 'round.  The key here is that indexing, contents[i], is a
VERY fast operation, taking a short constant time for any list and index
in it, while "looking for the index for an item" is a lengthy operation,
taking time proportional to the index of the item's first occurrence.


Alex




From mhammond at skippinet.com.au  Tue Apr 29 19:01:05 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Tue, 29 Apr 2003 23:01:05 GMT
Subject: win32 display jpg file
In-Reply-To: <irxra.10972$pK2.14342@news.indigo.ie>
References: <irxra.10972$pK2.14342@news.indigo.ie>
Message-ID: <RuDra.7018$lD4.74951@news-server.bigpond.net.au>

Ilariu Raducan wrote:

> Hi All,
> 
> Does anybody knows how to display a JPG file in pythonwin and draw lines 
> over it?

With a fair bit of work :)  You would need to use PIL to decode the jpg 
and convert it to a bitmap.  IIRC, Pythonwin still has a PIL demo in it. 
  To draw the lines you would then need to drop back to the Windows GDI API.

Mark.




From max at alcyone.com  Tue Apr  1 17:29:33 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 01 Apr 2003 14:29:33 -0800
Subject: The Whitespace Language
References: <283adf56.0304011143.4f4a6934@posting.google.com>
Message-ID: <3E8A12CD.6194A9A8@alcyone.com>

Kevin Dahlhausen wrote:

> I'm getting so much entertrainment out of the the joke on the jokers
> that created the Whitespace Language page -since we use indentation to
> delimit block strucutre, white-space has been part of Python's syntax
> since the beginning.  Ha!

Yeah, I found that slightly amusing when I read the blurb on Whitespace
as well.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ A good indignation brings out all one's powers.
\__/ Ralph Waldo Emerson
    Esperanto reference / http://www.alcyone.com/max/lang/esperanto/
 An Esperanto reference for English speakers.



From opengeometry at yahoo.ca  Tue Apr  1 01:37:59 2003
From: opengeometry at yahoo.ca (William Park)
Date: 1 Apr 2003 06:37:59 GMT
Subject: converting from shell script to python
References: <kfn0jaoig1.fsf@shaoz.activesky.com.au>
Message-ID: <b6bc47$3enbd$1@ID-99293.news.dfncis.de>

Shao Zhang <shao at cia.com.au> wrote:
> Hi,
> 
> I need to convert a very simple shell script to python in order get it
> run on both windows and unix. However I cannot quite work out a simple
> way of doing it without using lists/array/join etc...
> 
> #!/bin/sh
> ROOTDIR=/usr/lib
> CLASSPATH="$ROOTDIR/a/a.jar:$ROOTDIR/b/b.jar:$ROOTDIR/c/c.jar"
> java -classpath $CLASSPATH foo
> 
> What is the most python way of doing the above?

Do what?  You are calling Java, and that would be OS issue, ie.
    os.system(...)
If you're talking about variable substitution, then 
    ROOTDIR = "/usr/lib"
    CLASSPATH = "%s/a/a.jar:%s/b/b.jar:%s/c/c.jar" % (ROOTDIR, ROOTDIR, ROOTDIR)
is one solution.  Though, why you don't want to use list/array/join
would be another topic.


> Will perl be a better
> choice for things like this?
> 
> Thanks.
> 
> Shao.

-- 
William Park, Open Geometry Consulting, <opengeometry at yahoo.ca>
Linux solution for data management and processing. 



From danb_83 at yahoo.com  Wed Apr  9 03:51:04 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 9 Apr 2003 00:51:04 -0700
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com> <mailman.1049836334.25959.python-list@python.org>
Message-ID: <ad052e5c.0304082351.6cc33feb@posting.google.com>

"Mike Rovner" <mike at bindkey.com> wrote in message news:<mailman.1049836334.25959.python-list at python.org>...
> def split(S,N):
>   """S - source string
>      N - number of parts
>   """
>   lens=len(s)
>   l=lens/N

Don't you mean lens//N ?

>   return [S[i:i+l] for i in xrange(0,lens,l)]



From mwh at python.net  Wed Apr 16 17:30:17 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 21:30:17 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr>
Message-ID: <7h3y92ap1uk.fsf@pc150.maths.bris.ac.uk>

laotseu <bdesth at removethis.free.fr> writes:

> Paul Foley wrote:
> > On Tue, 15 Apr 2003 23:49:37 +0000, laotseu  wrote:
> > 
> >>Even for OO programmers, functionnal features in Python are IMHO a
> >>great plus, and BTW functionnal and OO paradigm does not have to
> >>conflict (that would be functionnal vs imperative and object vs
> >>procedural). CLOS is one of the great system objects out here, and
> >>it's been implemented on top of a functionnal language,
> > Oh yes?  Which functional language would that be?
> > 
> 
> CLOS means Common Lisp Object System.

Paul knows that :-) I'd wager the point he was making is that Common
Lisp isn't "really" a functional programming language...

Ocaml would be a better example, I'd have thought.

Cheers,
M.

-- 
  In general, I'd recommend injecting LSD directly into your temples,
  Syd-Barret-style, before mucking with Motif's resource framework.
  The former has far lower odds of leading directly to terminal
  insanity.                                            -- Dan Martinez



From alessandro.bottoni at infinito.it  Sun Apr  6 06:03:28 2003
From: alessandro.bottoni at infinito.it (Alessandro Bottoni)
Date: Sun, 6 Apr 2003 10:03:28 +0000
Subject: What did happen to PyXPCOM?
Message-ID: <200304061003.28906.alessandro.bottoni@infinito.it>

I recently visited the ASPN page devoted to PyXPCOM and discovered that the 
PyXPCOM tar.gz/zip files are not on the web site any more. The most obvious 
explanation is that PyXPCOM is now hosted by the cvs of Mozilla, as promised. 

But...

- Which releases of Mozilla contain PyXPCOM? Which was the first release of 
Mozilla that included PyXPCOM? Do _all_ currently available versions of 
Mozilla include it or just the "main" 1.0.X trunk?

- PyXPCOM is available in Mozilla _binary_ distributions or I have to downoald 
a source distribution and recompile it with some specific switch turned on 
(something like "--with-pyxpcom")?

Many thanks in advance for your attention.
 
---------------------
Alessandro Bottoni






From bdesth at removethis.free.fr  Tue Apr  1 08:46:32 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 01 Apr 2003 08:46:32 -0500
Subject: Parsing C [OT]
References: <Pine.GSO.4.44.0303311743400.19083-100000@amy4.Stanford.EDU>
Message-ID: <3E899838.8010500@removethis.free.fr>

Daniel Timothy Bentley wrote:
> Hey,
> I find myself having to write a C interpreter in Jython.

<mode="joke">
Writing a C interpreter for a Python interpreter implemented on top of a 
Java interpreter which is probably at least partially written in C !

Wow !

No Lisp or Smalltalk layer planned ?-)
</mode>

Laotseu




From corbet-python at lwn.net  Thu Apr 10 12:36:27 2003
From: corbet-python at lwn.net (Jonathan Corbet)
Date: Thu, 10 Apr 2003 10:36:27 -0600
Subject: SYSV IPC module maintainer?
Message-ID: <20030410163627.6866.qmail@eklektix.com>

I'm using the Python module for SYSV shared memory and semaphores; it's
part of a twisted bit of code that lets me run Oasis banner ads on
LWN.net.  The module is great, but, as it turns out, it's also responsible
for a memory leak I've been chasing for quite some time.  So I tried
dropping a note to the author (Vladimir Marangozov) with a patch, but it
came bouncing right back to me.

So...my question is, is anybody active maintaining this module at this
point?  If not, maybe I'll put up a patched version somewhere so others
don't have to go through the same process.

Thanks,

jon

P.S.  I'll attach the patch, in case it's useful to anybody.

Jonathan Corbet
Executive editor, LWN.net
corbet at lwn.net


--- shmmodule.c.~1.1.~	2002-05-29 10:00:59.000000000 -0600
+++ shmmodule.c	2003-04-10 10:13:56.000000000 -0600
@@ -1101,6 +1101,7 @@
 	}
 	Py_DECREF(o);	/* the owned reference in shm_dict doesn't count! */
     }
+    Py_DECREF(keyo);
     /* set up the status data */
     if (shmctl(o->shmid, IPC_STAT, &(o->ds)) == -1) {
 	Py_DECREF(o);




From exarkun at intarweb.us  Wed Apr  2 10:39:32 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 10:39:32 -0500
Subject: Best practice for object attributes?
In-Reply-To: <b6eusn$avn$1@nntp0.reith.bbc.co.uk>
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk> <3E8AEC44.2090909@removethis.free.fr> <b6eusn$avn$1@nntp0.reith.bbc.co.uk>
Message-ID: <20030402153932.GA4884@meson.dyndns.org>

On Wed, Apr 02, 2003 at 04:12:36PM +0100, Michael Sparks wrote:
> [snip]
> 
> I'm aware of the meaning of self :) and know that the first form gives
> you class attributes. The reason I was asking is because I'd noticed
> the following interesting behaviour, which if _intentional_ is useful.
> 
> >>> a=shrubbery1()
> >>> b=shrubbery1()
> >>> a.name,b.name,shrubbery1.name
> (None, None, None)                  # a.name, b.name exist - presumed
>                                     # references to shrubbery1.name
> >>> a.name="hello"
> >>> a.name,b.name,shrubbery1.name
> ('hello', None, None)               # Bzzzt! a.name seems to be
>                                     # independent
> >>> b.name="brian"
> >>> a.name,b.name,shrubbery1.name
> ('hello', 'brian', None)            # Hmm. So is b.name
> >>> shrubbery1.name="Surprise"
> >>> a.name,b.name,shrubbery1.name
> ('hello', 'brian', 'Surprise')      # Changing class leaves existing
>                                     # instances unaffected.
> >>> a=shrubbery1()                  # Get 2 fresh copies.
> >>> b=shrubbery1()
> >>> a.name,b.name,shrubbery1.name
> ('Surprise', 'Surprise', 'Surprise') # a.name, b.name cloned from
>                                      # shrubbery1.name
> 
> As far as I can tell the behaviour seems to be that you automatically
> gain instance attributes that clone the state of the class attributes
> at object creation time, which are then independent and can be modified
> via the usual self.<attr> syntax later on. (They cannot however be
> modified via the self.__dict__[<attr>] syntax since they're not stored
> there)
> 

  Not quite.  What really happens is simpler.  When an attribute lookup on
an instance fails, the lookup proceeds to the instance's class object (if it
fails there, it proceeds to the class object's bases, and so on).  The
attributes aren't cloned, and the time of creation matters not.  Here is an
example:

    class Foo:
        attr = 'String'

    f = Foo()
    Foo.attr = 10
    print f.attr


  You are still right that this behavior is useful. >:)  One important thing
to note, though, is that this is almost -never- what you want when the
attributes refer to mutable objects.  For example:

    class Foo:
        attr = []
        def update(self, value):
            self.attr.append('It is a value: %r' % (value,))

    f = Foo()
    g = Foo()
    g.update(f)
    print f.attr

  As you can see, f and g (and indeed Foo) are sharing the same list object. 
Sometimes this is what you want, but more frequently it is not, and your
only choice is to initialize the attribute to a new list object in some
instance method, usually __init__.

  Hope this helps,

  Jp

-- 
        "I quite agree with you," said the Duchess; "and the moral of
that is -- Be what you would seem to be' -- or, if you'd like it put
more simply -- Never imagine yourself not to be otherwise than what it
might appear to others that what you were or might have been was not 
otherwise than what you had been would have appeared to them to be
otherwise.'"       -- Lewis Carrol, "Alice in Wonderland"
-- 
 up 13 days, 12:00, 3 users, load average: 0.01, 0.02, 0.00




From aleax at aleax.it  Mon Apr 28 02:48:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 06:48:46 GMT
Subject: Date validation
References: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au> <mailman.1051257039.22404.python-list@python.org> <38ec68a6.0304272054.6e4d2eb2@posting.google.com>
Message-ID: <i94ra.13664$3M4.372132@news1.tin.it>

Asun Friere wrote:
   ...
> While checking this out, I noticed that calender.py in 2.3a2 imports
> 'datetime,'
> for which there is no python file.  Is this now the wrapper to the
> various clib implementations?

Module datetime, new in 2.3, is a C-coded extension dealing with
date and time computations, and it will hopefully be helpful in
making Python date and time handling less dependend on the quirks
of the underlying C runtime libraries.


Alex




From mertz at gnosis.cx  Tue Apr  1 13:50:13 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 01 Apr 2003 13:50:13 -0500
Subject: Terminal-based text editor with color syntax hilighting
References: <slrnb8hvka.pki.ramen@lackingtalent.com>
 <mailman.1049169002.28920.python-list@python.org>
 <slrnb8j93q.421.ramen@lackingtalent.com>
Message-ID: <l9di+kKkXgKL092yn@gnosis.cx>         

Dave Benjamin <ramen at lackingtalent.com> wrote previously:
|Are there any others? What about Fte, I wonder... is that editor/IDE still
|being actively maintained? According to their website it supports Python
|color hilighting, but I'm not sure if the s-lang version does. I'll check
|into it.

Unfortunately, AFAIK FTE only works in Unix-like systems over XTerm.  A
console session doesn't do FTE (under Linux, OS/2 has both console and
PM versions of FTE).  As with the OP, this rules out using FTE on most
of my connections--my web host doesn't run an X client, my OS/2, Windows
and MacOSX machines don't run an X server).

IMO, the best option for the rather common scenario of remote SSH
connection is Jed. But I would welcome learning of others.  I *have*
been using vim on one host recently, and have kinda-sorta learned basic
navigation, but it won't stick in my head after a month (I know from
experience).  Likewise with emacs.

Jed has this wonderful feature of having topbar menus that actually let
you visually select the most common actions (without needing to remember
a cryptic key-sequence).  This fabulous design had only been in DOS
text editors since 1982, so it's nice that Linux is starting to catch up
with that.

Yours, Lulu...

--
    _/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: Postmodern Enterprises _/_/_/
   _/_/    ~~~~~~~~~~~~~~~~~~~~[mertz at gnosis.cx]~~~~~~~~~~~~~~~~~~~~~  _/_/
  _/_/  The opinions expressed here must be those of my employer...   _/_/
 _/_/_/_/_/_/_/_/_/_/ Surely you don't think that *I* believe them!  _/_/





From romany at actimize.com  Tue Apr 29 03:59:59 2003
From: romany at actimize.com (Roman Yakovenko)
Date: Tue, 29 Apr 2003 10:59:59 +0300
Subject: Visual Basic - Python compare
Message-ID: <91BFE89EFFA2904E9A4C3ACB4E5F2DF5027AD4@exchange.adrembi.com>

    Hi. I wrote Python - VB compare. 
Preface for compare: we have client - server application, 
where server is written in C++ and client consist of 2 part. Part 1 is
wrapper between C++ classes 
to COM objects. Part 2 is GUI written in VB 6.0. Now we want to migrate
from VB 6.0 to VB Net.
The only reason for migration is: VB Net modern environment. ( The
program is going to be fully rewritten. 
We have to do it, so we may choose a different development environment
). I decided to propose 
python instead of VB. For this purpose I wrote small VB-Python compare.
There are a few question 
in the compare. Also any comments are welcome. If I missed something
please report me. Also I would like
To hear about problems that exist( I believe they do ) for writing
application where server side is written in C++
and client side in python.

Compare.

        Why Python is better then Visual Basic for me.

    Python is fully object oriented language (like C++). There are a few
power 
futures in this language: generators, clear syntax.

    XML RPC lib. XML-RPC is a "Remote Procedure Call" method that uses
XML 
passed via HTTP as a transport. With it, a client can call methods with 
parameters on a remote server (the server is named by a URI) and get
back
structured data. This module supports writing XML-RPC client code; it
handles 
all the details of translating between conformable Python objects and
XML on the 
wire. Also client side and server side can be written in different
languages. It 
means that there is no need in wrapper of C++ classes for the client
side. 

Question: XML RPC lib use url in order to establish connection. In my
case I have 
a little bit different requests: client - written in python, server
written in C++, 
connection made over TCP IP. Also I'd like to see secured connection.

    Boost.Python lib - C++ library which enables seamless
interoperability 
between C++ and the Python. It means there are no more complicated
wrappers. 
Example:  
    struct World
    {
        World(std::string msg): msg(msg) {} // added constructor
        void set(std::string msg) { this->msg = msg; }
        std::string greet() { return msg; }
        std::string msg;
    };
   
    enum choice { red, blue };

        In order to expose this class to VB I need:
    1. To use VC++ wizard. There are 3 files this wizard adds, also it
touches idl, 
        resource and registration files. 
    2. After this I need to describe the functionality of this class. It
means every call I need to redirect to 
       internal object. The problems I meet here are: different data
types definitions. VB True == -1, C++ true == 1. 
       It means that if you don't insert right value then boolean test
in VB may faild. BSTR may be NullString, EmptyString. 
       During  redirection of calls it is possible to make a few bugs
(I suppose that copy + paste make a good job in this area). 
    3. VC++ idl compiler can't work with C++ build-in enums and
constants. It means that almost every const in the system 
       is duplicated. There were a few bugs in this area also. Developer
changed original enum value in C++ code without 
       changing it in idl. 
    4. There is no easy way to path errors to the calling object. If you
want to pass errors you need some global variable to keep 
        error description their. The only allowed error description is:
error id,  error message, help file. It is an ugly way to treat errors.

        In order to expose this class to python I need to write only
this
    
        class_<World>("World", init<std::string>())
            .def(init<double, double>())
            .def("greet", &World::greet)
            .def("set", &World::set);
         
        enum_<choice>("choice")
            .value("red", red)
            .value("blue", blue);
        
    Export\import from (to) file is simpler - there is no more code that

translate object to string and string to object - python can handle it
itself. 
There were a few bugs in this area also, like forgotten property or
encoding 
problems.
        
    Every file is executable file - it means you may write testing code
within file. 
Also there is a "UNI Test" for python. 

    GUI - there are 5 or more different gui libraries for python: 
-Tk( build-in )
-wxWindows
-Qt (may be the best advertising of this library is KDE)
-GTK (may be the best advertising of this library is GNOME)
 
    Python could work with COM.

    IDE - there are a lot of them, some of them are free and good. Also
there is 
environment for Visual Net.

    There are also a few installers for python that makes executable
(exe) from 
sources. If you do it you don't have to worry weather python installed
on user 
computer or not.

Thanks for your attention. Any comments are apriciated and welcom.

Roman





From just at xs4all.nl  Wed Apr  9 13:38:16 2003
From: just at xs4all.nl (Just)
Date: Wed, 09 Apr 2003 19:38:16 +0200
Subject: no Python mail archiving utility?
References: <joe-17FB5F.10302009042003@orngca-news02.socal.rr.com>
Message-ID: <just-6A27B1.19381609042003@news1.news.xs4all.nl>

In article <joe-17FB5F.10302009042003 at orngca-news02.socal.rr.com>,
 Joe Strout <joe at strout.net> wrote:

> Yesterday morning I posted asking where to find a web mail 
> archive/search program written in Python.  I got nothing but the sound 
> of chirping crickets in response.
> 
> Does this mean that no such utility exists?  Are all the dozens of 
> mailing lists about Python-related topics serving up their archives via 
> programs written in C and Perl?

Mailman 2.0.x comes with pipermail, which is pure Python, but real basic 
(no pun intended ;-), it only does archiving, not searching (the latter 
is no problem if the archives are public, since Google will pick them 
up). Don't know if Mailman 2.1 ships the same or something better. I 
think there are some Python-based indexing tools out there.

Just



From jennyw at colorfulexpressions.com  Tue Apr  1 11:08:45 2003
From: jennyw at colorfulexpressions.com (jennyw)
Date: 1 Apr 2003 08:08:45 -0800
Subject: Trying to run Boa Constructor on Windows
References: <327f0ccd.0303311507.2ed2d53e@posting.google.com> <mailman.1049154948.9372.python-list@python.org>
Message-ID: <327f0ccd.0304010808.206c7bd7@posting.google.com>

Cliff Wells <LogiplexSoftware at earthlink.net> wrote in message news:<mailman.1049154948.9372.python-list at python.org>...

> Get the latest wxPython (2.4.0.7, IIRC) and get Boa Constructor from
> CVS.  There's been a lot of changes the last couple of weeks for both
> packages.

In the end, that's what I ended up doing. I was originally hesitant to
get stuff off CVS because I thought it might be unstable, but it seems
to be working well.

Thanks!

Jen



From martin at v.loewis.de  Sat Apr 12 04:49:40 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Sat, 12 Apr 2003 10:49:40 +0200
Subject: intern'ed strings and deepcopy()
In-Reply-To: <yfs8yugfsl0.fsf@black132.ex.ac.uk>
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk>
Message-ID: <3E97D324.8000009@v.loewis.de>

Alexander Schmolck wrote:
> But I'm not sure that that's what the OP's up to. 

The original poster asked (paraphrased)
"Am I right assuming that interned strings are not copied
by deepcopy?"

to which the short answer is

"Yes, you are right"

The slightly longer answer is

"Yes. Strings are never copied by deepcopy since deepcopy
knows they are immutable. Whether they are interned or not
is irrelevant".

When Eric Max said "its identity is unimportant", he really
meant "it's unimportant whether they are interned, and their
identity is preserved in deepcopy" :-)

Regards,
Martin




From mertz at gnosis.cx  Thu Apr  3 14:22:11 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Thu, 03 Apr 2003 14:22:11 -0500
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com>
 <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>
 <9383481e.0304030705.524cf0fa@posting.google.com>
Message-ID: <jnIj+kKkXQwc092yn@gnosis.cx>         

drew at astro.pas.rochester.edu (Drew Moore) wrote previously:
|voltage += 5  # (raise the voltage by 5 volts)
|voltage *= 2  # (double the current voltage)
|voltage = 3  # (set the voltage to 3 volts)
|my voltage object is clobbered with an integer object.

The case makes sense... but IMO, not nearly enough to change the
behavior of assignment.

For this particular case, you could simply coopt the meaning of some
other augmented operator to set voltage.  For example:

    voltage |= 3

The need to bitwise-and on voltage seems unlikely, so you should not
step on anything important.  Of course, experienced Python programmers
might do a little double take on the meaning, but documentation can
explain it.

Then again, I wouldn't mind adding the new operator ":=" which was
"generic augmentation"... the operator could do -absolutely nothing- for
standard types, but could be available for custom meaning in custom
types.

Yours, Lulu...

--
    _/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: Postmodern Enterprises _/_/_/
   _/_/    ~~~~~~~~~~~~~~~~~~~~[mertz at gnosis.cx]~~~~~~~~~~~~~~~~~~~~~  _/_/
  _/_/  The opinions expressed here must be those of my employer...   _/_/
 _/_/_/_/_/_/_/_/_/_/ Surely you don't think that *I* believe them!  _/_/





From peter at engcorp.com  Fri Apr 25 09:14:28 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 09:14:28 -0400
Subject: Char to ASCII and back ?
References: <roy-CF32FA.11400525042003@malibu.unice.fr>
Message-ID: <3EA934B4.6D740940@engcorp.com>

Jean-Paul Roy wrote:
> 
> How do you transform a char (or a one-char string) into the integer
> ascii code and back ? Say to code a message.

Others have answered the basic question, but I'm puzzled by your
example use case.  What do you mean, "to code a message"?  In
many cases, you would just want to send the string itself... it's
just a sequence of bytes in memory.  For example, if you are doing
socket communication, or talking through a serial port, you don't
do anything special to strings, you just send them as-is.

Or is it really just curiosity, in which case I can shut up again. :-)

-Peter



From max at cNOvSisiPonAtecMh.com  Thu Apr  3 13:49:43 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Thu, 03 Apr 2003 18:49:43 GMT
Subject: computer science with python
References: <m3istvtodl.fsf@moroni.pp.asu.edu>
Message-ID: <bn%ia.3718$n95.1212398@twister.nyc.rr.com>

http://www.tcfb.com/freetechbooks/bookphyton.html
see "how to think like a computer sci..." and some others

--
========================================
Max Khesin, software developer -
max at cNvOiSsPiAoMntech.com
[check out our image compression software at www.cvisiontech.com, JBIG2-PDF
compression @
www.cvisiontech.com/cvistapdf.html]


"David Bear" <david.bear at asu.edu> wrote in message
news:m3istvtodl.fsf at moroni.pp.asu.edu...
>
> I purchased the OReilly book "Computer Science with Perl" (or
> something similar and found it very fun -- even though I really hate
> Perl.  It made me think that python would be a terrific language to
> demonstrate some key computer science concepts.
>
> Anyone know if a book has already been done?  I find that many CS
> language concepts are debated/discussed here and have to be careful
> when I look at this group because I become too interested in reading
> all the conceptual stuff rather than getting a specific question
> answered.  Seems this list would provide a good starting point for
> such a book.  I image it could be many volumes...





From rmunn at pobox.com  Wed Apr 16 15:09:12 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 16 Apr 2003 19:09:12 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org>
Message-ID: <slrnb9qlrf.7eb.rmunn@localhost.localdomain>

Gerhard H?ring <gh at ghaering.de> wrote:
> * Dan Jacobson <jidanni at dman.ddts.net> [2003-04-15 13:59 +0800]:
>> I need to use one liners every day, especially in Makefiles.  I thus
>> need to know how to make any kind of python program into a big long one
>> liner. However, right out of the starting gates I run into
>> $ python -c 'print 2;for i in (1,4):print i'
>>   File "<string>", line 1
>>     print 2;for i in (1,4):print i
>>               ^
>> SyntaxError: invalid syntax
> 
> Try:
> 
> python -c """
> print 2
> for i in (1, 4):
>     print i
> """
> 
> It seems to work under a bash script at least. I once did something with
><<EOF but have since forgotten the little knowledge I had a about shell
> scripting ;-)

If you're in a shell script, you could do:

----- Cut here -----
# ... do some shell script stuff ...
rm temporary_script.py  # Make sure there isn't a symlink lying around
cat >temporary_script.py <<EOF
# Start of Python code
import math
print math.pi
for i in range(5):
    print i
print "All done"
EOF
python temporary_script.py
rm temporary_script.py
# ... more shell script stuff ...
----- Cut here -----

Notice the first "rm temporary_script.py" to make sure someone hasn't
"carelessly" left a symlink lying around pointing to a file like
/etc/passwd that you wouldn't want overwritten. There is still a race
condition here -- if someone manages to create a symlink between the rm
and the cat, you could be in trouble -- but it would be tricky to
exploit, and you could do away with the problem entirely by using random
(and therefore unpredictable) filenames for your temporary files.

This way you clean up the directory after you're done, and you don't
need to have a ton of small scripts lying around; you can consolidate
them all into one big file. This is the approach used by the GNU
autoconf suite -- take a look at the source of configure sometime.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From sholden at holdenweb.com  Wed Apr 16 09:57:52 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 13:57:52 GMT
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <b7ih78$fb2$1@panix1.panix.com>
Message-ID: <Ajdna.110595$It5.16632@news2.central.cox.net>

"Aahz" <aahz at pythoncraft.com> wrote in message
news:b7ih78$fb2$1 at panix1.panix.com...
> In article <slrnb9lsbe.4pa.marcus at infa.abo.fi>,
> Marcus Alanen <marcus at infa.abo.fi> wrote:
> >
> >On the other hand, I've always wished for a document "language X for
> >people know language Y / programming in general / think they know". We
> >could skip all that "let's hold hands while I take you though the
> >mysteries of a dictionary" nonsense, and just concentrate on the issue
> >at hand: how does our knowledge from language X map to this new
> >language? What pitfalls do we have in the language at the moment
> >(e.g. python default arguments are in my humble opinion still
> >counter-intuitive) ?
>
> Right now, I'm trying to get some momentum moving on a book titled
> _Python for VB6 Programmers_.  Anyone interested should subscribe to
> marketing-python.
>
> http://wingide.com/mailman/listinfo/marketing-python

You'd do better to call it "Python for VB programmers", expecially since the
change to VB6 is so radical that migrating to a new and more sensible
language might be an attractive option to some!

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From staschuk at telusplanet.net  Wed Apr  9 23:01:22 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 9 Apr 2003 21:01:22 -0600
Subject: a size-intelligent progress bar?
In-Reply-To: <1PYka.1140$RX1.361@newsfep3-gui.server.ntli.net>; from tomchance@gmx.net on Wed, Apr 09, 2003 at 06:20:02PM +0100
References: <1PYka.1140$RX1.361@newsfep3-gui.server.ntli.net>
Message-ID: <20030409210122.A2493@tibia.amotlpaa.bogus>

Quoth Tom Chance:
> I'm making a shell-based program that includes a progress bar (like wget's
> one). At the moment, I've got it printing an equals sign every 2% so I get
> a progress bar 50 "="s wide, which fits my Konsole perfectly. This isn't
> very handy for other shell sizes though :)
> 
> Does anyone have, or can anyone point me in the direction of a way of making
> a decent progress bar that will automagically detect the width of the shell
> open, even resizing it mid-"progress"?

I think the way to determine the width of the current terminal is
terminal-dependent.  You could use curses.

If memory serves, when a window gets resized the process(es?)
associated with it gets sent a WINCH signal.  I don't know whether
curses traps this for you; if so, just keep rechecking the
terminal width via curses at appropriate moments.

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From martin at v.loewis.de  Fri Apr 25 04:04:55 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 10:04:55 +0200
Subject: Problem getting zlib.o built on solaris 8 system
In-Reply-To: <mailman.1051199933.5003.python-list@python.org>
References: <mailman.1051199933.5003.python-list@python.org>
Message-ID: <3EA8EC27.8070000@v.loewis.de>

Reed, Judith wrote:
> Text relocation remains                         referenced
> 
>     against symbol                  offset      in file
> 
> I have zlib 1.3.1 installed, so I am baffled as to what is going on 
> here. Can anyone make any suggestions?

Your zlib installation does not feature position-independent code (PIC).
You either need to reinstall zlib with PIC, or you need to incorporate 
the zlib module statically into Python. I recommend to do the latter.
To do so, edit Modules/Setup.

Alternatively, remove your zlib installation, and use the one provided 
by Sun, in /usr/lib.

Regards,
Martin




From xxdigitalhellxx at hotmail.com  Tue Apr 15 10:29:35 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Tue, 15 Apr 2003 08:29:35 -0600
Subject: Speeding up: s += "string"
In-Reply-To: <v9ntbr3ioli8d3@corp.supernews.com>
References: <mailman.1050393203.4704.python-list@python.org> <UoPma.20694$T34.658757@news2.tin.it> <v9ntbr3ioli8d3@corp.supernews.com>
Message-ID: <b7h53f$cf3$1@heckle.cc.uregina.ca>

Cameron Laird wrote:
> I'm not optimistic.  I'll be more precise:  the problem
> is real, and good solutions are sensitive to data char-
> acteristics.  It's easy to improve performance of the
> s1 += s2 operations that characterize e-mail filtering
> while simultaneously degrading unacceptably the same
> operation embedded in an XML filter.  I am close to
> asserting the conjecture that every solution harms too
> many existing legitimate users.

Just adding my two cents..

People should keep in mind that s += a is shorthand for s = s + a. If you 
look at it the second way you can clearly see that a is appending to s and 
then s is *assigned* that value. In both cases two operations are being done.

Perhaps we could introduce .= as a true assign/append operation. Although 
this would still be a mutable operation, which contradicts the documents.

-- 
Ryan




From stuart at bmsi.com  Wed Apr  2 16:35:53 2003
From: stuart at bmsi.com (Stuart D. Gathman)
Date: Wed, 02 Apr 2003 16:35:53 -0500
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <ZDHha.4658$yh1.618117@news1.east.cox.net> <v8fa11rboem327@corp.supernews.com>
Message-ID: <ZIIia.2942$BL2.49486@news.uswest.net>

On Sun, 30 Mar 2003 21:35:04 -0500, Garen Parham wrote:

> Stuart D. Gathman wrote:
> 
> 
>>   1) no "safe" mode with GC, where memory overwrites, leaks, and
>> corruption are impossible.
>> 
>> 
> Sure there is, though not part of the language proper, the Boehm GC
> works great.

That provides conservative GC, but overwrites and corruption are still a
problem.

>>   2) Compiler is slow - even on 2.6Ghz Xeon.  Using header caching
>> compilers is complex. Extensive template use makes compiler slow - but
>> they are essential to make C++ programming as "safe" as possible.
> 
> This part isn't any inherent property of the language.  If your
> experience is based on using g++ I could understand why you'd think
> this. C++ is pretty difficult to parse though - but I've used plenty
> that are real fast; dmc++ being the fastest.

My experience is based on g++, bcc, and xlc (IBM).

> Some good info, but I think you're conflating C and C++ here.  The
> double-free issue is a mostly C thing (unnecessary in C++).  In C++ it's
> also common to use smart pointers or other library facilities to manage
> memory.

I use the tools you mention extensively, and they are a great help, but do
not by any stretch prevent double free and other memory management issues.
 You can try to be disciplined, but eventually you flub up and get
confused about which other object owns a given object.  As a discipline,
you could avoid all object sharing, and thus let C++ destructors handle it
all - but this introduces serious performance problems (might as well use
python).  I generally start out with no object sharing, then introduce
sharing for performance bottlenecks until it is fast enough.  Each shared
object is a potential memory problem.

Using STL iterators helps avoid most buffer overruns, but again does not
make them impossible.  The advantage of Java (and to a lesser extent
python) is that these things are enforced at the virtual machine level.

Smart pointers that do reference counting are slower than just using
Boehms collector for smaller objects.

-- 
Stuart D. Gathman <stuart at bmsi.com>
Business Management Systems Inc.  Phone: 703 591-0911 Fax: 703 591-6154
"Confutatis maledictis, flamis acribus addictis" - Mozart background
song for the Microsoft "Where do you want to go from here?" commercial.



From Peter.Havens at Level3.com  Fri Apr 11 12:52:36 2003
From: Peter.Havens at Level3.com (Havens, Peter)
Date: Fri, 11 Apr 2003 10:52:36 -0600
Subject: Confusion: assignments create object references.
Message-ID: <0C642BE4EE085149B52839125D4FC9D422BDA0@idc1exc0002.corp.global.level3.com>

// -----Original Message-----
// From: Jp Calderone [mailto:exarkun at intarweb.us]
//
// > So far so good. This works as I would expect given the rule above.
// > 
// > >>> L = 0
// > >>> M = L
// > >>> L
// > 0
// > >>> M = 1
// > >>> L
// > 1
// > >>>
// > 
// 
//   The above never happened, I assure you. >:)    Python 
// doesn't work this way.

I'm a silly rabbit. I hand copied the test incorrectly. The real test was:

>>> L = 1
>>> M = L
>>> L
1
>>> M = 0
>>> L
1
>>>

Those are the real results (I double checked), but now I'm really confused. Why does the former test (the one I lied about) above give expected results, and the later doesn't?




From jcao at cs.uchicago.edu  Thu Apr 17 23:12:46 2003
From: jcao at cs.uchicago.edu (Jing (George) Cao)
Date: Thu, 17 Apr 2003 22:12:46 -0500 (CDT)
Subject: question on buildinh /demo/embed
Message-ID: <Pine.GSO.4.52.0304172209090.9836@gargoyle.cs.uchicago.edu>

I try to build the example in /Demo/embed, but there is several compiling
errors. It is said in README that I should modify $LIBS as in
/Modules/Makefile.  But I can't find any Makefile in the /Modoles
directory.

I am lost here, anybody can help me?

Thanks,
--Jing




From martin.vonloewis at hpi.uni-potsdam.de  Tue Apr  8 05:19:34 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Tue, 8 Apr 2003 11:19:34 +0200
Subject: Updating CVS
References: <5Nhka.139$N5.25540@mencken.net.nih.gov> <pan.2003.04.07.19.52.22.351156.788@thomas-guettler.de> <l5jka.141$N5.25497@mencken.net.nih.gov>
Message-ID: <b6u476$h2g$1@zeppelin.rz.uni-potsdam.de>

> The line "sys.stderr = sys.stdin" redirects any errors to the browser. My
> question is less cgi and more cvs. I guess I didn't state it clear enough.
I
> just want to know why the previous script doesn't update my CVS directory.
> Is it because I trying to run in from a web script and not as myself being
> logged in?

Your question was clear, but you have not provided sufficient information
to answer it. You probably didn't because you did not have them yourself;
Thomas was suggesting mechanisms to collect more information.

The most likely cause of the problem is that cvs is asking for a password
which nobody enters.

Regards,
Martin





From philh at cabalamat.org  Wed Apr  2 15:11:02 2003
From: philh at cabalamat.org (phil hunt)
Date: Wed, 2 Apr 2003 21:11:02 +0100
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com> <b6esse$pge$1@heckle.cc.uregina.ca> <pan.2003.04.02.15.14.54.788223@mn.rr.com>
Message-ID: <slrnb8mgum.f98.philh@cabalamat.uklinux.net>

On Wed, 02 Apr 2003 15:13:45 GMT, Anna <revanna at mn.rr.com> wrote:
>>> OK cleverclogs, how do we write 39?
>
>XXXIX
>
>I think what you need here is:
>3. At most three Is, Xs or Cs may appear *consecutively* in any given
>literal. 

The Romans generally preferred not to use the subtractive principle, 
hence: XXXVIIII

-- 
Philip Hunt <philh at cabalamat.org>

Interested in adventure holidays in Spain? 
Look at: <http://www.cabalamat.org/advcon/>



From rmunn at pobox.com  Thu Apr 10 13:10:20 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 10 Apr 2003 17:10:20 GMT
Subject: atexit functionality
References: <mailman.1049821814.20582.python-list@python.org> <10074955979500@aluminium.rcp.co.uk> <mailman.1049914516.26770.python-list@python.org>
Message-ID: <slrnb9akld.bom.rmunn@localhost.localdomain>

viszneki at telerama.com <viszneki at telerama.com> wrote:
> Thanks for playing Duncan. You know I was in good spirits and happy to 
> contribute something until you and ol Stevey Holden decided to lecture me on 
> Internet Etiquette, and tell me how much of a jerk I was to not have time to 
> sign up for every mailing list I'd like to contribute something to. Gee I'm 
> such a jerk.

Steve (NOT Stevey) Holden was right, you know. Not that you are a jerk
-- he never said that -- but that it's good etiquette to ask for
*public* responses to your question, not private ones. See

    http://www.catb.org/~esr/faqs/smart-questions.html

and read the section entitled "Don't ask people to reply by private
email" for more details. Summary: Other people might benefit from the
answer to your question, so it should be answered on the mailing list.
It's rude to deny others the benefit of seeing those answers just for
your personal convenience.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From bignose-hates-spam at and-zip-does-too.com.au  Mon Apr  7 20:07:56 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Tue, 08 Apr 2003 00:07:56 GMT
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk>
Message-ID: <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake>

On Mon, 7 Apr 2003 10:45:58 GMT, Michael Hudson wrote:
>  Ben Finney <bignose-hates-spam at and-zip-does-too.com.au> writes:
> > Even if I [ensure a package is in the PYTHONPATH] on my computer,
> > it's not going to be so on someone else's computer when they
> > download my package tree.
>  
>  It is a Fact Of Life that Python packages (and modules therein) need
>  to know where they are in the global module namespace.

Thanks to all who have offered this; the verdict seems to be unanimous.

My question, then, becomes this:  How does one gracefully ensure Python
can find the (potentially) myriad top-level directories under which
one's in-development projects reside?

Is freedom in choosing the directory structure -- such a useful tool in
maintaining similar copies of a directory tree during development --
lost to the need for an ever-changing PYTHONPATH, or, worse,
hand-altering package names within the Python scripts whenever things
move about?

I feel dim for not being able to see the answer, but it seems absurd to
me that one can't rename a directory and expect the scripts within it to
continue working without outside help.

-- 
 \                   "The best is the enemy of the good."  -- Voltaire |
  `\                                                                   |
_o__)                                                                  |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From LogiplexSoftware at earthlink.net  Mon Apr 21 06:32:14 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 21 Apr 2003 03:32:14 -0700
Subject: Why functional Python matters - A kind of conclusion
In-Reply-To: <3EA08A7D.3010206@removethis.free.fr>
References: <mailman.1050527910.25653.python-list@python.org>
	 <3EA08A7D.3010206@removethis.free.fr>
Message-ID: <1050921134.1691.110.camel@software1.logiplex.internal>

On Fri, 2003-04-18 at 16:30, laotseu wrote:

for x in range(100):
>    print "I shall not say that Lisp is a functional language anymore"

Better yet:

for x in range(1000):
    print "I shall not invite those rude %$&!s at c.l.lisp to c.l.py anymore"

;)

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From mrsinger at usa.net  Wed Apr 16 10:57:53 2003
From: mrsinger at usa.net (Matt Singer)
Date: 16 Apr 2003 07:57:53 -0700
Subject: FTP
References: <e62b26d3.0304151727.6de2e2e4@posting.google.com> <mailman.1050460645.26676.python-list@python.org>
Message-ID: <e62b26d3.0304160657.7c31f532@posting.google.com>

Helps quite a bit... Thanks!


Nicodemus <nicodemus at globalite.com.br> wrote in message news:<mailman.1050460645.26676.python-list at python.org>...
> Matt Singer wrote:
> 
> >I have an application with a PYQT interface that uploads large files. 
> >Is there an async method to get how much data has been sent with
> >storlines to update a progress bar?  Or is there an alternate ftp
> >interface?
> >
> 
>  From the Python docs:
> 
> storlines(command, file)
>     Store a file in ASCII transfer mode. command should be an 
> appropriate "STOR" command (see storbinary()). Lines are read until EOF 
> from the open file object file using its readline() method to provide 
> the data to be stored.
> 
> So you can send any object with a readline method that support the same 
> semantics of a file (untested code):
> 
> 
> class ProgressFileReader:
>     
>      def __init__(self, filename, callback):
>          self.file = file(filename, 'r')
>          self.file_size = os.path.getsize(filename)
>          self.callback = callback # will be called with one parameter:
>                                   # the percentage of the file read so far
> 
>      def readline(self, *args):
>          result = self.file.readline(*args)
>          percent = self.file.tell()*100/self.file_size
>          self.callback(percent)
>          return result
> 
> And use it as follows:
> 
>     f = ProgressFileReader('myfile', mycallback)
>     ftpobj.storlines('STOR myfile', f)
>      
>  
> Another option would be to subclass the file object.
> 
> Hope that helps,
> Nicodemus.



From sjmachin at lexicon.net  Fri Apr  4 03:38:39 2003
From: sjmachin at lexicon.net (John Machin)
Date: 4 Apr 2003 00:38:39 -0800
Subject: split encloser
References: <435d40e0.0304031530.1c197742@posting.google.com>
Message-ID: <c76ff6fc.0304040038.4ca3d62f@posting.google.com>

aussie2010 at yahoo.com (Chris) wrote in message news:<435d40e0.0304031530.1c197742 at posting.google.com>...
> string.split() takes a delimiter and works fine as long as the
> delimiter isn't part of the data fields. But frequently they are.
> e.g. 'John Doe,135 South Main St.,#122, Springfield, Iowa' or
>       ' so long goodbye see ya'
> 
> Because the fields can contain the delimiter in some cases, an
> encloser is usually used (typically "") to handle those fields.
> 
> The above strings would be written:
> 'John Doe,"135 South Main St., #122", Springfield, Iowa'
> and 
> '"so long" goodbye "see ya"'
> 
> I don't understand regular expressions but I was wondering if anyone
> that did knew of a way to get re.split() to handle "enclosers" as used
> above.

Aussie cobber,

REs are about as useful as a hip pocket on a singlet when it comes to
data that includes the "encloser" as well. For example in a database,
yer old Auntie Mabel's address_line_1 might be

   "Emoh Ruo", 123 Smith St

but would appear as 

   """Emoh Ruo"", 123 Smith St"

in an exported file (or it should).

Get onta Dave Cole's you-beaut CSV module at:

http://www.object-craft.com.au/projects/csv/

Cheerio,
John



From janeaustine50 at hotmail.com  Mon Apr 28 19:19:41 2003
From: janeaustine50 at hotmail.com (Jane Austine)
Date: 28 Apr 2003 16:19:41 -0700
Subject: socket module's listen backlog problem
Message-ID: <ba1e306f.0304281519.1304ce5a@posting.google.com>

I'm running Python 2.3b1, which has been downloaded from python.org as
binary, on my Win XP Pro machine.

socket.SOMAXCONN is 5. I want it to be higher. However, listen with
higher backlog number doesn't work.

I'm using asyncore and removed the following code but it still doesn't
work with more than 5 backlogs.

#from asyncore
        if os.name == 'nt' and num > 5:
            num = 1

I suppose the socket module uses winsock2.h and it's SOMAXCONN is
0x7fffffff.

What's wrong? What can I do to increase listen backlog?



From exarkun at intarweb.us  Sat Apr 19 01:24:57 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sat, 19 Apr 2003 01:24:57 -0400
Subject: variable in Python
In-Reply-To: <DD4oa.113758$0X.26571104@twister.columbus.rr.com>
References: <b7lvbh$an3$1@news-reader13.wanadoo.fr> <mailman.1050586367.28668.python-list@python.org> <DD4oa.113758$0X.26571104@twister.columbus.rr.com>
Message-ID: <20030419052457.GA21954@meson.dyndns.org>

On Sat, Apr 19, 2003 at 04:53:55AM +0000, berg_ wrote:
> 
> > object it refers to; additionally, there is no concept of a "null"
> > reference.
> 
> Python does have the concept of null, it is represented by None.
> 
> cursor = None
> 

  This is merely semantic quibbling, of course - but None is an object.  Is
it often used like NULL?  Sure, but then, [] and () and '' are also often
used like NULL (They all have a false truth value, after all).  The concept
of a variable that doesn't point to anything, though, doesn't exist in
Python.  "cursor = None" binds cursor to the immutable singleton object
"None".

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 30 days, 1:02, 5 users, load average: 0.03, 0.01, 0.00




From iamshady at rediffmail.com  Thu Apr 17 06:47:21 2003
From: iamshady at rediffmail.com (Jim Shady)
Date: 17 Apr 2003 03:47:21 -0700
Subject: Parsing
Message-ID: <9afd3f36.0304170247.2a082573@posting.google.com>

Hello,

I have a file like this:

----
[tested]
101294029
101294030

----

The structure of the file is such, with different headers and varied
numbers. The blank line exists too, before the next entry starts.

What I'd like to do is run through each entry with x.readlines() and
put all 4 elements in an array for further processing. Right now I'm
trying with:

for chunk in psfile.readlines():
    if i <= 4:
        cur_chunk.append(chunk[:-2])
        i = i+1
        if i > 4:
            i = 0
            cur_chunk = []

for x in bigchunk:
    print x

This somehow doesn't work. I am a newbie to Python and would
appreciate any help.

Thanks in advance.

Regards,
Jim



From aleax at aleax.it  Fri Apr 18 15:21:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 19:21:52 GMT
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <kfYna.33951$LB6.766156@news1.tin.it>

<posted & mailed>

Bill Martin wrote:

> I'm wondering about the value of allowing a class definition like this:
> 
>              class C:
>                pass

Each instance x of such a class C is a separate, initially empty 
namespace.  You can manage the namespace by adding and removing
names, querying if a name is present and what value it has if so,
and the like -- there are various tools you can use for that, such
as direct assignments and deletions:
    x.aname = 23
    del x.aname
and builtins such as setattr, getattr, delattr, hasattr.

Further, all the namespaces you thus generate have a common root,
class C itself -- at any time you can (using exactly the same
tools) manage that common root namespace.

Pretty useful indeed, if you need namespaces.  There are quite a
few common additions to the idiom, such an initializer to allow
handy initialization of a set of names in a namespace (instance)
while you're generating it:

class Bunch:
    def __init__(self, **kwds): self.__dict__.update(kwds)

Here I've used the canonical name for this Python idiom, as found
in the Cookbook.  Now, you can initialize a namespace at the
same time you're instantiating it, e.g:

apoint = Bunch(x=23, y=45)

and the usefulness keeps growing.  Use Advanced Google Groups Search
on this newsgroup for word 'metabunch' and you'll find even niftier
versions of this idiom based on custom metaclasses -- etc, etc.


> Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or

Sure, that's using a and b as two separate namespaces, the most
fundamental and elementary use of the idiom.

> some such. If I try to print a.x2 or b.x1, I get an exception message
> basically saying those member variables don't exist. It seems to me this

Well, of course -- they're SEPARATE namespaces, by design!

> defeats one of the basic ideas behind OOP, that is to assign members to
> the class definition, thus ensuring that all instantiations of a class
> have the same member variables and methods. 

I'm not sure where you've read that this is "one of the basic ideas
behind OOP", but, wherever that may be, it's simply bunk -- it must
be an author having some hidden agenda and trying to pass off "static
typing" as somehow part of "OOP", presumably trying to convince you
that static typing must be good by association.  Fortunately, this
particular disinformation is easy to debunk, thanks to the fact that
Smalltalk, basically THE oldest OO language still in widespread use
(Simula67 was older, but it's not widespread these days), doesn't use
static typing -- when you send a message to an object, you can learn
at runtime that the object "doesn't understand" that message (i.e.,
doesn't have that method).  Why, some excellent OO languages don't
even HAVE classes -- just objects... if ANYTHING connected to classes
HAD to be part of OO, those language couldn't exist.  Bunk, as I say.

Don't believe everything you read...

> I know there must be a
> reason that python allows this kind of thing that I described, but I
> can't figure out what that reason could be. 

Practicality.

> To me, it looks like a very
> dangerous thing to do. 

Why so?  Please explain.  Python has dictionaries.  You can do:

    aa = {}
    bb = {}

and later

    aa['x1'] = 1.2
    ab['x2'] = 3.5

but now of course if you try to print aa['x2'] or bb['x1'], you 
get an exception message basically saying those items don't exist.

So: do you consider this "a very dangerous thing"?  How so?

And if that's not all that dangerous, then why should the syntax 
to do basically *exactly the same things* on top of a class (and
with even less redundant punctuation) suddenly become "very
dangerous"?!

If you don't see how these are exactly the same things, consider
that if you use your initial example and then do
    aa = a.__dict__
    bb = b.__dict__
you get exactly to this latter point, and viceversa.  If you do
not have any "behavior" (methods, i.e. executable attributes), a
class instance is basically a thin syntax wrapper around a dict,
with a few useful extra features that add absolutely no danger
whatsoever (and that you haven't even mentioned anyway).

> Can someone help me understand why we'd like to
> do it? Thanks,

Why we'd like to refer to a.x1 rather than aa['x1']?  Nicer and
smoother syntax.  Moreover, classes CAN have behavior, and thus
supply nicely polymorphic access to their functionality -- PLUS
they offer inheritance, so several instances can automatically
share some of their state (and behavior), etc, etc.  So, in fact,
there's a lot of practical, useful potential connected to that
"thin syntax wrapper".

You appear to think there are downsides and indeed "danger".
Can you help me understand what you think those downside are
and where the danger lies?  Thanks.


Alex




From belred1 at yahoo.com  Thu Apr 24 22:14:12 2003
From: belred1 at yahoo.com (Bryan)
Date: Fri, 25 Apr 2003 02:14:12 GMT
Subject: xmlrpc introspection
References: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com> <af6b5bdd.0304240434.fd7e6ad@posting.google.com>
Message-ID: <UR0qa.77049$gK.181217@rwcrnsc52.ops.asp.att.net>

i see a scalability problem with listMethods.  we have a server that
dynamically loads many providers.  each provider registers one or more
services.  each service handles one or more methods.  so a typical method
name would be service.method or for example calculator.add.  when the server
receives calculator.add, it knows to send it to all providers that
registered for calculator service.  now when our server receives a
listMethods request, it has to return all the method names in all providers.
this could be hundreds or even thousands.  it would be nice if it was
standard that there was a listServices request and listMethods took an
optional service parameter.  then you can scale this up.  i noticed that
it's typical to have method names in the format service.method so i'm too
far off with this.  i'm probably going to have to implement it this way, i
just wish it was the "standard" way.

bryan

"poiboy" <bitbucket at safe-mail.net> wrote in message
news:af6b5bdd.0304240434.fd7e6ad at posting.google.com...
> Quick answer: Register 'system.listMethods', 'system.methodSignature',
> and 'system.methodHelp' in your server and make sure they perform as
> stated in the docs (listMethods returns a list of strings, etc.). Yup
> - just do it all by hand.
>
> The output for these methods are typical <methodResponse>s (snippets
> available at http://www.xmlrpc.org/spec) - which *should* be taken
> care of automatically by the XMLRPC server; lists and/or strings
> converted to something like
>
> <?xml version="1.0"?>
> <methodResponse>
>     <params>
>         <param>
>             <value>
>                 <array>
>                     <data>
>
> <value><string>remoteMethodOne</string></value>
>
> <value><string>remoteMethodTwo</string></value>
>
> <value><string>remoteMethodThree</string></value>
>                     </data>
>                 </array>
>             </value>
>         </param>
>     </params>
> </methodResponse>
>
> ..for 'server.listMethods'. In other words, you *shouldn't* have to
> worry about the raw XML. Verify your server's output democratically
> via the debugger link above
> (http://www.dscpl.com.au/xmlrpc-debugger.php - added to my bookmarks,
> mahalo Graham!) or manually by testing 'server.listMethods' at
> http://xmlrpc.usefulinc.com/demo/server.php.
>
> A headstart on introspection can be found at
> http://xmlrpc-c.sourceforge.net/hacks/xmlrpc_registry.py, and there's
> a detailed account of a comprehensive 'system.describeMethods'
> implementation at
http://xmlrpc-epi.sourceforge.net/specs/rfc.system.describeMethods.php
> which I recommend for practical documentation consideration.
>
> Two unsolicited cents worth: Complimentary RPC test cases are
> invaluable (list all methods as links, click to test individually or
> click "test all"). Had to get that off my chest.
>
> Aloha, the poiboy





From abelikov72 at hotmail.com  Thu Apr 17 04:00:27 2003
From: abelikov72 at hotmail.com (Afanasiy)
Date: Thu, 17 Apr 2003 08:00:27 GMT
Subject: "pythonic"
Message-ID: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>

Too many times now I have seen the term "pythonic" used to vaguely
describe something which is never explicitly defined, as if it cannot be.

Now it seems I can just replace "pythonic" with "my idea of utopia".

I do not know if this ignorance, laziness, or a meme, but I would like
to know if there a real definition of "pythonic". Anyone?



From pan-newsreader at thomas-guettler.de  Wed Apr  2 03:58:07 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 02 Apr 2003 10:58:07 +0200
Subject: distutils: uninstall
Message-ID: <pan.2003.04.02.10.58.06.906512.810@thomas-guettler.de>

Hi!

I want to uninstall something I installed with distutils.
"setup.py clean" only removes the build directory. The "--all" option does
not help, too.

The only message I found related to this is quite old:
 http://mail.python.org/pipermail/distutils-sig/2000-March/000793.html

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From hs2l2ATyahoodotcom at yahoo.com  Thu Apr 17 13:08:02 2003
From: hs2l2ATyahoodotcom at yahoo.com (Haran Shivanan)
Date: 17 Apr 2003 17:08:02 GMT
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de> <mailman.1050586468.29594.python-list@python.org> <rozna.115488$It5.70451@news2.central.cox.net>
Message-ID: <b7mn1i$2icfr$2@ID-148859.news.dfncis.de>

module "Steve Holden" core dumped with the following output:
> 
>>>> class A:
> ...   a = 0
> ...   b = 1
> ...
>>>> class B(A):
> ...   pass
> ...
>>>> b = B()
>>>> A.a = 23
>>>> b.a
> 23
>>>>
> 
> I suspect the OP wants a separate class variable for each subclass. Right,
> Haran?
Right.
The variables in the subclass should not be bound to the parent.
Any ideas?
> 
> regards
> --
> Steve Holden                                  http://www.holdenweb.com/
> How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
> Python Web Programming                 http://pydish.holdenweb.com/pwp/
> 
> 
> 
> 


-- 
Kaa's Law: In any sufficiently large group of people, most are idiots.



From staschuk at telusplanet.net  Mon Apr 14 09:04:20 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 14 Apr 2003 13:04:20 -0000
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 14)
Message-ID: <v9lcekr9vkl734@corp.supernews.com>

QOTW:  "[I]f you only need something which quacks, don't worry about
checking that it's also a duck."             -- Andrew Dalke

"Source code with decent unit and acceptance tests is an asset.  Source
code without decent tests is a liability."  -- Peter Hansen


Discussion:
    The interaction of __slots__ and inheritance is explained, amid
    debate about the merits of __slots__ itself.
        http://groups.google.com/groups?th=3e450bb36a53e4fb

    Tim Peters proposes changing list.insert() to follow the negative
    index convention.
        http://groups.google.com/groups?th=192070c6605fb830#link10

    Gregory R. Warnes submits and solicits feedback on PEP 754: "IEEE
    754 Floating Point Special Values".
        http://groups.google.com/groups?th=5359401ba8ab5ffc

    Jason Tishler is still looking for Cygwin Python regression
    testers for Windows 95 and Windows Me.
        http://cygwin.com/ml/cygwin/2003-04/msg00314.html

    Oren Tirosh shows how to write something like a Python generator
    in C, using Duff's Device.
        http://groups.google.com/groups?th=ac5e4f70acfb521f#link9

Announcements:
    Capon 0.1: Capon blurs the line between a conventional make
    utility and a general-purpose programming language to provide the
    best of both worlds.
        http://www.dancontrol.net/share/capon/index.html

    elmer 1.0.2: Elmer allows function calls and various data types,
    both native and custom, to pass transparently between two
    languages.  In the current version, those languages are Python and
    either Tcl or C.
        http://elmer.sourceforge.net

    KirbyBase 1.0: KirbyBase is a simple, pure-python, flat-file
    database management system, with a few twists.
        http://www.netpromi.com/files/kirbybase1.0.zip

    wxPyPlot V1.0 Beta 2: An enhanced derivative version of
    wxPlotCanvas to provide simple lightweight plotting in wxPython.
    The emphasis is both on simple and lightweight.
        http://www.cyberus.ca/~g_will/wxPython/


========================================================================
Everything you want is probably one or two clicks away in these pages:

    Python.org's Python Language Website is the traditional
    center of Pythonia
        http://www.python.org
    Notice especially the master FAQ
        http://www.python.org/doc/FAQ.html

    PythonWare complements the digest you're reading with the
    daily python url
         http://www.pythonware.com/daily  
    Mygale is a news-gathering webcrawler that specializes in (new)
    World-Wide Web articles related to Python.
         http://www.awaretek.com/nowak/mygale.html 
    While cosmetically similar, Mygale and the Daily Python-URL
    are utterly different in their technologies and generally in
    their results.

    comp.lang.python.announce announces new Python software.  Be
    sure to scan this newly-revitalized newsgroup at least weekly.
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python.announce

    Brett Cannon continues the marvelous tradition established by 
    Andrew Kuchling and Michael Hudson of summarizing action on the
    python-dev mailing list once every other week.
	http://www.python.org/dev/summary/

    The Python Package Index catalogues packages.
        http://www.python.org/pypi/

    The somewhat older Vaults of Parnassus ambitiously collects references
    to all sorts of Python resources.
        http://www.vex.net/~x/parnassus/   

    Much of Python's real work takes place on Special-Interest Group
    mailing lists
        http://www.python.org/sigs/

    The Python Business Forum "further[s] the interests of companies
    that base their business on ... Python."
	http://www.python-in-business.org

    The Python Software Foundation has replaced the Python Consortium
    as an independent nexus of activity
        http://www.python.org/psf/

    Cetus does much of the same
        http://www.cetus-links.org/oo_python.html

    Python FAQTS
        http://python.faqts.com/

    The old Python "To-Do List" now lives principally in a
    SourceForge reincarnation.
        http://sourceforge.net/tracker/?atid=355470&group_id=5470&func=browse
        http://python.sourceforge.net/peps/pep-0042.html
     
    The online Python Journal is posted at pythonjournal.cognizor.com.
    editor at pythonjournal.com and editor at pythonjournal.cognizor.com
    welcome submission of material that helps people's understanding
    of Python use, and offer Web presentation of your work.

    *Py: the Journal of the Python Language*
        http://www.pyzine.com

    Tenth International Python Conference 
        http://www.python10.org            

    Archive probing tricks of the trade:
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python&num=100
        http://groups.google.com/groups?meta=site%3Dgroups%26group%3Dcomp.lang.python.*

Previous - (U)se the (R)esource, (L)uke! - messages are listed here:
  http://www.ddj.com/topics/pythonurl/         
  http://purl.org/thecliff/python/url.html (dormant)
or
  http://groups.google.com/groups?oi=djq&as_q=+Python-URL!&as_ugroup=comp.lang.python


Suggestions/corrections for next week's posting are always welcome.
E-mail to <Python-URL at phaseit.net> should get through.

To receive a new issue of this posting in e-mail each Monday morning
(approximately), ask <claird at phaseit.net> to subscribe.  Mention
"Python-URL!".


-- The Python-URL! Team--

Dr. Dobb's Journal (http://www.ddj.com) is pleased to participate in and
sponsor the "Python-URL!" project.




From pan-newsreader at thomas-guettler.de  Sat Apr  5 09:31:02 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Sat, 05 Apr 2003 16:31:02 +0200
Subject: Input from a file as a command line argument
References: <3E8EDBDA.7020701@pbezone.net>
Message-ID: <pan.2003.04.05.16.31.02.87447.878@thomas-guettler.de>

On Sat, 05 Apr 2003 15:36:26 +0200, Teemu Luojola wrote:

> I have read through Python documentation and tried several searches, but
> I can't find a simple answer to this question.
> 
> If I make an executable script in linux environment, how can I pass
> input to that script from a file on command line? Like the following:
> 
> $ script.py input.txt

do it like this:

import sys
input_txt=sys.argv[1]

> $ script.py < input.txt

import sys
lines=sys.stdin.readlines() # or .read() to get it in one string
for line in lines:
    .....

since the above code reads everything from stdin first and the enters
the loop, you might prefere an interation:

import sys
for line in sys.stdin.xreadlines():
    print line
    
 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From ianb at colorstudy.com  Sun Apr  6 22:10:22 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 06 Apr 2003 21:10:22 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <6ee58e07.0304061758.35070e7c@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <6ee58e07.0304061758.35070e7c@posting.google.com>
Message-ID: <1049681422.658.153.camel@lothlorien>

On Sun, 2003-04-06 at 20:58, Lothar Scholz wrote:
> No its not a naturally tailored web-programming language, because it
> is very hard to integrate in Server Pages. You have to change the
> indentation syntax/semantic of python to embedd it in a real useable
> way.  This is done by Webware but it looks very strange to normal python
> programmers.

OTOH, it's rather easy to integrate HTML into Python code -- easier than
many other languages (maybe Perl is better, PHP isn't).

> Of course you can use template systems. But this is not always a good
> economic decision.

If you want a quick solution, then I'd recommend using no template at
all.  If you want real templating, Server Pages aren't a particularly
good implementation.

  Ian






From rdsteph at earthlink.net  Tue Apr 29 09:24:14 2003
From: rdsteph at earthlink.net (Ron Stephens)
Date: Tue, 29 Apr 2003 13:24:14 GMT
Subject: Python Tutorials
Message-ID: <15451.8030105@earthlink.net>

I have newly organized the links to 67 online Python tutorials on my 
site, starting from easy to advanced, and then by subject. This list 
might be especially useful for extreme newcomers to progamming, as there 
are quite a few very easy, introductory tutorials, allowing one to work 
one's way up from the absolute beginning; but there is also good 
coverage of GUI programming, CGI, and many other specific and even 
advanced subjects, including scientific programming, bioinformatics, 
databases, XML and many other topics; all having to do with Python. It 
is amazing the number and quality of Pythonic tutorials the community 
has produced! We owe their authors many thanks...

www.awaretek.com/plf.html




From a.schmolck at gmx.net  Sun Apr  6 22:26:24 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 07 Apr 2003 03:26:24 +0100
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org>
Message-ID: <yfsr88fkppr.fsf@black132.ex.ac.uk>

Ian Bicking <ianb at colorstudy.com> writes:
> I came to appreciate Emacs more when I looked at THE
> http://humane.sf.net/the/ -- an editor made by Jef Raskin, former Mac
> interface designer.  The interface he describes, without realizing it,
> is Emacs, and THE is just a poor implementation of Emacs.  

He certainly knows about Emacs (have a look at the index of his book), but I
guess obtaining patents and attracting attention as some sort of visionary is
unnecessarily complicated by pointing out the prior art yourself. And although
the most crucial (and brilliant) idea, incremental search, is nicked (and
uncredited on his webpage), the quasi-modal character (i.e. that you have to
keep the search key pressed) of LEAP (tm, patented and what have you) seems
like an actual difference -- you don't have to press return/C-g/movement
commands to get out of it, which I would guess (I haven't tried it -- no mac)
makes it more natural for moving in text (it also places the cursor at the
beginning of the match, I think, which seems similarly useful for moving
around).

> Now, I'm not saying Emacs couldn't be improved.  But it's a far, far
> better starting point that "modern" text editors.  I'm thoroughly
> convinced that modality is the superior interface for most situations,
> and I think GUIs are leaning more in that direction as time goes on.  I
> think this is why vi remains popular -- even though I personally don't
> like it -- because it's modal.  At least anything that's keyboard
> friendly should be modal, and a text editor that isn't keyboard friendly
> is clearly stupid.

Huh?

> 
> But I'd still like a Python Emacs to make experimentation with UI
> possible.  It would also be neat if it could be an environment for
> application development, like Emacs is, like THE wants to be... Oberon
> was like this too, and while I didn't like many parts of Oberon, the
> overall environment was really neat.  Squeak keeps trying to move this
> way too. 

Squeak also happens to sport an UI element whose usefulness seems incredibly
obvious to me but which I haven't seen anywhere else (exchanging contents of
current and prior selection).

alex



From eppstein at ics.uci.edu  Fri Apr 18 19:01:47 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Fri, 18 Apr 2003 16:01:47 -0700
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net>
Message-ID: <eppstein-653307.16014718042003@news.service.uci.edu>

In article <C8_na.646$dP1.2863 at newsc.telia.net>,
 "Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote:

> > What's wrong with:
> >
> > t = 0
> > for y in x:
> >      t += y[1]
> >
> > The extra variable?  The number of lines?
> 
> I am really curious about that to, I would most definitley say that this is
> the most pythonic way. It is simple and easily readabel by anyone and I am
> sure it is just as fast any of the other way.

If you want the sum of a list of items, you should write it in a way 
that looks like "the sum of a list of items", not in a way that looks 
like "loop over these items, maintain another variable t, perform a 
sequence of additions".  Why do we have high level languages if not to 
express our intentions at a higher level and let the language worry 
about what low-level operations are needed to implement it?

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From sholden at holdenweb.com  Mon Apr 28 21:49:47 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 29 Apr 2003 01:49:47 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <20030417205405.A10672@tibia.amotlpaa.bogus> <mailman.1051310385.1948.python-list@python.org> <b8cgst$20ho$1@nntp6.u.washington.edu> <mailman.1051490038.11587.python-list@python.org> <NTjra.23897$WQ4.21608@news2.central.cox.net>
Message-ID: <%Skra.24045$WQ4.10941@news2.central.cox.net>

<lurk>
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/





From Anish.Mehta at enst-bretagne.fr  Mon Apr 21 05:55:28 2003
From: Anish.Mehta at enst-bretagne.fr (Mehta, Anish)
Date: Mon, 21 Apr 2003 11:55:28 +0200
Subject: Multiple assignment?
Message-ID: <3EA3C010.5090002@antares.enst-bretagne.fr>

Can someone tell me the what is being done here in this line of code.

dir, file = os.path.split(input)

Waiting for reply

Thanks

Anish MEHTA





From berkowit at silcom.com  Mon Apr 14 23:42:30 2003
From: berkowit at silcom.com (Paul Berkowitz)
Date: Mon, 14 Apr 2003 20:42:30 -0700
Subject: Built-ins vs. modules
Message-ID: <BAC0CDB6.35077%berkowit@silcom.com>

I see from the Python Library Reference (for 2.2) that a lot of things which
I thought required a module to be imported can also be accessed without
doing so:

    >>> "boo".capitalize()   #built-in
    'Boo'
    >>> import string
    >>> string.capitalize("boo")
    'Boo'


Why import the string module if you can do the same things anyway? The page
of built in String Methods bears an uncanny resemblance to the page on the
string module. Is there a historical reason? Are the built in type methods
quite recent?


-- 
Paul Berkowitz






From robin at jessikat.fsnet.co.uk  Thu Apr 17 13:14:54 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Thu, 17 Apr 2003 18:14:54 +0100
Subject: fixedpoint.py vs Python 2.1
References: <$uJOMKAPtqn+Ew9c@jessikat.demon.co.uk>
 <mailman.1050594427.26198.python-list@python.org>
Message-ID: <W6ZEJKAOEun+EwZA@jessikat.demon.co.uk>

In article <mailman.1050594427.26198.python-list at python.org>, Tim Peters
<tim.one at comcast.net> writes
.....
>
>Find a version of the code that predates the SF one, and it won't have this
>problem (it didn't originally use type(self) or self.__class__).
>
>I don't know whether the SF folks intended to support Pythons earlier than
>2.2.
>
>
well I looked at the one referred to by the SF page at

        http://fixedpoint.sourceforge.net/TimsMessage.txt

ie
        http://starship.python.net/crew/aahz/FixedPoint.py

and it seems quite a number of changes were made since then. In
particular _mkFP is passed the constructor attributes rather than a
class constructor.

Starting with the SF one I replaced the usage of type(self) referring to
the current instance class with self.__class__ and that seems to
work.....,

but I'm not sure if type(instance) is exactly equivalent to
instance.__class__ in 2.2 so could be missing something.
-- 
Robin Becker



From ckea25d02 at sneakemail.com  Tue Apr  1 20:14:03 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 13:14:03 +1200
Subject: Use Pyrex instead (Re: newbie at SWIG, help needed with typemaps)
In-Reply-To: <Xns935061D008C57duncanrcpcouk@127.0.0.1>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <3E87E64C.F7D2C11E@easystreet.com> <Xns935061D008C57duncanrcpcouk@127.0.0.1>
Message-ID: <b6dd1d$453c2$1@ID-169208.news.dfncis.de>

Duncan Booth wrote:
> I, of course, misread this and wrote:
> 
>      cdef class Spam:
>     	    	method = classmethod(Spam)

Hmmm, maybe I should make it clearer that I'm only talking about
Python classes here, not extension types.

> This gives an error from the C compiler: 'method' undeclared identifier.
> Fair enough that it doesn't work (the documentation sort of says it won't), 
> but not so good that it generated bad C code.

That appears to be a bug -- it shouldn't be letting you
put anything other than declarations inside a cdef class
in the first place. I'll see about fixing this.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ajsiegel at optonline.net  Tue Apr 22 12:47:42 2003
From: ajsiegel at optonline.net (Arthur)
Date: Tue, 22 Apr 2003 12:47:42 -0400
Subject: Paintshop Pro 8
Message-ID: <000801c308ee$e4d07050$0c02a8c0@Arts>

>So why is this?

>* Possible Linux/Mac versions in the pipelines?
>* Don't like COM?
>* Want to promote Python?

>Anyone any insight?

I am finding Python becoming ubiquitous in applications involving graphics.
Free, opensource  and commercial.  This happens to be the realm in which I
work with Python, and it was in searching for painless access to OpenGL that
I initially found it.  Now, more amd more when I search for some
functionality in this realm, without specific concern as to its
implementation details, I am finding Python in play.  Counter intuitive in
the sense that graphics has the need for speed, which is not Python's claim
to fame. But its not the engine that's in Python in any case.

Have some theories about why Python fits so well with graphics work, which I
will spare the group.

It would, however, be great to see the graphics related Python reesources
collected in one place.

"Computer Graphics with Python" - my book, I thought of it first.

Art









From intentionally at blank.co.uk  Sat Apr 26 10:13:04 2003
From: intentionally at blank.co.uk (Stephen Horne)
Date: Sat, 26 Apr 2003 15:13:04 +0100
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <slrnbajav2.e5n.ramen@lackingtalent.com> <pan.2003.04.25.17.49.27.407432@jerf.org>
Message-ID: <4o3lav4k04m9g38lm0k6e1mlau6lgma81m@4ax.com>

On Fri, 25 Apr 2003 21:49:36 GMT, "Jeremy Bowers" <jerf at jerf.org>
wrote:

>Implementation of the brain object left as an exercise for the reader. I
>know it can be done, as I am in possession of such an object right now.

Yeah, but my brain object must have been written in C++ - it's very
buggy indeed, and certainly suffering from multiple memory and
resource leaks. Security is very poor (I had a major headache dealing
with the 'flu' virus just the other week), and it's so unreliable I
get hours of downtime almost every single day.

I've been trying to contact the developer but it's no use - I just
can't get a reply.




From logistix at cathoderaymission.net  Tue Apr 15 13:53:28 2003
From: logistix at cathoderaymission.net (logistix at cathoderaymission.net)
Date: 15 Apr 2003 10:53:28 -0700
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org>
Message-ID: <3c91a864.0304150953.5f390761@posting.google.com>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> wrote in message news:<mailman.1050393203.4704.python-list at python.org>...

> 
> It seems to me that the behavior CAN be improved.  If strings were to
> use a preallocation strategy--similar to what lists do--MANY string
> appends could avoid a memory copy on the original string.  The intial
> value could stay where it was, and the extra characters could live in
> the already-available contiguous space.  Obviously, the utilized length
> would need to be stored somewhere.  Once in a while, of course, the
> preallocation would fill up; in that case you'd need to do a costly
> strcopy()... but this could be avoided MOST of the time, ISTM.
> 

This is the way the current list optimization works.  It's based on
(generally valid) assumption that re'malloc has three cases:
    You're re'mallocing to a smaller size, shrink allocation
    You're re'mallocing to the same size, do nothing
    YOu're re'mallocing to a bigger size, grab new memory and do a
memcpy()

The current optimization for list adjusts the malloc'ed memory based
on the size current size.  Even for something like a 512Meg list, it
only uses a maximum of 16 Meg for padding.

The real problem is the fact that strings are immutable.  From the
bytecode perspective, I don't think it's possible to reliably assume
that if string has a refcount = x in a '+' operation that it's not a
shared reference (where you couldn't optimize)

I submitted a patch to use the same allocation technique for
arraymodule, and it decreased my test case of 8 million appends from
12 hours to 4-5 seconds.  Needless to say, the patch was accepted and
will be 2.3 ;)

Someone else had a feature request to allow strings to be .appended()
to array.array("c") objects, but there's no patch yet.  If this were
implemented, array.array("c") would pretty much become a
"mutableString" class that'd be easy to use and run at reasonable
speeds.  [I was thinking about writing a patch, and then started
obsessing about what sort of unicode coercions should take place and
my brain froze up]



From DennisR at dair.com  Mon Apr 21 13:50:28 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Mon, 21 Apr 2003 17:50:28 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com>
Message-ID: <EbWoa.372$Fb7.28620918@newssvr21.news.prodigy.com>

> I'm thinkin in the 800k range for a simple program?

My anti-spam program is 797 KB.  800K is doable.  If you download, you can
see the files I include.  I use a browser for GUI and do not include a
separate GUI framework.
--

Dennis Reinhardt

DennisR at dair.com   http://www.spamai.com?ng_sig





From tstegen at cis.strath.ac.uk  Tue Apr 22 22:10:22 2003
From: tstegen at cis.strath.ac.uk (Thomas Stegen)
Date: Wed, 23 Apr 2003 02:10:22 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <871y01f2hw.fsf@deneb.enyo.de> <3E9F1FE8.5030409@removethis.free.fr> <pw-A79DD1.16064217042003@reader1.panix.com>
Message-ID: <3ea5e905$1@nntphost.cis.strath.ac.uk>

Paul Wallich wrote:
> 
> It's the second. In a strictest-sense functional language, it is 
> impossible to write a whole program that *does* have side effects. 

It would be easy to write such a program though. Without any sideeffects
the preconditions reduces to the postconditions and so the whole program
disappears. :) Now that would be very nice and simple, it might even be
the silver bullet.

-- 
Thomas.
        "How long is six months? Ten years?"
            - Denis Johnson, Fiskadoro




From marcos at renata.macondo.pri  Sat Apr  5 13:59:51 2003
From: marcos at renata.macondo.pri (Marcos =?iso-8859-1?q?S=E1nchez?= Provencio)
Date: Sat, 05 Apr 2003 20:59:51 +0200
Subject: FreeTDS vs Inline TDS: Problem with Inline TDS
References: <QhmQ8.215123$9F5.15887558@typhoon.austin.rr.com>
Message-ID: <pan.2003.04.05.18.59.50.129329@renata.macondo.pri>

On Thu, 20 Jun 2002 15:12:48 +0000, Sandeep Gupta wrote:

> I'm attempting to access a SQL Server 2000 DB from Linux using ODBC.  My
> setup is
> Python --> mxODBC --> unixODBC --> <ODBC Driver> --> SQL Server
> 
> When I use the Inline TDS driver directly from unixODBC, all queries
> return the correct results, but the final line is always "-1 rows
> affected".  The FreeTDS driver returns the correct results as well as the
> correct number of rows affected.  The incorrect "rows affected" from
> Inline TDS seems to be causing problems with mxODBC.
> 
> Has anyone experienced this same problem with Inline TDS's driver?
> 
> Are there any other suggestions for querying against SQL Server 2000 from
> Python running on Linux?
> 
> Thanks
> Sandeep
You may try
 http://www.object-craft.com.au/projects/sybase/
compiled with FreeTDS. It does not allow parameters in queries AFAIK. You
may ask at the db-sig list to get more precise answers.



From oblivious at web.de  Fri Apr 11 14:19:39 2003
From: oblivious at web.de (Dominic)
Date: Fri, 11 Apr 2003 20:19:39 +0200
Subject: Macros in Python?
In-Reply-To: <mailman.1050064218.12167.python-list@python.org>
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org>
Message-ID: <b770t8$2b8$00$1@news.t-online.com>

Beni Cherniavsky wrote:
> Dominic wrote on 2003-04-09:
> 
> 
>>(defmacro with-root-privs (() &body body)
>>   (let ((oldidsym (gensym)))
>>     `(let ((,oldidsym (geteuid)))
>>        (seteuid 0)
>>        (unwind-protect
>>	   (progn , at body)
>>	 (seteuid ,oldidsym)))))
>>
>>Is there an easy way to produce similar code
>>in Python which gurantees that certain things
>>happen around some code? (Maybe by using
>>generators?)
>>
> 
> To complete the list of solutions, there is an evil way to abstract
> try..finally that is not recommended due to it's unreadability to
> anybody that doesn't know the trick.  It also only works in CPython
> with reference counting that invokes __del__ immediately.
> 
> The basic idea was to use a for loop and a generator:
> 
> def with_root_privs():
>     oldid = geteuid()
>     seteuid(0)
>     yield None # dummy
>     seteuid(oldid)
> 
> for _ in with_root_privs:  # `_` is dummy
>     # Dohere whatever you wanted.  The generator will cause this to be
>     # executed exactly once, with pre- and post- code excuted...
> 
> The problem, of course is that the for loop's body can raise an
> exception, or exit the loop in other evil ways.  Luckily in CPython,
> exiting the loop immediately drops the only reference to the
> generator, so hooking it's __del__ would do the trick.  Writing it as
> an iterator class instead of a generator would be be ugly but we can
> only do it once, by a universal wrapper taking a naive generator and
> returning a safe one.  Google for "abstracting try..finally with
> generators" in c.l.py if you are interested.  Again, it's hard to
> recommend this hack.  Instead, consider pushing PEP 310 :-).

Actually I also considered using generators but I would never have
come by this idea to use ref-counting side effects to make
it work. ;-) This one is really great though it shouldn't be used...
Now I'll see what PEP 310 is all about.

Ciao,
  Dominic








From unendliche at hanmail.net  Wed Apr 23 11:44:21 2003
From: unendliche at hanmail.net (Seo Sanghyeon)
Date: 23 Apr 2003 08:44:21 -0700
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org>
Message-ID: <45e6545c.0304230744.1a0248b4@posting.google.com>

Gregoire Dooms wrote:

> I empirically discovered (Python 2.2.2) that None could be considered an
> infimum and any string could be considered a supremum. But this fact is not
> stated in the language reference. Could/should that order be hardcoded in
> the language?

I think the language will not be going to guarantee anything about the
comparison between incomparable types. I once read that None as supremum
instead of infimum in the previous version of Python.

If you really need them, you can do this:

class Infimum(object):
  def __lt__(self, x):
    return True
inf = Infimum()

Or, if you like an one-liner:

inf = type('Infimum', (object,), {'__lt__': lambda self, x: True})()

-- Seo Sanghyeon



From thedustbustr at aol.com  Sat Apr 19 21:51:50 2003
From: thedustbustr at aol.com (TheDustbustr)
Date: 20 Apr 2003 01:51:50 GMT
Subject: PyOpenGL glClearColor() trouble
Message-ID: <20030419215150.18878.00000431@mb-m18.aol.com>

from OpenGL.GL import *
import OpenGL.Tk as Tk

class Test:
    def __init__(self):
        self.ogl=Tk.Opengl(width=800,height=600,double=1)
        self.ogl.pack(side='top', expand=1, fill='both')
        self.ogl.redraw=self.draw

        self.setup()
        self.ogl.mainloop()

    def setup(self):
        """Why does the following line not work..."""
        glClearColor(0, 0, 0, 0)

    def draw(self, tkev=None):
        """but uncommenting this line and commenting the previous one DOES
work!?"""
        #glClearColor(0, 0, 0, 0)
        glClear(GL_COLOR_BUFFER_BIT)

Test()

The way the script is set up right now (second glClearColor commented), glClear
clears the screen to the default ClearColor (1,0,1,0 aka pink).  I can only
make the ClearColor black (0,0,0,0) when I call glClearColor in the ogl.redraw
function... This is inconvienent and I shouldn't need to be called every frame.



From mis6 at pitt.edu  Wed Apr  9 11:16:25 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 9 Apr 2003 08:16:25 -0700
Subject: detecting modifications to an instance? (for debugging)
References: <ef08387c.0304082116.e658c31@posting.google.com>
Message-ID: <2259b0e2.0304090716.48bf422c@posting.google.com>

fortepianissimo at yahoo.com.tw (Fortepianissimo) wrote in message news:<ef08387c.0304082116.e658c31 at posting.google.com>...
> I have a big chunk of Python code and I'd like to detect any run-time
> modification to instances of a particular class.
> 
> I'd like to know how to do it with a minimal addition of code, i.e.,
> it's much preferrable to make the code report any such instances
> instead of using a third-party debugging tool for it.
> 



> Note I just want to detect the modification, not to prevent them
> (which I think is impossible in Python).
> 
> Any suggestion is extremely welcome.

Not sure if I understand correctly what you want, but you can control
run-time modification of attributes via __setattr__:

class DetectModification(object):
    def __setattr__(self,x,y):
        print "Trying to set '%s' to %s ..." % (x,y),
        object.__setattr__(self,x,y)
        print "Done"

class C(DetectModification):
    def __init__(self,a):
        self.a=a

c=C(1)
c.a=2
print c.a # => 2

This snipped will warn you when the attribute 'a' is set (once in C.__init__
then in the line c.a=2). You may prevent the modification by raising
an exception in __setattr__.


            Michele



From gh at ghaering.de  Fri Apr 25 08:28:21 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Fri, 25 Apr 2003 14:28:21 +0200
Subject: GUI libs
In-Reply-To: <Kq9qa.9083$h%2.957726@read1.cgocable.net>
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
	<Kq9qa.9083$h%2.957726@read1.cgocable.net>
Message-ID: <3EA929E5.7000508@ghaering.de>

Mark Nenadov wrote:
> Alessio Pace wrote:
> 
> 
>>Hi, I'd like to make a GUI for a program written in python, meant to run
>>on mac, win and linux: which is the preferable solution? I saw around
>>tkinter, wxpython and pyqt....
>
> I'd strongly recommend wxPython. wxPython is quite a rich GUI library and I 
> have found it to be more advanced/modern/feature-rich than tkinter.

You'll get a Python interface with the error behaviour of the wrapped 
C++ library. The default is to crash. In my limited experience, anyway.

-- Gerhard




From jon+usenet at unequivocal.co.uk  Wed Apr 30 17:07:37 2003
From: jon+usenet at unequivocal.co.uk (Jon Ribbens)
Date: 30 Apr 2003 21:07:37 GMT
Subject: Web programming and a different "type" problem
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com> <slrnbarhfb.67t.jon+usenet@snowy.squish.net> <b8lj2q$l14$1@newshost.accu.uu.nl> <m3fzo1v561.fsf@ferret.object-craft.com.au> <mailman.1051654786.4587.python-list@python.org> <m365owlu6v.fsf@ferret.object-craft.com.au> <mailman.1051673506.20716.python-list@python.org>
Message-ID: <slrnbb0eoo.fb6.jon+usenet@snowy.squish.net>

In article <mailman.1051673506.20716.python-list at python.org>, Ian Bicking wrote:
> It might be better if the error is signaled earlier and in a more
> pleasant way, but IMHO throwing away duplicate values when they
> aren't expected just covers up bugs.  But I don't think that's what
> you're doing in this case (though it sounds like jonpy does this,
> for instance)...?

You think it would be better to throw an exception if two form
variables have the same name unexpectedly? I... disagree ;-)



From staschuk at telusplanet.net  Sun Apr 27 20:36:57 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 27 Apr 2003 18:36:57 -0600
Subject: Forgetting "()" when calling methods
In-Reply-To: <oprn83xbstmu3s0q@news.cis.dfn.de>; from km-news1@onlinehome.de on Sat, Apr 26, 2003 at 09:18:37PM +0200
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <mailman.1051376980.15852.python-list@python.org> <oprn83xbstmu3s0q@news.cis.dfn.de>
Message-ID: <20030427183657.A28461@tibia.amotlpaa.bogus>

Quoth Klaus Meyer:
  [cases when omitting () from a function call causes strange behaviour
   instead of an exception]
> You are right. I had made this error maybe 2 or 3 times and found the bug 
> fast and without problems.
> But because it's a simple "typing error" or so, i am not save to make this 
> error again... :-(

It's interesting that you think of omitting the () on a function
call as a "typing error".

The need for () might become easier to remember when it is
strongly associated with the conceptual distinction between
calling the function and mentioning it.  (This distinction will
become important to you when you start using functions as
first-class objects -- e.g., passing them as arguments to other
functions, producing them as return values from your own
functions, and the like.)  Then it becomes not a typing error but
a thinking error, which is less likely.

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From ramen at lackingtalent.com  Wed Apr 16 02:11:49 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Wed, 16 Apr 2003 06:11:49 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <mailman.1050438685.19213.python-list@python.org> <sggp9v8lpk0045klmquc8gec1ju7q4kmpe@4ax.com>
Message-ID: <slrnb9pt1o.iin.ramen@lackingtalent.com>

In article <sggp9v8lpk0045klmquc8gec1ju7q4kmpe at 4ax.com>, Courageous wrote:
> Anonymous functions are occasionally useful, as in Java's anonymous
> class expressions, really used as a poor man's lambda of sorts. I've
> seen this overused, though, that's for sure.

How frequently a feature is useful depends on the problem domain. Java's
anonymous classes are a poor substitute because they carry so much baggage.
For comparison, here's an implementation and usage of map in Java:

static interface Functor {Object call(Object[] args);}

static Object[] map(Functor func, Object[] array) {
    Object[] result = new Object[array.length];
    for (int i = 0; i < array.length; i++) {
        result[i] = func.call(new Object[] {array[i]});
    }
    return result;
}

public static void main(String args[]) {
    System.out.println(Arrays.asList(map(
        new Functor() {
            public Object call(Object[] args) {
                return args[0] + ", world!";
            }
        }, new String[] {"Hello", "Goodbye"}
    )));    
}

And the same usage in Python:

>>> print map(lambda x: x + ', world!', ['Hello', 'Goodbye'])
['Hello, world!', 'Goodbye, world!']

This is why Java is IMHO unfit for common functional tasks where Python
shines with elegance. As to whether or not anonymous functions can be
overused, I think any feature can be overused. Classes, for one. In fact, in
my day to day work, I find the two of similar usefulness, but maybe it's
just because of the way I program. =)

Peace,
Dave



From skip at pobox.com  Tue Apr  8 11:04:48 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 8 Apr 2003 10:04:48 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <3E92E00B.6000708@kyborg.dk>
References: <m2u1dcw22s.fsf@localhost.localdomain>
        <3E9040CF.2040606@removethis.free.fr>
        <m2k7e7addl.fsf@localhost.localdomain>
        <b6pgcm$1t3$1@panix3.panix.com>
        <m27ka7aae0.fsf@localhost.localdomain>
        <6ee58e07.0304061722.2472b3d0@posting.google.com>
        <3E9175A5.1060607@kyborg.dk>
        <yfsof3ijmua.fsf@black132.ex.ac.uk>
        <3E92E00B.6000708@kyborg.dk>
Message-ID: <16018.58640.617445.473239@montanaro.dyndns.org>

>>>>> "Kim" == Kim Petersen <kp at kyborg.dk> writes:

    >> (defun foo()
    >>   (message "when foo is called folding is: %s" case-fold-search))
    >> (defun bar()
    >>   (message "when bar is called folding is: %s" case-fold-search)
    >>   (let ((case-fold-search t))
    >>     (foo))
    >>   (message "now folding is: %s again" case-fold-search))
    >> (bar)

    Kim> hmmm - occured to me that you might mean that case-fold-search was
    Kim> a global variable first - then local.... but that wouldn't matter
    Kim> to this since we would need to do lookup in locals first _then_ in
    Kim> globals (as expected)

Look again.  In Elisp, when foo() looks up the value of case-fold-search it
will find the value defined in the let statement inside bar() because it's
lexically scoped (looks up the call chain for values).  If this was Python,
the reference to case-fold-search in foo() would be resolved at the global
scope because Python is statically scoped (looks in locals, globals, then
builtins only - ignoring nested functions which this case doesn't involve).
If the global value of case-fold-search was nil (False), elisp would would
print

    when bar is called folding is: nil
    when foo is called folding is: t
    now folding is: nil

The textually equivalent Python would print

    when bar is called folding is: False
    when foo is called folding is: False
    now folding is: False

Skip




From cnetzer at mail.arc.nasa.gov  Sun Apr 20 15:03:06 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 20 Apr 2003 12:03:06 -0700
Subject: A decent explanation of continuations (for lisp)
In-Reply-To: <b7ufg4$o1d$1@news.hccnet.nl>
References: <3E9E3040.5040409@tismer.com>
	 <mailman.1050607597.6802.python-list@python.org>
	 <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de>
	 <mailman.1050725072.21332.python-list@python.org>
	 <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl>
	 <ruadnZIgGvclOj-jXTWcpA@speakeasy.net>  <b7ufg4$o1d$1@news.hccnet.nl>
Message-ID: <1050865385.593.4.camel@sayge.arc.nasa.gov>

On Sun, 2003-04-20 at 08:42, Anton Vredegoor wrote:

> Thanks. It seems Tim is making a heroic effort to explain the new in
> terms of the old. This necessarily must fail, but what other option is
> there instead of rebuilding all terminology out of gotos?

FWIW, I only began to understand the implications of continuations after
reading "On Lisp", chapters 20 and 22.  I strongly recommend it
(especially chap. 22) if anyone wants to get more grounded in the
possibilities of continuations.

http://www.paulgraham.com/onlisp.html

download the pdf at:
http://www.paulgraham.com/onlisptext.html

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From juneaftn at orgio.net  Fri Apr 18 09:40:23 2003
From: juneaftn at orgio.net (June Kim)
Date: 18 Apr 2003 06:40:23 -0700
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com> <8b2d5523.0304180059.722de25a@posting.google.com>
Message-ID: <5699ebda.0304180540.487f6dca@posting.google.com>

qvx3000 at yahoo.com (Qvx) wrote in message news:<8b2d5523.0304180059.722de25a at posting.google.com>...
> blunck at gst.com (Christopher Blunck) wrote in message news:<1147e466.0304160630.5b510724 at posting.google.com>...
> > So how many of you guys use XP processes?
[snip]
> admit it: I don't know how to to use it. All work is done in Oracle
> database and Forms and they don't fit nicely with TDD (at least to
> me). How do you test a form? How do you test a lengthy database job
> whose input depends on dozens of tables and which modifies/creates
> records across dozens of tables?
[snip]

The essence is that you keep a simple rule of having GreenBarCycle
short, and in order to keep it do whatever available to you. Good
and suitable solution will emerge by itself or you'll get smart
enough to discover it.

For some hints:

To TDD database, go for the model first. Don't consider it's saved
on the disk. Consider it in-memory database or something. As soon
as you touch the persistence layer, which must be somewhat later,
you can use mock objects(google it).

see also http://groups.yahoo.com/group/testdrivendevelopment/files/TDDD.pdf 

I don't know what you mean by Form exactly, but guessing from my
experience that most XP beginners suffer from TDDing UI jobs, I'd
give you the following hints:

Make the presentation layer thiner and thiner. Have a look at MVP;
it is much easier to TDD than MVC.

see also:
 * http://www.sdmagazine.com/documents/s=7147/sdm0206b/0206b.htm
 * http://www.objectmentor.com/resources/articles/TheHumbleDialogBox.pdf

If you try these ways, your design will improve enourmously as a Throw-in.

With warmest regards,
June (Changjune Kim)



From staschuk at telusplanet.net  Tue Apr  8 21:47:55 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 8 Apr 2003 19:47:55 -0600
Subject: Best way to make a weighted sum
In-Reply-To: <86vfxo3b32.fsf@ieee.org>; from godoy@metalab.unc.edu on Tue, Apr 08, 2003 at 06:58:57PM -0300
References: <867ka46awj.fsf@ieee.org> <mailman.1049833916.16117.python-list@python.org> <86vfxo3b32.fsf@ieee.org>
Message-ID: <20030408194755.C1973@tibia.amotlpaa.bogus>

Quoth Jorge Godoy:
> Steven Taschuk <staschuk at telusplanet.net> writes:
> 
> > I don't know why you're using a class for this.  A simple function
> > which calculates (and/or verifies) the checksum seems more natural.
> 
> The class has several other things. As I said, this was a snippet of
> code. It showed you the __init__ and the other methods I was
> interested in discussing.

Ah.  Sorry, I thought you posted the complete class.  My mistake.

  [...]
> The line that does the mod 11 is this one, in case you want to see it:
> 
>         self.modulo = self.soma % 11
> 
> Note that self.soma had other operations before being submitted
> here. Some of our States do really weird things before the modulus
> calculation. And some do these weird stuff after that.

Hm.  Another alternative, then: a dict mapping from state to
checksum function.  For the simple cases in your original message,
something like this:

	checksumfuncs = {}

	def simplechecksum(ie, weights):
		w = weights[-len(ie):]
		return reduce(operator.add, [x*y for x, y in zip(ie, w)]) % 11

	checksumfuncs['CE'] = checksumfuncs['AP'] = \
		lambda ie: simplechecksum(ie, [9, 8, 7, 6, 5, 4, 3, 2])
	checksumfuncs['RJ'] = \
		lambda ie: simplechecksum(ie, [2, 7, 6, 5, 4, 3, 2])

A weird checksum function might look like, for example,

	def weirdchecksum(ie):
		s = simplechecksum(ie, [2, 7, 6, 5, 4, 3, 2])
		# do something weird to s
		return s
	checksumfuncs['AB'] = weirdchecksum

In this approach you might have

	def checksum(ie, state):
		return checksumfuncs[state](ie)

This might adapt better to the cases in which one state has many
ways of computing a checksum; the function for that state chooses
among the various methods by whatever means are appropriate.  (I
assume it's possible to determine the appropriate method by
inspecting ie.)

> > If uf is not in ['CE', 'AP', 'RJ'], then self.pesos simply does not
> > get set; this will cause CalculaSoma to fail later, in a mysterious
> > way.  Better to have an informative exception raised at this point,
> > surely.
  [...]
> Also there will be no exception since this is also bound to a domain
> at the database server.

If it were my program, I'd check anyway, just in case.  At the
least I'd use an assertion.
 
> You enlightened me a lot on this message, thanks.

Glad I could be of help.

> There is a green, multi-legged creature crawling on your shoulder.

Aaah!

-- 
Steven Taschuk                                        staschuk at telusplanet.net
"Study this book; read a word then ponder on it.  If you interpret the meaning
 loosely you will mistake the Way."         -- Musashi, _A Book of Five Rings_




From pan-newsreader at thomas-guettler.de  Wed Apr  2 01:34:28 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 02 Apr 2003 08:34:28 +0200
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de> <e9570f37.0304011121.1873dac2@posting.google.com>
Message-ID: <pan.2003.04.02.08.34.27.966495.810@thomas-guettler.de>

On Tue, 01 Apr 2003 21:21:47 +0200, Graham Fawcett wrote:

> Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote in message
> news:<pan.2003.04.01.13.58.42.127815.759 at thomas-guettler.de>...
>> Hi,
>> 
>> I think about moving from Zope to
>> ZODB and (Quixote or Twisted).
>
> 
> I'm working on such an application right now -- Quixote and ZODB. It's
> working out very well, and is far more manageable (and enjoyable!) than
> any of the Zope projects I've worked on.
> Do you have a specific concern that you're seeking feedback about?

Hi Graham!

I just saw your message called "Quixote Server on top of Twisted Python"
in Quixote-Users of last week.

What is your intend of doing this? Do you want to use PTL, or do you want
use it to use ZODB? I haven't looked at twisted in detail, but it should
be possible to use twisted with ZODB without quixotte, too.
 
 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From tjg at craigelachie.org  Sat Apr  5 17:29:15 2003
From: tjg at craigelachie.org (Timothy Grant)
Date: Sat, 5 Apr 2003 14:29:15 -0800
Subject: properties and get/set methods
Message-ID: <200304051429.15203.tjg@craigelachie.org>

Hi all,

Given the following class...

class monty(object):
    def __init__(self):
        self.__foo = 0
        self.__bar = 0

    def setvar(self):
        do_some_attribute_generic_thing()
        set_appropriate_attribute()        

    def getvar(self):
        do_some_attribute_generic_thing()
        return appropriate_attribute()

    foo = property(getvar, setvar)
    bar = property(getvar, setvar)

=====================================

Is it possible to determine which attribute was used to as the trigger for the 
call to setvar() or getvar()? so that set_appropriate_attribute() or 
appropriate_attribute() can set the correct attribute or get the correct 
attribute?

Thanks.

-- 
Stand Fast,
    tjg.

Timothy Grant
www.craigelachie.org





From edcjones at erols.com  Wed Apr 23 11:12:46 2003
From: edcjones at erols.com (Edward C. Jones)
Date: Wed, 23 Apr 2003 11:12:46 -0400
Subject: Pyrex pointer problem
Message-ID: <b86aio$2b3$1@bob.news.rcn.net>

I use Pyrex 0.7.2 and Python 2.2.2. For line "return self.t.as" in
"wrapped.pyx", I get the error message:

     Cannot convert 'A_Struct' to Python object

How do I fix this? How can I get a list of Python and C variables (or
objects) for a program?

--------------
astruct.h:

typedef struct {
     int iVal;
     struct A_Struct* as;
} A_Struct;

--------------
astruct.c

#include "astruct.h"

--------------
wrapped.pyx:

ctypedef int size_t

cdef extern from "stdlib.h":
     void* malloc(size_t size)
     void free(void* ptr)

cdef extern from "astruct.h":
     ctypedef struct c_A_Struct "A_Struct":
         int iVal
         c_A_Struct* as

cdef class A_Struct:
     cdef c_A_Struct* t

     def __new__(self, i):
         self.t = <c_A_Struct*> malloc(sizeof(c_A_Struct))
         if self.t == NULL:
             raise MemoryError
         self.t.iVal = i
         self.t.as = self.t

     def __dealloc__(self):
         free(self.t)

     def __getattr__(self, name):
         if name == 'iVal':
             return self.t.iVal
         if name == 'as':
             return self.t.as
         return object.__getattribute__(self, name)

     def __setattr__(self, name, val):
         if name == 'iVal':
             self.t.iVal = val
         if name == 'as':
             self.t.as = val
         else:
             object.__setattr__(self, name, val)




From forodejazz at yahoo.es  Tue Apr  8 03:49:15 2003
From: forodejazz at yahoo.es (=?ISO-8859-1?Q?David_Asorey_=C1lvarez?=)
Date: 8 Apr 2003 00:49:15 -0700
Subject: Python & Music: Performance question
References: <1759581e.0304060218.164d4cd4@posting.google.com> <1759581e.0304070549.3b3a4caa@posting.google.com> <3E91C3D9.E274C521@easystreet.com>
Message-ID: <1759581e.0304072349.656e94a1@posting.google.com>

achrist at easystreet.com wrote in message news:<3E91C3D9.E274C521 at easystreet.com>...
> Is there anything in Python like the Midi support in the java
> sdk?  (under MS Windows) It's pretty easy in java to send midi data 
> straight from the program to the midi sequencer or synthesizer
> (no need to write a file or deal with some external player).   
> Anything like that for Python?
> 
> Al
>      

  I've been looking for this, and I've found Solfege
(http://solfege.sourceforge.net/), an excellent eartraining software
written in Python.
  Solfege can "talk" with the sound card. It uses two modules,
winmidi.pyd in the win32 version and solfege_c_midimodule.so in the
linux version. There are some "pure python" code wrapping this
modules. I asked the author, Tom Cato Amundsen, some details and he
answered me fastly and kindly, giving me some useful advices. You can
ask him for the details and study solfege's source code.
  I've developed an small tuner (is only an _experiment_, with dirty
code and dirty design) using the Solfege's libraries
(http://www.geocities.com/forodejazz/programacion/pytuner/pytuner.tar.gz
--sorry, the web page is in Spanish). Maybe you would like see the
code.
   
  It would be very nice some standard midi module, but at this time,
there isn't anything like this. Actually, we need write a MIDI file
and use some external player.

David

P.S.: Again, excuse my poor English ;-)



From machowsk1 at attbi.com  Sun Apr 27 20:27:08 2003
From: machowsk1 at attbi.com (Zipper114)
Date: Mon, 28 Apr 2003 00:27:08 GMT
Subject: Sorting files
References: <ATZqa.667651$S_4.718911@rwcrnsc53> <Y7_qa.934482$sj7.37993441@Flipper>
Message-ID: <wz_qa.114422$gK.218859@rwcrnsc52.ops.asp.att.net>

Yup, that was basically it. I changed the "stat.ST_CTIME" to "stat.ST_MTIME"
to get the behaviour I wanted, but that was it. Thanks!


"Nicola Paolucci" <durdn at yahoo.it.oops!.invalid> wrote in message
news:Y7_qa.934482$sj7.37993441 at Flipper...
> Hi Zipper,
>
> Zipper114 wrote:
> > I'm a bit of a noob when it comes to Python, so this might seem a bit
> > trivial, but.....
> >
> > How can I get the listing of files of a particular type in a directory,
> > sorted by date? For example, I using the following call to get all the
> > JPEG's in a dir:
> >
> > Images = glob.glob1(MyPath+"\\images", "*.jpg")
> >
> >
> > I want to do just this, but have them sorted by date. Any ideas?
>
> You could use the os.stat function and the stat module like:
>
> import os
> import pprint
> import stat
>
> p = 'f:/mp3/italiane/Franco Battiato'
> tosort = []
>
> for f in os.listdir(p):
>      tosort.append((os.stat(os.path.join(p,f))[stat.ST_CTIME],f))
>
> tosort.sort()
> pprint.pprint([r[1] for r in tosort])
>
>
> I hope this helps,
> ciao,
> Nick
>
> -- 
> #Remove .oops!.invalid to email or feed to Python:
> 'Tmljb2xhIFBhb2x1Y2NpIDxuaWNrQG5vdGp1c3RjYy5jb20+'.decode('base64')
>





From mcfletch at rogers.com  Wed Apr 16 15:25:08 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Wed, 16 Apr 2003 15:25:08 -0400
Subject: wxPython & PyOpenGL
In-Reply-To: <pDcna.48562$t_2.4267@afrodite.telenet-ops.be>
References: <pDcna.48562$t_2.4267@afrodite.telenet-ops.be>
Message-ID: <3E9DAE14.7030506@rogers.com>

wxPython definitely works with PyOpenGL on Win2K (it's the primary 
development platform for OpenGLContext :) ).  What's wrong with your 
code is that you never tell OpenGL the size of the window into which it 
should render.  Try this:

    def OnPaint(self, event):
        dc = wxPaintDC(self)
        self.SetCurrent()
        size = self.GetClientSize()
        glViewport( 0,0, *size )
        DrawGL()
        self.SwapBuffers()

HTH,
Mike

Roel Schroeven wrote:

> Hi,
>
> I'm trying to make a small application using wxPython to provide a GUI 
> and PyOpenGL hardware acceleration. I know I have to create a 
> wxGLCanvas, which will be used by PyOpenGL functions. I wrote some GUI 
> code and adapted the code from the first example in the Red Book to 
> fit in, and it works. Sorta. I can set the background color, but 
> nothing else. No exceptions are thrown, the canvas just doesn't 
> display anything (besides the background color).
>
> In case it has anything to do with it, I'm using ActivePython on 
> Windows 2000 (though I'm planning to do most of the development on 
> Linux).
>
> Am I doing something wrong? Is there a problem with 
> ActivePyton/Windows/wxPython/PyOpenGL? Any toughts are more than welcome.
>
> The code I have so far:
>
>
> from wxPython.wx import *
> from wxPython.glcanvas import *
> from OpenGL.GL import *
>
> def DrawGL():
>     glClearColor(0.8, 0.1, 0.1, 0.0)
>     glClear(GL_COLOR_BUFFER_BIT)
>     # (so far it works)
>
>     glColor3f(1.0, 1.0, 1.0)
>
>     glMatrixMode(GL_PROJECTION)
>     glLoadIdentity()
>     glOrtho(-1.0, 1.0, -1.0, 1.0, -1.0, 1.0)
>
>     glBegin(GL_POLYGON)
>     glVertex2f(-0.5, -0.5)
>     glVertex2f(-0.5, 0.5)
>     glVertex2f(0.5, 0.5)
>     glVertex2f(0.5, -0.5)
>     glEnd()
>     # (but all this has no effect whatsoever)
>
>     glFlush()
>
>
> class MyGL(wxGLCanvas):
>     def __init__(self, parent):
>         wxGLCanvas.__init__(self, parent)
>         EVT_PAINT(self, self.OnPaint)
>
>     def OnPaint(self, event):
>         dc = wxPaintDC(self)
>         self.SetCurrent()
>         DrawGL()
>         self.SwapBuffers()
>
> class MyFrame(wxFrame):
>     def __init__(self, parent, ID, title):
>         wxFrame.__init__(self, parent, ID, title, wxDefaultPosition, 
> wxSize(700, 500))
>         self.GLCanvas = MyGL(self)
>
>         MainSizer = wxBoxSizer(wxVERTICAL)
>         MainSizer.Add(wxButton(self, -1, "top"), 0, wxEXPAND)
>
>         MiddleSizer = wxBoxSizer(wxHORIZONTAL)
>         MiddleSizer.Add(wxButton(self, -1, "left"), 1, wxEXPAND)
>         MiddleSizer.Add(self.GLCanvas, 2, wxEXPAND)
>         MiddleSizer.Add(wxButton(self, -1, "right"), 1, wxEXPAND)
>         MainSizer.Add(MiddleSizer, 1, wxEXPAND)
>
>         MainSizer.Add(wxButton(self, -1, "bottom"), 0, wxEXPAND)
>
>         self.SetSizer(MainSizer)
>         self.SetAutoLayout(True)
>         #MainSizer.Fit(self)
>
> class MyApp(wxApp):
>     def OnInit(self):
>         frame = MyFrame(NULL, -1, "Python/OpenGL")
>         frame.Show(True)
>         self.SetTopWindow(frame)
>         return True
>
> app = MyApp(0)
> app.MainLoop()
>
>
>

-- 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From m.faassen at vet.uu.nl  Mon Apr 28 16:37:36 2003
From: m.faassen at vet.uu.nl (Martijn Faassen)
Date: 28 Apr 2003 20:37:36 GMT
Subject: Web programming and a different "type" problem
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com>
Message-ID: <b8k3eg$l0s$1@newshost.accu.uu.nl>

Cameron Laird <claird at lairds.com> wrote:
> What's a better way? 

In Zope you can exploit the buit-in marshaller to convince it to do
the list thing:

<input type="text" name="foo:list" />

The ':list' gets picked up by Zope and then it does the right thing.

I've written a product for Zope called 'Formulator' that basically
solves this problem too, for most simple web forms. There are certainly
other systems for Zope (and elsewhere) that can do this kind of thing.
For Zope 3 work is being done on Zope 3 Schema/Forms that can tackle
this.

> Is there even a better way to 
> think about this, in terms of basic Web standards?

I don't know about web standards.. XForms perhaps, but I'd hardly call
that 'basic'. :)

Regards,

Martijn
-- 
History of the 20th Century: WW1, WW2, WW3?
No, WWW -- Could we be going in the right direction?



From duncan at NOSPAMrcp.co.uk  Thu Apr 10 06:00:58 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 10 Apr 2003 10:00:58 +0000 (UTC)
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com> <mailman.1049901770.14045.python-list@python.org>
Message-ID: <Xns93596E7C1B227duncanrcpcouk@127.0.0.1>

Jim Meyer <jmeyer at pdi.com> wrote in news:mailman.1049901770.14045.python-
list at python.org:

> Toss this one onto the heap; it feels a bit more pythonic and
> straightforward than some of the solutions offered:
> 
>   byLen = lambda a,b : cmp(len(a), len(b))
>   words = 'abcd/df/a/iiwk/abcdefghijkl/b/c'.split('/')
>   words.sort(byLen)
>   print words[-1]

I think that if you are going to do it this way it would be more 'pythonic' 
to use a decorated sort (left as an exercise for the reader).

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From selwyn at home.net.nz  Wed Apr 23 23:48:13 2003
From: selwyn at home.net.nz (selwyn)
Date: Thu, 24 Apr 2003 15:48:13 +1200
Subject: Using the MySQLdb module
In-Reply-To: <3EA6E75E.7070305@cis.ohio-state.edu>
References: <3EA6E75E.7070305@cis.ohio-state.edu>
Message-ID: <M%Ipa.4767$AB5.671501@news02.tsnz.net>

Michael Mayhew wrote:
> Greetings,
> 
>     I am trying to use the MySQLdb module on a simple database local to 
> my personal Linux machine. However, I keep running into trouble when I 
> execute SQL on my cursor objects, specifically when I execute the insert 
> command. I have been through numerous docs on the Python DB API v2.0 and 
> on the module itself but have not turned up anything useful. Any 
> assistance would be greatly appreciated. Here is the error message:

I found the following article from devshed pretty useful:

http://www.devshed.com/Server_Side/Python/PythonMySQL/page1.html

regards,
Selwyn





From abpillai at lycos.com  Mon Apr  7 07:01:35 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 7 Apr 2003 04:01:35 -0700
Subject: os.chmod() question ?
References: <bd993a2f.0304022308.19f36fec@posting.google.com> <slrnb8oqd5.174.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <bd993a2f.0304070301.588d1184@posting.google.com>

Yeah yeah. I was prefixing an 'o' instead of a '0'. 
Stupid me. No wonder I never got it right ! :-O

Thanks
Anand Pillai
Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> wrote in message news:<slrnb8oqd5.174.do-not-spam-ben.hutchings at tin.bwsint.com>...
> In article <bd993a2f.0304022308.19f36fec at posting.google.com>,
> Anand B Pillai wrote:
> > Hi pythonfolks,
> >  
> >     What numerical argument does os.chmod() accept on windoze .
> > For a unix box, I used to give the following for files:
> > 
> > chmod - 777 => All rwx permissions enabled (-rwxrwxrwx)
> > chmod - 644 => RW for owner, r for group/world (-rw-r--r--)
> > 
> > I attempted os.chmod() command on Windoze on similar lines
> > on a read-only file.
>  <snip> 
> >     Apparently, os.chmod() works differently from original chmod() command.
> > On a unix box, the results were totally different.
> > 
> > Do I need to pass an octal integer as the second argument ?
> 
> The Python interpreter can't interpret literals in different bases
> according to which function you call, so you must put a '0' at the
> beginning of octal numbers.
> 
> Anyway, Windows file permissions work in a totally different way
> from Unix permissions.  The only simple flag is the R (read-only)
> attribute.  In NT/2000/XP all other permissions are controlled by
> an ACL.  In toy Windows everything else is always permitted.
> 
> Under Windows, I think os.chmod() ignores everything except the
> owner-writable flag (0200) and sets the R attribute to the inverse
> of that.  Cygwin will show permissions as either 0777 or 0555
> depending on the state of the R attribute.
> 
> So 777 decimal = 01411 octal which means the R attribute is set
> and the permissions read as 0555, and 644 decimal = 01204 octal
> which means the R bit is cleared and the permissions read as 
> 0777.  All clear?



From dan at grassi.org  Tue Apr  1 16:58:12 2003
From: dan at grassi.org (Dan Grassi)
Date: Tue, 1 Apr 2003 21:58:12 +0000 (UTC)
Subject: How to obtain python file path
References: <20030401163212631-0500@news.covad.net> 
	<slrnb8jusc.5u1.ramen@lackingtalent.com>
Message-ID: <20030401165546102-0500@news.covad.net>

Not quite, that only provides the path to the initial .py file.

With a directory structure:
fa
    a.py
    fb
        b.py

cd .../fa
execute a.py
where a adds fb to sys.path
a imports b.py

I need b.py to report it's path which would be .../fa/fb

Dan


In <slrnb8jusc.5u1.ramen at lackingtalent.com> Dave Benjamin  wrote:
> import os, sys
> print os.path.abspath(sys.argv[0])
> 



From spam at mattbergin.co.uk  Sat Apr 12 14:08:06 2003
From: spam at mattbergin.co.uk (Matt Bergin)
Date: Sat, 12 Apr 2003 19:08:06 +0100
Subject: Reference Tracking
Message-ID: <b79km7$iur$1@wisteria.csv.warwick.ac.uk>

I am trying to write a class which gives out objects and keeps track of 
them so they can be returned to the class and given out to other code 
when the code they were given out to no longer has any reference to the 
objects (i.e. has finished using it).

The trouble is, the manual says not to try and stop objects from being 
killed by creating new references in __del__. My original thoughts were 
to write some code in __del__ of the given out object that returned it 
to the original class. Is this is a good way of going about this 
problem? Any ideas would be welcome!

-- 
Matt Bergin




From skip at pobox.com  Thu Apr 10 13:12:25 2003
From: skip at pobox.com (Skip Montanaro)
Date: Thu, 10 Apr 2003 12:12:25 -0500
Subject: python 2.3a2 Win32 expected failures in regression tests?
In-Reply-To: <200304101855.25404.harri.pasanen@trema.com>
References: <mailman.1049910795.4707.python-list@python.org>
        <e2d925a8.0304100803.46e63240@posting.google.com>
        <16021.39020.38164.886501@montanaro.dyndns.org>
        <200304101855.25404.harri.pasanen@trema.com>
Message-ID: <16021.42489.672529.853275@montanaro.dyndns.org>

    Harri> Are there CVS snapshot tarballs available?  Trying to log in, I'm
    Harri> getting

    Harri> cvs [login aborted]: reading from server: Connection reset by peer

I don't believe anyone's been making CVS snapshots.  I'll be happy to "cvs
up" and make it available for you though.  (Let me know.)

I believe the cvs errors are simply due to load at SourceForge.  They've
been having a lot of load problems lately.  Apparently, they give higher
priority to authenticated cvs access than to anonymous access.  I don't know
that there's a good solution.

Skip





From jepler at unpythonic.net  Fri Apr 18 14:33:43 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 18 Apr 2003 13:33:43 -0500
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <b7pchr$ig0$1@main.gmane.org>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <b7pchr$ig0$1@main.gmane.org>
Message-ID: <20030418183341.GA18062@unpythonic.net>

On Fri, Apr 18, 2003 at 10:28:18AM -0700, Mike Rovner wrote:
> "Martin v. L?wis" <martin at v.loewis.de> wrote in message
> 
> > > Is there a short way to say why Guido didn't want continuations?
> >
> > He didn't like the implementation strategy; I believe he was neutral
> > about the feature per se.
> 
> Second that.
> AFAIK one of the points was/is the ability to implement it in Jython,
> which is/was impossible while you are mangling with C stack frames.

AFAIK, generator functions have turned out to be quite difficult in Java
too, but that didn't seem to stop their acceptance.

Jeff




From ckea25d02 at sneakemail.com  Sun Apr  6 23:30:31 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 15:30:31 +1200
Subject: Calling C libraries from a Python script
In-Reply-To: <b6n4rn$6uitb$1@ID-139865.news.dfncis.de>
References: <e81e0f3d.0304050639.37b59655@posting.google.com> <qmCja.4148$Du.14013@newsc.telia.net> <b6n4rn$6uitb$1@ID-139865.news.dfncis.de>
Message-ID: <b6qqr9$82i1b$1@ID-169208.news.dfncis.de>

Dave Kuhlman wrote:
> Also look at SWIG.

Also have a glance at Pyrex:

   http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From sholden at holdenweb.com  Thu Apr 17 12:02:50 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 16:02:50 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr>    <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk>    <3E9EB3E9.1040208@removethis.free.fr> <87k7dt5m3j.fsf@interim.henrik-motakef.de>    <1azna.115402$It5.22896@news2.central.cox.net> <20030417114316.M13181@lain.cheme.cmu.edu>
Message-ID: <KeAna.115813$It5.26090@news2.central.cox.net>

"Matthew Danish" <mdanish at andrew.cmu.edu> wrote in message
news:20030417114316.M13181 at lain.cheme.cmu.edu...
> On Thu, Apr 17, 2003 at 02:49:33PM +0000, Steve Holden wrote:
> > I can see that, but wouldn't you agree that removing redundant features
also
> > removes cruft from the source, and eases the maintenance burden, in turn
> > reducing the likelihood of implementation bugs? While Guido is known to
> > regret the inclusion of lambda in Python, I don't think he would want to
> > take it out on purely ideological grounds. Indeed, the suggestion was
not so
> > much that it would be taken out of the existing language, rather that it
> > would not be implemented in some "futurePython", normally called Python
3 or
> > Python 3000.
> >
> > It would be nice to have One True Python, but that's about as likely as
One
> > True Lisp. The Python world is pragmatic enough to be able to live with
"one
> > *most obvious* way to do it", but if there are two similar features in
the
> > language they have to be maintained in lockstep as the language
develops.
>
> What's wrong with Python and CL!!!?  They have all these extraneous
> features which can accomplish the same thing!!! All you need is the
> <insert model of computability here>!!! Anything else is liable to
> develop cruft and cause bugs in the single implementation of our ONE
> PURE LANGUAGE!!!  I can't believe that people might want to model
> problems any other way!!!
>

Absolutely. Your pure approach causes me to shed tears of joy. If only
everyone else could see the Truth. In fact, were there endless time, all we
would need would be One True Turing Machine and an Infinite Number of
Monkeys.

Did I mention that you only need nine commands to write any program? ;-)

and-we-could-dispense-with-the-turing-machine-at-a-pinch-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From Kyler at news.Lairds.org  Thu Apr 10 18:23:16 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Thu, 10 Apr 2003 22:23:16 GMT
Subject: Zope large PDF question
References: <e09988ac.0304070214.4212e2e9@posting.google.com> <oqh8m-ldq.ln1@news.lairds.org> <e09988ac.0304100345.122b8ab4@posting.google.com>
Message-ID: <qo5hm-mpd.ln1@news.lairds.org>

On comp.lang.python, davec999 at yahoo.co.uk (David Currie) writes:

>> As I recall, Reader sends a request for pieces of a
>> large file
>> 	1-500, 1001-1500, 501-1000
>> and Zope decides to be efficient and join all of the
>> contiguous regions.  Unfortunately, that defeats the
>> purpose of requesting the ranges *and* breaks Reader.

>I see.  Is there any chance of explaining how to disable this
>behaviour, or even better, of sharing the patch? ;)

Oh, fine...dig, dig, dig...  Shouldn't this be online
somewhere?  I posted the patch on a Zope mailing list.
dig, dig, dig...  Grrr...it's not in my past year's 
mail.

Lessee...the problem is in OFS/Image.py.  It doesn't
look like my current Zopes have the patch, but you
should be able to disable byterange support by commenting
out all of these lines.
	RESPONSE.setHeader('Accept-Ranges', 'bytes')
You might also need to set range/request_range to None.

Sorry I'm not clearer on this.  I've copied the list in
case someone with more understanding feels like jumping
into the discussion.

--kyler



From jkraska at san.rr.com  Sat Apr 12 13:21:30 2003
From: jkraska at san.rr.com (Courageous)
Date: Sat, 12 Apr 2003 17:21:30 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net> <PAXla.4934$LB6.139412@news1.tin.it>
Message-ID: <njig9vovtcptn8eropgqpa8rlvfe9prq2r@4ax.com>

>Let's put it this way: in C++ you have to use this->foo SOME of the
>time (to avoid running afoul of extremely subtle and deuced bugs
>in some templates) but it's quite hard to predict exactly when you'll
>need to do it.  With Python, you ALWAYS use self.foo (one character
>less than C++'s this->foo!-) and thus your code is more regular and
>uniform, there is no risk of ugly conventions such as naming member
>variables m_this and m_that (popular in several C++ shops), no bugs
>due to accidental local "hiding" of a member variable by a local one
>(I've taught and consulted on C++ long enough to have seen THAT
>happen many, MANY times -- and not to idiots, either... rather, to
>very good professional programmers!) -- a HUGE amount of conceptual
>AND pragmatical simplification of your life.

Alex Martelli demonstrates once again a fine talent at burying the
hatchet. I rather doubt there can *be* much of a reply to this fine
retort.

C//




From hbl at st-andrews.ac.uk  Tue Apr  8 10:44:02 2003
From: hbl at st-andrews.ac.uk (Hamish Lawson)
Date: Tue, 08 Apr 2003 15:44:02 +0100
Subject: PEP305 csv package: from csv import csv?
Message-ID: <5.2.0.9.0.20030408140539.01cc6220@spey.st-andrews.ac.uk>

According to the documentation in progress at

     http://www.python.org/dev/doc/devel/whatsnew/node14.html

use of the forthcoming csv module (as described in PEP305) requires it to 
be imported from the csv package:

     from csv import csv

     input = open('datafile', 'rb')
     reader = csv.reader(input)
     for line in reader:
         print line

Is there some reason why the cvs package's __init__.py doesn't import the 
required names from cvs.py, so allowing the shorter form below?

     import csv

     input = open('datafile', 'rb')
     reader = csv.reader(input)
     for line in reader:
         print line


Hamish Lawson





From patrick at fraley.de  Fri Apr 18 05:11:16 2003
From: patrick at fraley.de (Patrick W. Fraley)
Date: Fri, 18 Apr 2003 11:11:16 +0200
Subject: Zope Sequences and Python Scripting
References: <pan.2003.04.17.12.34.15.777976@fraley.de> <GN6dnaRxi5queAOjXTWcpw@august.net> <pan.2003.04.17.19.26.22.654130@fraley.de> <mailman.1050614909.20594.python-list@python.org>
Message-ID: <pan.2003.04.18.09.11.15.501791@fraley.de>

On Thu, 17 Apr 2003 16:22:40 +0000, Evan Simpson wrote:

> 
> Yep - your feeling is correct, and you'll get *much* better answers to 
> Zope questions on the Zope mailing lists than here on c.p.g.

Thanks for the advice.  I just subscribed to it.  There are a lot of
lists, but I am assuming the general zope list is what I want right?

Patrick



From skip at pobox.com  Sun Apr 13 18:33:12 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sun, 13 Apr 2003 17:33:12 -0500
Subject: PYTHONPATH question
In-Reply-To: <mb3j9vkgifh5ghc38mja4mhanc9urmuu76@4ax.com>
References: <E1927CG-0007k5-00@fast.aub.edu.lb>
        <mailman.1049632149.9888.python-list@python.org>
        <5ai29vogfthpqvbmlic09pbllltun61d2l@4ax.com>
        <mailman.1049729471.8104.python-list@python.org>
        <mb3j9vkgifh5ghc38mja4mhanc9urmuu76@4ax.com>
Message-ID: <16025.58792.207161.8274@montanaro.dyndns.org>

    nomad> PYTHONPATH=$PYTHONPATH:<colon-separated list of directories>
    nomad> export PYTHONPATH

    >> What's the interpretation of an empty element in $PYTHONPATH?

    nomad> My thinking was rather that, if there was a system-wide
    nomad> $PYTHONPATH set in /etc/profile, it would be overwriten by any
    nomad> $PYTHONPATH statement in ~/.profile that didn't explicitly
    nomad> include it, and that might bugger up other 3rd party modules.

Sure enough.  (I generally only program on single user systems, so the
/etc/profile thing wouldn't sting me.)  How about:

    PYTHONPATH=$PYTHONPATH${PYTHONPATH:+:}<colon-separated list of dirs>

Skip





From a.schmolck at gmx.net  Tue Apr  8 19:42:33 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 09 Apr 2003 00:42:33 +0100
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com> <slrnb96fei.ul2.news@cipw08.physik.uni-bonn.de>
Message-ID: <yfsel4cimja.fsf@black132.ex.ac.uk>

Carsten Gaebler <news at snakefarm.org> writes:

> Heiko Henkelmann wrote:
> > I need to split a string into an array of equally long substrings, like
> > e.g.:
> > 
> > '1234567890'   -> ['12', '34','56','78','90']
> > 
> > What is the most efficient way to do this?
> 
> I don't know how efficient it is but this is what came to my mind:
> 
> import re
> filter(None, re.split('(.{2})', '1234567890'))
> 
> cg.
> 

why not just:

>>> import re
>>> re.findall('..', '123456')
['12', '34', '56']


alex



From dale at riverhall.NOTHANKS.co.uk  Mon Apr 28 14:59:31 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Mon, 28 Apr 2003 19:59:31 +0100
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <s4uqavg0fjuucqve2vl4g03atpu25pdn97@4ax.com>

Frantisek Fuka <fuka at fuxoft.cz> wrote:

>When I try to call methods, I sometimes forget to include the 
>parentheses. Instead of:
>
>if object.isGreen():
>	do something...
>
>i sometimes write:
>
>if object.isGreen:
>	do something...
>

It's the curse of Visual Basic.

I suggest you take your frustration out on the person who persuaded
you to learn VB in the first place.

Recovery may be found in hours of coding in just about any grown-up
language.

-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From aleax at aleax.it  Sat Apr 12 11:07:57 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 15:07:57 GMT
Subject: hex and unsigned longs
References: <xjvof3bu75x.fsf@margvisi.ifi.uio.no>
Message-ID: <hZVla.4413$LB6.126815@news1.tin.it>

Igor V. Rafienko wrote:

> The documentation for the built-in function hex() says that hex()
> always returns an unsigned literal. However, on my 2.2.1 python, I get

Yes, I believe you've identified a bug in the docs.

> the following results:
> 
>>>> hex( 1 )
> '0x1'
>>>> hex( 1L )
> '0x1L'
>>>> hex( -1 )
> '0xffffffff'
>>>> hex( -1L )
> '-0x1L'
>>>> 
> 
> The literal in the last case is obviously not unsigned. What am I
> missing? (I.e. why are longs treated differently than ints by hex?)

hex is trying to ensure eval(hex(x))==x (when hex and eval are taking
place on the same machine, or more generally on machines of the same
word size).  hex just can't ensure that when type(x) is long, except
by using an explicit - sign; therefore, the statement in the docs that
"this always yields an unsigned literal" is overbroad -- it should be
slightly nuanced to "this always yields an unsigned literal for any
int argument", IMHO.  You can submit a documentation patch (or even just
a bug report) to sourceforge, just like you can submit bug reports or
patches for the interpreter and library parts of Python (it's a good
idea to doublecheck here, as you've done, to ensure you've actually
found an outright bug, like in this case -- bravo!).


Alex




From ianb at colorstudy.com  Tue Apr  8 05:36:05 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 08 Apr 2003 04:36:05 -0500
Subject: ANN: SQLObject 0.3
In-Reply-To: <b6u2tj$9ea$1@news-reader12.wanadoo.fr>
References: <mailman.1049776336.2246.python-list@python.org>
	 <b6u2tj$9ea$1@news-reader12.wanadoo.fr>
Message-ID: <1049794565.735.28.camel@lothlorien>

On Tue, 2003-04-08 at 03:57, Philippe Lafoucri?re wrote:
> > SQLObject 0.3 has been release:
> >   http://sqlobject.org
...
> > What is it?
> > -----------
> > 
> > SQLObject is an object-relational mapper.  It allows you to translate
> > RDBMS table rows into Python objects, and manipulate those objects to
> > transparently manipulate the database.
> > 
> > SQLObject emphasizes convenience.  It's easy to learn and easy to use,
> > and doesn't require too much typing or use of any tools outside of your
> > editor and your Python source.
> > 
> It is quite similar to the Twisted Matrix Row object !?

They are both object-relational mappers, several of which exist for
Python.  I hadn't looked at the Twisted ORM before, but it's no more
similar to SQLObject than the other ORMs -- they all do similar things
so of course they all look somewhat similar.  I think SQLObject is
considerably more sophisticated and more convenient than Twisted's
RowObject.  

SQLObject also creates objects that persist with less effort -- you
don't add them to a store, or fetch them from a store, or explicitly
save them back to the store.  The store is the object.  This is
different from many of the other ORMs for Python.

  Ian






From no at spam.invalid  Wed Apr 30 14:02:12 2003
From: no at spam.invalid (Russell E. Owen)
Date: Wed, 30 Apr 2003 11:02:12 -0700
Subject: Some oddities in 2.3b1, esp. with Tkinter
References: <b8mpp1$25hm$1@nntp6.u.washington.edu> <3EAEF1D6.30900@v.loewis.de>
Message-ID: <b8p336$27pg$1@nntp6.u.washington.edu>

In article <3EAEF1D6.30900 at v.loewis.de>,
 "Martin v. Lowis" <martin at v.loewis.de> wrote:

>Russell E. Owen wrote:
>> 1) A filter to remove fonts with non-ascii names (used to filter out 
>> stuff that I could not display in a pop-up menu) broke.
>>             try:
>>                 unicode(fontName, 'ascii', 'strict')
>>                 return True
>>             except UnicodeError:
>>                 return False
>> 
>> Python 2.2 throws UnicodeError. Python 2.3b1 throws TypeError. The fix 
>> is trivial (e.g. "except (UnicodeError, TypeError)". The filter is 
>> probably not needed anymore anyway, since Tk 8.4.2 can apparently 
>> display menu items with non-ascii characters.
>
>I don't understand. Why do you ever need to convert the *font name*
>to a Unicode string, or why would Python ever fail to do so? Font
>names are ASCII only, all the time, so you can't possibly ever get
>an exception here.

This is not true on MacOS X. I found it out the hard way. The font names 
with unicode appear to be asian fonts, at least for the most part.

Originally I naively just grabbed all font names (as returned by a 
tkFont.Font object) and tried to display them in an OptionMenu. Boy did 
I see a lot of error messages (about undisplayable glyphs). As I say, I 
think Tk 8.4.2 can actually display unicode in menus now, but even 
so...I don't think any of my users will be wanting to use the fonts with 
non-ascii names and it keeps the list down to a semi-sane size.

>> 2) Tk file events are no longer supported:
>> 
>> RuntimeError: _tkinter.createfilehandler not supported for threaded Tcl
>> 
>> This is painful since it's hard to see how to write a proper networking 
>> GUI with Tkinter without file events for reading data. Perhaps I can get 
>> them back again by disabling the new threaded tcl support.
>
>Are you using Debian by any chance? Otherwise, you should not see
>a threaded Tcl.

Hmm...I installed Aqua Tcl/Tk 8.4.2 from a MacOS X binary installer I 
found on the net. I'm not sure what options it was built with (though 
I'd have guessed the default ones). I can certainly install it from 
source if that would help.

>If your Tcl installation is threaded, you should use 
>tkapp.createfilehandler instead.

Hmmm...so I can first try my usual way of creating a file handler, and 
if that fails, then try this? I'll do it. It'd be great to not have to 
overhaul my networking code.

>> 3) Tk returning things as objects is nice, but I had this break at least 
>> one thing of mine and it will break others (see 4 below):
>> 
>> I was assuming that the values in the dictionary returned by 
>> tkFont.Font.config() were strings. Fortunately this was easy to fix. A 
>> bit disconcerting -- I wonder how many any other little pitfalls like 
>> this still buried in my code -- but well worth the cost of the change.
>
>What is the nature of that change?
>
>> Example:
>> v = StringVar()
>> v is a <Tkinter.StringVar instance at 0x2c98a0>
>> l = Label(textvariable=v)
>> x = l["textvariable"]
>> x is reported as a <namespaceVarName object at 0x0168b048>
>> but is a _tkinter.Tcl_Obj based on the following traceback:
>> root.getvar(x)
>> Traceback...
>> TypeError: getvar() argument 1 must be string, not _tkinter.Tcl_Obj
>
>Please submit a bug report. This is easy to fix (just invoke
>str() of the Tcl_Obj)

I'd be happy to. Could you please tell me where the Tkinter project 
lives?

>> 5) Note that Tkinter's handling of font objects does not seem to have 
>> changed. I was hoping it would start returning font objects instead of 
>> strings, or offer some way to convert the string to a font object, but 
>> apparently not:
>
>Does Tk have the notion of "font objects"?

tkFont.Font is the Tkinter version of Tk's named fonts. I have no idea 
if Tk's named fonts are objects. In fact I'd never even heard of tcl 
objects until reading about the recent Tkinter changes.



From mwh at python.net  Wed Apr 16 16:14:33 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 20:14:33 GMT
Subject: compatibility problem
References: <1f4c702e.0304161116.5b0f6565@posting.google.com>
Message-ID: <7h33ckiqjx3.fsf@pc150.maths.bris.ac.uk>

chad.scherrer at pnl.gov (Chad Scherrer) writes:

> I'm having a compatibility problem, and I'm wondering if anyone might
> be able to point me in the right direction. I'm running some code was
> written for python 2.1, and I'd like to make it run under 2.2, so I
> can subclass built-in types. The code is about 15,000 lines, with lots
> of different classes running around, but the gist of the problem is
> finding a solution set to an equation based on input from a file.
> Python 2.1 gives the correct solution, but somehow 2.2.2 is only
> giving me a subset. Not only that, but the order is different! The
> ordering is arbitrary anyway, but I would still expect it to be
> consistent between the two.
> 
> So I'm wondering what I should look for to make this work under 2.2.2.
> My first thought was nested scopes, but the PEP I read suggests that
> Python 2.1 gives a warning for any case that might cause a problem in
> 2.2. I don't think it could be the method resolution order, because
> object (or any built-in type) is never subclassed. Is there anything
> else I should be looking out for?

The order of elements in dict.keys() springs to mind.  That changed
2.1 -> 2.2.

Cheers,
M.

-- 
  MAN:  How can I tell that the past isn't a fiction designed to
        account for the discrepancy between my immediate physical
        sensations and my state of mind?
                   -- The Hitch-Hikers Guide to the Galaxy, Episode 12



From sholden at holdenweb.com  Thu Apr 17 11:12:05 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 15:12:05 GMT
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl>
Message-ID: <9vzna.115527$It5.95833@news2.central.cox.net>

"Anton Vredegoor" <anton at vredegoor.doge.nl> wrote in message
news:b7m4ql$ril$1 at news.hccnet.nl...
> From the universe of all currently available webpages (where
> "currently available" means available on the world wide web at the
> runtime of the script) I want to select a page in such a way that each
> page has an equally likely chance of being selected. Any idea about
> how to accomplish this is equally likely to be welcome ;-)
>

In order to be able to generate all pages with exactly equal probability you
have to know the cardinality of the set of all pages. Which nobody does,
because it changes all the time.

Assuming you did know how many there were, you'd simply need to build a list
of them all and then use whrandom.choice() to select an arbitrary element.
Assuming you have enough memory to store a list of 12,000,000,000+ elements,
of course ;-)

If you ever do work out a suitable method, please mail me via
http://www.myrandomwebpage.com/ :-)

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From nav at adams.patriot.net  Wed Apr 30 09:09:15 2003
From: nav at adams.patriot.net (Nick Vargish)
Date: 30 Apr 2003 09:09:15 -0400
Subject: Python Tutorials
References: <15451.8030105@earthlink.net> <yyyznm9thov.fsf@adams.patriot.net> <a5415304.0304291603.4171ec39@posting.google.com>
Message-ID: <yyyn0i8rv2c.fsf@adams.patriot.net>

rstephens at vectron.com (Ron Stephens) writes:

> OK, good criticism, Nick. I have gone back and changed the page (and
> the rest of the web site) to use smaller size fonts, specifically 4
> where I had used 6 on the tutorials page.

My suggestion is to not use explicit size settings at all. Most
people's browsers have settings that let them specify a default font
size that's comfortable for them. If you want to make a bit of text
larger for emphasis, you can use a relative sizing argument, which
would look like <font size="+1">.

To your credit, you set the font too large -- most sites that
manipulate the font size make the fonts too small.

> I know I should use a good  CSS style sheet, but I am hopelessly
> inept at 'style' , 'fashion', or 'art' of any kind.

That's a great reason to just leave the defaults well enough
alone. :^)

> Does it look better now?

Looks much better, thanks for listening to my well-meaning, if
tactless, suggestions... 

Nick

-- 
# sigmask.py  ||  version 0.2  ||  2003-01-07  ||  Feed this to your Python.
print reduce(lambda x,y:x+chr(ord(y)-1),'Ojdl!Wbshjti!=obwAqbusjpu/ofu?','')



From sholden at holdenweb.com  Wed Apr 16 09:43:21 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 13:43:21 GMT
Subject: reading Unix mailbox
References: <hZbna.17434$DN.480144@tornado.fastwebnet.it>
Message-ID: <Z5dna.110590$It5.98167@news2.central.cox.net>

"Alessio Pace" <puccio_13 at yahoo.it> wrote in message
news:hZbna.17434$DN.480144 at tornado.fastwebnet.it...
> Hi all,
> I am using Python2.3 alpha and I thought there was something updated to
> iterate over the messages of a mailbox in mbox format, but the next()
> method still return a deprecated RFC822  message object reference, am I
> wrong? Is there a way to read mailboxes in mbox format but iterating over
> MIMEMessages?
> Thanks to all who can help me
>

I seem to rememeber that a recent version (probably 2.2) modified the
mailbox objects so you could provide an optional (callable) factory argument
when creating them, and this would then be called when the mailbox instance
needed to create a message. The default is rfc822.Message, hence the
behavior you are seeing.

Does this help any?

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From gh at ghaering.de  Wed Apr 23 15:34:14 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 23 Apr 2003 21:34:14 +0200
Subject: how to link python against static sys libs?
In-Reply-To: <m3znmhf28y.fsf@mira.informatik.hu-berlin.de>
References: <pan.2003.04.23.12.53.47.253065@nospam-freenet.org.nz>
	<m3znmhf28y.fsf@mira.informatik.hu-berlin.de>
Message-ID: <3EA6EAB6.9040903@ghaering.de>

Martin v. L?wis wrote:
> "David McNab" <david at nospam-freenet.org.nz> writes:
>>I'm having a bit of a nightmare trying to build python to link the sys
>>libs (eg glibc) statically.
> 
> I haven't check recently, but I think static libraries are not really
> supported on Linux anymore. The C libraries relies on dynamic loading
> for certain functions (such has netdb), and there is no way to build a
> statically-linked executable that uses these functions.

You could use an alternative libc in this case. Either an older glibc or 
something like dietlibc.

If you manage to build Python against dietlibc, please report back. I'd 
be interested in wether this works.

-- Gerhard




From brian at rk-speed-rugby.dk  Tue Apr 22 04:11:45 2003
From: brian at rk-speed-rugby.dk (Brian Elmegaard)
Date: 22 Apr 2003 10:11:45 +0200
Subject: Should I prefer an external database
Message-ID: <u3ckbymri.fsf@mail.afm.dtu.dk>

Hi,

I am working on a script for organizing a program for a conference. I
started out using a connection to mySQL, but then I thought that it
would be easier to keep it within python. So I implemented a class:

class Paper:
        def __init__(title,author,session):
                self.title=title
                self.author=author
                self.session=session
I then have a list to which each new paper instance is appended. I can
sort by any of the paper instances variables by using the
decorate-sort-undecorate method.

I now have it working, but I probably could do better, so:

Is this a more pythonic way to implement this?

To do a select I think I need to use for loops specifying the names of
the variables, i.e:

for paper in paperlist:
        if paper.session=='foo':
                print paper.title

Probably this might be done more elegantly, but how do I know which
variables (database columns) the instance has?

Would I be better of with a list of list than with a list of class
instances? 
What about a class of classes? Could such be made and inherit from the
list class to support index and sort?

tia,
-- 
Brian (remove the sport for mail)
http://www.et.dtu.dk/staff/be



From ckea25d02 at sneakemail.com  Tue Apr  1 20:45:54 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 13:45:54 +1200
Subject: The Missing Manual? (Re: Finding Python help...)
In-Reply-To: <x7k7eejgya.fsf@guru.mired.org>
References: <3E88310F.6090709@mq.edu.au> <b6aanj$39skv$1@ID-71831.news.dfncis.de> <b6am14$3f4rk$1@ID-169208.news.dfncis.de> <x7k7eejgya.fsf@guru.mired.org>
Message-ID: <b6det4$48se7$1@ID-169208.news.dfncis.de>

Mike Meyer wrote:
> Instead of a third manual, how about moving the first section of the
> Library Reference - "Built-in Functions, Types, and Exceptions" - into
> the Language reference manual? It's not really thick enough to be
> worth a book by itself.

That would be a possibility. Or maybe all three could be
just one manual with three sections.

I don't really think it matters whether they're thought of
as separate manuals or as sections of one manual. It's
all just a bunch of HTML or whatever files anyway...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From python at dylanreinhardt.com  Sat Apr 19 13:07:41 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 19 Apr 2003 10:07:41 -0700
Subject: Parent references: hot or not?
In-Reply-To: <1050771656.14717.48.camel@ida.dylanreinhardt.com>
References: <1050763980.14721.41.camel@ida.dylanreinhardt.com>
	 <20030419145756.GA25916@meson.dyndns.org>
	 <1050771656.14717.48.camel@ida.dylanreinhardt.com>
Message-ID: <1050772061.14712.55.camel@ida.dylanreinhardt.com>

On Sat, 2003-04-19 at 10:00, Dylan Reinhardt wrote:

> Sounds like I should be sure to write a __del__ method for Container
> that ensures things work correctly.

And upon re-reading your post you say *not* to do that... :-)

I'd like to understand that a bit better.  What would be wrong with:

class Container:

   def __del__(self):
      # ensure all Items are deleted first
      for item in self.items.keys():
         del self.items[item]


Thanks,

Dylan





From salvatore at wanadoo.fr  Fri Apr  4 16:00:26 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Fri, 04 Apr 2003 23:00:26 +0200
Subject: Python and web programming .. my way ?
In-Reply-To: <odfr8v8uuaf9vcgdom0vfmue7p5h4dv6or@4ax.com>
References: <odfr8v8uuaf9vcgdom0vfmue7p5h4dv6or@4ax.com>
Message-ID: <b6krsf$1eq$1@news-reader11.wanadoo.fr>

Svenne Krap wrote:
> Hi.
> 
> First of all, despite being a complete Python-newbie, I consider
> myself a fairly competent PHP-programmer (amongst other things) with
> more than 5 years of experience. 
> 
> My question is, how does you handle page generation nicely in python ?
> In PHP i have adopted a way of generating a gui.php file, which holds
> a gui-object, that takes care of drawing the "framework" php-page
> (that is top, menues and bottom with colors, semi-dynamic content
> (like username for the user) and so on....) 
> 
> Meaning that, a simple website, using my system could be : 
> 
> === example.php ===
> <?php
> include "basics.inc";  
> // includes gui.inc amongst other and gives a $gui object
> $gui->top("frontpage.aboutme.hobbies") ; 
> // the in the above memberfunction the argument 
> // is usually a logical path used for menues 
> ?>
> Hello wolrd.
> <?php
> $gui->bottom();
> ?>
> === example.php ===
> 
> The above code could produce litterally hundreds of lines of html
> (with database-powered dhtml-menues and so on...)...
> 
> While haveing some drawbacks (ie. its not easy to make major
> design-changes, it proves to be a very effective way to write a whole
> website, and the resulting pages are very homogene and the sources are
> very simple and unbelievable easy to maintain afterwards (even for
> people who had prevously never seen the aproach before). It also has
> the benefit that every logical webpage has a physical php-file and
> that deeplinking and such stuff is tremendously easy.
> 
> The mod_python way with a single .py file to handle the whole
> directory seems ackward and clumsey to me. By the way.. is there a way
> to make the PythonHandler to handle the whole website (ie. from the
> document-root and downwards) and if so, is it easy to get the
> requested url ? Then I think, i can figure out a likewise system
> myself.
> 
> Is there any way to do something likewise in Python ? Or is there any
> good reason why my way (tm) is not easily doable in python ? And how
> should it be made instead....
> 
> Svenne

try: CherryPy

A whole site including the server in a single file !!!

http://www.cherrypy.org/

Regards

Salvatore




From tim.one at comcast.net  Mon Apr 28 14:35:24 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 28 Apr 2003 14:35:24 -0400
Subject: Forgetting "()" when calling methods
In-Reply-To: <BfScnR2GHIkjxzCjXTWcoQ@comcast.com>
Message-ID: <BIEJKCLHCIOIHAGOKOLHAEMAFHAA.tim.one@comcast.net>

[Tim]
>> No, the current state makes no sense ... [and] ...is an inconsistent
>> mish-mash

[Terry Reedy]
> OK, we agree on that.  It a result of various local optimazations and
> compromises

It's really the result of Guido changing his mind about whether comparisons
should be allowed to raise exceptions.

> and needs ground up redesign.

Probably, yes.

> My point in relation to () is that I would hate to see a similar
> mishmash created in relation to truth values, especially not for the
> local optimazation of allowing people to forget to type '()' in boolean
> context.

> ...

> The question is whether .sort() should come with at least an option to
> work with a heterogenous items and not raise an exception -- without
> the user having to write a custom order function  -- so programmers who
> know what they are doing when trying to sort such a list can do so
> nearly as easily as they used to.

Without compelling concrete use cases, that's unlikely to happen.  The use
cases are needed to flesh out what guarantees are *expected* of a "senseless
sort".  Presumably proponents want something more than "don't care -- make
up anything at all", else they wouldn't be sorting to begin with <wink>.

I'm personally unhappy with ordering results that aren't guaranteed
consistent across platforms, program runs, and Python releases.  So, e.g.,
the result of

class C:
    pass
a = C()
b = C()
print a < b

today flunks all those criteria.  Comparing complex numbers
lexicographically passes them (so, IMO, default instance comparison (which
is by memory address) is much more mysterious and dangerous than the way
Python used to compare complex numbers -- "dangerous" refers to that
persistent database systems like ZODB using BTree-based data structures act
insanely when objects end up getting compared by accidents of memory address
(ZODB users would be much better off if Python complained whenever an
ill-defined (by the criteria above) compare was requested)).





From aleax at aleax.it  Mon Apr 21 12:11:16 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 21 Apr 2003 16:11:16 GMT
Subject: __setitem__ puzzle
References: <b80q7c$cmd$1@news.hccnet.nl>
Message-ID: <EKUoa.67108$T34.1782963@news2.tin.it>

Anton Vredegoor wrote:

> I have subclassed the builtin list type to accept tuples for indexing.

...apparently to deal with nested lists, if I read you right.  I don't
like that; I've long thought that L[seq] should be [L[x] for x in seq],
and L[seq]=seq1 should be like

    for i, v in zip(seq, seq1): L]i] = v

with the added constraint that len(seq) == len(seq1).


Alex




From dberlin at dberlin.org  Sat Apr 12 20:57:03 2003
From: dberlin at dberlin.org (Daniel Berlin)
Date: Sat, 12 Apr 2003 20:57:03 -0400 (EDT)
Subject: Speed of string += string
In-Reply-To: <Zw0ma.4690$883.4089@nwrdny02.gnilink.net>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <65pjlhdx.fsf@morpheus.demon.co.uk>
 <Zw0ma.4690$883.4089@nwrdny02.gnilink.net>
Message-ID: <Pine.LNX.4.53.0304122054150.11576@dberlin.org>

> Surprisingly, += doesn't appear to be a major time waster for small
> (meaning about 10000 bytes) strings.  I'm guessing it's because the
> malloc my system (Debian Linux) is optimized for small memory
> allocations.

I've got a gnats to bugzilla converter (rewritten from a perl version)
that was taking 15 minutes on a very large (21 meg) pr containing 300k
lines of preprocessed source code in one gnats fields.

The speed was because in my direct conversion, i had transformed .= into
+=

All of the time was spent in a line like:
	fields[hdr] += line

After remember strings are immutable (I've been out of practice with
python for a year or two, since i'm a compiler hacker, not a script
hacker or application developer).
 Appending to a list and joining, or appending to a array('c') takes 4
seconds, instead of the 15 minutes.
:)




From aleax at aleax.it  Tue Apr 29 08:14:26 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 12:14:26 GMT
Subject: Another itertool function?
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <45pra.20060$xw4.19534@nwrdny01.gnilink.net>
Message-ID: <C0ura.19455$3M4.563854@news1.tin.it>

Raymond Hettinger wrote:
   ...
> Here's another idea that was submitted:
> 
> def windowed(iterable, group=2):
>     "returns (i[0], i[1], ... i[group-1]), (i[1], i[2], ... i[group]),

I think this would be perfect if the overlap between successively
returned windows was also a parameter, rather than being fixed to 1.


Alex




From skip at pobox.com  Tue Apr  8 09:18:30 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 8 Apr 2003 08:18:30 -0500
Subject: Need help with Python porting
In-Reply-To: <78ada8a6.0304080156.6a780fe2@posting.google.com>
References: <78ada8a6.0304080156.6a780fe2@posting.google.com>
Message-ID: <16018.52262.760048.248188@montanaro.dyndns.org>

    Guru> I am currently porting Python to Netware. I would really
    Guru> appreciate if someone could tell me where I could find a reference
    Guru> implementation for Netware. I downloaded the 2.2.2 source from
    Guru> python.org for Linux.  Could someone give me a pointer as to how
    Guru> to go about modifying this code so that it works for NetWare.

Try searching Google for "Python Netware".  The first item returned is a
reference to a Python Conference paper from 2000 by Brad Clements, "NetWare
Programming with Python".  He has a section about porting.  If you then back
up to the Google results page and click the "Groups" tab you'll see a few
more interesting references.

Skip




From aleax at aleax.it  Sat Apr 12 05:50:30 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 09:50:30 GMT
Subject: instance creatiion as an "event"
References: <mailman.1050095779.7490.python-list@python.org>
Message-ID: <GjRla.2983$LB6.81868@news1.tin.it>

<posted & mailed>

Arthur wrote:

> I am trying to re-engineer, as lightly as possible, a Python app I
> wrote, to be make it more functional when used at the interactive prompt.
> 
> The only thing I seem to be missing is some kind of "hook" to trigger a
> general update cycle of all existing class instances upon the creation
> of a new class instance at the interactive prompt

I'm somewhat confused, I think -- is your problem:

[A] how do I distinguish whether an operation is being performed "at
    the interactive prompt" or elsewhere?

and/or is it:

[B] how do I find out when an instance of a class has been created (quite
    apart from the issue of whether it's created at the interactive prompt
    or within some module),

and/or is it:

[C] how do I find out all the existing instances of a given class at any
    given point (during an interactive session), whether that is the time
    of the creation of a new instance, or not

Each of these tasks is quite feasible -- some trivially easy (B), some
devilishly hard and tricky (A), some not too bad (C).  But unless you
clarify which one (or more) of these tasks you want to perform, it's
quite hard to offer you appropriate suggestions -- I wouldn't want to
get into the intricacies of possible architectures to meet [A] (which
depend on why exactly you want to distinguish between interactive and
non-interactive stuff -- so, I'd need more information on what you're
trying to achieve, to supply maximally helpful information) unless [A]
is indeed what you need.


> The ideas I come up with to explore do not seem very straight forward -
> though some variation of one or another of them will probably work. But
> am generally weak on introspection issues and am wondering, before
> embarking done a crooked road:
> 
> Am I missing a something straight forward.

What about "the need to specify precisely what you're trying to
accomplish" (and, in the case of a "what" that is particularly subtle
and "black magic" oriented, maybe the "why" too, in order to allow
experts to suggest possible alternative architectures...).

[B] is easy: instance creation is accompanied by calls to the class's
__new__ staticmethod (which returns the new instance object) and to the
__init__ method (to initialize the newly created instance).  You can
piggyback your operations on either or both of these (or, if you want
to not touch your class's source code, you can use a custom metaclass
and operate in the __call__ method of the metaclass).

[C] is reasonably easy: you need the class (or its custom metaclass)
to keep a container of all its extant instances in order to be able
to iterate on them.  As I've already seen suggested, weak references
are what you normally want (if your class is a newstyle one that uses
__slots__, this may be a problem), and a weakref.WeakValueDictionary
(with id(...) of each instance as the key and the instance as the value)
is more often than not the handiest way to accomplish this.

[A] is devilishly hard because distinguishing between "things that
happen at the interactive prompt" and "things that happen elsewhere"
is not normally needed, and it's hard to see exactly what you want
to accomplish with it.  Suppose your module foo.py is:

class Foo:
    "whatever"

def makFoo(): return Foo()

and at the interactive prompt you have:

>>> a = Foo()
>>> b = makFoo()
>>> def xx(): return Foo()
...
>>> c = xx()

which of these calls to Foo counts as being "at the interactive
prompt", and which don't?  If all of them count, can you give
examples of instance creations that you want to consider are NOT
"happening at the interactive prompt"?  Or clarify the real purpose
of the distinction so that we can perhaps suggest alternatives?


Alex




From pobrien at orbtech.com  Sun Apr  6 10:43:15 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 09:43:15 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <u65pr7qx6.fsf@online.no>
Message-ID: <m2of3jadq4.fsf@localhost.localdomain>

Syver Enstad <syver-en+usenet at online.no> writes:

> pobrien at orbtech.com (Patrick K. O'Brien) writes:
> 
> > Sound like fun?
> 
> Sounds like Pymacs ;-)
> http://www.iro.umontreal.ca/~pinard/pymacs/

Not quite.  I've downloaded Pymacs and intend to work with it.  From
what I've read it looks great.  But to me it appears to be a way to
link Python and Elisp together, as in elisp code can access Python
modules and a Python module could be written that acted as an Emacs
function.  (Or something like that.)  But what I want to do is go one
step farther and replace Emacs altogether with a similar program
written in Python, using Scintilla, wxPython and PyCrust.  Pymacs does
look like it will be a handy tool to use during this effort.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From imbosol-1049844119 at aerojockey.com  Tue Apr  8 19:32:52 2003
From: imbosol-1049844119 at aerojockey.com (Carl Banks)
Date: Tue, 08 Apr 2003 23:32:52 GMT
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com> <+6xk+kKkXQeJ092yn@gnosis.cx> <mailman.1049829736.3719.python-list@python.org> <45yk+kKkXAlY092yn@gnosis.cx> <mailman.1049835314.7390.python-list@python.org>
Message-ID: <E_Ika.1677$d8.1053@nwrdny02.gnilink.net>

Lulu of the Lotus-Eaters wrote:
> "Mike C. Fletcher" <mcfletch at rogers.com> wrote previously:
> |>    '"%s"' % repr(aString)[1:-1]
> |Except when there's a " character in the string ;) .
> 
> Ahh... what I MEANT to write was, of course:
> 
>    '"%s"' % repr(aString)[1:-1].replace('"',r'\"')

You misspelt

     '"%s"' % repr(aString)[1:-1].replace('"',r'\"').replace(r"\'","'")


-- 
CARL BANKS









From mhammond at skippinet.com.au  Tue Apr 22 20:40:32 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Wed, 23 Apr 2003 00:40:32 GMT
Subject: Embedding Python in an MFC application
In-Reply-To: <a7gpa.456$OK2.4518100@news-text.cableinet.net>
References: <a7gpa.456$OK2.4518100@news-text.cableinet.net>
Message-ID: <4ilpa.7413$8K2.64749@news-server.bigpond.net.au>

Darin Willits wrote:
> Hi,
> 
> I know that the first response to this will be "Why are you embedding
> when you could just extend?" but please believe that I have my reasons
> and for now this is what I have to do.  However I am something of a
> newbie with this stuff so bear with me.
> 
> Now for the question:  I have created a simple MFC application (with
> AppWizzard) which mearly calls Py_Initialize() and Py_Finalize() in the
> OnInit function.  When I build and run this app it starts the python
> interpreter fine, and shuts it down without a noticable problem. 
> However when I close the application I get a huge memory leak.  Now for
> the wieird part.  If I switch the above application from linking to MFC
> as a shared dll, to statically linking to MFC the memory dump is gone. 
> Why is this?  This is driving me nuts as I really need to use the MFC
> as a shared dll.
> 
> Please tell me I am missing something simple.  I am learning this stuff
> as I go and this one has me stumped.

You aren't seeing leaks, you are seeing memory that is allocated once by 
Python, and not freed.  MSVC doesn't distinguish between the two, 
whereas all commercial checkers do.  Just ignore them.  Note there is an 
issue that may cause a few problems when doing multiple Py_Init/Fini 
pairs - so just don't do that ;)

Mark.




From mike at bindkey.com  Tue Apr 22 21:25:19 2003
From: mike at bindkey.com (Mike Rovner)
Date: Tue, 22 Apr 2003 18:25:19 -0700
Subject: pattern match
References: <220420031949394828%greg.bland@valpo.edu>
Message-ID: <b84q3r$dsb$1@main.gmane.org>

"Mitch MF Cummstain" <greg.bland at valpo.edu> wrote in message
news:220420031949394828%greg.bland at valpo.edu...
>                I just started using python and I need help with a
> pattern match.
>
>    m = re.match("(.{30})(word)(.{30})",paragraph);
>
> I need to match 30 characters on each side of word.  The problem word
> is getting treated as a string not a variable like I need it to.  How
> can I make it treat word as a variable ?

m = re.match("(.{30})%s(.{30})" % re.escape(word),paragraph)







From hawkeye.parker at autodesk.com  Wed Apr  2 19:16:55 2003
From: hawkeye.parker at autodesk.com (haughki)
Date: 2 Apr 2003 16:16:55 -0800
Subject: Some newbie questions the FAQs didn't answer!
References: <6ffd97db.0304021132.62f86c7b@posting.google.com>
Message-ID: <b6e56f7a.0304021616.5b8f4749@posting.google.com>

stigwood4 at yahoo.com (Jason C.) wrote in message news:<6ffd97db.0304021132.62f86c7b at posting.google.com>...
> I was thinking about taking Python up this summer.  I was initially
> drawn towards it because I had heard how easy it was to program GUIs
> using the various packages.  But then I read that in order to run a
> python program, the machine had to have the python VM installed, and
> the user had to go to a command line and type "python <program
> name>.py".
> 
> What I had planned to do was write Win32 GUIs, but I would guess that
> most common users don't have Python installed.  Is there some way to
> write the program in C++ and then have it bring up the GUI written in
> Python without Python being installed?  Wouldn't that require the code
> being compiled into machine code?

there are various python extensions that allow you to 'package' your
python app so that a potential user doesn't have to have python
installed to use your app.  check out:

http://www.mcmillan-inc.com/install1.html



From alanpong at hkstar.com  Fri Apr 18 08:40:47 2003
From: alanpong at hkstar.com (Alan Pong)
Date: 18 Apr 2003 05:40:47 -0700
Subject: Problem in Accessing C++ Class Instance Global Variable (near success, help please!!)
References: <mailman.1050656618.32150.python-list@python.org>
Message-ID: <ea0e2298.0304180440.1fe777f@posting.google.com>

1. i did used Py_InitModule.
2. correction: wrapper.cpp compiled to wrapper.obj which is linked to
form myext1.dll.   test.exe doesn't contain wrapper.obj
3.  sorry i don't understand fix2 and fix3 mentioned below.

4. i re-think about: can i access a global variable like this???? ...

any sample code please?


"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1050656618.32150.python-list at python.org>...
> > From: Alan Pong [mailto:alanpong at hkstar.com] 
> > 
> > python 2.3, vc++6, win32 exe project
> > 
> > I has one global c++ class instance variable.  I want to call its
> > member
> > function through wrapper function which is accessed by python.
> > However, as seen in wrapper.cpp, it give me another new variable copy
> > so that phython actually accessing another instance.
> 
> First, PyImport AddModule( "myext1" ), doesn't load the module, but
> returns either an empty module or a module "myext1" that had allready
> been loaded. It returns a PyObject that you're supposed to do something
> with :-)
> 
> The symptoms imply that your "library" (wrapper.cpp) is most likely
> compiled statically against both your program (test.exe) and your dll
> (myext1.cpp). I.e. the dll is calling a method on an object that is
> "named" the same but exists in the dll memory space, while your
> application is reading an object (same "name") but in the executable's
> namespace. Try extending the dll to read back the value from its object
> to verify.
> 
> fix1: "Don't do that" <wink>.
> fix2: pass the object as an argument.
> fix3: make the object available through a module object both parties can
> "import" (meaning you'll need the PyRun * functions that take
> namespaces.
> 
> -- bjorn



From dickey at saltmine.radix.net  Tue Apr  1 06:35:30 2003
From: dickey at saltmine.radix.net (Thomas Dickey)
Date: 1 Apr 2003 11:35:30 GMT
Subject: Terminal-based text editor with color syntax hilighting
References: <slrnb8hvka.pki.ramen@lackingtalent.com> <mailman.1049169002.28920.python-list@python.org>
Message-ID: <b6bti2$c7f$1@news1.radix.net>

Sean 'Shaleh' Perry <shalehperry at attbi.com> wrote:
> On Monday 31 March 2003 19:39, Dave Benjamin wrote:
>> Hey folks,
>>
>> I often write Python over a terminal session, and I really enjoy working
>> with color syntax hilighting, so my editor of choise for the past year or
>> so has been Jed. I've heard that it's possible to get color hilighting to
>> work with Emacs, but so far I've been unable to get that working. (Note:
>> I'm talking about an SSH session, not an XTerm. Text only.)

ssh is an application that might run in an xterm

>> I think Jed is a nice editor, and it's very fast, but I'm wondering if
>> there are any other editors that support this feature. Does anyone know?
>>
>> Thanks,
>> Dave

> new gnu emacs (around 8 months old I think) and vim (a super vi clone) both 
> support color coding.

actually XEmacs has done color for a few years.  For vi-clones, elvis and
vile both do color as well.

-- 
Thomas E. Dickey <dickey at radix.net> <dickey at herndon4.his.com>
http://dickey.his.com
ftp://dickey.his.com



From sismex01 at hebmex.com  Wed Apr  2 10:13:17 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Wed, 2 Apr 2003 09:13:17 -0600 
Subject: PEP 313 - Roman numeral literals
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F062330437D35D@mail.hebmex.com>

> From: sik0fewl [mailto:xxdigitalhellxx at hotmail.com]
> Sent: Wednesday, April 02, 2003 8:42 AM
> 
> Although I honestly don't know how, I'm sure it's possible. 
> Here's how I would do it, although it's probably not the right
> way. My roman numerals are a bit rusty. :)
> 
> XIL
> 
> ?
>

No, the correct writing of 39 in roman numerals is "XXXIX",
the proposed rule for only 3 of X, I or C in a single number
is premature, or incomplete.

It should read "3 consecutive X, I or C digits".

Salutations.

-gustavo




From rmunn at pobox.com  Tue Apr  1 01:27:55 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 01 Apr 2003 06:27:55 GMT
Subject: try:else: w/o except: - why not?
References: <1c004553.0303311553.53a9a92b@posting.google.com> <9ee55987.0303312049.59e7ddfd@posting.google.com>
Message-ID: <slrnb8hnbn.8q6.rmunn@localhost.localdomain>

yvan <yvan_charpentier at hotmail.com> wrote:
> manus at python.net (Manus Hand) wrote in message news:<1c004553.0303311553.53a9a92b at posting.google.com>...
>> I know that if you want to use else: on a try: block, you need to
>> specify one or more except: clauses.  I guess my question is why
>> this should need to be.
>> 
>> Personally, I have cases where it would be nice to do this:
>> 
>> try:
>>     # some code that may except
>> else:
>>     # but if it didn't except, I want to do this
>> 
>> Instead of writing the code that way, I need to write this:
>> 
>> try:
>>     # some code that may except
>> except:
>>     # and if it does, then, okay, just ignore it
>>     pass
>> else:
>>     # but if it didn't except, I want to do this
>> 
>> I know it's a core language decision and therefore up to Guido,
>> but I am curious (Guido?) if there is something I am missing that
>> would make a simple try:else: construction (without requiring any
>> except: blocks) not make sense?
>> 
>> Manus
> 
> Unless your code throws, execution will resume, so the following
> should behaves as you want.
> 
> try:
>      # some code that may throws
>      # Some code that will be executed if it did not throw above, will
> not if it threw (this is what is in your else block)
> except:
>      # will get there if it threw, will not otherwise

I can see a subtle difference between that block and the OP's "try: ...
else: ..." proposal. For simplicity, I'm going to use single print
statements to represent the code blocks being talked about.

OP's proposal (fragment #1):

    try:
        print "Code that may raise exception"
    else:
        print "No exception was raised"

Your almost-equivalent solution (fragment #2):

    try:
        print "Code that may raise exception"
        print "No exception was raised"
    except:
        print "Oops, caught an exception"

This will have exactly the same behavior as fragment #1 if and only if
the "else:" block can raise no exceptions. If the "else:" block of
fragment #1 is capable of raising exceptions (let's say it writes to a
log file and might raise IOError, for example), then:

Fragment #1 would raise the exception normally.

Fragment #2 would find the exception caught by the "except:" block. This
is fine, but that "except:" block was written with the intent of
catching exceptions from the first print statement, not the second!
Catching IOError, for instance, would be misinterpreted as meaning that
something went wrong with a file in the first block, not the second.

I hope I'm making myself clear. It's late and I'm not having the easiest
time putting my thoughts into words right now.

I find myself opposed to the OP's suggestion, by the way. Explicit is
better than implicit, and adding two lines really does not make that
much difference. I find this:

    try:
        print "Doing stuff"
    except:
        raise
    else:
        print "No exceptions, hooray"

immediately obvious in intent, whereas this:

    try:
        print "Doing stuff"
    else:
        print "No exceptions, hooray"

leaves me with a feeling of "Wait a minute, where's the except: clause?"

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From skip at pobox.com  Fri Apr 25 11:30:30 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 25 Apr 2003 10:30:30 -0500
Subject: Linux SUID
In-Reply-To: <3EA951F2.10207@ghaering.de>
References: <3ea94eae$0$23154$ba620e4c@reader0.news.skynet.be>
        <3EA951F2.10207@ghaering.de>
Message-ID: <16041.21654.333496.299591@montanaro.dyndns.org>

    >> There one can turn on the SUID-exec permission bit which
    >> is honoured by Perl.
    ...
    >> Is there something similar in Python?

    Gerhard> Short answer: No.

    Gerhard> There's a demo suid wrapper in the source distribution ...

I think the Mailman distribution has/does something similar, using a
compiled C program to do the suid bit.  If my recollection is correct it's
likely to be better tested than the demo wrapper in the Python distribution.

Skip




From exarkun at intarweb.us  Sat Apr 19 10:57:56 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sat, 19 Apr 2003 10:57:56 -0400
Subject: Parent references: hot or not?
In-Reply-To: <1050763980.14721.41.camel@ida.dylanreinhardt.com>
References: <1050763980.14721.41.camel@ida.dylanreinhardt.com>
Message-ID: <20030419145756.GA25916@meson.dyndns.org>

On Sat, Apr 19, 2003 at 07:53:01AM -0700, Dylan Reinhardt wrote:
> Hi all,
> 
> I'm working on a project where I have multiple instances of a container
> class that each hold a large number of another class.
> 
> I need the "contained" object to be able to make use of methods and
> attributes in its container. Inheritance won't do the trick since the
> things I need to refer to will be pretty dynamic.
> 
> A very minimal example of what I'm doing:
> 
> ---------------
> 
> class Container:
>    def __init__(self):
>       self.items = {}
> 
>    def add_item(self, item_name):
>       self.items[item_name] = Item(self)
> 
>    some_attr = 'spam'
> 
> 
> class Item:
>    def __init__(self, parent):
>       self.parent = parent
> 
>    def some_method(self):
>       return self.parent.some_attr
> 
> ---------------
> 
> This seems to be working well, but I want to check my assumptions.
> 
> If I'm understanding correctly:
> 
> 1. Any Item's self.parent is only a *reference* to its parent.

  Currect.  All variables in Python are but references.

> 2. Holding a reference to a parent isn't some kind of recursive paradox.

  Correct - so long as you don't implement __del__ on either the container
or the item.  If you do, Python will be unable to free the cycle and it will
end up in gc.garbage.  Not that this will stop your program from running
but: 1) it will be a memory leak; 2) the code in your __del__ method won't
actually be run.

> 3. As I add Items, Container size will grow linearly, not exponentially

  Correct.

> 4. It will be impossible to fully delete a non-empty Container.

  I'm not sure what "fully delete" means here.  The container will certainly
be able to do "del self.items".  If "self.items" was the only reference to
the items in the container, then the will all be collected by Python and
de-allocated.

> 
> I think those are valid assumptions, but would appreciate any
> confirmation, contradiction, or insight anyone cares to offer.
> 
> Also, if there's some other consideration or reason why holding a
> reference to one's parent is a Bad Idea, I'd love to hear about it.

  Not really.  Someone else recently asked this, and I think one of the good
answers was that it's only a bad idea if it is possible to accomplish the
same thing without the reference without undue additional algorithmic
complexity.

  Jp

-- 
It is practically impossible to teach good programming style to
students that have had prior exposure to BASIC: as potential
programmers they are mentally mutilated beyond hope of
regeneration.        -- Dijkstra
-- 
 up 30 days, 11:02, 7 users, load average: 0.09, 0.19, 0.12




From aleax at aleax.it  Sat Apr 19 04:05:49 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 08:05:49 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <Gu_na.34381$LB6.779929@news1.tin.it> <OT0oa.690$l75.32@tornadotest1.news.pas.earthlink.net>
Message-ID: <xr7oa.35303$LB6.796736@news1.tin.it>

Tyler Eaves wrote:
   ...
> Wow, I feel vindicated now for never trying to wrap my brain around stuff
> like list comprehensions. *g*

And here, you're wrong - list comprehensions are VERY worth "wrapping
your brain around"!  They allow clear and direct expression and their
performance is generally quite OK.

E.g. "make a list L of N*N+1 for N in range(1000)":

[alex at lancelot python2.3]$ python2.2 -O timeit.py 'L=[]' 'for i in 
range(1000):' '  L.append(i*i+1)'
1000 loops, best of 3: 1.45e+03 usec per loop

[alex at lancelot python2.3]$ python2.2 -O timeit.py 'L = map(lambda i: i*i+1, 
range(1000))'
1000 loops, best of 3: 1.23e+03 usec per loop

[alex at lancelot python2.3]$ python2.2 -O timeit.py 'L = [i*i+1 for i in 
range(1000)]'
1000 loops, best of 3: 1.1e+03 usec per loop

the LC form expresses the specs directly and even has a little performance
advantage in this case over the map/lambda form and the direct loop.  Study
up on list comprehensions, you'll be glad you did!


Alex




From daniel at dittmar.net  Thu Apr 24 18:34:50 2003
From: daniel at dittmar.net (Daniel Dittmar)
Date: Fri, 25 Apr 2003 00:34:50 +0200
Subject: php equivalents?
In-Reply-To: <e9Ypa.3964$b67.980821@read2.cgocable.net>
References: <OHSpa.8996$h%2.915957@read1.cgocable.net> <b890vo$rpj$1@news1.wdf.sap-ag.de> <e9Ypa.3964$b67.980821@read2.cgocable.net>
Message-ID: <b89oqj$d20$1@news.eusc.inter.net>

Kevin Howe wrote:
> Thanks for the suggestions Daniel. I've since discovered an equivalent to
> PHP's register_shutdown_function() as follows:
> 
> def myFunc ():
>   """ perform cleanup just before the script exits """
> 
> import atexit
> atexit.register(myFunc)

If you're doing straight CGI, then atexit will work. But I guess if 
you're using the Apache module or anything with fastcgi, then the atexit 
functions won't be called at the end of the request but only when the 
Python interpreter is shut down. Which may or may not be exactly what 
you want.

Daniel





From alanmk at hotmail.com  Sun Apr 27 09:50:51 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Sun, 27 Apr 2003 14:50:51 +0100
Subject: please comment on technologies
References: <slrnbalplq.254.luc@trillian.dont-panic.info> <HqCqa.8931$3M4.241359@news1.tin.it> <slrnbalv50.254.luc@trillian.dont-panic.info>
Message-ID: <3EABE03B.41C50F07@hotmail.com>

luc wastiaux wrote:

> Ok, I thought XSL was the definitive answer to all templating needs, but
> you don't seem very excited about it ?

Luc,

I like the sound of your project. I'm sure it's going to be fun and a
good learning experience.

As for XSLT, I fell in love with it when it first appeared in 1998. It
seemed like the perfect solution for my web-site creation needs (which
were relatively simple back then).

I liked it so much because it allowed me, as I thought, to separate
content from format in a clean fashion. And although the syntax can be a
little tiresome sometimes, the functional programming aspect of XSLT can
be picked up. The most important thing to understand about writing XSLT
code is the Xpath data model.

However, over the years I've drifted away from XSLT, for the following
reeasons.

1. Memory hog. Although the XSLT transformation sheets themselves may
not consume too much memory, the documents that they are transforming to
and from can consume *enormous* amounts of memory. The output side can
be dealt with by not building an output tree, but instead outputting to
SAX2 events and then into a character buffer. But to do that with many
existing XLST processors, you're looking at some coding work. And there
is NO WAY to avoid constructing the huge input tree: it MUST be
constructed so that you can run Xpath queries on it.

2. CPU hog. You really have to be very careful what you're doing with
XSLT. When you use an Xpath expression to select nodes in a tree, the
order of (in)efficiency of those Xpath expressions can be much higher
than you think. This is generally not a problem for the people who are
thoroughly familiar with Xpath. But it can be a huge problem for
newbies. Especially when you're using XSLT in a web application and
you're trying to figure out why your HTML pages take 20 seconds to
render. (I'm facing this problem right now. I'm load-testing a web app
that uses JSP and XSLT to generate a series of 17-20 HTML pages to
interact with the user. It all works really well, until you get to the
particular stage when it actually uses XSLT to render a group of HTML
pages. For a single user, it takes 20 seconds, and for 20 simultaneous
users, it takes nearly 5 minutes each user! Completely unacceptable. I
reckon the reason for this is because the development team were told to
use XSLT, without having the relevant training or experience to use it
*properly*).

3. Poor design ingegration. Now that I work on web sites that have to be
*designed*, i.e. using GUI design tools, it is essential that whatever
template language I use be able to integrate with the likes of
DreamWeaver or FrontPage. With XSLT, this can be quite difficult to do.
One possible solution is to auto-generate XSLT templates from the
GUI-created HTML page, using a combination of "tidy"ing the page into
XHTML and using a private (XML) namespace of attributes to tell the
template generator which bits of the GUI-edited HTML it should be
putting into which XSLT templates. And there's still often plenty of
hand-editing to be done after this process is complete.

If you want to use the best and most pythonic markup templating language
in existence (IMHO ;-) look no further than Template Attribute
Lanaguage, aka TAL. This was created by the BDFL (can't get more
pythonic than that), and solves all of the problems mentioned above. TAL
used to be a ZOPE-only thing, but now there are multiple standalone
implementations available. A list of them can be found in this Usenet
posting. 

http://tinyurl.com/af8i

As far as I'm concerned, TAL is *the* best templating language, for the
following reasons

1. Efficient: in terms of memory and cpu.
2. Simple: easiest thing in the world for people to understand.
3. Design-integration: As described above.

Just my 0,02 euro.

Best of luck with your project.

regards,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From duncan at NOSPAMrcp.co.uk  Wed Apr 16 07:17:26 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 16 Apr 2003 11:17:26 +0000 (UTC)
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <d7aq9v0a5ecoue90gccfos9jtkkqcr73di@4ax.com>
Message-ID: <Xns935F7BC1FC59duncanrcpcouk@127.0.0.1>

Dale Strickland-Clark <dale at riverhall.NOTHANKS.co.uk> wrote in 
news:d7aq9v0a5ecoue90gccfos9jtkkqcr73di at 4ax.com:

> This is a class you use in place of a string and it emulates a string
> in all practical ways except that comparisons and lookups are case
> insensitive. All uses of the string for assignments, however, yield
> the original case.

I'm slightly surprised that your class doesn't include:

    def lower(self):
        return self.____lowerCaseMe

A quick timing test indicates that this does make a difference.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From Anish.Mehta at enst-bretagne.fr  Mon Apr 21 16:34:35 2003
From: Anish.Mehta at enst-bretagne.fr (Mehta, Anish)
Date: Mon, 21 Apr 2003 22:34:35 +0200
Subject: Regarding Dictionaries in python
Message-ID: <3EA455DB.9000300@antares.enst-bretagne.fr>

Hello !

I am having a small query regarding dictionaries in python. From 
Orelly's Programming Python

 >>>D= {'a' : 1, 'b' : 2, 'c' : 3, 'd' : 4}
 >>>D.items()
[ ('b', 2), ('c', 3), ('d', 4), ('a', 1)]


why it is not showing the items in order in which i have inserted. 
Please tell me why it is happing like this.

Thanks in advance.

Regards...







From tismer at tismer.com  Sun Apr 20 11:57:02 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sun, 20 Apr 2003 17:57:02 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <ruadnZIgGvclOj-jXTWcpA@speakeasy.net>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net>
Message-ID: <3EA2C34E.8070003@tismer.com>

A.M. Kuchling wrote:
> On Sun, 20 Apr 2003 09:23:12 +0200, 
> 	Anton Vredegoor <anton at vredegoor.doge.nl> wrote:
> 
>>days, unfortunately. I'd be interested in an explanation of what
>>continuations are and what their advantages would be. 
> 
> 
> See an old posting by Tim Peters, archived at  
> http://mail.python.org/pipermail/python-dev/1999-July/000467.html

Thanks for reminding of this nice paper. I will add
it to the documentation.

ciao - chris

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From duncan at NOSPAMrcp.co.uk  Mon Apr 28 09:06:10 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Mon, 28 Apr 2003 13:06:10 +0000 (UTC)
Subject: pyrex exceptions
References: <b8j67p$aqpk$1@as201.hinet.hr>
Message-ID: <Xns936B8F53CB9F8duncanrcpcouk@127.0.0.1>

"Haris Bogdanovic" <haris at email.hinet.hr> wrote in
news:b8j67p$aqpk$1 at as201.hinet.hr: 

> How do I handle exceptions in Pyrex ?
> I have a C++ wrapper around some C library and I'm doing a Python
> wrapper around that library (copying C++ wrapper object model).
> In C++ wrapper there is a class Exception which is derived from C++
> standard exception class.

Pyrex doesn't really know about C++. In particular it doesn't have any 
facilities for handling C++ exceptions. If you are linking to C++ from 
Pyrex you'll need to wrap the C++ code in C++ functions that catch and 
handle all exceptions. You must never allow a C++ exception to propogate 
back into the Python interpreter, otherwise bad things may happen.

In this case it sounds as though you might be better to wrap Pyrex around 
the C library directly. Still follow the C++ object model if you want, but 
don't use any of the C++ code.

Alternatively you could look at Boost/Python which does know about C++ 
exceptions.

> In Pyrex it is not yet possible to derive an extension type from
> Python class so I would have to create extension type for handling
> exceptions from scratch.
> If I want to raise an exception like this :
> raise ExceptionClass(error)
> how should then ExceptionClass be defined ?

exactly as you would define it in Python:

   class ExceptionClass(Exception):
       pass

and then you can raise it from Pyrex just as you would in Python.

   raise ExceptionClass(error)

> I'm using wxPython GUI so I want to show the message box (window)
> which displays error text when raising exception and then let user to
> choose from options of ending or continuing the program from some
> previous place in code or correcting error which caused the exception.
> 
> Can you please give me just the bare bones of how should that
> exception class look like.(in Pyrex, as extension type)

The exception class must have Exception as a base class. Exception is not 
an extension type so your exception class shouldn't be one either.

> I'm also little bit unfamiliar with what exactly goes on when program
> raises an error with class instance as attribute. Could you explain
> that just in a few words ?
> 
I'm not sure exactly which bit you want explained. At the C level, or the 
Python level?

When a Python exception is raised in C code, and therefore in Pyrex code, 
the exception that is to be raised is saved and the C/Pyrex function 
returns a value indicating that an error occurred. The caller checks the 
return code, and if it indicates an error either handles the exception or 
returns an error back to its caller. Within Pyrex the checking of return 
codes is usually automated.

If one of the functions wants to handle an exception it checks whether the 
exception that was thrown matches something it wants to handle. If so it 
can clear the exception state and continue processing. If not it continues 
to propogate the error by returning an error code.


-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From edo4 at emich.edu  Fri Apr 18 17:45:33 2003
From: edo4 at emich.edu (Ngoc)
Date: 18 Apr 2003 14:45:33 -0700
Subject: Numeric and Visual .Net
Message-ID: <99457f2.0304181345.2d5870da@posting.google.com>

I run the command line to build debug mode for Numeric module.
When I run it from python command line or embedded Numeric module in
my C++ code, it prompt up the error message as "required MSVCRTD.dll."
MSVCRTD.dll is old MS visual dll.  
I'm wondering if any way that I can build Numeric module from Visual
.Net
Any suggestion??
Thanks in advance.
Ngoc



From henrik.motakef at web.de  Fri Apr 18 17:45:26 2003
From: henrik.motakef at web.de (Henrik Motakef)
Date: Fri, 18 Apr 2003 23:45:26 +0200
Subject: Why functional Python matters - A kind of conclusion
References: <mailman.1050527910.25653.python-list@python.org> <3EA08A7D.3010206@removethis.free.fr>
Message-ID: <877k9rh43t.fsf@interim.henrik-motakef.de>

laotseu <bdesth at removethis.free.fr> writes:

> Due to my deep ignorance, I came to believe that Lisp was the first
> programming language based on the lambda calculus theory, the first
> language to promote functions as first class objects, and the first
> language to promote side-effect free programming style...

On the other hand, Common Lisp claims to be the first language with a
standardized (as in "ANSI standard") object system :)

> BTW, is Paul Graham still a Lisper?

Yes he is. He is currently designing a new dialect of Lisp called
Arc, see http://www.paulgraham.com.

> for x in range(100):
>    print "I shall not say that Lisp is a functional language anymore"

(prog ()
  10 (print "I will not make my programming language do my homework.")
  20 (go 10))

;-)

Regards
Henrik



From aahz at pythoncraft.com  Tue Apr 29 00:18:37 2003
From: aahz at pythoncraft.com (Aahz)
Date: 29 Apr 2003 00:18:37 -0400
Subject: POD vs. reST for standalone writing?
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <mailman.1051584944.30800.python-list@python.org>
Message-ID: <b8kuet$ftn$1@panix1.panix.com>

In article <mailman.1051584944.30800.python-list at python.org>,
Andrew Bennetts  <andrew-pythonlist at puzzling.org> wrote:
>
>  - ReST isn't extensible enough; how do you add markup for, e.g. API
>    references?  How do you plug that into your docutils installation?  You
>    can't.  (The Lore markup for APIs is 
>        <code class="API">twisted.internet.reactor.run</code>
>    )

You're wrong::

    :API:`twisted.internet.reactor.run`

Granted that isn't currently trivial to add, but it's definitely
possible.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From peter at engcorp.com  Fri Apr 25 19:02:03 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 19:02:03 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <3EA9BE6B.772D990F@engcorp.com>

Frantisek Fuka wrote:
> 
> When I try to call methods, I sometimes forget to include the
> parentheses. Instead of:
> 
> if object.isGreen():
>         do something...
> 
> i sometimes write:
> 
> if object.isGreen:
>         do something...

All I can offer is that if you are doing test-driven development,
or writing unit tests for your code, you will very quickly 
catch this kind of problem and it will largely cease to be a
big issue for you.  Not only that, but you'll probably find 
yourself cured of the root problem fairly soon.

Python is an *excellent* language for agile methods, including
test-driven development.  Check the archives for various threads
on the topic of unit-testing with Python.

-Peter



From claird at lairds.com  Fri Apr 25 12:38:53 2003
From: claird at lairds.com (Cameron Laird)
Date: Fri, 25 Apr 2003 16:38:53 -0000
Subject: Reading a non-standard floating point format
References: <mailman.1051286382.24604.python-list@python.org>
Message-ID: <vaip4ts74ajqd0@corp.supernews.com>

In article <mailman.1051286382.24604.python-list at python.org>,
Tim Peters  <tim.one at comcast.net> wrote:
>[Fidtz]
			.
			.
			.
>> While I think I can see the way to do this with various bit-grubbing
>> techniques, the potential for error is massive, esp considering my non
>> comp-sci background.
>
>Post a link to a definition of this format, and I bet someone will help.
>
>> Does anyone know of a library in python or even (fairly) easily
>> wrappable C that might let me specify the format in a more high level
>> way?
>
>Precise English will translate into Python easily enough <wink>.
>
>

It's true that getting floating-point formats correct
to the far-right bit is, in general, tedious; Fidtz
shows good instincts in his caution about that prospect.

The bad news is that, no, there are no widely-applicable
parametrizations of the range of interesting floating-
point formats.  This remains an area of "craftwork".
Learn to use a hand-chisel, or find someone who will do
so for you.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From aleax at aleax.it  Thu Apr 17 10:53:48 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 14:53:48 GMT
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com> <6e07b825.0304162200.34884d36@posting.google.com> <pan.2003.04.17.13.36.08.2131.744@nospam.net>
Message-ID: <0ezna.42488$T34.1181608@news2.tin.it>

Drazen Gemic wrote:

>> People have said it in the past, the Python community needs to pick ONE
>> of the Python wxWindow GUI builders currently under development and
>> really promote the hell out of it.
> 
> I don't think it will ever happen. You people with microsoft background
> do not understand some things. In Python, PHP, Perl, Linux comunities
> there are no 'winners'. People just use software they like the most.

It may or may not happen, but SOME of the time some free program
emerges as the clearly dominating one in its niche -- others often
survive (not always...) but with very minor levels of effort and
investment going into them any more.  E.g., apache's dominating the
niche of "open-source general purpose webserver" -- there may be
others, but does the amount of time and effort going into their use,
development, documentation, etc, reach to even 1/10th as that which
apache attracts?

So, there aren't _always_ "winners" (but then, the same "not always"
holds for the proprietary world -- e.g., Delphi and VB keep surviving
and flourishing side by side, one against the other), but often there
are, even in open-source worlds.


> Some use this GUI builder, other use that, other do not use GUI builder
> at all, without bothering each other.

Actually, when two or more products (opensource or not) are used in
vastly overlapping niches, flamewars between devotees of each are
quite typical - not inevitable, but quite typical.  Part of the root
cause is in human sociobiology, part is in the 'risk', from the POV
of somebody who's invested hugely in one of them, that another will
"emerge as winner" and suck interest and effort away from that one.

> People do not want to be "guided" by some kind of superior authority
> that decides what is the best for everybody.

Some do, some don't.  Standards emerge, often by voluntary action
of people who are very much in the open-source community, because
in many cases there ARE substantial advantages to most everybody in
having exactly such a "superior authority" as a standards-making
body -- those cases are roughly all of those (and only those) where
the shared advantages by "networking effects" (in the economics
sense) surpass the disadvantages of relative resulting unitormity.

So (for example) you're perfectly free to build your own Linux
distribution from scratch and completely disregard any indications
from (e.g.) the FHS about what filesystem hierarchy "is the best
for everybody" -- in practice, very few people do, because the
advantage of compatible filesystem hierarchy layouts are quite
considerable to everybody, the compensating advantages of doing
your own idiosyncratic and quirky hierarchy quite small.

So, I think your assertions are way over-broad the way you put
them.  What situation applies to the specific case of GUI builders,
well, time will tell -- personally, I have no idea how that issue
will turn out... continuing fragmentation, or what degree of
consolidation.  But as you don't limit your assertions to that
specific issue, they can be seen as false in their generality.


Alex





From duncan at NOSPAMrcp.co.uk  Wed Apr 23 04:24:52 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 23 Apr 2003 08:24:52 +0000 (UTC)
Subject: Off topic: What usenet news reader to use?
References: <mailman.1051036295.31460.python-list@python.org>
Message-ID: <Xns93665EDC8E0E0duncanrcpcouk@127.0.0.1>

"Wai-Yip Tung \(wtung\)" <wtung at cisco.com> wrote in 
news:mailman.1051036295.31460.python-list at python.org:

> I haven't used Usenet news for a long time. Now that I have subscribed
> to python-list I find that email does not scale well to the high volumne
> of discussion. Instead of deleting emails all day I think I should go
> back to Usenet news. Is there any good Windows news reader you would
> recommend. I'm happily using Opera as the browser. But it's
> revolutionary news and email reader eludes me. What are you using?

I use XNews in conjunction with Hamster.

XNews may be found at http://xnews.newsguy.com/
Hamster may be found at http://www.tglsoft.de/

XNews is arguably one of the best online newsreaders available for Windows, 
but it doesn't work so well in offline mode (i.e. not connected to a news 
server), hence the reason why I use Hamster. Hamster is a local news/mail 
server, so I use it to provide XNews with what looks like a single newsfeed 
although it contains newsgroups from several different news servers, and 
also mailing lists converted into local newsgroups.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From syver-en+usenet at online.no  Sun Apr  6 19:22:31 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 07 Apr 2003 01:22:31 +0200
Subject: Debugging embedded python
References: <u4r5bgyfx.fsf@online.no> <Ph2ka.2428$Tb6.22760@news-server.bigpond.net.au>
Message-ID: <uvfxrfbyg.fsf@online.no>

Mark Hammond <mhammond at skippinet.com.au> writes:

> Syver Enstad wrote:
> > I have the python interpreter embedded in a daemon/service (this is
> > win2k so it's a service). I'd really like to be able to debug the
> > python code that is running. I was thinking of popping up a console
> > window to be able to run pdb, how does one go about it to accomplish
> 
> > this? Are there any code doing similar things or articles.
> 


> You should be able to run the service in "debug" mode (ie, as a normal
> 
> program" by using the "--debug" option.  Then you can debug however
> you like.

I am fooling around with embedding the python interpreter in an ISAPI
extension under IIS, I assume that you meant running a python service
under PythonService.exe?

> But assuming you do need to debug as a service, you really need an
> out-of-process debugger, like Komodo and most of the commercial
> IDEs. Pythonwin does *not* do out-of-process debugging, so is no
> good.

I haven't needed to debug as a service yet, using regression tests of
the output and printf style debugging has worked well. But I think it
is very useful (and reassuring) that one can inspect the program
thoroughly if one needs to do so (A tricky bug f.ex).



-- 

Vennlig hilsen 

Syver Enstad



From mis6 at pitt.edu  Mon Apr 21 10:44:00 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 21 Apr 2003 07:44:00 -0700
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <b7q72i$nvh$1@slb6.atl.mindspring.net>
Message-ID: <2259b0e2.0304210644.1c1c0584@posting.google.com>

"Andrew Dalke" <adalke at mindspring.com> wrote in message news:<b7q72i$nvh$1 at slb6.atl.mindspring.net>...
> [Aahz, on Dijkstra's BASIC quote]
> 
> Peter Hansen:
> > Perhaps only once
> > she advances beyond the student stage, and has some years of actual
> > _practice_, will she discard the early mental mutilation of BASIC
> > and find her way to Good Style.
> 
> I too learned programming starting with BASIC.  I started with
> the manual which came with my TI 99-4/A some 20+ years ago,
> and didn't give BASIC up until 12 years ago.  I think my last big
> BASIC program was a X-ray crystallography program, with a
> GUI to lay out and assign the lines, etc.
> 
> One thing to note, which Dijkstra didn't predict, is that I stopped
> using a 'traditional' BASIC back in '86, when I got a copy of
> QuickBasic for my birthday.  No line numbers, real functions,
> better control structures, compiled code - while editing in the
> IDE, so the compile-link-run cycle is really fast.  Quite nice,
> and after that and Turbo Pascal and Turbo C, going to K&R
> C on a unix machine was like a step back into darkness.
> 
> Anyway, so it is possible to get Good Style even under
> BASIC.  (Shhh!  And I think my BASIC background is
> one reason I like Python so much. ;)
> 
>                     Andrew
>                     dalke at dalkescientific.com

+1

I started with a Casio pocket calculator and then switched to AmigaBasic,
but essentially my experience is the same of yours. I also felt programming
in C on Unix was a step back into darkness. Happily, I found Python, with
its nice 

print "Hello World" 

that made me feel at home ;)

                                             Michele



From mertz at gnosis.cx  Tue Apr  8 16:36:29 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 16:36:29 -0400
Subject: Best way to make a weighted sum
References: <867ka46awj.fsf@ieee.org>
Message-ID: <NLzk+kKkXcsb092yn@gnosis.cx>         

Jorge Godoy <godoy at metalab.unc.edu> wrote previously:
|    def CalculaSoma(self):
|        ie = self.ie
|        iter = len(ie)
|        for i in ie:
|            produto = i * self.pesos[iter]
|            self.soma = self.soma + produto
|            iter = iter - 1
|I don't know if I got it in the most pythonic way in my code.

Looks fine.  You could be somewhat more concise with:

    from operator import mul, add
    amts, weights = pesos.keys(), pesos.items()
    sum = reduce(add, map(mul, amts, weights))/reduce(add, weights)

I'll leave it to a native speaker to translate my var names to
Portuguese.

Yours, Lulu...

--
 mertz@   _/_/_/_/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY:_/_/_/_/ v i
gnosis  _/_/                    Postmodern Enterprises         _/_/  s r
.cx    _/_/  MAKERS OF CHAOS....                              _/_/   i u
      _/_/_/_/_/ LOOK FOR IT IN A NEIGHBORHOOD NEAR YOU_/_/_/_/_/    g s





From nomad*** at ***absamail.co.za  Sun Apr 13 17:08:33 2003
From: nomad*** at ***absamail.co.za (Nomad)
Date: Sun, 13 Apr 2003 23:08:33 +0200
Subject: PYTHON EDITOR!!
References: <mailman.1049690233.24009.python-list@python.org> <33803989.0304070300.661d1d9@posting.google.com>
Message-ID: <4n3j9vossbu1p0j6m2c42cmq5al6ing8aj@4ax.com>

On 7 Apr 2003 04:00:23 -0700, tebeka at cs.bgu.ac.il (Miki Tebeka) wrote:

>Hello Nanthan,
>
>> im Nanthan, a postgraduate student from Malaysia.im very new to 
>> python language but im already getting use of it.
>Have fun.
>
>> Currently,i 
>> have been working in windows, but my development actually is on 
>> Linux, so can u suggest me the best editor(i mean use friendly 
>> and graphical,i dont want command line editor)for python?
>http://www.python.org/cgi-bin/moinmoin/PythonEditors
>
>Personally I use Vim (http://vim.sf.net) but it far from being graphical.
>

There's always gVim ;-)

-- 
Nomad

Wondering of the vast emptyness of the 'net
in search of something cool.



From awheat at imaging.robarts.ca  Mon Apr  7 15:40:35 2003
From: awheat at imaging.robarts.ca (Andrew Wheatley)
Date: Mon, 7 Apr 2003 15:40:35 -0400 (EDT)
Subject: Tkinter - tracking toplevel windows
In-Reply-To: <20030407193732.GB27968@unpythonic.net>
Message-ID: <Pine.LNX.4.44.0304071540110.3969-100000@dandelion.irus.robarts.ca>

Jeff:

Thanks for your quick response.  I just figured it out 5 minutes
ago.

Andrew

On Mon, 7 
Apr 2003, Jeff Epler wrote:

> On Mon, Apr 07, 2003 at 11:23:13AM -0700, Andrew Wheatley wrote:
> > I have a Tkinter Gui with a button which generates other toplevel windows.
> > Whenever a toplevel is generated, I append its id to a global list.  What I'd like
> > to do with wm_protocol/WM_DELETE_WINDOW method is remove from the list
> > the window id currently being killed and then destroy the window.
> > 
> > My question is:
> > 
> > How can I tell which window is being destroy when I kill it with the X in the 
> > top right hand corner.
> 
> Arrange to include a reference to the window when in the callback.  For
> example:
> 
>     def close_clicked(t):
>         print "close was clicked on window", t
> 
>     t = Tkinter.Toplevel()
>     t.wm_protocol('WM_DELETE_WINDOW', lambda t=t: close_clicked())
> 
> When WM_DELETE_WINDOW is defined, using window-manager "close"
> functionality (such as ALT-F4, clicking the "X", etc) merely triggers
> the callback, which can override the close event.
> 
> There's also the <Destroy> binding, but this is a little more
> problematic.  You *can* use the .widget attribute of the event, but due
> to a detail of Tk internals*, this binding will be applied to any widget
> inside t.  So below, you'll get two messages:
> 
>     def destroyed(evt):
>         print "window", evt.widget, "was destroyed"
> 
>     t.bind("<Destroy>", destroyed)
>     b = Tkinter.button(t); b.pack()
>     t.destroy()
> 
> So if you're interested only in toplevel windows, do this:
>     def destroyed_smarter(evt):
>         if not evt.widget.wm_toplevel() is evt.widget:
>             return # not actually the window
>         print "window", evt.widget, "was destroyed"
> 
> The <Destroy> event fires after the destruction has begun, so there's no
> opportunity to veto the destruction.
> 
> Jeff
> * See the documentation on "bindtags" for an explanation of why this is
>   so
> 

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Andrew Wheatley
Imaging Research Laboratories         phone: (519) 663-5777 X34324
JP Robarts Research Institute         FAX  : (519) 663-3403
100 Perth Drive, P.O. Box 5015
London, Ontario, CANADA  N6A 5K8

http://www.imaging.robarts.ca/~awheat/index.html
email: awheat at imaging.robarts.ca
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

"Faith alone in Christ alone by grace alone." - Charles Stover





From staschuk at telusplanet.net  Sat Apr 19 13:16:10 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 19 Apr 2003 11:16:10 -0600
Subject: Are reference cycles bad form?
In-Reply-To: <3o11avop6vkh3ake8dgbtst93t1or771ct@4ax.com>; from news@manuelmgarcia.com on Fri, Apr 18, 2003 at 11:43:42PM +0000
References: <3o11avop6vkh3ake8dgbtst93t1or771ct@4ax.com>
Message-ID: <20030419111610.A587@tibia.amotlpaa.bogus>

Quoth Manuel Garcia:
> I would like to ask the group, are reference cycles bad form?
> 
> I am debating whether to give a 'child' object a reference to its
> 'parent' object during creation of the 'child'.

Imho this is an engineering question, not a style question.  It's
a trade-off between the benefits of the algorithms you can write
when the data structure has reference cycles and the slightly
inferior garbage collection properties of such structures.  The
right trade-off to make depends on the circumstances.

For example, if you create and discard structures frequently,
timeliness of collection might make a big difference to the memory
profile of your program.  In such a case, reference cycles are
probably a bad choice.  (In CPython at least, noncyclic structures
are collected by reference counting, which is generally more
timely than the cyclic garbage collector.)

For another, if you need to move rapidly from child to parent,
reference cycles may be a good thing.

There are alternatives, of course.  Perhaps the parent node can be
computed from properties of the child; how fast is that
computation, and how frequently must it be done?

Perhaps the code which needs to move from child to parent can keep
track of the entire list of ancestors from the root to the node
presently of interest; then moving to the parent of the node is
just del ancestors[-1], which is speedy.  How much memory does
that list take?  How many instances of it do you need?  What
should happen if other code moves nodes in the tree while such a
list is being used?

Perhaps you can keep the tree structure in two dicts, one mapping
from parent to a list of its children and one mapping from child
to parent.  How much memory does that take?  Discarding an entire
subtree is more costly in this approach; how often need that be
done?

These are all engineering questions.  Good style in such cases is
imho to use a data structure appropriate to the problem at hand.
Each of the solutions above is suitable in some cases.

  [...]
> Using the 'weakref' module seems silly.  I have complete confidence
> with Python's garbage collection, so why slow things down with extra
> indirection?

Perhaps because reference cycles in which (at least) one link is a
weak reference can be collected by reference counting.  Again,
whether this is a good trade depends on the circumstances.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"[T]rue greatness is when your name is like ampere, watt, and fourier
 -- when it's spelled with a lower case letter."      -- R.W. Hamming




From florian.konnertz at web.de  Wed Apr 30 09:26:10 2003
From: florian.konnertz at web.de (Florian Konnertz)
Date: Wed, 30 Apr 2003 15:26:10 +0200
Subject: pyshout: make fails
Message-ID: <b8ommf$esp$05$1@news.t-online.com>

Hi all,

I tried to install pyshout, it is needed for the ZRFP Zope product, an 
mp3 admin-, stream-, voting tool.

The README tells me:
To compile, copy the Makefile.pre.in from your Python installation and type
make -f Makefile.pre.in boot
make

When i do this i get:
root at grooveway:/usr/src/pyshout-0.0.6# make -f Makefile.pre.in boot
make: *** No rule to make target `boot'.  Stop.

I tried also with other makefiles:
Makefile.pre.in
Makefile.pre
Makefile
and with different targets (boot, install, etc.)

Unfortunately I'm not familiar yet with make etc. so i have no idea how
and where to get the correct makefile and how to adjust it.
I hope you can give me some tips about make, how to install C modules 
for python, etc.
Or has anybody already concrete experience with pyshout? Please report 
asap! - Google finds only very few resources about it..


TIA for your help!

Florian


-- 
Florian Konnertz   ---   http://www.florian-konnertz.de	
http://openspirit.homelinux.net/noowiki/FrontPage
ZWiki about all topics, especially consciousness research
and wisdom traditions




From blunck at gst.com  Thu Apr 17 09:42:05 2003
From: blunck at gst.com (Christopher Blunck)
Date: 17 Apr 2003 06:42:05 -0700
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk>
Message-ID: <1147e466.0304170542.a54eadb@posting.google.com>

Robin Becker <robin at jessikat.fsnet.co.uk> wrote in message news:<XiCE$IAjFmn+Ew43 at jessikat.fsnet.co.uk>...
> I get the impression that XP is intended as a design none/test all
> process where the code some grows organically into the requirement
> through bursts of coding and testing. It even seems as though the
> testing is primary.

XP definitely does not discount design, it just puts a slightly
different spin on design.  Let's simplify project development into two
categories:  design (D) and implementation (I).  Implementation
actually consists of coding, testing, documentation, etc.  Design
consists of requirements gathering, scoping, planning, and ...
designing.

The way that I illustrate the difference between XP and other
practices is to plot design and implementation activities with respect
to time.  Time grows as you move to the right:

Traditional:   DDDDDDDDDDDDDDDDDDDDIIIIIIIDIIIIIIIIIIIIIDIIIIIIIIIDIIIII
XP:            DDIIIDIIDIIIDDIIDIDDDIDIIDDIIIIIDDIDIIDIIDIIDIIDIIDIDIDII

I probably have not a very good job in my diagram, but in the end, an
XP guy may spend as much time designing as a traditional developer
(maybe not), but because the design and implementation are comingled
you (the manager or the developer or the customer), have much more
granular control over the evolution process.

Traditional methodologies try to pin down unknowns, and seek to
eliminate risk by making an upfront investment in design.  But this
behavior in and of itself creates risk:  you delay implementation
until you feel your design is sufficient, possibly eating into time
you'll need later on for testing, documentation, and oh ya I forgot -
development.  :)

One aspect of XP is postponement of decisions until they actually need
to be made.  That's not to say that we're flying by the seats of our
pants.  We just recognize that often "the passage of time" is often a
much better means by which unknowns can be solved.  An unknown is
identified early, and everyone scratches their heads and says 'Ya -
I'm not very clear on that....'.  The traditional crowd then huddles
into a room and tries to gather requirements, design it, talk about
it, and ultimately build it.  The XPers say 'Let's think about it, and
come back to it later when we have some time.'  During that time the
XPers run a  thread in the back of their head to process the unknown. 
Emails get sent out, the customer is talked to, time goes by.  Maybe
the requirement or source of the unknown goes away, or changes into
something completely different.  Or, as is often the case, the unknown
solidifies and the team has a much better understanding of the
requirement.  This can happen quite often, actually.

It's funny how traditional thinkers accuse XPers of trying to "jump
into the code too quickly", when a lot of time we actually forego
decision making until such time that it is really necessary.  I can
see an argument for a slightly different statement being made about
traditional programmers:  they "jump into UML too quickly".


> Can XP be reduced to choosing candidates from the pool of programs that
> best satisfy a set of tests? If so experience from GP suggests that the
> tests require care to avoid common pitfalls. As an example a program
> might complete all the tests perfectly, but fail on all other instances
> of data (this commonly happens to overspecialised solutions).

That's very true.  And is why testing is such an important aspect of
development.  Testing is just a safety net that hangs below you. 
Having a safety net is important, but if the net is 200 ft. below you,
you're going to croak when you fall.  The more hard coded (and thus
fragile) your unit tests
are, the lower you place the safety net.  Eventually it can get to a
point where one says 'why even have a safety net?'



> Do XPers combine differing solutions etc etc?

I've always found XP minded people to be extremely open minded, and
willing to try different solutions to see if they'll work.  I can't
say the same for my traditional colleaques.  They are extremely close
minded, skeptical of anything other than what they know (this
permeates everything:  languages, testing practices, frequency of
integration, etc), and very confrontational.

It's sort of like the irony that is Windows vs Linux:  Most people
that criticize Linux have never used Linux.  Most people that
criticize Windows have used Windows and have switched to Linux. 
Likewise, most people that criticize XP have never used XP. 
Similarly, most people that use XP have come from traditional worlds
and have moved on to something better.

Ok that was a generalization.  ;)


-c



From skip at pobox.com  Mon Apr 14 09:54:24 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 08:54:24 -0500
Subject: Python Jobs
In-Reply-To: <bd993a2f.0304140148.36298ef3@posting.google.com>
References: <bd993a2f.0304140148.36298ef3@posting.google.com>
Message-ID: <16026.48528.386872.224398@montanaro.dyndns.org>

    Anand> Can somebody give me a python job ? I am quite bored of my
    Anand> current job and I love python programming. I need a change
    Anand> desperately but I am not getting it. I am ready to relocate out
    Anand> of my country to U.S/U.K for the job.

Start here in your job search:

    http://www.python.org/Jobs.html

Skip




From roelschlaeger at twacs.com  Tue Apr  8 17:24:40 2003
From: roelschlaeger at twacs.com (Oelschlaeger, Robert)
Date: Tue, 8 Apr 2003 16:24:40 -0500 
Subject: Pythonwin and Vim
Message-ID: <B46DE74CC66AD61191CD000475996916DEDE19@mail.twacs.com>

Is there a known compatibility problem with Pythonwin and Vim?

Environment: 
	Win2k machine running Cygwin. 
	Pythonwin (PythonWin 2.2.2 (#37, Nov 26 2002, 10:24:37) [MSC 32 bit
(Intel)] on win32.)
	Vim (VIM - Vi IMproved 6.1 (2002 Mar 24, compiled Feb  9 2003
13:10:32))

I normally use Vim for editing Python code, and occasionally use Pythonwin
for debugging. 

If I 

-	open a file in Pythonwin, 
	-	perform some debugging, 
	-	edit the file using Vim, 
	-	continue debugging, 
-	then shut down Pythonwin, 

the edited file just disappears without a trace.

"DON'T DO THAT!"  Too late. I already did.

I was probably going to throw the code away anyway -- but I'd have like that
to have been MY choice.

Bob




From delete.this.part.alex.news at attbi.com  Wed Apr 30 12:01:07 2003
From: delete.this.part.alex.news at attbi.com (Alex)
Date: Wed, 30 Apr 2003 16:01:07 GMT
Subject: question on the compile built in function.  Filename ignored?
Message-ID: <7rSra.706428$S_4.744795@rwcrnsc53>

I am fairly to new to Python, so please bear with me.

I am trying to use the built in function compile.  The filename argument
seems to be ignored no matter what I try.  According to the docs, it should
be used in the output error message.  Here is a simple example:

Python 2.2.2 (#1, Apr 16 2003, 22:00:16)
[GCC 3.2.1 20021207 (Gentoo Linux 3.2.1-20021207)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> compile(':\n', 'abc', 'exec')
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "<string>", line 1
    :
    ^
SyntaxError: invalid syntax


>From the docs I would expect the line 

  File "<string>", line 1

to read

  File "abc", line 1


What am I missing?

Thanks in advance,

Alex





From grante at visi.com  Wed Apr 23 16:10:37 2003
From: grante at visi.com (Grant Edwards)
Date: 23 Apr 2003 20:10:37 GMT
Subject: Curses on Win32
References: <b86r8p$27n$1@gateway.northgrum.com>
Message-ID: <3ea6f33d$0$185$a1866201@newsreader.visi.com>

In article <b86r8p$27n$1 at gateway.northgrum.com>, Dang Griffith wrote:

> I'm running 2.3a2 and decided to spearmint with curses, only to
> find it's not available for import on Win32.  Is/was curses
> available in 2.2? Is there an alternative text user interface
> library available?

I was going to say I agree 100% with the subject line, but on
second thought Win32 is it's own curse.  It doesn't need them
added on... ;)

OT:  There is a PDCurses port to Win32, but I don't think
there's a Python wrapper for it.

-- 
Grant Edwards                   grante             Yow!  KARL MALDEN'S NOSE
                                  at               just won an ACADEMY AWARD!!
                               visi.com            



From martin at v.loewis.de  Sun Apr 13 13:52:56 2003
From: martin at v.loewis.de (=?ISO-8859-15?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Sun, 13 Apr 2003 19:52:56 +0200
Subject: intern'ed strings and deepcopy()
In-Reply-To: <yfsr886e3gt.fsf@black132.ex.ac.uk>
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk> <3E977928.B3A6DC7E@engcorp.com> <yfsznmueec1.fsf@black132.ex.ac.uk> <m3he92pef3.fsf@mira.informatik.hu-berlin.de> <yfsr886e3gt.fsf@black132.ex.ac.uk>
Message-ID: <b7c85g$slh$03$1@news.t-online.com>

Alexander Schmolck wrote:

>>Notice that Python does *not* (anymore) guarantee
>>
>>type(a) == type(b) == str and b == a 'implies' id(intern(b)) == id(intern(a))
> 
> 
> Huh? Really?
> 
> I thought  ``id(EXP1) == id(EXP2)`` implies ``EXP1 is EXP2`` and that's the
> only difference I can see between the two lines.
> 
> So does this no longer hold true in python 2.3: "The `is' operator compares
> the identity of two objects; the id() function returns an integer representing
> its identity"?

I meant to explain all this inthe paragraph on mortality of interned 
strings. If a is not interned (and no other string of a's equivalence 
class is), then intern(b) might return a newly-interned string (*). Then
its id() is taken, and then the result of the intern call is no longer 
used. So the newly-interned string may become uninterned. Then intern(a) 
is invoked.

In the original example, writing

   intern(b) is intern(a)

requires that the result of intern(b) is preserved atleast until the 
"is" operator is computed, so it is certain that the string won't die.

(*) In this specific example, this doesn't actually happen: intern(b) 
will return b itself, since no other string has been interned yet. So 
there are other references to the interned string (namely, the variable 
b), so intern(a) later returns b. To make my example more correct, I 
should write

type(a) == type(b) == str and b == a 'implies'
   id(intern(b+" s")) == id(intern(a+" s"))

so that the strings being interned are temporaries themselves. For some
reason, this still doesn't trigger the mortality of the interned string
in 2.3a2, but I haven't investigated why. In any case, there is no 
*guarantee* (anymore) that the id of an interned string stays the same 
over time; one should store the interned string itself, and not its id.

Regards,
Martin




From bokr at oz.net  Sun Apr 20 22:36:27 2003
From: bokr at oz.net (Bengt Richter)
Date: 21 Apr 2003 02:36:27 GMT
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com> <a3b5b608.0304080302.c26ad68@posting.google.com> <af6b5bdd.0304181001.558da6f@posting.google.com> <84of317cd6.fsf@boost-consulting.com>
Message-ID: <b7vlfb$27c$0@216.39.172.122>

On Sun, 20 Apr 2003 05:21:41 -0400, David Abrahams <dave at boost-consulting.com> wrote:
[...]
>
>I don't care much about XML programming (haven't had a use for it),
>but at PyConDC recently I got together with a couple of guys from the
>Twisted project who do care, and designed a nice little meta-language
>in Python for doing this sort of template processing.  Here's a little
>example of it in action:
>
Interesting ;-) Did you snip that from a working version? I didn't work one up,
but I'd guess the approach was to define classes or factory functions
for body, table, tr, td, etc. and have them return suitable objects with suitable
__getitem__ methods. So "body" seems like a predefined object, unlike the others?
Or are they all predefined instances with factory and normal __call__  methods?
Am I mistaken about the apparent extra ")" below?

>    template = body[
                body()[   # might be more consistent?
>        table(id="outer", width="100%", height="100%", border="0")[
>            tr(valign="bottom")[
>                td(id="output", width="75%", valign="top", model="latestOutput")[
>                    div(pattern="listItem", view="html")[
>                        "Foo"
>                    ]
>                ],
>                td(id="room-contents", valign="bottom")[
>                    strong[
>                        "Stuff you have"
>                    ],
>                    div(model="playerInventory", view="List")[
>
>                        # Note programmatic elements
>                        if_(not arg1)
>                        [    
>                            div(_class="item")["Nothing"]
>                        ].else_
>                        [
>                            for_each(arg1)
>                            [
>                                div(
>                                    style=["color: red", "color: blue", None]
>                                  , view="item"
>                                  , controller="look")[arg1]
                                   , controller="look" # or drop the ")" of the next line?
>                                )
                                 )[arg1]               # or drop the ")" of the previous line?
>                            ]
>
>                        ]
>                    ]
>                ]
>            ]
>        ] 
>    ]
>
>    # invoke the template with some data
>    template(['foo', 'bar', 'baz', 'another item'])
>
>which generates:
[ ... snip ...]
>I really loved this because it was the result of healthy
>cross-pollination between the cultures and techniques of two very
>different programming communities: Python and C++.
>
It's always nice to see people of different cultures able to
overcome xenophobia etc. and have a good time, but what did C++
per se actually have to do with the above? Was it done in C++ first?
;-)

Regards,
Bengt Richter



From gildemere at hotmail.com  Mon Apr 28 10:16:09 2003
From: gildemere at hotmail.com (Marc Floessel)
Date: 28 Apr 2003 07:16:09 -0700
Subject: Recursive Algorithm getting stuck (because of Python Limits?)
Message-ID: <236da4ed.0304280616.349e8bdd@posting.google.com>

I am trying to execute a simple Tree Building algorithm.

"
nTreeBreadth = 7
nTreeDepth = 7


class Node:
    def __init__(self):
        self.childNodes = []
    
# Create a recursive Sub-Tree with <nChildren> until we reach the
final <nDepth>
def CreateSubTree (oRoot, nChildren, nDepth):
    for i in range(nChildren):
        oNewNode = Node()
        oRoot.childNodes.append(oNewNode)
        if (nDepth > 1):
            CreateSubTree (oNewNode, nChildren, (nDepth-1))

oRoot = Node() # Create the Root Node
CreateSubTree (oRoot, nTreeBreadth, nTreeDepth-1)
"

This works. However increasing the Depth or Breadth to 8 seems to
break Python. Python rises to about 100 MB memory consumption (as
indicated by the Windows Task Manager) and then seems to drop and rise
randomly. The algorithm never finishes. Am I breaking the stack? I
tried executing the code (with a Depth of 8) in Stackless Python, but
it didn't finish either, so it doesn't appear to be the stack. Are
there any other hard limits?



From ianb at colorstudy.com  Sat Apr  5 02:58:36 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 05 Apr 2003 01:58:36 -0600
Subject: cgi.py -- anybody using it?
In-Reply-To: <Pine.LNX.4.44.0304050051160.1193-100000@alice>
References: <Pine.LNX.4.44.0304050051160.1193-100000@alice>
Message-ID: <1049529516.6022.11.camel@lothlorien>

On Fri, 2003-04-04 at 17:58, John J Lee wrote:
> I just tried testing some HTML form upload code using cgi.py, and
> discovered that it wants lower-case HTTP headers :-/

Have you actually seen this behavior, or are you getting this from
reading the code?  I don't even think the code handles headers -- the
CGI process turns headers into environmental variables, and case isn't
preserved in that process.

In some places you may read code that uses the mimetools module to parse
headers.  It has a dictionary interface, but keys are case-insensitive,
so if it uses m["content-type"] it's still work with Content-Type or
somesuch.

And yes, the cgi module is in very active use, not just for cgi
programs, but in many systems that handle the web (since it has
functions for query string parsing and many other things).

  Ian






From adalke at mindspring.com  Fri Apr 25 01:11:26 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Thu, 24 Apr 2003 23:11:26 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
References: <mailman.1051023275.29338.python-list@python.org> <b8420j$629$1@slb3.atl.mindspring.net> <mailman.1051043856.10118.python-list@python.org> <b86gim$60r$1@slb5.atl.mindspring.net> <mailman.1051127503.25664.python-list@python.org> <b87il2$3rf$1@slb6.atl.mindspring.net> <mailman.1051211247.5278.python-list@python.org>
Message-ID: <b8afri$oqt$1@slb4.atl.mindspring.net>

Steven Taschuk:
> Hm.  I'm not sure I understand what would make such a language not
> Python.

I believe apologies are due.  It appears I misread flippancy into
your response, so this one of mine is now more serious.

> I agree, of course, that no known Python has such special __and__
> and __or__ methods.  But that no *known* Python works as I
> suggested does not mean that no *possible* Python works thus.
> (Consider that not long ago the yield statement was not part of
> any known Python.  And imho generators are a much larger change to
> the language than two new special methods with unusual arguments.)

I don't think generators are that large a change.  Pre-generators I
wrote a lot of code which looked like this

class TwoLineIterator:
    def __init__(self, infile):
      self.infile = infile
      self._n = 0
    def next(self):
      line = self.infile.readline()
      if not line:
        return None
      self._n = self._n + 1
      return [line, self.infile.readline()]
    def __getitem__(self, i):
      if self._n != i:
        raise TypeError("forward iteration only!")
      x = self.next()
      if x is None:
          raise IndexError(i)
      return x

I got so I could crank out the code almost without thinking.

With iterators, I added

    def __iter__(self):
        return iter(self, None)

With generators I changed the whole thing to

def TwoLineIterator(infile):
    while 1:
        line = infile.readline()
        if not line:
          break
        yield [line, infile.readline()]

Externally, the interface is exactly the same

for data in TwoLineIterator(open("input.txt")):
    ...

Conceptionally then, I see yield (on top of the iterator protocol)
as a great way to simplify building an existing idiom.  And
in addition, it lets me turn highly stateful operations (like a
tree or graph traversal) into iterators.  When the element
generation is several loops deep, it's even harder.  Without
generators, I ended up using threads to get the iterator-like
feel I wanted.

But Python the language did let me get that feel, even in the
1.4 days, and adding generators to Python has not affected
that feel.

Consider though the "two new special methods with
unusual arguments."  Assume the details of getting the stack
frames, etc have been worked out, along with problems of
not knowing the __and__ method even exists until the code
is run.  (It could be added after the class is created.)

You're left with functions which behave different from any
other function, and where the behaviour cannot be used
anywhere else.

First, they behave different.  What's the object passed to
the function?  It's callable, it contains some execution stack
information, to get the locals.  It's a pretty complicated
thing.  Does that data type already exist?  Although I don't
know the internals that well, I don't think so, which means
it needs to be written, documented, available for introspection,
etc.

Can that object be saved?  Eg, what does this do?

class Spam:
    def __and__(self, other):
      self.saved = other
      return other()

spam = Spam()
try:
  a = spam and 1/0:
except ZeroDivisionError:
  pass
spam.saved()

For some languages, like Smalltalk and Ruby, this makes sense.
Those languages support code blocks, which can be passed
around and executed.  But Python does not.

Were Python to have code blocks, then it should be applied
more generally to Python code.  Consider the if/else PEP of
a couple months ago.  With code blocks, the following would
be possible

ifelse(a != 0, 1/a, None)

which would pass the 1/a and None as independent code
blocks, to be executed based on the evaluation of a != 0.

And Guido said that this sort of function would not be one
of the ways to resolve the PEP, which I take to mean that
it's not really a Python solution.

BTW, this would also require some sort of prototyping, to
specify which fields are to be passed as blocks vs. evaluated
before passing, as in

def ifself(cond, if_true is code, if_false is code):
  if cond:
    return if_true()
  return if_false()

but that just doesn't feel right, which I take to mean it
isn't really Python.  It looks like Python, but it doesn't
feel like it.

And that's why I say that code which handles this sort
of special __and__ and __or__ case, if added to Python,
would produce code which wasn't really Python.  It may
be related to Python -- given the feedback I've seen from
people who like code blocks in other languages, I can
understand why people would like them in a Python-like
language.  But it would not be Python.

                    Andrew
                    dalke at dalkescientific.com.





From aleax at aleax.it  Tue Apr  1 04:09:47 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 09:09:47 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift>
Message-ID: <vHcia.39470$Jg1.934856@news1.tin.it>

Julian Tibble wrote:
   ...
> his C++ program, since I'm not sure I can write a better hash table than
> the g++ STL developers.

I'm SURE you can write a hash-table that's a better hash-table than
C++'s std::map, because the latter CANNOT be a hash table -- it must
rely on ordering of keys, and is normally a red-black tree;-).


Alex




From hs2l2ATyahoodotcom at yahoo.com  Fri Apr 18 13:15:28 2003
From: hs2l2ATyahoodotcom at yahoo.com (Haran Shivanan)
Date: 18 Apr 2003 17:15:28 GMT
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de> <mailman.1050586468.29594.python-list@python.org> <rozna.115488$It5.70451@news2.central.cox.net> <b7mn1i$2icfr$2@ID-148859.news.dfncis.de> <LBBna.43583$T34.1210957@news2.tin.it> <b7nttn$2p743$3@ID-148859.news.dfncis.de> <2259b0e2.0304180555.6c063ec3@posting.google.com>
Message-ID: <b7pbrf$355p3$2@ID-148859.news.dfncis.de>

module "Michele Simionato" core dumped with the following output:
> Haran Shivanan <hs2l2ATyahoodotcom at yahoo.com> wrote in message news:<b7nttn$2p743$3 at ID-148859.news.dfncis.de>...
>> module "Alex Martelli" core dumped with the following output:
>> [snipped informative stuff]
>> > 
>> > This prints 11, as apparently required by your specs.  I hope
>> > this can help you make the custom metaclass you need to
>> > implement your actual, exact specifications, of course.
>> Thanks *a lot* for your help.
>> Do you know of any on-line resources where I can get info
>> on metaclasses and using them?
>> I was only dimly aware of their presence until your post.
>> Thanks.
> 
> Alex didn't want to mention it, but his "Python in a Nushell" contains a
> discussion of metaclasses and the metametaBunch example (plus many other
> interesting things not necessarely related to metaclasses ...)
> 
> Good luck,
Damn. Didn't realize he wrote the book.
The campus store doesn't carry any python books so I'm not even aware of what 
they are like.
Should try and get a copy of it.

Thanks for the heads-up.
> 
>                     Michele


-- 
We've heard that a million monkeys bashing away at a million keyboards
could produce a complete Shakespearian work.
Now, thanks to the internet, we know this is not true.



From jepler at unpythonic.net  Fri Apr  4 16:53:13 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 4 Apr 2003 15:53:13 -0600
Subject: how to test when correct output is "random"?
Message-ID: <20030404215312.GE7895@unpythonic.net>

My current project, Novelwriting[1] is a knock-off of something called
the Dada Engine[2].  Essentially, you write a file that describes the
structure of a document, and Novelwriting outputs a document that
matches the structure.

I'd like to use this project as an opportunity to teach myself something
about testing methodology, but since the output is based on a random
number generator, how can the normal "unit test" or "doctest" tools work
for me?

Some things are fairly easy to test.  For instance, there are
"concatenation" rules, which read like
    T: A B;
    A: "a";
    B: "b";
I can easily test that str(T) == "ab".

But other parts are based on randomness.  "Alternation rules" are the
foundation on which this is built.  An alternation rule might read
like this:
    S: "a" | "b" | "c";
Then, str(S) would yield one of the alternatives.  Over many runs, it
should yield each choice about an equal number of times.  So how do
I test this?  Run a large number of trials and do a statistical test?
This doesn't seem to fit in with the philosophy of software testing,
which stresses exact repeatability.

Jeff
[1] http://unpythonic.net/jeff/novelwriting/
[2] http://dev.null.org/dadaengine/




From adalke at mindspring.com  Mon Apr 14 17:32:49 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Mon, 14 Apr 2003 15:32:49 -0600
Subject: parser stack size limit
References: <mailman.1050341002.21628.python-list@python.org> <7h3of38sxp9.fsf@pc150.maths.bris.ac.uk>
Message-ID: <b7f96m$c9l$1@slb9.atl.mindspring.net>

Michael Hudson:
> I'd also consider a saner way of building data structures...

While I'm not sure if this is saner ... You might also try
converting the code directly.  I don't know if one of the
Python modules parses Python independent of the internal
C code, but even if not there are other parser packages
which include the ability to parse Python code, like SPARK.

If you really want to have fun, it's possible to build up
the pickle or marshalled form of the data and use that.

But I agree that these are all workarounds.

                    Andrew
                    dalke at dalkescientific.com





From just at xs4all.nl  Sun Apr  6 16:23:46 2003
From: just at xs4all.nl (Just)
Date: Sun, 06 Apr 2003 22:23:46 +0200
Subject: Is there swap function in Python?
References: <b6q0n6$1c55$1@mail.cn99.com> <3r119vosrdukh6til7bdmgredlhhehqg64@4ax.com> <b6q1so$1cm2$1@mail.cn99.com>
Message-ID: <just-25C845.22234606042003@news1.news.xs4all.nl>

In article <b6q1so$1cm2$1 at mail.cn99.com>, "Xin Liu" <lx1978 at 263.net> 
wrote:

> >
> > What do you mean by the swap function? What should it do
> > precisely?
> >
> > -- 
> > Ren? Pijlman
> It is a function like swap(a, b) . Use to swap the value of a and b.
> A example in C++ language maybe like:
> template <class T> void swap(T & a, T & b)
> {
>     T * tmp;
>     tmp = a;
>     a = b;
>     b = tmp;
>     return
> }

Try

   b, a = a, b

Just



From tdelaney at avaya.com  Tue Apr  1 16:53:38 2003
From: tdelaney at avaya.com (Delaney, Timothy C (Timothy))
Date: Wed, 2 Apr 2003 07:53:38 +1000
Subject: list index question
Message-ID: <338366A6D2E2CA4C9DAEAE652E12A1DE3DD0A8@au3010avexu1.global.avaya.com>

> From: Matt Singer [mailto:mrsinger at usa.net]
> 
> Can I wildcard? I don't know the values in the second column.  I want
> to find the first entry what has a particular value in the first
> column.

That's not at all what you said. Next time, please ask the question you want an answer to. Anna's response was perfectly correct for the question you asked.

http://www.catb.org/~esr/faqs/smart-questions.html

No, you cannot wildcard - it doesn't make sense for this.

You need to define you problem further.

1. Do you want the *index* of the first tuple which matches or
2. Do you want the first tuple that matches itself.
3. Is the data in the sequence sorted? If so, how?

In any case, since you already have the data in a sequence, looping is probably the best option here - it has the advantage of being O(N). If the data is already sorted in a way that you want, you may be able to use a more efficient search algorithm to find *one* of the entries that matches, then search backwards until you find one that doesn't match.

Tim Delaney




From jepler at unpythonic.net  Wed Apr  2 22:52:24 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Wed, 2 Apr 2003 21:52:24 -0600
Subject: plugin system?
In-Reply-To: <1049340834.8760.22.camel@musicbox>
References: <mailman.1048836529.7090.python-list@python.org> <KVVga.31670$i26.818965@news2.tin.it> <1049072379.21707.34.camel@musicbox> <20030331014306.GB21551@unpythonic.net> <1049340834.8760.22.camel@musicbox>
Message-ID: <20030403035224.GA8901@unpythonic.net>

On Thu, Apr 03, 2003 at 03:33:37PM +1200, Rob Brown-Bayliss wrote:
> 
> Whats with the following exec statment?
> 
[Jeff Epler:]
> > In the code, you would instantiate the class with a helper function:
> >     def get_plugin_factory(name):
> >         # If you have packages (eg monty_python.spam.VikingChorus
> >         # or spam.spam.spam.spam.baked-beans.spam) this won't work
> >         # as written
> >         m, f = name.split(".")
> >         d = {}
> >         exec "import m" in d
> >         return eval(name, d)
> 
> I admit I havn't tried it yet, but what does the exec "import m" in d
> do?

The idea in the (untested) code was that making packages work would be a
simple addition.  You'd make m get name up to the last '.', so that for the
plugin 'pack.mod.plugin' you'd 'import pack.mod' and then get the value of
pack.mod.plugin.  As written, though, only names of the form
'modulename.factoryname' are intended to work.  There's no reason you can't
do this with the __import__ approach too.

In my view, the thing returned should act as a factory (including the
possibility of just being a class), so if you call it more than once you
should get two independent instances of the plugin.

In your code, it looks like 'newcog' names a module, and you can't easily
make a copy of a module.  When you name a factory, it becomes easy to ask
for more than one, to parameterize, etc.

Jeff




From a.schmolck at gmx.net  Sun Apr 13 13:00:27 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 13 Apr 2003 18:00:27 +0100
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk> <3E97D324.8000009@v.loewis.de>
Message-ID: <yfsu1d2e3is.fsf@black132.ex.ac.uk>

"Martin v. L?wis" <martin at v.loewis.de> writes:

> Alexander Schmolck wrote:
> > But I'm not sure that that's what the OP's up to.
> 
> The original poster asked (paraphrased)
> "Am I right assuming that interned strings are not copied
> by deepcopy?"
> 
> to which the short answer is
> 
> "Yes, you are right"
> 
> The slightly longer answer is
> 
> "Yes. Strings are never copied by deepcopy since deepcopy
> knows they are immutable. Whether they are interned or not
> is irrelevant".
> 
> When Eric Max said "its identity is unimportant", he really
> meant "it's unimportant whether they are interned, and their
> identity is preserved in deepcopy" :-)


Ah, I see, thanks for clearing this up :)

'as



From tjreedy at udel.edu  Tue Apr 15 07:55:42 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 15 Apr 2003 07:55:42 -0400
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee>
Message-ID: <kMacnbnfl9jPbAajXTWcoA@comcast.com>

"Vahur Lokk" <vlokk at hot.ee> wrote in message
news:3e9bd039_2 at news.estpak.ee...
> Hi!
> What I want to do is write a small daemon that can catch Save File
type of
> events in the system and manipulate files being saved.
> Having mostly sorted out the rest of the program I cannot think of a
way to
> catch these saves.
> I must warn that I am complete newb and doing my first steps in
Python (and
> programming overall).

If you want to catch disk writes initiated by other programs, writen
in whatever language, then I believe you have taken on an expert-level
project. Operating systems generally do *not* allow application
programs to catch such low level system events.  One would have to
either patch the existing disk drivers or write a (pseudo)driver that
would sit on top of the existing one.

If you only want to catch file writes in Python programs, you would
only have to patch the Python interpreter on your system.

If you only want to catch file writes in your own programs, you should
simply call your function before or in the process of writing.

myfile.write(manipulate(bytes))

Or derive a myfile class from the builtin file class, give it a save
method, and call that method instead of the write method.

Terry J. Reedy





From tanzer at swing.co.at  Fri Apr 11 03:13:24 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Fri, 11 Apr 2003 09:13:24 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "Thu, 10 Apr 2003 17:27:33 GMT."
             <9Qhla.7802$d8.3212@nwrdny02.gnilink.net>
Message-ID: <E193siy-0006Np-00@tswings.swing.cluster>

Carl Banks <imbosol-1049990205 at aerojockey.com> wrote:

> > Maybe I'm missing something, but non-dynamically scoped globals just
> > seem like a bad idea to me.
>
> What you're missing is that Globals are Evil(tm), and furthermore,
> that if your program depends on changing a global in a dynamic scope,
> it badly needs to be redesigned.  (To wit, that functions that refer
> to the global you want to change need to be changed to accept it as an
> argument.)
>
> I think dynamic scpoing is a wholly bad idea, and simplifying a very
> misguided programming practice doesn't make it any less so.

You are of course right. Surprisingly though, it works very well for
Emacs. Passing arguments would *not* work in the context of Emacs.

Practicality-beats-purity-ly yr's,

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Wed Apr 30 09:43:07 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Wed, 30 Apr 2003 15:43:07 +0200
Subject: How to call built in function 'open' from inside object that
 has 'open' method?
In-Reply-To: <db81c492.0304300526.7b8d3c86@posting.google.com>
References: <db81c492.0304300526.7b8d3c86@posting.google.com>
Message-ID: <3eafd2eb$0$49100$e4fe514c@news.xs4all.nl>

vadim wrote:
> I am new to Python and do not know the answer to very
> simple how to 'open' from inside another object that
> has 'open' method. I know that in C++ I should use
> following syntax ::open. 


Best solution: rename the 'open' method of your class to
something that doesn't clash with a Python builtin, for
instance openWindow.

Other solution (ugly); call __builtins__.open

Third solution (ugly as well): pass in the 'open'
builtin function as a parameter to your method,
with anther name.


--Irmen




From tim.one at comcast.net  Sun Apr 27 22:17:44 2003
From: tim.one at comcast.net (Tim Peters)
Date: Sun, 27 Apr 2003 22:17:44 -0400
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <20030427183847.C28461@tibia.amotlpaa.bogus>
Message-ID: <LNBBLJKPBEHFEDALKOLCEEFJEEAB.tim.one@comcast.net>

[Steven Taschuk]
> ...
> (Not that I'm in any special position vis a vis changing the
> documentation.  Quite the contrary.)

I beg to differ, sir!  The people with the greatest power to change the docs
are those who actually submit thoughtful doc patches.  On that basis, you're
the most powerful fellow I've seen this month <wink>.





From ark at research.att.com  Mon Apr 28 18:27:26 2003
From: ark at research.att.com (Andrew Koenig)
Date: Mon, 28 Apr 2003 22:27:26 GMT
Subject: Slice confusion : a[n:p] is a list exclude the last  element p  (thus spoke Dijkstra)
References: <mailman.1051532982.25352.python-list@python.org> <b8k0fl$dsa$1@news.hananet.net> <yu99bryqjqwq.fsf@europa.research.att.com> <mailman.1051564607.15912.python-list@python.org>
Message-ID: <yu991xzmjm0h.fsf@europa.research.att.com>

andrew> otoh, pointers are annoyingly different to iterators (at least
andrew> in java - don't know about c++)

C++ iterators behave much like like pointers; there are several
categories of iterators that differ from one another in what
pointer-like operations they support.  So, for example, a ``forward
iterator'' supports ++ but not --, a ``bidirectional iterator''
supports both ++ and -- but not addition or subtraction of arbitrary
integers, and so on.

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From hemanexp at yahoo.com  Mon Apr 28 11:23:18 2003
From: hemanexp at yahoo.com (perl lover)
Date: Mon, 28 Apr 2003 08:23:18 -0700 (PDT)
Subject: Getting mouse position interms of canvas unit.
Message-ID: <20030428152318.80404.qmail@web41710.mail.yahoo.com>


hi,
  iam new to python and tkinter. I have created a
canvas of size 300m * 300m (in millimeter). I bind
mouse move method to canvas. when i move the mouse
over the canvas the mouse position gets printed in
pixel unit. But i want to get mouse position values
interms of canvas unit (ie, millimeter). How can i get
mouse position values interms of canvas unit?

   My program is given below.
   *****************************

from Tkinter import *
root = Tk()
c = Canvas(root,width="300m",height="300m",background
= 'gray')
c.pack()

def mouseMove(event):
        print c.canvasx(event.x), c.canvasy(event.y)

c.create_rectangle('16m','10.5m','21m','15.5m',fill='blue')
c.bind('<Motion>',mouseMove)
root.mainloop()

Tnanx



__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com




From rando-66507737.5.reiman at xoxy.net  Wed Apr 23 17:52:47 2003
From: rando-66507737.5.reiman at xoxy.net (Charlie Reiman)
Date: 23 Apr 2003 14:52:47 -0700
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org> <GXupa.83382$T34.2079161@news2.tin.it>
Message-ID: <5jadeg285s.fsf@xoxy.net>

Alex Martelli <aleax at aleax.it> writes:

> <posted & mailed>
> > 
> For your algorithm, why not just, for example:
> 
> class Infimum:
>     def __cmp__(self, other): return -1
> infimum = Infimum()
> 
> class Supremum:
>     def __cmp__(self, other): return 1
> supremum = Supremum()
> 
> as coded these have obvious anomalies, e.g. infimum < infimum -- but
> those are easy to remove if they give problems to your algorithm.


Are there any dangers in something like this:

a = [1, infimum, infimum]
a.sort()

Should the __cmp__ methods return 0 for equality (identity, in this
case)?

Charlie



From zyziu at interia.pl  Thu Apr 17 14:36:29 2003
From: zyziu at interia.pl (zyziu)
Date: Thu, 17 Apr 2003 20:36:29 +0200
Subject: scrollbar
References: <b7kh3g$51u$1@nemesis.news.tpi.pl> <MAlna.4040$ej5.2777@news01.roc.ny.frontiernet.net> <b7ltfi$49c$1@atlantis.news.tpi.pl> <mailman.1050575068.28377.python-list@python.org>
Message-ID: <b7ms8e$go6$1@atlantis.news.tpi.pl>

Uzytkownik "Martin Franklin" <MFranklin1 at gatwick.westerngeco.slb.com>
napisal w wiadomosci news:mailman.1050575068.28377.python-list at python.org...
> zyziu wrote:
> > ----- Original Message -----
> > From: "Lee Harr" <missive at frontiernet.net>
> > Newsgroups: comp.lang.python
> > Sent: Thursday, April 17, 2003 1:22 AM
> > Subject: Re: scrollbar
> >
> >
> >
> >>In article <b7kh3g$51u$1 at nemesis.news.tpi.pl>, zyziu wrote:
> >>
> >>>Hello. I made lanChat in Python. I have problem with Text widget. I
want
> >
> > to
> >
> >>>text from Entry goes to Text widget and scrollbar goes down, because I
> >
>
> <snip>
>
> >
> >     def Pisz(self):
> >          self.dane = self.tekst.get()
> >          self.tekst.delete(0,'end')
> >          if (self.dane<>''):
> >              self.okno.insert('end',strftime("%H:%M:%S")+'\n')
> >              self.okno.insert('end',self.dane+'\n'+'\n')
>
>   A call to yview_pickplace here should do the trick.....
>
> self.okno.yview_pickplace("end")
>
>
> Regards
> Martin
>
>
Thanks everyone for help. I need this pick of code:
self.okno.yview_pickplace("end")
and it works. Thanks once again. Greetings from Poland. Zyziu123





From jylimd at yahoo.com.removeMe  Tue Apr 15 09:58:07 2003
From: jylimd at yahoo.com.removeMe (J Li)
Date: Tue, 15 Apr 2003 09:58:07 -0400
Subject: any benchmark results for python?
Message-ID: <b7h40n$28s$1@skates.gsfc.nasa.gov>

Hi,

I am looking for benchmark test results for python, in comparison to other
languages like Perl or Java or C++

thanks

jylimd(at)yahoo(dot)com





From anton at vredegoor.doge.nl  Tue Apr 29 16:21:11 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Tue, 29 Apr 2003 22:21:11 +0200
Subject: Sorting without transitivity
References: <20030419213405.GA588@obstler> <mailman.1051641592.28102.python-list@python.org>
Message-ID: <b8mmve$d5s$1@news.hccnet.nl>

Frank Niessink <frank at niessink.com> wrote:

>Any comments? Suggestions for improvements?

There's probably an indentation error at the last else clause. However
the code happily runs, and pychecker doesn't complain about it either.
If I "correct" the indentation the code goes into an endless loop for
a simple test like:

print tlist([1,2,3])

Anton





From wesc at fuzzyorange.com  Fri Apr  4 02:11:20 2003
From: wesc at fuzzyorange.com (Wesley Chun)
Date: Thu, 3 Apr 2003 23:11:20 -0800 (PST)
Subject: ANN: BayPIGgies mtg Wed Apr 9 7:30pm
In-Reply-To: <Pine.LNX.4.31.0303111410220.1869-100000@emperor.deirdre.org>
Message-ID: <Pine.LNX.4.31.0304032309200.17780-100000@emperor.deirdre.org>

BayPIGgies: Silicon Valley-San Francisco Bay Area Python Users Group

When:     April 9, 2003 @ 7:30pm
Where:    Stanford University, Palo Alto, CA
Agenda:   Introduction to Machine Learning and Support Vector Machines
Speaker:  Asa Ben-Hur

The basic concepts of machine learning will be presented, with a focus
on Support Vector Machines (SVM), which have shown their effectiveness
in many domains of application. Examples of applying the methodology in
areas such as text categorization and bioinformatics will be discussed.

# Call For Talks: We are actively seeking speakers for BayPIGgies! If you
would like to give a talk at one of our 2003 meetings (any Python related
topic), contact us to coordinate!

more info including directions:    http://www.baypiggies.net

hope 2 c u tomorrow nite!

-wesley

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

"Core Python Programming", Prentice Hall PTR, ? 2001
    http://starship.python.net/crew/wesc/cpp/

Silicon Valley-San Francisco Bay Area Python Users Group (BayPIGgies)
    http://baypiggies.net

wesley.j.chun :: wesc at deirdre.org or wesc at fuzzyorange.com
cyberweb.consulting : henderson, nv : cyberweb at rocketmail.com
http://www.roadkill.com/~wesc/cyberweb/





From exarkun at intarweb.us  Thu Apr  3 11:45:16 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 3 Apr 2003 11:45:16 -0500
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304030705.524cf0fa@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com>
Message-ID: <20030403164516.GA30464@meson.dyndns.org>

On Thu, Apr 03, 2003 at 07:05:26AM -0800, Drew Moore wrote:
> "Anders J. Munch" <andersjm at dancontrol.dk> wrote in message news:<3e8c083b$0$10386$edfadb0f at dread11.news.tele.dk>...
> [snip]
> 
> > Take a step back and tell us what problem you are trying to solve.
> > Whatever it is, I'm sure we can think of a better solution than having
> > assignment depend on whatever object, if any, happened to be
> > previously bound to the same name.
> > 
> > python-is-not-c++-ly y'rs, Anders
> 
> my original need?
> I wanted to create an object that controls a voltage and
> use it at the python command line.
> by overloading operators, I can do:
> 
> voltage += 5  # (raise the voltage by 5 volts)
> voltage *= 2  # (double the current voltage)
> 
> but when I do 
> 
> voltage = 3  # (set the voltage to 3 volts)
> 
> my voltage object is clobbered with an integer object.
> I don't want to require the user to type any more than
> name = value # no name.val = newval, no name.setval(value)
> 

    class Conductor(object):
        def set_voltage(self, value):
            # Twiddle whatever needs to be twiddled
        def get_voltage(self):
            # Lookup whatever the voltage is
            return it
        voltage = property(get_voltage, set_voltage)

    c = Conductor()
    c.voltage = 3

  This is more along the lines of what you'd really do, anyway, right? 
After all, if "voltage" is simply a global, then your program or library
will be forever tied to controlling a single item across which an emf can
be put.  Not very scalable, right?

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 14 days, 12:00, 4 users, load average: 0.00, 0.03, 0.00




From staschuk at telusplanet.net  Thu Apr 17 15:06:05 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 13:06:05 -0600
Subject: Degree symbol (UTF-8 > ASCII)
In-Reply-To: <f7199550.0304170555.77773f8a@posting.google.com>; from pc451@yahoo.com on Thu, Apr 17, 2003 at 06:55:48AM -0700
References: <f7199550.0304160930.6cfeca65@posting.google.com> <3e9daad2$0$49115$e4fe514c@news.xs4all.nl> <f7199550.0304161732.48c75c42@posting.google.com> <mailman.1050548247.26781.python-list@python.org> <f7199550.0304170555.77773f8a@posting.google.com>
Message-ID: <20030417130605.A7968@tibia.amotlpaa.bogus>

Quoth Peter Clark:
> Steven Taschuk <staschuk at telusplanet.net> wrote in message news:<mailman.1050548247.26781.python-list at python.org>...
  [...]
> > So just do something like
> >     print >>fileobject, chr(176)
> >     print >>fileobject, u'\N{DEGREE SIGN}'.encode('latin-1')
> 
>     It worked fine when printing to a file (i.e., from the prompt),
> but I still got an error when I try to append it to a string destined
> to be included in a list:
> 
>     w += [u'\N{DEGREE SIGN}'.encode('latin-1') + scale.strip()]
> UnicodeError: ASCII decoding error: ordinal not in range(128)

And scale is a Unicode string, right?

    >>> '\xb0' + 'C'
    '\xb0C'
    >>> '\xb0' + u'C'
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    UnicodeError: ASCII decoding error: ordinal not in range(128)

(For the same reason as before: concatenating a normal string and
a Unicode string results in a Unicode string, which means that the
normal string has to be decoded into a Unicode string first.)

I'd suggest doing all your string assembly with Unicode strings,
and only encoding to ISO-8859-1 at the output stage.  Mixing the
two kinds of strings is asking for trouble, as you have seen.

-- 
Steven Taschuk                                                 o- @
staschuk at telusplanet.net                                      7O   )
                                                               "  (




From ianb at colorstudy.com  Sun Apr 20 16:23:32 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 20 Apr 2003 15:23:32 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <84of317cd6.fsf@boost-consulting.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <6ee58e07.0304061758.35070e7c@posting.google.com>
	 <a3b5b608.0304080302.c26ad68@posting.google.com>
	 <af6b5bdd.0304181001.558da6f@posting.google.com>
	 <84of317cd6.fsf@boost-consulting.com>
Message-ID: <1050870212.701.24.camel@lothlorien>

On Sun, 2003-04-20 at 04:21, David Abrahams wrote:
> I don't care much about XML programming (haven't had a use for it),
> but at PyConDC recently I got together with a couple of guys from the
> Twisted project who do care, and designed a nice little meta-language
> in Python for doing this sort of template processing.  Here's a little
> example of it in action:
> 
>     template = body[
>         table(id="outer", width="100%", height="100%", border="0")[
>             tr(valign="bottom")[
>                 td(id="output", width="75%", valign="top", model="latestOutput")[
>                     div(pattern="listItem", view="html")[
>                         "Foo"
>                     ]
>                 ],
>                 td(id="room-contents", valign="bottom")[
>                     strong[
>                         "Stuff you have"
>                     ],
>                     div(model="playerInventory", view="List")[
> 
>                         # Note programmatic elements
>                         if_(not arg1)
>                         [    
>                             div(_class="item")["Nothing"]
>                         ].else_

This could almost be plain Python, if Python allowed code blocks.  The
use of []'s isn't really necessary, and actually the expressions could
maybe be eval'd as well...

td(id='room-contents', valign='bottom', contents=[
  strong(contents=['Stuff you have']),
  div(model='playerInventory', view="List", contents=[
    if_(<not arg1>, 
        <[div(_class="item", contents=['Nothing'])]>,
        <[for_each("i", <arg1>,
                  <[div(style=['color: red', 'color: blue', None],
                        view="item", controller="look", 
                        contents=[arg1])]>)]>)])])

<> is the only nice nesting braces I could think of (`` wouldn't really
do, since it doesn't nest, not to mention it looks much to much like
'').  Of course, closing that nested structure is kind of crazy.

And actually <> wouldn't work, because of if <a > 10> -- it's not
entirely ambiguous, but it certainly *looks* ambiguous.  And << and >>
are also used, even if it's only for one of the more stupid operators. 
Maybe <()>...

It's all pipe-dreams, but code blocks sure would be nice.  I don't think
the argument that they allow New Control Structures is correct --
iterators and the traditional control structures actually cover
everything I've ever thought of.  Though now that I think about it, is
the thing above a new control structure?  A sort of inside-out control
structure, but also asort of lazy evaluation, like a callback.

I've played around with Python objects that allow lazily evaluated
expressions, by overloading all the comparison operators (though you
can't overload and, or, not, or is).  In some ways it pleases me, though
it also feels fragile -- and it's not really compatible with normal
Python, i.e., you couldn't call a normal function with a lazy argument
(though you could call a method on a lazy object easily enough).

  Ian






From claird at lairds.com  Mon Apr 28 15:22:19 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 28 Apr 2003 19:22:19 -0000
Subject: Web programming and a different "type" problem (was: Slice confusion : a[n:p] is a list exclude the last  element p)
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it>
Message-ID: <vaqvrb3ipqblf2@corp.supernews.com>

In article <UXara.22428$K35.745468 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
><posted & mailed>
			.
			.
			.
>It's an important design principle because of human-factors considerations,
>which Koenig expresses better than I can (I really wish he would one day
>write a "Python Traps and Pitfalls" book, but he seems too busy writing
>great C++ books instead... perhaps the Python one would be too thin?-).  
			.
			.
			.
I'd be happy to contribute "FMM" (which I've never
bothered trademarking, though I've been tempted 
several times.

I'm looking for an idiom.  I suspect it'll interest
some of the original readers of this thread; that
vague stylistic intuition is my justification for
this hijacking.

Here's an observation:  several Python-based Web
schemes (Zope's one, I know), share what I regard as
an infelicity.  The punch-line is this:  checkboxed
data has to treat single items different from all
other multiplicities.  Pages that naturally receive
such variants as
  ...mypage?value=value0
  ...mypage?value=value1&value=value2
are interpreted as
    ...
  valuelist = Web-scheme-function('value')
  for value in valuelist:
      my_process(value)
    ...
'Doesn't work, though.  Here's where the special case
enters:  if only one datum with the indicated name is
present as input, Zope and the others return it as a
simple string, rather than a list with that string as 
its sole member.  To iterate over all checkboxed items,
therefore, in our typical Web example, one must
    ...
  valuelist = Web-scheme-function('value')
  if valuelist == str(valuelist):
      valuelist = [valuelist]
  for value in valuelist:
      my_process(value)
    ...
That's certainly ugly.

What's a better way?  Is there even a better way to 
think about this, in terms of basic Web standards?
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From riemer at ppprs1.phy.tu-dresden.de  Mon Apr  7 17:29:02 2003
From: riemer at ppprs1.phy.tu-dresden.de (Tilo Riemer)
Date: Mon, 7 Apr 2003 23:29:02 +0200
Subject: Announce: apm.py 0.1.0
In-Reply-To: <20030407211858.GA19329@ppprs1.phy.tu-dresden.de>
References: <20030407211858.GA19329@ppprs1.phy.tu-dresden.de>
Message-ID: <20030407212902.GA19379@ppprs1.phy.tu-dresden.de>

Hello,

I forgot:
http://www.iapp.de/~riemer/projects/apm.py/


Best regards, Tilo




From max at alcyone.com  Tue Apr 22 17:03:00 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 14:03:00 -0700
Subject: Twisted 1.0.4
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local> <341410000.1051007646@leeloo.intern.geht.de> <mailman.1051014938.19603.python-list@python.org>
Message-ID: <3EA5AE04.CF4DB11@alcyone.com>

Andrew Bennetts wrote:

> No, gcc is GPLed.  I'm not sure what you mean by "gcc libs"?  I'm
> pretty
> sure libgcc1 is part of gcc and thus GPLed also.

Linking against a GPL'd library requires your project to be GPL'd. 
That's why the gcc (and g++) libraries are LGPL'd, because they do not
enforce this restriction.

gcc itself is under GPL.  Its libraries are under LGPL.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Do not stand in a place of danger trusting in miracles.
\__/ (an Arab proverb)
    PyQStat / http://www.alcyone.com/pyos/qstat/
 A Python wrapper around QStat, the realtime game server status tool.



From mb at muenster.de  Thu Apr 10 16:57:17 2003
From: mb at muenster.de (Martin Bless)
Date: Thu, 10 Apr 2003 20:57:17 GMT
Subject: popen(), sendmail: Success?
Message-ID: <3e95d63f.2635484@news.muenster.de>

To send mails from Python here's the example my provider provides:
This http://faq.puretec.de/skripte/python/3.html

The relevant lines are these:

sendmail = os.popen("/usr/lib/sendmail -t", "w")
sendmail.write(msg)
sendmail.close()

This works just fine. But is it enough? How do I know the operation
succeeded?

I've read the docs about os.popen[234] and the module popen2 but I'm
not sure where to go. And since experimentation isn't easy in this
case I'd appreciate some hints or a concrete example very much.

Do I have to use popen2.popen3() and its wait() method? I have to
admit that I don't understand too much of what's being said about this
in the docs ... :-(

Martin





From peter at engcorp.com  Thu Apr 17 09:13:27 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 17 Apr 2003 09:13:27 -0400
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com> <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk> <3E9D8B1B.24050D34@engcorp.com> <7h3brz6qmsc.fsf@pc150.maths.bris.ac.uk> <3E9DB4AC.E94829D9@engcorp.com> <3e9e5242_2@news.estpak.ee> <7h3u1cxpmlr.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3E9EA877.B2326F56@engcorp.com>

Michael Hudson wrote:
> 
> Vahur Lokk <vlokk at hot.ee> writes:
> 
> > Peter Hansen wrote:
> >
> > > Michael Hudson wrote:
> >
> > Ooops!
> > I remember my firstever post to linux mailing list caused a month-long
> > flamewar there ;-)
> > Did I do it again?
> 
> If this is your idea of a flame war, I suggest you try
> comp.lang.perl.misc :-)
> 
> As far as I can tell Peter & I are merely vigorously agreeing with
> each other...

YES, DAMMIT!  THAT'S EXACTLY WHAT WE'RE DOING!!!!!  WHY ANYONE WOULD
THINK THIS WAS A FLAME-WAR IS BEYOND ME... SOME PEOPLE JUST HAVE NO
IDEA HOW TO CARRY OUT A REAL CAMPAIGN OF USENET TERROR THESE DAYS.
IT'S TIME THE REGIME WAS CHANGED: THE PSU HAS BEEN GETTING TOO



From no at spam.invalid  Tue Apr 29 17:10:55 2003
From: no at spam.invalid (Russell E. Owen)
Date: Tue, 29 Apr 2003 14:10:55 -0700
Subject: Some oddities in 2.3b1, esp. with Tkinter
Message-ID: <b8mpp1$25hm$1@nntp6.u.washington.edu>

I finally got my app almost running in 2.3b1. I've run into the 
following:

1) A filter to remove fonts with non-ascii names (used to filter out 
stuff that I could not display in a pop-up menu) broke.
            try:
                unicode(fontName, 'ascii', 'strict')
                return True
            except UnicodeError:
                return False

Python 2.2 throws UnicodeError. Python 2.3b1 throws TypeError. The fix 
is trivial (e.g. "except (UnicodeError, TypeError)". The filter is 
probably not needed anymore anyway, since Tk 8.4.2 can apparently 
display menu items with non-ascii characters.

2) Tk file events are no longer supported:

RuntimeError: _tkinter.createfilehandler not supported for threaded Tcl

This is painful since it's hard to see how to write a proper networking 
GUI with Tkinter without file events for reading data. Perhaps I can get 
them back again by disabling the new threaded tcl support. However, I 
suspect it's time to try out the Twisted framework.

3) Tk returning things as objects is nice, but I had this break at least 
one thing of mine and it will break others (see 4 below):

I was assuming that the values in the dictionary returned by 
tkFont.Font.config() were strings. Fortunately this was easy to fix. A 
bit disconcerting -- I wonder how many any other little pitfalls like 
this still buried in my code -- but well worth the cost of the change.

4) Tk Variables are trickier to use and the change will break some 
existing code.

Tkinter used to return a variable name if the variable of a widget was 
requested. Not as convient as returning the original variable, but once 
one found the magic setvar and getvar functions (which work on variable 
names) things were fine.

Now Tk returns a _tkinter.Tcl_Obj instead, and and setvar and getvar 
don't seem to work on that. (I'm relieved that I never actually used 
setvar and getvar in my final code; despite many temptations I was 
always uncomfortable with it).

Example:
v = StringVar()
v is a <Tkinter.StringVar instance at 0x2c98a0>
l = Label(textvariable=v)
x = l["textvariable"]
x is reported as a <namespaceVarName object at 0x0168b048>
but is a _tkinter.Tcl_Obj based on the following traceback:
root.getvar(x)
Traceback...
TypeError: getvar() argument 1 must be string, not _tkinter.Tcl_Obj

Note that x.get() doesn't work either -- i.e. the object does not mimic 
a Tk Variable. I'm guessing there is some way to manage this, but have 
no idea what it might be.

5) Note that Tkinter's handling of font objects does not seem to have 
changed. I was hoping it would start returning font objects instead of 
strings, or offer some way to convert the string to a font object, but 
apparently not:
f = tkFont.Font(font="system")
l = Label(font=f)
l["font"] returns: 'font2922816'

-- Russell



From scott_list at mischko.com  Mon Apr 14 17:03:08 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Mon, 14 Apr 2003 14:03:08 -0700
Subject: more informative stack traces
In-Reply-To: <3e9ab723$0$66212$edfadb0f@dread11.news.tele.dk>
References: <e0bf36b5.0304132100.4324c3b8@posting.google.com> <3e9ab723$0$66212$edfadb0f@dread11.news.tele.dk>
Message-ID: <200304141403.08895.scott_list@mischko.com>

On Monday 14 April 2003 06:27, Anders J. Munch wrote:
> "L.C." <deja-google at wotevah.com> wrote:
> > I am proposing a different style of stack trace which I found more
> > useful in debugging problems with live code. The main change involves
> > showing the actual parameter values in function calls, gdb-style.
> > It sounds simple because it really is, but it saves lots of guesswork
> > when navigating through the stack to see what caused a problem.
> >
> > An implementation is available here: http://www.wotevah.com/code/log.py
> > To enable, just import it in your main program.
> >
> > It requires syslog (thus Unix-only) because it seemed to be the natural
> > mechanism for reporting errors and warnings, although it will also use
> > stderr if you are on a terminal.
>
> May I suggest you refactor your code into two modules: a
> platform-independent one that inspects tracebacks and generates the
> traceback text in some sort of container or generator.  And a
> syslogging module one with the same interface as you have now, which
> is a client of the first module.  And perhaps a third module, a client
> that uses portable stderr output.
>
> Not only will this give you a fighting chance of convincing people on
> syslog-less platforms of the merits of your approach, but it will also
> benefit the structure of your code.  Hardwiring the output mechanism
> puts obstacles in the way of applications such as CGI scripts and GUI
> traceback presenters, where consoles and logfiles may be irrelevant
> because user interaction goes through other channels.

Any possibility of getting this idea to be part of Python? It seems like the 
Right Thing to me.  It could be made user-configurable (on or off) so that 
backward compatibility could be maintained.

Scott





From daniele.trovato at teleport.ch  Wed Apr 16 15:00:55 2003
From: daniele.trovato at teleport.ch (Daniele Trovato)
Date: Wed, 16 Apr 2003 21:00:55 +0200
Subject: PyQT and Eric
References: <b7is7g$i1d$1@rex.ip-plus.net> <mailman.1050479366.18545.python-list@python.org>
Message-ID: <b7k998$2c9e$1@news.imp.ch>

@Phil   I hope the answers helpful enough !
> What version of PyQt?
3.5-54

> python -c "from qt import qVersion, PYQT_VERSION"
Traceback (most recent call last)
File "<string>", line 1, in ?
File "/usr/lib/python2.2/site-packages/qt.py", line 39, in ?
import libqtc
ImportError: /usr/lib/python2.2/site-packages/libqtcmodule.so: undefined
symbol: _ZNK16QAssistantClient9classNameEv

> Can you run any of the PyQt examples?
No but I have the same error output
Traceback (most recent call last)
File "<string>", line 1, in ?
File "/usr/lib/python2.2/site-packages/qt.py", line 39, in
.....................................................

@G?nter
>Install PyQt sources and recompile. Fixed the problem for me
I will try it, but I'm not an expert in recompile sources an installing it.
_____________________________________________________________________
Thanks
Daniele






From max at alcyone.com  Fri Apr 11 17:01:40 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 11 Apr 2003 14:01:40 -0700
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com>
Message-ID: <3E972D34.6B829FCC@alcyone.com>

Fortepianissimo wrote:

> I can't seem to find the exact description between these two in the
> Library Reference - so am I correct in assuming that deepcopy() will
> not actually duplicate an intern'ed string?

Strings are immutable, so what good would a copy (via copy or deepcopy)
do you?  You cannot possibly change an existing string object, so its
identity is unimportant.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The price of eternal vigilance is indifference.
\__/ Marshall McLuhan
    Crank Dot Net / http://www.crank.net/
 Cranks, crackpots, kooks, & loons on the Net.



From roy at panix.com  Mon Apr  7 20:20:13 2003
From: roy at panix.com (Roy Smith)
Date: Mon, 07 Apr 2003 20:20:13 -0400
Subject: exceptions == errors?
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
Message-ID: <roy-AB9C56.20201307042003@reader2.panix.com>

Mark Harrison <mh at pixar.com> wrote:
> Are exceptions always considered errors, or there the same feeling
> as in C++ that exceptions can be used for non-error situations?

I think of exceptions as being anything out of the normal flow of 
processing.

I'm more fluent in Python than I am in C++, so I may be wrong here, but 
my perception is that the C++ community treats exceptions with a lot 
more fear and lothing than the Python world does.  I suspect part of the 
problem is that they are relatively more expensive in C++.  They are 
also relatively non-portable, and somewhat complex to fully understand 
how they interact with threads and object creation/destruction.

In Python, exceptions seem much more natural and lightweight.  I don't 
see any reason not to use exceptions liberally if they make the control 
flow easier to understand
 
> 	while 1:
> 		try:
> 			read request
> 			process request
> 		catch MyException,e:
> 			send response contained in e
> 		catch Exception,e:
> 			send "internal error" response

That seems perfectly plausable.  I assume that your "process request" 
line includes sending some normal response to the client.  You've now 
got all three possible flow paths covered -- normal operation, abnormal 
operation handled by the server, and "should never happen" faults which 
at least you handle gracefully (better to tell the client that you've 
lost your mind than to just stop talking).

And, of course, "catch" is spelled "except" in Python :-)



From aussie2010 at yahoo.com  Wed Apr  2 14:32:56 2003
From: aussie2010 at yahoo.com (Chris)
Date: 2 Apr 2003 11:32:56 -0800
Subject: Problem with Idle
References: <435d40e0.0304011438.5b7b0595@posting.google.com>
Message-ID: <435d40e0.0304021132.a80fe3e@posting.google.com>

Gerrit and Thomas: Thanks for the info.



From missive at frontiernet.net  Wed Apr 16 19:22:20 2003
From: missive at frontiernet.net (Lee Harr)
Date: Wed, 16 Apr 2003 23:22:20 GMT
Subject: scrollbar
References: <b7kh3g$51u$1@nemesis.news.tpi.pl>
Message-ID: <MAlna.4040$ej5.2777@news01.roc.ny.frontiernet.net>

In article <b7kh3g$51u$1 at nemesis.news.tpi.pl>, zyziu wrote:
> Hello. I made lanChat in Python. I have problem with Text widget. I want to
> text from Entry goes to Text widget and scrollbar goes down, because I can't
> see what write another user. Now I must do it manually. How do it to
> scrollbar goes down automatically? Thanks for answer. Zyziu from Poland.
> 
> 

You should probably be a little bit more specific, and maybe include
some of the code around where you are having the problem.

That said, if this is Tkinter, maybe something here will help:


      self.wout = Tkinter.Tk()
      self.frame = Tkinter.Frame(self.wout, bg='brown')


      self.frame.sc = Tkinter.Scrollbar(self.frame, orient=Tkinter.VERTICAL)

      self.frame.tout = Tkinter.Text(self.frame, yscroll=self.frame.sc.set,
					 bg='black', fg='white')

      self.frame.sc['command'] = self.frame.tout.yview
      self.frame.sc.pack(side='right', fill='y')


         t = self.c.read_eager().replace('\r','')

         if t:
            self.w.frame.tout.config(state='normal')
            self.w.frame.tout.insert(Tkinter.END, t)
            self.w.frame.tout.config(state='disabled')
            self.w.frame.tout.yview(Tkinter.END)   




This is just bits and pieces of a file I wrote a couple of years ago... 
I remember having trouble with the same thing.  Hope this helps.





From staschuk at telusplanet.net  Wed Apr  9 23:31:35 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 9 Apr 2003 21:31:35 -0600
Subject: optimizing memory usage of large in memory structures
In-Reply-To: <aebfcc5a.0304090817.2c683c9e@posting.google.com>; from theory79@yahoo.com on Wed, Apr 09, 2003 at 09:17:40AM -0700
References: <aebfcc5a.0304090817.2c683c9e@posting.google.com>
Message-ID: <20030409213135.B2493@tibia.amotlpaa.bogus>

Quoth Hugo Liu:
> working with corpus analysis and large knowledge bases, i often have
> to load large arrays into memory.  each array element is just a text
> string.  although the text is only 5mb, python consumes 40+ mb.  i
> would suspect that the overhead is from storing type info for each
> element.  [...]

Have you tested this?  For example, if you replace your list of
elements with a same-sized list of Nones, how much memory is
consumed?  How much memory is consumed if you load the 5MB file as
one big string?

-- 
Steven Taschuk                                                 o- @
staschuk at telusplanet.net                                      7O   )
                                                               "  (




From donn at drizzle.com  Tue Apr 15 03:01:04 2003
From: donn at drizzle.com (Donn Cave)
Date: Tue, 15 Apr 2003 07:01:04 -0000
Subject: Using __repr__ or __str__ for own printable class?
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org> <b7di0n$dd6gj$1@ID-169208.news.dfncis.de> <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net>
Message-ID: <1050390062.742825@yasure>

Quoth Mads Orbesen Troest <mads at troest.NEVERMORE.dk>:
| ... That was what I was after in the first place, and it seemed like the 
| way to do it is to override the __str__ method. I would like to be able to 
| do this ...
| 	print someInstanceOfMyClass
|
| ... and then determine what should be printed. This works with __str__ but 
| I was/am afraid if this overriding of __str__ will conflict with other uses 
| of __str__ and/or __repr__.

As you have seen, there is some confusion over the principles.

I personally think the simplest and most useful way to look at it is

  __str__: your object's data as a string

  __repr__: your object as a string.

(The object's data may be generated during __str__, for example
random gibberish, contents of a file, whatever.  My point is that
one calls str() to get those file contents, random gibberish or
whatever, while one calls repr() to get information about the
object and not strictly about the data.)

For most objects they're the same, because there isn't any obvious
and distinct value of __str__.  For a string, clearly if you make
a string from its data, you've got the same string, so str(a) = a.
However the object as a string carries quotes to identify itself.
repr(instance of A) != repr(instance of B) - even if they contain
the same data, they are different as objects.

The idea that repr() must marshal the object is a sorely misguided
extension of this.  It is not consistently supported enough to be
clearly useful, and it has done some harm, e.g., it apparently
contributed to a move to make the repr for floats to display full
precision.  That behavior in turn seems to be responsible for a good
deal of the complaining over why str(list) has to use repr on its
items.

	Donn Cave, donn at drizzle.com



From aleax at aleax.it  Tue Apr 29 12:09:53 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 16:09:53 GMT
Subject: Raising objects
References: <453d10b2.0304290750.222fb453@posting.google.com>
Message-ID: <ltxra.31737$K35.1001968@news2.tin.it>

Carlos P. wrote:

> Hi all,
> 
> I have this short question: can the operator 'raise' raise exceptions
> which are instances of object? I'm getting the following behaviour
> from python 2.2.2:

No, in 2.2 (and 2.3 beta 1) exceptions must still be classic-classes
or instances thereof [or strings for backwards compatibility with the
dark and distant past], and should be derived from Exception (but this
is a recommendation, not enforced).  I do not know when (or if) this
restriction will be removed -- in particular, because I do not know
of "use cases" where the ability to raise new-style classes or instances
thereof would be a major boon to an application.  Maybe you have some...?


Alex




From mcherm at mcherm.com  Fri Apr 11 11:03:27 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Fri, 11 Apr 2003 08:03:27 -0700
Subject: __slots_ and inheritance
Message-ID: <1050073407.3e96d93fded70@mcherm.com>

As Alexander Schmolck has so eloquently explained, use of __slots__ (which is 
intended simply as a speed-and-memory optimization) breaks several things... 
among them:
 1. It breaks weakrefs (can't make a weakref to a __slots__ object).
 2. It breaks pickling (can't pickle a __slots__ object without custom code).
 3. It breaks introspection (eg: can't call vars() on a __slots__ object).
 4. It breaks "decoration" with new attributes (can't add new attributes
      if object has no __dict__).

Of these, #4 was intentional, but I don't see any reason for the others!
For instance, introspection (most of it anyhow) could be enabled for 
__slots__ objects by modifying vars() (and several other things) to
check the __slots__ for a list of additional values. Pickling too could
set and get values listed in __slots__. I haven't any idea what the
issue with weakrefs would be.

So can anyone out there tell me if these are (a) planned, but not done
yet, or (b) welcome, patches would be appreciated, or (c) difficult or
impossible for some reason I'm not aware of.


-- Michael Chermside


PS: There are also a few breakages that I really don't care about. Eg:

 5. Problems if a subclass re-declares a variable already in use.
 6. No "default values" for instances through class variables.
 7. Things don't work right if you modify __slots__.

I'd say "just don't do that". If you need such things, don't use 
__slots__.





From walter at livinglogic.de  Thu Apr 10 04:21:52 2003
From: walter at livinglogic.de (=?ISO-8859-1?Q?Walter_D=F6rwald?=)
Date: Thu, 10 Apr 2003 10:21:52 +0200
Subject: HTML -> HTMLGen (or similiar)
In-Reply-To: <aT1la.8930$8Z5.56436@sea-read.news.verio.net>
References: <aT1la.8930$8Z5.56436@sea-read.news.verio.net>
Message-ID: <3E9529A0.6060504@livinglogic.de>

VanL wrote:

> Hello,
> 
> I have a bunch of existing HTML that I need to integrate into a new web 
> app.  I would ideally like to use one of the existing HTML generation 
> systems (HTMLgen, HyperText, any others) but I don't want to have to 
> build up an equivalent representation for the many pages I need to work 
> with.

You can use XIST for that (available from
http://www.livinglogic.de/Python/xist/)

Code looks like this:

=====
from ll.xist import parsers, presenters
e = parsers.parseURL("http://www.python.org/", tidy=True)
print e.repr(presenters.CodePresenter())
=====

This dumps the Python website as an XIST DOM constructor, looking
something like this (with whitespace removed):

ll.xist.xsc.Frag(
    ll.xist.ns.xml.XMLStyleSheet('href="./css/ht2html.css" ..."'),
    ll.xist.ns.html.html(
       xsc.Comment(' THIS PAGE IS AUTOMATICALLY ... '),
       xsc.Comment(' Fri Apr  4 10:59:44 2003 '),
       xsc.Comment(' USING HT2HTML 2.0 '),
       xsc.Comment(' SEE http://ht2html.sf.net '),
       xsc.Comment(' User-specified headers:\nTitle: Python ...'),
       ll.xist.ns.html.head(
          ll.xist.ns.html.title(
             'Python Language Website'
          ),
       ...

You might want to shorten the class names (i.e. removing the
"ll.xist.ns." part). The final output code might look like
this:

from ll.xist import xsc
from ll.xist.ns import xml, html

e = xsc.Frag(
    xml.XML10(),
    html.DocTypeXHTML10transitional(),
    xml.XMLStyleSheet('href="./css/ht2html.css" ..."'),
    html.html(
       xsc.Comment(' THIS PAGE IS AUTOMATICALLY ... '),
       ...
    )
)

print e.asBytes(encoding="utf-8")

Hope that helps!

Bye,
    Walter D?rwald




From aleax at aleax.it  Thu Apr 17 06:36:40 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 10:36:40 GMT
Subject: Memory leak while looping
References: <1020cbf8.0304162100.6fc49e6f@posting.google.com> <n8t1n-ms4.ln1@isomorphism.org>
Message-ID: <Ysvna.38944$T34.1111479@news2.tin.it>

Brandon Beck wrote:

> It appears your "leak" is with the call to range().  range(1, 10000000)
> causes a list of 10000000 to be created in memory.  This has nothing to
> do with item being assigned in the body of the loop, and you should see
> the same behavior in a script without the for loop and just the call to
> range.  Consider using the generator (or maybe it's just generator like)
> version of range, xrange.  That should run in constant memory.

Yes.  xrange is "just generator-like".  If you only need to repeat the
action N times and don't need the loop index in the body, consider also
Python 2.3's itertools.repeat (if you can afford to wait for Python 2.3).

Here are some measurements with Python 2.3 alpha on my old Athlon 1.2GHz
machine with Linux...:

[alex at lancelot python2.3]$ python timeit.py -s'import itertools as it' 'for 
x in range(1000*1000): pass'
10 loops, best of 3: 2.99e+05 usec per loop
[alex at lancelot python2.3]$ python timeit.py -s'import itertools as it' 'for 
x in xrange(1000*1000): pass'
10 loops, best of 3: 1.06e+05 usec per loop
[alex at lancelot python2.3]$ python timeit.py -s'import itertools as it' 'for 
x in it.repeat(0,1000*1000): pass'
10 loops, best of 3: 8.68e+04 usec per loop
[alex at lancelot python2.3]$

i.e., range's overhead for a million items is 300 milliseconds, xrange's
is 100 milliseconds, itertools.repeat's is 87 milliseconds -- and if you
are in a truly serious bottleneck, every little bit helps;-).


Alex




From skip at pobox.com  Sat Apr 12 14:11:32 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sat, 12 Apr 2003 13:11:32 -0500
Subject: Adding item in front of a list
In-Reply-To: <ZUXla.7029$T34.191442@news2.tin.it>
References: <mailman.1050038237.25551.python-list@python.org>
        <ZUXla.7029$T34.191442@news2.tin.it>
Message-ID: <16024.22228.274095.619740@montanaro.dyndns.org>

    Alex> Why can't mylist.insert use just the same "first argument means
    Alex> different things" convention as range...?

Guido's feeling is that the convention used by range was a mistake, and
intends not to repeat it.  (I believe this line of reasoning was already
posted by Tim.)

Skip




From markscottwright at hotmail.com  Tue Apr  8 00:32:46 2003
From: markscottwright at hotmail.com (msw)
Date: Tue, 08 Apr 2003 04:32:46 GMT
Subject: How to protect python code ?
In-Reply-To: <mailman.1049769313.31229.python-list@python.org>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com> <mailman.1049769313.31229.python-list@python.org>
Message-ID: <Ohska.730$001.647543@monger.newsread.com>

Dave Brueck wrote:
> And your failed search proves my point:
> 
> http://tinyurl.com/91al
> 
> (found it in <30 seconds!)
> 
> I'm sure there are plenty of people who read this group who would be happy
> to do contract work for you to code this up for you, but all the people
> who are willing to do it for you for free have spoken.

?

*I* didn't ask for any help with this problem - I offered some advice to 
the original poster.  I suggested that he distribute a version of the 
interpreter with custom "import" code and that he look elsewhere for 
advice on this issue, since the signal/noise ratio is so low when it 
comes to this question.  The fact that your post is the first one that 
references any actual code, that it occurs about 20 posts into this 
thread, and that it is snide and insulting indicates that even if my 
first suggestion wasn't very good, my second one was spot on.

Mark




From mgerrans at mindspring.com  Thu Apr 17 02:45:20 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Wed, 16 Apr 2003 23:45:20 -0700
Subject: wmi ?
References: <b7khea$kt2$1@news-reader13.wanadoo.fr>
Message-ID: <b7liis$3s1$1@slb3.atl.mindspring.net>

There is a GetObject() in win32com.client, which works like the one in
JScript or VBScript.

Here is an interpreter session example of using it:

>>> import win32com.client
>>> WMIService =
win32com.client.GetObject(r"winmgmts:{impersonationLevel=impersonate}!\\.\ro
ot\cimv2")
>>> BIOS = WMIService.ExecQuery("Select * from Win32_BIOS")
>>> BIOS[0].Manufacturer
u'Award Software International, Inc.'






From asnd at triumf.ca  Thu Apr  3 23:13:45 2003
From: asnd at triumf.ca (Donald Arseneau)
Date: 03 Apr 2003 20:13:45 -0800
Subject: state bits of an event  - on unix and mac
References: <b6ibad$376$1@news.iucc.ac.il>
Message-ID: <yfillyq6gsm.fsf@triumf.ca>

Noam Raphael <noamr at users.sourceforge.org.replace.org.with.net> writes:

>    For this I need to interpret the state variable which is generated
> for each event. The problem is that I've seen no documentation on the
> meaning of the bits of event, and I want my class to work on unix, win
> and mac, so finding it from my experience only won't be too good.

The %s state value is OS/wm dependent, so there is no standard
meaning to the bits.

Much more portable is to bind all of the modified events
(<Shift-1>, <Alt-1>, <Ctrl-1>, <Ctrl-Shift-1>, etc)

It would be a very useful enhancement if Tk gave a uniform
state string with events (most conveniently, in exactly the
format used to create bindings: "Ctrl-Shift-", or even combined
with the keysym "Ctrl-Alt-Return")

Donald Arseneau                          asnd at triumf.ca



From gh at ghaering.de  Thu Apr 24 19:06:42 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 01:06:42 +0200
Subject: Exception and finally question
In-Reply-To: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com>
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com>
Message-ID: <3EA86E02.30603@ghaering.de>

Tung Wai Yip wrote:
> Hello,
> 
> How to I access the exception object in a handler?
> 
> try:
>     raise NameError, 'HiThere'
> except NameError:
>     #how to print the message 'Hi There' here?
> 
> Why except and finally can't be used together? I like to log a error
> message if an exception is throw. And I a file to be guaranteed to be
> close. What can't I do them together?

try:
	try:
		...
	except ...:
		...
finally:
	...


HTH,

-- Gerhard




From mwh at python.net  Tue Apr  8 06:44:58 2003
From: mwh at python.net (Michael Hudson)
Date: Tue, 8 Apr 2003 10:44:58 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049731219.27404.python-list@python.org> <b6teqm$956ai$1@ID-169208.news.dfncis.de> <mailman.1049776891.5548.python-list@python.org> <b6tllv$8uchs$2@ID-169208.news.dfncis.de>
Message-ID: <7h3el4di7yu.fsf@pc150.maths.bris.ac.uk>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> writes:

> Sean 'Shaleh' Perry wrote:
> > However the reason most of the emacs modes, vim syntax definitions,
> > etc all don't quite work is they are regex based.  To really get
> > where the  structure editor needs to be would require a parser for
> > each item type.  Not a trivial task.
> 
> No, but it would be a much better foundation to build
> this stuff on, I think.

Has anyone reading this thread tried out leo?  I haven't, but I have
this feeling it might be interesting...

Cheers,
M.

-- 
  My hat is lined with tinfoil for protection in the unlikely event
  that the droid gets his PowerPoint presentation working.
                               -- Alan W. Frame, alt.sysadmin.recovery



From jepler at unpythonic.net  Fri Apr 25 16:23:43 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 25 Apr 2003 15:23:43 -0500
Subject: Permanent Clock
In-Reply-To: <d613c14.0304251200.33f1b741@posting.google.com>
References: <d613c14.0304251200.33f1b741@posting.google.com>
Message-ID: <20030425202341.GH30690@unpythonic.net>

In general, there's not a clock that the user can't set.  You could use
a remote time protocol to get the time from a well-known server, which
might make things harder for a user to circumvent your count (but also
means it would be useless when offline!).  You also have the problem
that the user could easily replace or remove the count file.

Both of these problems are more or less independent of the Python
language--questions like "can I trust information X from the operating
system" and "can I prevent the user from tampering with the contents of
a file" apply to all languages.

For PCs, I know of a device called a "timeHASP" which apparently
contains a unique identifier and a tamperproof clock.  I don't know if
any of these devices can also store some amount of user-writable data,
but it seems plausible (I believe that in our software, the presence or
absence of certain license features is written into the HASP before we
distribute it, but I don't know the details).  We use HASP with the
flexlm license library on Windows, and on its own from Linux.  These are
all C-libraries, but wrapping them using a tool like ctypes, pyrex, or
swig would probably not be hard.

Of course, if what you want is to give a trial user a certain number of
days to try the product, you can't give them a dongle when they download
it from your website.. it's a hard question.

import pickle, time
def MauricioCounter(file):
    # time must be tamperproof
    now = time.localtime()[:3] # year, month, day

    if os.file.exists(file):
        f = open(file, "rb")
        # file contents must also be tamperproof
        # (and replayproof!)
        then, count = pickle.load(f)
        f.close()
        if now > then:
            count = count + 1
    else:
        count = 1

    f = open(file, "wb")
    pickle.dump((now, count), f, 1)
    f.close()

Jeff




From jkraska at san.rr.com  Thu Apr 17 00:06:05 2003
From: jkraska at san.rr.com (Courageous)
Date: Thu, 17 Apr 2003 04:06:05 GMT
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com>
Message-ID: <hr9s9v8pgqr154h15lsjbk1pjs5u5n6n65@4ax.com>

>Note that Guido deeply regrets permitting one-line blocks; this is
>something he may try fixing for Python 3.0.

Eh? Rationale?

C//




From john.abel at pa.press.net  Tue Apr 29 06:45:57 2003
From: john.abel at pa.press.net (John Abel)
Date: Tue, 29 Apr 2003 11:45:57 +0100
Subject: Passing Named Paramater To A C Func?
Message-ID: <3EAE57E5.5010202@pa.press.net>

Hi,

I'm writing a C function, which will have a Python wrapper, and I wish 
to be able to pass it an optional, named parameter.  For example:

def funcName( dirName, fileName=None ):

	if fileName is None:
		Some Other Stuff

	cCall( dirName, fileName )

Is the best to handle this in the Python wrapper, or is there a way of 
handling these via C.  None of the extending documents seem to mention 
anything about this.

I'd be grateful for any pointers.

Thanks

John





From boldyrev+nospam at cgitftp.uiggm.nsc.ru  Sat Apr 19 01:12:03 2003
From: boldyrev+nospam at cgitftp.uiggm.nsc.ru (Ivan Boldyrev)
Date: 19 Apr 2003 12:12:03 +0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <b7m546$2a83$1@news.peterlink.ru> <lyptnlxq1n.fsf@cartan.de> <m24r4xdyd6.fsf@mycroft.actrix.gen.nz> <lyd6jlxkpi.fsf@cartan.de> <usft9v01l8fcdg8erdl2kishg0cnr1t9dp@4ax.com> <lyadepp1ck.fsf@cartan.de> <1lt0avslergps3bdv1lpvr48kef6t03640@4ax.com>
Message-ID: <ce27nxk3b.ln2@elaleph.borges.cgitftp.uiggm.nsc.ru>

On 8353 day of my life Thaddeus L. Olczyk wrote:
> 
> >> >And you have to be a bit relaxed when using this requirement; for
> >> >instance, writing an Y-combinator in ML is very hard because of the
> >> >stupid type checker.

> >> No it is easy.

> >> In OCaml, 
> >>     let rec fix f x=f (fix f) x;;

> >That's cheating and you know it :-)

> No. It's not.
> It's a standard definition of Y in a combinatorial approach
> to computing.

But 'let rec' does Y-combination behind scene.  So, it is trick.

Try to implement Y-combinator without rec...

-- 
Ivan Boldyrev
PGP fp: 3640 E637 EE3D AA51 A59F  3306 A5BD D198 5609 8673   ID 56098673

        Outlook has performed an illegal operation and will be shut down.
        If the problem persists, contact the program vendor.



From steven at twitch.net  Sun Apr 27 12:57:18 2003
From: steven at twitch.net (Steven Hodgen)
Date: Sun, 27 Apr 2003 09:57:18 -0700
Subject: Textbooks?
Message-ID: <oproar1sjtx3zh5c@news.mindspring.com>

Hello,

I'm currently teaching a high-school introduction to programming course.  
I'm currently teaching in Java, but am seriously considering changing to 
Python.  The problem is that I cannot seem to find any good textbooks, much 
less ones for students with no prior programming experience.  This is the 
one area in which Java excels.  I'm using an awesome programming 
environment called "BlueJ", and there are lots of textbooks for students at 
all levels.

However, Java is not a great language for beginners.  Python, on the other 
hand, is ideal.  If I can find a textbook that is even ok, I may consider 
changing, since the language is so much better for students at this level.  
Doesn anyone know of a book like this?  Ideally, it would have examples and 
excercises, and perhaps even a web-site with other instructor and student 
resources.

Thanks!

--Steven
steven at twitch.net



From aussie2010 at yahoo.com  Sun Apr 13 06:12:42 2003
From: aussie2010 at yahoo.com (Chris)
Date: 13 Apr 2003 03:12:42 -0700
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com> <mailman.1050087439.11600.python-list@python.org>
Message-ID: <435d40e0.0304130212.5659bd4d@posting.google.com>

Inyeol Lee <inyeol.lee at siimage.com> wrote in message news:<mailman.1050087439.11600.python-list at python.org>...
> On Thu, Apr 10, 2003 at 05:53:53PM -0700, Chris wrote:
> > Jason Tiller <jtiller at sjm.com> wrote in message news:<mailman.1049417184.1536.python-list at python.org>...
> > 
> > > 
> > > Good luck!  I hope this pattern at least gives you a starting point
> > > for implementing the split() in Python.
> > > 
> > > ---Jason
> > > Sonos Handbell Ensemble
> > 
> > Jason, thanks a lot for the example and the explanation. I fired up
> > Perl and it worked like a charm. Now to try and understand it!  I
> > looked at stuff about Python regular expressions and there is
> > something called finditer and findall that might be useful for parsing
> > a string, but can't seem to find any documentation on them.
> 
> It's in Library Reference section 4.2.3. (I've checked this in 2.3a2
> document, but It'll be there in 2.2 also.)
> 
> Brief summary for findall and finditer is;
> 
> findall returns list of matched string.
> finditer returns iterator of match object. You can do a lot more than
> findall since it supports all match object attributes.
> 
> Inyeol...

Thanks Inyeol. Be nice if they gave a small example of using finditer.



From bignose-hates-spam at and-zip-does-too.com.au  Sun Apr  6 21:31:11 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Mon, 07 Apr 2003 01:31:11 GMT
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com>
Message-ID: <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake>

On Sun, 06 Apr 2003, Erik Max Francis wrote:
>  Ben Finney wrote:
> > In ork.py I want:
> >     import eek
>  
> If the toplevel package is put in a place where Python can find it,
>  	import monty.foo.eek

It's not.  Even if I make it so on my computer, it's not going to be so
on someone else's computer when they download my package tree.

-- 
 \     "If you were going to shoot a mime, would you use a silencer?"  |
  `\                                                  -- Steven Wright |
_o__)                                                                  |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From peter at engcorp.com  Mon Apr  7 08:58:02 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 07 Apr 2003 08:58:02 -0400
Subject: urllib and binding to a interface on a multihomed system
References: <3e915ca3@news.mt.net.mk>
Message-ID: <3E9175DA.652C58E5@engcorp.com>

?????? ?. wrote:
> 
> I have a computer that's connected to three networks.
> On that computer I have a python script that uses urllib or urllib2 to fetch
> urls and POST data...
> 
> The problem is, I want the outgoing IP address of the connections made by
> urllib[2] to be one of the interfaces and not the default selected by the
> kernel. I can do this to low-level sockets with the "bind" method... Can I
> do the same thing to urllib?

Judging by the source code for urllib.py, which calls httplib.py, which 
directly creates and immediately attempts to connect() the socket involved
in its HTTPConnection.connect() method, the answer is No.

-Peter



From ckea25d02 at sneakemail.com  Tue Apr  1 21:29:38 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 14:29:38 +1200
Subject: The Whitespace Language
In-Reply-To: <283adf56.0304011143.4f4a6934@posting.google.com>
References: <283adf56.0304011143.4f4a6934@posting.google.com>
Message-ID: <b6dhf2$43bo6$2@ID-169208.news.dfncis.de>

Kevin Dahlhausen wrote:
> I'm getting so much entertrainment out of the the joke on the jokers
> that created the Whitespace Language page -since we use indentation to
> delimit block strucutre, white-space has been part of Python's syntax
> since the beginning.  Ha!

Except that we usually try to *avoid* mixing
tabs and spaces... :-)

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mike at edgelinkrecruiters.com  Fri Apr 18 18:02:48 2003
From: mike at edgelinkrecruiters.com (Mike Miadich)
Date: Fri, 18 Apr 2003 15:02:48 -0700
Subject: Job Opportunity
Message-ID: <NHBBIABMOLGOBKFNCIFJKEHCCAAA.mike@edgelinkrecruiters.com>

Hi all,

I am a Portland-based recruiter looking for an embedded Linux software
engineer with Python, c/c++ programming skills.  Music interest/background
is a must.  The position is full time and located in the Beaverton/Hillsboro
area. Do you know anyone I could speak to?  Thanks.

Mike Miadich
Director, Recruiting Services
EdgeLink, LLC
(503)-246-3989 work
(503)-539-5371 cell
(503)-246-4375 fax
mike at edgelinkrecruiters.com






From peter at engcorp.com  Wed Apr  2 14:57:08 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 14:57:08 -0500
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n ...)
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com> <mailman.1049298845.7655.python-list@python.org>
Message-ID: <3E8B4094.F61515B3@engcorp.com>

Jp Calderone wrote:
> 
>   Now, in its current state, it is impossible for the Python interpreter to
> know this.  What if range() is rebound?  So this can't be done now.  But, if
> some of the talk about forbidding the shadowing of builtins bears fruit, it
> might someday be possible for CPython to make this optimization, and all
> without any ugly syntax hacks.

Ouch!  *Forbidding* it?!  That would cripple some of my most sophisticated
and effective unit and acceptance tests!  I dearly hope that "forbid" was
an overly strong choice of verb.

What would be wrong with "discouraging", or "making it clear to the 
compiler when it happens" or something?

Where is this discussion going on, and can I/need I participate to put
forth my perspective on this.  I'm coming to *count* on this as one of
the most effective solutions for certain kinds of testing.  (Specifically,
by creating a mock file system by intercepting open() calls in my own
method, to allow generating of various file system errors and situations
and to increase performance of unit tests by avoiding the real file
system.)

-Peter



From b.maryniuk at forbis.lt  Sat Apr  5 06:05:40 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Sat, 5 Apr 2003 13:05:40 +0200
Subject: Pythonic way of web-programming
In-Reply-To: <b6kfjv$4cc$1@panix2.panix.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <b6kfjv$4cc$1@panix2.panix.com>
Message-ID: <200304051305.40152.b.maryniuk@forbis.lt>

On Friday 04 April 2003 19:32, Aahz wrote:
> Probably the "standard" answers these days for the basic building block
> of a web application is either Zope (if you want a full-blown framework
> for content management) 

Zope ain't just a "full-blown framework for content management". This what 
CMF-like or KONTENTOR-like products does. Zope is an application server, 
where you can build anything for web you want, even intergrate it with J2EE 
via SOAP/XML-RPC.

-- 
Regards, Bogdan

A computer without any MS Windows is like a fish without a bicycle.
		-- With apologies to Gloria Steinem





From bwitherspoon at emseal.com  Tue Apr 29 15:59:07 2003
From: bwitherspoon at emseal.com (Bill Witherspoon)
Date: Tue, 29 Apr 2003 15:59:07 -0400
Subject: mxODBC puzzle
Message-ID: <3EAED98B.2090009@emseal.com>

I've been fooling around with mxODBC/mxDateTime for a couple of days 
now, trying to effect an Update query. Can anyone spot the error I'm 
making with the code below:


sql = "Update tblOEOrders set tblOEOrders.ScheduleDate = #" + 
newdate.strftime("%m/%d/%Y") + "# where tblOEOrders.OrderID = " + 
str(orderid) + ";"

myconn = mx.ODBC.Windows.Connect('production')
mycursor = myconn.cursor()
numlines_affected = mycursor.execute(qrystring)

The DB is Access97 via DSN (which is *not* read only), and query seems 
to return OK (1 record affected), but the record never gets updated. 
Using the same code I can do Selects without difficulty.
Do I have to do a fetchall() equivalent to make the update query run, or 
is something else wrong?

TIA,
Bill.





From peter at engcorp.com  Fri Apr 11 08:36:12 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 11 Apr 2003 08:36:12 -0400
Subject: collecting data from file
References: <3E9671A3.2050004@sympatico.ca>
Message-ID: <3E96B6BC.E9DBE74F@engcorp.com>

Mustafa Celik wrote:
> 
> I want to scroll thru a file;
>    * find lines that match a string (e.g. HELLO) on the 2nd column
>    * add up the 4th column (an integer) on each matching line, say the
> variable is TOTAL
>    * subtract the 4th column from TOTAL if another is string (e.g. BYE)
> is hit on 2nd column of a line
> 
> Any tips?

Yes, post an example showing what you intend.  The instructions are
decent, but implementable only by making a few assumptions about what
you mean.  Also, are there any error conditions to be handled, or do
you guarantee the input is always perfect (e.g., always at least four
columns, and so on)?

-Peter



From rob at zoism.org  Mon Apr  7 03:21:10 2003
From: rob at zoism.org (Rob Brown-Bayliss)
Date: 07 Apr 2003 19:21:10 +1200
Subject: PYTHON EDITOR!!
In-Reply-To: <03040711314900.15230nanthan@siswa.utm.my>
References: <03040711314900.15230nanthan@siswa.utm.my>
Message-ID: <1049700081.6819.56.camel@musicbox>

> is cooledit the right choice.

Only you can decide that.  I use Moleskine
(http://www.micampe.it/moleskine.html) which is written in Python, has
syntax highlighting, multiple tabbed windows etc etc.

-- 

*  Rob Brown-Bayliss
*  =================	
*      zoism.org






From vlokk at hot.ee  Thu Apr 17 03:10:15 2003
From: vlokk at hot.ee (Vahur Lokk)
Date: Thu, 17 Apr 2003 10:10:15 +0300
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com> <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk> <3E9D8B1B.24050D34@engcorp.com> <7h3brz6qmsc.fsf@pc150.maths.bris.ac.uk> <3E9DB4AC.E94829D9@engcorp.com>
Message-ID: <3e9e5242_2@news.estpak.ee>

Peter Hansen wrote:

> Michael Hudson wrote:

Ooops!
I remember my firstever post to linux mailing list caused a month-long 
flamewar there ;-)
Did I do it again?

Wahur



From gregory.p.green at boeing.com  Mon Apr  7 17:03:04 2003
From: gregory.p.green at boeing.com (Green, Gregory P)
Date: Mon, 7 Apr 2003 14:03:04 -0700
Subject: Problem setting COM property using win32com
Message-ID: <E1B4F80EB080194BBAD3092392F511B3DADEF1@XCH-NW-13.nw.nos.boeing.com>

EnsureModule doesn't do the trick.

Here is what I have:

from win32com.client.gencache import EnsureModule
MyModule = EnsureModule('{C85E7012-8B4F-11D1-8E20-00A02498EA3C}', 0, 1, 0)

That loads the module but it doesn't help. I'll show the relevant portions of the genpy module:

class ISmRecordList(DispatchBaseClass):
	"""An object used for storing data. A collection of headers and records, each composed of a collection of nodes.  In turn, each node can contain information of various types."""
	CLSID = pythoncom.MakeIID('{C85E7013-8B4F-11D1-8E20-00A02498EA3C}')

# This CoClass is known by the name 'SmRecList.SmRecordList'
class SmRecordList(CoClassBaseClass): # A CoClass
	# An object used for storing data. A collection of headers and records, each composed of a collection of nodes.  In turn, each node can contain information of various types.
	CLSID = pythoncom.MakeIID("{C85E7014-8B4F-11D1-8E20-00A02498EA3C}")
	coclass_sources = [
	]
	coclass_interfaces = [
		'{C85E7013-8B4F-11D1-8E20-00A02498EA3C}',
		'{C85E7015-8B4F-11D1-8E20-00A02498EA3C}',
	]
	default_interface = '{C85E7013-8B4F-11D1-8E20-00A02498EA3C}'

-----Original Message-----
From: Paul Prescod [mailto:paul at prescod.net]
Sent: Monday, April 07, 2003 1:34 PM
To: Green, Gregory P
Cc: python-list at python.org
Subject: Re: Problem setting COM property using win32com


I don't know why that's happening, but I would suggest you take more 
control of the instantiation by doing something like this:

from win32com.client.gencache import EnsureModule
MyModule = EnsureModule('{3BBEE982-742B-11D2-BD0A-00A024C36143}', 0, 1, 0)

reclist = MyModule.SmRecList()

In my recent COM project, I used the "dispatch" mechanism and pythoncom 
was smart enough to pick up my gen_py'd module. But the mechanism above 
is fine too. In fact, it is probably better by virtue of being more 
explicit...

Green, Gregory P wrote:
> I don't think the gen_py class is being used:
> 
> print obj: <COMObject SmRecList.SmRecordList>
> 
> print obj.__class__: win32com.client.CDispatch
> 
> print dir(obj): ['_ApplyTypes_', '_LazyAddAttr_', '_NewEnum', '_Release_', '__AttrToID__', '__LazyMap__', '__call__', '__cmp__', '__doc__', '__getattr__', '__getitem__', '__init__', '__int__', '__len__', '__module__', '__nonzero__', '__repr__', '__setattr__', '__setitem__', '__str__', '_builtMethods_', '_enum_', '_find_dispatch_type_', '_get_good_object_', '_get_good_single_object_', '_lazydata_', '_make_method_', '_mapCachedItems_', '_oleobj_', '_olerepr_', '_print_details_', '_proc_', '_unicode_to_string_', '_username_', '_wrap_dispatch_']






From aahz at pythoncraft.com  Wed Apr 16 20:00:40 2003
From: aahz at pythoncraft.com (Aahz)
Date: 16 Apr 2003 20:00:40 -0400
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it>
Message-ID: <b7kqr8$jsq$1@panix2.panix.com>

In article <Wrena.33096$T34.876856 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>Steve Holden wrote:
>>
>> You'd do better to call it "Python for VB programmers", expecially since
>> the change to VB6 is so radical that migrating to a new and more sensible
>> language might be an attractive option to some!
>
>Changes from VB5 to VB6 weren't all that radical.  It was the change
>from VB6 to VB.NET, aka VB7, that was the shocker (dozens of backwards
>incompatible changes -- mostly taking VB's semantics closer to Python's,
>btw, and therefore actually for the better, but still...).  I thought
>the window of opportunity -- millions of VB6 programmers reluctant to
>move to dotNET -- had more or less closed by now, but Aahz apparently
>has identified that it's still open.  If that perception is correct,
>then the book would indeed have quite a target-market, potentially.

Correct.  I was at the Waterside Conference last week (put on by my
agent's company, aimed at writers and publishers).  One presentation
claimed that a recent survey found 50% of VB6 programmers refusing to
upgrade to VB.net.  While that's likely overstated, that's a good
indication of where the market currently is, and even a tiny fraction of
that market would be a big boost for us.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From info at kestlerfinancial.com  Wed Apr 23 10:08:06 2003
From: info at kestlerfinancial.com (Kestler Financial Group)
Date: Wed, 23 Apr 2003 09:08:06 -0500 (CST)
Subject: Request-a-Quote
Message-ID: <LYRIS-78697267-291775-2003.04.23-09.11.31--python-list#python.org@xtinmail1.exacttarget.com>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030423/4ebed59b/attachment.html>

From cnetzer at mail.arc.nasa.gov  Sat Apr 19 16:31:31 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 19 Apr 2003 13:31:31 -0700
Subject: arg-passing style question
In-Reply-To: <A4ioa.87812$Vzu.36931@news02.bloor.is.net.cable.rogers.com>
References: <A4ioa.87812$Vzu.36931@news02.bloor.is.net.cable.rogers.com>
Message-ID: <1050784291.1950.4.camel@sayge.arc.nasa.gov>

On Sat, 2003-04-19 at 13:12, Guy Middleton wrote:
> I want to pass two values to a function that expects a single argument.
> So, I could wrap them in either a list or a tuple.  Is either way preferred
> over the other?

I'd go tuple, but some guidelines would suggest a list.

For more discussion, see these threads:

http://mail.python.org/pipermail/python-list/2003-March/152761.html
http://mail.python.org/pipermail/python-list/2003-March/152768.html

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From maxm at mxm.dk  Fri Apr 25 06:30:18 2003
From: maxm at mxm.dk (Max M)
Date: Fri, 25 Apr 2003 12:30:18 +0200
Subject: Data-driven testing
In-Reply-To: <3EA87475.F6A19BE@engcorp.com>
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com>
Message-ID: <058qa.30797$y3.2515229@news010.worldonline.dk>

Peter Hansen wrote:

> I'm asserting that I have actually seen "throwaway" scripts, even predicted 
> their imminent development, and subsequently have actually thrown them away.
> I'm not asserting I do this with 100% accuracy.  I'm asserting that
> I very infrequently do decide that a script will be throwaway, and
> that in those rare cases I often turn out to be right, and often 
> enough that it seems worth continuing to make such predictions in
> order to save myself some time.


Actually I write a LOT of "throwaway" scripts, that I know I will not 
need again, or that if I will use them again, I will just rewrite them.

It is far more important for me that these scripts are up and running 
fast than that they are correct.

I use them for massaging large amount of text. If ie. I need to make a 
website for a customer, and I have a large amount of similar texts. It 
could be something like a list of employees that is keept in a flat 
file, text document, excell sheet or similar.

Or it could be the texts from an older website.

Usually these files are pretty similar, but I know they are full of 
errors, misspellings etc. So I have to edit them manually anyway.

I write a small script that can save me a lot of copy-pasting, adding 
files etc. by automating most of the process. These scripts only have to 
be "good enough" as it is faster to manually edit minor errors in the 
resulting output than to rewrite the script.

I think testing would be absolutely useless in these cases.

Doing these scripts has saved me countless hours in the world of editing 
messy human managed data.

Actually this kind of usage turned me on to Perl, and later to Python.

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From edcjones at erols.com  Sun Apr  6 09:55:38 2003
From: edcjones at erols.com (Edward C. Jones)
Date: Sun, 06 Apr 2003 09:55:38 -0400
Subject: Pyrex, enumeration constants, and wrapping #define's
Message-ID: <b6soip$okk$1@bob.news.rcn.net>

Harbison and Steele ( 4th edition, Section 5.5) say that some C 
compilers will accept an enum where some constant expression is written 
in terms of a previously defined enumeration constant:

     enum xx {
         a = 2,
         b = a + 3
     }

Does Pyrex accept the following?

     enum
         a = 2
         b = a + 3

If not, how should I wrap this?

     #define IPL_DEPTH_SIGN 0x80000000
     #define IPL_DEPTH_8S  (IPL_DEPTH_SIGN| 8)




From tim.one at comcast.net  Sun Apr 27 17:35:36 2003
From: tim.one at comcast.net (Tim Peters)
Date: Sun, 27 Apr 2003 17:35:36 -0400
Subject: In which versions is list.sort stable?
In-Reply-To: <Pine.GSO.4.51.0304272112030.17285@techunix.technion.ac.il>
Message-ID: <LNBBLJKPBEHFEDALKOLCIEEPEEAB.tim.one@comcast.net>

[Beni Cherniavsky]
> Why don't we just grow a `list.stablesort` method that can be an alias
> to `list.sort` currently but is guaranteed to be some stable sort
> forever?  Explicit is better than implicit, so in the rare case that
> you do rely on it, why not spell it out?

Because Guido has shown no interest in saying that Python-the-language
guarantees to provide a stable sorting method.  Now that 2.3 has a stable
sort, I expect many apps will find that it's still more efficient to force
stability via injecting sequence indices into sort tuples.  That is,

    temp = [(obj.key(), i, obj) for i, obj in enumerate(list_of_objects)]
    temp.sort()
    result = [obj for key, i, obj in temp]

will often be faster than

    temp = [(obj.key(), obj) for obj in list_of_objects]
    temp.sort()
    result = [obj for key, obj in temp]

The hangup is that tuple comparison breaks ties in the first element
position by going on to compare more element positions.  In the latter
spelling, that can invoke arbitrarily expensive object comparison, but in
the former spelling is always resolved no latter than by comparison of
unequal ints.

This became painfully clear during timing tests run while the 2.3 sort was
being developed, on a database-sorting example contributed by an alpha
tester, where primary keys where often equal.  Forcing stability in that
test ran twice as fast as when leaving the sequence indices out of the
tuples.

stability-is-the-answer-to-fewer-questions-than-are-asked-ly y'rs  - tim





From shao at cia.com.au  Tue Apr  1 01:11:42 2003
From: shao at cia.com.au (Shao Zhang)
Date: Tue, 01 Apr 2003 16:11:42 +1000
Subject: converting from shell script to python
Message-ID: <kfn0jaoig1.fsf@shaoz.activesky.com.au>

Hi,

I need to convert a very simple shell script to python in order get it
run on both windows and unix. However I cannot quite work out a simple
way of doing it without using lists/array/join etc...

#!/bin/sh
ROOTDIR=/usr/lib
CLASSPATH="$ROOTDIR/a/a.jar:$ROOTDIR/b/b.jar:$ROOTDIR/c/c.jar"
java -classpath $CLASSPATH foo

What is the most python way of doing the above? Will perl be a better
choice for things like this?

Thanks.

Shao.



From matthew at newsgroups.com  Fri Apr 18 19:50:56 2003
From: matthew at newsgroups.com (matthew)
Date: Sat, 19 Apr 2003 11:50:56 +1200
Subject: hexadecimal calculations
In-Reply-To: <b7q0uv$hco$1@aphrodite.grec.isp.9tel.net>
References: <b7q0uv$hco$1@aphrodite.grec.isp.9tel.net>
Message-ID: <b7q2p6$8rc$1@lust.ihug.co.nz>

Andros wrote:
> Hello,
> how to make these hexadecimal calculations in python?
> A0 XOR 79 = D9
> D9 - A8   = 31
> 
> Thanks ;)
>
 >>> hex(0xa0 ^ 0x79)
'0xd9'
 >>> hex(0xd9 - 0xa8)
'0x31'
 >>>




From aleax at aleax.it  Mon Apr 14 03:36:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 07:36:39 GMT
Subject: Question regarding indirect function calls
References: <pan.2003.04.13.20.13.21.504295@maxwell.com> <pan.2003.04.14.03.01.20.731919@maxwell.com>
Message-ID: <bytma.13796$T34.444782@news2.tin.it>

Maxwell Hammer wrote:

> In reply to: Jack & Alex:
> 
> Thanks for your advice, I did get it working using "getattr" as you
> suggested.
> 
> Great! now I have another questions, since I have your attention :-)
> 
> What is the most efficient way of accessing global variables across
> modules?

Just do so _explicitly_.


> I have main.py which parses a config file and sets up config values
> appropriately. Different values are required by different functions in
> other modules.
> 
> I have tried defining the config values as global in main, AND in the
> other modules - but does not work (get undefine errors).

Each module's globals belong to that module only -- other modules
can access them explicitly.

> main.py:
>   global config_values
>   config_values = []
> ...
>  load config_values
> ...
> 
> Module x
>  global config_values
> 
> def function1():
>   try to access config_values[0] fails - not found

Change to:
module x:

import main
def function1()
    access main.config_values[0]


> So what I am currntly doing is passing what ever config values are needed
> through the function call
>   x.function1(config_values, etc)

This is quite a good approach, too -- even more explicit.

Yet another alternative is to use "from main import config_values" in
module x, but I do NOT recommend this -- just "import main" and explicit
use of "main.config_values" is FAR better, both stylistically and in
terms of functionality.  For example, with the "from" statement, you
may become vulnerable to the case in which module x is loaded before
(or worse _during_...) module main -- or to situations where at some
later point module main rebinds its config_values name to a different
value -- and so on.  The "import" statement, with the explicit use of
"main.config_values", is more resilient, as well as more explicit and
thus clearer and more readable.


> How would you do this??

Probably with "import main" and using main.config_values, although
explicit arguments can also be an excellent alternative.


Alex






From puccio_13 at yahoo.it  Wed Apr 23 05:04:01 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Wed, 23 Apr 2003 11:04:01 +0200
Subject: python lexer
Message-ID: <8Wspa.23568$DN.669072@tornado.fastwebnet.it>

Hi, I was wondering which is a good lexer written in Python, I heard about
SPARK and PLY. I need only to tokenize texts, not to build trees on their
structure. 
string.split() I guess it is not enough, because I can't use reg exp as
token separators..How could I do otherwise?
Thanks

-- 
bye
Alessio Pace



From mh at pixar.com  Mon Apr  7 19:47:59 2003
From: mh at pixar.com (Mark Harrison)
Date: Mon, 07 Apr 2003 23:47:59 GMT
Subject: exceptions == errors?
Message-ID: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>

Are exceptions always considered errors, or there the same feeling
as in C++ that exceptions can be used for non-error situations?

I am asking because I am writing a network server that has the
general outline

	while 1:
		read request
		process request
		send response

So, I have the option to

1.  embed the "send response" code in the request processing code

2.  Have the request processing code generate formatted responses
    that are returned to the main loop

As is usually the case, adding robust error checking and reporting
throws a kink into the works.   I have to do a lot of exception
wrapping in the request processing code in order to ensure that
there are no code paths where an exception short-circuits a
response generations.

Therefore, I'm thinking of structuring the code like this:

	while 1:
		try:
			read request
			process request
		catch MyException,e:
			send response contained in e
		catch Exception,e:
			send "internal error" response

Am I setting myself up for trouble if I follow this route?
All advice gratefully received.

-- 
Mark Harrison
Pixar Animation Studios



From mcfletch at rogers.com  Tue Apr 15 21:43:45 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 15 Apr 2003 21:43:45 -0400
Subject: Big-O notation
In-Reply-To: <yu99y92bh10b.fsf@europa.research.att.com>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com>
Message-ID: <3E9CB551.4030006@rogers.com>

Andrew Koenig wrote:

>Mike>     Characterisation of an algorithm's (as distinct from an
>Mike>     implementation's) performance in relation to the size of the input
>Mike>     set.  Normally of interest in determining the
>Mike>     scalability/suitability of a particular algorithm.  The static (e.g.
>Mike>     setup) and proportional (e.g. implementation) factors in the
>Mike>     algorithm are factored out under the assumption that they are
>Mike>     swamped by the relative effects of the size of the input set as that
>Mike>     goes to infinity.
>
>Also, big-O notation always refers to an upper bound, not an estimate
>that might be high or low.  So, for example, every quantity that is
>O(N) is also O(N log N), but the reverse is not true.
>
I seldom seem to see people invoking that part of the definition, though 
it's definitely it.

Mostly people seem to talk about the best, average and worst-case 
performance as if the O notation was relevant in each case, though IIUC 
the worst-case would be the "real" O description.  That is, the O 
notations seem to be commonly used as shorthand notation for "constant 
running time", "linear running time", "quadratic running time" 
etceteras.  Bad people, bad :) .

>Mike> Common examples of big-O performance types:
>
>Mike>     * O(1) -- time is linear, that is, it takes the same amount of time
>Mike>       to do something regardless of how large the input set is.
>
>O(1) doesn't mean linear, it means bounded by a constant.  That is,
>if the running time of an algorithm is O(1), it means that there exists
>a time interval t such that the algorithm always finishes in time < t.
>  
>
Yup, I'd meant to write constant and brain-fog occluded the shore

...

>You can also think of O(log N) as being no worse than proportional to
>the number of digits in N.
>
Good way of describing it.
...

>Mike> re-factor the equations to get a final big-O notation, but for
>Mike> the most part, practical coders can get away with recognising
>Mike> and avoiding O(N**X) where X > 1 situations.  Recognising the
>  
>
...

>O(N**1.5) is often bearable.
>

Good point.
...

Okay, so now we have "The Practical Coder's Guide to big-O Notation" :) 
.  We rock ;) ,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From st at ferd2.thristian.org  Thu Apr 24 11:19:40 2003
From: st at ferd2.thristian.org (Screwtape)
Date: Fri, 25 Apr 2003 01:19:40 +1000
Subject: Vim Newbie =?iso-8859-1?Q?=96?= Python Edit/Compile/Run cycle and integrating py help questions
References: <a791ab19.0304240532.5a025b81@posting.google.com>
Message-ID: <slrnbag04b.rnv.st@ferd2.thristian.org>

m_init(): spawning followupTo('Daniel K')...done.
>I can't seem to find how to ask Vim to "compile" the current buffer
>the way I could in emacs (using c-c, c-c.)  What I'm hoping for is to
>run the current buffer, show the output in a split window
>(horizontal?) and move the cursor in the buffer to the first error. 
>Is there a Vim way of doing things?

Eh, I normally just run the app in a seperate terminal.

If you really want to get clever about it, though, have a look into
the 'makeprg' and 'errorformat' settings (type ":help 'makeprg'" or
":help 'errorformat'").

A little research reveals:

	http://vim.sourceforge.net/scripts/script.php?script_id=477

>Is there any way to integrate some python language reference to be
>accessed from within Vim?  Something like hot-key keyword lookup to
>the docs.

Again, I normally have the Python library reference open in Mozilla -
tabbed-browsing is great for this. The Python language is simple
enough that you don't have to keep looking things up, and I'd hate to
try to teach Vim about all the documentation locations of all the
modules I ever use.

>Is asking for keyword completion in Vim going against the grain of
>what Vim stands for?  (Simplicity)

These days, I can't live without

	http://vim.sourceforge.net/tips/tip.php?tip_id=102

>Any suggestions or pointers to where I can find these answers would be
>greatly appreciated.  I'm really enjoying the power of Vim!!

Search for "python" in both the 'scripts' and 'tips' sections here:

	http://vim.sourceforge.net/search.php

My favourites:

	http://vim.sourceforge.net/scripts/script.php?script_id=515
	http://vim.sourceforge.net/scripts/script.php?script_id=386
	http://vim.sourceforge.net/scripts/script.php?script_id=273

-- 
 ___________ ____________________________
| Screwtape | Reply-To: munged on Usenet |________ ______ ____ __ _  _   _
|
| My rabbit is, decigram for decigram, the cutest creature on earth.
|



From claird at lairds.com  Wed Apr 16 07:22:57 2003
From: claird at lairds.com (Cameron Laird)
Date: Wed, 16 Apr 2003 11:22:57 -0000
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net> <mailman.1050437489.6762.python-list@python.org>
Message-ID: <v9qf8hkb024ab0@corp.supernews.com>

In article <mailman.1050437489.6762.python-list at python.org>,
Scott Chapman  <scott_list at mischko.com> wrote:
			.
			.
			.
>Anyone else have good recommends regarding OOP books, especially focusing on 
>Python?
			.
			.
			.
<URL: http://mindview.net/Books/TIPython >
<URL: http://greenteapress.com/thinkpython.html >
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From dj3000 at iinet.net.au  Mon Apr  7 00:32:41 2003
From: dj3000 at iinet.net.au (DJ_3000)
Date: Mon, 7 Apr 2003 14:32:41 +1000
Subject: cgi form handler question
Message-ID: <3e90ff67$0$13167$5a62ac22@freenews.iinet.net.au>

Hi,

im writing a form cgi handler that handles data inputted by a user.

I would like to know how do I output the name of the field that the user
forgot to input.

eg I have the fields:

name
email
address
age
salary


the user forgets to input his email and age
how would i write a for loop to output:

email
age

for name in form.keys():
        print name

would output the field names that the user did inpuit.. i want the opposite
of this ta!

Thanks!





From mis6 at pitt.edu  Mon Apr 21 10:32:37 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 21 Apr 2003 07:32:37 -0700
Subject: Is this a super bug?
References: <mailman.1050791852.9118.python-list@python.org>
Message-ID: <2259b0e2.0304210632.53e73b21@posting.google.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1050791852.9118.python-list at python.org>...
> M:\python>python
> Python 2.3a2 (#39, Feb 19 2003, 17:58:58) [MSC v.1200 32 bit (Intel)] on
> win32
> Type "help", "copyright", "credits" or "license" for more information.
> >>> class ff(str):
> ...    def foo(self):
> ...      print 'str.  getitem  ', str.  getitem  (self, 0)
> ...      print 'super(ff, self).  getitem  ', super(ff,
> self).  getitem  (0)
> ...      print 'super(ff, self)[0]', super(ff, self)[0]
> ...
> >>> f = ff('asdf')
> >>> f.foo()
> str.  getitem   a
> super(ff, self).  getitem   a
> super(ff, self)[0]
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "<stdin>", line 5, in foo
> TypeError: unsubscriptable object
> 

This is not a bug, since super(cls,self) is supposed to return a super
object, and not self. Nevertheless super has various bugs: it cannot retrieve
properties, cannot retrieve __name__, does not work always properly
in multiple inheritance of a class and a metaclass, etc. If you google a bit,
you will find various posts of mine about problems with super.

At the end I got really tired, started studying attribute descriptors
(completely undocumented, AFAIK), spent a lot of time and come with the
following code.

Warning: this is quite non-trivial code and I don't claim I have tested it
in any possible situation.

def ancestor(C,S=None):
  """Returns the ancestors of the first argument with respect to the 
  MRO of the second argument. If the second argument is None, then 
  returns the MRO of the first argument."""
  if C is object:
      raise TypeError("There is no superclass of object")
  elif S is None or S is C:
      return list(C.__mro__)
  elif issubclass(S,C): # typical case
      mro=list(S.__mro__)
      return mro[mro.index(C):] # compute the ancestors from the MRO of S
  else:
      raise TypeError("S must be a subclass of C")

class convert2descriptor(object): # used in _super
  """To all practical means, this class acts as a function that, given an
  object, adds to it a __get__ method if it is not already there. The 
  added __get__ method is trivial and simply returns the original object, 
  independently from obj and cls."""
  def __new__(cls,a):
      if hasattr(a,"__get__"): # do nothing
          return a # a is already a descriptor
      else: # creates a trivial attribute descriptor
          cls.a=a
          return object.__new__(cls,a)
  def __get__(self,obj,cls=None):
      "Returns self.a independently from obj and cls"
      return self.a

def _super(C,S,methname):
  """Internal function invoking ancestor. Returns an attribute descriptor 
  object."""
  if methname=='__name__': # special case
      meth=ancestor(C,S)[1].__name__
  else:
      for c in ancestor(C,S)[1:]:
          meth=c.__dict__.get(methname)
          if meth: break # if found
  if not meth: raise AttributeError,methname # not found 
  return convert2descriptor(meth) # if needed

class Super(object):
  """Invoked as Super(cls,obj).meth returns the supermethod of
  cls with respect to the MRO of obj. If obj is a subclass of cls,
  it returns the unbound supermethod of obj; otherwise, if obj is an
  instance of some subclass of cls, it returns the obj-bound method."""
  def __init__(self,cls,obj):
      self.cls=cls
      self.obj=obj
  def __getattribute__(self,name): # how to redefine __getattribute__ properly
      obj=object.__getattribute__(self,'obj')
      cls=object.__getattribute__(self,'cls')
      if hasattr(obj,'__bases__') and issubclass(obj,cls):
          # if obj is a subclass, return unbound method
          return _super(cls,obj,name).__get__(None,obj)
      else: # if obj is an instance, return bound method
          S=type(obj)
          return _super(cls,S,name).__get__(obj,S)

# few checks

class ExampleBaseClass(object):
  """Contains a regular method 'm', a staticmethod 's', a classmethod 
  'c', a property 'p' and a data attribute 'd'."""
  m=lambda self: 'regular method'
  s=staticmethod(lambda : 'staticmethod')
  c=classmethod(lambda cls: 'classmethod')
  p=property(lambda self: 'property')
  d='data'

class C(ExampleBaseClass): pass
c=C()
print Super(C,C).p #<property object>
print Super(C,c).p #property
print Super(C,c).__name__ # ExampleBaseClasss
print Super(C,C).__name__ # ExampleBaseClass



From ajsiegel at optonline.net  Tue Apr 22 18:35:58 2003
From: ajsiegel at optonline.net (Arthur)
Date: Tue, 22 Apr 2003 18:35:58 -0400
Subject: instance creatiion as an "event"
Message-ID: <000a01c3091f$8bd45350$0c02a8c0@Arts>

Alex writes:

>Not with the once-habitual Python idiom for that purpose. Try it!

>class TestPoint:
>   def __init__(self):
>      self.tp = 23
>      self.redraw()
>   def redraw(self):
>     print 'TestPoint.redraw', self.tp

>class PointChild(TestPoint):
>   def __init__(self):
>      self.pc = 42
>      TestPoint.__init__(self)
>   def redraw(self):
>     print 'PointChild.redraw', self.tp, self.pc

>p = PointChild()

Indeed.  The behavior I intuitively expected is indeed what one get's and I
was off somewhere in la-la land misundertanding my error messages. Deja vu,
again.

And as it turns out,  a little experimentation indicates more thought as to
when in sub's __init__ its parent's __init__ is called might indeed be the
full answer to my issue.

A little simpler than I had hoped ;)

Art





From peter at engcorp.com  Wed Apr  2 14:51:26 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 14:51:26 -0500
Subject: OT: top-posting and the blind (was Re: need help with game)
References: <3e88b3d3.30570106@news-server> <3E88C1D8.3070305@motorola.com> <3e8a57d0.55986091@news-server> <3E8B1602.72BB1A37@juno.com>
Message-ID: <3E8B3F3E.ADA82191@engcorp.com>

Jeff Sandys wrote:
> 
> Re: top-posting: Blind people with text to speech converters
> prefer top-posting so they don't have to listen to the same
> stuff over and over again.

Presumably blind people are not reading their own postings that
often, so unless you are replying to a blind person with a text
to speech converter (and one who doesn't match the profile 
posted by Robin on the issue) then that doesn't make it a good
idea for you to do it.

If *you* are blind and find it easier for us to top-post in our
replies to you, then we would likely readily accommodate that 
when talking to you.

-Peter



From staschuk at telusplanet.net  Wed Apr 16 17:54:38 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 15:54:38 -0600
Subject: Big-O notation
In-Reply-To: <a8b7f07a.0304161158.29c0ef6b@posting.google.com>; from alloydflanagan@attbi.com on Wed, Apr 16, 2003 at 12:58:30PM -0700
References: <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <a8b7f07a.0304161158.29c0ef6b@posting.google.com>
Message-ID: <20030416155438.H4266@tibia.amotlpaa.bogus>

Quoth A. Lloyd Flanagan:
  [...]
> This mathematical analysis of an algorithm's behavior (and best-case,
> average, and worst-case performance) can be trivial.  For example, x +
> 1 is O(1).  [...]

I would have thought it's O(log x), due to the possibility of
carries.

-- 
Steven Taschuk                                        staschuk at telusplanet.net
"Study this book; read a word then ponder on it.  If you interpret the meaning
 loosely you will mistake the Way."         -- Musashi, _A Book of Five Rings_




From carsten at sarum.dk  Tue Apr 29 03:32:24 2003
From: carsten at sarum.dk (Carsten Gehling)
Date: 29 Apr 2003 00:32:24 -0700
Subject: Function to convert fetchone() result from a list to a dictionary
Message-ID: <c31eab9b.0304282332.32799f80@posting.google.com>

I'm fairly new to Python, and have yet to discover alle the little
details, that may make my programs look leaner and perhaps become
faster...

Using the DB API to access databases, I've stumbled upon what I would
call a shortcoming in the API (IMHO). All the fetch commands return
the results as lists instead of dictionaries.

So I've made my own conversion function, using the cursor.description
attribute. It's pretty straight-forward, so I was wondering, if
there's a smarter way to do this?

The function is like this:
-------------------------------
def cursor_fetchone_dict(cursor):
	result = cursor.fetchone()
	if result == None: return None
	result_dict = {}
	for i in range(len(result)):
		result_dict[cursor.description[i][0]] = result[i]
	return result_dict
-------------------------------

Example of use:
-------------------------------
import MySQLdb

def cursor_fetchone_dict(cursor):
	result = cursor.fetchone()
	if result == None: return None
	result_dict = {}
	for i in range(len(result)):
		result_dict[cursor.description[i][0]] = result[i]
	return result_dict

con = MySQLdb.connect(host='localhost', port=3306, user='xxxxx',
passwd='xxxxxx', db='xxxxxx')

cursor = con.cursor()
sql = "select * from sometable "
cursor.execute(sql)

result = cursor_fetchone_dict(cursor)
print result
-------------------------------

Any comments are welcome, TIA

- Carsten



From mads at troest.NEVERMORE.dk  Sun Apr 13 15:26:37 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Sun, 13 Apr 2003 21:26:37 +0200
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <18lb3o2cs1se2$.isoixa5igmj6$.dlg@40tude.net>
Message-ID: <12aeh974i1kdh.dfbwdn5e18px.dlg@40tude.net>

On Sun, 13 Apr 2003 21:24:24 +0200, Mads Orbesen Troest wrote:

>> Or may be it is possible to emulated Protected Methods in Python.
> 
> It can be "emulated" by prepending the method/attribute name with 2 
> underscores. 

Whoops, I was too quick and didn't see that you wrote "protected" and not 
"private".

At any rate: the mangling scheme can be used for "protected" methods, since 
you know the mangled name is _classname__attribute/method. You can then be 
one of the "evil" persons I talked about in the previous post, and access 
the "private" member that way.


 /\/\\ads Orbesen Troest



From patrick.vrijlandt at tip.nl  Tue Apr  8 16:34:57 2003
From: patrick.vrijlandt at tip.nl (Patrick Vrijlandt)
Date: Tue, 8 Apr 2003 22:34:57 +0200
Subject: COM and MS Word
References: <2eb85cfb.0304070115.6aecb644@posting.google.com>
Message-ID: <b74j7h$3ui$1@reader1.tiscali.nl>

Did you try to record a macro while you were doing manually once what you
plan to do?
Then of course, all you would do is change VB to Py ...

--
Patrick Vrijlandt

"CTony" <anthony.evershed at npl.co.uk> schreef in bericht
news:2eb85cfb.0304070115.6aecb644 at posting.google.com...
> It appears that some data files I need are in WordPerfect (.dat)
> format, and I need to get them into Python for processing and later
> entry into a database. I haven't been able to find a way of
> interfacing to WordPerfect directly, but MS Word 2000 can read these
> files...
>
> Here's the snag, though. All the examples of Python/COM/Word usage
> I've seen have involved pushing data into Word, rather than extracting
> data *from* it. I know that the data field I need is x lines below a
> known heading, so getting to the field is no problem:
>
> finder = wd.Selection.Find
> finder.Text = 'foo'
> finder.Execute()
> for n in range(x):
>     wd.Selection.MoveDown
> wd.Selection.HomeKey
>
> get you to the beginning of the field, no problems. How, however, can
> I select (block mark) and extract that line of text?
>
> If it helps, I'm using Python 2.2, with win32all build 148, on Windows
> NT 4 (though the resulting program will be used on Windows 2000 as
> well)...
>
> Many thanks,
>
> Tony Evershed





From maciej at maciejsobczak.com  Fri Apr  4 04:10:44 2003
From: maciej at maciejsobczak.com (Maciej Sobczak)
Date: Fri, 4 Apr 2003 11:10:44 +0200
Subject: Distributed environment
References: <3e8d1edc$0$235@hades.is.co.za>
Message-ID: <b6ji3b$t4c$1@SunSITE.icm.edu.pl>

Hi,

"Gerrit van Dyk" <gerritvd at decillion.net> wrote in message
news:3e8d1edc$0$235 at hades.is.co.za...

> I am in the process of designing a system that will need to use a
> distributed environment.

Please try:
http://www.maciejsobczak.com/prog/yami/

It is very easy to use this infrastructure in Python (for example, native
Python lists are used to compose messages) and it also supports other
languages (C, C++ and Tcl).

> Several different services will be registered with the central server to
> supply different types of services (rate feeds, risk analysis, forecasting
> etc).

Take a look at the Event Service -- it allows to distribute messages accross
many services in the subscription model. Other models (including load
balancing over many servers) are easy to implement with the use of messag
forwarding.
Do not hesitate to contact me in case of any further questions.

> The system
> needs to be up 24/7/52.

:)

> Is this feasable?

Yes, I think.

--
Maciej Sobczak
http://www.maciejsobczak.com/

Distributed programming lib for C, C++, Python & Tcl:
http://www.maciejsobczak.com/prog/yami/





From cappy2112 at yahoo.com  Sun Apr 13 03:22:02 2003
From: cappy2112 at yahoo.com (Tony C)
Date: 13 Apr 2003 00:22:02 -0700
Subject: Problems installing PIL 1.1.3 on Windows 2000
Message-ID: <8d3e714e.0304122322.213b9407@posting.google.com>

When I run the installer for PIL 1.1.3 on Win2K, the screen which is
supposed to allow me to select which Python distribution to use, is
BLANK, and the installer just hangs there. The NExt button is
disabled, but I can click CANCEL.

I only have Py 2.2.2 on my system, I don't know why this program has a
problem with this. I have installed amny Python packages already,
without any problems at all.


I've already run the register script, on the download page, and it
tells me Py20 is already registered.

Is there a fix/workaround for this problem ?




thanks



Tony



From brian at rk-speed-rugby.dk  Tue Apr 22 04:13:47 2003
From: brian at rk-speed-rugby.dk (Brian Elmegaard)
Date: 22 Apr 2003 10:13:47 +0200
Subject: pickling a class definition
Message-ID: <uy923x83o.fsf@mail.afm.dtu.dk>

Hi,

I am pickling instances of a class, but in the unpickling script it
seems that I have to know how the class was defined. I cannot
pickle the class definition in one script and unpickle in another. 

Is that correct?


-- 
Brian (remove the sport for mail)
http://www.et.dtu.dk/staff/be



From staschuk at telusplanet.net  Sun Apr 27 20:37:28 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 27 Apr 2003 18:37:28 -0600
Subject: Forgetting "()" when calling methods
In-Reply-To: <b8ekv0$o3e$1@panix2.panix.com>; from aahz@pythoncraft.com on Sat, Apr 26, 2003 at 02:59:44PM -0400
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <mailman.1051376980.15852.python-list@python.org> <b8ekv0$o3e$1@panix2.panix.com>
Message-ID: <20030427183728.B28461@tibia.amotlpaa.bogus>

Quoth Aahz:
> In article <mailman.1051376980.15852.python-list at python.org>,
> Steven Taschuk  <staschuk at telusplanet.net> wrote:
  [...]
> >It would be nice to be able to point people at a good on-line
> >debugging tutorial for beginners; anybody know of such a thing?)
> 
> Doing a Google for "debugging tutorial" pulls this as the top link:
> http://heather.cs.ucdavis.edu/~matloff/debug.html

Yes, that's not bad.  A little too tool-centric for my tastes, but
it does give good general concepts (e.g., "Finding your bug is a
process of confirming the many things that you believe are true --
until you find one which is *not* true.") and excellent examples
for the language used, C.

Of course, if I'd wanted to know what Google thinks, I'd have
asked Google.  Perhaps you were wondering why I didn't; I'm hoping
that one of the experienced teachers who posts here might make a
recommendation backed by their experience.  That would carry much
more weight (with me, at least) than Google's page rank.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From martin at v.loewis.de  Sun Apr 13 12:16:15 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 18:16:15 +0200
Subject: How to subclass datetime?
References: <ba1e306f.0304130339.3bece039@posting.google.com>
Message-ID: <m3d6jqpe40.fsf@mira.informatik.hu-berlin.de>

janeaustine50 at hotmail.com (Jane Austine) writes:

> I'm interested in the new datetime module and I want to subclass
> date/datetime class to modify the behaviour. However, simply
> inheritting the classes(like "class mydate(date):" ) didn't work.

What do you mean "didn't work"? It works fine for me:

>>> import datetime
>>> class Tag(datetime.date):pass
...

Regards,
Martin



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Wed Apr 30 07:29:53 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Wed, 30 Apr 2003 13:29:53 +0200
Subject: bug in Jython 2.1? VerifyError
Message-ID: <3eafb3b0$0$49112$e4fe514c@news.xs4all.nl>

I've run into some trouble using Jython 2.1
(jdk 1.4, 1.3, windows, linux).

The following piece of Python code:
----
def method():
     try:
         for dummy in [1,2,3]:
             try:
                 return "result"
             except:
                 pass
     finally:
         pass

print method()
----

fails to compile in Jython, I get this error:

java.lang.VerifyError: java.lang.VerifyError: (class: test$py, method: 
method$1 signature: (Lorg/python/core/PyFrame;)Lorg/python/core/PyObject;) 
Register 3 contains wrong type


That piece of code above seems to be the minimal piece of code to trigger
the problem. Something to do with nested try blocks and a loop?

Can someone confirm this or help me? Thanks!

--Irmen de Jong.




From peter at engcorp.com  Wed Apr  9 15:59:55 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 09 Apr 2003 15:59:55 -0400
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com> <3E935B75.64968B9@engcorp.com> <b71s3u$lf3$01$1@news.t-online.com>
Message-ID: <3E947BBB.48B85048@engcorp.com>

Heiko Henkelmann wrote:
> Peter Hansen wrote:
> > Heiko Henkelmann wrote:
> > > What is the most efficient way to do this?
> >
> > Define "efficient" ...
>
> to be more precise: The fastest way...

I'd bank on re.findall in this case, but I think you'd have to time
each approach to be sure.  Just avoid the lambda cases....
Also note that the re.findall method as given probably does not return
a lone final character as a result, while some of the other approaches
probably do.  Presumably, since you didn't specify it, you are 
planning to ensure the input always splits evenly.

-Peter



From linux42 at freemail.c3.hu  Wed Apr 16 09:35:00 2003
From: linux42 at freemail.c3.hu (Gabor Nagy)
Date: Wed, 16 Apr 2003 15:35:00 +0200
Subject: variable visiblity
Message-ID: <20030416133500.GB2518@swordfish.capgemini.hu>

Hi!

How can I access the "parent" function's local variable from a nested
function?

example:
def foo(par):
	def f(x): print x, par
	f(0)

SyntaxWarning: local name 'par' in 'foo' shadows use of 'par' as global in
nested scope 'f'
when run, prints x, then
...in f NameError: global name 'par' is not defined

Why: I'd like to use filter( f, mysequence) in foo.
f can only have one parameter, the sequence element. I need to pass it
something else to compare to this would be par.

Right now I do things like:
filter( f, map( None, mysequence, [par] * len(mysequence)))
but I feel that this is quite hard to read, and is not a "good" way to do
the thing.

I can use a non-nested f, and it works. I would like to make it nested, if
possible.

So, as you see, I can accept if the solution is something different from
what I had in my mind, I simply want it simple, and quick.

Regards,
Gee




From godoy at metalab.unc.edu  Tue Apr  8 17:58:57 2003
From: godoy at metalab.unc.edu (Jorge Godoy)
Date: Tue, 08 Apr 2003 18:58:57 -0300
Subject: Best way to make a weighted sum
References: <867ka46awj.fsf@ieee.org> <mailman.1049833916.16117.python-list@python.org>
Message-ID: <86vfxo3b32.fsf@ieee.org>

Steven Taschuk <staschuk at telusplanet.net> writes:

> I don't know why you're using a class for this.  A simple function
> which calculates (and/or verifies) the checksum seems more natural.

The class has several other things. As I said, this was a snippet of
code. It showed you the __init__ and the other methods I was
interested in discussing.

Your approach made me think again on some other implementations I had
in mind before.

> And, despite the class name, you don't actually do a mod 11.  I
> assume this is a bug.

You haven't seen the rest of the class. This is not a bug.

The line that does the mod 11 is this one, in case you want to see it:


        self.modulo = self.soma % 11


Note that self.soma had other operations before being submitted
here. Some of our States do really weird things before the modulus
calculation. And some do these weird stuff after that.

>>     def __init__(self, ie=None, uf=None):
>>         
>>         self.estados = ['CE', 'AP', 'RJ'] 
>>         self.ie = ie
>>         self.soma = 0
>
> Why isn't the uf argument used?

I cut it out with the portuguese comments by mistake. Sorry.

      self.uf = uf

It is used later. 

> If uf is not in ['CE', 'AP', 'RJ'], then self.pesos simply does not
> get set; this will cause CalculaSoma to fail later, in a mysterious
> way.  Better to have an informative exception raised at this point,
> surely.

It will have all valid States. I wouldn't clutter the message with 27
States information and lots of different weights just to check an
approach :-) You could send me a better code without it :-)

Also there will be no exception since this is also bound to a domain
at the database server.

> It's also more natural imho to have the weights in a dict keyed by
> uf.  That dict can replace your self.estados, which I note is not
> otherwise used.

It is used later, by other methods that were not part of this part of
the code that I was interested in. But thanks for you comments. But,
on the other hand, using this dict keyed by uf approach seems
interesting. Makes the code cleaner.

> Also, the weights are conceptually a sequence, not a map; why not
> make them a list?

They are bound to the number position. Just that. A list would do fine
here. There's one state where they are: 1, 3, 4, 5, 6, 7, 8, 10. And
there are other variants of it too. You saw one in my original code,
where you asked if it was right starting with 2 and the going from 7
to 2 again... For some States I have to do that only once, for other
there are different rules for different categories of the
company... It's a mess.

>>     def CalculaSoma(self):
>> 
>>         ie = self.ie
>>         iter = len(ie)
>> 
>>         for i in ie:
>>             produto = i * self.pesos[iter]
>>             self.soma = self.soma + produto
>>             iter = iter - 1
>
> Why not return the sum?  See above.
>
> It's also not good form to use 'iter' as a variable name, since
> 'iter' is the name of a built-in function.

Oops. The full name is 'iteracao'. I changed it so that it would be
easier for you to understand what's it used for. I should have sent
the program without removing comments and without changing this name. 

I'm sorry for changing the code I posted. 

>> What I want is to sum the products of a certain 'ie' number where each
>> of its digits is multiplied by a certain weigth, depending on its
>> position inside the number. (This is for some documents verification
>> and 'ie' can vary in the amount of digits as well as its weigth can
>> vary from one place to another --- I know this is a mess, but it's the
>> way some Brazilian states did it...)
>
> An alternative:
>
> 	import operator
>
> 	weights = {}
> 	weights['CE'] = weights['AP'] = [9, 8, 7, 6, 5, 4, 3, 2]
> 	weights['RJ'] = [2, 7, 6, 5, 4, 3, 2] # is this right?

(weird isn't it? I also thought it was wrong but that's what's stated
at Rio de Janeiro's Treasury Secretary website:
http://www.sef.rj.gov.br/sub_adj_trib/sucief/calc_ie/index.shtml) 

> 	def checksum(ie, state):
> 		w = weights[state][-len(ie):]
> 		return reduce(operator.add, [x*y for x, y in zip(ie, w)]) % 11
>
> Imho this is much simpler, more straightforward, and easier to use.

Yes, it is. I didn't know zip(). It indeed facilitates a lot. I just
have to think, now, on how to add the several modifiers to this
checksum :-) It will be easy if there's no new weirdness here... 

The list approach also is interesting. I just have to think on how it
adapts to the cases where there are more possibilities in the same
State. 


You enlightened me a lot on this message, thanks.

-- 
Godoy.      <godoy at metalab.unc.edu>

There is a green, multi-legged creature crawling on your shoulder.



From andrew-pythonlist at puzzling.org  Thu Apr 17 04:14:16 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Thu, 17 Apr 2003 18:14:16 +1000
Subject: "pythonic"
In-Reply-To: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
Message-ID: <20030417081416.GB21962@frobozz.local>

On Thu, Apr 17, 2003 at 08:00:27AM +0000, Afanasiy wrote:
> Too many times now I have seen the term "pythonic" used to vaguely
> describe something which is never explicitly defined, as if it cannot be.
> 
> Now it seems I can just replace "pythonic" with "my idea of utopia".
> 
> I do not know if this ignorance, laziness, or a meme, but I would like
> to know if there a real definition of "pythonic". Anyone?

import this

-Andrew.





From grahamd at dscpl.com.au  Sat Apr  5 19:46:30 2003
From: grahamd at dscpl.com.au (Graham Dumpleton)
Date: 5 Apr 2003 16:46:30 -0800
Subject: Distributed environment
References: <3e8d1edc$0$235@hades.is.co.za>
Message-ID: <dc6f5c99.0304051646.5c97809e@posting.google.com>

Gerrit van Dyk <gerritvd at decillion.net> wrote in message news:<3e8d1edc$0$235 at hades.is.co.za>...
> Hi all,
> 
> I am in the process of designing a system that will need to use a 
> distributed environment. 
> 
> The system will consist of multiple (thin) clients += 100-200 that will need 
> to connect to a central server. The client applications is not 
> browser-based as they need to do a lot of data manipulation, draw charts, 
> run scenarios etc. The idea is to do most of the processing on servers 
> except for the GUI stuff. 
> 
> Several different services will be registered with the central server to 
> supply different types of services (rate feeds, risk analysis, forecasting 
> etc). 
> 
> The central server(HUB) needs to control all the services and clients. When 
> a request comes in from a client, the HUB needs to distribute this request 
> to the correct service. Multiple services could be registered for the same 
> type of service. This will enable multiple clients to use the same type of 
> service concurrently. The services could also make use of relational 
> databases or other types of data stores. The client applications will have 
> no concept of data storage and will have no direct access to the databases.
> 
> In some cases, some of the services will be needing the services of other 
> services, in essence also becoming clients for the other services.
> 
> The central server (system) needs to keep track of all the clients and their 
> requests. This is to ensure that if a machine or service fails the system 
> should be able to recover without affecting the clients at all. The system 
> needs to be up 24/7/52.
> 
> These services will also be distributed across multiple machines for load 
> balancing.
> 
> My questions are.
> 
> Is this feasable? 
> Any pointers or pitfalls I might step in?
> Which environments would you recommend I should look at, for doing this type 
> of thing?

You might consider looking at OSE (http://ose.sourceforge.net). It has
all the components which you would require for making such a system.
These include a distributed service agent model with presence so you
can query what services exist, when they come into existance and when
they dissappear. Combined with this is a request/reply framework so
that service agents can communicate with each other. Each service agent
also has a means of publishing reports about it state, what it is doing and
for making information generally available. The request/reply framework
is also accessible using gateways from XML-RPC (and SOAP if you really
must). The core of OSE is C++ but provides Python wrappers. You can
write code exclusively in Python, or you can write services in C++ as well.

You can find a 120 page manual for using the Python wrappers for OSE
on the web site, just follow the "Python Manual" link.



From DavidMLewis.NO at SPAM.mac.com  Tue Apr 15 19:31:00 2003
From: DavidMLewis.NO at SPAM.mac.com (David Lewis)
Date: Tue, 15 Apr 2003 23:31:00 GMT
Subject: Printing Columns
References: <mailman.1050426466.11198.python-list@python.org> <150420031341416871%DavidMLewis.NO@SPAM.mac.com> <3E9C788F.74CDC1FA@engcorp.com>
Message-ID: <150420031631008308%DavidMLewis.NO@SPAM.mac.com>

In article <3E9C788F.74CDC1FA at engcorp.com>, Peter Hansen
<peter at engcorp.com> wrote:

> David Lewis wrote:
> > 
> > This should give you a start:
> > 
> > >>> x = ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h']
> > >>> x[0: :3]
> > ['a', 'd', 'g']
> > >>> x[1: :3]
> > ['b', 'e', 'h']
> > >>> x[2: :3]
> > ['c', 'f']
> 
> This kind of response, requiring *alpha* software that is not yet
> released and unlikely to be used by the majority of people, deserves
> a caveat:
> 
> The above will not work with standard versions of Python prior to 2.3.
> 
You are correct and I apologize. I'd forgotten that extended slice
notation was one of the additions.

Best,
   David



From anton at vredegoor.doge.nl  Fri Apr 18 05:22:21 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Fri, 18 Apr 2003 11:22:21 +0200
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl> <3E9EAA36.BFB1018D@engcorp.com> <IQmdnQg6xagDdQOjXTWcpA@august.net> <b7neqo$knk$1@news.hccnet.nl> <N_Nna.50877$t_2.4682@afrodite.telenet-ops.be>
Message-ID: <b7og5c$h8b$1@news.hccnet.nl>

Roel Schroeven <j4nff5m02 at sneakemail.com> wrote:

>> But why do the DNS-lookup if it's possible to load a page using an ip?
>
>Because not all websites can be reached that way, due to virtual 
>hosting. Take my website for example (don't look at it, it's really not 
>interesting :) ). http://roelschroeven.net gives a normal webpage, while 
>http://64.191.37.195 (which is the ip-address for that domain) gives a 
>generic page (in this case something that shouldn't really be there IMHO).

Thanks a lot for pointing this out. I did the following test to
corroborate your explanation:

>>> socket.gethostbyaddr("64.191.37.195")
('kontiki.ion-web.com', [], ['64.191.37.195'])
>>> socket.gethostbyname("roelschroeven.net")
'64.191.37.195'
>>> socket.gethostbyname("kontiki.ion-web.com")                   
'64.191.37.195'
>>> 

This is quite perplexing. My view of the internet is that it is like
a phonebook where it is possible to link each name to a number. I
forgot that sometimes a lot of people are living on the same address!
Also thanks to Terry for bringing this up, although it took another
post to make me understand what was meant. I guess I am back to the
drawing board now :-|

Anton



From rmunn at pobox.com  Mon Apr 21 18:00:54 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 21 Apr 2003 22:00:54 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com> <b7kvuh$20aem$1@ID-169208.news.dfncis.de> <3EA03AFF.1030309@removethis.free.fr> <5LVna.126512$It5.11343@news2.central.cox.net>
Message-ID: <slrnba85p3.shc.rmunn@localhost.localdomain>

Steve Holden <sholden at holdenweb.com> wrote:
> "laotseu" <bdesth at removethis.free.fr> wrote ...
>> Greg Ewing (using news.cis.dfn.de) wrote:
>> > Tj wrote:
>> >
>> >> Lambda could also stand renaming;
>> >> Pythonistas probably think it's an arrogant name.  (Or exotic?  I
>> >> can't tell.)
>> >
>> >
>> > Actually, we think it's kind of cute.
>> >
>> >   Mary had a little lambda,
>> >   Its syntax white as snow,
>> >   And every program Mary wrote,
>> >   She wrote in Lisp, you know.
>>
>> lol !-)
>>
> This *finally* gives me the chance to ask: when a lambda grows up, does it
> become a sheepda?

Only if it is stored in RAM.

Otherwise, it's a sheepma, don't ewe know.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From aleax at aleax.it  Sat Apr 26 06:05:42 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 10:05:42 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com>
Message-ID: <WRsqa.9333$K35.305964@news2.tin.it>

Terry Reedy wrote:

> 
> "Frantisek Fuka" <fuka at fuxoft.cz> wrote in message
> news:b8cadp$1fj9$1 at ns.felk.cvut.cz...
>> When I try to call methods, I sometimes forget to include the
>> parentheses. Instead of:
> 
> In Python, '()' is, in the proper context, the 'call me with the
> contained (possible none) arguments' operator.  So the above is much
> like saying, "When I try to calculate products, I sometimes forget to
> include the asterisk' -- like mathemeticians are in the habit of
> doing -- and then asking that interpreter start guession that missing
> binary operators are implied multiplication.

Hmmm -- not really: what the OP's asking is, rather, "much like" asking
that, when he does forget an asterisk, Python should give him an error
message rather than silently accepting his errant code and doing nothing
much with it.

And indeed, if you do forget an asterisk when coding a multiplication,
you do normally get an error except in very unlucky combinations of
circumstances -- e.g. if you ALSO omit the whitespace, are trying to
multiply variables named 'foo' and 'bar', and there is ALSO a variable
named 'foobar' that's accessible from the current scope.  In practice,
therefore, omitted asterisks aren't much of a problem for new Python
users -- they get error messages and soon learn to use the asterisk,

Omitted parentheses (to call an argument-less function or method) are 
more of a problem because the "error" is subtler -- some action that
should be performed just isn't, and the effects of the bug show up,
typically, much later.

Errors should not pass silently, unless explicitly silenced.  Yet,
that is exactly what this specific error -- omitting the trailing
parentheses when calling an argument-less function or method -- appears
to do to the Python beginner (who's coming from languages where the
parentheses aren't needed in this case, typically Pascal or something
derived from Pascal -- popular languages for teaching beginners, at
least around here).

I don't know what to DO about it, but some minor tweaks might help.
For example, I cannot think of any code I've ever written that would
break if methods (and functions) spouted a __nonzero__ special method
(or, of course, the C-coded equivalent thereof) raising an exception
with a clear error message -- this would at least provide prompt
error messages for usages such as "if x.amethod:" and the like.  And
MAYBE PyChecker could be enhanced to optionally warn about code that
gets an attribute from an object and does absolutely nothing with the
attribute it's gotten -- yes, this WOULD give warnings for the
reasonably common idiom of
    try: x.something
    except AttributeError: ...
but this isn't an idiom a beginner's all that likely to use, therefore
the *optional* nature of the warning might help -- beginners would get
the warning, advanced users would suppress it.

I think this may be important enough to warrant a PEP, as I have
already suggested to the OP.  Of course, like most all developments
in open-source software, it requires somebody interested enough in
the issue to push it forwards -- if it scratches nobody's itch, it's
best dropped.  But, the problem is not with the _code_ necessary to
enhance Python's error-diagnostics for these cases (can't speak about
PyChecker, which I've never hacked on, but the mod to the Python core
for what I've mentioned above IS quite simple) -- rather, the problem
is with defining exactly what issues we can solve, with what mods to
Python as it stands, giving the pro's and con's of each possible
solution (including that of doing nothing), and letting the BDFL
Pronounce about it all.  Exactly what a PEP is good for!


> the object resulting from the conditional expression.  The boolean
> value of function objects is always True.

Yes, and the latter is exactly the feature which I think we might
do away with -- I've never written any code USING it.

>> You can say to me "Don't forget to always include the parenteses"
> 
> Just as you should not forget the asterisk in multiplications.

But he'd reliably get prompt error messages for the latter error --
he doesn't for the former -- THAT is the problem.

>> I'm still curious if this cannot be somehow configured,
> 
> The consistency of Python expression evaluation is one of its basic
> features.  Fighting it is the path to frustration.

Yes, but there would be no inconsistency in decreeing that function
and method objects can't be transparently evaluated as true or false
in boolean contexts, any more than, say, in decreeing that complex
numbers cannot be compared for lesser/greater (a change that was in
fact done moving 1.5.2 -> 2.0) -- indeed the latter change, apparently
inspired by purism rather than pragmatism, did give me a little bit
of trouble (the keys.sort() step broke in a piece of code I had, that
was displaying a dictionary in an ordered way, when some keys were
complex) while the former (proposed) change wouldn't.


> Have you never yet typed in 'obj.meth.__doc__' at the interactive
> interpreter prompt?  That is possible precisely because 'obj.meth' is
> precisely what it claims to be.  Also, it should not be too long
> before you want to start abbreviating obj.meth and do something like
> 
> stack = []
> push = stack.append
> pop  = stack.pop
> #code using these three names, simple example being
> push(1); print stack
> # [1]
> print pop(), stack
> # 1, []

None of the changes I'm considering would in the least affect either
of these important use cases, nor the many other common ones (such
as putting functions and methods as values in lists or dicts, etc).


Alex




From aleax at aleax.it  Tue Apr 15 19:16:42 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 23:16:42 GMT
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1> <3e9c1413$1@news.012.net.il> <jJUma.23762$T34.707989@news2.tin.it> <3e9c4a74$1@news.012.net.il>
Message-ID: <up0na.19865$LB6.500421@news1.tin.it>

Avner Ben wrote:

> "Alex Martelli" <aleax at aleax.it> wrote in message
> news:jJUma.23762$T34.707989 at news2.tin.it...
> 
>> You'll have to measure the performance for yourself, but you may
>> find it quite good -- do, and let us know!
>>
> 
>     Thank you all for the help. Actually, the problem is a little more
> complicated and involves topological sort of the names to make sure that
> names within names are not processed twice, but never mind. I got some

OK, but consider that the \b feature of re's can also help, if you're
interested in word-boundaries.  Also, re's | matches left to right --
I don't think you need a topological sort here, just order your names
longest-first and you should be fine.  Just hopefully-helpful hints...

> interesting hints. give me a few days and I'll let you know about the
> performance.

OK!


Alex




From aleaxit at yahoo.com  Thu Apr 10 02:46:40 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: Thu, 10 Apr 2003 08:46:40 +0200
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz>
Message-ID: <b7340h0r7m@enews1.newsguy.com>

Blair Hall wrote:
   ...
> In this case the attribute 'c' can be created. Perhaps a __slots__
> in the base class does not prevent attributes from being added
> to the derived class (I guess this is good). But,

Exactly.

> 
> # Third case: class A has no slots
> class A(object):
>      def __init__(self):
>          pass
> 
> class B(A):
>      __slots__ = ('z',)
>      def __init__(self):
>          A.__init__(self)
> 
> b = B()
> b.c = 3 # passes
> 
> in this case, although the class B has slots its base class
> does not. This seems to allow instances of B to ignore their
> own __slots__ requirements. This, to my mind, is a problem:
> it seems to require knowledge of the __slots__ in all
> base classes.
> 
> Can anyone steer me towards a better understanding of __slots__?

__slots__ is, essentially, an optimization (in terms of memory
consumption) intended for classes that are likely to be created 
in very large numbers: you give up the normal flexibility of
being able to add arbitrary attributes to an instance, and what
you gain in return is that you save one dictionary per instance
(dozens and dozens of bytes per instance -- that will, of course,
not be all that relevant unless the number of instances that are
around is at the very least in the tens of thousands).

If class X inherits class Y, and also has X.__slots__, then an
instance x of X can have any attribute that could be had by an
instance y of Y, PLUS attributes whose names are in X.__slots__.
This rule appears to me to be quite simple and regular.

Thus, the behavior in your "third case" should be obvious: if
the base class has no __slots__ then its instances can have any
attribute, and therefore so can instances of subclasses.

I suspect that any confusion may be caused by seeing __slots__
as something OTHER than an optimization -- for example, as a
way to reproduce the behavior of other languages where you
can, or must, specify possible attributes for every class.  It
isn't, really -- it's a memory-consumption optimization.

You complain that the simple rule for __slots__ "seems to
require knowledge of the __slots__ in all base classes", but
fail to specify what PURPOSE you think is going to be served
by meeting that "requirement" for knowledge.  Saving memory?
Sure, in order to save memory you do need to know all ways in
which memory could be consumed, at some level -- but in fact
all that you require is that every base class HAVE __slots__,
no knowledge of WHAT slots are used in every base class is at
all necessary in order for the purpose of __slots__ to be
accomplished - saving the per-instance dict otherwise required.


Alex




From nirina at mail.blueline.mg  Wed Apr  9 23:40:43 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Thu, 10 Apr 2003 07:40:43 +0400
Subject: Importing problem
References: <2737271.1049712504@dbforums.com> <2745665.1049862516@dbforums.com>
Message-ID: <001301c2ff13$4e8389a0$fba2383e@raseliar>

"nitintz" wrote
>
> Actually i need to import a Qt project (Its a PyQt Project where i need
> to execute a Qt project and do some manipulations on it).

That's not enough to catch your problem.
Have you already wrtite some scripts for your project?
When you run it, what kind of problems are you facing?
Do you want to run a Qt project and open on it you program or do you want
only some features of Qt to your project?
What kind of maipulations do you want to do, image manipulation, big data
manipulation ...?
Give us more details.
Write some scripts, try and post error you encounter.

nirinA







From aleax at aleax.it  Tue Apr 29 08:37:17 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 12:37:17 GMT
Subject: SV: Function to convert fetchone() result from a list to a dictionary
References: <mailman.1051616752.7559.python-list@python.org>
Message-ID: <1mura.29970$K35.954645@news2.tin.it>

<posted & mailed>

Carsten Gehling wrote:

>> -----Oprindelig meddelelse-----
>> Fra: python-list-admin at python.org
>> [mailto:python-list-admin at python.org]Pa vegne af Alex Martelli
>> Sendt: 29. april 2003 11:09
> 
>> Yes.  I think what you REALLY want (even though you have no way to
>> know you want it:-) is Greg Stein's wonderful dtuple.py, see:
>>
>> http://www.lyra.org/greg/python/dtuple.py
>>
>> The best explanation and set of examples I know for this is in Steve
>> Holden's "Python Web Programming" (New Riders) -- I highly recommend
> 
> I already have this book, but unfortunately (for me) the chapters about
> database manipulation is about the only section of the book, that I
> haven't shuffled through. :-)

Excellent choice in books -- but if you're into databases, strange
choice of chapters;-).

> Looking at dtuple.py I think it may be shooting a bit over target. I could
> be wrong - after all, I'm very new to Python :)

Ah, but you can USE dtuple even without fully GRASPING it yet, see!


>> if you can't purchase it I think you can at least get its examples
>> free off the web (start looking at Steve's site, www.holdenweb.com).
> 
> Will do this!
> 
>> Pretty good (except that you're using tabs, which don't show up
>> when I reply with KDE 3's KNode:-)  You can do a bit better (for
> 
> Heh - I like using tabs. It makes it easier for me to control indentation.

...but not everybody's using your tools of choice.  Mail and news
readers, editors, printers & their drivers, etc, are all liable to
do weird things with tabs.  Using spaces you DO control indentation;
using tabs you don't, really, unless you can control the programs
used by everybody you might like to share your code with (e.g. to
ask for help about it).


>> So here's a step-by-step version (using SPACES, not TABS:-)...:
>>
>>
>> def cfd(cursor):
>>     values = cursor,fetchone()
>>     if values is None:
>>         return None
>>     keys = [d[0] for d in cursor.description]
>>     result_dict = dict(zip(keys, values))
>>     return result_dict
> 
> Hmm... (looking up zip() in the documentation...). Okay I understand
> everyting except this contruct:
> 
> keys = [d[0] for d in cursor.description]
> 
> Let me guess: it loops through cursor.description (obvious), for each of
> the items (which are tuples) it "outputs" element 0. The output is then
> automatically formatted in a way, that conforms to the list creation
> syntax.

Good guessing!  It's called a "list comprehension", and it's one of
Python's powerful constructs.

> Or something like that. Kind of neat - haven't seen that in any other
> language, that I've used.

Python mutuated it from the functional-programming language Haskell,
but in Haskell it looks a bit less readable (to me, at least) because
it uses puntuation instead of keywords such as "for" and "in".

>>     return dict(zip([d[0] for d in cursor.description],values))
>>
>> if you're into one-liners!-)
> 
> I'm not - I'd like to be able to read my own code 6 months from now... ;-)

Yep.  There's a balance between terseness and verbosity that makes
for best readability.  Me, I'd prepare the list of keys separately
(as in the cfd function above) and then compact the last TWO lines,
only, into a single return statement:

    return dict(zip(keys, values))

as being the best compromise in this specific case -- but, where
exactly to put the threshold IS a matter of tastes.

> BTW: Wish me welcome - I've just subscribed. And the way I continue to be
> marveled by Python, I'm not likely to unsibscribe again. :-)

Welcome!  Python IS indeed marvelous.


Alex




From revanna at mn.rr.com  Tue Apr 15 16:54:08 2003
From: revanna at mn.rr.com (Anna)
Date: Tue, 15 Apr 2003 20:54:08 GMT
Subject: book on Python for Fortran programmers
References: <3064b51d.0304150458.595afc30@posting.google.com> <vhbo9vgvsqjg8l7s0aluijrran56cg8mlm@4ax.com>
Message-ID: <pan.2003.04.15.20.55.25.783394@mn.rr.com>

On Tue, 15 Apr 2003 12:15:18 +0000, Michael Prager wrote:

> beliavsky at aol.com wrote:
> 
>>Python has been suggested by several posters to comp.lang.fortran as a
>>scripting language worth learning. What are some good book(s) and web
>>site(s) to get an experienced Fortran 95 programmer started in Python?
> 
> Of the books that I've looked at, I prefer the "Python 2.1 Bible" by
> Brueck and Tanner.  It's rather complete and doesn't start by explaining
> what a bit is.
> 
> An . . . essential reference . . . is "Python Essential Reference, Second
> Edition" by David Beazley.  It's a concise version of everything
> (reminiscent in its brevity of Metcalf and Reid) whose main drawback is
> the very small type size used.
> 
> Whatever you buy, make sure it's current at least to Python 2.1. (The
> current version is 2.2.2.)  AVOID any reference that is 1.5.x or 1.6.x --
> some worthwhile features have been added..

Well, if you're looking for up-to-date, from the reviews I've seen,
_Python in a Nutshell_ would be worth looking at for this. I've gotten the
impression that it's particularly suited to experienced programmers. 

Just my $.03 worth,
Anna




From ta-meyer at ihug.co.nz  Wed Apr 30 01:53:19 2003
From: ta-meyer at ihug.co.nz (Tony Meyer)
Date: Wed, 30 Apr 2003 17:53:19 +1200
Subject: Temporary Files
In-Reply-To: <1ED4ECF91CDED24C8D012BCF2B034F13015F21B0@its-xchg4.massey.ac.nz>
Message-ID: <1ED4ECF91CDED24C8D012BCF2B034F1318CE41@its-xchg4.massey.ac.nz>

I need to update a file - I need to write the new version to a temp
file, and then remove the old version and replace it with the new one.
This is what I'm doing at the moment:

>>> import shutil
>>> import tempfile
>>> out = tempfile.TemporaryFile("w")
>>> out.write(data)
>>> out.flush()
>>> shutil.copyfile(out.name, permanent_name)
>>> out.close()

This works fine on Windows, but on (some, at least) linuxes it fails
because out.name doesn't give the correct name.

Could someone point me towards the right way of doing this?  I realise I
can just do:
>>> out = file(tempfile.mktemp())
but it seems from the documentation that TemporaryFile is the better
method.

Is there perhaps a different method of copying?  One that doesn't take
filenames, but file objects?

I've tried googling, but everything just comes up with guff about
security flaws in Python temp files...

TIA.

=Tony Meyer





From news at datatailors.xs4all.nl  Sat Apr 26 10:19:05 2003
From: news at datatailors.xs4all.nl (Peter van Kampen)
Date: Sat, 26 Apr 2003 16:19:05 +0200
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com> <b8b7kj$g74$1@news.hccnet.nl>
Message-ID: <slrnbal5ap.jc7.news@datatailors.com>

In article <b8b7kj$g74$1 at news.hccnet.nl>, Anton Vredegoor wrote:
> Peter Hansen <peter at engcorp.com> wrote:
> 
><throwaway scripts>
> 
>>I can't see any point in continuing the discussion, can you?
> 
> This discussion about whether throwaway scripts tend to get thrown
> away or not is too heavily based on the wrong Dutch quotes. A more
> fitting quote would be "ieder voordeel heb zijn nadeel" from a famous
> Dutch soccer player (and later soccer trainer in Italy). It means as
> much as "every advantage has its disadvantage".

Just nitpicking: He (Johan Cruyff) went to Barcelona, Spain. And the
exact quote is: "Elk nadeel heeft zijn voordeel" (Elluk nadeel hep se
fordeel)

PterK

-- 
Peter van Kampen
pterk -- at -- datatailors.com



From tim_one at email.msn.com  Sat Apr 12 20:46:29 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sat, 12 Apr 2003 20:46:29 -0400
Subject: Speed of string += string
In-Reply-To: <16024.39557.236558.700373@montanaro.dyndns.org>
Message-ID: <LNBBLJKPBEHFEDALKOLCIENEEGAB.tim_one@email.msn.com>

[Carl Banks]
> Surprisingly, += doesn't appear to be a major time waster for
> small (meaning about 10000 bytes) strings.  I'm guessing it's
> because the malloc my system (Debian Linux) is optimized for small
> memory allocations.

[Skip Montanaro]
> Actually, if you're using 2.2.n or later, it's Python's small object
> allocator (see Objects/obmalloc.c), which is providing most of
> the speedup.  I'm not sure if n is 0, 1, or 2.

obmalloc isn't enabled by default before Python 2.3, where it has given
dramatic speedups in some cases (where "dramatic" depends on the platform
and the case at hand; it's generally faster everywhere, but not always
dramatically so).

There was a ton of tuning and reworking to get it to work well with
new-style classes & new dict allocation strategies, to make it 100% binary
compatible at the C API level with older Python releases, and to plug a deep
security hole.  In all, that was too much work to backport to the 2.2 line.
Enabling the optional obmalloc before 2.3 can't be recommended, because "the
security hole" mentioned allowed writing pure Python code that could
scribble over the platform malloc's internal memory.  You had to know a lot
about obmalloc internals to do that on purpose, but you didn't have to know
anything about 'em to do it by accident (a "magic cookie" distinguished
obmalloc's memory from the platform malloc's, and before 2.3 obmalloc relied
on the unlikelihood of platform malloc memory containing the magic bit
pattern by accident).

WRT joining short strings, memcpy() is quite zippy on Pentium-class boxes,
and an O(N**2) algorithm can be quicker than an O(N) one across a large
range of small N if the latter's overhead is smaller than the former's.  For
example, if a simple algorithm takes N**2 units of time, and a fancy one
100*N units, the simple algorithm is quicker for N < 100, despite being
horrid for N much larger than that.

Another kind of common string catentation task is joining a small number of
strings, where some are fixed.  Like

    '(' + guts + ')'

versus
    ''.join(['(', guts, ')'])

versus
    '(%s)' % guts

The second and third ways are about twice as fast as the first as len(guts)
increases, but they've got more fixed overhead than the first way (for
example, the third way always has the fixed runtime overhead of parsing and
interpreting the intent of the format string).  Because of that, the first
way can actually be the quickest for all len(guts) an app is likely to see,
and especially in 2.3 (because obmalloc's fixed overhead for small-object
allocation in 2.3 is indeed small).

worrying-about-small-speed-differences-is-a-way-of-life-but-not-a-
    good-one-ly y'rs  - tim





From edreamleo at charter.net  Wed Apr 23 16:49:35 2003
From: edreamleo at charter.net (Edward K. Ream)
Date: Wed, 23 Apr 2003 15:49:35 -0500
Subject: Locale settings
References: <3ea437a1_1@dns.sd54.bc.ca> <b81k5m$822$06$1@news.t-online.com> <fc3b2087.0304221550.5ebdf985@posting.google.com>
Message-ID: <vadv2v39519m2b@corp.supernews.com>

> Guess I  can use the included version for awhile
> and then, once 2.3 become common, just tell users to use that python
> version.

No need.  Try to use getpreferredencoding, and fall back to your copied
code.  For example, after reading this thread I added a global
getpreferredencoding function to my app like this:

try:
 # Use Python's version of getpreferredencoding if it exists.
 # It is new in Python 2.3.
 import locale
 getpreferredencoding = locale.getpreferredencoding
except:
 # Use code copied from locale.py in Python 2.3alpha2.
 if sys.platform in ('win32', 'darwin', 'mac'):
  << define getpreferredencoding using _locale >>
 else:
  << define getpreferredencoding for *nix >>

This code is at the top level of the module.

Edward

** Please note my new email address and Leo's new URL below **

--------------------------------------------------------------------
Edward K. Ream   email:  edreamleo at charter.net
Leo: Literate Editor with Outlines
Leo: http://webpages.charter.net/edreamleo/front.html
--------------------------------------------------------------------





From mlsj at earthlink.net  Thu Apr 17 13:23:34 2003
From: mlsj at earthlink.net (Josh)
Date: 17 Apr 2003 17:23:34 GMT
Subject: Downloading TV listings with urllib
Message-ID: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4>

Can someone give me general pointers on how it could be done? I am trying 
to go to tvguide.com and download the TV listings for my area. The problem 
is how to specify my location and cable provider. I used Proxomitron to 
look at the traffic generated while downloading the listings with my 
browser. When I used httplib with:

host = 'www.tvguide.com'
pathn = '/Listings/index.asp'

def tvg():
    try:
        h = httplib.HTTPConnection(host)
        h.putrequest('GET', pathn)
        h.putheader('Accept', 'text/html')
        h.putheader('Accept', 'text/plain')
        h.putheader('User-Agent:', 'Mozilla/5.0 (Windows; U; Windows NT 
5.1; en-US; rv:1.3) Gecko/20030312')
        h.putheader('Accept:', 'application/x-shockwave-
flash,text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/p
lain;q=0.8,video/x-mng,image/png,image/jpeg,image/gif;q=0.2,*/*;q=0.1')
        h.putheader('Accept-Language:','en-us,en;q=0.5')
        h.putheader('Accept-Encoding:', 'gzip,deflate,compress;q=0.9')
        h.putheader('Accept-Charset:', 'ISO-8859-1,utf-8;q=0.7,*;q=0.7')
        h.putheader('Keep-Alive:', '300')
        h.putheader('Cookie', 
'SITESERVER=ID=fd7d0ca25acdd2439ab3f4b48b0827b4; GBA=18; 
TVGID=9CE64992B85F4F9A82B57EBC092D11B2; nat=0; ServiceID=75347; zip=63130; 
ptfc=yfki')
        h.endheaders()
        err = h.getresponse()
    	  return err
    except socket.error, er:
        print 'socket error ', er

when I use err.read() to read the object returned I get everything in Hex.

I tried using urllib, but there I can't figure out a way to pass the 
location information etc.

As I said, I just need just general pointers on how to deal with this. Any 
help would be greatly appreciated.

Thanks

Josh

    	




From richardjones at optushome.com.au  Sat Apr  5 01:46:54 2003
From: richardjones at optushome.com.au (Richard Jones)
Date: Sat, 5 Apr 2003 16:46:54 +1000
Subject: cgi.py -- anybody using it?
In-Reply-To: <Pine.LNX.4.44.0304050051160.1193-100000@alice>
References: <Pine.LNX.4.44.0304050051160.1193-100000@alice>
Message-ID: <200304051646.55018.richardjones@optushome.com.au>

On Sat, 5 Apr 2003 09:58 am, John J Lee wrote:
> I just tried testing some HTML form upload code using cgi.py, and
> discovered that it wants lower-case HTTP headers :-/

Yep, I'm using it for a number of large user-base projects (Roundup, PyPI, 
amongst others), and I've never run into the issue you're talking about.


    Richard





From buzzard at urubu.freeserve.co.uk  Thu Apr 10 19:57:14 2003
From: buzzard at urubu.freeserve.co.uk (Duncan Smith)
Date: Fri, 11 Apr 2003 00:57:14 +0100
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com> <b73vov$d92$00$1@news.t-online.com> <HzGdnQNo1Z2-AAijXTWcog@comcast.com> <b74ckl$itb$1@newsg4.svr.pol.co.uk> <8c2dnXyBdvemeQijXTWcqw@comcast.com>
Message-ID: <b750e8$86q$1@news6.svr.pol.co.uk>

"Terry Reedy" <tjreedy at udel.edu> wrote in message
news:8c2dnXyBdvemeQijXTWcqw at comcast.com...
>
> "Duncan Smith" <buzzard at urubu.freeserve.co.uk> wrote in message
> news:b74ckl$itb$1 at newsg4.svr.pol.co.uk...
> > Yes, you do have the link function as well as the linear predictor,
> and the
> > cut-off would be 0.5 rather than 1.  But in terms of absolute
> efficiency
> > there is (according to the literature) little to choose between
> linear
> > discriminant analysis and logistic regression.  Changing the cut-off
> would
> > be a simple way of attempting to minimise Alex's 'error-cost'.
>
> For at least some types of data distributions, I can believe that.  My
> comment about stepwise construction of a feature set versus
> over-fitting with everything you can think of applies equally well to
> logistic regression.  As far as I know, logistic regression cannot be
> used for more than two mutually-exclusive groups.  I don't know
> whether Alex might ever need to.
>
> Terry J. Reedy
>

There is nominal logistic regression for handling more than two groups. I've
never used it personally (that I can remember), but it is available in eg.
Minitab.  There's no stepwise procedure in Minitab (other than the usual one
implemented by a sensible user, by examining the output and refitting).
Even so, I'd still advocate leaving some data out of the fitting process, so
that there are observations to validate the model against, then refitting
the chosen model's parameters using all the data at the end.  (Personally
I'd favour a jacknife procedure because it also helps to identify unusual /
influential observations.)  That said, a naive Bayes classifier would be
much easier to code, and probably have similar predictive performance (could
be jacknifed to avoid overfitting).  And, a simple cut off adjustment would
still be possible.

Duncan






From gustav at morpheus.demon.co.uk  Thu Apr 17 15:40:09 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Thu, 17 Apr 2003 20:40:09 +0100
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
 <mailman.1050438927.20366.python-list@python.org>
 <yu99y92bh10b.fsf@europa.research.att.com>
 <mailman.1050457621.2105.python-list@python.org>
 <slrnb9q90a.mhg.news@datatailors.com> <3E9DDC81.932475AF@alcyone.com>
Message-ID: <3ckgewva.fsf@morpheus.demon.co.uk>

Erik Max Francis <max at alcyone.com> writes:

>> But these are not really 'algorithms'. Most algorithms are a lot
>> more complex even when reduced to it's essentials. I understand why
>> quicksort is better than bubblesort but I don't see an obvious way
>> to label quicksort O(??).
>
> But there is always some set of behaviors that are core to the
> algorithm, and that's what you look at as the number of elements goes to
> infinity.  Bubble sort is O(n^2), but quicksort is O(n ln n).

Just to clarify this point a bit, the measure usually used in sorting
is the number of swaps of two elements (or the number of comparisons -
these are usually equivalent in terms of O() calculations).

So, you look at bubblesort. Basically, each element gets compared
against each other - that's n * n comparisons, hence O(n**2).

But for quicksort you partition, and compare each element against the
partitioning element. That's n comparisons against log(n) partitioning
elements (each partitioning step halves the size of the chunks, so you
do this log2(n) times). So the complexity is O(n*log(n)).

Disclaimer: This is all *very* rough, and from memory as well. But I
hope it's enough that you get the idea.

Paul.
-- 
This signature intentionally left blank



From bdesth at removethis.free.fr  Fri Apr 18 13:50:55 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Fri, 18 Apr 2003 17:50:55 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com> <b7kvuh$20aem$1@ID-169208.news.dfncis.de>
Message-ID: <3EA03AFF.1030309@removethis.free.fr>

Greg Ewing (using news.cis.dfn.de) wrote:
> Tj wrote:
> 
>> Lambda could also stand renaming;
>> Pythonistas probably think it's an arrogant name.  (Or exotic?  I
>> can't tell.)
> 
> 
> Actually, we think it's kind of cute.
> 
>   Mary had a little lambda,
>   Its syntax white as snow,
>   And every program Mary wrote,
>   She wrote in Lisp, you know.

lol !-)

Laotseu




From peter at engcorp.com  Tue Apr 15 17:24:31 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 15 Apr 2003 17:24:31 -0400
Subject: Printing Columns
References: <mailman.1050426466.11198.python-list@python.org> <150420031341416871%DavidMLewis.NO@SPAM.mac.com>
Message-ID: <3E9C788F.74CDC1FA@engcorp.com>

David Lewis wrote:
> 
> This should give you a start:
> 
> >>> x = ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h']
> >>> x[0: :3]
> ['a', 'd', 'g']
> >>> x[1: :3]
> ['b', 'e', 'h']
> >>> x[2: :3]
> ['c', 'f']

This kind of response, requiring *alpha* software that is not yet
released and unlikely to be used by the majority of people, deserves
a caveat:

The above will not work with standard versions of Python prior to 2.3.


-Peter



From xxdigitalhellxx at hotmail.com  Tue Apr  8 13:26:37 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Tue, 08 Apr 2003 11:26:37 -0600
Subject: executable won't run on another machine
In-Reply-To: <slrnb95sur.t8.do-not-spam-ben.hutchings@tin.bwsint.com>
References: <v90mh6ppf9bc09@corp.supernews.com> <slrnb95sur.t8.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <b6v0qj$mhl$2@heckle.cc.uregina.ca>

Ben Hutchings wrote:
> In article <v90mh6ppf9bc09 at corp.supernews.com>, Lance wrote:
> 
>>Hi All,
>>
>>I created an executable using McMillan's Installer. It runs great on the
>>development machine. I copied it to another machine and it doesn't run. I
>>type the executable name from a command prompt, hit return, and get another
>>command prompt.
> 
> <snip>
> 
> Did you put './' in front of the executable name?  Most Unix systems are
> configured not to include the current directory in the search path.  If
> you forgot to do that then you could be running some other program that is
> in the search path.

This is especially true if you call your executable 'test' (which I've 
done more than once) because the Unix command 'test' prints nothing to 
stdout or stderr, so you just get another command prompt.

-- 
Hope this helps,
Ryan




From adalke at mindspring.com  Fri Apr 25 15:43:43 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Fri, 25 Apr 2003 13:43:43 -0600
Subject: xmlrpc introspection
References: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com> <af6b5bdd.0304240434.fd7e6ad@posting.google.com> <UR0qa.77049$gK.181217@rwcrnsc52.ops.asp.att.net> <dc6f5c99.0304250005.43ffbfcc@posting.google.com>
Message-ID: <b8c2v5$8mr$1@slb0.atl.mindspring.net>

Graham Dumpleton:
> That this presents a problem is not so much a deficiency of the the
introspection
> API but that many XML-RPC server side implementations only really easily
allow
> you to have XML-RPC services hosted out of a single URL. To my mind,
different
> web services should not all be at the same URL if they do different
things.

There is a web services philosophy called ReST ("representational state
transfer", as I recall) which takes that viewpoint quite a bit further, so
that
the function calls of XML-RPC end up being requests to move from one
URL to another (which might be created on-demand).  Done that way,
the introspection shifts to using HTTP queries rather than XML-RPC on
top of  HTTP.

It's an interesting idea, but mostly irrelevant to this discussion.

                    Andrew
                    dalke at dalkescientific.com





From duncan at NOSPAMrcp.co.uk  Wed Apr 30 08:37:04 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 30 Apr 2003 12:37:04 +0000 (UTC)
Subject: Self modifying systems question
References: <mailman.1051685036.19081.python-list@python.org> <33803989.0304300426.402cd1ef@posting.google.com>
Message-ID: <Xns936D8A6027ABduncanrcpcouk@127.0.0.1>

tebeka at cs.bgu.ac.il (Miki Tebeka) wrote in 
news:33803989.0304300426.402cd1ef at posting.google.com:

>> Self modifying code is a well known technology. Has anyone done this in
>> Python? What I'm seeking for is examples of software that adds to
>> itself (methods and variables) in runtime. I'd like an instantiated 
>> object to
>> be able to add a method to itself.
>>>> class A:
>      pass
> 
>>>> a = A()
>>>> a.f = lambda x: x + 1
>>>> a.f(100)
> 101

That doesn't really add a method to 'a', it adds a function, but the 
function doesn't have any way to determine which instance it was called on. 
You need to simulate the same binding that you get when calling a method 
defined in the class.

Here is one way to add a method to an instance (Python 2.2 and on):

>>> def f(self):
	return self.x + 1

>>> class A: pass

>>> a.x = 5
>>> a.f = f.__get__(a)
>>> a.f()
6
>>>

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From lekishvili at python.qartu.com  Thu Apr 10 13:01:12 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 10 Apr 2003 10:01:12 -0700
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <hfYka.2414$h%2.256728@read1.cgocable.net>
Message-ID: <a3b5b608.0304100901.58ced140@posting.google.com>

myk <mykbourassa at cogeco.ca> wrote in message news:<hfYka.2414$h%2.256728 at read1.cgocable.net>...
> Don't really need a tool.  Just:
> 
> a)  "centre" your data (translate for zero mean and scale for unity 
> variance);
> b)  do svd (in NumPy I think) on the data set resulting from a);
> c)  eigenvectors are columns of U and eigenvalues are diagonal of S.  
> PCA scores are just the values in the columns of U i.e. 1st column is 
> first PC scores, etc.

Just a small correction:
The difference between the SVD and PCA is that the column vectors of
U, unlike to T score matrix of PCA, are normalized. In other words, we
have:

X = U*S*V (SVD)
X = T*P (PCA)

from which follows that T = U*S.

P is loading matrix used for the initial variable importance
diagnostics.

GRTZ,
Giorgi

V and P are in fact the same, as normalization has place during the
decomposition


> 
> You have much more flexibility if you don't use a "package".  HTH
> 
> majb
> 
> sebastien wrote:
> 
> >Hi,
> >
> >Is there any PCA analysis tools for python ?
> >If it does, do you have any idea on how well it would scale ?
> >
> >I have already seen PyClimate (but it is not available for Windows
> >which will be one of the target). Is there some LAPACK like packages ?
> >
> >Thanks for reading this mail.
> >
> >Sebastien.
> >  
> >



From mhammond at skippinet.com.au  Sun Apr 13 09:29:17 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Sun, 13 Apr 2003 13:29:17 GMT
Subject: win32all odbc/MS SQLServer/bigint
In-Reply-To: <mailman.1050177799.27007.python-list@python.org>
References: <mailman.1050177799.27007.python-list@python.org>
Message-ID: <NCdma.2939$hF.24271@news-server.bigpond.net.au>

Bjorn Pettersen wrote:
> MS <groan>). My other options are wrapping our db lib as a Python
> extension (mostly done, but not DB API compliant), or recoding in C++
> (should be fast, I've got a working prototype :-)

Don't forget "fixing the ODBC module" ;)  I would if only I was setup to 
test it :(

Mark.




From staschuk at telusplanet.net  Fri Apr 25 11:06:12 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 09:06:12 -0600
Subject: why not say just what wants an integer?
In-Reply-To: <20030423075316.GA13945@upl.cs.wisc.edu>; from hupp@cs.wisc.edu on Wed, Apr 23, 2003 at 02:53:16AM -0500
References: <87d6jeeyo3.fsf@jidanni.org> <20030423075316.GA13945@upl.cs.wisc.edu>
Message-ID: <20030425090612.D261@tibia.amotlpaa.bogus>

Quoth Adam Hupp:
> On Wed, Apr 23, 2003 at 10:27:08AM +0800, Dan Jacobson wrote:
  [...]
> > TypeError: an integer is required
  [...]
> > BTW, couldn't python be improved by saying just what wants an
>   integer?
> 
> It's generally bad form to smash everything on one line like that.  A
> message like "range requires an integer" would be somewhat more useful
> though.

Though only somewhat.  It's easy to imagine cases in which the
non-integer had been passed down a few layers of call and only
detected as invalid in a function which the user didn't think
they'd called.  For a simple case:

    import random

    def choose(n, k):
        """Choose k random values from [0,n), in some random order."""
        unselected = range(n)
        selected = []
        for _ in range(k):
            val = random.choice(unselected)
            selected.append(val)
            unselected.remove(val)
        return selected

Then the exception raised by choose(3, 'foo') would not be more
perspicuous for mentioning range.

It *would* be more perspicuous if it mentioned the value which is
not an integer but should be, say by
    raise TypeError('%r is not an integer' % val)
Even better would be for tracebacks to include argument values, as
was suggested lately.

-- 
Steven Taschuk                staschuk at telusplanet.net
"I tried to be pleasant and accommodating, but my head
 began to hurt from his banality."   -- _Seven_ (1996)




From aahz at pythoncraft.com  Wed Apr  2 13:12:13 2003
From: aahz at pythoncraft.com (Aahz)
Date: 2 Apr 2003 13:12:13 -0500
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com> <slrnb8lbo0.ccp.rmunn@localhost.localdomain> <3E8AFFB7.9516A49F@engcorp.com> <b6f7ib$4rc09$1@ID-139865.news.dfncis.de>
Message-ID: <b6f95t$16p$1@panix1.panix.com>

In article <b6f7ib$4rc09$1 at ID-139865.news.dfncis.de>,
Dave Kuhlman  <dkuhlman at rexx.com> wrote:
>
>But suppose that I *do* implement double linked lists.  I'll have two
>objects that point at each other, in other words circular references.
>
>In older versions of Python, circular references were not garbage
>collected.  Is this still true in Python 2.2 and 2.3?  Do I need to use
>the gc module to *enable* garbage collection?  Or, is not *disabling*
>it good enough?  Do I still need to worry about this?  And, suppose you
>tell me that circular references are collected but I don't believe you.
>How can I check?

There would only be a circular reference if the doubly-linked list were
not referred to from outside.  When all external references are deleted,
garbage collection will reap the cycle.  (Note that I'm not answering the
rest of your questions; you can find that info in the docs.)
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From mcfletch at rogers.com  Sun Apr 20 01:11:19 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 20 Apr 2003 01:11:19 -0400
Subject: red hat 8.0 and python
In-Reply-To: <200304200005.43862.dreed@capital.edu>
References: <3EA21326.7070100@sympatico.ca> <200304200005.43862.dreed@capital.edu>
Message-ID: <3EA22BF7.8040402@rogers.com>

You could also try the libraries/python group on rpmfind.net

    http://rpmfind.net/linux/RPM/Libraries_Python.html

which has RPMs named "tkinter-2.2.1-19" and "tkinter-2.2.2-1" (as well 
as a number of other Python RPMs).

Enjoy,
Mike

Dave Reed wrote:

>On Saturday 19 April 2003 23:25, Peter wrote:
>  
>
...

>It's on one of the CDs - install the tkinter-2.2.1-17 rpm file.
>  
>
...

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From john.burton at jbmail.com  Mon Apr 14 06:40:41 2003
From: john.burton at jbmail.com (John Burton)
Date: 14 Apr 2003 03:40:41 -0700
Subject: wxHtmlWindow image question
Message-ID: <2402252b.0304140240.ae74443@posting.google.com>

I'd like to use wxHtmlWindow to render some html in my python
application which I generate at run time. I can use .SetPage('test')
to set the text, but I want to have some images in the html. I want to
generate the images on the fly. Is there any way to do this other than
by generating the images to disk files which seems a bit unncessary. I
mean is there any way to directly provide the image data to the
wxHtmlWindow component on request?



From vze4rx4y at verizon.net  Wed Apr  9 16:02:56 2003
From: vze4rx4y at verizon.net (Raymond Hettinger)
Date: Wed, 09 Apr 2003 20:02:56 GMT
Subject: python 2.3a2 Win32 expected failures in regression tests?
References: <e2d925a8.0304090914.2fd8c154@posting.google.com>
Message-ID: <Q%_ka.1951$d8.1087@nwrdny02.gnilink.net>

> When running the regression tests, the following 3 fail:
>
> test_tarfile, test_timeout, test_zipimport
>
> (test_zip passes)
> All in all, I have 201 tests that pass, but those 3 fail.
>
> I wonder if this is normal state of the affairs for Python 2.3a2 on
> win32, or something in my build that does not click.

Two of those may fail if bz2 compression extensions haven't
been built.  Try importing bz2 to see if that is the problem.

I also have sporadic failures on test_timeout.
It seems to occur when I run the tests in background
but runs fine when all the system resources are free.


Raymond Hettinger





From exarkun at intarweb.us  Thu Apr  3 11:56:52 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 3 Apr 2003 11:56:52 -0500
Subject: How to protect python code ?
In-Reply-To: <3E8BE455.4000904@novagrid.com>
References: <3E8BE455.4000904@novagrid.com>
Message-ID: <20030403165652.GC30464@meson.dyndns.org>

On Thu, Apr 03, 2003 at 09:35:49AM +0200, christophe grimault wrote:
> Hello,
> 
> Before trying my own solutions, I wonder if someone already has a solution
> to this. We have been programming a quite large application in python and
> it contains many original algorithm and other things. The application will
> be sent to customers in a few month and since it is a commercial app, we
> need some protection against copy (illegal copies) and we need to make
> sure that critical parts of the code cannot be viewed and studied.
> 

  Why?  Is your competition going to steal your algorithm and publish their
own software (which people will then buy exclusively, instead of yours), and
which is obfuscated infinitely better than yours, so it is impossible to
prove it is based on stolen property?

  Or are your algorithms stolen? ;)

  Protect against illegal copies?  Take a look around, there is almost no
scheme that is successful here.  People warez software, it's a shame, but
it's a fact of life.

> One way is to deliver only .pyc files, but I guess this is quite a poor
> protection, uh ?

  Defeatable in a minute or two, yes.

> 
> An other way is to move some parts to a C API ext module and deliver only
> the .so, but this is quite time consuming ...
> 

  Pyrex might speed it up a bit.

> Does someone has a mean to do this, knowing that we don't seek absolute
> protection, but something that discourages people that will just "try a
> little"

  Proper licensing and a tank of attack lawyers.  Alternatively, approach
your competition before the release, and offer to sell them what you think
they're going to try to steal.

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 14 days, 12:00, 4 users, load average: 0.00, 0.03, 0.00




From duncan at NOSPAMrcp.co.uk  Mon Apr 14 10:37:18 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Mon, 14 Apr 2003 14:37:18 +0000 (UTC)
Subject: accessing a functions var from a subfunction.
References: <a3cae241.0304140616.1833e5c4@posting.google.com>
Message-ID: <Xns935D9DC5F2A38duncanrcpcouk@127.0.0.1>

seppi at seppi.de (Sebastian Wilhelmi) wrote in 
news:a3cae241.0304140616.1833e5c4 at posting.google.com:

> Now my question: Is there some way to achieve this with a nicer
> syntax?

The usual way is to define a class and store your counter in the class.
You don't actually *have* to use nested functions everywhere.

e.g.
class Counter:
    def __init__(self):
        self.count = 0
    def inc(self):
        self.count += 1
    def show(self):
        print self.count

def test():
    c = Counter()
    c.inc()
    c.inc()
    c.show()

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From cnetzer at mail.arc.nasa.gov  Thu Apr  3 18:20:59 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 03 Apr 2003 15:20:59 -0800
Subject: Tkinter question
In-Reply-To: <E1910ny-0003Rc-00@fast.aub.edu.lb>
References: <E1910ny-0003Rc-00@fast.aub.edu.lb>
Message-ID: <1049412059.675.18.camel@sayge.arc.nasa.gov>

On Thu, 2003-04-03 at 01:14, Ali Dada wrote:
> hi all:
> 
> i have a window made of a couple of buttons stacked above each other.i need to
> any a widget on the top that opens a submenu to the right when clicked. how can
> i do that? see the code below and tell me how to correct it (the submenu is
> opening downwards over the other buttons, not to the right)

Well, since you asked so nicely...

You are trying to have a menubar menu open to the right rather than
down, which I don't think is possible with Tk (or if it is, it is a lot
of kludgery).

However, you may be able to use a menubutton, instead.  This will have
different semantics than a Menu in the menubar, but it allows you to
specify the direction you want a pop-up menu to open.

Here is my reworked version of your code, to show what I mean.  (one
tip- do not pack an derived frame widget in it's own init() method.  Let
the code that created that derived frame decide how and where to pack
it.)


#!/usr/bin/env python
import sys
from Tkinter import *
from tkMessageBox import *
from tkSimpleDialog import askfloat

class Index(Frame):
    def __init__(self, parent=None):
        Frame.__init__(self, parent)
        self.makemenu(self)
        Button(self, text='Repository').pack(side=TOP, fill=BOTH)
        Button(self, text='Regression').pack(side=TOP, fill=BOTH)
        Button(self, text='Code Manage').pack(side=TOP, fill=BOTH)

    def makemenu(self, parent):
        pref = Menubutton(parent,
                          text='Preference >',
                          underline=0,
                          direction='right')

        pref1 = Menu(pref, tearoff=0)
        pref1.add_command(label='Cut', underline=0)
        pref1.add_command(label='Paste', underline=0)
        pref.configure(menu=pref1)
        pref.pack(side=TOP)

if __name__ == '__main__':
    root=Tk()
    root.title('TAGZ')

    i = Index()
    # NOTE - do NOT do a self.pack() in __init__.  That is changing
    # the behavior of the widget you are inheriting, and is very bad
    # form.
    i.pack()

    i.mainloop()






From db3l at fitlinxx.com  Mon Apr  7 18:12:36 2003
From: db3l at fitlinxx.com (David Bolen)
Date: 07 Apr 2003 18:12:36 -0400
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no>
Message-ID: <ufzouhs8b.fsf@fitlinxx.com>

"Nils Grimsmo" <nilsgri at idi.ntnu.no> writes:

> i have a problem getting wxPython to update the gui while the thread is
> running user procedures. i have tried calling wxWindow.Refresh() and
> wxWindow.Update() in various orders both on the wxFrame and the wxButton, 
> but the gui doesn't update.

The only time that the GUI will be updated is when events are being
processed, which is normally blocked while you are processing a
current event.  You can use wxYield() to process any pending events
while still inside a current event handler if necessary to get GUI
updates without returning from your current event handler.

If you don't yield (or return from processing the current event), the
exact behavior may be platform and/or version dependent, as some
widgets on some platforms may immediately update some of their
attributes when they are changed, while others may just wait for an
eventual paint event to do so.

> i have posted source for a test program below. what i would like it to do,
> is to update the button with the new label before control is returned from
> run().
> (...)

When you have your time.sleep(1.0) you are sleeping, but doing so in
an event handler such that wxPython does not continue to process any
events.  If instead you added a call to wxYield() after the update,
you should consistently see the change.

Normally, this sort of thing shouldn't be necessary, because as soon
as your run() returns, events will be processed and the button
updated.  So I'm assuming that in your larger application, run() is
actually going to run a lengthy process which may need the GUI updated
along the way.  In such a case there are a few general approaches you
can take (including putting your processing into a separate thread so
the GUI thread remains responsive).  You might want to check out
http://wiki.wxpython.org/index.cgi/LongRunningTasks to see if it
provides a little more insight into approaches you might take.

> another thing i am wondering about is to disable the button
> temporarily. must i use wxEvtHandler::Disconnect?

I suppose you could, but the more typical way would be to just disable
the button window itself (with the Enable(False) method), which should
"grey out" the button and no longer respond to any mouse clicks.

Of course, there's still some potential that someone might manage to
click on the button twice before your code runs and disables it (this
is also true with the disconnect option), so it's not a bad idea to
sanity check your run() operation so that it only starts up a process
if one isn't already running, if that is one of your constraints.

--
-- David
-- 
/-----------------------------------------------------------------------\
 \               David Bolen            \   E-mail: db3l at fitlinxx.com  /
  |             FitLinxx, Inc.            \  Phone: (203) 708-5192    |
 /  860 Canal Street, Stamford, CT  06902   \  Fax: (203) 316-5150     \
\-----------------------------------------------------------------------/



From uj797 at victoria.tc.ca  Sat Apr 12 13:05:57 2003
From: uj797 at victoria.tc.ca (Arthur T. Murray)
Date: 12 Apr 2003 09:05:57 -0800
Subject: The Hundred Year Language
References: <krPla.57087$It5.27500@news2.central.cox.net>
Message-ID: <3e983965@news.victoria.tc.ca>

"Charles Pritchard" <chuck at visc.us> wrote on Sat, 12 Apr 2003:
>
> The article is a bit long, but I thought it was a good read:
> http://www.paulgraham.com/hundred.html
>
> First Paragraph:
>  It's hard to predict what life will be like in a hundred years.
> There are only a few things we can say with certainty. We know
> that everyone will drive flying cars, that zoning laws will be
> relaxed to allow buildings hundreds of stories tall, that it will
> be dark most of the time, and that women will all be trained
> in the martial arts. Today I want to zoom in on one detail of
> this picture. What kind of programming language will they use
> to write the software controlling those flying cars?

A pure language of thought will obviate the need for
http://mind.sourceforge.net/lisp.html -- LISP -- and
all the other archaic programming languages.

Some ineradicably necessary embedded systems may still use
http://mind.sourceforge.net/forth.html -- Forth-- but other
programming languages will have been replaced with neurons.

Mentifex
-- 
http://mentifex.futureai.com/theory5.html -- AI4U Theory of Mind;
http://mentifex.futureai.com/jsaimind.html -- Tutorial "Mind-1.1" 
http://mentifex.futureai.com/mind4th.html -- Mind.Forth Robot AI;
http://mentifex.futureai.com/ai4udex.html -- Index for book: AI4U



From wester at ilt.fhg.de  Fri Apr  4 05:52:59 2003
From: wester at ilt.fhg.de (Rolf Wester)
Date: Fri, 04 Apr 2003 12:52:59 +0200
Subject: Severe c-module problem
Message-ID: <b6jo6v$715$1@nets3.rz.RWTH-Aachen.DE>

Hi,

I have a c-Module "pyopt" with a function:

static PyObject *_wrap_c_nlo_init_Hs(PyObject *self, PyObject *args) {
     PyObject *resultobj;
     char * arg0;

     cout << "in init_Hs" << endl;

     if(!PyArg_ParseTuple(args,(char *)"ss:c_nlo_init_Hs", &arg0))
             return NULL;

     cout << "in init_Hs: " << arg0 << endl;

     c_nlo_init_Hs(arg0);
     Py_INCREF(Py_None);
     resultobj = Py_None;
     return resultobj;
}

If I invoke this function (iteractively from within Xemacs):
import pyopt

pyopt.c_nlo_init_Hs("name")
pyopt.c_nlo_init_Hs("name", 1.0)
pyopt.c_nlo_init_Hs("name", "1.0")

I get:

 >>> ## working on region in file /usr/tmp/python-aEF22U...
in init_Hs
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "/usr/tmp/python-aEF22U", line 1, in ?
     pyopt.c_nlo_init_Hs(nlo.get_name())
TypeError: c_nlo_init_Hs() takes exactly 2 arguments (1 given)
 >>> ## working on region in file /usr/tmp/python-aEFDBb...
in init_Hs
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "/usr/tmp/python-aEFDBb", line 1, in ?
     pyopt.c_nlo_init_Hs(nlo.get_name(), 1.0)
TypeError: c_nlo_init_Hs() argument 2 must be string, not float
 >>> ## working on region in file /usr/tmp/python-aEFQLh...
in init_Hs

Process Python segmentation fault


I have no idea what could be wrong. I would be very appreaciative for 
any help.

Thanks in advance

Rolf




From avner at skilldesign.com  Tue Apr 15 10:15:44 2003
From: avner at skilldesign.com (Avner Ben)
Date: Tue, 15 Apr 2003 16:15:44 +0200
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1>
Message-ID: <3e9c1413$1@news.012.net.il>

"Duncan Booth" <duncan at NOSPAMrcp.co.uk> wrote in message
news:Xns935E66BE3E498duncanrcpcouk at 127.0.0.1...
> "Avner Ben" <avner at skilldesign.com> wrote in
> news:3e9b0e03$1 at news.012.net.il:
>
> That seems very strange to me as lower casing the strings has the
> potential to be much more efficient than a case insensitive compare even
> when the compare is done by a C routine. Perhaps if you posted a general
> outline of your code we could suggest ways to speed it up. The biggest
> potential for code speedup is often to use a different algorithm rather
> than trying to optimise individual operations.
>

    I am searching for one or more members of a list of names (that may get
near 1000 pieces in a normal application and may average some 40 characters)
in a string (that may average 100 characters). For each name found in the
string, I do something to the string, using the original name (in its proper
case). At present, I am looking for the exact name as stated, with agreeable
performance. However, this makes me miss many cases where the author of the
string did not care for the casing (which happens frequently). I have little
problem with either lowering or uppering the entire list once, because it is
not expected to change often and when that happens once in a while, a
subscription mechanism can take care of that. However, this requires holding
two lists (I still need the original names as are) and lowering or uppering
each string for comparison (and keep its original version for
restructuring).

    Obviously, the efficient algorithm here would be to skip the standard
comparisons altogether and write an algorithm that advances on the string
character by character and - for each - attempts to parse using the list. I
have written this kind of algorithm in C++ before and it performed
reasonably even with quantities larger than the present ones. However, when
I wrote this algorithm in Python, the performance became unacceptable. The
lesson I learnt from the experience was that the built-in string search and
comparison of the standard Python library (which, I take it, is written in
C) is at least four times faster than anything you can do for yourself, by
running character by character in Python.

    Avner.





From mis6 at pitt.edu  Mon Apr 28 12:13:21 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 09:13:21 -0700
Subject: __base__ and __bases__[0]
Message-ID: <2259b0e2.0304280813.5a94c41f@posting.google.com>

I thought __base__ and __bases__[0] where just equivalent, but it
seems it is not so:

>>> class C(object): pass
...
>>> class M(C,type): pass
...
>>> M.__bases__[0]
<class '__main__.C'>
>>> M.__base__
<type 'type'>

Question: how __base__ is computed and what's the rationale about it ?

TIA,

                                           Michele



From jason at tishler.net  Mon Apr 28 12:43:30 2003
From: jason at tishler.net (Jason Tishler)
Date: Mon, 28 Apr 2003 12:43:30 -0400
Subject: mingw python module
In-Reply-To: <000501c30cf6$865e9720$ce14d4d3@june>
References: <000501c30cf6$865e9720$ce14d4d3@june>
Message-ID: <20030428164329.GF1864@tishler.net>

June,

Please post instead of sending private email.

On Mon, Apr 28, 2003 at 04:52:24AM +0900, Changjune Kim wrote:
> Dear Jason Tishler,
> 
> I read your post on c.l.p about your compiling readline module on
> win32 using mingw at
> http://groups.google.com/groups?selm=mailman.1016455560.10460.python-list%40py
> thon.org
> 
> I seem to have compiled it with mingw32 2.0.0 and Python 2.3b1. I can
> even use it for rlcompleter. However, there is a problem. When I
> ctrl-C, python shell doesn't throw exception but segment fault.

See below...

> Strangely, the shell quits with ctrl-D like in linux.

AFAICT, the above is the expected behavior.

> Has this happend to you, too?

I never tried this before.  I just did so and can confirm your findings.

If this is important to you, then you will have to debug this yourself.
Since I have never typed a Ctrl-C into a Win32 Python interpreter in
over 2.5 years, it is safe to assume that my motivation to investigate
this corner case is extremely low. :,)

> Thank you in advance.

You are welcome.

> Sincerely,
> 
> June

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From bokr at oz.net  Sun Apr 27 14:56:26 2003
From: bokr at oz.net (Bengt Richter)
Date: 27 Apr 2003 18:56:26 GMT
Subject: WML examples using mod_python?
References: <slrnbals94.2qp.grante@tuxtop.visi.com> <uLFqa.9353$h%2.1023630@read1.cgocable.net> <slrnbamprf.2th.grante@tuxtop.visi.com>
Message-ID: <b8h94q$g7v$0@216.39.172.122>

On 27 Apr 2003 05:10:45 GMT, Grant Edwards <grante at visi.com> wrote:
>In article <uLFqa.9353$h%2.1023630 at read1.cgocable.net>, Mark Nenadov wrote:
[...]
>> If you are looking for something to make WML-generation
>> quicker, you might want to check out "pywml"
>> (http://sourceforge.net/projects/pywml).
>
>Thanks!  That's what I was looking for.  I was sure there had
>to be something like that, but all my searches found were a
>Zope/WML howto.
I was curious, but I found just an empty project.
Maybe there used to be a Weapon of Mass Lunacy there,
and it had to be removed quickly ;-/

Regards,
Bengt Richter



From dale at riverhall.NOTHANKS.co.uk  Wed Apr 23 12:45:20 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 23 Apr 2003 17:45:20 +0100
Subject: execute a file
References: <jcfdav4pr3ebi52pf5cdt1lfjmipgjuh7n@4ax.com>
Message-ID: <pmgdavgi81udqaellbpgkcbae1i60hkaga@4ax.com>

Giacomo <kylnas at tiscalinet.it> wrote:

>Hi everybody, i would like to know how can i execute a generic file.
>I mean execute a file like *.txt, *.jpg launching the application to
>see that file (an editor for txt file, for example paint shop pro for
>jpg file, and so on).
>bye and thanks

For Windows, you probably want:

 os.startfile(path)

I don't know about other platforms.
-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From mis6 at pitt.edu  Mon Apr 21 16:58:55 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 21 Apr 2003 13:58:55 -0700
Subject: Is this a super bug?
References: <mailman.1050811779.16487.python-list@python.org>
Message-ID: <2259b0e2.0304211258.482fd800@posting.google.com>

> "Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message 
> news:<mailman.1050811779.16487.python-list at python.org>...
> 
> > <snip> this is the sequence that "should" happen:
> > 1  super(ff, f)[0]  # f == self
> > 2  superinst(type:<class ff>, obj:f)[0]
> > 3  si.__getitem__(0)
> > 4    si.__getattr__('  getitem  ')
> > 5    .. return str.__dict__['  getitem  '].__get__(f)
> > 6    boundmethod(str.__getitem__, f)
> > 7  bmeth(0)
> > 8   'a'
> 

This is the sequence I would expect:

1. super(ff, f)[0] 
2. superinst(type:<class ff>, obj:f)[0]
3. type(si).__getitem__(si,0)
4. <type 'super'>.__getitem__(si,0)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AttributeError: type object 'super' has no attribute '__getitem__'

Nevertheless one gets

>>> super(ff,f)[0]
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: unsubscriptable object

Therefore I am confused :-(

On the other hand if I do

>>> class C(object): pass
...
>>> c=C()
>>> c[0] # I would expect to be C.__getitem__(c,0)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: unindexable object

I get the unindexable error, not the attribute error. The
problem has nothing to do with super, but I agree there is
something strange in how __getitem__ is invoked.
Maybe the problem is simply in the displayed error
message: displaying 

  AttributeError: <class '__main__.C'> has no attribute '__getitem__'

could be more confusing for newbies than 

  TypeError: unindexable object

and therefore the AttributeError has been changed to a TypeError.

Is this the answer ? Or am I entirely wrong on my understanding of
__getitem__ ?

                                   Michele



From jack at performancedrivers.com  Thu Apr 17 12:35:58 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Thu, 17 Apr 2003 12:35:58 -0400
Subject: "pythonic"
In-Reply-To: <qait9vsttagq0c8gk5f40ou991bitistfe@4ax.com>; from jkraska@san.rr.com on Thu, Apr 17, 2003 at 03:39:44PM +0000
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com> <qait9vsttagq0c8gk5f40ou991bitistfe@4ax.com>
Message-ID: <20030417123557.B1553@localhost.localdomain>

On Thu, Apr 17, 2003 at 03:39:44PM +0000, Courageous wrote:
> 
> >I do not know if this ignorance, laziness, or a meme, but I would like
> >to know if there a real definition of "pythonic". Anyone?
> 
> First you have to understand the word "unpythonic".
> 
> This is kind of like the word "communist". You know? During the
> 50's, your life was over if you were labeled a "communist"? 
> Guido, Tim, and team sit on a the Technical Committee for Un[ythonic
> Activities.

I know you were being tounge-in-cheek, but to be a Communist in the 50s was to
advocate the violent overthrow of the US government.  Today it just means
you like hanging out in espresso bars.  McCarthy might have been a self 
promoting dick, but there really _was_ a secret cabal out to get us. The fact
that they didn't succeed doesn't mean they didn't exist.

Back to humor, at a similar level of severity, "Unpythonic" would mean secretly
advocating (static typing | implicit is better than explicit | Guido getting 
hit by a bus and being replaced by Larry Wall).  To be "purged" would mean
having your CVS write permissions revoked.

can't-think-of-anything-clever-about-dead-and-red-ly,

-jack




From donn at u.washington.edu  Mon Apr 28 13:22:14 2003
From: donn at u.washington.edu (Donn Cave)
Date: 28 Apr 2003 17:22:14 GMT
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com> <zAara.22169$K35.739509@news2.tin.it> <b8jls7$qve$1@panix2.panix.com>
Message-ID: <b8jo06$32n0$1@nntp6.u.washington.edu>

Quoth aahz at pythoncraft.com (Aahz):
| In article <zAara.22169$K35.739509 at news2.tin.it>,
| Alex Martelli  <aleax at aleax.it> wrote:
|> Aahz wrote:
|>> In article <b8hc64$6n2$1 at slb9.atl.mindspring.net>,
|>> Andrew Dalke <adalke at mindspring.com> wrote:
|>>>
|>>>        i = i + 1
|>> 
|>> I'm a little curious why you didn't write this as
|>> 
|>>     i += 1
|>> 
|>> Any reason (i.e. still need to support 1.5.2) or just habit?  If the
|>> latter, why haven't you changed your habit?
|>
|> There is no special reason to "change that habit", e.g. not performance:
|>   [...]
|> No statistically significant / measurable difference.  The tiny difference
|> in conciseness is hardly compelling one way or another, either.
|
| However, I do think the difference in readability is somewhat compelling.
| Using augmented assignment, I can think to myself, "Ah, this is mutating
| an object."  (This aside from the fact that numbers can't actually be
| mutated -- it's a convenient mental shorthand.)

Hm, so strictly speaking, you can think to yourself, ``ah, this may
be mutating an object, depending on the implementation.''  Readability
doesn't seem very compelling here, since the point is ambiguous.
Maybe to avoid confusion one should never use augmented assignment
at all, and invoke a method instead when the intent is to mutate.

	Donn Cave, donn at u.washington.edu



From jidanni at dman.ddts.net  Tue Apr 22 22:27:08 2003
From: jidanni at dman.ddts.net (Dan Jacobson)
Date: Wed, 23 Apr 2003 10:27:08 +0800
Subject: why not say just what wants an integer?
Message-ID: <87d6jeeyo3.fsf@jidanni.org>

Say, does lack of ability to figure out how to make this work 
$ python -c 'print range(map(lambda x:x+1,(11,131,11)))'
Traceback (most recent call last):
  File "<string>", line 1, in ?
TypeError: an integer is required

indicate 1. newbie, 2. lazy newbie, 3. can't really blame me?

BTW, couldn't python be improved by saying just what wants an integer?

Yes, it happened after I added range(), so range wants the integer, I
can tell that.  So list() won't help etc.

P.S. "pydoc print" gives nothing. Should it give some clue to the user
wanting to know about "print", or is that some violation of principles?
-- 
http://jidanni.org/ Taiwan(04)25854780



From Jacek.Generowicz at cern.ch  Wed Apr  2 10:33:42 2003
From: Jacek.Generowicz at cern.ch (Jacek Generowicz)
Date: 02 Apr 2003 17:33:42 +0200
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <tyf65pwvrqh.fsf@pcepsft001.cern.ch>

Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> writes:

> In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer wrote:
> <snip>
> >     3.  At most three Is, Xs and Cs may appear in any given literal.
> <snip>
> 
> OK cleverclogs, how do we write 39?

IXL. 

(That ignores some of his other restrictions, but as you omitted them ...)



From tismer at tismer.com  Mon Apr 21 22:26:11 2003
From: tismer at tismer.com (Christian Tismer)
Date: Tue, 22 Apr 2003 04:26:11 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <20030421121843.A596@tibia.amotlpaa.bogus>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <3EA2C43F.2080802@tismer.com> <20030421121843.A596@tibia.amotlpaa.bogus>
Message-ID: <3EA4A843.2050507@tismer.com>

Steven Taschuk wrote:
> Quoth Christian Tismer:
>   [...]
> 
>>Quoting Bjorn Pettersen about continuations:
>>
>>"""
>>They also seem to have the property that before you understand them, no
>>explanation will improve your understanding; and after you understand
>>them you lose the ability to explain it <wink>.
>>"""
> 
> 
> Hm...
> 
> Before studying continuations, function calls are function calls
> and loops are loops.  While studying continuations, function calls
> are not function calls and loops are not loops.  After studying
> continuations, once again function calls are function calls and
> loops are loops.

But then, you should have understood how to build function
calls and loops from continuations, and that in fact the is no
"call" or "return" in the first place, but these are artificial,
from a POV where continuations are the basic building block.

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From a.schmolck at gmx.net  Mon Apr 14 09:25:48 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 14 Apr 2003 14:25:48 +0100
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10 <7h3vfxhpl23.fsf@pc150.maths.bris.ac.uk>
Message-ID: <yfsu1d1cisj.fsf@black132.ex.ac.uk>

Michael Hudson <mwh at python.net> writes:

> Paul Foley <see at below.invalid> writes:
> 
> > [It should do that for lexical variables, too...conflating binding and
> > assignment the way Python does is a major design error, but it's too
> > late to fix it now]
> 
> I'm unconvinced about the "major"... but anyway: how would *you* do

So am I, but...

> it?  It seems unfeasible syntactically, for starters.
> 

... what's wrong with e.g:

def bar():
   # declares locals, similar to ``global``
   var a,b,c = 3
   x = 3 # error
   a = 4 # OK

or 


def bar():
   let x=3, y=4:
     print x, y # OK
   x, y # ERROR

?

'as



From sdieselil at yahoo.com  Thu Apr 10 18:30:54 2003
From: sdieselil at yahoo.com (sdieselil)
Date: 10 Apr 2003 15:30:54 -0700
Subject: Can I define "static" (like in C++) methods in Python?
Message-ID: <169a7283.0304101430.2bec922b@posting.google.com>

Hi

Can I define methods which are attributes of class and not attributes of
instance?



From syver-en+usenet at online.no  Tue Apr  1 19:38:05 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 02 Apr 2003 02:38:05 +0200
Subject: stop exeuction and give my a prompt
References: <b6dara$5mv$1@news.asu.edu>
Message-ID: <u3cl1d98y.fsf@online.no>

David Bear <david.bear at asu.edu> writes:

> I would like to have the ability to insert a statement in my code that
> would
> 
> stop program execution at that point and give me a python prompt.  I
> don't
> 
> know what to call this. 
> 
> What I would like to do is to stop program execution and inspect the
> state
> 
> of certain dictionaries --- I don't want to insert a bunch of print 
> statements and watch output fly of the screen.  I'd also like to be
> able to
> 
> fiddle with these dictionaries when my program stops.  
> 
> Any pointers?

Use the python debugger. In the most primitive way is to use it from
the interactive prompt.

Do:

import pdb # at the top of your sourcefile

Do:

pdb.set_trace() # at the point you want to debug.

I use some magic that I found at the python cookbook that gives me a
debugger prompt if the python program I am running crashing (so called
Just In Time debugging).

For a more pleasant debugging environment check out Emacs or Pythonwin.
-- 

Vennlig hilsen 

Syver Enstad



From chrtoeller2 at freenet.de  Mon Apr 14 08:49:58 2003
From: chrtoeller2 at freenet.de (Chris)
Date: 14 Apr 2003 05:49:58 -0700
Subject: conversion of INTEGER into string
References: <a01b8ad6.0304140110.2b055605@posting.google.com> <3E9A7BA1.AEEB589B@alcyone.com>
Message-ID: <a01b8ad6.0304140449.32970c68@posting.google.com>

> You want the str function/type:
> 
> >>> str(2000)
> '2000'


I knew the way is an easy one...

thank you

and sorry for boring the rest of the group with such a trivial question


Chris



From canrocks at yahoo.ca  Mon Apr 21 13:14:39 2003
From: canrocks at yahoo.ca (Denis Agar)
Date: 21 Apr 2003 10:14:39 -0700
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
Message-ID: <25a672ee.0304210914.1cc2049c@posting.google.com>

Hi, I'm extremely new to python. I've spent the majority of my waking
hours this easter reseaching Python GUIs, compilers, and the language
in general. It's been arduous trying to piece together knowledge from
different places but I'm slowly getting the hang of it. A little about
me: My only other (major) experience with programming was a semester
of Visual Basic in school. I really wanted to get away from the
hodgepodge of bad vibes coming from Microsoft, and was suggested
almost unanimously Python by a forum (http://www.shareaza.com/forum)

I'm blabbering! Anyway, my problem is this. I tried to compile a
simple Hello World program into something runnable on non-python
computers, and it came out being 7MB in size. Is there something I'm
doing wrong? Just so you know, I do know somewhat what I'm doing. I
know that it's incorporating all of the GUI. So is there a way to not
incoporate all of the gui, and still have a functional program? I'm
thinkin in the 800k range for a simple program? The reason I ask is
because I'm thinking of porting something I made in VB to Python
before I develop it any further.

PS: The program is 92k in Visual Basic.



From dmgriffith at tasc.com  Wed Apr 23 11:45:02 2003
From: dmgriffith at tasc.com (Dang Griffith)
Date: Wed, 23 Apr 2003 11:45:02 -0400
Subject: pattern match
References: <220420031949394828%greg.bland@valpo.edu>
Message-ID: <b86cf0$oe7$1@gateway.northgrum.com>

Mitch MF Cummstain wrote:

>                I just started using python and I need help with a
> pattern match.
>       
>    m = re.match("(.{30})(word)(.{30})",paragraph);
> 
> I need to match 30 characters on each side of word.  The problem word
> is getting treated as a string not a variable like I need it to.  How
> can I make it treat word as a variable ?

I'm using 3 instead of 30, but I think you'll get the idea.

   >>> import re
   >>> paragraph = "abcdefghijklmnopqrstuvwxyz"
   >>> word = "lmn"
   >>> pat = "(.{3})(%s)(.{3})" % word
   >>> # or, pat = "(.{3})(" + word + ")(.{3})"
   >>> m = re.search(pat, paragraph)
   >>> m.groups()
   ('ijk', 'lmn', 'opq')

Something to keep in mind is that re.match only matches at the beginning
of the string.  You might consider using re.search instead, as I did
above.  It also returns match objects you can play with.
      --dang



From DennisR at dair.com  Wed Apr 16 15:10:43 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Wed, 16 Apr 2003 19:10:43 GMT
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com>
Message-ID: <TUhna.928$cq2.100423716@newssvr13.news.prodigy.com>

> I'm working with a xml document which doesn't include an encoding,

The lack of an encoding is your problem.  Pick an encoding which defines 256
characters, not 128, if you want to represent chr(176).in a single byte.

> I would like to insert the degree symbol (chr(176)), but
> because this is outside the bounds (chr(128) is the limit), Python
> raises an XML error.

UTF-8 has no chr(128) limit.  Your document is UTF-8 encoded (because you
have not specified an encoding) and you need to meet UTF-8 encoding until
you explicitly specify some other encoding.  More likely is that chr(176)
and perhaps what follows are not legal UTF-8.
--

Dennis Reinhardt

DennisR at dair.com   http://www.spamai.com





From tug at wilson.co.uk  Tue Apr 29 13:18:28 2003
From: tug at wilson.co.uk (John Wilson)
Date: Tue, 29 Apr 2003 18:18:28 +0100
Subject: Problem subclassing tuple
References: <mailman.1051631814.18940.python-list@python.org> <b8m8rm$sef$1@panix3.panix.com>
Message-ID: <097a01c30e73$59f17490$640a0b0a@handel>

----- Original Message ----- 
From: "Aahz" <aahz at pythoncraft.com>
[snip]

>
> Use __new__ instead of __init__.
> -- 

I tried:

...    def __new__(cls, *args, **kargs):
...         instance = object.__new__(cls)
...         tuple.__init__(instance, args)
...

But I got:

>>> Holder(1, 2)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "<stdin>", line 3, in __new__
TypeError: object.__new__(Holder) is not safe, use tuple.__new__()

I'm sure I'm being dense. Can you advise me as what to do in the __new__
method?

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From martin at v.loewis.de  Tue Apr 22 01:48:13 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 22 Apr 2003 07:48:13 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <erk5avkh423eh8bc13ctvtbjqt9f6kt4kv@4ax.com> <84smscm5df.fsf@boost-consulting.com> <pf58avcm8u8r84183tia6alfo8gctc4fdu@4ax.com> <84el3vptmf.fsf@boost-consulting.com> <6999avcasif2smdq6q117vrbcouh2id7gd@4ax.com>
Message-ID: <m3fzobdqw2.fsf@mira.informatik.hu-berlin.de>

Courageous <jkraska at san.rr.com> writes:

> Did it, long time ago. It was either on a HPUX or Solaris machine.
> IIRC. Perhaps I'm no longer current, though. :)

Atleast for Solaris, it still works fine. See

http://sourceforge.net/projects/odem

for a library that uses that mechanism.

Regards,
Martin



From spahievi at vega.bg  Sat Apr  5 05:32:05 2003
From: spahievi at vega.bg (Niki Spahiev)
Date: Sat, 5 Apr 2003 13:32:05 +0300
Subject: What are the differences between SWIG and Boost.Python?
In-Reply-To: <b6l0eq$tf0$1@main.gmane.org>
References: <169a7283.0304040821.9ad0976@posting.google.com>
 <b6l0eq$tf0$1@main.gmane.org>
Message-ID: <103610962.20030405133205@vega.bg>

4/5/2003, 1:20:42, Mike Rovner wrote:

MR> FWIW, I decided for myself:

MR> - wrap small C libs with Pyrex,
MR> - wrap big C libs with SWIG,
MR> - wrap small C++ libs with boost.python,
MR> - wrap big C++ with (I still have to do that ;)

pyste = (python + gcc-xml + boost.python)

-- 
Best regards,
 Niki Spahiev





From mathias-usenet at valpo.de  Wed Apr 16 08:14:52 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 16 Apr 2003 12:14:52 GMT
Subject: Problem Cron-Job using Python-Script
References: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de> <mailman.1050493585.1066.python-list@python.org>
Message-ID: <b7jhfs$1i2r1$1@ID-15245.news.dfncis.de>

Gerhard Haering <gh at ghaering.de> wrote:

> Cron normally doesn't set any environment variables.

Thats not right, cron sets at least HOME, SHELL and LOGNAME. 

Mathias



From cben at techunix.technion.ac.il  Wed Apr  9 13:35:21 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Wed, 9 Apr 2003 20:35:21 +0300 (IDT)
Subject: STRING TO LIST
In-Reply-To: <Xns9358627B73228duncanrcpcouk@127.0.0.1>
References: <mailman.1049843053.19630.python-list@python.org>
 <sNicneNGn5EZ6g6jXTWcpQ@comcast.com> <Xns9358627B73228duncanrcpcouk@127.0.0.1>
Message-ID: <Pine.GSO.4.51.0304091923570.1222@techunix.technion.ac.il>

Duncan Booth wrote on 2003-04-09:

> While this works, the original poster should be aware that unless
> he/she has full control over the origin of the string, it could have
> nasty side effects, since the string could do things like opening
> files or even executing arbitrary Python code.

Indeed.  One other attack, except for calling builtins, is mutating
variables local to the caller of `eval` with list comprehnsions.
This means that even untrusted safely evaled code that doesn't itself
have access to builtins, better use a safe eval for third code that it
doesn't trust (sandbox in sandbox).

> Of course this may not be relevant, but in case it is, a better solution
> is:
>
> >>> x=str("[[12,13],[14,15]]")
> >>> print eval(x, {'__builtins__': {}})
> [[12, 13], [14, 15]]
> >>>
>
Is this completely safe?  First, note that you must catch all
exceptions, because of e.g.:

>>> eval('''1/0''', {'__builtins__': {}})
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "<string>", line 0, in ?
ZeroDivisionError: integer division or modulo by zero

Futhermore, can't one do some nasty thing through attributes of
builtin types?  Here is a simple security hole (probaly highly
dependant on specific python versions):

Python 2.2.2 (#1, Jan 30 2003, 21:26:22)
[GCC 2.96 20000731 (Red Hat Linux 7.3 2.96-112)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> "".__class__
<type 'str'>
>>> "".__class__.__class__
<type 'type'>
>>> "".__class__.__base__
<type 'object'>
>>> dir(type)
['__base__', '__bases__', '__basicsize__', '__call__', '__class__',
'__cmp__', '__delattr__', '__dict__', '__dictoffset__', '__doc__',
'__flags__', '__getattribute__', '__hash__', '__init__',
'__itemsize__', '__module__', '__mro__', '__name__', '__new__',
'__reduce__', '__repr__', '__setattr__', '__str__', '__subclasses__',
'__weakrefoffset__', 'mro']
>>> type.__subclasses__(object)
[<type 'type'>, <type 'list'>, <type 'NoneType'>,
<type 'NotImplementedType'>, <type 'module'>,
<type 'posix.stat_result'>, <type 'posix.statvfs_result'>,
<type 'dict'>, <type 'function'>, <type 'str'>, <type 'file'>]
>>> eval('''[t for t in
...          "".__class__.__class__.__subclasses__(
...                                     "".__class__.__base__)
...          if t.__name__ == "file"][0]''',
...      {'__builtins__': {}})
<type 'file'>
>>> eval('''[t for t in
...          "".__class__.__class__.__subclasses__(
...                                     "".__class__.__base__)
...          if t.__name__ == "file"][0]('CRACK', 'w')''',
...      {'__builtins__': {}})
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "<string>", line 1, in ?
IOError: file() constructor not accessible in restricted mode

What?!  They think they are smarter than I am :-)?  OK, they got me
this time.  How is this "restricted mode" detected anyway?  Does
`file` look up the stack frame to see if it's called from eval, or if
it has no `__builtins__`?  Is it a permanent mode of the interpreter
itself?

In any case, this is something special coded into `file`.  But the
`type..__subclasses__(object)` list above shows all classes your
runtime has, from any modules.  In any big program there would be one
that is not protected:

>>> class foo(object):
...   def __init__(self, fname):
...     file(fname, 'w')
...
>>> type.__subclasses__(object)
[<type 'type'>, <type 'list'>, <type 'NoneType'>,
<type 'NotImplementedType'>, <type 'module'>,
<type 'posix.stat_result'>, <type 'posix.statvfs_result'>,
<type 'dict'>, <type 'function'>, <type 'str'>, <type 'file'>,
<class '__main__.foo'>]

Indeed, any class appears!  Now, if happen to know about this class in
your code, or write some byte-code disassembling thingie for finding
such classes automatically (I'm pretty sure that lambdas together with
list comprehensions are turing-complete), I'm in:

>>> eval('''[t for t in
...      "".__class__.__class__.__subclasses__(
...                                 "".__class__.__base__)
...      if t.__name__ == "foo"][0]('CRACK')''',
...      {'__builtins__': {}})
<__main__.foo object at 0x811c4ec>
>>>
[1]+  Stopped                 python2
cben at sc8-pr-shell1:~$ ls CRACK
CRACK

See?

Another point is that it's trivial to write fairly short list
comprehensions that will block for very long or infinite times and
possibly explode you memory.  Actually you don't need list
comprehensions for this, something as simple as ``eval('2**(2**20)')``
took too long for my patience...  So the safest `eval` will always be
open to denial-of-service attacks!  I wonder whether timeout and a
memory limit can be implemented to avoid it...

Still, frequently you don't need the flexibility of allowing real code
(with more or less restrictions), you just want to read some constant
data.  It's the most natural thing for a LISPer: read without eval.
OK, in Python *there isn't* any read syntax for lists, tuples and
dicts, there are only "dysplays" (the analogue of the list function as
compared to the quote special form).  But it'd be trivial to define.
So why doesn't python have some builtin to parse it?  (Don't tell me
about pickle, it's not very readable <wink> and certainly not
secure...).

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From eric.brunel at pragmadev.com  Wed Apr 30 10:43:59 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Wed, 30 Apr 2003 16:43:59 +0200
Subject: Three questions about Tkinter
References: <BAD59F7C.B1B%arnaud@tribu.ch>
Message-ID: <b8on3n$8au$1@news-reader13.wanadoo.fr>

Arnal wrote:
> Hi,
> 
> I'm new at python programming (as well as english writing!), and I'm trying
> to make a simple text editor.
> 
> I'm using Tkinter, but I've the following problems:
> 
> -I can't figure out how to use a printer to print under Windows.

Short answer: you can't. That's one of the biggest issues in Tkinter and tk 
(IMHO...): there is no access to the printing facilities on the current 
platform, whatever that means. Note it's a limitation of tk itself, not of 
Tkinter. I hope the tcl/tk guys will take care of this issue some day.

You can access the regular Windows printing facilities via the win32ui et al 
modules, but it's far from easy and it may even crash your application on Win9x. 
The usual workaround is to generate in a "common" format like .doc or .pdf, then 
   print the generated file via an external application.

> -I use a "text" control (I dislike the "widget" term, sorry), but I can't

(NB: you really should learn to like the "widget" term, since it's quite the 
standard name for these things... :-) )

> resize it on the fly (the width and height properties are measured in
> characters and lines, not in pixels). My current workaround is to make the
> window not resizable but it would be nice to do it.

I don't exactly understand what you're trying to do here. What's the problem? 
Can you post a piece of code showing it?

> -I've a sub menu for recents files. When I read in my pref file to list
> every file paths stored in it, I would like to add them in the recents
> submenu. I can't find how to do it. I tried, inside a "while" with the
> variable i:
> 
> MRecents.insert_command(0,label=GetFileName(file),accelerator=acc,command=Op
> enRecent(i))
> 
> MRecents.insert_command(0,label=GetFileName(file),accelerator=acc,command=Op
> enRecent i)
> 
> (MRecents is a global variable representing my submenu, GetFileName is a
> function, no special meaning in the problem,acc is defined). In the first
> example, the OpenRecent function is called directly, that's the normal
> behavior, and there is a syntax error in the second example.
> How to define an array of menu items with all calling the same function with
> a parameter?

The most common way to do it is via a lambda function. If you don't know what it 
is, look at:
http://www.python.org/doc/current/tut/node6.html#SECTION006740000000000000000

Practically, you'd write what you want to do like follows:

MRecents.insert_command(0, label=GetFileName(file),
                            accelerator=acc,
                            command=lambda i=i: OpenRecent(i))

There are more sophisticated solutions but this one should be enough for you.

HTH
-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From jepler at unpythonic.net  Mon Apr  7 15:37:32 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Mon, 7 Apr 2003 14:37:32 -0500
Subject: Tkinter - tracking toplevel windows
In-Reply-To: <d44e8ee7.0304071023.1d571c2b@posting.google.com>
References: <d44e8ee7.0304071023.1d571c2b@posting.google.com>
Message-ID: <20030407193732.GB27968@unpythonic.net>

On Mon, Apr 07, 2003 at 11:23:13AM -0700, Andrew Wheatley wrote:
> I have a Tkinter Gui with a button which generates other toplevel windows.
> Whenever a toplevel is generated, I append its id to a global list.  What I'd like
> to do with wm_protocol/WM_DELETE_WINDOW method is remove from the list
> the window id currently being killed and then destroy the window.
> 
> My question is:
> 
> How can I tell which window is being destroy when I kill it with the X in the 
> top right hand corner.

Arrange to include a reference to the window when in the callback.  For
example:

    def close_clicked(t):
        print "close was clicked on window", t

    t = Tkinter.Toplevel()
    t.wm_protocol('WM_DELETE_WINDOW', lambda t=t: close_clicked())

When WM_DELETE_WINDOW is defined, using window-manager "close"
functionality (such as ALT-F4, clicking the "X", etc) merely triggers
the callback, which can override the close event.

There's also the <Destroy> binding, but this is a little more
problematic.  You *can* use the .widget attribute of the event, but due
to a detail of Tk internals*, this binding will be applied to any widget
inside t.  So below, you'll get two messages:

    def destroyed(evt):
        print "window", evt.widget, "was destroyed"

    t.bind("<Destroy>", destroyed)
    b = Tkinter.button(t); b.pack()
    t.destroy()

So if you're interested only in toplevel windows, do this:
    def destroyed_smarter(evt):
        if not evt.widget.wm_toplevel() is evt.widget:
            return # not actually the window
        print "window", evt.widget, "was destroyed"

The <Destroy> event fires after the destruction has begun, so there's no
opportunity to veto the destruction.

Jeff
* See the documentation on "bindtags" for an explanation of why this is
  so




From bokr at oz.net  Thu Apr  3 15:52:42 2003
From: bokr at oz.net (Bengt Richter)
Date: 3 Apr 2003 20:52:42 GMT
Subject: Is there a python realaudio player? For windows?
Message-ID: <b6i6uq$rcs$0@216.39.172.122>

I recently wanted to play a .ram file without having to accept
all the baggage around the minimal player capability. Is there
such a thing? Even in C/C++, in case I got motivatged to write
a wrapper? TIA.

Regards,
Bengt Richter



From Jewlz408 at aol.com  Tue Apr  1 16:38:04 2003
From: Jewlz408 at aol.com (Jewlz408)
Date: 1 Apr 2003 13:38:04 -0800
Subject: I need help
Message-ID: <31ef0f43.0304011338.7ae3d45b@posting.google.com>

I need a lot of help
What do you know about Arrays
How many storage Locations will be allocated for each of the given arrays?
1) Dim A(1 to 10)
2) Dim B(-3 to 5)
3) Dim C$( 12)
4) Dim D( 1 to 4, -2 to 2)
5) Dim E(4,6)

And What does the following programs do with its output?

Dim Table( 1 to 5, 1 to 3)
For I = 1 to 5
????? For J = 1 to 3
????? Let Table(I,J) = 1
????? Next J
Next I
End

What does the program do?
And what is the output?


Thankyou so much for your time 
Please help me!

E-mail me with your answer please at
Jewlz408 at aol.com
Thanks again



From staschuk at telusplanet.net  Wed Apr 30 09:58:27 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 30 Apr 2003 07:58:27 -0600
Subject: running a PyUnit TestSuite
In-Reply-To: <b8nrlv$dvn$07$1@news.t-online.com>; from achim.domma@syynx.de on Wed, Apr 30, 2003 at 08:49:34AM +0200
References: <b8nrlv$dvn$07$1@news.t-online.com>
Message-ID: <20030430075827.A334@tibia.amotlpaa.bogus>

Quoth Achim Domma:
> I have a TestSuite which depends on the execution order of the tests. If I
> use the 'main' object provided by unittest.py I can not specify the
> execution order, so I set up a TestSuite by hand. But how should I run it? I
> found no easy way to pass a TestSuite for unittest.main. Any hint?

One way is to use the defaultTest argument to unittest.main:

    #!/usr/bin/env python
    import unittest

    class TestCase1(unittest.TestCase):
        def testFoo(self):
            self.assert_(1)

    class TestCase2(unittest.TestCase):
        def testBar(self):
            self.assert_(0)

    def suite():
        loader = unittest.TestLoader()
        s = unittest.TestSuite()
        s.addTest(loader.loadTestsFromTestCase(TestCase2))
        s.addTest(loader.loadTestsFromTestCase(TestCase1))
        return s

    if __name__ == '__main__':
        unittest.main(defaultTest='suite')

But I do wonder why your tests need a given execution order.  This
is contrary to the doctrine of unit tests, which holds that each
test should be independent.

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From martin at v.loewis.de  Wed Apr 23 01:29:16 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 23 Apr 2003 07:29:16 +0200
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <3EA5B173.9420324C@alcyone.com> <mailman.1051056037.8077.python-list@python.org> <3EA5DE1E.22704437@alcyone.com>
Message-ID: <m3r87tdbo3.fsf@mira.informatik.hu-berlin.de>

Erik Max Francis <max at alcyone.com> writes:

> but at this point it's looking awfully ugly.

That's why it will take a long time until Python warns about
assignments to True/False. Essentially, when Python 2.3 is considered
an old Python version (i.e. at Python 4.0), the warning may be
enabled. If the warning hasn't been seen for some time (say, at Python
6.2), it becomes an error. 

So the typical process is that people will have to actively break
backwards compatibility at some time by removing the assignment.
There is still a point in writing

try:
  True
except NameError:
  True, False = 1, 0

as this gives you the "proper" boolean objects in 2.3+.

People which want to be absolutely forwards compatible have two
options:

a) use some other names for internal boolean constants, e.g.

try:
  true, false = True, False
except NameError:
  true, false = 1, 0

b) import True, False from a module that doesn't get installed
   in Python 2.3+
try:
   from boolcompat import *
except ImportError:
   pass

   boolcompat would be shipped only for 2.1 and earlier.

Regards,
Martin



From ianb at colorstudy.com  Sat Apr  5 03:22:32 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 05 Apr 2003 02:22:32 -0600
Subject: Pythonic way of web-programming
In-Reply-To: <a3b5b608.0304040904.146dfff5@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
Message-ID: <1049530952.6019.34.camel@lothlorien>

I won't answer you point for point, but maybe talk about this
generally.  These things have been very much on my mind lately.

There are several options for Python web programming -- far more than
you list, actually, even when you take out the options that aren't being
actively developed.  There's many ways to categorize them, but here I
might say:

* Zope, a world of its own.  Certainly it could be a point of much 
  debate, but lots of people don't like Zope, often for the reasons they
  do like Python.
* Python in language-neutral or foreign language frameworks (like 
  Jython/J2EE, or ASP .NET -- Python works fine in pre-.NET ASP too).
* Minimal, i.e., cgi.
* Some from-scratch implementation of a Python web environment.  This is
  where the diversity is right now.

Zope had its chance at being the canonical Python web framework -- there
was a time when all the focus of using Python for web applications was
towards Zope.  It had 100% mindshare.  But it's not good at being a
general-purpose platform for web applications, which is why all sorts of
people have created alternatives.  And you have to create *something*,
because cgi doesn't cut it, you will want to work on a higher level.

Anyway, that said, we now have all this diversity.  Maybe some of it is
called for -- I'd certainly be unhappy if we had unity around an option
I didn't like.  But we have a lot of diversity that isn't useful too. 
What's the One True Templating Language?  I don't know.  But I don't
think we need more than four -- there are several valid and useful
different ways to do templating.  But there's also a limit.

The same is true in the from-scratch implementations.  There are some
valid differences between, say, Webware and Quixote.  But there's a
limit to that as well.

Right now, no, there isn't any Right Way to do Python on the web.  Zope
might be the closest thing, but there is no consensus.  I can't really
figure out what a consensus would be, but I imagine there's at least
some things we can come to consensus on... maybe underlying
infrastructure, or ways we can share code (like making templating
separate from the frameworks).  There's still a lot of work to be done
on that path, though.

  Ian






From riknessa at hotmail.com  Tue Apr 15 15:30:45 2003
From: riknessa at hotmail.com (Rikko)
Date: Tue, 15 Apr 2003 21:30:45 +0200
Subject: [Q] can PySerial work with a USB to serial adapter ?
Message-ID: <xZYma.26921$c42.17977@fe01.atl2.webusenet.com>

Hello python minded,
I'm trying to use PySerial with Python 2.2.2 on an XP system
with a USB to RS232 adapter.
Upon opening the port (making an instance of the Serial class)
the system return the error message "unable to open file",
thereby referring to the COM1 port used.

Is it ME who's goofing thing up or what?
With kind regards, Rik.






From sheershion at mailexpire.com  Mon Apr  7 19:33:12 2003
From: sheershion at mailexpire.com (Robert Amesz)
Date: Mon, 07 Apr 2003 23:33:12 -0000
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no>
Message-ID: <Xns9357101E083EFrcamesz@amesz.demon.nl>

Nils Grimsmo wrote:

> i have a problem getting wxPython to update the gui while the
> thread is running user procedures. i have tried calling
> wxWindow.Refresh() and wxWindow.Update() in various orders both on
> the wxFrame and the wxButton, but the gui doesn't update.

Have a look at wxApp::Yield() or ::wxSafeYield(). (I've used C++ syntax 
here because the docs do; Yield() is in a member function of your wxApp 
class instance, and wxSafeYield() is a function in the wxPython.wx 
module.) _Do_ read the docs, as you need to use those functions with 
some care.


Robert Amesz



From vlokk at hot.ee  Wed Apr 16 05:10:09 2003
From: vlokk at hot.ee (Vahur Lokk)
Date: Wed, 16 Apr 2003 12:10:09 +0300
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org>
Message-ID: <3e9d1cdd_2@news.estpak.ee>

Beni Cherniavsky wrote:

> Perhaps you should rethink your goals.  Why would you want to catch
> file modifications and how do you want to manipulate them?  Perhaps
> there are better ways?  What is your real goal?
> 
I wanted to build a simplistic? document management system. For example, 
I've got folder for letters out as well as project-based folders. I would 
save the letter in Letters Out folder and a symlink created in respective 
project folder etc. Programs that create such files are different - office 
suite, mail client (attachments coming in), browser (files from web).

I have learned programming some 15 years ago (BASIC) and wanted to re-learn 
it and am looking for some small project for a start. 
File system drivers do not sound like a newbie project, though. Polling file 
system for changes did come to my mind as a solution, but it seems like 
very resource-hungry way to do things.

Wahur



From bdesth at removethis.free.fr  Mon Apr 21 20:46:45 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 22 Apr 2003 00:46:45 +0000
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org> <QxAoa.540832$3D1.296696@sccrnsc01> <a5415304.0304210544.1342b436@posting.google.com>
Message-ID: <3EA490F5.30606@removethis.free.fr>

Ron Stephens wrote:
> Boa Constructor is, in my opinion, a great project. It will give
> Python an open source Delphi- or Visual Studio- like IDE with visual
> gui-creator.
> 
> PythonCard , in my opinion, goes beyond this in giving us an extremely
> easy to learn and use version of the same thing. Boa Constructor, like
> VS and Delphi, is quite complex. Learning Visual Studio or Boa
> Constuctor is a majopr undertaking all by itself. PythonCard, on the
> other hand, is extremely user friendly. I think PythonCard is
> revolutionary in empowering programmers who are not prefessional
> computer scientists.
> 

Yep... There is another aspect of PythonCard that, imvho, makes it more 
pythonic (than Boa), it's the fact that it's somehow a 'contained' 
environnement. I mean that there is no clear distinction between the 
environnement and the application, all the developpement tools being 
parts of the final application - even if hidden. It's somehow like a 
smalltalk environnement, where you modify an existing system until it 
fits the application needs.

(oh shit, I'm sure I've just said another stoopidity that will end up in 
a long cross-post with c.l.smalltlak now)

Laotseu




From mgerrans at mindspring.com  Sun Apr 13 17:36:31 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Sun, 13 Apr 2003 14:36:31 -0700
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <b7aj86$q9i$1@slb6.atl.mindspring.net> <j%9ma.9030$T34.282951@news2.tin.it>
Message-ID: <b7cl9v$63t$1@slb2.atl.mindspring.net>

"Alex Martelli" wrote:
> > Learning Python is an excellent book, but unfortunately quite out of
date.
>
> Yes to both.  Fortunately, a new version IS in the works, and indeed
> at quite an advanced stage...

That's great news!   I can't wait.    But don't worry, it won't stop me from
buying a copy of your recently released Python in a Nutshell -- which has
been garnering rave reviews -- in the meantime.  ;-)

- Matt





From list-python at ccraig.org  Mon Apr 14 10:25:15 2003
From: list-python at ccraig.org (Christopher A. Craig)
Date: 14 Apr 2003 10:25:15 -0400
Subject: Speed of string += string
In-Reply-To: <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
	<a9Zla.7472$T34.203357@news2.tin.it>
	<roy-06B012.20145112042003@reader1.panix.com>
	<bJ2ma.5447$883.4511@nwrdny02.gnilink.net>
	<iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net>
Message-ID: <t1w4r51upf8.fsf@kermit.wreck.org>

Mads Orbesen Troest <mads at troest.NEVERMORE.dk> writes:

> But is that /guaranteed/ (as it is with STL), or does it just happen to be 
> that way with the current implementation (however unlikely it is to 
> change)?

It's technically not guaranteed even in the current implementation.
All those operations are actually O(N), but that's a very unlikely
worst case.  

-- 
Christopher A. Craig <list-python at ccraig.org>
"Going to school make a person educated, any more than going to a 
garage makes a person a car" Slashdot




From BPettersen at NAREX.com  Fri Apr 18 05:01:07 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Fri, 18 Apr 2003 03:01:07 -0600
Subject: Problem in Accessing C++ Class Instance Global Variable (near success, help please!!)
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFD7@admin56.narex.com>

> From: Alan Pong [mailto:alanpong at hkstar.com] 
> 
> python 2.3, vc++6, win32 exe project
> 
> I has one global c++ class instance variable.  I want to call its
> member
> function through wrapper function which is accessed by python.
> However, as seen in wrapper.cpp, it give me another new variable copy
> so that phython actually accessing another instance.

First, PyImport_AddModule( "myext1" ), doesn't load the module, but
returns either an empty module or a module "myext1" that had allready
been loaded. It returns a PyObject that you're supposed to do something
with :-)

The symptoms imply that your "library" (wrapper.cpp) is most likely
compiled statically against both your program (test.exe) and your dll
(myext1.cpp). I.e. the dll is calling a method on an object that is
"named" the same but exists in the dll memory space, while your
application is reading an object (same "name") but in the executable's
namespace. Try extending the dll to read back the value from its object
to verify.

fix1: "Don't do that" <wink>.
fix2: pass the object as an argument.
fix3: make the object available through a module object both parties can
"import" (meaning you'll need the PyRun_* functions that take
namespaces.

-- bjorn





From pan-newsreader at thomas-guettler.de  Thu Apr 10 17:50:37 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Thu, 10 Apr 2003 23:50:37 +0200
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de>
Message-ID: <pan.2003.04.10.23.50.37.62608.750@thomas-guettler.de>

On Thu, 10 Apr 2003 22:57:17 +0200, Martin Bless wrote:


> To send mails from Python here's the example my provider provides: This
> http://faq.puretec.de/skripte/python/3.html
> 
> The relevant lines are these:
> 
> sendmail = os.popen("/usr/lib/sendmail -t", "w") sendmail.write(msg)
> sendmail.close()

popen can you give an stdout and stderr (Standard out/error) stream. Read
them after writing to stdin. Look at their content. I think if they are
empty all is fine. You can get the return value of the command with
wait(), too.

If you are using windows you can test popen with popen("dir").

  thomas




-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From skywalkerpackage at hotmail.com  Tue Apr 15 09:33:44 2003
From: skywalkerpackage at hotmail.com (Johannes Eble)
Date: Tue, 15 Apr 2003 13:33:44 GMT
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de>
Message-ID: <3e9be2ff.101896993@news.online.de>

Hello again,

thanks very much for your advice.

I am amazed that there are so many alternatives out there. After
thinking over the pros and cons I decided that I will follow the
PythonWin solutions. I haven't used PythonWin, but I have tried the
tutorial and the sample demo and I am impressed about what can be done
with the toolkit (also compared to TkInter which I am used to). The
Object Graphics Library (OGL) seems to be very nice. I haven't tried
PyUt, but I know that it is also based on PythonWin (It probably uses
OGL as well?) and I will try it too.

>From what I have found out about GTK+, it seems rather complicated to
me. This is also due to the pseudo object oriented structure in c.
PyGTK is probably better and more pythonic, but I found the wxPython
toolkit easier and cleaner.


Best regards,

Johannes







From andersjm at dancontrol.dk  Thu Apr  3 05:09:34 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Thu, 3 Apr 2003 12:09:34 +0200
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com>
Message-ID: <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>

"Drew Moore" <drew at astro.pas.rochester.edu> wrote:
> Howdy! 
> 
> I submitted a pre-PEP on overloadable assignment. 
> 
> The idea is:
> 
> In situations where the assignment token "=" appears, and an augmented
> assignment symbol (such as +=, -=, *=, etc..) would also be
> syntactically correct, the assigned-to object would be checked for an
> __assign__(self, other) method. 

You terminology is off.  There is no such thing as an "assigned-to"
object in Python.  Assignments change bindings, not objects.

Take a step back and tell us what problem you are trying to solve.
Whatever it is, I'm sure we can think of a better solution than having
assignment depend on whatever object, if any, happened to be
previously bound to the same name.

python-is-not-c++-ly y'rs, Anders





From mertz at gnosis.cx  Wed Apr  2 12:54:08 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Wed, 02 Apr 2003 12:54:08 -0500
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
 <b68jti$2mj66$1@ID-169208.news.dfncis.de>
 <RbCia.6782$9I.555309@news20.bellglobal.com> <3E8AFB77.38D0EDDF@engcorp.com>
Message-ID: <APyi+kKkXANI092yn@gnosis.cx>         

Peter Hansen <peter at engcorp.com> wrote previously:
|I just scanned a few tens of thousands of lines of our code, and as near
|as I can tell only 13 cases out of over 160 instances of "for" loops are
|using range() or xrange().

In my estimation, the enumerate() function introduced in 2.3 will cover
many of those 13 cases too.

In my own code, in the rare case I use range(len(lst)) for a loop, I
still generally index into lst within the body.  As in:

    for i in range(len(lst)):
        if i > ...:
            lst[i] = ...

Or that sort of thing.  I don't like the above construct especially, but
enumerate fixes it.

Admittedly, once in a while I really do want to loop over some integers
as such (i.e. not considered as indices).  But for that, range(N)
doesn't feel unnatural.

Yours, Lulu...

--
---[ to our friends at TLAs (spread the word) ]--------------------------
Echelon North Korea Nazi cracking spy smuggle Columbia fissionable Stego
White Water strategic Clinton Delta Force militia TEMPEST Libya Mossad
---[ Postmodern Enterprises <mertz at gnosis.cx> ]--------------------------





From a.schmolck at gmx.net  Sun Apr 13 18:58:48 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 13 Apr 2003 23:58:48 +0100
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net>
Message-ID: <yfsk7dydmxj.fsf@black132.ex.ac.uk>

Carl Banks <imbosol-1050078176 at aerojockey.com> writes:

First of all, I'd better clear up some potential misunderstanding: I *do not*
of course advocate that lexical scoping in python (or anywhere else, for the
matter) be ditched for dynamic scoping. If that was what you thought I was
arguing, I can't blame you for not taking it seriously and I think I should
have been clearer (at least in the last paragraph of my second post).

What I *am* claiming is that dynamic scope -- and not global scope in the form
currently implemented in python -- is the right thing to maintain implictly
shared state (cf. Common Lisp where global variables are either dynamically
scoped or constant whereas everything else is statically scoped, unless
explicitly declared otherwise -- which seems the right idea to me).

To avoid further confusion, I'll use global scope to refer to the module level
scope that is currently available in python and lexical and dynamic scope
according to their normal definitions and as one finds them in CL (and the
former also in python).



> Alexander Schmolck wrote:
> > Carl Banks <imbosol-1049990205 at aerojockey.com> writes:
> >
> > You did notice, of course, that you don't only need to change that particular
> > function foo, but also everything (directly or indirectly) calling foo? As
> > well as everything that is called (directly or indirectly) by foo and that
> > uses this global state information? Including 3rd party code and builtin
> > constructs of the language?
> 
> Yes.  Since the idea here is to rewrite the Emacs in Python, you have
> to rewrite all of this anyways.

I'm simply interested in the question as to whether globals of the type C,
Java and most languages, but specifically python, have them are a good idea or
whether dynamically scoped variables are superior. So I'd prefer not get too
sidetracked by Emacs and its uses of dynamic scope (some bad, some good, IMO).
The above describes by no means a problem limited to refactoring code in
dynamically scoped legacy languages but anyway, since you bring it up: there
is rewriting and rewriting -- not that anyone who's got the necessary
ressources would be stupid enough to rewrite emacs anyway[1] -- I'd personally
rather rewrite leveraging more rather than fewer powerful abstractions.


> 
> 
> >> I think dynamic scpoing is a wholly bad idea, and simplifying a very
> >> misguided programming practice doesn't make it any less so.
> > 
> > How much time did you actually spend thinking about global variables and
> > dynamic scope or what I wrote before you posted?
> > 
> > Why are global variables evil?
> [snip reasons globals are evil]
> 
> I didn't say globals are evil. I said Globals are Evil(tm). Big difference.
> (You should learn the meaning of (tm); it means the writer fully realizes
> that it's only a general truth, even though others consider it an absolute
> truth.)

I can't see were I attribute the claim that "globals are evil" to you -- *I*
happen to think globals of the broken kind python suffers from are evil (plain
and simple, no (tm)) and dynamic scope would provide a much better solution.
I'm always willing to hear why I'm wrong.

> I do not oppose global variables in general.

That was never in doubt. The question is not so much whether or not one should
have implicitly shared variables (in python and similar languages), but how
they should work and how they should be used. Your answer to my question on
your opinion on why python provides global variables seems to be (correct me
if I'm wrong):

I)  to provide *one*[2] implicitly shared state that should only be altered in
    an irrecoverable fashion.

My answer would be:

II) to provide a poor substitute for implicitly shared state that can be
    shared at different granularities and so that the same name can
    simultanously be exposed with different values to different values to
    different parts of the program -- all with automatic stack-popping style
    recovery of former values on natural or unnatural block exit.

 
> I do oppose global variables where there are situations where it feels
> necessary to change it and restore it.  That tells you right away that
> your functions should not use a global, but should have it passed in.
> Designing such a function to refer to the global anyways is Bad
> Programming(tm).

Repetition doesn't make this claim doesn't make it more or less true. Why not
instead back it up in some way or attack a provided example where I feel that
you *do* want to modify and restore state?

You say:

> These issues should be handled by global variables if they indeed are
> global in nature.  The stuff you mentioned usually is.

and then:

> I have not said lexical scoping handles the issues you have referred
> to better than dynamic scoping.  I have said that the issues you refer
> are only issues in the presense of Bad Programming(tm).  Anyone using
> globals discriminately will not have such issues to deal with.

I infer this to mean that it's OK to have logging and to use global state to
control logging, as long as I don't want one part of the program to be treated
differently. So to come back to an example from my previous post, a
particularly critical routine that should be logged at a higher level without
affecting the rest of the logging operations, would be Bad Programming(tm)?
Same for output precision, assertions and other debugging options (or
localized fine-tuning of e.g. floating point behavior or error handling etc.)?

Would you mind elaborating a bit? I do not yet fully appreciate why you'd
always want to share state either in a fully parameterized fashion or with
zero granularity and no shadowing or why my examples epitomize Bad and
indiscriminate Programming(tm).

> I have not said lexical scoping handles the issues you have referred
> to better than dynamic scoping.  I have said that the issues you refer
> are only issues in the presense of Bad Programming(tm).  Anyone using
> globals discriminately will not have such issues to deal with.

A bold claim. I bet a few people in the Lisp community would disagree that
their use of dynamic variables reflects their limited coding and design
abilities.

Do you ever use try-finally?

> 
> But to answer your question: yes, dyanmic scoping is so awful it
> doesn't justify this particular benefit, especially when the benefit
> only arises in Bad Programming(tm).

Indeed, wickedness should be its own reward -- if only we could agree that Bad
Programmers(tm) are the sole beneficiaries.

> In fact, dynamic scoping is so awful is doesn't justify any benefits.

Have you had traumatic childhood experiences with early perl or lisp dialects?
:)
 
> It's not just me; meny, many people regard dynamic scoping as a bad
> mistake.

[As explained, lexical scope as default for local variables is not questioned]

Well many, many people regard C as the ulitmate programming language and I
think those insightful individuals also have every right to regard dynamic
scope as an aberration, for I indeed C, to my knowledge, is almost utterly
devoid of anything resembling it (as long as you don't happen to shun any
posix pollutions like forking etc.).

Nonetheless as quite a few smart people also regard non-dynamically scoped
globals as a bad mistake, so I'd prefer to to see further support apart from
these arguments from authority (your's and the anonymous masses').

Also note that, as pointed out in Beni Cherniavsky's informative posting,
python already shows various signs of dynamic infiltration such as exceptions.

[I assume the following was written under the impression that I advocated
global scope for all variables, but I'll reply anyway just in case]
 
> I don't wish to go into detail here.  The main problems are: It forces
> you to use defensive naming schemes, even for your local variables.

Not if you have lexical variables as default. I guess you could even simply
adapt the existing scheme (where you have to declare globals explicitly to be
able to set them) for dynamic scope. Thus (for illustrative purposes):

def bar():
  global foo      # shares state, modifies last version of foo in according to
                  # call stack;
  global quux = 3 # shadows quux (or introduces a new dynamically scoped
                  # variable, but only visible to callees)

  my_var = 3      # no declaration => lexically scoped (no need to look beyond
                  # this block)

> You cannot determine the variables in effect just by looking at a function
> in context. You cannot do lexical closures. It's less efficient.

All equally true of globals as they are in python. BTW, not that it matters,
but dynamically scoped variables may well be more efficient than passing
parameters (the reason that it doesn't matter is of course that such design
decisions shouldn't be influenced by minor performance characteristics).

> 
> > Rather you
> > just pretended the issue didn't exist and I was just being stupid
> > for not realizing this to start with.
> 
> Don't be presumptuous.

I'm trying.

'as


Footnotes: 
[1]  That is not to say that writing a *different* and (noticeably) *better*,
     emacs-inspired editor wouldn't be a worthwhile undertaking.

[2]  Although with (arguably) different visiblility.




From marc at informatik.uni-bremen.de  Tue Apr 22 06:34:06 2003
From: marc at informatik.uni-bremen.de (Marc Recht)
Date: Tue, 22 Apr 2003 12:34:06 +0200
Subject: Twisted 1.0.4
In-Reply-To: <20030422013947.GB7666@frobozz.local>
References: <20030421110438.4991f713.itamar@itamarst.org>
 <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local>
Message-ID: <341410000.1051007646@leeloo.intern.geht.de>

> to Twisted and Twisted's users.  Similarly, Twisted has some C extensions
> which are typically built with gcc, but gcc's licence (GPL) also has no
> relevance to Twisted.
IIRC then gcc libs are LGPL'ed, so it doesn't affect Twisted. I wouldn't be 
so sure if they were GPL'ed.

Regards,
Marc

mundus es fabula




From aleax at aleax.it  Thu Apr 17 13:18:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 17:18:39 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050529915.4666.python-list@python.org> <_Fzna.27993$LB6.643892@news1.tin.it> <eppstein-F15082.09415717042003@news.service.uci.edu>
Message-ID: <PlBna.28604$LB6.655794@news1.tin.it>

David Eppstein wrote:

> In article <_Fzna.27993$LB6.643892 at news1.tin.it>,
>  Alex Martelli <aleax at aleax.it> wrote:
> 
>> Steven Taschuk wrote:
>>    ...
>> > trade-off as being between, say, an algorithm which runs in O(n)
>> > time but needs O(n^2) memory to do it, and an alternative which
>> 
>> I don't think there can be any such algorithm -- in O(n) time,
>> the algoritm cannot USE more than O(n) memory, so how could it
>> NEED it...?  The reverse, sure, but not this way, I think.
> 
> A simple example would be an algorithm that takes as input a sequence of
> values that are all (for whatever reason) guaranteed to be integers in
> range(n**2), and tests whether any two of them are the same:
> 
> def anysame(seq):
>     allocate array A, of dimension n**2, WITHOUT INITIALIZING IT
>         (is this possible in pure python?)

I don't think it is, but probably you could write a C-coded
extension that does this.  It would have to be a C-coded type
in order to be able to ensure returning the x-th item as a
valid Python object without having undergone any initialization,
though.  Perhaps on some architectures mmap applied to some
appropriate /dev/??? would let you do this (but I'm not sure).

>     for i,x in enumerate(seq):
>         if A[x] is an integer and 0 <= A[x] < i and A[A[x]] == x:
>             return True
>         A[x] = i
>     return False

Neat!  Takes a while to convince oneself of this, but, yes,
if "allocating N cells of uninitialized memory" exists as a
O(1) primitive, this will work.  I had not considered this
possibility when I expressed my doubts on Steven's words!


> Initializing A would require too much time, but the algorithm is capable
> of working with an uninitialized array.  The usual trick for reducing
> the memory requirements of such an algorithm is to replace the array by
> a hash table, and of course in Python the hash table (dictionary) based
> duplicate detection algorithm is much more preferable, but this makes
> theoreticians unhappy by turning worst-case time bounds into randomized
> expected time bounds.

Oh yes, absolutely.


Alex




From andy at wild-flower.co.uk  Thu Apr 17 18:10:15 2003
From: andy at wild-flower.co.uk (Andy Jewell)
Date: Thu, 17 Apr 2003 23:10:15 +0100
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
In-Reply-To: <h3qna.143255$vs.16245277@news3.calgary.shaw.ca>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <b7kqr8$jsq$1@panix2.panix.com> <h3qna.143255$vs.16245277@news3.calgary.shaw.ca>
Message-ID: <200304172310.15908.andy@wild-flower.co.uk>

On Thursday 17 Apr 2003 5:27 am, Parzival Herzog wrote:
> "Aahz" <aahz at pythoncraft.com> wrote in message
> news:b7kqr8$jsq$1 at panix2.panix.com...
>
> > In article <Wrena.33096$T34.876856 at news2.tin.it>,
> >
> > Alex Martelli  <aleax at aleax.it> wrote:
> > >Steve Holden wrote:
> > >> You'd do better to call it "Python for VB programmers", expecially
> > >> since the change to VB6 is so radical that migrating to a new and more
> > >> sensible language might be an attractive option to some!
>
> Maybe when you're done that, you could do "Python for RPG/400"
>
>      H*
>      H* RPG/400 VERSION OF THE BOTTLES PROGRAM *
>      H*
>      FSCREEN  O   F      80            WORKSTN
>      C                     MOVE 100       X       30
>      C           X         DOWGE0
>      C                     EXCPT
>      C                     SUB  1         X
>      C                     END
>      C                     SETON                     LR
>      OSCREEN  E
>      O                         X          3
>      O                                   26 'BOTTLES OF BEER ON THE'
>      O                                   31 'WALL,'
>      O                         X         36
>      O                                   53 'BOTTLES OF BEER'
>      O        E
>      O                                   22 'TAKE ONE DOWN AND PASS'
>      O                                   32 'IT AROUND'
>
> P.S. Can you believe it, at version 400 kicks at the can, they came up with
> this?
>
> Parzival

Hmmm...

When I look at Python compared to RPG now (after a little AS/400 training - my 
RPG course is the week after next) I begin to see that RPG is not as ugly as 
I at first thought.  It's worse than that - it's practically assembler code, 
but without the mnemonics! :-P .  What HAVE I let myself in for?

I'll take this to my course and throw it round a bit to see the reaction...

It's a pity I can't recode our legacy apps at work in Python/400 - I have 
toyed with it, but it's impractical to rewrite nearly 15 years of coding 
strata, with its roots in System 36 RPG, in Python...

It's a dirty job, but someone's gotta do it.  

regards, 
-andyj




From member27857 at dbforums.com  Wed Apr  9 02:47:08 2003
From: member27857 at dbforums.com (onefatman)
Date: Wed, 09 Apr 2003 06:47:08 +0000
Subject: sorting strings written to files
Message-ID: <2745895.1049870828@dbforums.com>

Can someone help me?

I have a small phone directory .dat file that has the entries
written to it as strings with three fields in each, separated by \t
and ended by \n.

e.g.
phbook.write("Joe\tNobody\t62907243\n")
phbook.write("Daniel\tStevens\t62962328\n")
phbook.write("Ellen\tWay\t62910758\n")
phbook.write("Jason\tGreig\t62314587\n")

it has a function that allows you to write new entries, but i need to
know how to sort them by first name, and then by last name when the
first names are the same.

please help.
Daniel

>>>>>>>>>>>>>>>>>>>>
Also, does anyone know how to delete details and change details? it
would be a big help. thanks.

--
Posted via http://dbforums.com



From martin at v.loewis.de  Sun Apr  6 03:40:10 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 06 Apr 2003 09:40:10 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <m3r88g9iqt.fsf@mira.informatik.hu-berlin.de>

pobrien at orbtech.com (Patrick K. O'Brien) writes:

> If you were crazy enough to think that you could create a program
> along the lines of Emacs, but written in Python, how would you go
> about it?  How would you design the domain model for files, buffers,
> windows, and frames?  How would you allow the same level of
> customizability?  How would you map functions (or methods or whatever)
> to keys?  Any thoughts?

This problem is underspecified, atleast for the actual implementation
strategy that follows: Do you need to support both terminal and GUI
mode? Multiple GUI libraries? Do you need API compatibility with Emacs?

> Here is the background.  I'm working on a file editor (called
> PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> PyCrust code base.  I've got a basic version working, and it supports
> editing multiple files and all the usual, basic features.  But I want
> it to be as customizable and extensible as Emacs.  In fact, I want to
> model it after Emacs, so that elisp code could be rewritten in Python
> and work with PyAlaMode.  To do that, I've got to support the same
> primitives as Emacs; expose files, buffers, regions, and windows the
> same way Emacs does; handle the Emacs keybindings and ability to remap
> keys, the same way Emacs does, etc.

I think you don't have much design choices then, if you want to expose
the very *same* functionality. In fact, you have to chose whether you
want to make it compatible with Emacs or with XEmacs, and you have to
specify a specific release you target, as each release has its own API
modifications.

I recommend that you don't; there are certain aspects in Emacs that
are there for historical reasons only and that should be dropped. My
favourite here is the Emacs Mule representation, which represents
characters in multi-byte form, using ISO 2022 escape codes to switch
between character sets. I recommend that you use Unicode
instead. This, of course, means that an entire suite of elisp cannot
be readily ported to your editor, but must be rewritten.

I would suggest a bottom-up strategy rather than a top-down strategy:
Take your favourite elisp package (say, gnus, or python-mode), and try
to make it run. 

To do so, write an automated translator that translates all required
elisp files to Python code. Don't make that translator
perfect. Instead, make it just translate the files you are looking at,
and extend it as you find it fails to translate existing elisp code.
Special-case builtin functions in the translator as appropriate
(e.g. 'require).

Make the translator translate all elisp code. You will then find that
the package doesn't work, because there are primitive functions which
are defined in C, not in elisp. Try to write these functions in
Python, using the Scintilla and Python API.

That approach will require constant refactorization, but it will tell
you early on what you can expect and what is impossible to achieve.

Regards,
Martin



From bjarke-news2003 at trebe.dk  Tue Apr  8 15:01:56 2003
From: bjarke-news2003 at trebe.dk (Bjarke Dahl Ebert)
Date: Tue, 8 Apr 2003 21:01:56 +0200
Subject: Loading a class the name of which is only known runtime
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net> <3BDka.4759$Du.16741@newsc.telia.net> <1ruvtw54py90d.1sob8ay5wclmb.dlg@40tude.net>
Message-ID: <d1Fka.7728$y3.941559@news010.worldonline.dk>

"Mads Orbesen Troest" <mads at troest.NEVERMORE.dk> wrote in message

> Actually, what I am looking for is more alone the lines of:
>
> moduleToImport = "SomeModule" # <- Retrieved as CGI argument
> classToInstantiate = "SomeClass" # <- Retrieved as CGI argument
>
> import moduleToImport
> someInstance = classToInstantiate()

Have a look at the builtin function __import__.

themodule = __import__(themodulename)
theclass = getattr(themodule, theclassname)
a = theclass()

There is a slight complication, though:
If themodulename is of the form "foo.bar", then module foo is returned. Some
themodulename.split('.'), and additional getattr calls should help.

Of course, there are all kinds of security issues with doing __import__ and
getattr like this in a CGI script...

Bjarke








From michaels at rd.bbc.co.uk  Wed Apr  2 07:20:50 2003
From: michaels at rd.bbc.co.uk (Michael Sparks)
Date: Wed, 02 Apr 2003 13:20:50 +0100
Subject: Best practice for object attributes?
Message-ID: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk>

Hello,


If I want to model objects with 3 attributes based on the behaviour of
python 2.2/2.3 I'ved noticed I can use either of the following idioms:

class shrubbery1(object):
   name = None
   knightowner = "ni"
   flowering = False

or:
class shrubbery2(object):
   def __init__(self):
      self.name = None
      self.knightowner = "ni"
      self.flowering = False

In both cases I gain objects which have the same 3 attributes.
Differences between
them I find are:
   * shrubbery1 has 3 class attributes as well, with the initial values
     indicated. This isn't the case with shrubbery2.

     >>> [x for x in dir(shrubbery1) if not x[0] == "_"]
     ['flowering', 'knightowner', 'name']

     >>> [x for x in dir(shrubbery2) if not x[0] == "_"]
     []

   * shrubbery2 has 3 entries added to self.__dict__ which act as the
     attribute stores. This is not the case with shrubbery1.

     >>> shrubbery1().__dict__
     {}
     >>> shrubbery2().__dict__
     {'flowering': False, 'name': None, 'knightowner': 'ni'}


Personally I find the former (shrubbery1) clearer, but I'm posting to
find out what people think is generally the better of the two. Do people
find any inherent advantages to one versus the other? (I note that the
shrubbery1 also makes the names of object attributes available to
metaclasses as well)

Thanks for any comments,


Michael.
-- 
Michael.Sparks at rd.bbc.co.uk    
British Broadcasting Corporation, Research and Development
Kingswood Warren, Surrey KT20 6NP

This message (and any attachments) may contain personal views
which are not the views of the BBC unless specifically stated.





From pan-newsreader at thomas-guettler.de  Fri Apr 11 03:25:53 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Fri, 11 Apr 2003 09:25:53 +0200
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de> <pan.2003.04.10.23.50.37.62608.750@thomas-guettler.de> <slrnb9bq9h.113.news@cipw08.physik.uni-bonn.de>
Message-ID: <pan.2003.04.11.09.25.52.584216.890@thomas-guettler.de>

On Thu, 10 Apr 2003 23:59:13 +0200, Carsten Gaebler wrote:

> In article <pan.2003.04.10.23.50.37.62608.750 at thomas-guettler.de>,
> Thomas Guettler wrote:
> 
>> popen can you give an stdout and stderr (Standard out/error) stream.
>> Read them after writing to stdin.
> 
> No, you can't read the stdout and stderr of a process spawned by popen()
> because its stdout and stderr are connected to the stdout and stderr of
> the calling process.

Sorry, I meant popen2 and friends

  thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From anton at vredegoor.doge.nl  Mon Apr 21 08:33:16 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 21 Apr 2003 14:33:16 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050857614.6299.python-list@python.org>
Message-ID: <b80q7v$cme$1@news.hccnet.nl>

Tim Peters" <tim_one at email.msn.com> wrote:

[this link, from different viewpoints]
 http://mail.python.org/pipermail/python-dev/1999-July/000467.html

>I argued in favor of adding generators to Python starting in about 1991.
>Every time I tried, the effort got hijacked by people going on about
>continuations.  The post's primary point was that generators *can* be
>explained "in terms of the old", and easily.  I think it succeeded at that.

Yes, the generator part was very good.

>Coroutines are harder, but can still be seen as a natural extension.
>Continuations require a different mindset entirely, and I believe that post
>was instrumental in (finally) getting generators taken seriously as a
>feature that didn't require buying into continuations first.

Yes, seen from this perspective the post was probably ok. If seen as
an explanation of continuations however, it is diminished by these
issues.

Anton




From wilk-spamout at flibuste.net  Fri Apr 25 13:24:30 2003
From: wilk-spamout at flibuste.net (Wilk)
Date: 25 Apr 2003 19:24:30 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
Message-ID: <87ptnalcc1.fsf@flibuste.net>

Alessio Pace <puccio_13 at yahoo.it> writes:

> Hi, I'd like to make a GUI for a program written in python, meant to run on
> mac, win and linux: which is the preferable solution? I saw around tkinter,
> wxpython and pyqt....

A web interface !

-- 
William Dode - http://flibuste.net



From dave at pythonapocrypha.com  Mon Apr  7 11:22:00 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Mon, 7 Apr 2003 08:22:00 -0700 (PDT)
Subject: How to protect python code ?
In-Reply-To: <u8bka.651$001.591065@monger.newsread.com>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org>
 <u8bka.651$001.591065@monger.newsread.com>
Message-ID: <Pine.LNX.4.51.0304070748470.29279@ns0.tbye.com>

On Mon, 7 Apr 2003, msw wrote:

> Jp Calderone wrote:
> > On Thu, Apr 03, 2003 at 09:35:49AM +0200, christophe grimault wrote:
> >>Before trying my own solutions, I wonder if someone already has a solution
> >>to this. We have been programming a quite large application in python and
> >>it contains many original algorithm and other things. The application will
> >>be sent to customers in a few month and since it is a commercial app, we
> >>need some protection against copy (illegal copies) and we need to make
> >>sure that critical parts of the code cannot be viewed and studied.
> >>
>
> I'm really interested in this as well, and I think that it's an
> important issue for python.  Unfortunately, as I'm sure you've
> discovered by now, it's a question that comp.lang.python is
> constitionally incapable of answering.  The only answers you will get in
> this group are:
>
> 1. intellectual property law is sufficient
> 2. no protection scheme is perfect

Not true - check the archives. In the past people, including myself, have
posted the general steps to do to "protect" the code. It's not worth the
effort in most cases, IMO, but answers *have* been given.

[snip]
> I think it ought to be relatively easy to distribute a hacked version of
> the python interpreter with your application that substitutes the
> standard "import" code with code that first looks for .pys files (or,
> what the hell, .tcl - let's not make this any easier) that are
> des-encrypted .py files.

It *is* relatively easy to do this - there are various ways to overload
imports, import from libraries, add new extensions and import handlers,
and you don't need a "hacked" interpreter - the simplest example of
running Python from C code will do.

> But unfortunately, if anyone's done this, you'll never find out about it
> on c.l.p.  :-)

Not true. What you *will* find on c.l.p however are many people who ask
for this but *appear* to misunderstand how much benefit it really gives
them as well as what costs are involved. Also, many of these people refuse
to do any investigation into how to do it. Even after given a few hints
they moan and groan because nobody will do it for them. I don't think the
OP was doing these things, but I do think that's why you don't see others
lining up to help.

For most programs, you're kidding yourself if you think the code is *that*
valuable. It has value, sure, but very few programs have some great secret
that needs to remain hidden. If you want to hide your code, use py2exe or
McMillian's Installer - that will keep out just about everyone without
adding too much complexity or maintenance costs. Beyond that, the law of
diminishing returns starts to win because pretty quickly you get into
guarding against people who are willing to invest large amounts of time
cracking it or, worse, people who crack for fun.

Even once you have the source code, there's still enormous work involved
before you can use it to generate revenue (to create your own knock-off
product). In the case of protecting the code to protect the revenue stream
(anti-piracy), you have to realize things like many/most pirates would not
have purchased your program anyway, your selling price may be too high,
having your customer relationship depend completely on the product itself
is a weak position and tough to defend, etc.

My two cents,
Dave




From jkraska at san.rr.com  Mon Apr 14 21:46:31 2003
From: jkraska at san.rr.com (Courageous)
Date: Tue, 15 Apr 2003 01:46:31 GMT
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>	 <mailman.1049530929.9225.python-list@python.org>	 <a3b5b608.0304051049.7048d76@posting.google.com>	 <mailman.1049571010.25625.python-list@python.org>	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>	 <mailman.1049653811.11491.python-list@python.org>	 <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
Message-ID: <qtom9vopignqfueuke26m2lbd1pn39mjcd@4ax.com>

>I think it's sad that Python has this slogan "there's only one way
>to do it", but that nowhere is this less true than web development.
>At Python-UK I opined that this might be holding us back in comparison
>to other languages and tools.

Not a bad point overall.

Of course, you're sort of leveraging the wrong quip. "There's only
one way to do it" applies to the *language*. It's never applied to
API's, really.

At the risk of boldly tossing incidental mud up in the face of
the entirety of all Pythonistas, I really don't think our community
has enough cohesion or clout to put together a uniform .NET/Web Services
style package with standards-establishing authority. We're not Microsoft,
after all, with a cool $billion to dispose of at a sneeze.

What you'll have to do instead is wait for the competitive feeding
frenzy to end and allow the technically superior platform to present
itself. This is both the bane and boon of open software in action,
IMO, and at a guess it very much applies here.

C//




From mhammond at skippinet.com.au  Fri Apr 25 07:40:14 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Fri, 25 Apr 2003 11:40:14 GMT
Subject: event IDs for win32evtlogutil.ReportEvent
In-Reply-To: <b8aq99$1l79$1@news.vol.cz>
References: <b8aq99$1l79$1@news.vol.cz>
Message-ID: <y89qa.10885$8K2.105071@news-server.bigpond.net.au>

Jaroslav Jakl wrote:
> Hello all,
> I need to write events to windows event log.
> In the win32evtlogutil is writen that win32evtlog.pyd can be source of
> messages,
> but i don't know what messages are there and what are their IDs.
> Can anybody help me.
> Thanks in advance
> Jaroslav Jakl

Check out win32evtlog.mc in the CVS source tree (or buy "Python 
Programming on Win32" <wink>.

Mark.




From ckea25d02 at sneakemail.com  Tue Apr 15 19:05:56 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 11:05:56 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <slrnb9oajn.493.rmunn@localhost.localdomain>
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <m2fzokfgni.fsf@mycroft.actrix.gen.nz> <slrnb9oajn.493.rmunn@localhost.localdomain>
Message-ID: <b7i2jt$170ph$1@ID-169208.news.dfncis.de>

Robin Munn wrote:
> Paul Foley <see at below.invalid> wrote:
>>
>>What if another thread (B) does the same thing, and changes sys.stdout
>>while this thread (A) is still inside the "try" block?  And if thread
>>A runs its "finally" clause and resets sys.stdout while thread B is
>>still in its "try" block, they screw each other.
> 
> This is the classic problem that arises when two threads need to use any
> shared resource, isn't it? 

No, actually it's the classic problem that arises when
two threads spuriously share something (a global variable)
that shouldn't have been shared in the first place.

This is *not* an argument for dynamic scoping. It's an
argument for explicitly passing things the state that
they need!

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From hupp at cs.wisc.edu  Thu Apr 10 17:44:27 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Thu, 10 Apr 2003 16:44:27 -0500
Subject: popen(), sendmail: Success?
In-Reply-To: <3e95d63f.2635484@news.muenster.de>
References: <3e95d63f.2635484@news.muenster.de>
Message-ID: <20030410214427.GA17320@upl.cs.wisc.edu>

On Thu, Apr 10, 2003 at 08:57:17PM +0000, Martin Bless wrote:
> 
> To send mails from Python here's the example my provider provides:
> This http://faq.puretec.de/skripte/python/3.html
> 
> The relevant lines are these:
> 
> sendmail = os.popen("/usr/lib/sendmail -t", "w")
> sendmail.write(msg)
> sendmail.close()
> 
> This works just fine. But is it enough? How do I know the operation
> succeeded?
> 
> I've read the docs about os.popen[234] and the module popen2 but I'm
> not sure where to go. And since experimentation isn't easy in this
> case I'd appreciate some hints or a concrete example very much.
> 
> Do I have to use popen2.popen3() and its wait() method? I have to
> admit that I don't understand too much of what's being said about this
> in the docs ... :-(

Basic popen should be fine.  To check for sucess use 

if sendmail.close():
   # Handle error....

since .close() returns None on sucess, otherwise returns a wait() type
error.  You might want to use smtplib instead though.  I don't know
what docs you were looking at but
http://www.python.org/doc/current/lib/os-newstreams.html was fairly
clear.

-Adam




From gh at ghaering.de  Thu Apr 24 22:02:04 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 04:02:04 +0200
Subject: Errors during xmldiff install on xp
In-Reply-To: <b4002bf3.0304241750.7b2a840c@posting.google.com>
References: <b4002bf3.0304241750.7b2a840c@posting.google.com>
Message-ID: <3EA8971C.6010905@ghaering.de>

M Ross wrote:
> Getting the following error msgs when trying to install the xmldiff
> module:
> 
> c:\Python22\include\pyconfig.h(37) : fatal error C1083: Cannot open
> include file: 'io.h': No such file or directory
> 
> running 
> 
>>python setup.py install
> 
> from the xmldiff directory.
> 
> io.h is on my machine-I have VS.NET installed. [...]
                                ^^^^^^

That might be the problem. Python is built with VC6 and distutils is 
geared towards VC6 (alternatively, Borland or mingw).

-- Gerhard




From tim.one at comcast.net  Sun Apr 27 23:55:05 2003
From: tim.one at comcast.net (Tim Peters)
Date: Sun, 27 Apr 2003 23:55:05 -0400
Subject: Dict handling of floats
In-Reply-To: <b94nn-i04.ln1@grendel.myth>
Message-ID: <LNBBLJKPBEHFEDALKOLCGEFOEEAB.tim.one@comcast.net>

[Jim Richardson]
>>> Linux, because eventually, you grow up enough to be trusted
>>> with a fork()

[Tim]
>> OTOH, I'm not sure anyone grows up enough to be trusted with a
>> float <wink>:
>>
>>     http://www.python.org/doc/current/tut/node14.html

[Jim Richardson]
> I figured it was something like that, It doesn't make any difference in
> my application, I just found it kind of odd.
>
> I have no real schooling in programming, all I learn is as I go,
> sometimes, stuff like this, trips me up.

Oh sure.  I hope you didn't read that as a dig!  I can testify that experts
screw up in the face of floating-point, and testify with confidence since
I'm one of the experts who so screws up <wink>.

> <Jim, who *still* doesn't grokk self.>

You will!  Floating point is doomed to remain esoteric to those who don't
make a concerted study of the topic, but self has a naturalness that will
eventually seem obvious.

self-is-what's-left-when-you're-not-looking-at-others-ly y'rs  - tim





From mcherm at mcherm.com  Fri Apr  4 09:20:50 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Fri,  4 Apr 2003 06:20:50 -0800
Subject: Zope style (was: tuning our site (was: help! advocacy resources needed fast))
Message-ID: <1049466050.3e8d94c2d048b@mcherm.com>

Dylan Reinhardt writes:
> Hello, <dtml-var fname>.  This is demonstration of MySoft version
> <dtml-var software-version>.  Just think, <dtml-var fname>, now 
> you can shift paradigms and think outside the box at 
> <dtml-var company> without leaving your office.
> 
> vs.
> 
> my_output = """Hello, %s.  This is demonstration of MySoft
>          version %s.  Just think, %s, now you can shift 
>          paradigms and think outside the box at %s without 
>          leaving your office.""" % (fname, version, fname, company)
> return my_output
> 
> --------
> 
> I find the former to be far more readable and more easily maintained...
> particularly as the size of the sample increases.  Inserting additional
> variables or moving around existing ones is easy in the first sample and
> must be done carefully in the second.

I think you just need a minor style change in your python code. I find
the following about as readable as your DHTML code:

  my_output = """Hello, %(fname)s.  This is demonstration of MySoft
           version %(software_version)s.  Just think, %(fname)s, now 
           you can shift paradigms and think outside the box at 
           %(company)s without leaving your office.""" % nameMap
  return my_output

Of course, PEP 292 makes this marginally better:

  my_output = """Hello, $fname.  This is demonstration of MySoft
           version $software_version.  Just think, $fname, now 
           you can shift paradigms and think outside the box at 
           $company without leaving your office.""".sub(nameMap)
  return my_output

Now that's really quite readable and maintainable.

-- Michael Chermside





From glenfant at NOSPAM.bigfoot.com  Tue Apr  1 07:44:11 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Tue, 1 Apr 2003 14:44:11 +0200
Subject: HTMLView widget for Tkinter
References: <mailman.1049197502.20990.python-list@python.org>
Message-ID: <b6c0rm$1o95$1@biggoron.nerim.net>

"Egor Bolonev" <ebolonev at front.ru> a ?crit dans le message de news:
mailman.1049197502.20990.python-list at python.org...
> Hello all,
>
> I need to show html e-mail in my Tkinter program. Is it possible?
>

There's no HTML panel in Tkinter like wxPython (and perhaps others)
provides. So you hould hard code HTML rendering in a Tkinter panel yourself.
This is difficult but possible.
You could get enlightenment on rendering HTML in a Tkinter panel hacking
Grail.

http://sourceforge.net/projects/grail

Note however that rendering HTML  from in a wxPython widget is easier.
There's already a Python/wxPython mail client at sourceforge (don't remember
its name)

--Gilles




From cpl.19.ghum at spamgourmet.com  Wed Apr 30 06:04:53 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Wed, 30 Apr 2003 12:04:53 +0200
Subject: HTTP application streams a response to the client gradually - how   deal with it?
References: <Xns936CB5697A6D4cpl19ghumspamgourmet@62.153.159.134> <3EAEC97C.1EADEBF7@hotmail.com> <Xns936D814A89Fcpl19ghumspamgourmet@62.153.159.134> <3EAF956F.CF1D9A26@hotmail.com>
Message-ID: <Xns936D7A9DFE6E5cpl19ghumspamgourmet@62.153.159.134>

Alan,

> I.E. You want to be able to read the user input channel 
> at the same time as retrieving updates from the server. And because 
> urllib and  httplib encapsulate the entire request/response inside 
> one method call, your client is blocked while the server call is 
> executing, and thus very unresponsive?

Not exactly the problem ... as you say, that could be easily solved with 
threading.

The exact problem is:

the data has to be processed one chunk at a time. Without the response 
being complete.

Example: The chat server sends:

Bart says: nice to see you all here
Miriam says: yes, I also enjoy the talk
Lisa says: cool evening, yeah
...

the response channel stays open. All the time. When I wait for it to 
finish, so that I can start processing the Response-Objekt - no good.

And while the responses get in from the server, I also have to do 
additional POST requests

POST user=Icke says="i also think it's a nice evening"

> There is a simple solution to that problem: use multiple threads. One
> thread reads input from the user, another thread does nothing but
> contact the server periodically for updates.
Yeah, that would be perfect. But: it seems that the server does not 
finish his response. So until httplib.read() has a timeout ...

> Given what you've said, I'd say it's probably overkill to go down to 
> the socket level. 
I surely hope so, Alan. Posting this question is based on my thinking, 
that I can't beat Timbot or Guidocode by any means - and better use it.

> functionality you need, and (mostly) bug free, having been used by
> thousands of people over the years. Rolling your own HTTP protocol
> would be an enormous amount of work, due to the wide range of weird and
> wonderful ways in which HTTP is used and abused.

Yeah ... and I think I have just stumbled into a very abusive one.

Thanks for all your feedback... maybe there is an easier way?

Harald





From bvdpoel at kootenay.com  Mon Apr 21 14:22:07 2003
From: bvdpoel at kootenay.com (Bob van der Poel)
Date: Mon, 21 Apr 2003 11:22:07 -0700
Subject: Locale settings
Message-ID: <3ea437a1_1@dns.sd54.bc.ca>

I've got a minor problem with locale settings and tkinter. My program 
has done a test to determine the current locale setting for the user:

loc = locale.getdefaultlocale()
if loc[1]:
   userEncoding = loc[1]
     try:
        ''.encode(userEncoding, 'replace')
     except:
        userEncoding = ''

if userEncoding == '':
   userEncoding='ascii'

Now, for the most part this works. However, one user of my program has 
just upgraded to Linux Mandrake 9.1 and (I think) he has set his 
language to "de". In any event, the getdefaultlocal() seems to return 
'de' which isn't a valid python encoding (??) and the above code then 
converts to his 'ascii'.

Strings entered into entry boxes, etc. are converted to a useable 
encoding with:

if type(s) != type(''):
   s=s.encode(userEncoding, 'replace')

But, with the 'ascii' encoding accented characters are converted to '?' 
(which is as expected). Now, on my system, with language set to 
'english' I get a locate setting of ISO8859-1. And this does display 
accented characters properly.

My question, after all this, is what to tell my German user as to the 
correct way to set his locale so that these characters display. Or, if 
my code is not correct, a proper manner to do all this. BTW, this is all 
done to convert the unicode strings sometimes returned by tk-entryboxes 
to normal, 8bit strings, which can be sorted, etc. ... avoiding the dreaded:
	
	UnicodeError: ASCII decoding error: ordinal not in range(128)

Thanks.


-- 
Bob van der Poel ** Wynndel, British Columbia, CANADA **
EMAIL: bvdpoel at kootenay.com
WWW:   http://www.kootenay.com/~bvdpoel





From vjb at vuit.com  Thu Apr 17 19:40:30 2003
From: vjb at vuit.com (Vince Buonassisi)
Date: Thu, 17 Apr 2003 16:40:30 -0700
Subject: error in 
Message-ID: <F9B97989-712D-11D7-9F51-000A95682CD2@vuit.com>

hi,

i am getting the following error when i load my test script in my web  
server:

Mod_python error: "PythonHandler reqhandler"

Traceback (most recent call last):

   File "/usr/lib/python2.2/site-packages/mod_python/apache.py", line  
181, in Dispatch
     module = import_module(module_name, _req)

   File "/usr/lib/python2.2/site-packages/mod_python/apache.py", line  
335, in import_module
     module = imp.load_module(mname, f, p, d)

   File "/usr/local/apache/htdocs/reqhandler.py", line 9, in ?
     from xml.dom.ext.reader.Sax2 import FromXmlStream, FromXml

   File  
"/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/__init__.py",  
line 21, in ?
     encoder = codecs.lookup("utf-8")[0] # encode,decode,reader,writer

   File  
"/usr/src/build/143041-i386/install/usr/lib/python2.2/encodings/ 
__init__.py", line 51, in search_function

RuntimeError: cannot unmarshal code objects in restricted execution mode

the script i am running is very simple.  i just send a text message to  
the browser.  i am using the following versions of software:

apache 1.3.26
python 2.2.1
mod python 2.7.8
PyXML 0.7.1
linux redhat 8

any help would be appreciated.

thanks

vince
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 1304 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030417/aa17c492/attachment.bin>

From peter at engcorp.com  Wed Apr  2 19:16:09 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 19:16:09 -0500
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
	        <b68jti$2mj66$1@ID-169208.news.dfncis.de>
	        <RbCia.6782$9I.555309@news20.bellglobal.com>
	        <mailman.1049298845.7655.python-list@python.org>
	        <3E8B4094.F61515B3@engcorp.com>
	        <ZpHia.23$Sd.4@fed1read01>
	        <3E8B48E0.C786D192@engcorp.com>
	        <mailman.1049316186.31356.python-list@python.org>
	        <3E8B55B2.D8D7D2F2@engcorp.com>
	        <mailman.1049321730.7572.python-list@python.org>
	        <nFKia.141$Sd.25@fed1read01> <mailman.1049328065.20513.python-list@python.org>
Message-ID: <3E8B7D49.829034FF@engcorp.com>


Skip Montanaro wrote:
> 
>     Tim> I think that nearly this exact use case was discussed here:
>     Tim> http://mail.python.org/pipermail/python-dev/2003-March/034272.html
>     Tim> . If you go forward a few messages, GVR suggests adding "from
>     Tim> __builtins__ import open" to modules that you might want to fiddle
>     Tim> with.
> 
> Yeah, but that's quite different from Peter's use case, where he wants to
> unit test a module more-or-less from the outside.  He shouldn't have to add
> hooks in the module itself to unit test it.
> 
>     Tim> This made me recall that the proposal was slightly more lenient
>     Tim> than I initially remembered:
> 
>     Tim> some_random_module.open = my_open
> 
>     Tim> would be OK if open is already shadowed in some_random_module. The
>     Tim> compiler then can figure out that optimizing access to open is not
>     Tim> going to work.
> 
> That won't help Peter either.  I suspect all of the modules he wants to test
> don't actually shadow open().

Very true.  There is definitely NO desire to have to munge the internals
of the code under test, just to let it be tested.

I hope Greg's comments about being able to disable this are on the mark,
because otherwise this would be a clear case of a desire for (unnecessary,
IMHO) optimization being given higher priority than maintaining one of
Python's strongest advantages.  Speed will *never* be Python's strongest
suit, but its dynamicism clearly is; let's not bugger that up!

-Peter



From mcherm at mcherm.com  Wed Apr 30 12:35:02 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Wed, 30 Apr 2003 09:35:02 -0700
Subject: Problem subclassing tuple
Message-ID: <1051720502.3eaffb367a905@mcherm.com>

John Wilson writes:
[This doesn't work:
> class Holder(tuple):
>     def __init__(self, *values):
>         tuple.__init__(self, values)

    [...]

> class Holder(tuple):
>     def __new__(cls, *args, **kargs):
>         return tuple.__new__(cls, args)

> is the way to do it for a tuple :))))))

   [...]

> I'm really quite interested in why this doesn't work. I *thought* I'd
> grasped the concept of the new style classes but this seems to demonstrate
> that I haven't!

Actually, you've discovered the original reason for the invention
of __new__. Guido invented __new__ because __init__ wouldn't work
for classes like tuple that were immutable. The combination of __new__
and __init__ is what is called "two stage initialization" -- the 
__new__ method is the "factory", that just creates the empty object,
while __init__ is the "initializer" that sets up the object. Normally,
you want to override __init__ -- the only exceptions are those cases
where you NEED __new__. For instance, if you want to return objects
of different types depending on some parameters, then you'll need to
use __new__ (or just use a factory function instead of a constructor!).
Or if you want to "initialize" an immutable object, then it'll have
to be done in __new__ (that's your situation).

For more about it, including some discussion of why two stage
initialization is so good, see 
http://cocoa.mamasam.com/MACOSXDEV/2003/03/1/58013.php (and I thought
I'd heard a discussion of the subject in Python recently, but couldn't
find a link).

-- Michael Chermside







From andrew.henshaw at gtri.gatech.edu  Sun Apr 27 18:33:20 2003
From: andrew.henshaw at gtri.gatech.edu (Andrew Henshaw)
Date: Sun, 27 Apr 2003 18:33:20 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <ZFyqa.7824$3M4.220335@news1.tin.it>
Message-ID: <vaomliqpl26u6c@corp.supernews.com>

<posted & mailed>

Alex Martelli wrote:

> Klaus Meyer wrote:
> 
>>> I don't know what to DO about it, but some minor tweaks might help.
>> 
>> I also have made this mistake sometimes.
>> 
>> On reason, i think, is the mix of function-call and values, example:
>> 
>> import time
>> print time.time()
>> print time.timezone
>> 
>> You have to remember the doc of the module carefully to make no mistake.
> 
> Well, in this case it's no big deal, actually --
> 
>>>> print time.time
> <built-in function time>
> 
> i.e. if you forget parentheses you get an immediate reminder of that.
> 
> 
This reminds me of a recent bug that bit me for a few minutes:

import time

endTime = time.time() + RUN_TIME
while time.time < endTime:
    do_some_processing()


and do_some_processing is never called.

--Andy



From gustav at morpheus.demon.co.uk  Sat Apr  5 05:35:00 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Sat, 05 Apr 2003 11:35:00 +0100
Subject: output buffering--am I missing something obvious?
References: <yu99istu9ds9.fsf@europa.research.att.com> <16013.63756.223973.484125@montanaro.dyndns.org>
 <mailman.1049492108.15272.python-list@python.org>
Message-ID: <8yupdygb.fsf@morpheus.demon.co.uk>

Andrew Koenig <ark at research.att.com> writes:

> Skip> I tried using 1.5.2 (!) on a Solaris 8 machine (it's what comes on the
> Skip> Software Companion CD for Sol8) I get the same results:
>
>>>>> import sys
>>>>> sys.stdout.write("foo")
> foo> sys.stdout.flush()
>>>>> sys.stdout.write("bar\n")
> Skip> bar
>
> Skip> Bit rot somewhere?  It certainly shouldn't do that.
>
> In a follow-up message, I suggested that interactive input tossed
> unwritten output.  I'm now not so sure it's that simple, because if I
> do this:
>
>>>> sys.stdout.write("foooooooooooooooooo"); sys.stdout.flush()
>
> it appears to write the argument string *and then to backspace over it*,
> erasing the characters from view.
>
> Very odd.

Readline (or lack of readline) strangeness...?

Paul.
-- 
This signature intentionally left blank



From sholden at holdenweb.com  Wed Apr 16 17:06:10 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 21:06:10 GMT
Subject: Web Hosting
References: <2rjna.171$hT2.182473@news2.news.adelphia.net>
Message-ID: <6Bjna.111823$It5.95226@news2.central.cox.net>

"Bob X" <bobx at linuxmail.org> wrote in message
news:2rjna.171$hT2.182473 at news2.news.adelphia.net...
> Can you name a good (reasonably priced) Web Hosting company that supports
> Python (either CGI or mod_python)?
>

See http://www.python.org/cgi-bin/moinmoin/PythonHosting

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From revanna at mn.rr.com  Wed Apr  2 19:15:13 2003
From: revanna at mn.rr.com (Anna)
Date: Thu, 03 Apr 2003 00:15:13 GMT
Subject: A few errors in "Python in a Nutshell" - Numeric related
References: <b6frd5$uq$1@peabody.colorado.edu>
Message-ID: <pan.2003.04.03.00.14.30.217865@mn.rr.com>

On Wed, 02 Apr 2003 16:23:16 +0000, Fernando Perez wrote:

> Hi,
> 
> I recently got Python in a Nutshell, and as I expected, I've found it to
> be extremely informative and well written.  I haven't read it in detail,
> but I'm glad I got it.
> 
> However, I'd like to point out a few problems I found in Chapter 15,
> Numeric processing.  I did not read it in excruciating detail, so please
> don't consider this an exhaustive bug hunt of that chapter.  And by the
> way, that chapter is a fabulous reference to NumPy, far better for most
> things than the standard manual which is organized in a 'cute' (read
> random) way and for some things is nearly useless.  Ch. 15 actually
> explains many things which the NumPy docs gloss over, and the function
> reference is *alphabetized*!! (how hard would it be to do that in the
> normal manual??)
> 
> Anyway, my few nits for future reference:

<snip>

> Anyway, these are just some small nits which readers which use Numeric
> may want to be aware of.  If there is an errata page for the book, it
> may be worth adding them to it.

Alex Martelli is at Python-UK.

You might wish to submit this to the online errata form. Here's the url
for doing so.

http://www.oreilly.com/cgi-bin/errata.form/pythonian

cordially,
Anna



From andersjm at dancontrol.dk  Thu Apr 10 06:01:03 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Thu, 10 Apr 2003 12:01:03 +0200
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1049901964.20840.clpa-moderators@python.org> <b72fo5$bq6$1@panix1.panix.com>
Message-ID: <3e9540c1$0$31939$edfadb0f@dread12.news.tele.dk>

"Aahz" <aahz at pythoncraft.com> wrote:
> 
> So how does Capon differ from SCons?

SCons has lots of features.  The main thrust of Capon is on providing
the means to add features.  Capon is a language by itself, and Capon
scripts are not Python scripts.  That sounds complex but isn't, the
language is just Python in disguise.

It's not inconceivable that I could find a straightforward way to
bridge Capon and SCons, providing "Caponic" style access to SCons
features.  It's not inconceivable that someone else could, either,
given the extensible nature of Capon.

I'm not going to attempt a comparison with SCons, I don't know SCons
well enough for that, and besides I'm taking Bjarne Stroustrup's
advice on comparing languages.  Instead let me illustrate what it is
that sets Capon apart.

Take a look at this implementation of 'clean', a function to delete
intermediary files.

| import os  # import a Python module
| .clean():  # lazy notation; think of this as "def clean(self):"
|     # dependency_tree() is a generator that lists dependencies recursively
|     for dep in .dependency_tree():
|         dep.clean_this() 
| 
| .clean_this():
|     pass
|
| .file.gen/clean_this():
|     if .exists():
|         if .verbose:
|             print "clean() deleting %s" % (.filename,) 
|         os.unlink(.filename)

'clean' is a simple function that walks a generator and calls
'clean_this' on each node.  The nodes typically represent files but
don't have to.  Except for implicit self argument lists and function
bodies are straight Python.

'clean_this' comes in two incarnations: The simple ".clean_this()"
that doesn't do anything, and ".file.gen/clean_this()" which deletes a
file.  It's what I call a category qualified function.  This means
that clean_this will do different things for nodes in different
categories.  For example, if you call clean_this on a node in category
"file.gen.obj.c", which would be a typical category for an object file
which is compiled from a C file, then the second version will be
called, and the file will be deleted.  

Calling clean_this on a source file in category "file.source.c" will
call the first version and do nothing.

This is the clean and clean_this functions as I have written them.
They happen to be a part of the distribution, but you could just as
well have written them in your own scripts.  Either way, you can
override their behaviour for specific nodes, and in more than one way.
By specialising a category:

| .file.gen.map/clean_this():
|        pass # never delete map files

Or by overriding for a single node or group of nodes:

| Gen('systemfile.lib'):  # a node representing the systemfile.lib file
|       .clean_this():    # overrides inherited versions
|               send_email(to="BOFH", subj="Delete systemfile.lib, please?")

All top-level function are automatically available from the command
line.  So you can write:

  cap .clean()

to run the 'clean' function on the default node.

More interesting than 'clean' is the 'compile' function, which serves
the same purpose as build commands in a makefile; but you can read all
about that in the tutorial.  I hope I've peaked your curiosity and
you'll give it a try.

caponically-y'rs, Anders





From staschuk at telusplanet.net  Thu Apr 17 15:06:53 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 13:06:53 -0600
Subject: How many of you are Extreme Programmers?
In-Reply-To: <1147e466.0304160630.5b510724@posting.google.com>; from blunck@gst.com on Wed, Apr 16, 2003 at 07:30:46AM -0700
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <20030417130653.B7968@tibia.amotlpaa.bogus>

Quoth Christopher Blunck:
  [...]
> The natural response a non-Python programmer has to this statement is,
> "what?!  you __trust__ the programmer?!  that doesn't work in most
> environments."  Despite the ignorance this response demonstrates, I
> none-the-less thought about the statement and the answer I came up
> with was "we write lots of really thorough tests to demonstrate
> functionality."

There is another point:  Although there are indeed lots of
programmers who do not deserve this trust and will betray it every
time, the feeling among Pythonistas is, I think, that (a) this is
not sufficient justification for hamstringing the competent, and
(b) no language feature will make an appreciable dent in the
incompetent's ability to screw everything up.

Dealing with incompetent programmers is an organizational problem,
not a language design problem.

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From pinard at iro.umontreal.ca  Sun Apr  6 15:43:55 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 06 Apr 2003 15:43:55 -0400
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2u1dcw22s.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <oqvfxrh0n8.fsf@alcyon.progiciels-bpi.ca>

[Patrick K. O'Brien]

> If you were crazy enough to think that you could create a program along
> the lines of Emacs, but written in Python, how would you go about it?

Hello, Patrick, and all pythoneers. :-)

I once was crazy enough to at least ponder doing something like this,
being conscious of the huge amount of work such a project might represent.
I did starve for an Emacs environment scriptable in Python, but also was
aware of the massive amount of knowledge in the Emacs distribution and
project, and the rapid pace of Emacs development.  This would have been
fairly difficult to follow and cope with in the long run.  No doubt to me
that seriously tackling such a project could have swallowed me not only
whole, but for a good part of my life.  I still felt too young to die! :-)

Pymacs seemed to me like a much more acceptable compromise, as it could
be written relatively quickly, giving me a good part of all I wanted.
Oh, I'm surely not satisfied with the complexity of the mix, Emacs plus
Python just cannot be simpler than a pure Python solution would have been,
at least seen from a reasonable distance -- another planet, say? :-).
And moreover, there is a cost to the communication protocol between Emacs
and Python which is better not be abused if you want good response times.

> Sound like fun?

In a word, I would love to see a pure Python implementation of Emacs,
but have the prejudice that it requires an extra-ordinary involvement,
making that goal unreachable in practice for most mortal programmers...
So, even if Pymacs is far from ideal, for me, it has been a blessing.

All this being said, I sincerely wish that your project will succeed!

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From hancock at anansispaceworks.com  Thu Apr 17 13:19:00 2003
From: hancock at anansispaceworks.com (Terry Hancock)
Date: Thu, 17 Apr 2003 10:19:00 -0700
Subject: Zope REQUEST question
Message-ID: <OG2dnWc4TKANfAOjXTWcrg@august.net>

Okay, this is a little tricky to explain.  This is technically a Zope 
question, but it has to do with programming products in Python.

I'm doing a Forbidden Thing with a Zope product, which is overloading its 
__getattr__ function.  It's a necessary Evil, though. ;-)

And actually, it works really well, as far as it goes.  It *is* a nightmare 
to debug, which is why I've moved nearly all of the code out of this part 
so it can be tested independently.

Anyway, it turns out, I need to be able to access the REQUEST object
from there (I'm trying to implement an HTTP_REFERER block).

So far, I've tried:

getattr(self, 'REQUEST')

and

ObjectManager.__getattr__(self, 'REQUEST')

which should do the same thing, except the ObjectManager.__getattr__ 
is much safer (it's not recursive).  I'm not sure why I thought that REQUEST
would be an attribute of the object at this point, but I'm pretty sure I 
read this somewhere and didn't just think it up out of thin air. However, 
in this case at least, it's wrong -- these both give me attribute errors 
(at least my object returns "not found" which is what happens when you
raise an attribute error from inside of __getattr__ -- other errors usually
give you the traceback you'd normally expect).

Does anyone have a better idea for how to get the REQUEST?

Cheers,
Terry

-- 
Terry Hancock
Anansi Spaceworks http://www.AnansiSpaceworks.com/




From bokr at oz.net  Thu Apr 17 13:35:23 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 17:35:23 GMT
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net> <3E9ED903.79B5F164@sympatico.ca>
Message-ID: <b7mokr$br$0@216.39.172.122>

On Thu, 17 Apr 2003 12:40:35 -0400, "Colin J. Williams" <cjw at sympatico.ca> wrote:
>user at domain.invalid wrote:
[...]
>> What do the back ticks in method "__str__" mean?
>
>The back ticks, as in `self.value`,  are equivalent to repr(`self.value`).
ITYM ...                                                repr(self.value).

>
>ie. it provides the  ``official'' string representation of self.value.
>
>Colin W.
>
>

Regards,
Bengt Richter



From lorenzo at mysurname.net  Sun Apr 27 20:12:06 2003
From: lorenzo at mysurname.net (Lorenzo Bolognini)
Date: Mon, 28 Apr 2003 00:12:06 GMT
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <ql_qa.18612$K35.626741@news2.tin.it>

"Steven Hodgen" <steven at twitch.net> ha scritto nel messaggio
news:oproar1sjtx3zh5c at news.mindspring.com...

> However, Java is not a great language for beginners.  Python, on the
other
> hand, is ideal.  If I can find a textbook that is even ok, I may
consider
> changing, since the language is so much better for students at this
level.

I'm a beginner and I liked very much

Practical Python
By Magnus Lie Hetland

And you could find the tut's at this page useful for the very
beginning
http://hetland.org/python/

Lorenzo





From webmaster at quanta1.world--vr.com  Tue Apr  1 23:52:52 2003
From: webmaster at quanta1.world--vr.com (David Sfiligoi)
Date: Tue, 1 Apr 2003 23:52:52 -0500
Subject: nntplib, huge xover object
Message-ID: <ZCidndFDJdNGwBejXTWc3g@giganews.com>

Hi,

I built a small script that use the xover function in the nntplib module.   
The problem that I came across is that what xover can return a huge tuple 
when there is 1000s of article in a newsgroup(which is frequent)

testxover_resp,testxover_subs = s.xover(start,end)

On my system it's not an issue I have 768 Mb of RAM.... but I have to 
believe that there is a way to optimise this while keeping all this simple.

How can I limit the amount of memory xover would take. the other I did an 
xover of a huge group and the python process was taking about 650Mb Res. 
memory.  

Then I go on .and put  the articlenumbers and subjects into a massive 
dictionary which I agree it's not  a very inteligent way to do this...but 
right now it works, I will later get rid of this huge dictonary and replace 
it with something more memory efficient.

maybe somebody can teach me a technique in python to deal with object that 
may take lots of memory? Maybe I can just put the tuple returned by xover 
into a file(I don't have a clue how)?

Dave






From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Mon Apr 28 15:18:32 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Mon, 28 Apr 2003 21:18:32 +0200
Subject: saving output to file
In-Reply-To: <MUera.658292$L1.187296@sccrnsc02>
References: <MUera.658292$L1.187296@sccrnsc02>
Message-ID: <3ead7e89$0$49116$e4fe514c@news.xs4all.nl>

Eric Anechiarico wrote:
> I am still not as proficient as I would like to be in Python, so this may be
> a trivial question to most.
> 
> I want to have the output I get in my IDE window (using MacPython 2.2.2
> primarily in Carbon) to be written to a file instead of having to copy the
> output and then paste it to a text file myself.  How would I go about doing
> this?
> 
> Mainly I would be using the code to do this with programs that calculate
> outputs of large numbers (primes, Fibonacci, etc.) and they sometimes
> overflow the buffer and I cannot get back very far to view the beginning.
> Make sense?  Basically since I am doing this in the IDE window anyway.
> 
> Any suggestions?

I'm guessing that you are now "print"-ing the results of your calculations
to the screen, as in:

def fib(n):
    a,b = 0,1
    while b<n:
        print b,
        a,b = b,a+b 

Instead of printing them to the screen, print them to a file:

def fib(n, resultfile):
    a,b = 0,1
    while b<n:
	print >>resultfile, b,
        a,b = b,a+b 

fib(2000, open("fibonacci_to_2000.txt",'w'))


You might also want to check the following parts of the manual:
http://www.python.org/doc/current/tut/node9.html#SECTION009200000000000000000
http://www.python.org/doc/current/lib/module-StringIO.html

If you want you can even replace the 'default' standard output (your
screen) with your own file object:

sys.stdout=open("my_stdout.txt",'w')

Does this make sense?

--Irmen




From ckea25d02 at sneakemail.com  Sun Apr  6 23:19:23 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 15:19:23 +1200
Subject: pyCUPS still alive?
In-Reply-To: <1Bsja.72740$OV.186811@rwcrnsc54>
References: <1Bsja.72740$OV.186811@rwcrnsc54>
Message-ID: <b6qq6f$82pho$1@ID-169208.news.dfncis.de>

djw wrote:
> Is the pyCUPS project sill alive?

I don't know about pyCUPS, but You might like to look
at Michael JasonSmith's Quick Guide to Pyrex, wherein
he presents a Pyrex wrapper for part of CUPS:

   http://www.cosc.canterbury.ac.nz/%7Empj17/pyrex-guide/

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From tim at bladerman.com  Tue Apr  8 13:26:28 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Tue, 08 Apr 2003 17:26:28 GMT
Subject: PyQT
References: <3e92e21e$0$1986$a729d347@news.telepac.pt>
Message-ID: <8DDka.4761$Du.16503@newsc.telia.net>

"Carlos Sousa" <linux_pt2003 at netcabo.pt> >

> Can someone tell me who to install eric3 with PyQT ?
> I try from the sources but something goes wrong

you must tell a lot more.

What platform are you using?
What have you tried already?
What goes wrong?
What kind of errormessages do you get?
Have you searched google groups for your particular error?

Tim





From jjl at pobox.com  Thu Apr 24 14:09:26 2003
From: jjl at pobox.com (John J. Lee)
Date: 24 Apr 2003 19:09:26 +0100
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com> <mailman.1050978337.8048.python-list@python.org> <87wuhm2rsu.fsf@pobox.com> <Bkipa.1154$Hm6.104194064@newssvr14.news.prodigy.com>
Message-ID: <87adefkbs9.fsf@pobox.com>

"Dennis Reinhardt" <DennisR at dair.com> writes:

> JpC> >   Is the app that generates these available (commercially or for
> free)?
> JJL> Probably py2exe or McMillan Installer, I imagine.
> 
> Nope.... neither one.  I used Python itself and other programs, as described
> in separate posting.

Yes, I read your post.  You said you did some kind of manual hack to
package your app, but clearly you have some system for your web page
packager (neat, BTW :-), which you didn't specify.

So, does your thing work better on space than the alternatives?  Maybe
it's worth making it available if so...


John



From tipton51 at subdimension.moc  Thu Apr  3 19:52:46 2003
From: tipton51 at subdimension.moc (Tipton Bandy)
Date: Fri, 04 Apr 2003 00:52:46 GMT
Subject: Is there a python realaudio player? For windows?
References: <b6i6uq$rcs$0@216.39.172.122>
Message-ID: <yH4ja.7115$ey1.562521@newsread1.prod.itd.earthlink.net>


A *.ram file is a "real audio metafile". It does not contain any
sound at all.  If you save the file and open it up in a text
editor, you will find a line that looks like this:
rtsp://server/path/file.rm
This is a streaming media file.  I was hoping to be able to copy
the *.rm file to my hard disk so that I could listen offline, but
so far I haven't been able to do that.  Google hasn't helped.
Tipton

Bengt Richter wrote:
> I recently wanted to play a .ram file without having to accept
> all the baggage around the minimal player capability. Is there
> such a thing? Even in C/C++, in case I got motivatged to write
> a wrapper? TIA.
> 
> Regards,
> Bengt Richter




From thorsten at thorstenkampe.de  Wed Apr  2 06:15:12 2003
From: thorsten at thorstenkampe.de (Thorsten Kampe)
Date: Wed, 2 Apr 2003 13:15:12 +0200
Subject: intro to OOP using Python, tutorial?:newbie
References: <3E8A5AE4.1080902@netscape.net>
Message-ID: <a5331ndv9u9p.dlg@thorstenkampe.de>

* vector (03-04-02 05:37 +0100)
> is there an introduction to OOp using python tutorial doc,file,pdf etc 
> anywhere on the net.

* Object Oriented Programming With Python
  http://www.devshed.com/Server_Side/Python/OOPWithPython/OOPWithPython1/print_html
  http://www.devshed.com/Server_Side/Python/OOPWithPython/OOPWithPython2/print_html
  (http://www.devshed.com/Server_Side/Python/OOPWithPython/OOPWithPython1/OOPWithPython1.pdf)
  (http://www.devshed.com/Server_Side/Python/OOPWithPython/OOPWithPython2/OOPWithPython2.pdf)

Thorsten
-- 
 Content-Type: text/explicit; charset=ISO-8859-666 (Parental Advisory)
 Content-Transfer-Warning: message contains innuendos not suited for
 children under the age of 18



From eric.brunel at pragmadev.com  Fri Apr 11 05:48:13 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Fri, 11 Apr 2003 11:48:13 +0200
Subject: Help with text editing!
References: <mailman.1050048017.19472.python-list@python.org>
Message-ID: <b762k9$ds0$1@news-reader11.wanadoo.fr>

jinal jhaveri wrote:
> Dear All
> 
> 
> I have written following code for the file open wizard of my editor
> 
> which sets the text.
> where text arg is one read from a file.
> 
> 
>  def setAllText(self,text):
>                 self.text.delete('1.0',END)
>                 self.text.insert(END,text)
>                 self.text.mark_set(INSERT,'1.0')
>                 self.text.see(INSERT)
> 
> Now I want two things to be donw
> 
> 1) I want to display this text such that alternate rows have same color and alternate columns have same color
> 
> for eg. 
> assume columsn are alternatively in blue and green.
> 
> BGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGB (ASSUMING THIS IS BLACK)
> WGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGW (ASSUMING THIS IS WHITE)
> 
> Can anyone give me an idea how do I modify this text to do that?
> 
> 2) Instead of allowing the edit for the whole text I have entered, I want editing possible only line per line basis.
> for eg. 
> 
> name1 JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
> name2 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
> name3 KKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKK
> 
> so when I select this text , it should allow me to edit only the line  (only line not even name1,name2 etc) on which I do double click and it shouldn't allow me to edit the whole text.
> 
> 
> I will highly appreciate if someone can give me hints on how to go about

Let me answer first to the second point: you won't be able to do that with a 
single Text widget. The Text widget is intended to display a _text_, which will 
be modifiable as a whole.

So my advice would be to use as many text widgets as there are lines to modify. 
And do not include the name in the text, but put it outside in a label. The next 
thing is that you may be annoyed by the fact that Text widgets are natively 
multi-line, so a press on Return will insert a new line. You can change that by 
explicitely doing a binding on the Return key. My first idea was to use entries 
(much easier to use for what you want to do) but your first point seem to 
indicate that you want your texts to have several colours, and entries can't do 
that.

Here is an example:

-------------------------------------------
from Tkinter import *

class SingleLineText(Text):

   def __init__(self, *args, **options):
     ## Height = 1 line + explicitely ignore the 'wrap' option
     options['height'] = 1
     options['wrap'] = 'none'
     apply(Text.__init__, (self,) + args, options)
     ## Ignore binding for 'Return' key
     self.bind('<Return>', lambda evt: 'break')

root = Tk()
for i in range(20):
   name = 'name%s' % i
   l = Label(root, text=name)
   l.grid(row=i, column=0, sticky='w')
   t = SingleLineText(root, width=24, bd=0, bg='white')
   t.grid(row=i, column=1, sticky='w')
root.mainloop()
-------------------------------------------

Not very cute, but you get the idea.

To do what I can understand for your first point, you can use the following 
sub-class of SingleTextLine:

--------------------------------------------
class AltColouredSingleLineText(SingleLineText):

   def __init__(self, *args, **options):
     ## Get specific options for this widget
     colour1, colour2 = 'blue', 'green'   # Whatever...
     if options.has_key('colour1'):
       colour1 = options['colour1']
       del options['colour1']
     if options.has_key('colour2'):
       colour2 = options['colour2']
       del options['colour2']
     ## Super-init
     apply(SingleLineText.__init__, (self,) + args, options)
     ## Create/configure tags in text
     self.tag_configure('COLOUR1', foreground=colour1)
     self.tag_configure('COLOUR2', foreground=colour2)
     ## Binding on any key release to display the colours
     self.bind('<KeyRelease>', self.__updateColours)

   def __updateColours(self, evt):
     ## Remove former tags
     self.tag_remove('COLOUR1', 1.0, END)
     self.tag_remove('COLOUR2', 1.0, END)
     ## Get text length
     textLen = len(self.get(1.0, '1.end'))
     ## Browse all text to update colours
     for i in range(0, textLen, 2):
       self.tag_add('COLOUR1', '1.%s' % i)
       self.tag_add('COLOUR2', '1.%s' % (i+1))
--------------------------------------------

Then use a line like:

t = AltColouredSingleLineText(root, width=24, bd=0, bg='white',
                               colour1='red', colour2='blue')

instead of the line:

t = SingleLineText(root, width=24, bd=0, bg='white')

in the code above.

For more information on what you can do with the Text widget, please refer to:
http://www.pythonware.com/library/tkinter/introduction/text.htm

HTH
-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From i.linkweiler at web.de  Sun Apr 13 05:52:58 2003
From: i.linkweiler at web.de (Ingo Linkweiler)
Date: Sun, 13 Apr 2003 11:52:58 +0200
Subject: PyWin: Redraw during long loops?
References: <3E8EA676.90008@web.de> <nd2ka.2425$Tb6.22760@news-server.bigpond.net.au>
Message-ID: <3E99337A.1000709@web.de>

> Calling win32ui.PumpWaitingMessages() generally avoids this.

no:

 >>> for i in range(90):
... 	time.sleep(0.1)
... 	x=win32ui.PumpWaitingMessages()

no redraw of the PyWin IDE while running this script.

Ingo




From mcherm at mcherm.com  Thu Apr  3 12:00:17 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Thu,  3 Apr 2003 09:00:17 -0800
Subject: else: w/o except: - why not?
Message-ID: <1049389217.3e8c68a1d9b33@mcherm.com>

David Bolen writes:
> I think you need to be careful about blanket statements like "hardly
> EVER" - it really depends on the sort of target environment being
> worked on.

Yes, on consideration I think you're right. It's rare in the kind
of development that I do, but it depends on what you're doing. Even
on projects that DO use this, though, I'd expect it to happen in a
few well-defined locations, so I wouldn't mind if the way to spell
it were somewhat more verbose (so newbies wouldn't use it by
mistake) but that's already done and not worth crying over.

-- Michael Chermside





From carroll at tjc.com  Thu Apr  3 03:15:37 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Thu, 03 Apr 2003 00:15:37 -0800
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com>
Message-ID: <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com>

On Tue, 1 Apr 2003 23:52:52 -0500, David Sfiligoi
<webmaster at quanta1.world--vr.com> wrote:

>I built a small script that use the xover function in the nntplib module.   
>The problem that I came across is that what xover can return a huge tuple 
>when there is 1000s of article in a newsgroup(which is frequent)
>
>testxover_resp,testxover_subs = s.xover(start,end)
>
>On my system it's not an issue I have 768 Mb of RAM.... but I have to 
>believe that there is a way to optimise this while keeping all this simple.
>
>How can I limit the amount of memory xover would take. the other I did an 
>xover of a huge group and the python process was taking about 650Mb Res. 
>memory.  

This is a tough one; the problem is, if you were actually issuing the
XOVER command to the NNTP server, you could handle it line-by-line as
it come in.  But nntplib does that on your behalf and hands you the
whole tuple, filled with all the headers.

So as long as you're using nntplib instead of rolling your own (and I
think that using nntplib the way to go), you're going to have the
constraint that you have to have the resources to handle that big
tuple.

>Then I go on .and put  the articlenumbers and subjects into a massive 
>dictionary which I agree it's not  a very inteligent way to do this...but 
>right now it works, I will later get rid of this huge dictonary and replace 
>it with something more memory efficient.

I'm doing something on a smaller scale, with the Unicode Unihan table.
I have one program that reads in the entire Unihan.txt file (about 26
meg), which takes about 15 seconds or so.  I save only the things
entries and attributes I need in a big dictionary, and then pickle the
dictionary to a file.  Later, when I run programs against the Unihan
data, first thing I do is load in the pickled file into the
dictionary, and then run against the dictionary.

Mind you, then you're still back to having the entire dictionary back
in memory.

To avoid this, you might want to take a look at the various dbm
modules.  I haven't used them yet, but I'd start at
http://www.python.org/doc/current/lib/module-anydbm.html

Caveat: I've only been playing with Python for a  month or so.  YOu
may get some better responses from some of the veterans.



From Vector180W at netscape.net  Tue Apr  1 22:37:08 2003
From: Vector180W at netscape.net (vector)
Date: Wed, 02 Apr 2003 13:37:08 +1000
Subject: intro to OOP using Python, tutorial?:newbie
Message-ID: <3E8A5AE4.1080902@netscape.net>

Hi all
is there an introduction to OOp using python tutorial doc,file,pdf etc 
anywhere on the net. something that explains, the interaction 
between,objects,classes,def etc.
Id prefer something using python examples




From cybersamurai at mac.com  Mon Apr  7 10:03:29 2003
From: cybersamurai at mac.com (cybersamurai at mac.com)
Date: Mon, 7 Apr 2003 11:03:29 -0300
Subject: select a class in execution time
Message-ID: <B5E651DD-6901-11D7-8CA9-000393B10A78@mac.com>

have some like that on Python?

Object x = Class.forName( "mypackage.TheClass" ).newInstace();
(this is a java code)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 219 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030407/530c79eb/attachment.bin>

From belred1 at yahoo.com  Tue Apr 22 12:23:32 2003
From: belred1 at yahoo.com (Bryan)
Date: 22 Apr 2003 09:23:32 -0700
Subject: xmlrpc introspection
Message-ID: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com>

i have some python clients that are using xmlrpc to connect to a
server written in non-python code.  i'm interested in implementing the
xmlrpc introspecton methods in 11.20.1 of the python manual on the
server side so python clients can use this feature.  but i can't seem
to find an example of the raw xml request and response or detailed
documentation for these methods.  can someone point me to the docs for
this.

thanks,

bryan



From fredrik at pythonware.com  Wed Apr 23 11:45:57 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 23 Apr 2003 17:45:57 +0200
Subject: python lexer
References: <8Wspa.23568$DN.669072@tornado.fastwebnet.it>
Message-ID: <b86ce0$8oq$1@main.gmane.org>

Alessio Pace wrote:
> Hi, I was wondering which is a good lexer written in Python, I heard about
> SPARK and PLY. I need only to tokenize texts, not to build trees on their
> structure.
> string.split() I guess it is not enough, because I can't use reg exp as
> token separators..How could I do otherwise?

it may be easier (and sometimes more efficient) to use re.findall
to find the tokens, instead of using split to find the separators.

more here:

    http://effbot.org/guides/xml-scanner.htm

</F>

<!-- (the eff-bot guide to) the python standard library (redux):
http://effbot.org/zone/librarybook-index.htm
-->







From aleax at aleax.it  Sun Apr 13 12:54:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 16:54:46 GMT
Subject: split encloser
References: <mailman.1050107298.22320.python-list@python.org> <435d40e0.0304130208.1e51b4db@posting.google.com>
Message-ID: <qDgma.8761$LB6.237749@news1.tin.it>

Chris wrote:
   ...
> it's just that my brain freezes trying to understand regular
> expressions (beyond the basics). 

Then maybe you shouldn't be straining to use re's beyond the basics --
you can get a LOT done in Python without them, after all!

> I have the book you mention and maybe
> I can dust it off and give it a go!

Make sure you use Friedl's _second_ edition, btw.  And, of course,
it will exemplify re's, but not the specific ways they're used in
Python (perhaps a secondary issue, but when you mention such re
methods as finditer and sub, it may make a difference).


Alex




From markscottwright at hotmail.com  Mon Apr  7 20:41:28 2003
From: markscottwright at hotmail.com (msw)
Date: Tue, 08 Apr 2003 00:41:28 GMT
Subject: How to protect python code ?
In-Reply-To: <v938h26mdl9s8f@corp.supernews.com>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <v938h26mdl9s8f@corp.supernews.com>
Message-ID: <3E921B51.5030401@hotmail.com>

Cameron Laird wrote:
> I exaggerate a bit.  There are indeed legitimate roles
> for obfuscating technologies.  Managers think they're
> far more common than is true, though, and programmers
> far, FAR overestimate the objective value of source
> code.  It's one of our occupational hazards.
> 
> Dave's right.

Maybe, but since I'm the only one here who's read the requirements 
document, I submit that I'm the only one qualified to suggest a software 
design, no matter how limited my intelligence.

Besides, my original post was not that this should be done, but that no 
one on comp.lang.python would be willing to discuss the matter.

I'm right.  ;-)

msw




From whisper at oz.net  Thu Apr 10 22:25:27 2003
From: whisper at oz.net (David LeBlanc)
Date: Thu, 10 Apr 2003 19:25:27 -0700
Subject: Can I define "static" (like in C++) methods in Python?
In-Reply-To: <fnsb9vcog73uj94f3ncq2tuqcf07v8s2ku@4ax.com>
Message-ID: <GCEDKONBLEFPPADDJCOEGECAJHAA.whisper@oz.net>

> -----Original Message-----
> From: python-list-admin at python.org
> [mailto:python-list-admin at python.org]On Behalf Of Rene Pijlman
> Sent: Thursday, April 10, 2003 15:41
> To: python-list at python.org
> Subject: Re: Can I define "static" (like in C++) methods in Python?
>
>
> sdieselil:
> >Can I define methods which are attributes of class and not attributes of
> >instance?
> --

http://www.python.org/doc/FAQ.html#4.84

> Ren? Pijlman
> --

I think Ren?'s link is out of date wrt current Python.

Starting with at least Python 2.2, you can:

class myclass:
	myclassvar = 0 # standard class variable

	def myclassmeth(myclass): # class as first arg
		pass

	myclassmeth = classmethod(myclassmeth)

	def mystaticmeth(): # don't use self as first arg!
		pass

	mystaticmeth = staticmethod(mystaticmeth)

Then:
	myclass.myclassmeth()

	myclass.mystaticmeth()

are calls on the class, not an instance.

"As a brief explanation of these new kinds of methods, static methods aren't
passed the instance, and therefore resemble regular functions. Class methods
are passed the class of the object, but not the object itself."
(I'm not clear on the nuances from this...)

You can find additional info about this in the "What's New" section of the
Python doc included with the release IIRC. Also available at
http://www.python.org/doc/current/whatsnew/sect-rellinks.html

Personally, I think this is cumbersome: why not use classdef and staticdef
instead of def in the above to cause the needed class wrapping to occur?

Dave LeBlanc
Seattle, WA USA





From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Wed Apr 30 12:01:45 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Wed, 30 Apr 2003 18:01:45 +0200
Subject: bug in Jython 2.1? VerifyError
In-Reply-To: <mailman.1051713834.1595.python-list@python.org>
References: <3eafb3b0$0$49112$e4fe514c@news.xs4all.nl> <mailman.1051713834.1595.python-list@python.org>
Message-ID: <3eaff369$0$49100$e4fe514c@news.xs4all.nl>

David Ripton wrote:

> I can reproduce the problem, with Jython 2.1 on JDK 1.4.2 beta on Linux.
> 
> 2.1 is pretty old -- you might want to try building Jython from CVS and 
> see if this has been fixed before reporting it.

I did. (Jython 2.2a0 from CVS). Same problem.
I reported this as a bug;

http://sourceforge.net/tracker/index.php?func=detail&aid=730156&group_id=12867&atid=112867

(somebody else has submitted a duplicate report sometime later, # 730171)

Hope this gets fixed quickly.

--Irmen de Jong




From peter at engcorp.com  Wed Apr  2 15:22:59 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 15:22:59 -0500
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com> <slrnb8lbo0.ccp.rmunn@localhost.localdomain> <3E8AFFB7.9516A49F@engcorp.com> <b6f7ib$4rc09$1@ID-139865.news.dfncis.de>
Message-ID: <3E8B46A3.3DA1E10F@engcorp.com>

Dave Kuhlman wrote:
> 
> Peter Hansen wrote:
> 
> > Robin Munn wrote:
> >>
> >> sdieselil <sdieselil at yahoo.com> wrote:
> >> > Hi
> >> >
> >> > How can I implement double-linked lists and other complex tree-like
> >> > structures in Python?
> >>
> >> Do you *need* doubly-linked lists? Is there something that Python's
> >> builtin list type can't do for you? Consider whether it's worth taking
> >> the time to code up a data structure, with all the attendant edge cases,
> >> when you've already got such nice useful builtin types...
> >
> > I can think of at least one example where one would *need* doubly-linked
> > lists... a homework assignment.  ;-)
> 
> But suppose that I *do* implement double linked lists.  I'll have two
> objects that point at each other, in other words circular references.

Uh, only if you've implemented some magic technique of deleting a range
of items from the list, and are not planning to maintain references to 
them from elsewhere.  And even then as Aahz points out, the GC will find
the cycles and remove them from memory (unless the Nodes have __del__ methods).

But why would you do that?  Normally you remove items one at a time from
a list...  or are you talking about in the case where you delete the 
reference to the entire list without first emptying the list?  It sounds
like you're fishing here... why not just build it and see how it works?
(I realize you weren't the OP...)

> In older versions of Python, circular references were not garbage collected.
> Is this still true in Python 2.2 and 2.3?  Do I need to use the gc module
> to *enable* garbage collection?  Or, is not *disabling* it good enough?  Do
> I still need to worry about this?  And, suppose you tell me that circular
> references are collected but I don't believe you.  How can I check?

Uh.... allocate a zillion linked lists, fill them with objects, and then
delete the references to them and watch the memory get reclaimed.  

This is _Python_: just try it!  There's no excuse for someone not 
experimenting with something like this in Python... it takes all of a few
minutes to try it out and see how it behaves.  In XP, you call that a
"spike" or spike solution, and it tells you just enough that you can 
carry on with design decisions based on the results.  Python is very
very very good for this sort of approach, to the point that if you 
don't believe me about the garbage collection, you have a very clear 
path ahead of you. :-)

-Peter



From eric.brunel at pragmadev.com  Fri Apr 11 05:09:11 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Fri, 11 Apr 2003 11:09:11 +0200
Subject: Help with text editing!
References: <mailman.1050048017.19472.python-list@python.org>
Message-ID: <b760b1$8ab$1@news-reader10.wanadoo.fr>

jinal jhaveri wrote:
> Dear All
> 
> 
> I have written following code for the file open wizard of my editor
> 
> which sets the text.
> where text arg is one read from a file.
> 
> 
>  def setAllText(self,text):
>                 self.text.delete('1.0',END)
>                 self.text.insert(END,text)
>                 self.text.mark_set(INSERT,'1.0')
>                 self.text.see(INSERT)
> 
> Now I want two things to be donw
> 
> 1) I want to display this text such that alternate rows have same color and alternate columns have same color
> 
> for eg. 
> assume columsn are alternatively in blue and green.
> 
> BGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGB (ASSUMING THIS IS BLACK)
> WGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGW (ASSUMING THIS IS WHITE)
> 
> Can anyone give me an idea how do I modify this text to do that?
> 
> 2) Instead of allowing the edit for the whole text I have entered, I want editing possible only line per line basis.
> for eg. 
> 
> name1 JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
> name2 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
> name3 KKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKK
> 
> so when I select this text , it should allow me to edit only the line  (only line not even name1,name2 etc) on which I do double click and it shouldn't allow me to edit the whole text.
> 
> 
> I will highly appreciate if someone can give me hints on how to go about
> 
> 


-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From zf_acs at DELETETHIShotmail.com  Tue Apr 15 07:18:36 2003
From: zf_acs at DELETETHIShotmail.com (zif)
Date: Tue, 15 Apr 2003 07:18:36 -0400
Subject: hex -> 16bit signed int (newbie)
Message-ID: <TTRma.1502$KH1.321574@news20.bellglobal.com>

How can I get 16bit signed int from a hex string?
Eval('0x...') gives me 16bit unsigned,
struct.pack('i', '\x..\x..') works fine, but
Python doesn't allow me to concatenate strings
with '\' (str(92) gives me '\\', which doesn't
work as well).

Thanks.





From pan-newsreader at thomas-guettler.de  Tue Apr  1 06:58:42 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Tue, 01 Apr 2003 13:58:42 +0200
Subject: ZODB: Quixote vs. Twisted
Message-ID: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de>

Hi,

I think about moving from Zope to
ZODB and (Quixote or Twisted).

I looked at zope3 but it seems very bloated.

Up to now Zope seems the best http server for ZODB.

Has anyone experience with serving an application 
which uses ZODB as database backend but does not use Zope?

 thomas


-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From hllee678 at hotmail.com  Mon Apr 21 08:11:53 2003
From: hllee678 at hotmail.com (=?big5?B?p/UgvuWswg==?=)
Date: Mon, 21 Apr 2003 20:11:53 +0800
Subject: M2Crypto problem
Message-ID: <Sea2-F27zRPTNZStgLz0000c9bc@hotmail.com>

Hi,
I am using M2Crypto to write the SSL server and client.
My server was based on Medusa.
The client built a SSL connection to server,
and then created a new thread to receive the response from server.

The code of client:
rfile=sock.makefile('rb',-1) #sock is the ssl connection between server and 
client
try: 
   data = rfile.read(100)
   if not data:
      print "no data, maybe try to close connection"   
except:
   print "exception, maybe try to close connection"   

When the server shutdown and close the SSL connection between server and 
client,
the client receiving thread raised a windows exception on rfile.read 
procedure:
Unhandled exception in python.exe (PYTHON21.DLL): 0xC0000005: Access 
Violation

The code of server:
sock.shutdown(2)         # socket is the ssl connection between server and 
client
sock.close() 

Please help me if you know how to solve this problem.
Thank you very much!


_________________________________________________________________
???? MSN ???????????????????????????????
? http://msn.com.tw 





From andrew-pythonlist at puzzling.org  Thu Apr 24 07:47:41 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Thu, 24 Apr 2003 21:47:41 +1000
Subject: Question about accessing class-attributes.
In-Reply-To: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com>
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com>
Message-ID: <20030424114741.GA15612@frobozz.local>

On Thu, Apr 24, 2003 at 01:29:44PM +0200, Nikolai Kirsebom wrote:
> Given the following class hierarchy:
>       
> class A(object):
>       InstCount = 0
>       def __init__(self):
>               self.__class__.__dict__['InstCount'] += 1

A simpler way to do this line would be:

              self.__class__.InstCount += 1

-Andrew.





From drobinow at yahoo.com  Thu Apr  3 22:04:09 2003
From: drobinow at yahoo.com (David Robinow)
Date: 3 Apr 2003 19:04:09 -0800
Subject: Cygwin Python regression testers wanted
References: <mailman.1049385789.4709.python-list@python.org>
Message-ID: <59a4b541.0304031904.1cfb2918@posting.google.com>

Jason Tishler <jason at tishler.net> wrote in message news:<mailman.1049385789.4709.python-list at python.org>...
> I'm in the process of updating the Cygwin section of the official Python
> README.  Since I only have access to Windows NT 4.0 and 2000, I am
> seeking Cygwin Python regression testers for the following Windows
> platforms:
>     XP Home
 That's me!
> 
>    python -E -tt /usr/lib/python2.2/test/regrtest.py -l 2>&1 | tee regrtest.out
> 
> 3. post any failures and/or hangs other than test_commands
172 tests OK.
1 test failed:
    test_commands
20 tests skipped:
    test_al test_cd test_cl test_curses test_dbm test_dl test_email
    test_email_codecs test_gl test_imgfile test_largefile
    test_linuxaudiodev test_locale test_nis test_socket_ssl
    test_socketserver test_sunaudiodev test_unicode_file test_winreg
    test_winsound
Ask someone to teach regrtest.py about which tests are
expected to get skipped on cygwin.



From peter at engcorp.com  Thu Apr 24 19:34:13 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 24 Apr 2003 19:34:13 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it>
Message-ID: <3EA87475.F6A19BE@engcorp.com>

Alex Martelli wrote:
> 
> Peter Hansen wrote:
> > throwaway than the ones I "know" are not.  If I'm wrong... no big deal:
> > I rewrite as a real script with proper tests.  Since the script was
> > only twenty or thirty lines of code (any more and it could not possibly
> > be throwaway, right?) it isn't really a big deal.
> 
> Wrong.  "Throwaway" means "I [think I] am going to need to run this
> only once".  Whether it's 30 or 60 lines has nothing to do with it.

This is getting us nowhere.  

I'm asserting that I have actually seen "throwaway" scripts, even predicted 
their imminent development, and subsequently have actually thrown them away.
I'm not asserting I do this with 100% accuracy.  I'm asserting that
I very infrequently do decide that a script will be throwaway, and
that in those rare cases I often turn out to be right, and often 
enough that it seems worth continuing to make such predictions in
order to save myself some time.  These I believe are facts about *me*.

You're asserting that you have never accurately predicted this, or
at least not accurately enough to make prediction worthwhile, or 
at least that even when you predict correctly it you discount the 
prediction and proceed on the assumption that you are wrong.  (Or 
something like that... I'm unsure it's actually relevant to my point.)

Neither of us can assert anything about the other's real situation,
nor presumably about Aahz' (though I admit I did... maybe I was wrong).

I can't see any point in continuing the discussion, can you?

-Peter



From martin.vonloewis at hpi.uni-potsdam.de  Wed Apr 23 05:07:13 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Wed, 23 Apr 2003 11:07:13 +0200
Subject: i18n in a library
References: <b85hjt$6l2na$1@ID-15245.news.dfncis.de> <jdspa.47509$LB6.1087058@news1.tin.it>
Message-ID: <b85l43$dhf$1@zeppelin.rz.uni-potsdam.de>

> I think what you want is gettext.translation, which offers you
> finer-grained control than gettext.install.  I.e., within the
> modules of library L, you could use
>
>     import gettext
>     _ = gettext.translation('L').ugettext
>
> so that in those modules function _ will look things up in domain
> 'L', rather than in whatever domain is "installed" globally.

Alternatively, gettext.dgettext is the Python version of the
equivalent C API, so you could also do

def _(msg):
   return gettext.dgettext('L', msg)

This has the added advantage of returning the string itself if
no catalog can be found (instead of giving IOError as
gettext.translation does); with Python 2.x (for some x),
you can pass the fallback=True argument to translation
to achieve the same advantage.

The d in dgettext is for 'domain'; with 'L' being a textual domain.
The notion of a textual domain is somewhat independent from the
notion of a software package; different packages may share the same
domain, yet a single package may consist of several domains.

Regards,
Martin





From bdesth at removethis.free.fr  Tue Apr 22 14:18:59 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 22 Apr 2003 18:18:59 +0000
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org> <QxAoa.540832$3D1.296696@sccrnsc01> <a5415304.0304210544.1342b436@posting.google.com> <3EA490F5.30606@removethis.free.fr> <mailman.1051002218.12703.python-list@python.org>
Message-ID: <3EA58793.4000301@removethis.free.fr>

Oktay Safak wrote:
> Very rude of you, Laotseu ;)

Yes, I know, I shouldn't have mispelled 'smalltalk' !-)

> 
>>(oh shit, I'm sure I've just said another stoopidity that will end up in 
>>a long cross-post with c.l.smalltlak now)
> 








From aahz at pythoncraft.com  Mon Apr  7 16:56:38 2003
From: aahz at pythoncraft.com (Aahz)
Date: 7 Apr 2003 16:56:38 -0400
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <m2k7e75aze.fsf@localhost.localdomain>
Message-ID: <b6som6$c3p$1@panix1.panix.com>

In article <m2k7e75aze.fsf at localhost.localdomain>,
Patrick K. O'Brien <pobrien at orbtech.com> wrote:
>
>Is it enough to support the following:
>
>* default Emacs keybindings
>* ability to remap keybindings
>* multiple buffers and multiple windows
>* multiple modes
>* keyboard friendliness (no need for a mouse)
>* minibuffer-style control of environment
>
>If not, what's missing and what's the definition of a "good editor"
>written in Python?  Should it allow customizations and extensions,
>like Emacs, but be a completely new and Pythonic API?

What's missing is a set of default vi keybindings.  Another missing
feature is working from the console.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From arosado at softhome.net  Sun Apr 27 21:06:13 2003
From: arosado at softhome.net (Andres Rosado)
Date: Sun, 27 Apr 2003 21:06:13 -0400
Subject: How to write RTF from Python?
Message-ID: <5.2.0.9.0.20030427210611.00bde860@mail.softhome.net>

On 02:31 AM 4/26/2003 -0400, python-list-request at python.org feed this fish 
to the sharks:
>Another idea: how about PDF?  I don't know if the formatting gets
>preserved at all if you cut'n'paste, though.

Unfortunately, it won't work. We tried here and it was a PITA getting 
something from a PDF to a Word Document.


-- 
Andres Rosado
Email: andresr at despammed.com
ICQ: 66750646
Homepage: http://andres980.tripod.com/
Get my PGP key at http://andres980.tripod.com/pgp-key.asc

Some are born great, some achieve greatness, and some have greatness
thrust upon 'em.
         -- Twelfth Night  





From duncan-news at grisby.org  Sun Apr  6 16:37:03 2003
From: duncan-news at grisby.org (Duncan Grisby)
Date: Sun, 06 Apr 2003 20:37:03 +0000
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <a626bbd7.0304031146.6f98fc28@posting.google.com> <m3of3n6vfu.fsf@mira.informatik.hu-berlin.de>
Message-ID: <af319$3e908fef$516049d2$26016@nf1.news-service.com>

In article <m3of3n6vfu.fsf at mira.informatik.hu-berlin.de>,
 Martin v. L?wis <martin at v.loewis.de> wrote:

>> c) size of the "library"
>
>That depends on the implementation you use. For a client, it is
>certainly true that XML-RPC libraries are significantly (factor 10)
>smaller than the Fnorb libraries.

It's also worth pointing out that, although Fnorb and omniORB are both
a megabyte or two in size, that's still significantly smaller than
Python itself, so size is not an issue for the vast majority of
applications written in Python.

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --



From newspost at lolmc.com  Mon Apr 28 18:24:25 2003
From: newspost at lolmc.com (Lol)
Date: 28 Apr 2003 15:24:25 -0700
Subject: Arg names in class and function code - Newbie question
Message-ID: <64e13ba5.0304281424.52cb6b5d@posting.google.com>

HI all,
	Could anyone let me know - do the names you give as arguments HAVE to be
actual variable names?
e.g def readFile(filename,dir):
	blah blah
	return blah
or even class FileRead(filename,dir):
	def doSomething (self,this):
		yadda yadda
		return yadda

In the above examples do filename and dir have to be the name of the
variables to pass as args to the class/functions OR are they descriptive
so I can pass the real name of the file and directory?



From duncan at NOSPAMrcp.co.uk  Thu Apr 24 12:21:49 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 16:21:49 +0000 (UTC)
Subject: php equivalents?
References: <OHSpa.8996$h%2.915957@read1.cgocable.net> <b890vo$rpj$1@news1.wdf.sap-ag.de>
Message-ID: <Xns9367AF961CB45duncanrcpcouk@127.0.0.1>

"Daniel Dittmar" <daniel.dittmar at sap.com> wrote in 
news:b890vo$rpj$1 at news1.wdf.sap-ag.de:

>>> var_export($var)
>>
>> This function converts a variable to a text string of valid code
>> which can be reevaluated at any time. Python has pickle/cPickle, and
>> also has modules that can dump/load to XML, however, none of these
>> create actual code, but instead create a formatted data string.
> 
> repr () is the equivalent function, at least for the builtin types.

repr() won't give the desired effect if there are multiple references to 
the same object involved. e.g.

>>> aList = [1, 2, 3]
>>> aDict = {'x': aList, 'y': aList }
>>> aDict
{'y': [1, 2, 3], 'x': [1, 2, 3]}

Evaluating the repr of aDict won't give you an equivalent structure, and if 
you do:

>>> aList[1] = aDict
>>> aDict
{'y': [1, {...}, 3], 'x': [1, {...}, 3]}

then you no longer have something you can eval at all.

(Mind you, I have no idea whether PHP's var_export handles the equivalent 
of these cases either.)

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From python at dohao.org  Sun Apr 20 22:06:08 2003
From: python at dohao.org (Wenshan Du)
Date: 20 Apr 2003 19:06:08 -0700
Subject: Zope and FreeBSD
Message-ID: <e786d63d.0304201806.4c5293a6@posting.google.com>

Hi,

I've built a website with Zope 2.5.1 and FreeBSD 4.3, it's better than
other solutions.

The server got so slow recently that I have to reboot every 3 days,
and Zope stop sometimes.

I don't know why, is zope not stable on FreeBSD? or FreeBSD got virus?



From max at alcyone.com  Thu Apr  3 17:58:17 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 03 Apr 2003 14:58:17 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3E8CBC89.733D01D1@alcyone.com>

Michael Hudson wrote:

> I certainly did.

Me too.  I don't have any huge opinion on the proposition one way or
another, but I must say I do find it reassuring in Python to know that
(at present, without the proposal) whenever I see an assignment, I
always know that there's going to be a rebinding/attribute assignment
taking place.  Assignment seems so fundamental an operation that I'm not
sure whether it would be net good or net bad to have it overridable.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The decree is the Sultan's, the mountains are ours.
\__/ Dadaloglu
    Official Omega page / http://www.alcyone.com/max/projects/omega/
 The official distribution page for the popular Roguelike, Omega.



From christopher.saunter at durham.ac.uk  Mon Apr 28 05:31:48 2003
From: christopher.saunter at durham.ac.uk (Christopher Saunter)
Date: Mon, 28 Apr 2003 09:31:48 +0000 (UTC)
Subject: GUI testing with python  (SendKeys)
References: <b8c6c9$p2m$0@198.102.102.163> <b8ecgs$47i$1@sirius.dur.ac.uk> <3EAAC082.1DD45058@engcorp.com> <3EAAC0F6.81B70C87@engcorp.com>
Message-ID: <b8ise4$l2m$1@sirius.dur.ac.uk>

Peter Hansen (peter at engcorp.com) wrote:

: Never mind... dict.org handles that too.  I should have checked first.

: noddy
:   
:      /nod'ee/ [UK: from the children's books] 1. Small and
:      un-useful, but demonstrating a point.  Noddy programs are
:      often written by people learning a new language or system.
:      The archetypal noddy program is hello, world.  Noddy code
:      may be used to demonstrate a feature or bug of a compiler.
:      May be used of real hardware or software to imply that it
:      isn't worth using.  "This editor's a bit noddy."
:   
:      2. A program that is more or less instant to produce.  In this
:      use, the term does not necessarily connote uselessness, but
:      describes a hack sufficiently trivial that it can be written
:      and debugged while carrying on (and during the space of) a
:      normal conversation.  "I'll just throw together a noddy awk
:      script to dump all the first fields."  In North America this
:      might be called a mickey mouse program.  See toy program.


: Actually, which one did you mean?

I'm leaning towards their second definition, although without much in the 
way of instructions the code leans towards the first ;-)

cds




From km-news1 at onlinehome.de  Sat Apr 26 11:57:07 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Sat, 26 Apr 2003 17:57:07 +0200
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it>
Message-ID: <oprn8ulhb1mu3s0q@news.cis.dfn.de>

> I don't know what to DO about it, but some minor tweaks might help.

I also have made this mistake sometimes.

On reason, i think, is the mix of function-call and values, example:

import time
print time.time()
print time.timezone

You have to remember the doc of the module carefully to make no mistake.

In this case time() must be a function, because it has to read the actual 
timer value, but must timezone be a value?

My question about style:
Should a Module/Class etc. return values, if possible, or better use 
function to return values?

If PyChecker could check such cases this would be very helpful.

-- 
Gru? - regards Klaus :-)



From aleax at aleax.it  Sat Apr 19 04:13:25 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 08:13:25 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu>
Message-ID: <Fy7oa.53694$T34.1483263@news2.tin.it>

David Eppstein wrote:

> In article <C8_na.646$dP1.2863 at newsc.telia.net>,
>  "Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote:
> 
>> > What's wrong with:
>> >
>> > t = 0
>> > for y in x:
>> >      t += y[1]
>> >
>> > The extra variable?  The number of lines?
>> 
>> I am really curious about that to, I would most definitley say that this
>> is the most pythonic way. It is simple and easily readabel by anyone and
>> I am sure it is just as fast any of the other way.
> 
> If you want the sum of a list of items, you should write it in a way
> that looks like "the sum of a list of items", not in a way that looks

"The sum" is so frequently needed that I wouldn't mind at all if
Python singled it out as a built-in.  But "reduce(operator.add, ..."
just isn't a great way to express it, in my opinion (and yet as an
old APL'er, and FP-liker, I _should_ like it -- but I don't).

> like "loop over these items, maintain another variable t, perform a
> sequence of additions".  Why do we have high level languages if not to
> express our intentions at a higher level and let the language worry
> about what low-level operations are needed to implement it?

We use VHLL's to enhance our programming productivity.  There's no
*mandate* that each idiom MUST be very abstract and require higher
maths; on the contrary, Python thrives on simplicity and directness.

"sum([y[1] for y in x])" would be very simple and direct, IMHO; the
"same" thing expressed with "reduce(operator.add,[y[1] for y in x])"
may only feel "simple and direct" to PhD's in maths -- it feels
_abstruse_ to most people, in my opinion.

Python isn't about being abstruse.


Alex




From cpl.19.ghum at spamgourmet.com  Tue Apr 29 11:51:46 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Tue, 29 Apr 2003 17:51:46 +0200
Subject: HTTP application streams a response to the client gradually - how deal with it?
Message-ID: <Xns936CB5697A6D4cpl19ghumspamgourmet@62.153.159.134>

Hello,

I want to connect to an HTTP application, that streams its response to the 
client gradually.

(it is a "web chat")

>From reading documentes and google, I learned that urllib will not fit.

Also httplib --- I only see methods that provide me with whole HTTP-
responses.

Is there any way to deal with "streamed http" without going to socket-
programming?


Tnx,

Harald




From Kyler at news.Lairds.org  Mon Apr  7 22:23:07 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Tue, 08 Apr 2003 02:23:07 GMT
Subject: How to protect python code ?
References: <mailman.1049758213.26256.python-list@python.org> <slrnb945tu.p4.bignose-hates-spam@rose.localdomain.fake> <5ppka.23067$It5.15751@news2.central.cox.net>
Message-ID: <prm9m-biv.ln1@news.lairds.org>

"Steve Holden" <sholden at holdenweb.com> writes:

>Of course, the businesses that make a good living out of proprietary
>software encourage all sorts of FUD about open source. Consequently, large
>organizations implement rules forbidding the use of open source software
>products because "they aren't supported" when in fact the support for some
>open source products is actually superior to that of most proprietary
>products.

Regardless of the current state of support for two products, my
preference is going to be for the one that *can't* become
unsupported.  That means open source/Free software.  I can
always find someone to support it.

--kyler



From mwh at python.net  Wed Apr 16 09:12:11 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 13:12:11 GMT
Subject: Python 2.3a2: sockettimeout() buggy?
References: <slrnb9qki8.7p2.news@cipw08.physik.uni-bonn.de>
Message-ID: <7h3smsir3h1.fsf@pc150.maths.bris.ac.uk>

Carsten Gaebler <news at snakefarm.org> writes:

> I just tried out the new socket timeouts in Python 2.3a2. Seems that the
> timeout does not work if you make a file from the socket.

I'm pretty sure Skip's fixed this in CVS.

Cheers,
M.

-- 
  Well, you pretty much need Microsoft stuff to get misbehaviours
  bad enough to actually tear the time-space continuum.  Luckily 
  for you, MS Internet Explorer is available for Solaris.
                              -- Calle Dybedahl, alt.sysadmin.recovery



From carroll at tjc.com  Tue Apr 29 02:56:39 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Mon, 28 Apr 2003 23:56:39 -0700
Subject: Determining the volume ID for a file/drive?
Message-ID: <uv7savkms6lsgchcv6jh8ahtccu9lon3oi@4ax.com>

Is there any way to determine the volume serial number for a disk,
given the filename (or drive, since I can determine the drive from the
filename with os.path.splitdrive(os.path.abspath(file))?

My environment is Windows & Python 2.2.2, and I don't need for my code
to be portable. 




From bokr at oz.net  Sat Apr 26 13:41:03 2003
From: bokr at oz.net (Bengt Richter)
Date: 26 Apr 2003 17:41:03 GMT
Subject: Reading a non-standard floating point format
References: <mailman.1051286382.24604.python-list@python.org> <vaip4ts74ajqd0@corp.supernews.com> <Xns9327331C4783Bfidtzxyzabc@195.8.68.209>
Message-ID: <b8egbf$o21$1@216.39.172.122>

On Fri, 25 Apr 2003 21:09:28 +0100, Fidtz <fidtz at clara.co.uk> wrote:
[...]
>Thanks for the info, I was worried that might be the situation!
>
>The closest thing I have to a definition from the AlphaBasic programming
>manual: The reason for this is that floating point numbers occupy six
>bytes of storage. Of the 48 bits in use for each 6-byte variable, the
>high order bit is the sign of the mantissa. The next 8 bits represent
>the signed exponent in excess-128 notation, giving a range of
>approximately 2.9*10^-39 through 1.7*10^38. The remaining 39 bits
>contain the mantissa, which is normalized with an implied high-order bit
>of one. This gives an effective 40-bit mantissa which results in an
>accuracy of 11 significant digits. 
>
>I had a go this afternoon and managed to read the sign and something
>approximating the exponent, though I will have to test the edge
>conditions carefully :) 
>
Got some test data? Is it big-endian or little-endian? Or do you need to handle both?

Regards,
Bengt Richter



From a.schmolck at gmx.net  Tue Apr 15 17:14:14 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 15 Apr 2003 22:14:14 +0100
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de> <Xns9358723026D8Cduncanrcpcouk@127.0.0.1> <3e9bdd3d.100422968@news.online.de>
Message-ID: <yfsbrz7cvkp.fsf@black132.ex.ac.uk>

skywalkerpackage at hotmail.com (Johannes Eble) writes:

> >A quick look at the Dia website seems to indicate that DiaPython exists.
> 
> Not really. There is a way to use Python plug-ins, but it is undocumented.
> It is impossible to write a plug-in without a knowlede of the c code. Also,
> I don't like GTK+ because this pseudo object oriented c style is very ugly
> IMHO let apart the problems you face with Windows.

You can just Right-mouse click on canvas,
Dialogs->Python Console. And tatah, you get a python console. dir() will then
Dialogs->reveal what's lying around:


>>> dir()
['__builtins__', '__doc__', '__history__', '__name__', 'dia']
>>> dia
<module 'dia' (built-in)>
>>> dir(dia)
['DiaArrow', 'DiaBezPoint', 'DiaColor', 'DiaConnectionPointType',
'DiaDiagramData', 'DiaDiagramType', 'DiaDisplayType', 'DiaError',
'DiaExportFilter', 'DiaFont', 'DiaHandleType', 'DiaImage', 'DiaLayerType',
'DiaObjectType', 'DiaObjectTypeType', 'DiaPoint', 'DiaProperties',
'DiaProperty', 'DiaRectangle', 'DiaText', '__doc__', '__name__',
'active_display', 'diagrams', 'get_object_type', 'load', 'register_callback',
'register_export', 'update_all']

etc. 

I have only looked at it for a minute (and succesfully zoomed my document :),
but I think it might be worth a little tinkering because even if the python
interface is undocumented and not completely up-to-scratch because dia itself
certainly seems to work rather well to me.

'as



From donn at drizzle.com  Thu Apr 17 03:00:52 2003
From: donn at drizzle.com (Donn Cave)
Date: Thu, 17 Apr 2003 07:00:52 -0000
Subject: Using __repr__ or __str__ for own printable class?
References: None <mailman.1050538113.22063.python-list@python.org>
Message-ID: <1050562849.887077@yasure>

Quoth Tim Peters <tim.one at comcast.net>:
| [Tim, quotes the __repr__ docs
|  http://www.python.org/doc/current/ref/customization.html
| ]
|
| [Donn Cave]
|> It's clear that the docs promote this idea, but they also hint at
|> its half-baked quality - repr() evals to the original value, unless
|> it doesn't.  I don't think the idea is clear at all.
|
| I do, but the variety of objects is messy.  The spirit of __repr__ is "do
| the most faithful job you can, within reason".  Some objects simply can't be
| reproduced faithfully with a comprehensible Python expression, and the docs
| admit to that reality up front.
|
|> Taken together with the __str__ documentation, the original poster saw
|> this stuff and did not find it clear.
|
| I believe that, but am not concerned by it.  There's usually more than one
| way to convert an object to string that may be useful, and for some objects
| there may be many distinct ways, and Python has two convert-to-string
| special methods to cover two frequent uses:  friendly (str) and anal (repr).
| I think these work very well for the builtin scalar types, but not so well
| for the builtin container types (because str(container) applies repr() to
| the containees).  It's also sometimes unfortunate that the interactive
| prompt is wedded to __repr__ by default.  So it goes.  What to do with each
| method is in large part art, and the teachable seem to do OK learning by
| example.  Here's an example new in 2.3:
|
| >>> import datetime
| >>> t = datetime.date.today()
| >>> print repr(t)
| datetime.date(2003, 4, 16)
| >>> print str(t)
| 2003-04-16
| >>>
|
| Fred Drake decided to make repr() and str() do those specific things for
| dates, and it was instantly obvious to me, and to Guido, that those are
| appropriate things for each to do.  Is it actually hard for you to see that
| both match the doc's intents?  It can't be hard to see that the repr() does,
| so maybe you can't see that the str() result is "more convenient or
| concise"?  Hard to believe <wink>.

I did concede that the documentation allows you to read into it what
you already understand to be true.  I find that these implementations
satisfy my own preconception of how repr and str should work.  That
happens not to conflict with the __str__ documentation, but it's so
vague that it's hard to imagine a real solid conflict (convenient to
whom, or what?  Just convenient in some way, your choice?)


|> The requirements for repr conflict, and the primary one is conceded to perhaps
|> not be practicable.
|
| That's because sometimes it's not.  So what?  Because it may not always be
| possible, you want the language to cut off the possibility of trying?

Not at all.  There's no harm in mentioning that repr could work that
way and sometimes does.  It just should not be stated as a requirement,
it's silly and confusing to say you must but only if you can.


|> The requirements for str are on the other hand silent on its intended
|> application.
|
| Applications are the user's business, not the language spec's.  If you can't
| think of an application for "more convenient or concise; doesn't need to try
| to be a Python expression", stay away implementing __str__.

I think the original poster was really quite sensibly wondering under
what circumstances str() might be invoked on his object.  I won't pretend
that there's a great deal that can be said about that on any specific
level, but words like ``informal'' and ``convenient'' are worse than
nothing when not attached to some context.

|> The only kind of clarity this documentation possesses is when it lets you read
|> into it what you already thought.  Not the fault of the documentation, though,
|> but of the unsound ideas behind it.
|
| It's evident that more than one way of converting to string may be useful,
| and that some ways are more useful for developers and others for end users.
| Those are "the unsound ideas behind it".  They're not precise ideas, nor do
| I believe they can be made precise.  Please tell us how *you* would design
| the to-string facilities in a language -- sticking strictly to "sound"
| ideas, of course.

Well, we've been over this, but I would be delighted.  I think it's
going to help to acknowledge that all Python features are exclusively
useful to developers, who are actually the end users of Python.

The fact that end users of a Python application are exposed more
commonly to the output of str(), says something about str's application
in a general sense - it's about transforming data, and that's what
users experience, data.

repr is about objects, which users do not experience.  I could go
on, but I think that's really enough.  It's a string representation
that is about the object qua object, and from this you can infer its
applications.  Objects implement data, and where you just want that
data, you use str.

Am I creating a new language here?  Obviously I am not, this is not
how Python ought to have worked, it's how it does work.  It's how
it ought to be documented.

	Donn Cave, donn at drizzle.com



From mathie at wossname.org.uk  Wed Apr  2 09:11:57 2003
From: mathie at wossname.org.uk (Graeme Mathieson)
Date: Wed, 02 Apr 2003 15:11:57 +0100
Subject: Python equivalent of _tempnam?
References: <3e8ab45b$1@mail.hmgcc.gov.uk>
Message-ID: <sa6k7ed6laq.fsf@kerouac.projectcolo.org.uk>

In <3e8ab45b$1 at mail.hmgcc.gov.uk>,
on Wed, 2 Apr 2003 10:58:51 +0100, "Richard" <richardd at hmgcc.gov.uk> writes:

> I have been told that a way of doing this in C is by using _tempnam( ).
> However I want to write this in Python, so is there an equivalent function?

os.tempnam() by any chance?  Though, as the documentation[1] points out,
you'd be better off using os.tmpfile() instead.

Footnotes: 
[1]  http://www.python.org/doc/current/lib/os-file-dir.html
-- 
Right now, there are scr1pt k1dd13s plotting to DDoS my network, my co-lo
server is not responding  to pings and  the people that I IRC with may be
involved in both.  I'm  sysadmin Graeme Mathieson and this is the longest
day of my life.                               http://www.wossname.org.uk/



From ckea25d02 at sneakemail.com  Mon Apr  7 23:44:10 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 15:44:10 +1200
Subject: Cross-reference 'import' in a class hierarchy
In-Reply-To: <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de>

Ben Finney wrote:
> I feel dim for not being able to see the answer, but it seems absurd to
> me that one can't rename a directory and expect the scripts within it to
> continue working without outside help.

It's been suggested before (and I happen to agree) that
Python should provide a general way of making *relative*
references to modules or packages. You can do this
already for a module that resides directly in the same
package as the module doing the importing, but for
anything else you must use an absolute path from the
"root package".

However, the response so far from the BDFL has only
been "Why would you want such a thing?", which is
unfortunate, because I find it hard to understand how
someone could think you *wouldn't* want it.

I'm hoping that someday he will see the light...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From xantor at linux.be  Sat Apr 19 08:57:45 2003
From: xantor at linux.be (Michael Anckaert)
Date: 19 Apr 2003 14:57:45 +0200
Subject: getting an item's list index
Message-ID: <1050757065.735.2.camel@carpathia>

Hello all,

I have a list called contents. If one of the elements has the value
"item1" for example, how can I get that elements index?


for item in contents:
	# here I need to get that items index

-- 
Greetings,
Michael Anckaert, aka The XanTor
OpenPGP key: F7A6C3AB
Fingerprint: A329 43FC 3953 A944 5DDC  2E05 8E5D AD60 F7A6 C3AB
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 196 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/python-list/attachments/20030419/00be999c/attachment.sig>

From scott_list at mischko.com  Thu Apr 17 15:59:10 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Thu, 17 Apr 2003 12:59:10 -0700
Subject: HTML form pre-populating functionality or hints on building myown
In-Reply-To: <1050608556.720.298.camel@lothlorien>
References: <200304161939.13004.scott_list@mischko.com> <Hhwna.114818$It5.51036@news2.central.cox.net> <1050608556.720.298.camel@lothlorien>
Message-ID: <200304171259.10272.scott_list@mischko.com>

On Thursday 17 April 2003 12:42, Ian Bicking wrote:
<snip>

> I think what the OP was talking about was different, meant to be used on
> the server side.  Like you'd have a pseudo-template something like:

<snip>

> And that subform function goes in, finds the <input...> tags, and fills
> them in from the form dictionary, so that the user is presented with
> their input and a chance to make corrections.

Ian,
You hit it on the head.  I want to be able to do this for correcting errors 
and for initially populating forms.  I started writing my own last nite just 
for a fun first project in Python.  It's coming along fairly well already 
(testimony to Python's greatness).

Thanks!
Scott





From dsavitsk at e-coli.net  Thu Apr 24 14:07:28 2003
From: dsavitsk at e-coli.net (dsavitsk)
Date: Thu, 24 Apr 2003 18:07:28 GMT
Subject: [Question] I want to know what script more speedy....  VB script and  Python Script...
References: <b87h29$bju$1@news1.kornet.net> <3EA75010.D9F9C857@engcorp.com>
Message-ID: <AJVpa.3145$%_3.2498395@newssrv26.news.prodigy.com>

"Peter Hansen" <peter at engcorp.com> wrote in message
news:3EA75010.D9F9C857 at engcorp.com...
> "??????" wrote:
> >
> > Plz... answer me...
>
> It sort of depends on the type of application.
>
> Please read http://www.catb.org/~esr/faqs/smart-questions.html
> and try posting again...

That is a truly painful essay.  My favorite part: "Those are loser
attitudes."  The jist, however, can be stated as, "try to ask clear
questions in the future," and is probably good advice.

Anyhow, to answer the OP, assuming that you are comparing the languages to
be used in an active server context, neither should be faster or slower.
Python, however, has some language niceties which may make your code more
efficient, and ultimately faster.  On the downside, Python for asp is not
always ideal and can work (or not work) in unexpected ways.

-d





From nicodemus at globalite.com.br  Tue Apr 15 22:36:30 2003
From: nicodemus at globalite.com.br (Nicodemus)
Date: Tue, 15 Apr 2003 23:36:30 -0300
Subject: FTP
In-Reply-To: <e62b26d3.0304151727.6de2e2e4@posting.google.com>
References: <e62b26d3.0304151727.6de2e2e4@posting.google.com>
Message-ID: <3E9CC1AE.5040508@globalite.com.br>

Matt Singer wrote:

>I have an application with a PYQT interface that uploads large files. 
>Is there an async method to get how much data has been sent with
>storlines to update a progress bar?  Or is there an alternate ftp
>interface?
>

 From the Python docs:

storlines(command, file)
    Store a file in ASCII transfer mode. command should be an 
appropriate "STOR" command (see storbinary()). Lines are read until EOF 
from the open file object file using its readline() method to provide 
the data to be stored.

So you can send any object with a readline method that support the same 
semantics of a file (untested code):


class ProgressFileReader:
    
     def __init__(self, filename, callback):
         self.file = file(filename, 'r')
         self.file_size = os.path.getsize(filename)
         self.callback = callback # will be called with one parameter:
                                  # the percentage of the file read so far

     def readline(self, *args):
         result = self.file.readline(*args)
         percent = self.file.tell()*100/self.file_size
         self.callback(percent)
         return result

And use it as follows:

    f = ProgressFileReader('myfile', mycallback)
    ftpobj.storlines('STOR myfile', f)
     
 
Another option would be to subclass the file object.

Hope that helps,
Nicodemus.






From andrew at acooke.org  Tue Apr 29 19:41:34 2003
From: andrew at acooke.org (andrew cooke)
Date: Tue, 29 Apr 2003 19:41:34 -0400 (CLT)
Subject: Textbooks?
In-Reply-To: <1051658594.16051.0@demeter.uk.clara.net>
References: <oproar1sjtx3zh5c@news.mindspring.com> 
     <80dac0f4.0304281047.9143cd0@posting.google.com> 
     <1051658594.16051.0@demeter.uk.clara.net>
Message-ID: <4415.10.2.0.2.1051659694.squirrel@127.0.0.1>

a printed version, pdf and postscript files are available from
http://www.greenteapress.com/thinkpython/

if you're having problems with the ps version (i don't know what cups is)
then try using psresize from the psutils package to make sure the paper
size is correct (europe uses A4, america uses letter).

andrew

Ken Parkes said:

> On Mon, 28 Apr 2003 11:47:24 +0000, Ionutz Borcoman wrote:
>
>> Hi,
>>
>> Why don't you try "How to Think Like a Computer Scientist: Learning with
>> Python"
>>
>> http://www.ibiblio.org/obp/thinkCSpy/
>>
>> It's free and it's easy reading for people with no programming
>> experience.
>>
>> Ionutz
>
> I hate reading on a screen and when I try to print this book it cuts off
> the last two characters on each line, and the last two lines on each page.
> Does it work for other people?  I've never had this problem with anything
> else.  ( using cups on Mandrake 9 )
>
> Ken.
>
> --
> http://mail.python.org/mailman/listinfo/python-list
>
>


-- 
http://www.acooke.org/andrew




From tjreedy at udel.edu  Mon Apr 21 14:20:28 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Mon, 21 Apr 2003 14:20:28 -0400
Subject: __setitem__ puzzle
References: <b80q7c$cmd$1@news.hccnet.nl>
Message-ID: <dOudnc1k_I7TrjmjXTWcqw@comcast.com>

"Anton Vredegoor" <anton at vredegoor.doge.nl> wrote in message
news:b80q7c$cmd$1 at news.hccnet.nl...
> I have subclassed the builtin list type to accept tuples for
indexing.

As written (and modified by me), it *requires* a sequence for
indexing.
(To relax this, one would have to text index for type==int and branch
accordingly.)
As Aahz correctly said, you should specify what l[tuple] is supposed
to mean.
Reading code and usage reveals the intention that
l[(a,b,c)]==l[a][b][c], etc.

> class tlist(list):
>
>     def __getitem__(self,x):
>         y = self[:]

This is your problem.  By itself, making a shallow copy wastes time
and space.

>         for z in x:
>             y = y[z]

The reason for the copy is to upcast self from a tlist to a list so
that y[x] calls list.__getitem__ instead of recursively calling this
method with x not a sequence.  I believe this would fail anyway if any
of the sublists of self were tlists themselves instead of lists.

>         return y
>
>     def __setitem__(self,x,val):
>         y = self[:]

Here the copying is fatal when len(x)==1

>         for z in x[:-1]:
>             y = y[z]
>         y[x[-1]] = val

because then you only change the temporary instead of self itself.

...
>     L[(1,1)] = 0
>     print L
>     #this doesn't work

'this' is ambiguous -- line above or below?  (ans.: below)
so is "doesn't work" -- wrong answer or exception? (ans.: wrong
answer, I think)

>     L[(1,)] = 0

As explained above, L is not modified, just the internal copy.

The solution is to directly do what you tried to do indirectly by
making the copy.
This also works with tlists as sublists.

class tlist(list):
    def __getitem__(self, x):
        lget = list.__getitem__
        y = self
        for z in x:
            y = lget(y, z)
        return y
    def __setitem__(self, x, val):
        lget = list.__getitem__
        y = self
        for z in x[:-1]:
            y = lget(y, z)
        list.__setitem__(y, x[-1], val)

def test():
    L = tlist(['aa',['bb',tlist(['cc','dd']),'ee'],'ff',['gg']])
    print L
    print L[(1,1,0)]
    L[(1,1,0)] = 0
    print L
    L[(1,)] = 0
    print L

>>> test()
['aa', ['bb', ['cc', 'dd'], 'ee'], 'ff', ['gg']]
cc
['aa', ['bb', [0, 'dd'], 'ee'], 'ff', ['gg']]
['aa', 0, 'ff', ['gg']]

Terry J. Reedy





From pinard at iro.umontreal.ca  Sun Apr  6 20:54:10 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 06 Apr 2003 20:54:10 -0400
Subject: Handling slices
Message-ID: <oqhe9bgma5.fsf@alcyon.progiciels-bpi.ca>

Hi, Python friends.

I'm currently using Python 2.2.1.  How does one support slices when given
to __getitem__?  Suppose I get a slice object.  How can I use it to get
a slice of a local array?  In other words, how one does apply a slice?

The documentation says that __getslice__ is deprecated since 2.0, but then,
I need an alternative way for using it.  What is it?

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From mcfletch at rogers.com  Sun Apr 20 03:21:30 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 20 Apr 2003 03:21:30 -0400
Subject: Sorting without transitivity
In-Reply-To: <rQjoa.37575$LB6.855365@news1.tin.it>
References: <mailman.1050788132.16618.python-list@python.org> <rQjoa.37575$LB6.855365@news1.tin.it>
Message-ID: <3EA24A7A.10900@rogers.com>

For those interested in the algorithm... My 2 toposort implementations 
(one is based on Tim's code, the other's one that I worked up without 
realising it was a topological sort until after it was complete) are 
available here:

    http://members.rogers.com/mcfletch/programming/toposort.py

Much longer than the kjbuckets implementations, however.  The code's 
been stable for a few years now, used in the vrml processing libraries, 
(and if it's still there, the Boa class-graph layout code).

Enjoy all,
Mike

Alex Martelli wrote:

>Frank Niessink wrote:
>  
>
...

>I think what you want is what's known as a *topological
>sort* -- a sort obtaining any ordering that respect certain
>given pairwise constraints that don't, however, define a
>complete order.
>  
>
...

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From mwh at python.net  Mon Apr 14 15:06:51 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 14 Apr 2003 19:06:51 GMT
Subject: Parsing strings?
References: <698f09f8.0304132224.6a28c76f@posting.google.com> <7h3znmtpo0r.fsf@pc150.maths.bris.ac.uk> <ogAma.15965$U34.9541@nwrddc02.gnilink.net>
Message-ID: <7h3smsksxtg.fsf@pc150.maths.bris.ac.uk>

Neil Schemenauer <nas-usenet at arctrix.com> writes:

> Michael Hudson <mwh at python.net> wrote:
> > tweedgeezer at hotmail.com (Jeremy Fincher) writes:
> > 
> >> Is there any way to turn a string (in Python format, escapes and all)
> >> into an actual Python string without using eval()?
> >> 
> >> I.e., I'd like to turn the four-character string "\\x00" into the
> >> one-character Python string "\x00" without using eval().  Is there any
> >> way to do that?
> 
> In Python 2.3 courtesy of MvL:
> 
>     >>> '\\x00'.decode('string-escape')
>     '\x00'

Oh yeah.

> > Don't think so -- and if there is, Lib/compiler/transformer.py would
> > like to know...
> 
> I guess someone should tell it.

Ah, now I remember the problem: PEP 263-style unicode issues.  I still
don't think there's a nice way of solving that one.

Cheers,
M.

-- 
  Screaming 14-year-old boys attempting to prove to each other that
  they are more 3133t than j00.
         -- Reason #8 for quitting slashdot today, from
            http://www.cs.washington.edu/homes/klee/misc/slashdot.html



From skip at pobox.com  Mon Apr  7 20:20:45 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 7 Apr 2003 19:20:45 -0500
Subject: exceptions == errors?
In-Reply-To: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
Message-ID: <16018.5597.933725.584314@montanaro.dyndns.org>

    Mark> Are exceptions always considered errors, or there the same feeling
    Mark> as in C++ that exceptions can be used for non-error situations?

Exceptions are frequently used to signal non-error conditions.  For example,
they can be used to break out of nested loops:

    class Finished(exception): pass

    try:
        for foo in some_list:
            for bar in some_other_list:
                ...
                if conditions_met:
                    raise Finished
                ...
    except Finished:
        pass

Skip




From schulz.online at gmx.de  Thu Apr 24 14:19:23 2003
From: schulz.online at gmx.de (Birgit Lachner)
Date: Thu, 24 Apr 2003 20:19:23 +0200
Subject: Problem with PyQt on SuSE 8.2
References: <b842hl$dsa$1@online.de> <b844nl$fst$1@online.de> <ae5c48d7.0304240631.465b9b0e@posting.google.com>
Message-ID: <b899tl$etc$1@online.de>

Stefan Quandt wrote:

> Birgit Lachner <schulz.online at gmx.de> wrote in message
> news:<b844nl$fst$1 at online.de>...
>> > we try ro run several Python programs which are using PyQt on SuSE 8.2.
>> > Every time the same error message occurs:
>> >  :
>> > We tried the original PyQt packages which comes with SuSE 8.2. We also
>> > have compiled PyQt from the source RPMs, but the same error occurs.
>> > 
>> Okay ... as we compiled the original riverbank-sources the programms are
>> running now.
>> 
>> Bye ... Birgit & Bernd Lachner
> 
> I try to compile the PyQt-x11-gpl-3.5 sources from riverbank but build.py
> gives me many X11 related linker errors from when it tries to link a
> (generated) qttest.cpp (also on SuSE 8.2).
> 
> How did you do this??
Hallo Stefan,

we did the following as root:
>> python build.py -q /usr/lib/qt3/ -l qt
... then call "make" and "make install".

Bye ... Birgit und Bernd Lachner.




From duncan at NOSPAMrcp.co.uk  Tue Apr  1 03:03:22 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 1 Apr 2003 08:03:22 +0000 (UTC)
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <bd993a2f.0303310333.5a95ecb3@posting.google.com> <roy-D262FA.09230131032003@reader2.panix.com> <b6aam1$7o6$1@news.island.liu.se>
Message-ID: <Xns93505AB7F581Cduncanrcpcouk@127.0.0.1>

Thomas Bellman <bellman at lysator.liu.se> wrote in 
news:b6aam1$7o6$1 at news.island.liu.se:

>> I've heard C++ guys bash Python as not 
>> object oriented because Python doesn't support private data.  "You mean 
>> I can bypass the accessor functions and touch an object's data 
>> directly!?"
> 
> Of course, that is possible to do in C++ too.  I would even contend
> that it is actually *easier* to do so by mistake in C++ than in
> Python; but it is easier to do so willingly in Python than in C++.
> (And here is a hint for those that don't believe it can be done in
> C++: pointer arithmetics.)

I found '#define private public' before including the header file worked 
for me and was less messy than pointer arithmetic. (Pedants may wish to 
point out that this isn't guaranteed to work, but in the particular case 
where I had to do it, it worked just fine.)

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From usenet_spam at janc.invalid  Mon Apr 28 21:54:05 2003
From: usenet_spam at janc.invalid (JanC)
Date: Tue, 29 Apr 2003 01:54:05 GMT
Subject: Installer Problem WinNT to Win 98
References: <3EADD5C2.ABC48D5E@easystreet.com>
Message-ID: <Xns936C27AC56317JanC@213.224.134.81>

achrist at easystreet.com schreef:

> Even so, the executable ran fine on my [WinNT] machine.  The executable
> crashed on friend's Windows 98 machine. 

Did you use the Unicode version of wxPython?  When wxPython-Unicode crashes 
on Win98 (it did on my Win98SE), you should use wxPython-ANSI instead...

-- 
JanC

"Be strict when sending and tolerant when receiving."
RFC 1958 - Architectural Principles of the Internet - section 3.9



From peter at engcorp.com  Mon Apr 14 14:06:53 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 14 Apr 2003 14:06:53 -0400
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl>
Message-ID: <3E9AF8BD.BC918E5E@engcorp.com>

Irmen de Jong wrote:
> 
> [offtopic]
> 
> Erik Max Francis wrote:
> 
> >>>>Course(lowEarthOrbit, lowMarsOrbit).transfers()
> >
> > [<ExtractionTransfer @ 0x82893ac (out of `low orbit around `Earth'')>,
> > <HohmannTransfer @ 0x812cef4 (Hohmann from `Earth' to `Mars')>,
> > <InsertionTransfer @ 0x82adb64 (into `low orbit around `Mars'')>]
> >
> 
> Is this actual code? What does this do? I'm interested.

Given that Python is an "agile language", it seems very likely that
it must be an implementation of Robert Zubrin's XP-style "Mars Direct" 
program for a manned mission to Mars.  

<0.9 wink but I wish it were lower-ly y'rs> 

-Peter



From ianb at colorstudy.com  Thu Apr 17 02:39:05 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 17 Apr 2003 01:39:05 -0500
Subject: Python and Schools
In-Reply-To: <3E9E0FBA.6050503@indy.rr.com>
References: <qGqla.61$Tj7.91655@news.uswest.net>
	 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
	 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
	 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
	 <bDSma.22753$T34.687158@news2.tin.it> <slrnb9o8s8.iea.news@datatailors.com>
	 <mailman.1050427771.31620.python-list@python.org>
	 <3E9E0FBA.6050503@indy.rr.com>
Message-ID: <1050561545.710.233.camel@lothlorien>

On Wed, 2003-04-16 at 21:24, Tim Ottinger wrote:
> Donnal Walter wrote:
> > I am a physican first and programmer second. I've never taken a course 
> > in computer science, and I do not know the meaning of O(N). OTOH, just 
> > from following comp.lang.python I think I have a pretty good idea why
> > 
> >     s += str(i)
> > 
> > would be a bad idea. Nevertheless, I certainly would not be opposed to 
> > deepening my understanding. (PythonIAN is great, BTW.)
> 
> I came in late. If we're talking about schools, I'm not sure whether 
> we're talking about colleges and adult ed, or high school, or elementary 
> school. Are we talking about teaching people who know nothing about
> programming?

It's also a question of what you're trying to teach.  Something that has
always impressed me with Logo is that it's more of a teaching
philosophy, and a language goes with it.  The intention behind Logo
isn't to teach programming... rather, programming is a medium in which
to teach all sorts of other great stuff.

I personally believe that programming is *the* way pre-algebra and
algebra should be taught.  Logo shows it works very well for geometry as
well.  But you really have to get rid of the idea that you're teaching
programming -- it does the children a disservice anyway, because
programming is a rather niche skill in the larger picture.  When you're
teaching *with* programming, big-O notation and many other programming
notions are just a distraction.

> Maybe not too bad for 5th or 6th grade, since they're far from writing
> any professional code. You tolerate more with rank beginners, and you
> correct it in time.

I'd go further -- if you're not trying to teach them to become
programmers it's not a matter of tolerating, it's completely okay to be
inefficient.  It'd be like criticizing a student's handwriting on the
caption they made for a picture in art class, or criticizing the
aesthetic of a graph made for algebra.

  Ian






From aleax at aleax.it  Tue Apr 15 10:31:43 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 14:31:43 GMT
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1> <3e9c1413$1@news.012.net.il>
Message-ID: <jJUma.23762$T34.707989@news2.tin.it>

Avner Ben wrote:

> "Duncan Booth" <duncan at NOSPAMrcp.co.uk> wrote in message
> news:Xns935E66BE3E498duncanrcpcouk at 127.0.0.1...
>> "Avner Ben" <avner at skilldesign.com> wrote in
>> news:3e9b0e03$1 at news.012.net.il:
>>
>> That seems very strange to me as lower casing the strings has the
>> potential to be much more efficient than a case insensitive compare even
>> when the compare is done by a C routine. Perhaps if you posted a general
>> outline of your code we could suggest ways to speed it up. The biggest
>> potential for code speedup is often to use a different algorithm rather
>> than trying to optimise individual operations.
>>
> 
>     I am searching for one or more members of a list of names (that may
>     get
> near 1000 pieces in a normal application and may average some 40
> characters) in a string (that may average 100 characters). For each name
> found in the string, I do something to the string, using the original name
> (in its proper case). At present, I am looking for the exact name as

Sounds like a good application of regular expressions.

import re
namesre = re.compile('(%s)'%')|('.join(listofnames), re.I)

for m in namesre.finditer(astring):
    dosomething(astring, listofnames[m.lastindex-1])


You'll have to measure the performance for yourself, but you may
find it quite good -- do, and let us know!


Alex




From jjl at pobox.com  Tue Apr 22 16:07:38 2003
From: jjl at pobox.com (John J. Lee)
Date: 22 Apr 2003 21:07:38 +0100
Subject: Off topic: What usenet news reader to use?
References: <mailman.1051036295.31460.python-list@python.org>
Message-ID: <8765p647p1.fsf@pobox.com>

"Wai-Yip Tung \(wtung\)" <wtung at cisco.com> writes:

> I haven't used Usenet news for a long time. Now that I have subscribed
> to python-list I find that email does not scale well to the high volumne
> of discussion. Instead of deleting emails all day I think I should go
> back to Usenet news. Is there any good Windows news reader you would
> recommend. I'm happily using Opera as the browser. But it's
> revolutionary news and email reader eludes me. What are you using?

Just started using GNUS -- very nice if you already use emacs.  Better
than pine for reading news over a dialup connection, and does
threading better, too.  Not Windows-specific, but both work fine in
Windows.


> By the way I'm also looking for a free or low cost email client. My
> company is using Outlook but I want to find one for personal use. Thank
> you for you suggestions.

Depends what you like.  I like to use IMAP so I can get at my mail
from anywhere.  Pine and Mulberry do IMAP OK.  Pine is very easy to
use (it is once set up, anyway), does (almost) everything I want and
is text menu-based; also does newsreading.  Mulberry is a nice, but
non-free, graphical client with masses of bells and whistles,
including disconnected operation (disconnected mode was a little buggy
about five months ago, not sure how it is now).

fastmail.fm is a good IMAP service provider -- once-off fee for decent
bandwidth and space plus a webmail system to fall back on, repeating
fee if you want support and more bandwidth / space.  Good filtering
system using the Sieve language.


John



From peter at engcorp.com  Wed Apr  2 15:25:45 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 15:25:45 -0500
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org>
Message-ID: <3E8B4749.E778873F@engcorp.com>

danny at iamlearning.com wrote:
> 
> I'm new on this, please bear with we.
> 
> Let's say I have this script:
> 
> --------------------------
> #!/usr/bin/env python
> x = 'Hello Python'
> 
> print x,
> 
> --------------------------
> 
> How can I compile this script to *.pyo?

What's the name of the script?



From afd00 at aub.edu.lb  Thu Apr  3 04:14:42 2003
From: afd00 at aub.edu.lb (Ali Dada)
Date: Thu, 3 Apr 2003 09:14:42 GMT
Subject: Tkinter question
Message-ID: <E1910ny-0003Rc-00@fast.aub.edu.lb>

hi all:

i have a window made of a couple of buttons stacked above each other.i need to
any a widget on the top that opens a submenu to the right when clicked. how can
i do that? see the code below and tell me how to correct it (the submenu is
opening downwards over the other buttons, not to the right)

#!/usr/bin/env python
import sys
from Tkinter import *
from tkMessageBox import *
from tkSimpleDialog import askfloat


def makemenu(root):
    pref = Menu(root)
    root.config(menu=pref)

class Index(Frame):
	def __init__(self):
		Frame.__init__(self)
		self.pack()
		Button(self, text='Repository').pack(side=TOP, fill=BOTH)
		Button(self, text='Regression').pack(side=TOP, fill=BOTH)
		Button(self, text='Code Manage').pack(side=TOP, fill=BOTH)
		

	def makemenu(self, root):
		pref = Menu(root)
		root.config(menu=pref)
		pref1 = Menu(pref, tearoff=0)
		pref1.add_command(label='Cut', underline=0)
		pref1.add_command(label='Paste', underline=0)
		pref.add_cascade(label='Preference >',  menu=pref1,  underline=0)

if __name__ == '__main__':
	root=Tk()
	root.title('TAGZ')
	i = Index()
	i.makemenu(root)
	i.mainloop()









From ianb at colorstudy.com  Tue Apr 22 13:07:21 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 22 Apr 2003 12:07:21 -0500
Subject: Overloading and?  was <RE: Should I prefer an external
	database>
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DE065@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE065@admin56.narex.com>
Message-ID: <1051031240.697.138.camel@lothlorien>

On Tue, 2003-04-22 at 09:52, Bjorn Pettersen wrote:
> > for paper in Paper.select(AND(Paper.q.title == 'foo',
> >                               Paper.q.author.startswith('Bob'))):
> 
> Does anyone know the reason for not allowing an overload of the and
> operator? (of course, the only use case I've seen is the one above
> [thrice, first time in 97] and alternate logical calculi [e.g. fuzzy
> logic] -- so not much different from some of the other operators? :-)

Certain operators (and, or, not, is) can't be overloaded.  (Actually, up
until 2.1, the kind of overloading I'm doing with the comparison
operators wasn't possible)

For and, or, and not you can overload (I believe the name is)
__nonzero__, and they use that to determine if the value should be
considered true or false.  But for an expression generator like this I
need to overload each operator so it returns a specific value, not just
True or False.  In this case a == b returns an object that evaluates to
"a = b" (for use in SQL).

  Ian






From martin.vonloewis at hpi.uni-potsdam.de  Thu Apr 24 09:13:14 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Thu, 24 Apr 2003 15:13:14 +0200
Subject: SGMLParser questions
References: <igmbavg5ughls6ksiden2rgnael1s0ub8q@4ax.com> <m3n0ihdb80.fsf@mira.informatik.hu-berlin.de> <698f09f8.0304240053.1b9e38e9@posting.google.com>
Message-ID: <b88ntd$5uf$1@zeppelin.rz.uni-potsdam.de>

> It seems it'd be easy enough to add a "case sensitive" keyword
> argument to __init__ and sprinkle a few if statements before the
> .lower()s on strings.  Would such a patch be accepted?

That would be good enough to make it case sensitive, yes;
it would not be sufficient to support XML namespaces.

Adding case sensitivity also affects the names of
the handlers (do_*): should they be case-sensitive as well,
or should they still get lower-cased?`

If this issue is resolved, and the patch comes with
documentation and test cases, I would accept it (not for 2.3
though, anymore)

Regards,
Martin





From jimhill at swcp.com  Sun Apr 27 19:12:12 2003
From: jimhill at swcp.com (Jim Hill)
Date: Sun, 27 Apr 2003 23:12:12 +0000 (UTC)
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <a5415304.0304271335.2f569879@posting.google.com>
Message-ID: <b8ho4c$9ct$1@iruka.swcp.com>

Ron Stephens wrote:

>you can find [my Python book] reviews at www.awaretek.com/plf.html

"Gee, Gramma, what big text you got!"

"The better to blind you with, my darling!"


Jim
-- 
The freedom of speech guaranteed by the First Amendment means 
you're free to write "speach" and "ammendment", but it doesn't
mean the rest of us won't snicker at you for it.



From aleax at aleax.it  Wed Apr 23 06:57:10 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 10:57:10 GMT
Subject: Setting the Title of the script window
References: <1814331d.0304230219.3c6a82de@posting.google.com>
Message-ID: <akupa.83076$T34.2073301@news2.tin.it>

Xnedra wrote:

> Hi there,
> 
> is there a possibility to set the title of the window the python
> script was started in?
> I know it is possible with dos with these lines:
> 
> if os.name == "nt":
>   cmdline = 'TITLE TestTitle'
>   os.system(cmdline)
> 
> but i would like to set it independent of the operating system. Is
> there a python command for this or does anyone know the corresponding
> commandline for unix?

Unfortunately (in this context), "unix" does not have a single way
to implement "text windows" -- there are many such beasts!

Fortunately, most such beasts emulate to some extent the oldest of
them all, 'xterm'.  To set the window title under xterm, you must
print to the terminal a string with some escape sequences --

def set_xterm_title(title):
    print '\033]0;%s\007' % title

This will also work on a KDE "Konsole" and no doubt several other
terminal emulators -- but, I think, not ALL of them.


Alex




From peter at engcorp.com  Sat Apr 26 13:25:10 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sat, 26 Apr 2003 13:25:10 -0400
Subject: GUI testing with python  (SendKeys)
References: <b8c6c9$p2m$0@198.102.102.163> <b8ecgs$47i$1@sirius.dur.ac.uk> <3EAAC082.1DD45058@engcorp.com>
Message-ID: <3EAAC0F6.81B70C87@engcorp.com>

Peter Hansen wrote:
> 
> Christopher Saunter wrote:
> >
> >   You might want to check out Girder - http://www.girder.nl
> > It offers more robust and flexible ways of remote controlling Win32 GUI
> > stuff than sendkeys.  There is a very noddy Python client for talking to
> > Girder at http://www.dur.ac.uk/christopher.saunter/soft.html
> 
> "Noddy"?  Please translate. :-)

Never mind... dict.org handles that too.  I should have checked first.

noddy
  
     /nod'ee/ [UK: from the children's books] 1. Small and
     un-useful, but demonstrating a point.  Noddy programs are
     often written by people learning a new language or system.
     The archetypal noddy program is hello, world.  Noddy code
     may be used to demonstrate a feature or bug of a compiler.
     May be used of real hardware or software to imply that it
     isn't worth using.  "This editor's a bit noddy."
  
     2. A program that is more or less instant to produce.  In this
     use, the term does not necessarily connote uselessness, but
     describes a hack sufficiently trivial that it can be written
     and debugged while carrying on (and during the space of) a
     normal conversation.  "I'll just throw together a noddy awk
     script to dump all the first fields."  In North America this
     might be called a mickey mouse program.  See toy program.


Actually, which one did you mean?

-Peter



From vze4rx4y at verizon.net  Fri Apr 25 00:07:20 2003
From: vze4rx4y at verizon.net (Raymond Hettinger)
Date: Fri, 25 Apr 2003 04:07:20 GMT
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth> <+xKq+ks/K/pQ089yn@the-wire.com>
Message-ID: <Yv2qa.32862$ot1.31621@nwrdny02.gnilink.net>

[Jim]
> >I have come across a weirdness. I have a number of floats, (lat long
> >pairs) which I stuff into a dict, but when I take them out, my nice %2f4
> >number, (say, 47.5894) is something long and *almost* right, like
> >47.589333333333333339. What gives?

[Mel]
>    Chances are that the float hasn't changed, but that
> you're printing it in a different way.


The Python FAQ has a pretty good explanation:

http://www.python.org/cgi-bin/faqw.py?req=show&file=faq04.098.htp


Raymond Hettinger





From anton at vredegoor.doge.nl  Tue Apr 15 21:51:17 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Wed, 16 Apr 2003 03:51:17 +0200
Subject: Big-O notation (was: Re: Python and Schools)
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <3E9C6C62.8010808@rogers.com> <mailman.1050441947.6447.python-list@python.org>
Message-ID: <b7id3q$3s4$1@news.hccnet.nl>

Chad Netzer <cnetzer at mail.arc.nasa.gov> wrote:

>Also, one further class that is import to know is O( 2**N ), such as the
>brute-force solution to the "travelling salesman problem".  These are

Although it's important to not set too severe limitations I have the
impression this salesman is underestimating the problem.

>>> def fac(n):
...     return reduce(operator.mul,range(2,n+1),1)
... 
>>> fac(10)
3628800
>>> 2**10
1024

Anton



From alanpong at hkstar.com  Thu Apr 17 23:06:39 2003
From: alanpong at hkstar.com (Alan Pong)
Date: 17 Apr 2003 20:06:39 -0700
Subject: How to Import dll inside other directory
Message-ID: <ea0e2298.0304171906.37cb0b8a@posting.google.com>

a C extension dll (myext1.dll) compiled by vc++ and placed in c:\

run python 2.3.exe under c:\ in dos prompt, 
>>from myext1 import *
ok

then, move the dll to c:\temp, run python 2.3.exe again under c:\,
>>from temp.myext1 import * (according to the help doc.)
error:
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
ImportError: No module named temp.myext1

Does anybody know how to import dll inside other directory?
Thanks.
Rgds.
Alan



From max at alcyone.com  Wed Apr 30 04:11:58 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 30 Apr 2003 01:11:58 -0700
Subject: Self modifying systems question
References: <mailman.1051685036.19081.python-list@python.org>
Message-ID: <3EAF854E.79D4F4A@alcyone.com>

Sten Kvamme wrote:

> Self modifying code is a well known technology. Has anyone done this
> in
> Python? What I'm seeking for is examples of software that adds to
> itself (methods and variables) in runtime. I'd like an instantiated
> object to
> be able to add a method to itself.

Oh yeah, Python can do this just fine.  Dynamic attribute addition is
actually the normal way things are done; after all, that's all that an
__init__ method does (provided the class doesn't have __slots__) -- it
adds attributes.  It's customary, for maintainability's sake, to add
everything all at once, but there's no reason you can't add more later.

Same thing with "methods" -- all you really need is an attribute that's
callable.  If you want something that has access to the self instance,
you can just pass that along as well.

>>> class C: pass
... 
>>> c = C()
>>> c
<__main__.C instance at 0x8159244>
>>> dir(c)
['__doc__', '__module__']
>>> c.x = 1
>>> c.x
1
>>> c.f = lambda t: t**2
>>> c.f(2)
4
>>> c.f(3)
9
>>> c.f(4)
16
>>> c.xSquared = lambda self=c: self.x**2
>>> c.xSquared()
1
>>> c.x = 10
>>> c.xSquared()
100


-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Once the people begin to reason, all is lost.
\__/ Voltaire
    Bosskey.net: Quake III Arena / http://www.bosskey.net/q3a/
 A personal guide to Quake III Arena.



From ChuckEsterbrook at yahoo.com  Mon Apr 14 02:24:54 2003
From: ChuckEsterbrook at yahoo.com (Chuck Esterbrook)
Date: Sun, 13 Apr 2003 23:24:54 -0700
Subject: properties & exceptions
References: <3e9a0144$1@news.microsoft.com> <mailman.1050282441.10913.python-list@python.org>
Message-ID: <Tusma.2587$%A2.75@fed1read06>

On Sunday 13 April 2003 05:30 pm, Saveen Reddy [MSFT] wrote:
> I was suprised by a python behavior I saw concerning exceptions thrown
> when accessing a class through a property. Essentially, I am seeing an
> unexpected exception, one that implies the property doesn't exist when
> in fact it does. I haven't found any docs that explain this behavior
> and so my questions are (1) is this behavior "by-design"? and (2)
> Besides not using properties, is there a workaround?

On Sunday 13 April 2003 06:05 pm, Mike C. Fletcher wrote:
> Any chance you don't have a __metaclass__ = type declaration?  You
> don't subclass object in the example you cite, and without one of
> those properties just don't work.

I reported this behavior as a bug some time ago, but was told it wasn't 
a bug because I didn't subclass object. However, it's very deceptive 
and Python usually either works or gives a meaningful exception, so 
let's call it a "flaw". For example, if properties just don't work in 
classes not derived from object, then the Pythonic thing to do would be 
to raise a CannotUsePropertiesInOldClasses exception when trying to 
create them.

In any case, the reason I'm replying here is to report that in Python 
2.3a2, the bug, er I mean the flaw, has been fixed: exceptions in 
properties propagate correctly regardless of the base classes. That's 
great since it will reduce confusion. Plus, I like to use properties in 
wxPython subclasses, but cannot currently change wxObject to inherit 
object.

-- 
Chuck
http://ChuckEsterbrook.com




From nilsgri at idi.ntnu.no  Mon Apr  7 07:36:03 2003
From: nilsgri at idi.ntnu.no (Nils Grimsmo)
Date: Mon, 07 Apr 2003 13:36:03 +0200
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no> <3E915BA0.A4B5286@engcorp.com>
Message-ID: <pan.2003.04.07.11.36.03.78556@idi.ntnu.no>

On Mon, 07 Apr 2003 07:06:08 -0400, Peter Hansen wrote:
> Nils Grimsmo wrote:
>> 
>> class MainWindow(wxFrame):
>>     label = ["foo", "bar"]
>>     button = None
>>     def __init__(self, parent, id, title):
>>         wxFrame.__init__(self, parent, -4, title, size = (100, 100))
>>         self.button = wxButton(self, 13, self.label[0])
>>         EVT_BUTTON(self, 13, self.run)
>>         self.Show(true)
>> 
>>     def run(self, event):
>>         for c in range(2):
>>             for x in [0, 1]:
>>                 self.button.SetLabel(self.label[x])
>>                 # please update. pretty, please!
>>                 time.sleep(1.0)
> 
> When do you think/hope run() will be called?  I don't think that
> method is looked for by wxPython...  Did you base this on 
> some sample code?

run() is called when you push the button. run() sets the label for the
button every second for four seconds, but the button label is not updated
before you return from run().

that is not the behavior i want. any suggestions?


klem fra nils
 
-- 
nils grimsmo
nilsgri at idi.ntnu.no




From jack at performancedrivers.com  Wed Apr  2 19:17:41 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Wed, 2 Apr 2003 19:17:41 -0500
Subject: When Python outruns C++ ...
In-Reply-To: <miXIvJAi02i+Ew2O@jessikat.fsnet.co.uk>; from robin@jessikat.fsnet.co.uk on Thu, Apr 03, 2003 at 12:07:14AM +0100
References: <mailman.1048973984.17441.python-list@python.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com> <wBIia.2000$y3.310439@news010.worldonline.dk> <b6fmcq$dh0$1@panix2.panix.com> <miXIvJAi02i+Ew2O@jessikat.fsnet.co.uk>
Message-ID: <20030402191741.C936@localhost.localdomain>

On Thu, Apr 03, 2003 at 12:07:14AM +0100, Robin Becker wrote:
> Interestingly attempts to get an integer for loop (without range/xrange)
> are met with extreme disapproval by the language gurus and the fastest
> looping constructs of all map, filter et al are considered bad form.

I prefer map and filter all the time, but they aren't always the fastest.
List comps are faster for simple tasks and if you need both the IN list
and the NOT IN list, using a for loop is faster than two filters (unsuprisngly)

-jackdied




From exarkun at intarweb.us  Sat Apr 19 12:59:23 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sat, 19 Apr 2003 12:59:23 -0400
Subject: Parent references: hot or not?
In-Reply-To: <1050772061.14712.55.camel@ida.dylanreinhardt.com>
References: <1050763980.14721.41.camel@ida.dylanreinhardt.com> <20030419145756.GA25916@meson.dyndns.org> <1050771656.14717.48.camel@ida.dylanreinhardt.com> <1050772061.14712.55.camel@ida.dylanreinhardt.com>
Message-ID: <20030419165923.GA26681@meson.dyndns.org>

On Sat, Apr 19, 2003 at 10:07:41AM -0700, Dylan Reinhardt wrote:
> On Sat, 2003-04-19 at 10:00, Dylan Reinhardt wrote:
> 
> > Sounds like I should be sure to write a __del__ method for Container
> > that ensures things work correctly.
> 
> And upon re-reading your post you say *not* to do that... :-)
> 

  Glad you noticed that ;)

> I'd like to understand that a bit better.  What would be wrong with:
> 
> class Container:
> 
>    def __del__(self):
>       # ensure all Items are deleted first
>       for item in self.items.keys():
>          del self.items[item]
> 

    >>> import gc
    >>> gc.set_debug(1)
    >>> class Foo:
    ...     pass
    ...
    >>> a, b = Foo(), Foo()
    >>> a.b, b.a = b, a
    >>> del a, b
    >>> gc.collect()
    gc: collecting generation 2...
    gc: objects in each generation: 0 0 1215
    gc: done, 4 unreachable, 0 uncollectable.
              ^^^^^^^^^^^^^  ^^^^^^^^^^^^^^^

  Now, with __del__:

    >>> class Foo:
    ...     def __del__(self):
    ...         print 'Foo.__del__'
    ...
    >>> a, b = Foo(), Foo()
    >>> a.b, b.a = b, a
    >>> del a, b
    >>> gc.collect()
    gc: collecting generation 2...
    gc: objects in each generation: 0 0 1217
    gc: done, 4 unreachable, 4 uncollectable.
              ^^^^^^^^^^^^^  ^^^^^^^^^^^^^^^
    >>> gc.garbage
    [<__main__.Foo instance at 0x81b9d6c>, <__main__.Foo instance at 0x81b9d94>]


  As you can see, __del__ is never called, and a and b still exist as
members of the gc.garbage list.  There are some good reasons behind this,
but they are beyond the scope of this post ;)  If you pick up any good
reference on garbage collection, this should be covered.

  What it is possible to do, if you are very careful about it, is to examine
the objects in gc.garbage, break their cycles, remove them from gc.garbage,
then call gc.collect() again.  Since they are no longer in reference cycles,
they can be finalized and destroyed normally.

  I'm not sure if I've answered your original question.  Without more
details, I'm not sure what else to say but this: if you don't want things to
be collected and finalized, don't delete your references to them; if you do
want them to be collected (and maybe finalized), don't put them in cycles,
or don't implement __del__.  Or, go to the trouble described in the above
paragraph.

  Hope this helps,

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 30 days, 13:02, 4 users, load average: 0.29, 0.22, 0.16




From theory79 at yahoo.com  Wed Apr  9 12:17:40 2003
From: theory79 at yahoo.com (Hugo Liu)
Date: 9 Apr 2003 09:17:40 -0700
Subject: optimizing memory usage of large in memory structures
Message-ID: <aebfcc5a.0304090817.2c683c9e@posting.google.com>

hi,

i'm hoping that someone out there has a solution to hack this problem.

working with corpus analysis and large knowledge bases, i often have
to load large arrays into memory.  each array element is just a text
string.  although the text is only 5mb, python consumes 40+ mb.  i
would suspect that the overhead is from storing type info for each
element.  if the strings were all fixed length, i'd just store them
ordered and fixed width into a single string and use bisect, but they
aren't fixed width.

does anyone know if there is a way to fake static typing for this
situation or any other way to make this more efficient (without
sacrificing operating system interoperability or having to write
external C code?)  any suggestions or pointers would be much
appreciated!

hugo



From scottprive at earthlink.net  Wed Apr 16 09:01:46 2003
From: scottprive at earthlink.net (Scott Prive)
Date: Wed, 16 Apr 2003 09:01:46 -0400
Subject: win32net.NetUseAdd syntax, possible bug
Message-ID: <003301c30418$568c7e40$0601a8c0@scottprive.com>

Hello,

I'm attempting to use win32net.NetUseAdd, had a problem and I think there's
a bug in the API, or the docs (or me! :-)

The ActiveState made it clear most of the properties need to be in a
dictionary. It lacked an example use, but referred me to the MSDN site, and
at MSDN could see there the data structure. In theory I should combine the
two docs and know all the things I can pass into the call via dictionary.

Small problem: the MSDN data structure calls the PW key "password", and
win32net.NetUseAdd does not understand this. It looks like the function
needs a key called "passwd", and if not provided it falls upon some default
value (or the current credentials), making it appear that your user/pass
values are not valid on the server when in fact they were.

This is my first post to the list (cc: me, I'm not subscribed). Is this the
appropriate place for possible bug reports? How do I determine a module's
"owner", and should I mail them directly?

The code looks like:

    import win32net
    data = {}
    data['remote'] = r'\\mypdc.example.com\myshare'
    data['local'] = 'Z:'
    data['password'] = 'foo' #Hmm.. I should call this     data['passwd'] =
'foo'
    data['user'] = 'foo'
    data['asg_type'] = 0
    win32net.NetUseAdd(None,1,data)

-Scott








From olczyk at interaccess.com  Sun Apr 20 22:23:42 2003
From: olczyk at interaccess.com (Thaddeus L Olczyk)
Date: Mon, 21 Apr 2003 02:23:42 +0000 (UTC)
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <b7m546$2a83$1@news.peterlink.ru> <lyptnlxq1n.fsf@cartan.de> <m24r4xdyd6.fsf@mycroft.actrix.gen.nz> <lyd6jlxkpi.fsf@cartan.de> <usft9v01l8fcdg8erdl2kishg0cnr1t9dp@4ax.com> <lyadepp1ck.fsf@cartan.de>
Message-ID: <b7vkne$cg0$1@wolfberry.srv.cs.cmu.edu>

On 17 Apr 2003 17:53:15 +0200, Nils Goesche <cartan at cartan.de> wrote:

>Thaddeus L Olczyk <olczyk at interaccess.com> writes:
>
>> On 17 Apr 2003 16:27:53 +0200, Nils Goesche <cartan at cartan.de> wrote:
>>=20
>> >A definition that seems sensible to me would be something like: A
>> >language is functional if it somehow contains the lambda calculus.
>> >If you can take lambda calculus formulae and readily translate them
>> >into working code.
>
>> That makes nearly every modern language functional.
>
>Certainly not Java or C# ;-)=20
Yes. Both contain lambda calculus. As does any language
which supports objects.

In ( a sort hand wavy way ),
class Lambda
{
public:
        retval exec(arglist){...}
}

Java ( can't say I know enough about C# ) may lack some syntactic
sugar to make it pretty, but it's there.

>But even if that's true -- then so be it.
>If it is true that modern languages tend to be functional, too, then
>you could call that progress.  Why should we change the definition of
>``functional=B4=B4 every year in order to keep the ratio of functional t=
o
>non-functional languages constant over time?
>
While it's true that in math and CS, many definitions evolve over
time ( ones that come to mind right now are function/procedure
[which were distinct, now considered the same thing], fiber bundles,
sheaves, topological space ... ), that's all just bullshit thrown out
to distract people from *your stupidity*.

That's because the definition that you have suggested is not some
formal definition appearing in seminal "Introduction to Functional
Programming" type books or papers. It's *your* definition, and it
sucks. Because it's your definition and not a formal one, there is
no suggestion that I am changing the definition over time. As I said,
it's simply a distraction to hide the fact that your definition
*sucks*.

As for your definition, it sucks because it fails to make any
distinction. Almost everyone would agree that languages like
Haskell, ML, Clean, Scheme and  ( sometimes ) Lisp are functional.
And your definition includes those. But your definition would say that
languages like C, Eiffel and Smalltalk ( not to mention C++, Java and
C# ) are functional, which almost no one would agree with.

>> >And you have to be a bit relaxed when using this requirement; for
>> >instance, writing an Y-combinator in ML is very hard because of the
>> >stupid type checker.
>>=20
>> No it is easy.
>>=20
>> In OCaml,=20
>>     let rec fix f x=3Df (fix f) x;;
>
>That's cheating and you know it :-)
>
No. It's not.
It's a standard definition of Y in a combinatorial approach
to computing.

>> For a more traditional approach.
>> type 'a rf =3D RF of ( 'a rf -> 'a );;
>>=20
>> let y f x=3D=20
>>   let g (RF h) z=3D
>>       f( h (RF h) )  z in
>>   (g (RF g)) x;;
>>=20
>> Of course the example above looks like:
>>    let fact x=3D y ( fun q->
>>                              (fun n->=20
>>                                 if n<2 then 1=20
>>                                 else n* (q (n-1)))) x;;
>>=20
>> When I did it this way the first time it was a bit difficult getting
>> the type right.
>
>I know: Been there done that ;-\  Took me some time. =20
What can I say. Some people are slower than others.

>Certainly /much/
>longer than simply translating the lambda calculus formula to Lisp, so
>I refuse calling it easy, even if the ML code that finally works is
>short.
And know we see why some people are slower.
I never even thought of "translating" the code.
I did what every one is taught in every math class.
I rederived it in ML. In the process I found it easy.
The only sticking point was that I did not understand
user defined types well enough to write HOFs. Once
I understood that it was trivial.

>
>> But the reason was that I was new to the OCaml typing system. It
>> wasn't what the type should look like, it was just getting the
>> declaration right.
>
>If it's so easy, why did they even bother putting it into the ML FAQ?
Because there are lazy people who don't write code from first
principles, or bother to learn enough about user defined types
to write HOFs ( the true FP practitioner doesn't just use HOFs
he writes them too ).

>
>> While not an expert, I should say this type of infinite cascading
>> type and data is typical of a class of ML programs. I believe the
>> little MLer has an example early on.
>>=20
>> This should also be easy for a Lisper/Schemer to grasp as it just
>> extends on some of their ideas.
>
>Sure, once you know how to do it, it seems easy and obvious.  But that
>is true about almost all problems.
No. It's easy and obvious because I learned enough OCaml to make it
easy and obvious. These are aspects of the language which are
fundamental, and anyone taking a professional approach to programming
in ML should be familiar with it enough to understand how to write
HOF's. It's only hackers who have a hard time implementing Y in ML.


--------------------------------------------------
Thaddeus L. Olczyk, PhD
Think twice, code once.




From spamfilter2 at mupp.net  Thu Apr 17 12:28:43 2003
From: spamfilter2 at mupp.net (j2)
Date: Thu, 17 Apr 2003 16:28:43 GMT
Subject: FInd nic speed/duplex (Win32)
Message-ID: <Xns9360BC0A648C4spamfilter2muppnet@195.67.237.51>

Is there any sample code to be able to find NIC settings (speed and duplex)   
for a NIC in a Win98 system?



From max at alcyone.com  Fri Apr  4 17:36:44 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 04 Apr 2003 14:36:44 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122> <3E8CF4B5.1E221DD7@alcyone.com> <9383481e.0304040445.a3752dd@posting.google.com>
Message-ID: <3E8E08FC.FB89E901@alcyone.com>

Drew Moore wrote:

> Can we follow up on this a little?
> I'm trying to understand this concern in more detail..
> 
> I'd like this PEP to go one of two ways..
> either implementation,
> or a good solid technical reason why its a bad idea.
> 
> so far, the response seems to be:
> "it is technically possible, but my scripts might
> run slower.. and it just makes me nervous."

Well, your scripts would run slower, but I think the real concern is
about the readability problems this might cause.

> I suspect there is some basis for this nervousness, but
> I still haven't seen an example of how this proposed feature
> would break things.
> 
> since no current code uses __assign__,
> nothing would break immediately.
> 
> however, once it started to be used.. what would go wrong??

It isn't that it wouldn't break preexisting code, it's that after its
introduction, suddenly previously meaningful code can take on entirely
new meanings.  And since Python is dynamic, there's no way to know in
other code how this change might affect it.  Take for instance:

	def f(x=0):
	    if not x:
	        x = 1
	    # ... do something with x

While at first this appears a contrived example, this kind of thing
happens all the time -- say, you're getting some input externally, and
you're clamping it to within bounds, for example.  It is simple and safe
because you can trust in the "x = 1" statement rebinding the _local
name_ x and not having any effect outside the scope of the function call
f.

Now let's say your proposal is passed and implemented.  Now this "x = 1"
statement might _not_ do the rebinding after all.  One would hope that a
well-behaved __assign__ method would make this all okay, but then again
the point of an __assign__ method is that you can make it do whatever
you want.

The down side isn't that this will immediately break huge amounts of
code, it's that suddenly very simple, almost ubiquitous code can
suddenly take on a subtly different meaning, and because of Python's
dynamicism, it's hard to say when the effect will change the meaning.

I suspect this falls into the same class of "Okay this changes the
meaning of everything" like lazy evaluation or macros.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If the sky should fall, hold up your hands.
\__/ (a Spanish proverb)
    The laws list / http://www.alcyone.com/max/physics/laws/
 Laws, rules, principles, effects, paradoxes, etc. in physics.



From ianb at colorstudy.com  Thu Apr 17 02:17:36 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 17 Apr 2003 01:17:36 -0500
Subject: Python in HTML files
In-Reply-To: <_2ednU2VIv5QqQOjXTWcqw@comcast.com>
References: <mailman.1050550654.3607.python-list@python.org>
	 <_2ednU2VIv5QqQOjXTWcqw@comcast.com>
Message-ID: <1050560256.719.188.camel@lothlorien>

On Thu, 2003-04-17 at 00:00, Em wrote:
> I'm trying to get Python commands to run from within HTML files (a la 
> PHP) or at least simulate that behavior.
[...]
> So my question is, do you know of any technologies like this that are 
> not Servlets and do not copy files outside a locked-down user's profile?

You might try Spyce: spyce.sf.net

In your situation you'd probably be forced to put a line like
    #!/home/me/cgi-bin/spyce_something.py
at the top of every file.  Though maybe there's other tricks.

  Ian






From NineOfSix at gmx.de  Sat Apr 26 06:50:48 2003
From: NineOfSix at gmx.de (Axel Grune)
Date: Sat, 26 Apr 2003 12:50:48 +0200
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com> <S5tqa.9420$K35.308092@news2.tin.it>
Message-ID: <b8do9j$91iv5$1@ID-126101.news.dfncis.de>

Alex Martelli wrote:
> we all know how teenagers tend to rebel against parental authority), the

...and then they start to do drugs and program Perl! Jesus, save them! ;-)

Axel




From skip at pobox.com  Tue Apr 29 09:52:41 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 29 Apr 2003 08:52:41 -0500
Subject: SV: Function to convert fetchone() result from a list to a dictionary
In-Reply-To: <BHEFKIOIMEDGILGDACKDKEBEDFAA.carsten@gehling.dk>
References: <b8lfrr$21c$1@nntp0.reith.bbc.co.uk>
        <BHEFKIOIMEDGILGDACKDKEBEDFAA.carsten@gehling.dk>
Message-ID: <16046.33705.479451.992753@montanaro.dyndns.org>

    Carsten> Hmm... Nice. However, is this MySQLdb.cursors.DictCursor
    Carsten> available in other DB modules - say, if I someday changed to
    Carsten> (shudder) Postgres or Oracle?

Yes, though in a slightly different manner.  In Psycopg (one of a couple
PostgreSQL wrappers), instead of specifying the type of cursor at database
connection time, you call a different fetch function, e.g., c.dictfetchone()
or c.dictfetchall().

Skip




From kkto at csis.hku.hk  Fri Apr 18 04:38:29 2003
From: kkto at csis.hku.hk (Isaac To)
Date: 18 Apr 2003 16:38:29 +0800
Subject: Memory leak while looping
References: <1020cbf8.0304162100.6fc49e6f@posting.google.com> <1020cbf8.0304171900.110a1675@posting.google.com>
Message-ID: <7ihe8wfbei.fsf@enark.csis.hku.hk>

>>>>> "Roger" == Roger Hancock <roger_hancock at hotmail.com> writes:

    Roger> Thanks for the help.  The xrange call solved my problem.
    Roger> However, I'm still a little puzzled by the behavior I'm seeing
    Roger> with the range call.  With my now better understanding of the
    Roger> range call, I would expect to see my function take a big chunk of
    Roger> memory until the function exited.  Instead, the function
    Roger> continues to grab more and more memory.  First, it takes most of
    Roger> my physical memory (256 MB) and then proceeds to eat up my swap
    Roger> memory.  Eventually my machine reboots.  Pretty impressive, huh?

In your original program, the range() call never terminate.  It tries to
grab a huge (100M integers with each 16-bytes would mean 1.6G bytes) memory
which is allocated in the VM, and then try to initialize every of the cell
by writing different numbers into them.  If your computer does not try to
limit the amount of memory allocated for each process, all the time will end
up used for swapping, and at the end the swap is filled and your program
will be terminated.  But at the time when the swap is filled, other program
might also fails, because they never expects a full swap space.  This might
explain why you see the computer reboots.

    Roger> Just imagine the damage I could do with 4 lines of code.

You don't need 4 lines.  You need just the expression range(1,100000000).
Nothing more.  You can try running Python and type that into the prompt to
see the behaviour.

If you don't like that, you can limit the amount of memory for each process.
Different OS has different way to do that, and I only know that for
Unix/Linux (which is the only OSes I use) I will use "ulimit -v 40960" to
limit all process to use no more than 40Mi bytes of virtual memory.  Then
instead of killing other programs, the OS just kill the program saying that
the memory limit has been exceeded.

Regards,
Isaac.



From do-not-spam-ben.hutchings at businesswebsoftware.com  Wed Apr  2 08:58:18 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 2 Apr 2003 13:58:18 GMT
Subject: How to obtain python file path
References: <20030401163212631-0500@news.covad.net> <slrnb8jusc.5u1.ramen@lackingtalent.com> <20030401165546102-0500@news.covad.net>
Message-ID: <slrnb8lr3q.15s.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <20030401165546102-0500 at news.covad.net>, Dan Grassi wrote:
> Not quite, that only provides the path to the initial .py file.
> 
> With a directory structure:
> fa
>     a.py
>     fb
>         b.py
> 
> cd .../fa
> execute a.py
> where a adds fb to sys.path
> a imports b.py
> 
> I need b.py to report it's path which would be .../fa/fb

Assuming the current directory hasn't changed since the file was
loaded, or the file was loaded with an absolute filename:

    def mydir():
        import os.path
        return os.path.abspath(os.path.dirname(__file__))



From aleax at aleax.it  Tue Apr 22 08:51:12 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 12:51:12 GMT
Subject: looping and deleting
References: <3cf72d52.0304210801.2128eca2@posting.google.com> <b815os$618$1@panix1.panix.com> <b8176u$a2e$1@slb9.atl.mindspring.net>
Message-ID: <4Vapa.44460$LB6.1000653@news1.tin.it>

Matt Gerrans wrote:

> "Aahz" wrote:
>> Are you trying to modify the file in-place?  If so, look at the
>> fileinput module.  If not:
> 
> I thought the fileinput module was just for gathering all the files on the
> command line into one virtual stream (that is, reading them all without
> respect to filename, as one long series of lines).

Well, then, it's a good thing to learn that fileinput is MUCH more
generally useful than that, isn't it?-)  Look it up...


> I don't know if this is the best or most Pythonic way, but this is what I
> usually do to modify a text file in place (unless I know it will be a
> really big file):
> 
> lines = file( filename ).readlines()
> f = file( filename, 'w' )
> for line in lines:
>    f.write( modifyTheLine(line) )
> f.close()

Sure, if everything fits comfortably in memory this is an excellent
solution, too!


Alex




From news at exultants.org  Fri Apr  4 05:01:20 2003
From: news at exultants.org (Van Gale)
Date: Fri, 04 Apr 2003 10:01:20 GMT
Subject: Distributed environment
In-Reply-To: <3e8d1edc$0$235@hades.is.co.za>
References: <3e8d1edc$0$235@hades.is.co.za>
Message-ID: <QJcja.793$vo1.55965849@newssvr21.news.prodigy.com>

As Maciej pointed out, YAMI might do what you need.

I also have a feeling that Spread Toolkit (http://www.spread.org) is 
exactly what you need (unless you need load balancing).  Python 
interface is available at http://www.python.org/other/spread/

Van




From abcdebl2nonspammy at verizon.net  Tue Apr 22 00:18:51 2003
From: abcdebl2nonspammy at verizon.net (David Lees)
Date: Tue, 22 Apr 2003 04:18:51 GMT
Subject: regular expressions
In-Reply-To: <TAadnXAyb8p8CzmjXTWcow@speakeasy.net>
References: <b81lq4$17n8$1@galaxy.us.dell.com> <1050961701.8396.0@despina.uk.clara.net> <TAadnXAyb8p8CzmjXTWcow@speakeasy.net>
Message-ID: <Lo3pa.5735$vs2.4989@nwrdny01.gnilink.net>

A.M. Kuchling wrote:
> On Mon, 21 Apr 2003 23:00:21 +0100, 
> 	Garry Knight <garryknight at gmx.net> wrote:
> 
>>The best resource (IMO) is Jeffrey Friedl's "Mastering Regular Expressions",
>>published by O'Reilly, ISBN 1-56592-257-3.
> 
> 
> Note that the second edition of this book seems to have dropped all the
> Python material, judging by the table of contents.  The first edition only
> covered the regex module, not the re module.  It's worth a look if you want
> to learn more about regexes, but it won't be useful as reference material.
> 
> --amk
Not quite all, but pretty skimpy.  There is a section starting on page 
237 titled 'Benchmarking with Python' and the index does have 9 
different page references under 'Python'.

David Lees




From Kyler at news.Lairds.org  Mon Apr  7 12:23:12 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Mon, 07 Apr 2003 16:23:12 GMT
Subject: Zope large PDF question
References: <e09988ac.0304070214.4212e2e9@posting.google.com>
Message-ID: <oqh8m-ldq.ln1@news.lairds.org>

davec999 at yahoo.co.uk (David Currie) writes:

>I am having some problems serving large(ish) PDF files from Zope. 
>Basically, Acrobat Reader seems to hang trying to open a PDF directly
>from a Zope server when used as an IE plugin.  Downloading the file
>with right-click/save-as works fine, opening the file in the plugin
>when it is stored on an Apache server on the same machine and opening
>the file from the same place also works without any problems.  So it
>looks like the Reader plugin does not play well with Zope.  Does
>anyone have any ideas on why this could be happening?

Yes, it's almost certainly due to byterange reading.

>Does Zope have
>problems transferring large files under some circumstances?

Zope has problems dealing with byteranges the way that
Acrobat Reader specifies them for "large" files.  I 
reported this a long time ago, but haven't followed it
lately.  (I just use my own patch to disable it.)  

As I recall, Reader sends a request for pieces of a
large file
	1-500, 1001-1500, 501-1000
and Zope decides to be efficient and join all of the
contiguous regions.  Unfortunately, that defeats the
purpose of requesting the ranges *and* breaks Reader.

--kyler



From roy at panix.com  Mon Apr 28 18:20:41 2003
From: roy at panix.com (Roy Smith)
Date: Mon, 28 Apr 2003 18:20:41 -0400
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com> <zAara.22169$K35.739509@news2.tin.it>
Message-ID: <roy-621090.18204128042003@reader1.panix.com>

Alex Martelli <aleax at aleax.it> wrote:
> No statistically significant / measurable difference.  The tiny difference
> in conciseness is hardly compelling one way or another, either.

I disagree (respectfully).

I'm all about writing code that's easier to read.  I'll admit that the 
+= form is only marginally simplier to read, but marginal is better than 
nothing.

I tend to pick-and-choose from the XP menu, but one of the items I 
really like is "refactor mercilessly".  Why write a variable name twice 
when you can write it once?  Easier to read, and less error prone.  Have 
fun finding the typo:

cpqHeEltTolPowerSupplyRedundant = cpqHeFltTolPowerSupplyRedundant + 1

Yes, that's a real variable name from real production code.  No, we 
didn't have a choice about the name.  It's an object in a MIB and our 
coding standards require we use MIB object names literally, which was 
never a problem until we started working with this MIB :-(.



From roy at panix.com  Wed Apr  2 14:03:42 2003
From: roy at panix.com (Roy Smith)
Date: 2 Apr 2003 14:03:42 -0500
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <435d40e0.0304011457.56e32bd0@posting.google.com> <sKSdnfwAUb0mrxejXTWcog@comcast.com> <435d40e0.0304021052.736293b3@posting.google.com>
Message-ID: <b6fc6e$1fe$1@panix2.panix.com>

>>> What are the reasons for not having multi-line comments?

Multi-line comments are evil.  I grep up on C's /* */ comments, so
this is not a case of "you like what you grew up with".

The first time you spend a half-hour trying to understand a block of
ugly code and finally get to the end and find a "*/", you'll
understand why they're evil.

The first time you accidentally type "* /" instead of "*/" and wonder
why your program still compiles but is suddenly doing strange stuff,
you'll understand why they're evil.

Any decent editor will let you stick "//" or "#" or ";" in front of
every line in a block of code as quickly as you could surround it with
/* */ markers.



From BPettersen at NAREX.com  Tue Apr 22 20:16:56 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Tue, 22 Apr 2003 18:16:56 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE08A@admin56.narex.com>

> From: Andrew Dalke [mailto:adalke at mindspring.com] 
> 
> Bjorn Pettersen:
> > From: Ian Bicking [mailto:ianb at colorstudy.com]
> [... + SQLObject example]
> 
> > for paper in Paper.select(AND(Paper.q.title == 'foo',
> >                               Paper.q.author.startswith('Bob'))):
> 
> > Does anyone know the reason for not allowing an overload of the and
> > operator?
> 
> The 'and' and 'or' operators work on the boolean-ness of the objects,
> and apply short-circuit behaviour.  This is different than 
> other binary
> operators, which don't short-circuit.

Ah! Thanks.

-- bjorn




From max at alcyone.com  Tue Apr 22 23:14:58 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 20:14:58 -0700
Subject: Twisted 1.0.4
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local> <341410000.1051007646@leeloo.intern.geht.de> <20030422121836.GA8567@frobozz.local> <1051031635.18957.24.camel@software1.logiplex.internal> <mailman.1051064256.4034.python-list@python.org>
Message-ID: <3EA60532.6ADC21C@alcyone.com>

Andrew Bennetts wrote:

> Using gcc doesn't require you to link against anything, let alone
> LGPLed
> libraries -- although of course that is the most common situation;
> most
> systems using gcc will also be using glibc.  But the libraries linked
> against
> are a seperate issue to the point I was making about gcc; using gcc
> doesn't
> necessitate using LGPLed libraries, ...

While that's in some sense technically true, any use of gcc to build an
application -- that is, actually create an application rather than
simply compile to object code via the -c option -- _will_ silently link
in those pesky LGPL libraries like libc or libg++.

In special forms of embedded systems where you in effect have your own
operating system, one might use gcc in a way where no such linking takes
place, but for any typical application it will, whether you realize it
or not.  (I'm not sure whether or not you're aware of this, but I
thought I'd point it out anyway.)

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ 'Tis man's to fight, but Heaven's to give success.
\__/ Homer
    Maths reference / http://www.alcyone.com/max/reference/maths/
 A mathematics reference.



From BPettersen at NAREX.com  Sat Apr 12 02:59:23 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sat, 12 Apr 2003 00:59:23 -0600
Subject: win32all odbc/MS SQLServer/bigint
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEAF@admin56.narex.com>

We were using the win32all odbc module, trying to fix some records in
our database. Our account IDs are of SQL type bigint, i.e. __int64 if
you're on MS. After seeing incorrect results (and spending some time
looking at our own code), it turns out that the odbc module silently
corrupts data in paramters passed to cursor.execute(). Has anyone seen
this/got a workaround for it (we're currently cast'ing to varchar on
selects, and doing string substitution elsewhere, which is killing
performance, and we can't get it to work in subselects...)? I looked at
the mxODBC documentation, and it says it's doing a conversion
[automatic] to string for bigint data... (again too slow). I'm pretty
flummoxed, the ODBC C calls for bigint/__int64 are pretty straight
forward(?)

As an added bonus, someone had the pleasure of showing me that a simple
select/fetchall (i.e. all C code for Python), was over three times
faster in C#..., and the win32all odbc module has crashed python
somewhere deep down in sql three times in the last two hours (we're
expecting this app to run for ~4 days...). Looks like it's back to C++
(*sigh*).

-- bjorn




From mgerrans at mindspring.com  Tue Apr  8 12:43:40 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Tue, 8 Apr 2003 09:43:40 -0700
Subject: help modules Python
References: <mailman.1049747592.21028.python-list@python.org>
Message-ID: <b6uu8h$dom$1@slb3.atl.mindspring.net>

Looks like you just need the latest Python (I think object was introduced
with version 2.0, but others on this list would know more precisely -- in
any case, the latest download of Python) and the win32 extensions, so just
go to download link at www.python.org and you'll find these.






From jjl at pobox.com  Mon Apr  7 21:14:44 2003
From: jjl at pobox.com (John J Lee)
Date: Tue, 8 Apr 2003 02:14:44 +0100
Subject: 'Catastrophic failure's trying to use OCX
In-Reply-To: <b6stun$mcf$1@lust.ihug.co.nz>
References: <b6stun$mcf$1@lust.ihug.co.nz>
Message-ID: <Pine.LNX.4.44.0304080212180.1201-100000@alice>

On Tue, 8 Apr 2003, Gary Stephenson wrote:

> I am trying to interface to an .OCX file.  After successfully running makepy
> over it, this is what happens:
[...]
> pywintypes.com_error: (-2147418113, 'Catastrophic failure', None, None)
> >>>
>
> Exactly the same thing occurs if I use late-binding instead (via
> win32.com.client.Dispatch.dynamic('....').
>
> What am I doing wrong?

You're not -- it's a bug in the server.  IIRC, it means the server didn't
catch an exception, or caused an access violation, neither of which a COM
server is supposed to allow to happen.  It's really a COM question you're
asking, not a Python question, though.


John




From mb at muenster.de  Sun Apr 13 15:50:47 2003
From: mb at muenster.de (Martin Bless)
Date: Sun, 13 Apr 2003 19:50:47 GMT
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de>
Message-ID: <3e99b0f8.10658656@news.muenster.de>

Thanks a lot to all who answered. To summarize:

(1) Yes, its all about a cgi environment. Only single mails are being
sent when people submit a form.  Therefore unbearable system load
isn't an issue. Nevertheless I had been using smtplib before and was
quite happy with it. Unfortunately Puretec, a big german provider with
several million accounts changed the rules recently when they moved to
their new and wonderful computing centre. All of a sudden
S=smtplib.SMTP('any.smtp.server') would bring about error 111,
"connection refused". Connecting to 'localhost' is an interesting idea
I hadn't thought of - I'll give it a try. But I'm not very optimistic.

(2) With success I mean: Did sendmail queue my mail without
complaints? From what's being said I conclude that sendmail.close()
provides a yes or no answer and moreover sendmails exit code.
Therefore I'll probably go this direction (untested, modified from
what Carsten Gaebler suggested):

import os, sys

try:
  sendmail = os.popen("/usr/lib/sendmail -t", "w")
  sendmail.write(msg)
  status = sendmail.close()
  done = 1
except:
  done = 0

if not done:
  print "something went wrong:", sys.exc_type, sys.exc_value
elif status:
  exitcode = status >> 8
  signal = status & 0xFF
  print "sendmail exit code %d (killed by signal %d)" % (exitcode,
signal)

Of course I won't have the print statements actually print into the
resulting html page.

(3) Looks like I only have to use popen2.popen3(...) if I want to
retrieve sendmails stdout and stderr streams. Maybe I give it a try
and do some experimentation on what's possible in this specific cgi
environment.

(4) For completeness my original question:

>
>To send mails from Python here's the example my provider provides:
>This http://faq.puretec.de/skripte/python/3.html
>
>The relevant lines are these:
>
>sendmail = os.popen("/usr/lib/sendmail -t", "w")
>sendmail.write(msg)
>sendmail.close()
>
>This works just fine. But is it enough? How do I know the operation
>succeeded?
>
>I've read the docs about os.popen[234] and the module popen2 but I'm
>not sure where to go. And since experimentation isn't easy in this
>case I'd appreciate some hints or a concrete example very much.
>
>Do I have to use popen2.popen3() and its wait() method? I have to
>admit that I don't understand too much of what's being said about this
>in the docs ... :-(
>
>Martin
>

bye again,
Martin




From aleax at aleax.it  Sun Apr 13 15:24:32 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 19:24:32 GMT
Subject: Creating a PyObject* from a C double*
References: <20cb355e.0304122315.c068128@posting.google.com> <YJ9ma.6928$LB6.196819@news1.tin.it> <20cb355e.0304131110.55240129@posting.google.com>
Message-ID: <QPima.9382$LB6.251834@news1.tin.it>

Joze wrote:

> Exactly what i was looking for. Thanks a lot !!
> 
> Is there any way to avoid copying the array element by element and
> simply pass a pointer to it?

If you KNOW the array will never, NEVER be freed (typically a
static array), then you can use PyArray_FromDimsAndData, I think.
Personally, I've never actually used it -- copying stuff (typically
one blast of memmove) seems a tiny overhead, and keeping descriptor and
data closer together may give more performance when you're _using_ the
array, for all I know.

But one thing I HAVE done -- instead of keeping the array elsewhere,
just START with the PyArray_FromDims, and use the (double*)parr->data
as "the array itself" right from the start.


Alex




From aleax at aleax.it  Tue Apr 15 18:00:25 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 00:00:25 +0200
Subject: pread/pwrite (was Re: files: direct access - NOT sequential)
In-Reply-To: <20030415182141.GC30739@unpythonic.net>
References: <mailman.1050398560.8881.python-list@python.org> <200304151925.57736.aleax@aleax.it> <20030415182141.GC30739@unpythonic.net>
Message-ID: <200304160000.25702.aleax@aleax.it>

On Tuesday 15 April 2003 08:21 pm, Jeff Epler wrote:
   ...
> pread() call.  (if you *always* want to append to the file, there's a way
> to do that now with write(), isn't there? opening with os.open(,O_APPEND)

Yep, but if you sometimes want to append and sometimes to write to
a specific offset?

> / file(,"a").  Hm, what does pwrite() do on a file opened with
> O_APPEND?)

No idea (not familiar with pwrite &c).

> I was really thing in "wrap C routines to Python" mode, not "what would
> be a useful way to expose this kind of functionality to Python" mode...

Then Guido will likely not be all that interested, since the redesign I was
mentioning is about getting _away_ from dependencies on the underlying
C libraries.  Maybe there's a space for these platform-dependent syscalls
in posix.* or whatever for specific platforms, I dunno.

> that library in another thread you'd want this.  Blech.  But if you're
> writing a Python program you'd probably never use something as primitive
> as a single file object shared in multiple threads.

Me?  I sure wouldn't, right.  But if others are doing it, perhaps module
posix should expose this, I dunno.


Alex





From jubafre at brturbo.com  Tue Apr  8 13:56:08 2003
From: jubafre at brturbo.com (jubafre at brturbo.com)
Date: Tue, 8 Apr 2003 14:56:08 -0300 (GMT-03:00)
Subject: STRING TO LIST
Message-ID: <316753154.1049824568087.JavaMail.nobody@webmail1>

Hello, 

I need to change a variable from string to list. 
> 
> >>> x=str("[[12,13],[14,15]]") 
> >>> x 
> '[[12,13],[14,15]]' 
> 
> i have x and i want to change to a list, but not in this way: 
> 
> >>> x=list(x) 
> >>> x 
> ['[', '[', '1', '2', ',', '1', '3', ']', ',', '[', '1', '4', ',', '1', '5', ']', ']'] 
> 
> must be a list like this: 
> 
> >>> x=[[12,13],[14,15]] 
> >>> x 
> [[12, 13], [14, 15]] 
> 
> someone know??? 


Juliano Freitas

From nirina at mail.blueline.mg  Mon Apr 14 12:07:31 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Mon, 14 Apr 2003 20:07:31 +0400
Subject: Python 2.2.2 & Tix on win32 - What must I do?
References: <b7666e$oto$1@newstree.wise.edt.ericsson.se> <mailman.1050061697.14672.python-list@python.org> <3e995853$0$24701$edfadb0f@dread14.news.tele.dk>
Message-ID: <012001c302a0$ac39d500$f5a2383e@raseliar>

"Frithiof Andreas Jensen"
> Raseliarison nirinA
>
> > You must have Tix widgets installed.
> > http://tix.sourceforge.net/
> > and maybe also the Tix library
>
> Easier said than done ;-)
>

Yes, but ...

> The Tix home page makes some vague statements on installing Tixapps and
> Idestudio for binary installs yet the home-pages for said applications
refer
> back to tix.sourceforge.net ...
>

... win**** binary installation issues?
How about installing a new and free operating system?
Tix is a part of tcl package with most Linux distribution.

> Perhaps I shall wait until Tix is properly integrated into Python; I do
not
> wish do damage my development environment.
>

Why wait? I'm sure you will run your Linux box before win binary will be
released. I always think, i may be wrong, that installing from source is
better than installing binary package.

Regards,

nirinA






From rmunn at pobox.com  Thu Apr  3 05:16:40 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 03 Apr 2003 10:16:40 GMT
Subject: Python equivalent of _tempnam?
References: <3e8ab45b$1@mail.hmgcc.gov.uk> <b6fj2e$hd3$1@gateway.northgrum.com>
Message-ID: <slrnb8ndge.eaq.rmunn@localhost.localdomain>

Dang Griffith <dmgriffith at tasc.com> wrote:
> How about using os.tempnam(...) ?
> 
>>>> import os
>>>> x = os.tempnam("", "output_")+".txt"
>>>> print x
> C:\TEMP\output_2.txt"
>>>> x = os.tempnam("", "output_")+".txt"
>>>> print x
> C:\TEMP\output_3.txt
> 
> In my few minutes of spearminting, I couldn't get the first parameter
> (the directory) to behave the way I interpret the documentation as
> saying it should.  Could be my interpretation.  :-)  I'm using Win2k,
> Python 2.2.2.
>    --dang

[snip fully-quoted post under response. Please don't top-post.]

Looks like it's putting the "common location for temporary files" in
when the dir parameter is an empty string. My guess is that internally,
it's checking dir for truth; if dir is false, then the temp dir is used.
If you want temporary files in the current directory, try specifying a
dir parameter of "." -- see below for Unix example:

>>> os.tempnam()
__main__:1: RuntimeWarning: tempnam is a potential security risk to your program
'/tmp/fileAjhgek'
>>> os.tempnam("", "foo")
'/tmp/foo6x64Kw'
>>> os.tempnam("", "foo")
'/tmp/fooKusOXN'
>>> os.tempnam(".", "foo")
'./fooxT9AB8'

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From max at alcyone.com  Sun Apr  6 21:22:35 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sun, 06 Apr 2003 18:22:35 -0700
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <3E90D2DB.1A799612@alcyone.com>

Ben Finney wrote:

> In ork.py I want:
> 
>     import eek

If the toplevel package is put in a place where Python can find it, then

	import monty.foo.eek

or

	from monty.foo import eek

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Sit loosely in the saddle of life.
\__/ Robert Louis Stevenson
    Maths reference / http://www.alcyone.com/max/reference/maths/
 A mathematics reference.



From goodger at python.org  Tue Apr 29 00:17:18 2003
From: goodger at python.org (David Goodger)
Date: Tue, 29 Apr 2003 00:17:18 -0400
Subject: POD vs. reST for standalone writing?
In-Reply-To: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
Message-ID: <3EADFCCE.7080504@python.org>

eichin at metacarta.com wrote:
> Now I start seeing mention of reStructuredText (especially in comments
> about pyCon.)  reST files look very "ascii art" on the input side, but
> after a little experiment, that seems to make them *harder* to write,
> if somewhat (and I think only a little) easier to read. 

That's significant: "somewhat (and I think only a little) easier to 
read".  I designed reStructuredText to be a *lot* easier to read. 
Opinions vary of course, I think mostly determined by what you're used 
to.  (Some people think in XML, others dream in TeX.  Not me.)

As for being harder to write, that may be true, especially if the author 
isn't used to the markup yet.  In reStructuredText, readability is 
deemed far more important than writability.  But with increasing 
familiarity comes increasing ease of writing.

> Am I missing something?

Just have an open mind, and enlightenment may come to you.  Where it 
comes *from*, I cannot say. ;-)

> Is there actually a reST clear short hand,
> and a tool that expands it to full-blown ascii-art?

No, although there is some work being done on a reStructuredText Emacs mode.

> Or is at least fuzzy in the counting of dashes and other punctuation
> for marking different levels?

I don't understand the question; please clarify.

-- 
David Goodger    http://starship.python.net/~goodger    Projects:
   * Python Docutils: http://docutils.sourceforge.net/
     (includes reStructuredText: http://docutils.sf.net/rst.html)
   * The Go Tools Project: http://gotools.sourceforge.net/






From karaatanasov at hotmail.com  Sun Apr 20 09:39:57 2003
From: karaatanasov at hotmail.com (Kiril Karaatanasov)
Date: 20 Apr 2003 06:39:57 -0700
Subject: Parent references: hot or not?
References: <mailman.1050762692.5175.python-list@python.org>
Message-ID: <71a1c515.0304200539.3f62b672@posting.google.com>

Hi this looks much as a case where aquisition is the answer to your questions.

See this example:

    import ExtensionClass, Acquisition

    class C(ExtensionClass.Base):
      color='red'

    class A(Acquisition.Implicit):

      def report(self):
        print self.color

    a=A()
    c=C()
    c.a=A()

    c.a.report() # prints 'red'

    d=C()
    d.color='green'
    d.a=a

    d.a.report() # prints 'green'

    a.report() # raises an attribute error

Aquisition is widely used in Zope. Here is alink to the standart guide:

http://debian.acm.ndsu.nodak.edu/doc/python-extclass/Acquisition.html

Hope this helps :)



From aahz at pythoncraft.com  Wed Apr  2 11:39:31 2003
From: aahz at pythoncraft.com (Aahz)
Date: 2 Apr 2003 11:39:31 -0500
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com> <b6esse$pge$1@heckle.cc.uregina.ca> <3E8B0030.E3B0DE00@engcorp.com>
Message-ID: <b6f3o3$t1n$1@panix3.panix.com>

In article <3E8B0030.E3B0DE00 at engcorp.com>,
Peter Hansen  <peter at engcorp.com> wrote:
>
>I believe Google could provide more detail with a search using 
>"stainless roman numerals" or something like that.

You're using Roman numerals for underwear?
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From gry at ll.mit.edu  Tue Apr 29 13:25:19 2003
From: gry at ll.mit.edu (george young)
Date: Tue, 29 Apr 2003 13:25:19 -0400
Subject: how to parse numeric data files
Message-ID: <20030429132519.4e92271e.gry@ll.mit.edu>

[python 2.3a1]
We have several electronic testing machines(of various ages and
manufacturers) that spew out testing data files in various ascii formats.
Currently we have a nasty mess of awk/shell/C/fortran programs that
extract and process some data from these files.  I have a dream of
a suite of simple, clear, maintainable python programs to do these tasks.  

Ideally, for each file format(there may be a dozen or so), a clear,
concise descriptor file (yacc-like language definition? whatever)
would drive the parsing process for reading that kind of data file,
providing a nice clear API to the data.  

Is there a PyYacc or some such parser module?  Or should I use
the regular-expression module or some such lower level facility
to hard code each format?  

The trick is I hope to come up with something that our hardware
engineers can understand and maintain easily without studying
things like BNF, LALR etc. (they won't).

I hope then to use Numeric and ScientificPython to do the crunching.

I know I'm trying to do something complex very simply, but
hey! this is python right?

[Below is a sample of one of the worst formats, shortened from a 40MB file!]
-- George


PH_lot_id:			YES_2_1.25V_1.0V
Device_id:			4Mb_dc1
Operator_id:			EEA
PFD:			YES_2_1.25V_1.0V 4Mb_dc1 4M_digit_capture1 VDD=1.25V VLEAKLOW=0V V_TCSETP=1.0V
Date:			01/30/03
Test_type:			Functional_Test
Prober_id:			EG.3001
Probe_Card:			ye_sram_4Mb_1
Test_description:			4Mb SRAM1

Binlist: wafer_log
Bin: 	Opn	reprobe	 00	FAIL	Open Pins Test
Bin: 	DC	no_reprobe 	01	PASS	DC
Bin: 	db	no_reprobe	-01	FAIL	

Data_format: X,Y  SW_Bin  HW_Bin  P_F  Date  Time  Temp  FF/TF/TT  Bin_desc
Begin_data:
                         
3,2 	DC 01 	PASS 	20030411 133414 -9999.0 3/3/38 DC
  Data_format_test: suite  function  result [ res[0] unit[0] res[1] unit[1] res[2] unit[2] res[3] unit[3] ]
  Begin_data_tf:
  speccalues  spec_values PASS
     Test_number: 0
    Begin_user_data:
LEVEL SPEC VARIABLES
Eqn = 1 Sps = 1 Description = ye_sram_64Kb_vdd_1500mV_w_0mV
VDD =  1.250 V WAFER =  0.000 V V_TCSEPT =  1.000 V VLEAKLOW =  0.000 V 
Eqn = 1 Sps = 2 Description = ye_sram_64Kb_vdd_100mV_w_0mV
VDD =  0.100 V WAFER =  0.000 V V_TCSEPT =  0.000 V VLEAKLOW =  0.000 V 
TIMING SPEC VARIABLES
Eqn = 1 Sps = 1 Description = ye_sram_4Mb
t_setup =   333.3 ns t_pw_enable =   333.3 ns t_wait =   333.3 ns Period =  1000.0 ns

    End_user_data
  Connectivity_T1  leakage PASS   -10.000000 	uA     -8.999900 	uA      0.000000 	uA      0.000000 	uA 
     Test_number: 1
     Begin_limit_data:
SPRM 101,1,102001,102001
SQSL "ye_sram_4Mb"
leakage;addr_in00,addr_in01,addr_in02,addr_in03,data_out0,data_out1,data_out2,data_out3,data_out4,data_out5,data_out6,data_out7,enable,set0bar,set1bar,writebar; V;    -1.0;     -1.0;uA;   -10.0;   -10.0;uA;   -15.0;   -15.0;uA;    -1.0;     -1.0;ms;     200;     200;3;Opens ($P);10; V;    -1.0;    -1.0;;;;0;
     End_limit_data
    Data_format_pins: pin_name  PFNU0 [ value0 unit0 ] PFNU1 [ value1 unit1 ]
    Begin_data_pins:
    addr_in00 P   -10.000000 	uA  N     0.000000 	uA 
    addr_in01 P   -10.000000 	uA  N     0.000000 	uA 
    addr_in02 P   -10.000000 	uA  N     0.000000 	uA 
    addr_in03 P   -10.000000 	uA  N     0.000000 	uA 
    data_out0 P   -10.000000 	uA  N     0.000000 	uA 
    data_out1 P   -10.000000 	uA  N     0.000000 	uA 
    data_out2 P   -10.000000 	uA  N     0.000000 	uA 
    data_out3 P   -10.000000 	uA  N     0.000000 	uA 
    data_out4 P   -10.000000 	uA  N     0.000000 	uA 
    data_out5 P   -10.000000 	uA  N     0.000000 	uA 
    data_out6 P   -10.000000 	uA  N     0.000000 	uA 
    data_out7 P   -10.000000 	uA  N     0.000000 	uA 
    enable P   -10.000000 	uA  N     0.000000 	uA 
    set0bar P   -10.000000 	uA  N     0.000000 	uA 
    set1bar P   -10.000000 	uA  N     0.000000 	uA 
    writebar P   -10.000000 	uA  N     0.000000 	uA 
    End_data_pins
  Idd_at_100mV_T2  standby_current FAIL     0.004769 	 A      0.004769 	 A      0.000000 	 A      0.000000 	 A 
     Test_number: 94
     Begin_limit_data:
SPRM 101,1,102001,102001
SQSL "ye_sram_4Mb"
standby_current;vdd;uA;     0.01;uA;    1.00;Idd_at_100mV ($P);0;
     End_limit_data
    Data_format_pins: pin_name  PFNU0 [ value0 unit0 ] PFNU1 [ value1 unit1 ]
    Begin_data_pins:
    vdd F     0.004769 	 A  N     0.000000 	-
    End_data_pins
  Idd_at_Plus_1250mV_T3  standby_current PASS     0.071565 	 A      0.071565 	 A      0.000000 	 A      0.000000 	 A 
     Test_number: 100
     Begin_limit_data:
SPRM 101,1,104001,104001
SQSL "ye_sram_4Mb"
standby_current;wafer;uA;     -1.0;uA;       0;Wafer_current_at_2000mV ($P);0;
     End_limit_data
    Data_format_pins: pin_name  PFNU0 [ value0 unit0 ] PFNU1 [ value1 unit1 ]
    Begin_data_pins:
    wafer F     0.000002 	 A  N     0.000000 	-
    End_data_pins
  M02831EO  M02831EO PASS
     Test_number: 102
    Begin_user_data:
bit31_col_00-031_01 254 258 254 258 254 258 254 258 254
bit30_col_31-063_01 258 254 258 254 258 254 258 254 258

digit_cap_M0_b28-b31_even
ffeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
...
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeee00eeeeeeeeeeeeeeeeee
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee

bit31_col_00-031_10 256 256 256 256 256 256 256 256 256
bit30_col_31-063_10 256 256 256 256 256 256 256 256 256

M02831 NFE = 32768 NFO = 32768 NF = 65408

 digit_cap_M0_b28-b31_odd
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
...



From dbrown2 at yahoo.com  Mon Apr  7 21:19:08 2003
From: dbrown2 at yahoo.com (dbrown2 at yahoo.com)
Date: 7 Apr 2003 18:19:08 -0700
Subject: Beginner: How to copy a string?
References: <mailman.1048813246.30445.python-list@python.org> <d524bdb3.0304031822.2ed781be@posting.google.com> <3E8CF439.D4D48B8A@engcorp.com>
Message-ID: <d524bdb3.0304071719.3154645c@posting.google.com>

Thanks for the clear and concise answer.  It would likely take me a
long time if ever to understand this point on my own looking through
written docs.  Gotta love comp.lang.python and it's inhabitants!

It looks like I'm done copying strings (or at least trying).  I'm a
convert.  In fact, just wait 'til someone asks this question again.  I
hope I can be the first in line to show utter shock at the silly
notion. ;-)

Regards,

David


Peter Hansen <peter at engcorp.com> wrote in message news:<3E8CF439.D4D48B8A at engcorp.com>...

>clip...
> Excellent try, but I'm afraid it wouldn't work that way (unless someone
> really devious deliberately tried to break how things should work).
> The data returned by the external library, in order to be interfaced to 
> Python, would have to create a *Python* string using calls into the 
> Python interpreter (see http://www.python.org/doc/current/ext/ext.html
> for more on this sort of thing).  Python would already have copied the
> external data into a new area, which would then be preserved forever
> in the face of everyone's futile :-) attempts to copy it.
> 
> So long as Python strings are not exposed via a pointer to some 
> malicious external routine, I doubt you'll ever need to worry about
> the copying thing, but I salute your attempt!
> 
> -Peter



From max at alcyone.com  Sun Apr 27 03:07:29 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sun, 27 Apr 2003 00:07:29 -0700
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com>
Message-ID: <3EAB81B1.D706C567@alcyone.com>

Terry Reedy wrote:

> "Erik Max Francis" <max at alcyone.com> wrote in message
> news:3EAB6A14.707AD826 at alcyone.com...
>
> > # f holds a callable object or None
> > if f:
> >     f(...)
> 
> or
> print f and f() or 'cannot do that' # assumes boo(f()) always True

Shorter is not always better.  This being an awfully good example.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Death to all fanatics!
\__/ Malaclypse the Younger
    Kepler's laws / http://www.alcyone.com/max/physics/kepler/
 A proof of Kepler's laws.



From juneaftn at REMOVETHIShanmail.net  Wed Apr 16 13:30:15 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Thu, 17 Apr 2003 02:30:15 +0900
Subject: How many of you are Extreme Programmers?
References: <1050508138.3e9d7b6a1d520@mcherm.com> <mailman.1050509219.31421.python-list@python.org>
Message-ID: <b7k3u9$4k5$1@news.hananet.net>

"Christopher Blunck" <blunck at gst.com> wrote in message
news:mailman.1050509219.31421.python-list at python.org...
> On Wed, Apr 16, 2003 at 08:48:58AM -0700, Michael Chermside wrote:
> > Christopher writes:
> > > So how many of you guys use XP processes?
> > Well, I'm not sure if I qualify or not. I really LIKE most of the
points
> > that XP makes, but I find many of its proponants overly exuberant. I
> > think exhaustive, automated tests are a wonderful idea. I believe in
> > refactoring. I think pair programming works very well most of the time.
> > I think super-short release cycles (with code that grows feature-by-
> > feature) is a great idea. I think having a customer on-site is
extremely
> > helpful.
>
> It's quite funny, but if you sit an XP and RUP guy down at the same table
> you'll find that they agree with most principals.  They'll disagree on
some
> things (like design), but I've never seen them disagree on iterative
> development, frequent testing, frequent integration, etc.  They may
disagree
> on the frequency of "frequent testing" (XP favoring daily testing), or on
> the implementation details of iterative development.  But for the most
part,
> I've found would-be opposites to share very similar views.
>
>
[snip]

Grady Booch, for example, is a board member of the Agile Alliance, and he
said, in my and other interviews, that he believes in and likes it. You
might know that Rational Rose has XP plug-in.

See also Mark Paulk's write-up Extreme Programming from a CMM Perspective
at http://www.sei.cmu.edu/cmm/papers/xp-cmm-paper.pdf




From rodrigob at elo.utfsm.cl  Mon Apr 28 09:30:46 2003
From: rodrigob at elo.utfsm.cl (Rodrigo Benenson)
Date: Mon, 28 Apr 2003 13:30:46 +0000
Subject: Pyrex Hard Case
Message-ID: <3EAD2D06.7010505@elo.utfsm.cl>

I had allready done that. This command create an

extern MACRO_NAME(EXPECTED_ARGUMENTS)

Then macro name is remplaced by something like

extern double * C2F(stack).STK + x - 1 (EXPECTED_ARGUMENTS);

and the C compile is missed, this is the line I have comment via hand edit.

rodrigob.

 >
 > De: Francois Pinard <pinard at iro.umontreal.ca>
 > Fecha: 2003/04/28 Mon AM 08:55:10 GMT-04:00
 > Para: rodrigob at elo.utfsm.cl (Rodrigo Benenson)
 > CC: python-list at python.org
 > Asunto: Re: Pyrex Hard Case
 >
 > [Rodrigo Benenson]
 >
 > > Basically there is a Macro function that I can't map to Pyrex, so I
 > > create an equivalent dummy function and hand edit the .c file erasing
 > > the definition.
 >
 > I'm not fully sure I'm replying to your question, but this might work:
 >
 > ---------------------------------------------------------------------->
 > cdef extern from *:
 >     MACRO_NAME(EXPECTED_ARGUMENTS)
 > ----------------------------------------------------------------------<
 >
 > If you do that, every time you use MACRO_NAME as a Python function, the
 > net effect will be to call the macro, and Pyrex will make no attempt to
 > declare it in the generated C code.  Of course, you may replace `*' above
 > by the header file defining the macro, if it needs an explicit #include.
 >
 > --
 > Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard
 >





From writeson at earthlink.net  Thu Apr 24 08:07:09 2003
From: writeson at earthlink.net (Doug Farrell)
Date: 24 Apr 2003 05:07:09 -0700
Subject: ftp problem on solaris machines with python
Message-ID: <88bc63c6.0304240407.674220cb@posting.google.com>

Hi all,

We've written an installation tool in Python2.2 that tars up
directories and ftp's the files over to the destination machines. This
works great from linux machines to linux machines. However, going from
solaris to solaris machines the process croaks if the tar files being
transferred are to big. Has anyone seen problems like this using
Python on solaris? And if so, any suggestions?

Thanks in advance,
Doug Farrell



From spam04 at kuwest.de  Wed Apr  9 17:37:17 2003
From: spam04 at kuwest.de (Jens Kubieziel)
Date: Wed, 9 Apr 2003 21:37:17 +0000 (UTC)
Subject: Faqwiz - strange error
Message-ID: <b723qd$gj9$2@QBI050102.kubieziel.de>

Hi,

I have a strange error within the faqwizard-script. I changed following
string:

 ,-----[ entry in Faqwizard ]
| PROLOGUE = '''
| <HTML>
| <HEAD>
| <TITLE>%(title)s</TITLE>
| </HEAD>
| 
| <BODY BACKGROUND="http://www.python.org/pics/RedShort.gif"
|       BGCOLOR="#FFFFFF"
|       TEXT="#000000"
|       LINK="#AA0000"
|       VLINK="#906A6A">
| <H1>%(Title)s</H1>
| <p>added some text</p>
| '''
 `--------------

Below the <H1>-line I added some text enclosed in <p>-tags. When I called
the "new" script (and even if I remove those tags) I get following error:

Traceback (most recent call last):

  File "faqw.py", line 26, in ?
   import faqwiz
  File "/www/users/web70/html/cgi-bin/faqwiz.py", line 841, in ?
   wiz.go()
  File "/www/users/web70/html/cgi-bin/faqwiz.py", line 389, in go
   meth()
  File "/www/users/web70/html/cgi-bin/faqwiz.py", line 409, in do_home
   self.prologue(T_HOME)
  File "/www/users/web70/html/cgi-bin/faqwiz.py", line 403, in prologue
   emit(PROLOGUE, entry, kwdict=kw, title=escape(title))
  File "/www/users/web70/html/cgi-bin/faqwiz.py", line 63, in emit
   f.write(_interpolate(format, args, kw))
  File "/www/users/web70/html/cgi-bin/faqwiz.py", line 53, in _interpolate
   return format % m
  File "/www/users/web70/html/cgi-bin/faqwiz.py", line 173, in __getitem__
   value = d[key]
  TypeError: unsubscriptable object

Original at http://www.kubieziel.de/cgi-bin/faqw.py

Where is the problem and how can I solve it?

Thanks
-- 
Jens Kubieziel                                   http://www.kubieziel.de
Better the prince of some inferior court,
Than second, or less, in beatific light.
		-- Lucifer, Joost van den Vondel's "Lucifer"



From billw at witherspoon-design.com  Tue Apr 29 23:16:23 2003
From: billw at witherspoon-design.com (Bill Witherspoon)
Date: Tue, 29 Apr 2003 23:16:23 -0400
Subject: mxODBC puzzle
In-Reply-To: <3EAEE43B.1040307@lemburg.com>
References: <3EAED98B.2090009@emseal.com> <3EAEE43B.1040307@lemburg.com>
Message-ID: <20030430031623.GA3438@localhost.localhost>

Thanks guys, I read that in the docs but had assumed that it didn't
apply to my situation, because I (foolishly) thought Access doesn't
support transactions. I thought the instruction was for more 'full
featured' databases!

(I may have had the blinders on - the project I'm working on is a
prototype to prove we can transition away from MS exclusive technologies
like Access. If I'm successful we may migrate from Access Db/Apps to
Python/WxPython with mxODBC (religion neutral) to some DB that
scales a bit better)

Thanks again,
Bill.


* M.-A. Lemburg wrote:
> Bill Witherspoon wrote:
> >I've been fooling around with mxODBC/mxDateTime for a couple of days 
> >now, trying to effect an Update query. Can anyone spot the error I'm 
> >making with the code below:
> >
> >
> >sql = "Update tblOEOrders set tblOEOrders.ScheduleDate = #" + 
> >newdate.strftime("%m/%d/%Y") + "# where tblOEOrders.OrderID = " + 
> >str(orderid) + ";"
> >
> >myconn = mx.ODBC.Windows.Connect('production')
> >mycursor = myconn.cursor()
> >numlines_affected = mycursor.execute(qrystring)
> 
> This should read:
> 
> mycursor.execute(...)
> numlines = mycursor.rowcount
> 
> >The DB is Access97 via DSN (which is *not* read only), and query seems 
> >to return OK (1 record affected), but the record never gets updated. 
> >Using the same code I can do Selects without difficulty.
> >Do I have to do a fetchall() equivalent to make the update query run, or 
> >is something else wrong?
> 
> mxODBC defaults to transactional operation. To see the updates
> in the database you have to commit your work, i.e. call
> 
> myconn.commit()
> 
> HTH,
> -- 
> Marc-Andre Lemburg
> eGenix.com
* M.-A. Lemburg wrote:
> Bill Witherspoon wrote:
> >I've been fooling around with mxODBC/mxDateTime for a couple of days 
> >now, trying to effect an Update query. Can anyone spot the error I'm 
> >making with the code below:
> >
> >
> >sql = "Update tblOEOrders set tblOEOrders.ScheduleDate = #" + 
> >newdate.strftime("%m/%d/%Y") + "# where tblOEOrders.OrderID = " + 
> >str(orderid) + ";"
> >
> >myconn = mx.ODBC.Windows.Connect('production')
> >mycursor = myconn.cursor()
> >numlines_affected = mycursor.execute(qrystring)
> 
> This should read:
> 
> mycursor.execute(...)
> numlines = mycursor.rowcount
> 
> >The DB is Access97 via DSN (which is *not* read only), and query seems 
> >to return OK (1 record affected), but the record never gets updated. 
> >Using the same code I can do Selects without difficulty.
> >Do I have to do a fetchall() equivalent to make the update query run, or 
> >is something else wrong?
> 
> mxODBC defaults to transactional operation. To see the updates
> in the database you have to commit your work, i.e. call
> 
> myconn.commit()
> 
> HTH,
> -- 
> Marc-Andre Lemburg
> eGenix.com
> 
> Professional Python Software directly from the Source  (#1, Apr 29 2003)
> >>> Python/Zope Products & Consulting ...         http://www.egenix.com/
> >>> mxODBC, mxDateTime, mxTextTools ...        http://python.egenix.com/
> ________________________________________________________________________
> EuroPython 2003, Charleroi, Belgium:                        56 days left
> 
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
Listen, you big, stupid space-creature.  Nobody, but nobody, eats the
Simpsons!

       		-- Homer Simpson
		   Treehouse of Horror




From durdn at yahoo.it.oops!.invalid  Sun Apr 27 19:57:44 2003
From: durdn at yahoo.it.oops!.invalid (Nicola Paolucci)
Date: Sun, 27 Apr 2003 23:57:44 GMT
Subject: Sorting files
In-Reply-To: <ATZqa.667651$S_4.718911@rwcrnsc53>
References: <ATZqa.667651$S_4.718911@rwcrnsc53>
Message-ID: <Y7_qa.934482$sj7.37993441@Flipper>

Hi Zipper,

Zipper114 wrote:
> I'm a bit of a noob when it comes to Python, so this might seem a bit
> trivial, but.....
> 
> How can I get the listing of files of a particular type in a directory,
> sorted by date? For example, I using the following call to get all the
> JPEG's in a dir:
> 
> Images = glob.glob1(MyPath+"\\images", "*.jpg")
> 
> 
> I want to do just this, but have them sorted by date. Any ideas?

You could use the os.stat function and the stat module like:

import os
import pprint
import stat

p = 'f:/mp3/italiane/Franco Battiato'
tosort = []

for f in os.listdir(p):
     tosort.append((os.stat(os.path.join(p,f))[stat.ST_CTIME],f))

tosort.sort()
pprint.pprint([r[1] for r in tosort])


I hope this helps,
ciao,
	Nick

-- 
#Remove .oops!.invalid to email or feed to Python:
'Tmljb2xhIFBhb2x1Y2NpIDxuaWNrQG5vdGp1c3RjYy5jb20+'.decode('base64')




From sismex01 at hebmex.com  Thu Apr 10 12:26:40 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Thu, 10 Apr 2003 11:26:40 -0500
Subject: Python compared toJava
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F0623304501F2B@mail.hebmex.com>

> From: blaktyger at hotmail.com [mailto:blaktyger at hotmail.com]
> 
> Hi!
> 
> I recently read that the Python interpreter is like the Java 
> Virtual Machine.
> Is that true?
>

Almost but not quite totally unlike.

;-)

-gus




From jjl at pobox.com  Tue Apr 22 14:42:50 2003
From: jjl at pobox.com (John J. Lee)
Date: 22 Apr 2003 19:42:50 +0100
Subject: debugging fileinput
References: <pan.2003.04.20.06.14.16.600197@privacy.net>
Message-ID: <874r4q5q6t.fsf@pobox.com>

"Kyoskyu" <me at privacy.net> writes:

[...]
> import fileinput, string
> for line in fileinput.input():
> 	process(strip(line))
> 
> However, for some reason, this snippet of code gets to the end of the file
> and starts again at the beginning.  I'm at a loss as to how to debug this
> problem.  If someone has any ideas or knows the reason, I'd appreciate the
> help.  Thanks.

Don't know the reason offhand, but here are some debugging ideas:

1. State the file to iterate over explicitly (as an argument to
   fileinput.input, IIRC: for line in fileinput.input(myfilename))

2. Does process() do any sys.stdout.seek()s?  Or call functions that
   might do that?


John



From warlock at eskimo.com  Mon Apr 28 00:39:43 2003
From: warlock at eskimo.com (Jim Richardson)
Date: Sun, 27 Apr 2003 21:39:43 -0700
Subject: Dict handling of floats
References: <b94nn-i04.ln1@grendel.myth> <mailman.1051502306.13602.python-list@python.org>
Message-ID: <ftnun-jtc.ln1@grendel.myth>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sun, 27 Apr 2003 23:55:05 -0400,
 Tim Peters <tim.one at comcast.net> wrote:
> [Jim Richardson]
>>>> Linux, because eventually, you grow up enough to be trusted
>>>> with a fork()
> 
> [Tim]
>>> OTOH, I'm not sure anyone grows up enough to be trusted with a
>>> float <wink>:
>>>
>>>     http://www.python.org/doc/current/tut/node14.html
> 
> [Jim Richardson]
>> I figured it was something like that, It doesn't make any difference in
>> my application, I just found it kind of odd.
>>
>> I have no real schooling in programming, all I learn is as I go,
>> sometimes, stuff like this, trips me up.
> 
> Oh sure.  I hope you didn't read that as a dig!  I can testify that experts
> screw up in the face of floating-point, and testify with confidence since
> I'm one of the experts who so screws up <wink>.
> 

Nah, I take correction pretty well, I get it a lot :)

>> <Jim, who *still* doesn't grokk self.>
> 
> You will!  Floating point is doomed to remain esoteric to those who don't
> make a concerted study of the topic, but self has a naturalness that will
> eventually seem obvious.
> 
> self-is-what's-left-when-you're-not-looking-at-others-ly y'rs  - tim
> 
> 

I'll look forward to my future enlightenment re: self    :)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+rLCNd90bcYOAWPYRAuCwAJ463K2Wxv9Y6OmW3yNNND8vw6u+jQCfYYw3
nkMidtn0y4Xa2bVjW668zTo=
=FRVL
-----END PGP SIGNATURE-----

-- 
Jim Richardson         http://www.eskimo.com/~warlock

Linux, because eventually, you grow up enough to be trusted with a fork()



From mads at troest.NEVERMORE.dk  Sat Apr 12 13:19:29 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Sat, 12 Apr 2003 19:19:29 +0200
Subject: Speed of string += string
Message-ID: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>

Hi;

Given that strings are immutable, is it exceedingly slow (and memory 
spiking) to do a lot of "string += string" operations in one's code? IOW, 
does each and every += evaluation lead to a new concatenated copy of the 
previous string (now freed for garbage collection) and the new string, so 
that the longer the string to which stuff is appended is, the longer times 
each += operation takes?

TIA,
   /\/\\ads Orbesen Troest



From max at cNOvSisiPonAtecMh.com  Wed Apr  2 13:25:26 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Wed, 02 Apr 2003 18:25:26 GMT
Subject: PIL altrenatives?
Message-ID: <qWFia.668$n95.847183@twister.nyc.rr.com>

I am looking for a python func to do edge detection. A good function would
be a version of im.point(func), where func would be passed the pixel's
neighbourhood, so that different edges can be produced. As far as I can tell
PIL does not support this directly. Of course, this could probably be done
with PIL the long way, by getting individual pixels. But are there
alternatives? Also, is there a good source for PIL code samples?

thanks,

--
========================================
Max Khesin, software developer -
max at cNvOiSsPiAoMntech.com
[check out our image compression software at www.cvisiontech.com, JBIG2-PDF
compression @
www.cvisiontech.com/cvistapdf.html]






From aleax at aleax.it  Sat Apr 26 06:23:54 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 10:23:54 GMT
Subject: Forgetting "()" when calling methods
References: <mailman.1051324540.14913.python-list@python.org>
Message-ID: <_6tqa.9432$K35.308270@news2.tin.it>

Brian Quinlan wrote:

> Erik Max Francis wrote:
>> One thing that PyChecker, at least, is more likely to catch is "code
> has
>> no effect" statements like
>> 
>> sys.exit
> 
> How do you know that the code has no effect?

You cannot KNOW, but, this is tricky, unreadable code -- I'd be quite
happy to have PyChecker optionally warn me against it, should I ever
happen to code this way.


Alex




From donnal at donnal.net  Mon Apr 21 10:36:13 2003
From: donnal at donnal.net (Donnal Walter)
Date: Mon, 21 Apr 2003 09:36:13 -0500
Subject: wxBoxSizers in window and two embedded panels.
In-Reply-To: <4933657516.20030421190214@tyumbit.ru>
References: <4933657516.20030421190214@tyumbit.ru>
Message-ID: <b80vhk$1am$1@main.gmane.org>

Igor Prischepoff wrote:
> Hi!
>  Sorry for my english, i'm not native english speaking man.
>  Recently switched to python and it's wonderful wxPython toolkit.
>  Having trouble with wxBoxSizer.
>  I want this heirarchy
> 
>       wxWindow (a must! i've don't need to attach panels to wxFrame directly.)
>        /    \
>     Panel2  Panel1
>     
> Panel1 and panel2 must occupy all available screen space on window1.
> Trouble is: they don't.
>     
> 
>  ------------------------------  -----
>  !  panel2                     !      !
>  !-----------------------------!      !
>  !  panel1                     !      !
>  !                             !      !
>  ------------------------------!      !
>  !                 window1            !
>  !------------------------------------

See if this code does what you want. Remember that wxPanel is derived 
from wxWindow.

#!/usr/bin/env python
from wxPython.wx import *

class MyPanel(wxPanel):
     def __init__(self, parent):
         # create panel (the outer window)
         wxPanel.__init__(self, parent, id=-1)
         # create the inner panels
         panel1 = wxPanel(self, -1)
         panel2 = wxPanel(self, -1)
         panel2.SetBackgroundColour("BLUE")
         panel1.SetBackgroundColour("RED")

         # create the sizer
         sizer = wxBoxSizer(wxVERTICAL)

         # add panels to sizer
         sizer.Add(panel1, 1, wxEXPAND)
         sizer.Add(panel2, 1, wxEXPAND)

         # enable sizers
         self.SetSizer(sizer)
         self.SetAutoLayout(1)
         sizer.Fit(self)
         sizer.SetSizeHints(self)

class MyFrame(wxFrame):
     def __init__(self):
         wxFrame.__init__(self, None, -1, 'Sizer Demo',
                          size=(400,200))
         outerWin = MyPanel(self)
         outerWin.SetBackgroundColour("BLACK")

class MyApp(wxApp):
     def OnInit(self):
         frame = MyFrame()
         frame.Show(1)
         return 1

if __name__ == '__main__':

     app = MyApp(0)
     app.MainLoop()






From ianb at colorstudy.com  Sun Apr  6 14:39:53 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 06 Apr 2003 13:39:53 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <a3b5b608.0304060321.6f17ab@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
	 <mailman.1049571010.25625.python-list@python.org>
	 <a3b5b608.0304060321.6f17ab@posting.google.com>
Message-ID: <1049654393.658.57.camel@lothlorien>

On Sun, 2003-04-06 at 06:21, Giorgi wrote:
> > But I'm not saying it wouldn't be nice to have more cohesion... that's
> > just not the way it is now.  It's harder to make it happen, because
> > unlike Java or .NET, the Python community is not a top-down community. 
> > Guido has some authority, but I think he's almost as much a decision
> > process as a decider -- he decides on things where there's some level of
> > consensus in the community, and doesn't make dictates otherwise.  The
> > standard library has some authority, but its fairly neutral on these
> > things.  Besides that there's no authority at the moment.
> > 
> Well, maybe we should establish an advisory board? :)

I just recently proposed a web-sig, which is just a mailing list, but
from there I'm hoping we can get developers from the various frameworks
talking together and maybe make some progress.  I'm just waiting on the
mailing list to be created.  Here's my proposal:
  http://mail.python.org/pipermail/meta-sig/2003-March/001246.html

You can email meta-sig at python.org if you think it's a good idea.

  Ian






From aleax at aleax.it  Wed Apr 30 04:13:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 08:13:46 GMT
Subject: plugin system
References: <WHIra.27759$su3.3307778@twister.nyc.rr.com>
Message-ID: <_ALra.23341$3M4.663663@news1.tin.it>

<posted & mailed>

xam wrote:

> I am interested in writing a plugin system where the main app is python
> but it is capable of accepting helper "plugins" which can be either python
> or C++. I guess I want some sort of an interface spec for the plugins that
> can be implemented in either language. Which technologies should I
> research? Are there other projects that I can refer to as a model -

Boost Python (www.boost.org) lets you code Python-loadable modules in C++;
they will be basically undistinguishable from python-coded modules from
the point of view of the "main app in python" that imports them.

There are many other ways to achieve the same end-result (the "Python
C API", CXX, SWIG, SIP, ...) but if you're committed to C++ anyway for
the other language (i.e. don't care about being able to implement such
modules in C, Fortran, and so on), then I think Boost is the best.  And
anyway, when your specs are that the plugins must be "Python-extension
modules", then somebody else can later choose to implement plugins in
other ways, and "from the outside" (from your main Python app) they'll
work just fine.

So, I recommend you go with this choice, after exploring Boost Python
to convince yourself that it does make things as easy as pie for the
C++ coder.  Boost is a large library (mostly template-based) that also
offers a LOT of pluses to the C++ coder, by the way, in addition to
being able to interface with Python so easily.  Much of Boost may see
its way into the next release of the C++ standard, particularly given
that the population of Boost developers and the membership of the C++
standards committee have substantial overlap (the Boost Python part in
particular is of course unlikely to, given there is no ISO standard
for Python itself nor is any such thing likely to emerge!-).



Alex




From peter at engcorp.com  Mon Apr 14 12:16:18 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 14 Apr 2003 12:16:18 -0400
Subject: A python server
References: <15f445fb.0304140751.170d410e@posting.google.com>
Message-ID: <3E9ADED2.280A671F@engcorp.com>

dansan wrote:
> 
> I'm trying to write a little server that will allow connections via
[snip]
> I guess I'm having dificult with the higher-level desing of such a
> system.  Right now my desing is very primitive, and has a LOT of
> issues:
> 
>  # start
>  # loop-waiting-for-connections (can't break out of this from cli yet)
>  # make tournament schedule
>  # run-games
> 
> I have program in other programming languages before, and lately I've
> been doing some python.  So far I have enjoyed it a lot, so I wanted
> to give this project a try in py.  Any help would be greatly
> appreciated, in higher level desing as well as in lower-level code
> snippets.  URL to documentation are also invaluable :)

I believe the best advice for you at this stage might be the
now-ubiquitous "Twisted does that".  Although at your level 
figuring out how to work with Twisted (http://www.twistedmatrix.com)
might be non-trivial, it's a certainty that rolling it yourself at
a lower level (i.e. sockets) is going to be much harder.

Five or ten lines of code on top of Twisted (and there are lots
of good examples and How-Tos at that web site) is enough to get
you past the socket issues, at which point the job is entirely in
your hands to program the tournament scheduling parts, the
"run games" part, etc.... in other words, it will let you focus
on the application logic itself, and not on the infrastructure.

-Peter



From tebeka at cs.bgu.ac.il  Wed Apr 30 08:26:34 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 30 Apr 2003 05:26:34 -0700
Subject: Self modifying systems question
References: <mailman.1051685036.19081.python-list@python.org>
Message-ID: <33803989.0304300426.402cd1ef@posting.google.com>

Hello Sten,

> Self modifying code is a well known technology. Has anyone done this in
> Python? What I'm seeking for is examples of software that adds to
> itself (methods and variables) in runtime. I'd like an instantiated 
> object to
> be able to add a method to itself.
>>> class A:
	pass

>>> a = A()
>>> a.f = lambda x: x + 1
>>> a.f(100)
101

HTH.
Miki



From fredrik at pythonware.com  Sun Apr 27 06:42:05 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sun, 27 Apr 2003 12:42:05 +0200
Subject: newbie question about forms in tkinter
References: <3eaba626@clear.net.nz>
Message-ID: <b8gc42$ajd$1@main.gmane.org>

Eamon Reyn wrote:

> I am creating a little app which needs to have multiple forms pop up but
> whenever I try to create a second form in my application the content I want
> to appear on the new form just gets placed into my existing one. Here is som
> sample code which looks vaguely sane to me but can someone please tell me
> what is wrong with it.

> class Interface2(Frame):
>     def __init__(self, master = None):
>         Frame.__init__(self, master)

here, master=None means that you're adding stuff to the root
window.

>     def newForm(self):
>         myInter2 = Interface2()
>         myInter2.mainloop();

to open the interface in a separate window, create a Toplevel
window, and pass it as the master:

    top = Toplevel()
    top.title("my other window")
    myInter2 = Interface2()

see chapters 8 and 10 of my "introduction to tkinter" book for
more information on toplevels:

    http://www.pythonware.com/library/tkinter/introduction/application-windows.htm
    http://www.pythonware.com/library/tkinter/introduction/dialog-windows.htm

</F>







From peter at engcorp.com  Thu Apr  3 13:52:23 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 03 Apr 2003 13:52:23 -0500
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org>
	        <b6fj57$626$1@news-reader11.wanadoo.fr>
	        <3E8B53BA.DBB0C84C@engcorp.com>
	        <mailman.1049319846.7316.python-list@python.org>
	        <3E8B67AD.7AD74C0C@engcorp.com> <mailman.1049392327.15686.python-list@python.org>
Message-ID: <3E8C82E7.CD249521@engcorp.com>

danny at iamlearning.com wrote:
> 
> > danny at iamlearning.com wrote:
> >>
> >> I want my users run *.pyo or *.pyc files instead of *.py.
> >
> > To expand on Jp's response a little: please explain why you think
> > you want this, so we can better understand the need that's driving it.
> > Maybe you are trying to do something you shouldn't be trying
> > to do, or maybe you're just going about it the wrong way.  What
> > makes you think doing what you are asking is a good thing?
> >
> > -Peter
> 
> I am working in test environment, and I didn't like testers alter my
> script (this happened to me). Basicly, I want my scripts are under control
> with my knowledge. If the problem come up, I know where to start
> troubleshooting.

Hmm... when we have a situation where someone does something inappropriate, 
we ask them not to, and explain why.  Generally they don't do it again...

Better than using individual .py[oc] files then would be to package
the program using py2exe, wouldn't it?  Are these people testing this
code itself, or using the code in the process of testing something 
else?  

If they're testing this code, then they're idiots and should
be fired if they go changing it on you. 

If they are using the code as a tool for testing something else,
you should probably have a proper release process for the tool, 
involving revision control, version numbers, and a nice little
packaging and installation sequence which prevents the whole issue
from happening in the first place.

Anyway, what makes you think that they won't do things like install
their own replacement .py files which do whatever they want, rather
than use your .pyc file, if they're such tricky types that they would
change your .py files on you?

(My advice: use education of the users to solve your problem, not 
fancy technical solutions.)

-Peter



From dirk at gerrits.homeip.net  Tue Apr 22 01:06:11 2003
From: dirk at gerrits.homeip.net (Dirk Gerrits)
Date: Tue, 22 Apr 2003 07:06:11 +0200
Subject: IDLE question
In-Reply-To: <20030421222056.5204.qmail@web11602.mail.yahoo.com>
References: <20030421222056.5204.qmail@web11602.mail.yahoo.com>
Message-ID: <b82iie$kec$1@main.gmane.org>

Stan Baptista wrote:

> I'm running IDLE on Win2K.  Is it possible to change
> the default subdir on a file open (menu
> File->Open...)?

I think you can do it by changing the working directory in the IDLE 
shortcut's properties.

Regards,
Dirk Gerrits






From erno-news at erno.iki.fi  Wed Apr 23 12:24:47 2003
From: erno-news at erno.iki.fi (Erno Kuusela)
Date: 23 Apr 2003 19:24:47 +0300
Subject: dbm concurrency
References: <6c492956.0304230623.74b92d7f@posting.google.com>
Message-ID: <kufzo9mbao.fsf@lasipalatsi.fi>

In article <6c492956.0304230623.74b92d7f at posting.google.com>,
haribeau at gmx.de (Clemens Hermann) writes:

| Might this be a problem when the C-program requests read-only access
| while python is writing to the dbm or vice versa? If so, how can this
| be solved?

yes. you should lock the file in both access cases, with shared lock
for readers and exclusive lock for the writer. in lockf terms,
LOCK_SH and LOCK_EX respectively.

  -- erno



From mcfletch at rogers.com  Tue Apr  8 17:20:50 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 08 Apr 2003 17:20:50 -0400
Subject: Surprise: del weakref.WeakKeyDictionary[ item ] raises RuntimeError
Message-ID: <3E933D32.8060500@rogers.com>

Just started getting this very surprising RuntimeError popping up when 
my application exits (occuring during a weakref callback, actually).  
Turns out that weakref.WeakKeyDictionary uses this idiom to do a deletion:

        for ref in self.data.iterkeys():
            o = ref()
            if o == key:
                del self.data[ref]
                return

which will raise a RuntimeError if the size of the data dictionary 
changes during the iterkeys loop.  Using keys seems to avoid the 
problem, but with the obvious disadvantages of keys vs. iterkeys.

Somewhat surprising behaviour when you're used to most 
dictionary/mapping behaviours being pretty close to atomic :) .  Anyway, 
just thought there might be others out there in time and space who might 
find it helpful to know about it.

Enjoy,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From aahz at pythoncraft.com  Fri Apr  4 13:02:37 2003
From: aahz at pythoncraft.com (Aahz)
Date: 4 Apr 2003 13:02:37 -0500
Subject: Why deleting large variables doesn't free up any memory?
References: <169a7283.0304040936.69e2b9a1@posting.google.com>
Message-ID: <b6khbt$ftf$1@panix2.panix.com>

In article <169a7283.0304040936.69e2b9a1 at posting.google.com>,
sdieselil <sdieselil at yahoo.com> wrote:
>
>I'm using Python 2.2.2 under FreeBSD. When I create a large variable
>(using for example "a = range(1000000)") 'top' command indicates that
>Python ate 12MB of swap space. But when I execute "del a" nothing
>happens, it doesn't release these 12MB! Using "gc.collect()" also
>doesn't help. What's the problem? What happened to Python's garbage
>collector?

That's got nothing to do with Python; generally speaking, once the OS
allocates memory to a process, the process keeps it until it exits.
However, if you allocate and de-allocate big chunks of memory, you'll
notice that Python's size stays static -- the memory gets re-used.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From python at dylanreinhardt.com  Wed Apr 23 12:54:05 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 23 Apr 2003 09:54:05 -0700
Subject: Zope and FreeBSD
In-Reply-To: <e786d63d.0304201806.4c5293a6@posting.google.com>
References: <e786d63d.0304201806.4c5293a6@posting.google.com>
Message-ID: <1051116844.19783.115.camel@ida.dylanreinhardt.com>

You didn't say if you were packing the ZOBD regularly.

If you aren't, you should be... at least daily.  Perhaps more often if
you've got objects that get updated regularly.

Also note that Zope is in version 2.6.1, though I used 2.5.1 until just
a couple weeks ago and never had this problem.

Dylan



On Sun, 2003-04-20 at 19:06, Wenshan Du wrote:
> Hi,
> 
> I've built a website with Zope 2.5.1 and FreeBSD 4.3, it's better than
> other solutions.
> 
> The server got so slow recently that I have to reboot every 3 days,
> and Zope stop sometimes.
> 
> I don't know why, is zope not stable on FreeBSD? or FreeBSD got virus?





From ckea25d02 at sneakemail.com  Fri Apr 11 02:30:35 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 11 Apr 2003 18:30:35 +1200
Subject: class too big for one file! How can I split it?
In-Reply-To: <3E95797F.3A24944F@engcorp.com>
References: <e956b5f1.0304100512.597654d3@posting.google.com> <3E95797F.3A24944F@engcorp.com>
Message-ID: <b75mrb$be69r$1@ID-169208.news.dfncis.de>

Peter Hansen wrote:
> HHaegele wrote:
> 
>>Now, OTLib.py has more then 32000 lines. And that doesn?t work. Python
>>1.5 can?t execute then. 
> 
> I presume you've already eliminated the alternative of *not* using
> the as-good-as-obsolete Python 1.5?

Assuming you have, you could split it into a number
of smaller classes, and then create one big class
that inherits from all of them.

You also might be able to reduce the number of
lines by putting multiple statements on one line
wherever possible, e.g.

   def Att(self): global XX; print XX

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From evan at 4-am.com  Mon Apr 28 14:14:28 2003
From: evan at 4-am.com (Evan Simpson)
Date: Mon, 28 Apr 2003 13:14:28 -0500
Subject: Zope, Relative Object Calls
In-Reply-To: <f1844358.0304281001.41a2d4b2@posting.google.com>
References: <f1844358.0304281001.41a2d4b2@posting.google.com>
Message-ID: <3EAD6F84.8070203@4-am.com>

John Withers wrote:
> In the application I am working on, tags like the following appear:
> <dtml-var "apps.profile.forms.rpt_fetchallrefs(_.None,_,aboutnum=_['serialnum'])">
> 
> I am trying to get something like a relative anchor, where:
> <a href="forms/rpt_fetchallrefs">.</a> would work

You'll be much more likely to get answers to this sort of question on 
the Zope mailing list than here.

That said, have you tried:

<dtml-var "forms.rpt_fetchallrefs(_.None,_,aboutnum=_['serialnum'])">

If the DTML is being called in the correct context, acquisition should 
make this work.

Cheers,

Evan @ 4-am






From m.faassen at vet.uu.nl  Wed Apr  2 00:48:39 2003
From: m.faassen at vet.uu.nl (Martijn Faassen)
Date: 2 Apr 2003 05:48:39 GMT
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de> <b6d5s3$lvo$2@newshost.accu.uu.nl> <pan.2003.04.02.07.45.41.185943.810@thomas-guettler.de>
Message-ID: <b6dtjn$unc$1@newshost.accu.uu.nl>

Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote:
> On Wed, 02 Apr 2003 01:03:31 +0200, Martijn Faassen wrote:
> 
>> Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote:
>>> I looked at zope3 but it seems very bloated.
>> 
>> It isn't done yet, obviously. How does it seem bloated? Too much
>> framework for your tastes?
> 
> Yes, too much framework. 
> 
> I like ZODB, but don't like ZPT, DTML, Acquisistion, ZClasses,
> Interfaces and Components.

Zope 3 doesn't have much in the way of DTML (it's there but completely 
optional) or ZClasses. Acquisition is far less prominent as well. It does have
ZPT and interfaces and components.

Also it isn't fully baked yet. :)

Regards,

Martijn
-- 
History of the 20th Century: WW1, WW2, WW3?
No, WWW -- Could we be going in the right direction?



From ckea25d02 at sneakemail.com  Mon Apr  7 00:02:00 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 16:02:00 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1049662570.29853.python-list@python.org>
References: <m2u1dcw22s.fsf@localhost.localdomain>	 <6ee58e07.0304061144.35c5ab92@posting.google.com>	 <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org>
Message-ID: <b6qsmb$85j4g$1@ID-169208.news.dfncis.de>

Ian Bicking wrote:
> No, don't change the interface! ...  The
> minibuffer is vastly superior to popup dialogs.

One thing I would suggest is to give each window its own
minibuffer (appearing when needed). I find it very annoying
when I start doing something with the minibuffer, temporarily
want to go and do something in another window which requires
using the minibuffer, and get "Minibuffer already in use,
use control-meta-cokebottle to abort"... Aaaarghhh!

> I think this is why vi remains popular -- even though I personally don't
> like it -- because it's modal.

Er, I'm afraid I can't agree that modality is a good thing
in general. It has its uses, but only in limited contexts, I
find. As in the above -- having each window modal with respect
to its own minibuffer is okay; having the *entire* application
modal with respect to *one* minibuffer is a nuisance.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mwm at mired.org  Tue Apr  1 11:55:09 2003
From: mwm at mired.org (Mike Meyer)
Date: Tue, 01 Apr 2003 16:55:09 GMT
Subject: The Missing Manual? (Re: Finding Python help...)
References: <3E88310F.6090709@mq.edu.au> <b6aanj$39skv$1@ID-71831.news.dfncis.de> <b6am14$3f4rk$1@ID-169208.news.dfncis.de>
Message-ID: <x7k7eejgya.fsf@guru.mired.org>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> writes:
> Maybe there should be a THIRD manual, called something like
> "Built-in Objects Reference". Or maybe even "Built-in Objects
> and Functions Reference", since the builtin functions seem
> pretty fundamental as well, and not something you'd
> intuitively expect to find in a "library".

Instead of a third manual, how about moving the first section of the
Library Reference - "Built-in Functions, Types, and Exceptions" - into
the Language reference manual? It's not really thick enough to be
worth a book by itself.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From rmunn at pobox.com  Fri Apr 18 16:24:10 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 18 Apr 2003 20:24:10 GMT
Subject: Which documentation for 2.2?
References: <mailman.1050685171.4364.python-list@python.org>
Message-ID: <slrnba02vt.lc9.rmunn@localhost.localdomain>

Paul Berkowitz <berkowit at silcom.com> wrote:
> Can someone tell me why there are two separate sets of Documentation at
> python.org for Python 2.2? One was released 21 December 2001, the other on
> 29 March 2002. When you go to the latter it says it's Python 2.2p1. Is this
> an updated (corrected?) documentation for the same Python 2.2 release as the
> other (in which case why is the earlier set still published), or is it
> documentation for a 2.2p1 update of Python itself (in which case why doesn't
> it say so on the main page?)
> 
> Thanks for any clarification which can be provided. I'm trying to discover
> which set applies to the Python 2.2 installed by Mac OS 10.2. I will ask at
> PythonMac SIG list as well.

What are the contents of sys.version_info? On my copy of 2.2.2:

$ python
Python 2.2.2 (#1, Jan 12 2003, 12:07:20) 
[GCC 3.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.version_info
(2, 2, 2, 'final', 0)
>>> 

What does your copy show in version_info?

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From grahamd at dscpl.com.au  Tue Apr 22 20:37:13 2003
From: grahamd at dscpl.com.au (Graham Dumpleton)
Date: 22 Apr 2003 17:37:13 -0700
Subject: xmlrpc introspection
References: <mailman.1051031252.20283.python-list@python.org>
Message-ID: <dc6f5c99.0304221637.586dfc4c@posting.google.com>

Even if you can find the API definition, you are still going to have a couple
of decisions to make as to how you do it in Python. The main issue is that
the XML-RPC introspection API has the intent of expressing the method
signatures using strong typing. Since Python isn't strongly typed, in the worst
case you have to provide a way of manually enumerating what all the
possible argument sets for a method might be. This can be a pain and you
might instead loosen up your introspection API implementation a bit and
simply pass back the names of the arguments the exported methods have.

For an example implementation based on a messaging framework which
doesn't actually use the Python xmlrpc module but still provides a gateway
for XML-RPC, check out:

  http://ose.sourceforge.net/browse-example.php?example=introspection-service.py

There are two examples of how to do the introspection API in that file
corresponding to the choice explained above. Both methods use a methods
doc string for the method help.

Obviously, you will need to adapt this to whatever XML-RPC module you
are using but it should give you a start.

When you are done, if you want to see if it works and you can make your
service contactable on the net, use:

  http://www.dscpl.com.au/xmlrpc-introspect.php

and:

  http://www.dscpl.com.au/xmlrpc-debugger.php

Hmmm, I hope these two things still work, the firewall where I am working
clags when I use the introspection page.

Charlene and Bryan <belred1 at yahoo.com> wrote in message news:<mailman.1051031252.20283.python-list at python.org>...
> thanks skip,
> 
> but that site didn't have what i'm after.  the
> introspection API link is dead.  i went to frontier
> since i read that they updated their docs, but i can't
> find the updates with introspection anywhere.  i did
> google around first, but no luck anywhere.
> 
> bryan
> 
> --- Skip Montanaro <skip at pobox.com> wrote:
> > 
> >     Bryan> i'm interested in implementing the xmlrpc
> > introspecton methods
> >     Bryan> ...
> > 
> > Google is your friend.  Searching for "xmlrpc
> > introspection" yields this
> > promising item.
> > 
> >    
> >
>  http://xmlrpc-c.sourceforge.net/xmlrpc-howto/xmlrpc-howto-api-introspection.html
> > 
> > Skip
> 
> 
> __________________________________________________
> Do you Yahoo!?
> The New Yahoo! Search - Faster. Easier. Bingo
> http://search.yahoo.com



From tjreedy at udel.edu  Tue Apr 15 08:13:54 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 15 Apr 2003 08:13:54 -0400
Subject: files: direct access - NOT sequential
References: <mailman.1050398560.8881.python-list@python.org>
Message-ID: <33WdnaWr_sALaAajXTWcpQ@comcast.com>

"Patrick Carabin" <Patrick.Carabin at sciencesnaturelles.be> wrote in
message news:mailman.1050398560.8881.python-list at python.org...
i search a means to get " direct access " inside files, eg mixing
read & write operaions for the same file
example:
given a file with contents : "qwerty" , positionning on the "w", and
writing a "N" , so the file becomes "qNerty"
How can i achieve this in Python ?
----------
open in 'w' mode or (probably better for Windows,  'wb') mode and use
tell() and seek() methods.  tjr





From bdesth at removethis.free.fr  Tue Apr  8 17:18:05 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 08 Apr 2003 23:18:05 +0200
Subject: Newbie: Python + C Question
References: <3e92ebe8$1@rmcnt03a.rmc.ca>
Message-ID: <3E933C8D.5020700@removethis.free.fr>

bourassa wrote:
> Hi there.  I have a question about Python and the interaction with C. 
>  From Python I can use C extensions.  But does doing so speed up 
> execution speed? 

Yes, unless your C code is braindead buggy

> When Python uses the C extension does it just process 
> it in an interpreted fashion

Python is not a C interpreter.

> or does the code written in C run like a 
> compiled C program?  

newbie to programming, uh ?-)

How do you think this all works ?

the Python interpreter byte-compile and execute python source code, not 
C source code. C being a compiled language, C source code is just text, 
nothing you can execute before compilation (ie : transformation of 
source code into machine code).

When you use a "C extension" in Python, you use in fact a 
machine-language compiled extension. The 'C extension' term is faulty. 
The fact is that the extenstion has generally been written in C before 
compilation, hence the term, but  open such an extension with your 
favorite code editor, I can tell you you wont find anything like a C 
source in it !-)

BTW, Python (I mean the Python interpreter) itself is written in C.

Laotseu.




From ianb at colorstudy.com  Wed Apr 16 16:58:12 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 16 Apr 2003 15:58:12 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <dl8na.20637$LB6.515421@news1.tin.it>
References: <cQ_la.5597$LB6.158861@news1.tin.it>
	 <mailman.1050210449.24337.python-list@python.org>
	 <23891c90.0304140221.20e9e7fb@posting.google.com>
	 <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
	 <B8Pma.20554$T34.656139@news2.tin.it>
	 <mailman.1050471334.22000.python-list@python.org>
	 <dl8na.20637$LB6.515421@news1.tin.it>
Message-ID: <1050526692.30714.586.camel@lothlorien>

On Wed, 2003-04-16 at 03:18, Alex Martelli wrote:
> Possible, yes.  Feasible, I dunno.  Compare Cheetah and preppy, which
> both process templates to produce Python modules.  preppy does it in
> a very simple, rather Pythonic way -- inserting "directives" enclosed
> in {{...}} that are Python expressions or a few Python statements
> modified to have explicit endif/endfor/&c termination rather than
> relying on indentation.  Any Python programmer should have no problem
> understanding preppy in 10 minutes and then happily using e.g.
>     {{myobject.func()}}
> or
>     {{myobject['index']}}
> or
>     {{myobject.attr}}
> &c &c in a fully appropriate manner -- since by definition a Python
> programmer already understands perfectly the difference between
> calling a function (and KNOWS you need empty parentheses when you
> call a function without arguments -- something that keeps tripping
> beginners up, btw), indexing a container, accessing an attribute.
> 
> But suppose your design intention is different -- you want your
> templates to be maintained by NON-programmers, people which DO
> have problems understanding the differences between the above
> distinct concepts!  In Cheetah, the above constructs would be
>     $myobject.func
>     $myobject.index
>     $myobject.attr
> i.e., totally polymorphic.  The Cheetah designer, on the basis of
> a LOT of experience trying to have non-programmers maintain
> templates, believe this extreme polymorphism makes things VERY
> much easier for non-programmers (perhaps particularly ones that
> have some previous exposure to languageoids such as Javascript
> and VBScript that may also try to hide such differences...).

That's okay though, because both just need a namespace to evaluate in. 
Cheetah's handling of function calls and whatnot don't change its
interface, just the template language.  A DOM-like templating language
obviously needs a different interface, but otherwise all the other
templates basically take a namespace and are otherwise self-executing.

So a template needs a source text and a namespace, and probably a way to
indicate certain other options (say, some languages may have a
case-insensitivity option, but others won't -- Cheetah allows the
NameMapper stuff to be turned off, for instance, making it more like
preppy).  But there's still lots that they could share.

For a difference in interface, Cheetah takes a list of namespaces that
get folded together.  I haven't looked at preppy, but let's say it takes
a dictionary or a single namespace.  That's a different interface, but
it can be turned into one interface, and then both templating languages
will benefit.

  Ian






From aleax at aleax.it  Sat Apr 26 16:59:19 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 20:59:19 GMT
Subject: please comment on technologies
References: <slrnbalplq.254.luc@trillian.dont-panic.info>
Message-ID: <HqCqa.8931$3M4.241359@news1.tin.it>

luc wastiaux wrote:

> Hello, I am writing the specifications for a school project (the subject
> is free for us to choose), I was thinking about doing a news-server
> archive database just like google groups, but on a smaller scale (it would
> only archive a limited number of newsgroup, not all of USENET). I got this
> idea because my school uses a news server and a lot of valuable
> information is lost when old messages are erased from the server's spool.

Sounds like a cool and feasible idea to me.


> The project would be developped by a group of 4 to 6 CS students
> (including me), over a three months timespan (but not full time, we have
> to go to classes and other stuff)

The most valuable thing you'll learn will be how to structure your
work.  I heartily recommend incremental development (most crucial
features first), test-driven design, pair programming, and a healthy
diffidence towards "big design up front" - you may get the temptation
of designing in some big chunks of infrastructure preparing for future
enhancements... that may never come.


> I have a small experience with web stuff, but mostly with PHP, and some
> perl. and I'm learning python right now.
> 
> I'm (almost) sure we will make use of python as the main language, and use

Sounds like an excellent idea to me.

> either Mysql or postgresql for the database (I'm familiar with mysql but
> maybe it's too limited for what I want to do?). I just found out about

MySQL has limitations, but it's reasonably fast and simple for use
within its limitations.  You might want to program most of your app
to a "database engine independent layer" -- so you could start using
MySQL but still be able to easily switch to e.g. PostgreSQL if you
do turn out to need something that's a problem in MySQL, eg. Views.

> Spyce today and it looks like I could use it for the web interface.

Sure, that's one possibility.  Personally I'd choose Webware, which
offers an amount of flexibility I appreciate, but Spyce's OK too.

> Then I have this idea about using XSL for templating, the idea is not only
> to separate the data from the representation, but also allow programmers
> to "leech off" the search engine (it's easier to do this if the output is
> XML).

The ability for other programmers to send "queries" to your search
engine and get XML output is good.  However, personally, I would avoid
XSL (unless you guys are already well familiar with it -- or NEED to
get very familiar with it anyway to pursue your studies).  It's simpler
to have separate URLs, or a form field requesting XML or HTML output,
and just use different templates with the same contents -- doing your
templating in Cheetah (or preppy/Spyce/whatever).


Alex




From aleax at aleax.it  Wed Apr 23 17:49:51 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 21:49:51 GMT
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <NKxpa.85054$T34.2112670@news2.tin.it> <b86s4k$as7$0@216.39.172.122>
Message-ID: <3UDpa.87972$T34.2186535@news2.tin.it>

Bengt Richter wrote:

> On Wed, 23 Apr 2003 14:50:21 GMT, Alex Martelli <aleax at aleax.it> wrote:
> [...]
>>However, such new functionality doesn't get back-ported to previous
>>releases of Python, such as 2.2.*.  In all Python releases from 1,6
>>included to 2.3 excluded, "needle.find(haystack) > 0" is the idiom.
> 
> If needle is a substring to be found in haystack, that seems reversed.

Ooops, yes -- "haystack.find(needle) >= 0" is what I meant.  Pretty
bad thinko there -- sorry!


> Were you recently using re to find needles? E.g.,
> 
>  >>> import re
>  >>> needle = re.compile('needle')
>  >>> needle.findall(haystack)
>  ['needle', 'needle']

Yep.  If you're in no special hurry, that works, too.

Finding just one (leftmost) occurrence, string vs re:

[alex at lancelot src]$ ./python -O Lib/timeit.py -s'import re' 
-s'haystack="needle in haystack with another needle somewhere"; 
needle="needle"' -s'reneedle=re.compile(needle)'  'if haystack.find(needle) 
>= 0: pass'
1000000 loops, best of 3: 1.04 usec per loop

[alex at lancelot src]$ ./python -O Lib/timeit.py -s'import re' 
-s'haystack="needle in haystack with another needle somewhere"; 
needle="needle"' -s'reneedle=re.compile(needle)'  'if 
reneedle.search(haystack): pass'
1000000 loops, best of 3: 1.41 usec per loop

Finding / counting all occurrences, string vs re:

[alex at lancelot src]$ ./python -O Lib/timeit.py -s'import re' 
-s'haystack="needle in haystack with another needle somewhere"; 
needle="needle"' -s'reneedle=re.compile(needle)'  'if 
haystack.count(needle): pass'
1000000 loops, best of 3: 1.77 usec per loop

[alex at lancelot src]$ ./python -O Lib/timeit.py -s'import re' 
-s'haystack="needle in haystack with another needle somewhere"; 
needle="needle"' -s'reneedle=re.compile(needle)'  'if 
reneedle.findall(haystack): pass'
100000 loops, best of 3: 4.23 usec per loop


but of course, in Python 2.3, the 'in' operator is fastest as well
as handiest (and it, also, requires the needle on the left):

[alex at lancelot src]$ ./python -O Lib/timeit.py -s'import re' 
-s'haystack="needle in haystack with another needle somewhere"; 
needle="needle"' -s'reneedle=re.compile(needle)'  'if needle in haystack: 
pass'
1000000 loops, best of 3: 0.272 usec per loop


RE's have the advantage of easily allowing case-insensitive search -- that's
my most common reason for using them.  Second most common is \b, the "word
boundary" marker.


Alex




From roy at panix.com  Sat Apr 12 20:14:51 2003
From: roy at panix.com (Roy Smith)
Date: Sat, 12 Apr 2003 20:14:51 -0400
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it>
Message-ID: <roy-06B012.20145112042003@reader1.panix.com>

Alex Martelli <aleax at aleax.it> wrote:
> Yes -- it's THE one major performance trap in naively coded Python,
> since, when "a lot" is a sufficiently large N, a lot of += on short
> strings ends up being O(N squared) where O(N) approaches would be
> available.

One of the very few things I like about C++ over Python is that the STL 
is very careful to document the complexity of the various operations.  
AFAICT, no such documentation exists for Python.

True, it's documented that strings are immutable, and from that you 
should be able to infer that string addition requires lots of creating 
and destroying of objects, which sounds like it should work out to being 
quadradic, but that's not quite the same as having the algorithmic 
complexity spelled out for you.

As another example, I can guess that Python dictionaries work like STL 
maps on the inside, but that would be just a guess.  Is inserting an 
element logN?  Is keys() linear?  Is del on an element logN?  Is 
has_key() logN?  Those would be my guesses.  But they're just guesses.



From member27604 at dbforums.com  Tue Apr  8 05:29:53 2003
From: member27604 at dbforums.com (dolephi9080)
Date: Tue, 08 Apr 2003 09:29:53 +0000
Subject: [help]Is this a bug or my code's fault
Message-ID: <2741546.1049794193@dbforums.com>

import Tkinter
import sys
import os
import os.path



if __name__ == "__main__":
    rt = Tkinter.Tk()
    btn = Tkinter.Button(rt,text='one')
    btn.setvar('v','one')  #<=here I set a var to the btn object
    with v='one'
    btn.pack()
    btn1=Tkinter.Button(rt,text="aa")
    btn1.setvar('v','aa')   #<=here I set a vat to the btn1 object
    v='aa'
    btn1.pack()
    print btn1.getvar('v') #<= print the btn1 v
    print btn.getvar('v') #<= print the btn v

    rt.mainloop()




please look at the comment, I set the value to two buttons, but when I
print it at last, two buttons got the same data. I don't know why when I
set the btn1's value, it will rewrite the btn's value, is this a bug? by
the way I'm using activestate python 2.2.2

--
Posted via http://dbforums.com



From i.linkweiler at web.de  Sat Apr  5 04:48:38 2003
From: i.linkweiler at web.de (Ingo Linkweiler)
Date: Sat, 05 Apr 2003 11:48:38 +0200
Subject: PyWin: Redraw during long loops?
Message-ID: <3E8EA676.90008@web.de>

Hello,

I am using PythonWin.
When my programs are in a long loop, the redraw of Pywin is blocked.
PyWin only redraws its window and handles events, when a print (with 
newline) or input is done.

Is there a way to avoid this?

Ingo




From max at alcyone.com  Thu Apr 10 20:19:17 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 10 Apr 2003 17:19:17 -0700
Subject: Can I define nested classes in Python?
References: <169a7283.0304101432.13f5786c@posting.google.com>
Message-ID: <3E960A05.AD1EF8A9@alcyone.com>

sdieselil wrote:

> Can I define nested classes in Python and what access do they have to
> outer
> classes?

I answered this question in comp.os.linux.misc, where you originally
asked it.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ We're here to preserve democracy, not to practice it.
\__/ Capt. Frank Rasmey
    REALpolitik / http://www.realpolitik.com/
 Get your own customized newsfeed online in realtime ... for free!



From sholden at holdenweb.com  Thu Apr 17 09:30:31 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 13:30:31 GMT
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com> <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk> <3E9D8B1B.24050D34@engcorp.com> <7h3brz6qmsc.fsf@pc150.maths.bris.ac.uk> <3E9DB4AC.E94829D9@engcorp.com> <3e9e5242_2@news.estpak.ee> <7h3u1cxpmlr.fsf@pc150.maths.bris.ac.uk>
Message-ID: <X%xna.114962$It5.68496@news2.central.cox.net>

"Michael Hudson" <mwh at python.net> wrote ...
> Vahur Lokk <vlokk at hot.ee> writes:
>
> > Peter Hansen wrote:
> >
> > > Michael Hudson wrote:
> >
> > Ooops!
> > I remember my firstever post to linux mailing list caused a month-long
> > flamewar there ;-)
> > Did I do it again?
>
> If this is your idea of a flame war, I suggest you try
> comp.lang.perl.misc :-)
>
> As far as I can tell Peter & I are merely vigorously agreeing with
> each other...
>

Quick, somebody, throw a bucket of water on them!

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From pc451 at yahoo.com  Wed Apr 16 21:20:28 2003
From: pc451 at yahoo.com (Peter Clark)
Date: 16 Apr 2003 18:20:28 -0700
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com> <TUhna.928$cq2.100423716@newssvr13.news.prodigy.com>
Message-ID: <f7199550.0304161720.702ac0f6@posting.google.com>

"Dennis Reinhardt" <DennisR at dair.com> wrote in message news:<TUhna.928$cq2.100423716 at newssvr13.news.prodigy.com>...
> > I'm working with a xml document which doesn't include an encoding,
> 
> The lack of an encoding is your problem.  Pick an encoding which defines 256
> characters, not 128, if you want to represent chr(176).in a single byte.
    I'm sorry, upon re-reading my first message, I realize that I did
not describe the problem adequately enough. Let me see if I can try
again.
    When I try chr(176).encode('latin-1') I get the following message:
UnicodeError: ASCII decoding error: ordinal not in range(128). But
when I try 'print chr(176)' I get the degree symbol, which is what I
want. Yet when I try inserting this into the XML stream, I get the
following error:

    w += [chr(176) + scale.strip()]
UnicodeError: ASCII decoding error: ordinal not in range(128)

    I've tried changing its encoding to 'utf-8' as well with
chr(176).encode('utf-8'), but that just returns the same error. This
doesn't make sense: if 'print chr(176)' works fine, why doesn't it
work later on? In a nut shell, all I want to do is add the degree
character to a string contained within a list. That doesn't seem hard
at all, but I'm completely stumped.



From pjdurai at hotmail.com  Fri Apr 25 16:38:33 2003
From: pjdurai at hotmail.com (pj)
Date: 25 Apr 2003 20:38:33 GMT
Subject: GUI testing with python  (SendKeys)
Message-ID: <b8c6c9$p2m$0@198.102.102.163>

Greetings

Is there a python module similar to Perl's Win32::GuiTest?
Basically something that would implement SendKeys and friends on Win32 ?

Before I went ahead and rolled my own, I just wanted to check.
Web searches didnt turn up anything useful.

cheers
pj






From peter at engcorp.com  Fri Apr 11 20:47:32 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 11 Apr 2003 20:47:32 -0400
Subject: ASN.1
References: <mailman.1050092902.7843.python-list@python.org>
Message-ID: <3E976224.C0A82103@engcorp.com>

> "Batista, Facundo" wrote:
> 
> Hi people!
> 
> I'm searching a ASN.1 parser, to load the structure of a file to memory and then be able to parse the file.
> 
> Do you know something like this for python?

There's twisted.conch.ssh.asn1.py in Twisted.  
See http://www.twistedmatrix.com.

-Peter



From sismex01 at hebmex.com  Tue Apr 29 11:56:17 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Tue, 29 Apr 2003 10:56:17 -0500
Subject: Problem subclassing tuple
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F062330489D361@mail.hebmex.com>

> From: John Wilson [mailto:tug at wilson.co.uk]
> Sent: Tuesday, April 29, 2003 10:55 AM
> 
> Given this class:
> 
> class Holder(tuple):
>     def __init__(self, *values):
>         tuple.__init__(self, values)
> 
> 
> When I try to create an instance:
> 
> Holder(1, 2)
> 
> I get:
> 
> TypeError: tuple() takes at most 1 argument (2 given)
> 
> It looks like the constructor for the superclass is being 
> called rather than the __init__ method on Holder.
> 
> I'm obviously missing something here. Do I have to supply a 
> __new__ method? Any help would be appreciated.
> 
> John Wilson
>

Howdy.

First of all... What are you trying to do? :-)

Second.  If that's the full definition and intention
of "Holder", then you might want to know (or are
overlooking) the fact that "values", inside __init__,
is already a tuple.

Instead of a class Holder, you could build a function
holder() which simply returns the tuple it receives:

def holder(*values):
    return values

One of the most useful functions I've encountered
which simplified my (programming) life in small but
noticeable ways, is the following:

def Dict(**values):
    return values

This allows me to define dictionaries like this:

d = Dict(one=1, two="Two", three=...)

instead of this:

d = { "one":1, "two":"Two", ... }

So...

hth.

-gus




From tyler at cg1.org  Sun Apr 13 16:48:31 2003
From: tyler at cg1.org (Tyler Eaves)
Date: Sun, 13 Apr 2003 20:48:31 GMT
Subject: Help with CGI
References: <pan.2003.04.13.19.41.42.420658@soul.net>
Message-ID: <z2kma.284$8c4.159@tornadotest1.news.pas.earthlink.net>

On Sunday 13 April 2003 15:41, lost wrote:

> form = cgi.FieldStorage

There's your problem. 
That line should read: 

form = cg1.FieldStorage()

The () is the difference between what you want, and simply getting a copy of 
the class.




From exarkun at intarweb.us  Thu Apr 24 02:09:49 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 24 Apr 2003 02:09:49 -0400
Subject: Creating dynamically named lists/dictionaries
In-Reply-To: <ydFpa.15$IO5.10324@newsfep2-gui.server.ntli.net>
References: <ydFpa.15$IO5.10324@newsfep2-gui.server.ntli.net>
Message-ID: <20030424060949.GA17587@meson.dyndns.org>

On Thu, Apr 24, 2003 at 12:22:18AM +0100, Rogue 9 wrote:
> Hi all,

  Hi

> [snip]

    http://perl.plover.com/varvarname.html

  Jp

-- 
"Minerals are inexhaustible and will never be depleted. A stream of
investment creates additions to proved reserves, a very large in-ground
inventory, constantly renewed as it is extracted... How much was in the
ground at the start and how much will be left at the end are unknown and
irrelevant."  -- Morry Adelman, World Renowned Economist
-- 
 up 35 days, 3:03, 9 users, load average: 0.60, 0.73, 0.93




From amk at amk.ca  Mon Apr 21 21:12:01 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Mon, 21 Apr 2003 20:12:01 -0500
Subject: regular expressions
References: <b81lq4$17n8$1@galaxy.us.dell.com> <1050961701.8396.0@despina.uk.clara.net>
Message-ID: <TAadnXAyb8p8CzmjXTWcow@speakeasy.net>

On Mon, 21 Apr 2003 23:00:21 +0100, 
	Garry Knight <garryknight at gmx.net> wrote:
> The best resource (IMO) is Jeffrey Friedl's "Mastering Regular Expressions",
> published by O'Reilly, ISBN 1-56592-257-3.

Note that the second edition of this book seems to have dropped all the
Python material, judging by the table of contents.  The first edition only
covered the regex module, not the re module.  It's worth a look if you want
to learn more about regexes, but it won't be useful as reference material.

--amk



From wesc at fuzzyorange.com  Tue Apr  8 14:09:53 2003
From: wesc at fuzzyorange.com (Wesley Chun)
Date: Tue, 8 Apr 2003 11:09:53 -0700 (PDT)
Subject: ANN: BayPIGgies mtg Wed Apr 9 7:30pm
In-Reply-To: <Pine.LNX.4.31.0304032309200.17780-100000@emperor.deirdre.org>
Message-ID: <Pine.LNX.4.31.0304081104430.16054-100000@emperor.deirdre.org>

final reminder for this month's meeting tomorrow nite...

come join us if u can

-wesley


On Thu, 3 Apr 2003, Wesley Chun wrote:

> BayPIGgies: Silicon Valley-San Francisco Bay Area Python Users Group
>
> When:     April 9, 2003 @ 7:30pm
> Where:    Stanford University, Palo Alto, CA
> Agenda:   Introduction to Machine Learning and Support Vector Machines
> Speaker:  Asa Ben-Hur
>
> The basic concepts of machine learning will be presented, with a focus
> on Support Vector Machines (SVM), which have shown their effectiveness
> in many domains of application. Examples of applying the methodology in
> areas such as text categorization and bioinformatics will be discussed.
>
> # Call For Talks: We are actively seeking speakers for BayPIGgies! If you
> would like to give a talk at one of our 2003 meetings (any Python related
> topic), contact us to coordinate!
>
> more info including directions:    http://www.baypiggies.net
>
> -wesley
>
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> "Core Python Programming", Prentice Hall PTR, (c) 2001
>     http://starship.python.net/crew/wesc/cpp/
>
> Silicon Valley-San Francisco Bay Area Python Users Group (BayPIGgies)
>     http://baypiggies.net
>
> wesley.j.chun :: wesc at deirdre.org or wesc at fuzzyorange.com
> cyberweb.consulting : henderson, nv : cyberweb at rocketmail.com
> http://www.roadkill.com/~wesc/cyberweb/





From florian.proff.schulze at gmx.net  Sun Apr 27 03:54:42 2003
From: florian.proff.schulze at gmx.net (Florian Schulze)
Date: Sun, 27 Apr 2003 09:54:42 +0200
Subject: Reading a non-standard floating point format
In-Reply-To: <24944f78.0304250631.1778654e@posting.google.com>
References: <24944f78.0304250631.1778654e@posting.google.com>
Message-ID: <oprn92xgqjttxc4i@news.t-online.de>

Could it be a pascal real value? I think the real values from turbo pascal 
where 6 byte.

Maybe this helps,
Florian

On 25 Apr 2003 07:31:49 -0700, Fidtz <fidtz at clara.co.uk> wrote:

> I'm trying to do a database extraction program. Part of this is
> reading a large number of files with a 48bit floating point format
> that I think from the definition and searching on google, is unlike
> any on any other system (at least current ones anyway).
>
> While I think I can see the way to do this with various bit-grubbing
> techniques, the potential for error is massive, esp considering my non
> comp-sci background.
>
> Does anyone know of a library in python or even (fairly) easily
> wrappable C that might let me specify the format in a more high level
> way?
>
> Dominic
>



-- 
Florian Schulze



From andrew-pythonlist at puzzling.org  Sun Apr 27 01:02:09 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 15:02:09 +1000
Subject: Looping around a simple list of tuples
In-Reply-To: <3EAB7213.4030209@magma.ca>
References: <3EAB7213.4030209@magma.ca>
Message-ID: <20030427050209.GJ26530@frobozz.local>

On Sun, Apr 27, 2003 at 04:45:16AM +0000, Martin d'Anjou wrote:
> I have a list of tuples, each tuple is a (scalar , list)
> 
> a=[("r1",[1,2,3]),("r2",[8,9,10]),("r5",[12,14,16])]
> 
> for el in a:
>    print el
>    for r,l in el:
>       print "R "+`r`
>       print "L "+`l`
> 
> ('r1', [1, 2, 3])
> R 'r'
> L '1'
> Traceback (innermost last):
>   File "<stdin>", line 3, in ?
> ValueError: unpack list of wrong size
> 
> In my mind, the 'el' implicit variable is a tuple, and the for r,l loop 
> should just work but it does not. How come?

I think what you want is simply:

    for r,l in a:
        print "R "+`r`
        print "L "+`l`

Which could be spelt out as:

    for el in a:
        r, l = el
        print "R "+`r`
        print "L "+`l`

Does this make sense?
    
-Andrew.





From sholden at holdenweb.com  Tue Apr 15 10:00:57 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 15 Apr 2003 14:00:57 GMT
Subject: DB API for ADO/DAO?
References: <2402252b.0304142353.2b4711ae@posting.google.com>
Message-ID: <tgUma.99329$It5.30169@news2.central.cox.net>

"John Burton" <john.burton at jbmail.com> wrote in message
news:2402252b.0304142353.2b4711ae at posting.google.com...
> I want to be use an .MDB database from a python application on
> windows.
>
> The problem with odbc is that you need to manually set up data sources
> using the control panel which isn't something I want the users to have
> to do.
>
> So I'm using DAO via com to access the database and it's working fine
> except that I have to use the DAO com api instead of the standard
> python api. So I was wondering if there is a suitable DB API wrapper
> for DAO or ADO I can use? Or a better solution to my problem?

As well as the other responses pointing you to ADO code, also note that
mxODBC allows you to use DSN-less data sources.  From the mxODBC docs: """If
you want to connect to a file data source (without having to configure it
using the ODBC manager), you can do so by using the FILEDSN= parameter
instead of the DSN= parameter:
     DriverConnect('FILEDSN=test.dsn;UID=test;PWD=test')

This is sometimes useful when you want to dynamically setup a data source,
e.g. a MS Access database."""

Thanks to Marc-Andre Lemburg for pointing this out (and also for a fine
database module).

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From ckea25d02 at sneakemail.com  Sun Apr  6 22:46:29 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 14:46:29 +1200
Subject: output buffering--am I missing something obvious?
In-Reply-To: <8yupdygb.fsf@morpheus.demon.co.uk>
References: <yu99istu9ds9.fsf@europa.research.att.com> <16013.63756.223973.484125@montanaro.dyndns.org> <mailman.1049492108.15272.python-list@python.org> <8yupdygb.fsf@morpheus.demon.co.uk>
Message-ID: <b6qo8n$832vf$2@ID-169208.news.dfncis.de>

Paul Moore wrote:
> Readline (or lack of readline) strangeness...?

Perhaps readline prints a CR (without LF) before
writing the prompt, to make sure it's at the
beginning of a line before getting input?

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From sdhyok at yahoo.com  Wed Apr 16 22:02:09 2003
From: sdhyok at yahoo.com (sdhyok)
Date: 16 Apr 2003 19:02:09 -0700
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <Pine.OSX.4.53.0304161637310.801@mac10.cdc.noaa.gov>
Message-ID: <420ced91.0304161802.4a5e0b85@posting.google.com>

Jeffrey.S.Whitaker at noaa.gov wrote in message news:<Pine.OSX.4.53.0304161637310.801 at mac10.cdc.noaa.gov>...
> On Mon, 14 Apr 2003, sdhyok wrote:
> 
> The mx.DateTime module will do what you ask
> (http://www.egenix.com/files/python/mxDateTime.html).  For example, if you
> want hours since Jan 1 1900:
> 
> import mx.DateTime
> year = 1900
> month = 1
> day = 1
> hour = 0
> refdate = mx.DateTime.Date(year,month,day,hour)
> curdate = mx.DateTime.now()
> datediff = curdate - refdate
> print datediff.hours
> 
> 
> -Jeff
> 
>  --
> Jeffrey S. Whitaker         Phone : (303)497-6313
> NOAA/OAR/CDC  R/CDC1        FAX   : (303)497-6449
> 325 Broadway                Web   : http://www.cdc.noaa.gov/~jsw
> Boulder, CO, USA 80305-3328 Office: Skaggs Research Cntr 1D-124

Thanks, Jeff. 
This is exactly what I sought.



From peter at engcorp.com  Mon Apr  7 06:54:02 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 07 Apr 2003 06:54:02 -0400
Subject: cgi form handler question
References: <3e90ff67$0$13167$5a62ac22@freenews.iinet.net.au> <mailman.1049691790.26318.python-list@python.org>
Message-ID: <3E9158CA.C53791E5@engcorp.com>

Ian Bicking wrote:
> 
> > the user forgets to input his email and age
> > how would i write a for loop to output:
> >
> > email
> > age
>
> You'd have to define the fields you required, like:
> 
> requiredFields = ['name', 'email', 'address', 'age', 'salary']
> for field in requiredFields:
>     if not form.has_key(field):
>         print field
> 
> Or you might be able to do:
> 
> for name in form.keys():
>     if not form[name].value:
>         print name

The latter suggestion won't usually work, as the data sent
from the browser does not even contain the names of the fields
which are left empty. 

-Peter



From brian at sweetapp.com  Mon Apr  7 22:29:25 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Mon, 07 Apr 2003 19:29:25 -0700
Subject: How to protect python code ?
In-Reply-To: <slrnb945tu.p4.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <003301c2fd76$ac6b0be0$21795418@dell1700>

> I didn't read facetiousness; rather, I read in the original question 
> a sentiment perhaps better expressed as:
> 
>   "If a business *pays* to become dependent on software they don't
>   control, what can one conclude about that business?"
> Thus, a rhetorical question perhaps.  The "What kind of business" was
> not a query for what *industry* the business is in, but rather in a
> spirit of "what kind of idiot".

By my response, I meant to say that anyone who defends on the power
company, a telecom or a computer vendor is depending on software they
don't control. 

I cannot inspect the software that ensures that power reaches my office,
but I (and millions of other customers) am dependant on that software.

I would conclude that there are very few businesses that do not depend
on software that they have no control over. Closed systems of
non-technological people perhaps.

Cheers,
Brian





From henrik.motakef at web.de  Thu Apr 17 08:45:20 2003
From: henrik.motakef at web.de (Henrik Motakef)
Date: Thu, 17 Apr 2003 14:45:20 +0200
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr>
 <m27k9vfcz3.fsf@mycroft.actrix.gen.nz>
 <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk>
 <3E9EB3E9.1040208@removethis.free.fr>
Message-ID: <87k7dt5m3j.fsf@interim.henrik-motakef.de>

laotseu <bdesth at removethis.free.fr> writes:

> Lispers friends, let's have you're opinion on that point : is Common
> Lisp a functionnal language ?

Common Lisp is whatever you want it to be: You can write
functional-style programs in it if you want, but is certainly not a
"pure" functional language. It supports several other paradigms
equally well, and even more important, it will let you integrate
yet-unkown paradigms when you think that they are useful.

There are really ideas to remove the functional stuff from Python? I
wonder why people frequently think that making things impossible will
improve programming languages.

just-because-it's-flexible-doesn't-mean-it's-perl-ly y'rs
Henrik



From gry at ll.mit.edu  Wed Apr 23 14:26:50 2003
From: gry at ll.mit.edu (george young)
Date: Wed, 23 Apr 2003 14:26:50 -0400
Subject: Exception.__init__ needed in derived class?
References: <20030423140754.1946afea.gry@ll.mit.edu>
Message-ID: <20030423142650.443c044f.gry@ll.mit.edu>

On Wed, 23 Apr 2003 14:07:54 -0400
george young <gry at ll.mit.edu> threw this fish to the penguins:
> [python-2.3a1]
> Does class Exception have a constructor __init__ that needs
> to be called?  I couldn't find this in the docs.  I see many
> uses of inheriting from Exception without invoking
> Exception.__init__(self), but nobody ever *said* that this 
> was safe to leave out...
> 
> Pychecker complains about this, which is what brought it to
> mind.  I can certainly invoke it just to make pychecker 
> happy, but I wondered...

To elaborate, I often want to define a __init__ for my
Exception subclass because I want to give it named data, e.g.:

class BadRun(Exception):
	def __init__(self, name, id):
		self.name=name
		self.id=id
...
	raise BadRun(newname, current_id)
...
	try:
		self.run = Run('implant_test', self.step_id)
	except BadRun, x:
		print 'Run %s, id=%s can not be created',x.name, x.id

Would it be better to do:
class BadRun(Exception): pass
...
	ex = BadRun()
	ex.name='implant_test'; ex.id=self.step_id
	raise ex



From jerf at jerf.org  Fri Apr 25 17:49:36 2003
From: jerf at jerf.org (Jeremy Bowers)
Date: Fri, 25 Apr 2003 21:49:36 GMT
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <slrnbajav2.e5n.ramen@lackingtalent.com>
Message-ID: <pan.2003.04.25.17.49.27.407432@jerf.org>

On Fri, 25 Apr 2003 21:42:25 +0000, Dave Benjamin wrote:
> Speaking as one who experiences memory loss for various reasons, I will
> admit that there are certain memories which I can't seem to conjour up, even
> though I have "learned" them in the past. However, that doesn't mean they're
> gone. Perhaps they just weren't indexed properly. I have yet to learn how to
> do a linear scan on my brain (anyone have a code snippet for that?)

for memory in brain.iter_memories():
	print memory


Duh.

Implementation of the brain object left as an exercise for the reader. I
know it can be done, as I am in possession of such an object right now.



From aleax at aleax.it  Sat Apr 12 11:11:44 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 15:11:44 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <mailman.1050156199.10180.python-list@python.org>
Message-ID: <Q0Wla.4427$LB6.127243@news1.tin.it>

Steven Taschuk wrote:
   ...
> Pickle doesn't use __repr__; it has other, more reliable, ways of
> getting object state.

Correct.

> __repr__ is used by: (1) '%r' % foo and repr(foo); (2) '%s' % foo,
> str(foo), and print foo if __str__ isn't provided; and (3) the
> interactive interpreter, when printing the value of an expression
> entered at the prompt.  (And no doubt, other places that I've
> forgotten.)

IMHO the most maddening use of __repr__ is that done by the 
implementation of __str__ for all built-in container types:

>>> class X:
...   def __str__(self): return 'str'
...   def __repr__(self): return 'rep'
...
>>> print [X()]
[rep]
>>>

There's supposed to be a good reason for this to print [rep],
rather than [str] as common sense would suggest, but I keep
forgetting it (so it's probably nor very compelling;-).


Alex




From ckea25d02 at sneakemail.com  Mon Apr  7 21:28:28 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 13:28:28 +1200
Subject: ANN: Pyrex 0.7
Message-ID: <mailman.1049901963.20761.clpa-moderators@python.org>

Pyrex 0.7 is now available:

    http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/

Main new features:

  * C attributes of extension types can be exposed directly
    to Python code instead of using accessor methods.

  * C functions, etc. can more easily be wrapped with
    Python objects having the same name.

There are also many other improvements and bug fixes -- see the
web page and CHANGES file for more.


What is Pyrex?
--------------

Pyrex is a new language for writing Python extension modules.
It lets you freely mix operations on Python and C data, with
all Python reference counting and error checking handled
automatically.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg





From brian at sweetapp.com  Mon Apr  7 19:33:44 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Mon, 07 Apr 2003 16:33:44 -0700
Subject: How to protect python code ?
In-Reply-To: <XHnka.22636$It5.16795@news2.central.cox.net>
Message-ID: <002301c2fd5e$217564d0$21795418@dell1700>

>> What kind of business *pays* to become dependent on
>> software they don't control?!
>>
 
> Can we say "Microsoft customers"? "Oracle customers"? ... 
> the list is endless!

I'm sure that the original poster was being facetious but...power
customers, telecom customers, computer customers.

Cheers,
Brian





From ryomanga at aol.com  Sun Apr  6 12:47:25 2003
From: ryomanga at aol.com (Ryomanga)
Date: 06 Apr 2003 16:47:25 GMT
Subject: pb py32exe and PIL
References: <brzj4oqc.fsf@python.net>
Message-ID: <20030406124725.26136.00000748@mb-fd.aol.com>

>>   File "Image.pyc", line 960, in open
>> IOError: cannot identify image file
>> 

>You have to import the plugins you need in your script so that
>py2exe knows they are used. You should be able to google for this
>questions and find some answers.
>

I have imorted them my script work perfectly using pyhon myscript.py, its only
with the exe it's not working. Maybe you think py2exe doesn't import the PIL
library it's false there is the Imge.pyd file with the exe. Or maybe i don't
understand what you want to say..
    Guilh?m



From lekishvili at python.qartu.com  Thu Apr 10 13:13:36 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 10 Apr 2003 10:13:36 -0700
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <yfs7ka3iom2.fsf@black132.ex.ac.uk>
Message-ID: <a3b5b608.0304100913.7c6f994b@posting.google.com>

Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfs7ka3iom2.fsf at black132.ex.ac.uk>...
> s.thuriez at laposte.net (sebastien) writes:
> 
> > Hi,
> > 
> > Is there any PCA analysis tools for python ?
> 
> What is the analysis tool supposed to do?
> 
> Maybe this will do what you want, once you (downloaded and installed Numeric):
> 
> # Warning: hackish and not properly tested ripped out bit of code ahead
> # so no guarantees whatsoever
> # Anyway, it should at lesat sort of give you the idea
> # try pca(X); if that doesn't do what you want try pca(t(X))
> 
> from Numeric import take, dot, shape, argsort, where, sqrt, transpose as t
> from LinearAlgebra import eigenvectors
> 
> def pca(M):
>     "Perform PCA on M, return eigenvectors and eigenvalues, sorted."
>     T, N = shape(M)
>     # if there are less rows T than columns N, use
>     # snapshot method
>     if T < N:
>         C = dot(M, t(M))
>         evals, evecsC = eigenvectors(C)
>         # HACK: make sure evals are all positive
>         evals = where(evals < 0, 0, evals)
>         evecs = 1./sqrt(evals) * dot(t(M), t(evecsC))
>     else:
>         # calculate covariance matrix
>         K = 1./T * dot(t(M), M)
>         evals, evecs = eigenvectors(K)
>     # sort the eigenvalues and eigenvectors, decending order
>     order = (argsort(evals)[::-1])
>     evecs = take(evecs, order, 1)
>     evals = take(evals, order)
>     return evals, t(evecs)
> 
> 

Nice, but if we have big matrices, than it might be discouraging to
comput all of the PCs once we shall drop most of them. Implementation
of NIPALS with Numeric is 10 min work.




> You can download Numeric and use it to compute the eigenvalues and
> eigenvectors of an array.
> 
> 
> > If it does, do you have any idea on how well it would scale ?
> 
> It should scale fine. If you experience speed problems, configure Numeric 23
> with ATLAS support (you have to install ATLAS and LAPACK first, of course).
> For large matrices, this should be *much* faster than handwritten C code that
> doesn't use ATLAS.
> 
> > 
> > I have already seen PyClimate (but it is not available for Windows
> > which will be one of the target). Is there some LAPACK like packages ?
> 
> Yes, Numeric and scipy. (www.numpy.org, www.scipy.org, I should think)
> 
> 'as



From jkraska at san.rr.com  Thu Apr 17 11:46:11 2003
From: jkraska at san.rr.com (Courageous)
Date: Thu, 17 Apr 2003 15:46:11 GMT
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com> <hr9s9v8pgqr154h15lsjbk1pjs5u5n6n65@4ax.com> <Inyna.27603$LB6.635831@news1.tin.it>
Message-ID: <5qit9v4pcrn4haj40gfi61egrg1pi0cg7d@4ax.com>

>Having MTOWTDI ain't Pythonic.  The ability to write either:
>
>    if finished: break
>
>OR quite equivalently
>
>    if finished:
>        break
>
>is therefore a wart.

Hrm. Conceputalization? I consider these two equivalent. I suppose
it's bias. Years of using languages where WSDM (white spaces don't
matter), allows me to elide away some under some circumstances. :)

BTW, there's a thing I do in some programming environments, that
I personally really like to do, but I see few other people do. I'll
show you:

	if (alpha)       { DoSomething();        return x; }
    if (beta)        { DoSomethingElse();    return y; }
    if (delta)       { DoOtherThing();       return z; }

I like to use *columns*, in other words. I find this to be neat and
readible, and considerably superior to what would otherwise pass as
the defacto alternative.

I think you'll see how the example applies to the thread's subject...

C//




From R.Brodie at rl.ac.uk  Mon Apr 14 09:23:03 2003
From: R.Brodie at rl.ac.uk (Richard Brodie)
Date: Mon, 14 Apr 2003 14:23:03 +0100
Subject: Cryptography (was Re: How to protect python code ?)
References: <3E8BE455.4000904@novagrid.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com> <OIqdnV1p3aljSw-jXTWcpg@comcast.com> <b7bvft$os4$1@panix1.panix.com>
Message-ID: <b7ecno$n7m@newton.cc.rl.ac.uk>

"Aahz" <aahz at pythoncraft.com> wrote in message news:b7bvft$os4$1 at panix1.panix.com...

> Not that I'm disagreeing with your main point (about protecting Python
> code), but cryptographic systems that are developed privately are much
> easier to hack (historically, at least) than ones developed in public.

In general, yes. Whether experts working in classified establishments can't
do better than the public state of the art is much harder to quantify.
Consider the history of PKCS, for example.





From zhitenev at cs.vsu.ru  Mon Apr 14 08:24:06 2003
From: zhitenev at cs.vsu.ru (Lexy Zhitenev)
Date: Mon, 14 Apr 2003 16:24:06 +0400
Subject: Using Jython to process HTML
References: <662cede5.0304130905.6938703d@posting.google.com>
Message-ID: <b7e996$8vq$1@serv.vrn.ru>

"Tzahi" <tzahij at hotmail.com> wrote in message:
news:662cede5.0304130905.6938703d at posting.google.com...

> I want to use Jython for the HTML processing power, so that I can
> access other project and Java system classes. I hope this way I can
> read pointers to documents from standard Queues, created by other
> parts of the system, and process them with Python power.
>
> I have never used Jython. Can you please tell me if this design make
> any sense?
I think you won't have any trouble using Jython. The only limitation can be
if your programs will run on the client side: they'll have to download
Jython jar or something else. This wouldn't be acceptable. This is very good
on server side or standalone.

> Do other people use Jython I real-life projects?
Though the Jython project is somewhat frozen, I can see many people using
it, at least in c.l.py. Don't worry to use it if you need to. If your
solution will work and it will be easier to implement/support than Java,
then just do it.

Regards, Lexy.





From scott_list at mischko.com  Mon Apr 14 16:48:18 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Mon, 14 Apr 2003 13:48:18 -0700
Subject: Seperating HTML content from logic ( was: Re: Inserting Python commands into HTML documents)
In-Reply-To: <xrtma.10655$LB6.301836@news1.tin.it>
References: <zpqdne0Kztcb0AejXTWcow@comcast.com> <xrtma.10655$LB6.301836@news1.tin.it>
Message-ID: <200304141348.18694.scott_list@mischko.com>

On Monday 14 April 2003 00:29, Alex Martelli wrote:

> Sure, there are many ways in which you can choose to use this
> debatable "let's mix logic right inside presentation so we can't
> easily separate them any more" approach (perhaps excusable when
> one's in a deuced hurry, if one's used to it and thus can deploy
> it fastest).  

I've been dealing with web pages for a while, using the "embed the scripting 
language into the HTML" model.  I'd like to find some good docs on how to 
implement the "lets seperate the logic and presentation" model.

I'd also like to know which Python toolkits/frameworks that are aimed at 
implementing this model.

Any pointers?

Scott




From rs55862 at NOSPAMyahoo.com  Wed Apr  9 03:54:02 2003
From: rs55862 at NOSPAMyahoo.com (Rami Saarinen)
Date: Wed, 09 Apr 2003 10:54:02 +0300
Subject: Making python module using Borland C++ 5
In-Reply-To: <b6ujd4$pe5$1@news.hccnet.nl>
References: <b6u7fp$d2c$1@news.cc.tut.fi> <b6ujd4$pe5$1@news.hccnet.nl>
Message-ID: <b70jiq$men$1@news.cc.tut.fi>

Anton Vredegoor wrote:
>>ImportError: dynamic module does not define init function (initpyclips)
> 
> 
> Please have a look at:
> 
> http://www.cyberus.ca/~g_will/pyExtenDL.shtml
> 
> Anton.
> 

I finally managed to build the module using bcpp. Thank you Anton and 
Anders. It was all about including the EXPORT initpyclips=_initpyclips 
in a def file (see the link above).

Thanks again for the help.

--
Rami Saarinen




From hschilling at grc.nasa.gov  Thu Apr  3 11:50:02 2003
From: hschilling at grc.nasa.gov (Herbert Schilling)
Date: 3 Apr 2003 08:50:02 -0800
Subject: Motif window with interactive Python shell
Message-ID: <77fbed4e.0304030850.7ed2188c@posting.google.com>

Hi,
 
I have a Motif application that I extended and embedded with Python. 

I would really like to have a Motif window where I could interact with
the application via a Python shell.  I looked around for ideas on how
to do that but I didn't see much. Most of the ideas I saw involved
using GUI toolkits other than Motif. We really don't want to include
another GUI toolkit.

Any ideas? 

 Thanks!



From markulus00 at yahoo.com  Wed Apr  9 01:42:33 2003
From: markulus00 at yahoo.com (Mark Erikson)
Date: 8 Apr 2003 22:42:33 -0700
Subject: SSL/HTTPS POST proxy login issues
Message-ID: <4d3df955.0304082142.33446cb@posting.google.com>

Ooookay.  I've been wrestling with a problem for the last several
weeks, and if someone out there can help me solve it, I'll be
extremely grateful.

My school uses Novell Bordermanager as a proxy to the web.  On
school-owned computers, the Novell Proxy Key runs and notifies the
proxy that you're logged into the network and can get outside.  On
student-owned computers, an initial attempt to go to a website via the
proxy brings up a proxy login page, in which you enter the username
and password you use to log onto the Novell network.  This page is SSL
encrypted, and performs an SSL-encrypted POST to the proxy with the
username, password, destination address, and a couple of other fields.

Since this login procedure must be performed on an initial request as
well as after a half-hour of inactivity, all of which gets annoying,
I've been trying to work on a small utility that would run on startup,
perform the POST necessary to login to the proxy, then sit in the
Windows tray and do a simple GET of, say, Google every 20 minutes or
so to keep the login alive.  A fairly simple project, right?

Unfortunately, I have been COMPLETELY unable to execute a successful
SSL connection to any of our four proxy servers using Python.  Using
Python 2.2, I've tried the built-in httplib.HTTPSConnection, as well
as the one supplied by M2Crypto 0.09.

Here's the tail end of a typical traceback, along with the code that
produced it:

from M2Crypto import SSL, httpslib
ctx = SSL.Context()
c = httpslib.HTTPSConnection(SERVER_IP, 443, ssl_context=ctx)
params = urllib.urlencode({ various fields from the login page with
proper values } )
headers = { a whole bunch of stuff I copied from what Phoenix sends to
a server }
c.request('POST', '/BM-Login/auth-cup', params, headers)

File "C:\Python22\Lib\site-packages\M2Crypto\SSL\Connection.py", line
96, in connect_ssl
    return m2.ssl_connect(self.ssl)
SSLError: (0, 'Error')

The proxy servers have a certificate signed by the school, which I
believe I can export from Internet Explorer in PKCS7 format (not that
I'm completely sure about all that involves, I just know I can do it).

Meanwhile, a friend of mine running a Linux box has been using a
simplified version of the login page with the Javascript command
"loginform.submit()" as part of the body to do a sort of auto-login.

Sooo... I guess I have a few questions.

1) Has anyone ever tried to create something similar?
2) Do I need some sort of a client certificate/key in order to do an
SSL connection to the server?
3) Do I need to do some sort of verification of the server's
certificate, since it's self-signed?
4) Does anyone have ANY other ideas as to how to get this thing to
work?

Anyway, I hope the explanation is detailed enough, and I'll try to
provide more details if necessary.  Thanks in advance for any help!

Mark Erikson



From mis6 at pitt.edu  Fri Apr 11 12:12:35 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 11 Apr 2003 09:12:35 -0700
Subject: a module who knows its caller
Message-ID: <2259b0e2.0304110812.4eea37de@posting.google.com>

I have a module (say m.py) that can be called by different scripts (say a.py 
and b.py). I want m.py to be able to tell if it has been called by a.py or
by b.py.

What I do now is to pass the global name __file__ to a function m.beencalledby:
something like

# a.py
import m
m.beencalledby(__file__)  # __file__ is a.py


# b.py
import m
m.beencalledby(__file__)  # __file__ is b.py

# m.py

def beencalledby(fname): print 'I have been called by',fname

it works, but I wonder if there is a more elegant solution.

TIA,

                                   Michele



From igor at tyumbit.ru  Mon Apr 21 09:02:14 2003
From: igor at tyumbit.ru (Igor Prischepoff)
Date: Mon, 21 Apr 2003 19:02:14 +0600
Subject: wxBoxSizers in window and two embedded panels.
Message-ID: <4933657516.20030421190214@tyumbit.ru>

Hi!
 Sorry for my english, i'm not native english speaking man.
 Recently switched to python and it's wonderful wxPython toolkit.
 Having trouble with wxBoxSizer.
 I want this heirarchy

      wxWindow (a must! i've don't need to attach panels to wxFrame directly.)
       /    \
    Panel2  Panel1
    
Panel1 and panel2 must occupy all available screen space on window1.
Trouble is: they don't.
    

 ------------------------------  -----
 !  panel2                     !      !
 !-----------------------------!      !
 !  panel1                     !      !
 !                             !      !
 ------------------------------!      !
 !                 window1            !
 !------------------------------------
                                      
# Here is Boa generated app.(slightly changed.)
from wxPython.wx import *

def create(parent):
    return wxFrame1(parent)

[wxID_WXFRAME1, wxID_WXFRAME1PANEL1, wxID_WXFRAME1PANEL2, 
 wxID_WXFRAME1WINDOW1, 
] = map(lambda _init_ctrls: wxNewId(), range(4))

class wxFrame1(wxFrame):

    def __init__(self, parent):
        # generated method, don't edit
        wxFrame.__init__(self, id=wxID_WXFRAME1, name='', parent=parent,
              pos=wxPoint(252, 204), size=wxSize(768, 558),
              style=wxDEFAULT_FRAME_STYLE, title='wxFrame1')
        self.window1 = wxWindow(id=wxID_WXFRAME1WINDOW1, name='window1',
              parent=self, pos=wxPoint(0, 0), size=wxSize(760, 531), style=wxEXPAND)

        self.panel1 = wxPanel(id=wxID_WXFRAME1PANEL1, name='panel1',
              parent=self.window1, pos=wxPoint(16, 8), size=wxSize(536, 32),
              style=wxEXPAND)

        self.panel2 = wxPanel(id=wxID_WXFRAME1PANEL2, name='panel2',
              parent=self.window1, pos=wxPoint(24, 56), size=wxSize(528, 376),
              style=0)
# Just to see where which colour.
        self.window1.SetBackgroundColour("BLACK")
        self.panel2.SetBackgroundColour("BLUE")
        self.panel1.SetBackgroundColour("RED")
#       
# Am i miss something about Sizers?
# 
        box=wxBoxSizer(wxVERTICAL)
# put into sizer panel2 and panel1 with proportion factor 1:50
        box.Add(self.panel2,1,wxEXPAND)
        box.Add(self.panel1,50,wxEXPAND)
# Attach sizer to window1
        self.window1.SetSizer(box)
        self.window1.SetAutoLayout(true)
        self.window1.Layout()

class BoaApp(wxApp):
    def OnInit(self):
        wxInitAllImageHandlers()
        self.main = create(None)
        self.SetTopWindow(self.main)
        self.main.Show();
        return True


if __name__ == '__main__':
    application = BoaApp(0)
    application.MainLoop()
# end of code
# 
When running this app i've got initial sizes of panel1 and panel2
ok, but when resizing main window with mouse,my panels not resize.
In short - there must not be BLACK colour in my window.

I've installed sizer and attached it to window1.
Put there 2 panels, panel1 and panel2.
How to force my panels to rezise properly, i.e. to occupy all available
space in window1?
Sorry if it's wrong list, i've know there is wxPython list but can not
subscribe to it. There are trouble with their subscriber software.

Thanks in advance.

P.S. win2k + python 2.2.2. + wxPython - last stable release.

                     mailto:igor at tyumbit.ru





From frithiof.jensen at remove_this.jensen.tdcadsl.dk  Tue Apr 15 06:34:19 2003
From: frithiof.jensen at remove_this.jensen.tdcadsl.dk (Frithiof Andreas Jensen)
Date: Tue, 15 Apr 2003 12:34:19 +0200
Subject: Sv: Python compared toJava
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
Message-ID: <3e9bdf88$0$24726$edfadb0f@dread14.news.tele.dk>

Blaktyger <blaktyger at hotmail.com> skrev i en
nyhedsmeddelelse:c9e6ac92.0304100824.eec0f5a at posting.google.com...
> Hi!
>
> I recently read that the Python interpreter is like the Java Virtual
Machine.
> Is that true?

Not quite - Python actually *delivers* on "write once, run everywhere" ;-)





From jtiller at sjm.com  Thu Apr  3 19:45:12 2003
From: jtiller at sjm.com (Jason Tiller)
Date: Thu, 3 Apr 2003 16:45:12 -0800 (Pacific Standard Time)
Subject: split encloser
In-Reply-To: <435d40e0.0304031530.1c197742@posting.google.com>
Message-ID: <Pine.WNT.4.44.0304031631340.2016-100000@WS-USSV-50244>

Hi, "Aussie", :)

On 3 Apr 2003 aussie2010 at yahoo.com wrote:

> string.split() takes a delimiter and works fine as long as the
> delimiter isn't part of the data fields. But frequently they are.
> e.g. 'John Doe,135 South Main St.,#122, Springfield, Iowa' or
>       ' so long goodbye see ya'

> Because the fields can contain the delimiter in some cases, an
> encloser is usually used (typically "") to handle those fields.

> The above strings would be written:
>
> 'John Doe,"135 South Main St., #122", Springfield, Iowa'
>    and
> '"so long" goodbye "see ya"'

> I don't understand regular expressions but I was wondering if anyone
> that did knew of a way to get re.split() to handle "enclosers" as
> used above.

Hmm.  I am not yet a knowledgable user of Python's regex features, but
I *do* know Perl's pretty well.  With Perl, you might split up your
fields like this:

my $data = "Tiller, Jason, \"39177 Sundale Dr., Apt. 12\", Fremont, CA";

while( $data =~ /\G(?:\"(.+?)\",? ?)|\G(.+?), ?/g ) {
   my $field = $1 || $2;
   print "$field\n";
}

This says (essentially):
   Starting from the previous match point,
      Look for either:
         A series of characters surrounded by the "enclosers" or
         A series of characters up to the split character
      If found:
         Set our field to which of the matches succeeded
         Set the previous match point
      Else:
         Exit

The RE breaks down like this:
   /\G        # Anchored to the previous match location:
    (?:       # Don't capture the following group.
       \"     # Find an encloser
       (.+?)  # followed by a set of any characters (non-greedy)
       \"     # bounded by another encloser
       ,? ?)  # and potentially a split character and a space
    |         # OR
    \G        # Again achored at the previous match location
    (.+?)     # Find a set of any characters (non-greedy)
    , ?       # That end with the split character and possibly a space
   /gx        # Do this until the match fails and ignore whitespace in
              # the pattern

>From what I was reading in a modern edition (2.0) of "Programming
Python," Python has adopted many of Perl's RE enhancements
(non-capturing subexpressions, non-greedy quantifiers, additional
anchors, etc.), so the pattern may work in Python without much effort.
I dunno.

Good luck!  I hope this pattern at least gives you a starting point
for implementing the split() in Python.

---Jason
Sonos Handbell Ensemble





From garycramblitt_NOSPAM at comcast.net  Sat Apr  5 13:21:41 2003
From: garycramblitt_NOSPAM at comcast.net (Gary Cramblitt)
Date: Sat, 05 Apr 2003 13:21:41 -0500
Subject: [Survey] Yet Another Open Source RSS News Reader?
Message-ID: <khudnWYcCKiwghKjXTWcpg@comcast.com>

Because this message is somewhat off-topic, please do not respond to this
newsgroup.  Instead, send replies directly to garycramblitt at comcast.net.
(See below for explanation why I chose to post this survey here.)

This purpose of this posting is to assess the level of interest in a
proposed open source project for an RSS news reader I have written.  For
the time being, I am calling this application News Kiosk.  If you have an
interest in News Kiosk, either as a user, or as a possible developer,
please respond to the e-mail address above.

News Kiosk is (would be) a free open source RSS news reader written in
Python 2.2.1 and wxPython 2.4.0.7.  While the program is brand new, it is
already quite functional, and offers the following features:

1.  Reads any RSS format using the Ultra-liberal RSS parser by Mark
Pilgrim. (http://diveintomark.org/projects/rss_parser/)

2.  Retrieves and parses lists of feeds from syndication services such as
Syndic8 and NewsIsFree using the OPML parser by Fredrik Lundh.
(http://www.effbot.org/zone/effnews.htm)

3.  Fast and lightweight.  Configurable asynchronous retrieval from
multiple feeds.

4.  Architecture based on Principle of Separation.  See _The Art of Unix
Programming_ by Eric Steven Raymond.
(http://www.catb.org/~esr/writings/taoup/html/)  The GUI is separate from
Fetch Agents.  Fetch Agents are non-GUI subprocesses responsible for
retrieving and parsing news resources (RSS feeds).  The GUI is responsible
for displaying the results and interacting with the user. This
architecture also permits News Kiosk to be extended by adding additional
Fetch Agents designed to retrieve and parse information of almost any
type.

Although it is written using cross-platform libraries, this application is
targeted primarily towards the Linux platform.  In theory, it should run
under Windows, BSD, and Max OS X, but I have not tried it under these
platforms.  My sense is that these platforms already have many fine RSS
news readers available anyway.

Why another open source RSS news reader?  I wanted an RSS news reader for
my RedHat 8 system.  I began searching the web for a free open source
reader.  I found several possible packages but rejected each of them for a
variety of reasons as outlined below:

1.  NewsMonster (www.newsmonster.org).  This program appears to have many
fine capabilities.  Perhaps too many leading to bloat?  While it is open
source, it requires the Java Runtime Environment, which is itself not open
source.  I am trying not to install *any* software on my system for which
I cannot obtain the source code.  If I wanted closed-source solutions, I'd
go back to Windows.

2.  EffBot by Fredrik Lundh (http://www.effbot.org/zone/effnews.htm). A
basic news reader, it currently does not offer lists of feeds from
syndication services, although Fredrik clearly intends to add that
capability, since I'm using his OPML parser.  Uses Python and Tkinter. Has
some trouble with non-standard RSS feeds.  Fredrik seems to primarily
target the Windows platform.

3.  nntp/rss (http://www.sourceforge.net/projects/nntprss).  An
interesting approach that parses RSS feeds and makes them available to any
NNTP (Usenet) news reader application.  Uses Mark Pilgrim's Ultra-liberal
RSS Parser.  Does not seem to support syndication lists. Written in Java,
so I have the same objection to this one as NewsMonster.

4.  Straw (www.nongnu.org/straw/).  Try as I might, I simply could not get
this program to run on my system.  It has quite a number of dependencies. 
It appears that it has become a victim of GTK library creep.

5.  Amphetadesk (http://www.disobey.com/amphetadesk/).  An open source
project coded in Perl and XML.  Amphetadesk runs in a web browser and
therein lies the problem, at least for me.  The interface just isn't quite
what I'm looking for.  There are a number of other applications that also
run in a web browser.  See http://blogspace.com/rss/readers.

So after looking at these applications and not being fully satisfied with
any of them, I decided to write my own RSS news reader.

Perhaps it would be appropriate to say what News Kiosk is *not*:

1.  It does not keep a database of past articles.  No permanent storage.

2.  No blogging capability.  News Kiosk does not permit you to post to
your personal weblog.

3.  No built-in web browser.  News Kiosk displays a simple summary of each
article as provided by the RSS source, then permits user to read the full
article by launching the link in their favorite browser. This was done to
avoid security issues (cookies, javascript, spyware, etc.)

4.  Not designed for any particular Linux desktop.  Because the GUI is
written using wxPython (wxWindows), it does not support KDE and GNOME
themes, skins, etc.  However, the architecture of News Kiosk does permit
the substitution of other GUIs programmer's might want to write.

5.  Not designed for retrieval from huge numbers of feeds.  A few dozen
subscriptions are appropriate.

So what do you think?  Do you have an interest in News Kiosk, either as a
user or a possible developer?  Do you object to "yet another" RSS news
reader and think my objections to the packages above are unjustified?

Please respond to garycramblitt at comcast.net.

BTW, after writing News Kiosk, I considered applying for project space on
SourceForge.net.  I wondered, however, if yet another RSS news reader
would be well-received.  There has been some discussion recently about the
uncontrolled proliferation of "yet another <fill in blank here>" in the
open source world.  (See http://freshmeat.net/articles/view/774/ and
http://slashdot.org/articles/03/04/04/1313203.shtml?tid=117&tid=185). I
searched quite a while and could find no forums or resources where I could
float the idea of News Kiosk.  I sent an e-mail to the admins at
SourceForge.net asking where they thought I should post this survey.  They
suggested posting in the language-specific group for my application. 
Hence, this posting here in comp.language.python.

Thankyou for your time and consideration.

Gary Cramblitt




From alienoid at is.lg.ua  Mon Apr  7 08:16:26 2003
From: alienoid at is.lg.ua (Ruslan Spivak)
Date: Mon, 07 Apr 2003 15:16:26 +0300
Subject: how to convert from unicode to another codepage
Message-ID: <3E916C1A.80808@is.lg.ua>

Hello, python users!

I need to convert string from utf-8 in 'cp1251' codepage.
How to achieve that?

Thanks in advance.
Ruslan





From b.maryniuk at forbis.lt  Fri Apr 25 14:59:42 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Fri, 25 Apr 2003 20:59:42 +0200
Subject: What means exactly "Memory error"?
In-Reply-To: <b8bshn$d00$1@main.gmane.org>
References: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net> <slrnbai52u.8cv.hat@se-46.wpa.wtb.tue.nl> <b8bshn$d00$1@main.gmane.org>
Message-ID: <200304252059.42162.b.maryniuk@forbis.lt>

On Friday 25 April 2003 19:51, Fredrik Lundh wrote:
> an alternative is to use an incremental tree builder, and process the
> subtrees as they arrive.
> here's an example, using my elementtree module:

ElementTree is OK, but I need process HTML, not XML. With XML is no problem at 
all. Or I just something missed? 8-|

-- 
Regards, Bogdan

"I may not understand what I'm installing, but that's not my job. I
just need to click Next, Next, Finish here so I can walk to the next
system and repeat the process"
		-- Anonymous NT Admin





From Olivier.POYEN at clf-dexia.com  Fri Apr 11 07:49:29 2003
From: Olivier.POYEN at clf-dexia.com (POYEN OP Olivier (DCL))
Date: Fri, 11 Apr 2003 13:49:29 +0200
Subject: Getting source code from interpreter-defined function
Message-ID: <8963D6370B323E4AA3241200F0EAF7318C0DA2@FCEXVEXM002.dcl.int.dexwired.net>

> Other solution: given any function, even defined in 
> interpreter, I can get function.func_code.co_code, which is 
> byte compiled code for this function. Perhaps I could use it 
> and decompile it. But how....? 
> 
> 

This one works with decomyler
>>>from decompyler import decompyle
>>>import sys
>>>def foo():print "coole"
>>>pythonVersion='2.2'
>>>decompyle(pythonVersion,foo,sys.sdtout)

print "coole"
#plus other stuff, code dependent

One way to go.
Is this the best way ?  
-------------- next part --------------
-----------------------------------------------------------------

Ce message est confidentiel ; son contenu ne represente en aucun

cas un engagement de la part de Dexia Credit Local ou de Dexia

CLF Banque. Toute publication, utilisation ou diffusion, meme

partielle, doit etre autorisee prealablement par l'emetteur. Si

vous n'etes pas destinataire de ce message, merci d'en avertir

immediatement l'expediteur.



This message is confidential ; its contents do not constitute a

commitment by Dexia Credit Local or Dexia CLF Banque. Any

unauthorised disclosure, use or dissemination, either whole or

partial, is prohibited. If you are not the intended recipient of

the message, please notify the sender immediately.

-----------------------------------------------------------------

Consultez notre site internet www.dexia-clf.fr

La cle d'une gestion optimisee du secteur public local

-----------------------------------------------------------------

From titanrebel at comcast.net  Wed Apr 30 16:36:01 2003
From: titanrebel at comcast.net (kkennedy)
Date: 30 Apr 2003 13:36:01 -0700
Subject: network programming problem??
Message-ID: <14bd1399.0304301236.7fe19be5@posting.google.com>

I have written a small script to retrieve the python.org web page 7
times. I did this to try and provide data for a percieved problem in
another Python program. I ran this script on Windows XP and SuSE Linux
8.2 both running on the same machine (Athlon XP 2000+, 512MB RAM,
D-Link DFE-530TX+ NIC).  Here is the script:
-------------------- START -------------------------
from urllib import urlopen
import time

for x in range(7):
    start = time.time()
    doc = urlopen("http://www.python.org").read()
    end = time.time()
    print "Run %d . %0.2f" % (x, end - start)
--------------------- END --------------------------

Here are the results of running the script:

-------------------- LINUX --------------------------
Run 0 . 16.69
Run 1 . 16.54
Run 2 . 17.04
Run 3 . 16.55
Run 4 . 16.55
Run 5 . 16.54
Run 6 . 16.55

-------------------- WIN XP -------------------------
Run 0 . 2.69
Run 1 . 0.59
Run 2 . 0.61
Run 3 . 0.66
Run 4 . 0.59
Run 5 . 0.61
Run 6 . 0.59


Does anyone have a clue why these results are soooo different.  The
SuSE system doesn't "feel" sluggish.  My first clue was that a Linux
user complained about my program being slow, but none of my Windows
users complained.  My second clue was that I thought I could tell a
pretty big difference while browsing the web on Win XP vs. SuSE.  I
just thought it was my imagination, but I guess not!  This is quite a
difference.

I was under the impression that the TCP/IP stack on Linux was second
to none. Could it be that my NIC has poor driver support?  Any
suggestions would be welcome.

Thanks.

P.S. I have continued to get almost exactly the same results every
time I run the script on each OS, so I absolutely do not believe that
internet traffic is making any difference.



From nas-usenet at arctrix.com  Sun Apr  6 19:02:08 2003
From: nas-usenet at arctrix.com (Neil Schemenauer)
Date: Sun, 06 Apr 2003 23:02:08 GMT
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <e9570f37.0304060944.79564f88@posting.google.com> <jQ_ja.12641$7w2.9369@nwrddc01.gnilink.net> <mailman.1049659211.32170.python-list@python.org>
Message-ID: <Ql2ka.16821$aQ3.12700@nwrddc02.gnilink.net>

Chuck Swiger <cswiger at mac.com> wrote:
> Do your queries against FedEx's system need to be atomic, and in what 
> ways?

They should be but I don't think they actually are.  I'm getting
out of my depth here because I'm not an expert on the FedEx API.

> If you want is to make multiple outstanding transactions
> against a server asyncronously using an event-driven system
> without blocking

We don't need to do that since we are currently not using an
event driven model.  Like I said previously, making the system
work using an event driven model is possible but would require
making changes.

> I certainly agree with your general notion: programming
> paradigms may be well-suited for some tasks and not well-suited
> for others.  Perhaps it's not clear to me what else you had in
> mind: non-blocking I/O like select() or poll()?  What kind of
> SMP dispatch system are you using now?

We using a "pre-fork" model, similar to what Apache 1.3 uses.
There are multiple worker processes that receive requests from a
manager process.  The application interfaces with the web server
using mod_scgi.  Most requests are handled quicky so one process
handles almost all requests.  The other processes are there for
CPU intensive requests or for requests that block on other
resources.

> Network server architectures tend to be event driven in design
> because most flavors of network IPC are quantized into discrete
> network packets by the lower levels of the system;

Maybe.  I think a more important reason is that most network
servers are network IO bound.  For a static HTML file server or
FTP server, event driven is the way to go, IMHO.

  Neil



From skip at pobox.com  Wed Apr  2 15:42:13 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 2 Apr 2003 14:42:13 -0600
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
In-Reply-To: <3E8B48E0.C786D192@engcorp.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
        <b68jti$2mj66$1@ID-169208.news.dfncis.de>
        <RbCia.6782$9I.555309@news20.bellglobal.com>
        <mailman.1049298845.7655.python-list@python.org>
        <3E8B4094.F61515B3@engcorp.com>
        <ZpHia.23$Sd.4@fed1read01>
        <3E8B48E0.C786D192@engcorp.com>
Message-ID: <16011.19237.373550.57379@montanaro.dyndns.org>

    >> range = my_custom_range_function
    >> 
    >> would be OK. While,
    >> 
    >> some_random_module.range = my_custom_range_function
    >> 
    >> would not.

    Peter> Hmm.... that is exactly what is needed to make this effective in
    Peter> unit testing.  One imports the module under test, then fakes the
    Peter> open method in that other module:
    ...
    Peter> I'm not sure what other approach could be as simple, elegant, and
    Peter> effective and I worry about losing it.  Surely there are many
    Peter> other people doing similar things who would be hurt by this.

How about calling new.module() to create a new module object from the one
you want to test which has suitably munged globals?

Skip




From a.schmolck at gmx.net  Thu Apr 10 17:46:05 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 10 Apr 2003 22:46:05 +0100
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <7xwui295xq.fsf@ruckus.brouhaha.com> <yfsel4ah3oc.fsf@black132.ex.ac.uk> <mailman.1049987720.5681.python-list@python.org>
Message-ID: <yfs1y0agh5u.fsf@black132.ex.ac.uk>

Jim Meyer <jmeyer at pdi.com> writes:

> On Thu, 2003-04-10 at 06:39, Alexander Schmolck wrote:
> > I'd recommend against using __slots__ (unless really needed for optimization),
> > because it cripples the usefulness and generality of your classes to your and
> > also *to other people's* code (weakrefs, pickling, reflection -- all
> > compromised). The first two you could fix (but are unlikely to, unless you
> > *yourself* need to use weakrefs or pickling of instances of those classes),
> > the last one you can't.
> 
> I'm feeling very slow this morning; could you give a few explicit
> examples of how __slots__ interferes with weakrefs, pickling, and
> reflection?

class Normal(object): pass
class Normal2(object): pass    
class A(object):
    __slots__ = 'slot_a slot_b'.split()
    
class B(object):
    __slots__ = 'slot_a slot_b'.split()

>>> n = Normal()
>>> vars(n)
>>> n.__class__ = Normal2
>>> n
<__main__.Normal2 object at 0x8c7df6c>
>>> weakref.ref(n)
<weakref at 0x8b89154; to 'Normal2' at 0x8c7df6c>


now let's try with __slots__

{'slot_a': 23}
>>> a = A()
>>> vars(a)
TypeError: vars() argument must have __dict__ attribute
>>> a.__class__ = B
TypeError: __class__ assignment: 'B' object layout differs from 'A'
>>> weakref.ref(a)
TypeError: cannot create weak reference to 'A' object

(same for pickling)

maybe we *could* find out about a by looking at a.__slots__?

class A(object):
    __slots__ = 'slot_a slot_b'.split()
A.__slots__ = 'a b'.split()
>>> a = A()
>>> a.__slots__
['a', 'b']
>>> a.a = 1
AttributeError: 'A' object has no attribute 'a'
>>> a.slot_a = 1

Fun, isn't it?

And there is yet more fun:


"You cannot use a class attribute to define a default value for an instance
 variable defined by __slots__."


"If a class defines a slot that's also defined in a base class, the instance
 variable defined by the base class slot is inaccessible (except by retrieving
 its descriptor directly from the base class; this could be used to rename it).
 Doing this renders the meaning of your program undefined"

The real purpose is apparently efficiency but if so, it is remarkably ill
conceived. The main person who should have to pay for efficiency features is
the one who writes a particular piece of code that uses them, not others. And
if he imposes a burden on others, it should be quite clear to him and not too
onerous a burden either.

With __slots__ none of this holds true. I bet most people will want to use
__slots__ to catch typos (which seems attractive enough), as did Paul Rubin,
presumably without realizing at all the ill effects of using __slots__. So
mostly they won't bother making their classes pickleable or weakrefable. In
consequence you'll have plenty of code around that needlessly breaks, cripples
or at the very least complicates, other code, largely because people blithly
use some dubious "feature", most likely for the wrong reasons and without
understanding the consequences.

Fundamental reflective capabilities, simplicity and cleanness of the language
are all compromised, just for some silly efficiency gain.


'as




From tjreedy at udel.edu  Tue Apr  1 16:49:14 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 1 Apr 2003 16:49:14 -0500
Subject: I need a lot of help!
References: <31ef0f43.0304011336.1d2777ac@posting.google.com>
Message-ID: <f_2cnTUxwtvYlRejXTWcqQ@comcast.com>

"Jewlz408" <Jewlz408 at aol.com> wrote in message
news:31ef0f43.0304011336.1d2777ac at posting.google.com...
> I need a lot of help
> What do you know about Arrays
> How many storage Locations will be allocated for each of the given
arrays?
> 1) Dim A(1 to 10)
> 2) Dim B(-3 to 5)
> 3) Dim C$( 12)
> 4) Dim D( 1 to 4, -2 to 2)
> 5) Dim E(4,6)
>
> And What does the following programs do with its output?
>
> Dim Table( 1 to 5, 1 to 3)
> For I = 1 to 5
> For J = 1 to 3
> Let Table(I,J) = 1
> Next J
> Next I
> End
>
> What does the program do?
> And what is the output?
>
>
> Thankyou so much for your time
> Please help me!

You appear to be programming in some version of Basic.  This is the
newsgroup for Python, which is quite different.  You need to ask your
questions elsewhere.

TJR





From jmdeschamps at cvm.qc.ca  Mon Apr  7 09:22:02 2003
From: jmdeschamps at cvm.qc.ca (jmdeschamps)
Date: 7 Apr 2003 06:22:02 -0700
Subject: cgi form handler question
References: <3e90ff67$0$13167$5a62ac22@freenews.iinet.net.au> <mailman.1049691790.26318.python-list@python.org> <3E9158CA.C53791E5@engcorp.com>
Message-ID: <3d06fae9.0304070522.2f9e7059@posting.google.com>

Peter Hansen <peter at engcorp.com> wrote in message news:<3E9158CA.C53791E5 at engcorp.com>...
> Ian Bicking wrote:
> > 
> > > the user forgets to input his email and age
> > > how would i write a for loop to output:
> > >
> > > email
> > > age
> >
> > You'd have to define the fields you required, like:
> > 
> > requiredFields = ['name', 'email', 'address', 'age', 'salary']
> > for field in requiredFields:
> >     if not form.has_key(field):
> >         print field
> > 
> > Or you might be able to do:
> > 
> > for name in form.keys():
> >     if not form[name].value:
> >         print name
> 
> The latter suggestion won't usually work, as the data sent
> from the browser does not even contain the names of the fields
> which are left empty. 
> 
> -Peter

...left empty, -unless you use the form
cgi.FieldStorage(keep_blank_values = 1) to get your form info as in
http://www.python.org/doc/current/lib/node297.html

and as someone else might have said, 'Happy CGI-ing',

Jean-Marc



From xxdigitalhellxx at hotmail.com  Mon Apr 14 11:07:58 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Mon, 14 Apr 2003 09:07:58 -0600
Subject: Python and Schools
In-Reply-To: <3e9813e4_2@127.0.0.1>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <3e96b088_2@127.0.0.1> <xPzla.367$xm2.20@news02.roc.ny.frontiernet.net> <mailman.1050070881.6673.python-list@python.org> <3e9813e4_2@127.0.0.1>
Message-ID: <b7eiv9$9e0$1@heckle.cc.uregina.ca>

Jason Smith wrote:
> Bryan L. Fordham wrote:
>>It would make perfect sense for a teacher to require you to use
>>something more than list.sort() if you're studying sorting... but after
>>that, why make you write your own sorting routines if what's provided
>>works just fine?
> 
> I agree.
> 
> Still, what's wrong with this?
> 
>   a.the_sort_that_i_learned_last_week()
> 
> Anyway, this decision is all up to the teacher of course.  So then we see
> that Python gives a teacher flexibility depending on the focus of study. 
> And that sounds like another big plus to me.

Yeah. That's very interesting. For example, one week you could teach 
your students something that involves/requires sorting (can't think of 
any decent examples off the top of my head) and not actually teach them 
sorting until the next week.

-- 
Ryan




From news at titanic.co.uk  Sun Apr  6 06:28:36 2003
From: news at titanic.co.uk (Simon Faulkner)
Date: Sun, 06 Apr 2003 11:28:36 +0100
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com> <3E8F8C10.A7F921CF@engcorp.com>
Message-ID: <m6009vst7snfmo645pd9h8clbgio2dp0lp@4ax.com>

Thanks for all the help on this chaps,

The file has changed and I am now importing 25,000 15 digit numbers!

I am compiling a report as I go on each number, wether is is duped
within this batch and then looking it up in a MySQL database.

It seems to slow down as it goes and I guess that this is the overhead
of checking each number against the dictionary?

Simon



From news at nospam.eml.cc  Wed Apr 30 23:03:14 2003
From: news at nospam.eml.cc (Mahesh Padmanabhan)
Date: Thu, 01 May 2003 03:03:14 GMT
Subject: test (please ignore)
Message-ID: <Xns936DD680B820Anewsnospamemlcc@207.217.77.22>

test 



From pan-newsreader at thomas-guettler.de  Wed Apr  9 10:59:36 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 09 Apr 2003 16:59:36 +0200
Subject: Adding item in front of a list
References: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de> <mailman.1049892916.24803.python-list@python.org>
Message-ID: <pan.2003.04.09.16.59.35.949871.811@thomas-guettler.de>

On Wed, 09 Apr 2003 14:54:22 +0200, Skip Montanaro wrote:

>     Thomas> I think it would be nice if
>     Thomas>  l=[2, 3, 4]
>     Thomas>  l.insert(1)
> 
>     Thomas> would insert in front of the list, as opposed to append()
>     which Thomas> adds new elements at the end.
> 
> What about
> 
>     >>> lst = [2,3,4]
>     >>> lst.insert(0, 1)
>     >>> lst
>     [1, 2, 3, 4]

Hi,

insert(0, 1) behaves like insert(-1, 1). It is better, but
insert(1) would be best. 

This is not about the functionality, but about the beauty of python.

Would be nice if handling lists like queues and stacks would be a bit
better. 

I know it is not important, I just wanted to know if I have overlooked
something and would you think about it.

 thomas




-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From donn at drizzle.com  Thu Apr 24 00:27:09 2003
From: donn at drizzle.com (Donn Cave)
Date: Thu, 24 Apr 2003 04:27:09 -0000
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <f6a31cb8.0304220810.547f5acd@posting.google.com> <oRKdnYDxiYyb5TijXTWcpA@comcast.com> <1051075309.87191@yasure> <84d6jdbmpq.fsf@boost-consulting.com>
Message-ID: <1051158427.693034@yasure>

Quoth David Abrahams <dave at boost-consulting.com>:
| "Donn Cave" <donn at drizzle.com> writes:
[... re class whose __repr__ can break on bad data ...]
|> So here's this object.  You can't repr() it, including implicitly
|> from the interpreter prompt - you get a traceback instead.  Of course
|> that also goes for the list object that contains this object.  To
|> me, this is not right, and maybe it ought to be another point to
|> consider if we're reconsidering the documentation for __repr__.
|> Wish I could specify this better, but I believe there's a minimalist
|> way to think about implementing __repr__ that will tend to make it
|> less interesting but more reliable.
|
| One idea might be:
|
|     # pseudocode
|     class object:
|          def __repr__(self):
|             if hasattr(self, '__getinitargs__'):
|                 return (
|                     repr(self.__class__)
|                   + repr(self.__getinitargs__())
|                 )
|             else:
|                 return # whatever we do today
|
| I know that would only handle a small class of types, since most
| aren't picklable, but it might encourage people to implement
| __getinitargs__ to get the right __repr__ behavior.

If you haven't been following this thread, note that it isn't
clear to everyone that this would always be the right behavior
even when it's possible.  Not saying there's any problem with
this implementation, but just to draw a fine point, not necessarily
"the right" behavior.

	Donn Cave, donn at drizzle.com



From tjreedy at udel.edu  Tue Apr  1 20:04:52 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 1 Apr 2003 20:04:52 -0500
Subject: The Whitespace Language
References: <283adf56.0304011143.4f4a6934@posting.google.com> <3E8A12CD.6194A9A8@alcyone.com>
Message-ID: <eGWdnSuSpbdBqBejXTWc3w@comcast.com>

"Erik Max Francis" <max at alcyone.com> wrote in message
news:3E8A12CD.6194A9A8 at alcyone.com...
> Kevin Dahlhausen wrote:
>
> > I'm getting so much entertrainment out of the the joke on the
jokers
> > that created the Whitespace Language page -since we use
indentation to
> > delimit block strucutre, white-space has been part of Python's
syntax
> > since the beginning.  Ha!
>
> Yeah, I found that slightly amusing when I read the blurb on
Whitespace
> as well.

Do you mean this WhiteSpace (XML-centered Java extension)?
http://www.bluering.nl/whitespace/faq.html





From ckea25d02 at sneakemail.com  Mon Apr 14 01:15:53 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 14 Apr 2003 17:15:53 +1200
Subject: Pyrex: wrapping #defines with a type name argument
In-Reply-To: <b76ivu$eri$1@bob.news.rcn.net>
References: <b76ivu$eri$1@bob.news.rcn.net>
Message-ID: <b7dfi7$dguav$1@ID-169208.news.dfncis.de>

Edward C. Jones wrote:
> A common type of C macro has a type as an argument. For example:
> 
>     #define GET_ELEM(type,x,n) (*(type*)((char*)(x) + n))
> 
> What is a good way to wrap this in Pyrex (if any)?

I don't think there's any good way. The only things I can
think of offhand are:

(1) Write a bunch of external C functions, one for each
     type you want to deal with, which call the macro,
     and then wrap those.

(2) Don't use the macro at all, and do the casting
     yourself in Pyrex each time you need it.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From aleax at aleax.it  Fri Apr 25 18:56:14 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 22:56:14 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <vuiqa.3522$%_3.2810696@newssrv26.news.prodigy.com> <3EA9B7A4.59F9CE04@easystreet.com>
Message-ID: <i2jqa.8349$K35.249387@news2.tin.it>

achrist at easystreet.com wrote:
   ...
> So, again, wrt the larger nut, how to write RTF from Python? Any ideas?

A bit of Googling shows a Python module for RTF output is part of:
    http://ptommi.sourceforge.net/

Not sure how general-purpose it is, but it may be worth a look.


Alex




From matloff at laura.cs.ucdavis.edu  Tue Apr 29 00:32:09 2003
From: matloff at laura.cs.ucdavis.edu (Norm Matloff)
Date: Tue, 29 Apr 2003 04:32:09 +0000 (UTC)
Subject: pdb and threads
Message-ID: <slrnbas092.qkk.matloff@laura.cs.ucdavis.edu>

Is there a not-too-complicated workaround to get pdb to execute within
threads?

Norm Matloff




From news at snakefarm.org  Fri Apr 11 12:31:45 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 11 Apr 2003 16:31:45 GMT
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de> <v29im-ia2.ln1@nb2.stroeder.com> <slrnb9dma7.1vh.news@cipw08.physik.uni-bonn.de> <slrnb9doke.180.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <slrnb9drfh.228.news@cipw08.physik.uni-bonn.de>

Ben Hutchings wrote:
> Carsten Gaebler wrote:
>> 
>> Right. But if you are a really big ISP (and the original poster's ISP *is*
>> big) you do not want to let your web servers write error logs.
><snip>
> 
> Big enough not to care about errors?!  I would leave that ISP as quickly
> as possible.

Trust me: if you get 5 million requests a day, there are other more efficient
ways of tracing errors than scanning error logs. But I think this is getting
off-topic now.

cg.




From ianb at colorstudy.com  Mon Apr  7 01:33:00 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 07 Apr 2003 00:33:00 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	 <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
	 <b6qs04$7eqi1$1@ID-169208.news.dfncis.de>
	 <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
Message-ID: <1049693579.9724.217.camel@lothlorien>

On Mon, 2003-04-07 at 00:23, Carl Banks wrote:
> Third, there are a *ton* of functions in Emacs that operate on a
> certain point in a certain buffer (delete-char, kill-region, etc.).
> These kinds of functions typically get bound to keystrokes and other
> sorts of commands, which a user expects to work in the current buffer.
> If all these functions had to be passed the current buffer as an
> argument, it would just complicate things.

Maybe instead they should be methods of the buffer, and a keystroke
automatically calls a method on the current buffer.

Managing so many methods from diverse sources might be difficult.  I
want generic methods, I keep thinking of places where they'd be mighty
useful... like Mixins but better.

  Ian






From mwilson at the-wire.com  Thu Apr 24 11:01:48 2003
From: mwilson at the-wire.com (Mel Wilson)
Date: Thu, 24 Apr 2003 11:01:48 -0400
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org>
 <GXupa.83382$T34.2079161@news2.tin.it> <5jadeg285s.fsf@xoxy.net>
 <b87iqe$e9v$1@slb4.atl.mindspring.net>
Message-ID: <cx/p+ks/KHnc089yn@the-wire.com>

In article <b87iqe$e9v$1 at slb4.atl.mindspring.net>,
"Andrew Dalke" <adalke at mindspring.com> wrote:
>Charlie Reiman:
>> Are there any dangers in something like this:
>>
>> a = [1, infimum, infimum]
>> a.sort()
>>
>> Should the __cmp__ methods return 0 for equality (identity, in this
>> case)?
>
>As Alex pointed out, it should.
>
>However, I recall a post by Tim Peters some years ago saying
>that he fixed the sort code so it would finish, even in the face of
>absurdities like this.
>
>But I'm too lazy to find the post, or even test the code.  ;)

I've been playing:

#=======================================

"""limitcases.py
Provides objects which compare less-than and greater-than any other object.
Useful for sentinels, or for initial values in maximum- or minimum-finding
routines.
"""

class LowLimit:
    "Instances of LowLimit compare < any object except each other."
    def __cmp__ (self, other):
        if isinstance (other, LowLimit):
            return 0
        return -1

    def __hash__ (self):
        return id (LowLimit)
# class LowLimit


class HighLimit:
    "Instances of HighLimit compare > any object except each other."
    def __cmp__ (self, other):
        if isinstance (other, HighLimit):
            return 0
        return 1

    def __hash__ (self):
        return id (HighLimit)
# class HighLimit


if __name__ == '__main__':
    print LowLimit()
    print HighLimit()
    print

    # An application can provide its own printable representation.
    LowLimit.__str__ = lambda (x): "-Infinity"
    HighLimit.__str__ = lambda (x): "+Infinity"

    x = [1, HighLimit(), 2, LowLimit(), 0, HighLimit(), LowLimit()]
    x.sort()
    print 'X = ',
    for v in x:
        print v,
    print
    print
    print'X:', x
    print

#=======================================


   I hope I'm not spoiling the game for people who should be
learning.  I should probably look up singleton classes which
LowLimit and HighLimit should probably be.

        Regards.        Mel.



From vlindberg at verio.net  Tue Apr  1 23:29:39 2003
From: vlindberg at verio.net (VanL)
Date: Wed, 02 Apr 2003 04:29:39 GMT
Subject: swig help
Message-ID: <TGtia.6963$8Z5.50471@sea-read.news.verio.net>

Hello,

I have a c file that has a function

void myfunc(int x, int y, int z, float *new_x, float *new_y, float *new_z);

I can create an interface to it fine, but the typemap stuff keeps 
tripping me up -- I cannot seem to get it so that from python it looks like

 >>> result = myfunc(1,2,3)
 >>> result
(1.0, 2.0, 3.0)

Instead, it complains that it expects six arguments, the last three of 
which need to be pointers to ints.

I have the following interface file:

%typemap(argout) float *OutValue {
     PyObject *o, *o2, *o3;
     o = PyFloat_FromFloat(*$1);
     if ((!$result) || ($result == Py_None)) {
         $result = o;
     } else {
         if (!PyTuple_Check($result)) {
             PyObject *o2 = $result;
             $result = PyTuple_New(1);
             PyTuple_SetItem(target,0,o2);
         }
         o3 = PyTuple_New(1);
         PyTuple_SetItem(o3,0,o);
         o2 = $result;
         $result = PySequence_Concat(o2,o3);
         Py_DECREF(o2);
         Py_DECREF(o3);
     }
}

%typemap(in,numinputs=0) double *OutValue(double temp) {
     $1 = &temp;
}


void next(int x, int y, float z,
           float *new_x, float *new_y, float *new_z);


Could anyone tell me what I have wrong?

Thanks,

VanL




From duncan at NOSPAMrcp.co.uk  Tue Apr 15 11:06:32 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 15 Apr 2003 15:06:32 +0000 (UTC)
Subject: any benchmark results for python?
References: <b7h40n$28s$1@skates.gsfc.nasa.gov>
Message-ID: <Xns935EA2B55DF6Dduncanrcpcouk@127.0.0.1>

"J Li" <jylimd at yahoo.com.removeMe> wrote in
news:b7h40n$28s$1 at skates.gsfc.nasa.gov: 

> I am looking for benchmark test results for python, in comparison to
> other languages like Perl or Java or C++

See http://www.ipd.uka.de/~prechelt/Biblio/jccpprtTR.pdf

The author doesn't just compare running speed, he also compares memory 
consumption, development time and a bunch of other statistics.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From skip at pobox.com  Mon Apr 14 23:05:58 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 22:05:58 -0500
Subject: more informative stack traces
In-Reply-To: <200304141403.08895.scott_list@mischko.com>
References: <e0bf36b5.0304132100.4324c3b8@posting.google.com>
        <3e9ab723$0$66212$edfadb0f@dread11.news.tele.dk>
        <200304141403.08895.scott_list@mischko.com>
Message-ID: <16027.30486.863490.67122@montanaro.dyndns.org>

    >> > I am proposing a different style of stack trace which I found more
    >> > useful in debugging problems with live code. The main change
    >> > involves showing the actual parameter values in function calls,
    >> > gdb-style.

    >> May I suggest you refactor your code into two modules: a
    >> platform-independent one that inspects tracebacks and generates the
    >> traceback text in some sort of container or generator.  

    Scott> Any possibility of getting this idea to be part of Python? It
    Scott> seems like the Right Thing to me.  It could be made
    Scott> user-configurable (on or off) so that backward compatibility
    Scott> could be maintained.

Everyone involved should take a look at the cgitb module in the standard
distribution.  I submitted a patch for cgitb a long time ago which allows
plain text traceback generation:

    http://python.org/sf/569574

Note Ping's comment appended to that patch that the core code belongs in the
traceback module.

If you can get this sorted out quickly so that the core code is in the
traceback module I suspect you can still squeeze it into 2.3.

Skip






From ianb at colorstudy.com  Thu Apr 17 02:22:04 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 17 Apr 2003 01:22:04 -0500
Subject: HTML form pre-populating functionality or hints on building my
	own
In-Reply-To: <200304161939.13004.scott_list@mischko.com>
References: <200304161939.13004.scott_list@mischko.com>
Message-ID: <1050560524.715.198.camel@lothlorien>

On Wed, 2003-04-16 at 21:39, Scott Chapman wrote:
> Hi!
> I'm wanting to find (or make) a module that will take a HTML file, locate the 
> input-related form tags (<INPUT, <TEXTAREA, <SELECT, <OPTION) and modify them 
> by putting in values from a dictionary (i.e. pre-populate the form). I want a 
> module that is implemented 100% in python and does _not_ use any special tags 
> to do this (such as webware, albatross, spyce, etc.)

Huh... for some reason it's never occured to me to do it this way, I
always emit the tags with the values filled in, rather than parsing the
HTML.  An interesting idea.

Of course someone may correct me, but I haven't seen anything like this
for Python, and I at least try to keep an eye out for such stuff.

> If I should build this myself, I'd like to know, would be better to use a text 
> parsing module such as Plex to do the parsing?  I've never used any text 
> parsing tool in Unix so it'll be all new to me.  I want to keep it simple as 
> a standalone module to only populate form elements from a variable 
> dictionary, so I'm thinking a parser might be overkill? If you think a parser 
> would be the way to go on this, which one would you recommend? (Simple and 
> quick is best IMHO).

You'll probably want to use htmllib, which is a standard Python module.

  Ian






From bokr at oz.net  Fri Apr 25 21:09:42 2003
From: bokr at oz.net (Bengt Richter)
Date: 26 Apr 2003 01:09:42 GMT
Subject: What means exactly "Memory error"?
References: <mailman.1051182397.14916.python-list@python.org>
Message-ID: <b8cm8m$p7n$0@216.39.172.122>

On Thu, 24 Apr 2003 14:06:20 +0200, "Bo M. Maryniuck" <b.maryniuk at forbis.lt> wrote:

>Hello, all.
>
>I made an application which uses 4DOM. I have a big (~10M) autogenerated HTML 
>file from an Oracle tool. The problem is to public it. Nobody want to open 
>10MB HTML file in their browser. 
>
>OK, so the main idea was to split it using frequently repeated <HR> tag. Then 
>I got ~1500 parts. OK, now I parse each part with HtmlLib and keep it tree 
>somewhere in the variable (in memory). Then after each part is parsed I want 
>to do a lot of changes inside the document to adapt it to using with Zope in 
>the cycle.
>
Are you essentially extracting data that originated from Oracle, so as to
be able to re-publish it wrapped in your own HTML? If so, can you connect to
Oracle and just get the data? Or does the HTML solve access and
interfacing problems you don't want to deal with?

Is the data frequently updated, so caching is not going to help?
Is there a time stamp in the HTML that you can use as a quick check
against repeat processing if you are serving many pages between updates?

How would just extracting the data from the HTML and putting it in your
own simplified data base work out? Could you generate your various pages
that way? Doesn't Zope support various data bases? How much traffic do
you have to support? Someone who uses/develops Zope can advise.

How much of the original 10mb winds up copied as is to your output?

It's hard to guess what a good approach would be, not knowing what
you are actually doing.

Just a few thoughts...

Regards,
Bengt Richter



From yaipa at yahoo.com  Thu Apr 17 12:11:59 2003
From: yaipa at yahoo.com (yaipa h.)
Date: 17 Apr 2003 09:11:59 -0700
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it> <o1pna.3380$lf3.595806@cletus.bright.net>
Message-ID: <6e07b825.0304170811.129cf701@posting.google.com>

"George Kinney" <gk_2345 at yahoo.com> wrote in message news:<o1pna.3380$lf3.595806 at cletus.bright.net>...
> "Alex Martelli" <aleax at aleax.it> wrote in message
> news:Wrena.33096$T34.876856 at news2.tin.it...
> > need to understand dotNET to profit from this book".  I don't know if
> > there are substantial slices of programmers still stuck on VB5, VB4, ...,
> > and so on backwards, that might be scared off by the "6" in the title.
> 
> Believe it or not, I am the not so proud possessor of an application written
> in a mish-mash of VB5 and dBase.  Its creator claims that he can't upgrade
> to VB6, because VB6 is broken.  Yes, it is 8 years old and the same
> half-finished
> POS it was in '95, but he just won't budge.
> 
> The beautiful thing about it is his own insistance on antiquated tools gave
> me the
> argument I needed convince management to let me re-implement this particular
> package...in python. :)
> 
> What's the fastest way to convert a VB drone to python in my opinion?
> Make them use python for a project, big or small,  they'll never go back.


+1 ( or maybe +(n) ) ;-)



From theller at python.net  Thu Apr 17 12:20:18 2003
From: theller at python.net (Thomas Heller)
Date: 17 Apr 2003 18:20:18 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <d6jmqr2q.fsf@python.net> <Xns936072232AA84duncanrcpcouk@127.0.0.1> <he8xgr3g.fsf@python.net> <Xns9360A249561CBduncanrcpcouk@127.0.0.1> <Xns9360A663C7BA5duncanrcpcouk@127.0.0.1>
Message-ID: <znmpf64d.fsf@python.net>

Duncan Booth <duncan at NOSPAMrcp.co.uk> writes:

> > The difference code could actually be made much faster quite easily.
> > If I added a search for the first difference, then I could do the
> > ellipsis and truncating first and only pass what is left to difflib.
> 
> Old version took 0.18 seconds on Python 2.3, gave up waiting on Python 2.2. 
> This version takes 0.18 on Python 2.2, 0.08 on Python 2.3.

Cool! I can now use 10 million character strings, and the test only
takes 1.5s in 2.2, and 0.7s in CVS python.

Thomas



From tjreedy at udel.edu  Tue Apr 29 12:22:58 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 29 Apr 2003 12:22:58 -0400
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com>
Message-ID: <qqednfRtZYo_OzOjXTWcog@comcast.com>

"Christopher Blunck" <blunck at gst.com> wrote in message
news:1147e466.0304290638.287687f6 at posting.google.com...
> However, when I download a python module, it frequently has unit
> tests.

This is now a requirement for acceptance of new modules.  Reviewing
the test module along with the modules itself and doc patches is part
of the review process.   Bug tests accompanying bug patches are also
encouraged.

> What advice can you offer on spreading the testing bug around?

Perhaps note that systematic testing has been aroung awhile.  I
believe the ADA validation test suite was written pretty early.
Python had a systematic syntax test at least by 1.3 (my first version)
and I presume much earlier.

> What success stories do you have?

As an observer, Python itself.  It has gone thru numerous releases
(15+) and has been ported to numerous platforms (20+? depending on how
count) over more than a decade without succumbing to bit rot.

It is my impression that the current code version gets tested daily on
a variety of platforms.  Every week or two, someone posts on PyDev
mailing list 'Test X failed on system Y.  Sometimes a patcher forgot
to run the appropriate test, sometimes there is negative interaction
with other modules or compilation units, sometimes there is a platform
specific problem (on a system different from the one used by the
patcher).

Usually, mis-patches are fixed within a day or two.  Sometimes they
are withdrawn for longer reconsideration.  In either case, there is
another enhancement bug that most of us never see and another success
for relentless retesting.

Terry J. Reedy





From j4nff5m02 at sneakemail.com  Sat Apr 19 09:19:29 2003
From: j4nff5m02 at sneakemail.com (Roel Schroeven)
Date: Sat, 19 Apr 2003 13:19:29 GMT
Subject: getting an item's list index
In-Reply-To: <mailman.1050757111.12171.python-list@python.org>
References: <mailman.1050757111.12171.python-list@python.org>
Message-ID: <B1coa.52684$t_2.4923@afrodite.telenet-ops.be>

Michael Anckaert wrote:
> Hello all,
> 
> I have a list called contents. If one of the elements has the value
> "item1" for example, how can I get that elements index?
> 
> 
> for item in contents:
> 	# here I need to get that items index
> 

One could do:

for item in contents:
	index = contents.index(item)

But that's quite inefficient (0(n^2) I think). If you need the index 
anyway, I think you'd better do it like this:

for index in range(len(contents)):
	print index, contents[index]


-- 
"Codito ergo sum"
Roel Schroeven
http://roelschroeven.net




From aleax at aleax.it  Thu Apr 17 11:10:29 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 15:10:29 GMT
Subject: check to see if a symbol in a string has a binding
References: <kfed6jlv06e.fsf@alumni.cse.ucsc.edu>
Message-ID: <Ftzna.27929$LB6.642543@news1.tin.it>

Jinsong Zhao wrote:

> Sorry if the question is too simple. I want to see if the name inside
> the string has a binding and callable. For example,
> 
> 1. func_string="func"
> 2. func_string="func_string"
> 3. func_string="dir"
> 
> So the in case 1, "func" does not have a binding; case 2 "func_string"
> has a binding but not callable; in case 3 "dir" has a binding and is
> callable. How do I proceed based on the content in func_string? This
> is rather baffling.

It IS a rather weird specification, yes -- to look for a binding
ANYwhere, whether it be in the current function's locals, in the
current module's globals, or in the builtins?!  Why would one
ever want to do something like that...?

Anyway, first of all, you want the string's content to be "a
name": you can check that with modules cStringIO, tokenize,
and keyword -- perhaps overkill, but that *ensures* you're 
using the same rules Python itself uses to define what's 
a 'name'...:

import cStringIO, tokenize, keyword

def isaname(s):
    toks=list(tokenize.generate_tokens(cStringIO.StringIO(s).readline))
    if len(toks)!=2: return False
    if tokenize.tok_name[toks[0][0]] != 'NAME': return False
    return not keyword.iskkeyword(s)


Once you DO know s is a name, you can then safely try to evaluate
it as an expression, catching any possible errors, and check if 
the resulting value is callable.

So, overall:

def checkJZ(s):
    if not isaname(s): return False
    try: obj = eval(s)
    except NameError: return False
    else: return callable(obj)

So, does this meet your peculiar specs...?


Alex




From peter at engcorp.com  Wed Apr 16 15:02:21 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 15:02:21 -0400
Subject: How many of you are Extreme Programmers?
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com> <mailman.1050504987.840.python-list@python.org> <3E9D8A80.FD76BDE9@engcorp.com> <b7k3kj$4gj$1@news.hananet.net> <3E9DA883.CE0C38B4@engcorp.com>
Message-ID: <3E9DA8BD.250A3566@engcorp.com>

Peter Hansen wrote:
> 
> Changjune Kim wrote:
> >
> > Check out the Standish Group CHAOS report.
> >
> > http://www.softwaremag.com/archive/2001feb/CollaborativeMgt.html (refers to
> > 2001 report)
> > http://www.pm2go.com/sample_research/chaos_1994_1.php
> 
> The first link redirects to http://www.softwaremag.com/magazine.css,
> which is not found (404).

Sorry, I take that back.  Netscape 4.7 barfs on it, but IE handles
it fine...

-Peter



From edreamleo at charter.net  Sat Apr 19 08:39:41 2003
From: edreamleo at charter.net (Edward K. Ream)
Date: Sat, 19 Apr 2003 07:39:41 -0500
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050569128.32513.python-list@python.org>
Message-ID: <va2gsi3q5kl0b3@corp.supernews.com>

Thanks, Christian, for your pamphlet.  The excitement you exhibit shows just
why the sorrows, doubts and labors of creative endeavor are so worthwhile.
I myself wouldn't trade them for anything.

Sometimes earlier "false starts" are absolutely essential to a great
project.  I am reminded of how Andrew Wiles patched up his proof of Fermat's
Last Theorum.  The reason the penultimate proof failed was precisely the
reason a previously "failed" approach (the Taniyama-Shimura  conjecture)
could be made to work.

Edward
--------------------------------------------------------------------
Edward K. Ream   email:  edreamleo at charter.net
Leo: Literate Editor with Outlines
Leo: http://webpages.charter.net/edreamleo/front.html
--------------------------------------------------------------------





From brandon at geerbox.com  Thu Apr 17 16:48:12 2003
From: brandon at geerbox.com (Brandon Lewis)
Date: Thu, 17 Apr 2003 13:48:12 -0700
Subject: fromfd failure on Mac OS X
Message-ID: <E7DD7802-7115-11D7-BB40-000393AA8F54@geerbox.com>

fromfd function is apparently failing to return a valid socket on Mac 
OS X.

This is the error I get:

Traceback (most recent call last):
   File "/usr/share/spyce/run_spyceCGI.py", line 21, in ?
     execfile(sys.argv[0])
   File "/usr/share/spyce/verchk.py", line 35, in ?
     execfile(sys.argv[0])
   File "/usr/share/spyce/spyceCGI.py", line 45, in ?
     main()
   File "/usr/share/spyce/spyceCGI.py", line 34, in main
     more = cgi.accept()
   File "/usr/share/spyce/fcgi.py", line 183, in accept
     result = self.recv()
   File "/usr/share/spyce/fcgi.py", line 217, in recv
     self.connection, address = self.socket.accept()
socket.error: (22, 'Invalid argument')
[Thu Apr 17 11:33:28 2003] [error] [client 12.22.222.50] Premature end 
of script headers: /usr/share/spyce/run_spyceCGI.py


Running Mac OS X 10.2.4, Apache 1.3.27, Python 2.2, Spyce 1.3.9

Has anyone run into this problem and fixed it in the past?





From timh at zute.net  Tue Apr  8 02:48:21 2003
From: timh at zute.net (Tim Hoffman)
Date: Tue, 08 Apr 2003 14:48:21 +0800
Subject: Pythonwin COM and GDI
Message-ID: <3e926fb4$1@news>

Hi

I have been using python for several several years on unix platforms
but now I have to try and get my head around all the windows extensions
etc.. (I hardly ever use windows, so I have to come up to speed
on a whole bunch of stuff that is unfamiliar to me ;-)

I need to create a WMF file using a GDI calls. At first thought
it seemed that if I where to run makepy on gdi32.dll I should be able to
then use win32com.client. stuff to then use the gdi library direct.

But makepy say "could not locate a type library matching 
'c:\windows\system\gdi32.dll'

So I suppose I am missing something in my understanding of what
constitutes a valid DLL/ActiveX component that would suit makepy being
run against.

I have gone through the activestate docs, I must be thick,
cause I can't quite fathom the difference. Also it seems some of the GDI
stuff is in the win32ui module but unfortunately not the things I need
like CreateEnhWMF.

Any pointers or a link to a good FM for me to read further would be 
appreciated.

See ya

Tim




From rajarshi at presidency.com  Tue Apr 22 13:56:48 2003
From: rajarshi at presidency.com (Rajarshi Guha)
Date: Tue, 22 Apr 2003 13:56:48 -0400
Subject: passing a C struct to python
Message-ID: <pan.2003.04.22.17.56.48.365702.10507@presidency.com>

Hi,
  I've recently looked into calling some Python functions from my C code.
I've worked out the ways of calling a Python function
(PyObject_CallMethodObjArgs etc). However these are OK if I want to pass
a list, tuple, string etc.

However in some cases I might want to pass  an 'object' (which would be a
C struct) to the Python function. Alternatively, the Python function
might require me to pass an instance of some class. Am I correct in
thinking that these two problems are equivilant?

And my question is: what do I need to do to pass structs/classes etc? I
tried looking for pointers in the Embedding section of the Python manual,
but I could'nt really find anything :(

Any pointers would be appreciated

Thanks,



From BPettersen at NAREX.com  Tue Apr 22 10:52:50 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Tue, 22 Apr 2003 08:52:50 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE065@admin56.narex.com>

> From: Ian Bicking [mailto:ianb at colorstudy.com] 
[... + SQLObject example]

> for paper in Paper.select(AND(Paper.q.title == 'foo',
>                               Paper.q.author.startswith('Bob'))):

Does anyone know the reason for not allowing an overload of the and
operator? (of course, the only use case I've seen is the one above
[thrice, first time in 97] and alternate logical calculi [e.g. fuzzy
logic] -- so not much different from some of the other operators? :-)

-- bjorn





From staschuk at telusplanet.net  Wed Apr 23 15:53:21 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 23 Apr 2003 13:53:21 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
In-Reply-To: <b86gim$60r$1@slb5.atl.mindspring.net>; from adalke@mindspring.com on Wed, Apr 23, 2003 at 10:58:18AM -0600
References: <mailman.1051023275.29338.python-list@python.org> <b8420j$629$1@slb3.atl.mindspring.net> <mailman.1051043856.10118.python-list@python.org> <b86gim$60r$1@slb5.atl.mindspring.net>
Message-ID: <20030423135321.A1892@tibia.amotlpaa.bogus>

Quoth Andrew Dalke:
> Steven Taschuk
> > __and__ could be passed other as a callable which computes and
> > returns the second operand.  Then the standard __and__ could be
> >     def __and__(self, other):
> >         if bool(self):
> >             return self
> >         return other()
> 
> So a new idiom and new code (to generate a function from a set of
> unevaluated code), just to allow overriding booleans?  Seems
> rather excessive to me.

Yes.

I do not (and did not) endorse this notion for __and__.  I simply
do not agree with the previous claim that there's no way to
implement overloading and short-circuiting for the same operator.

No *good* way given how Python works now, perhaps.

> Gets pretty complicated with code like
> 
> print value or sys._getframe().f_lineno
> 
> since the function introduces a new frame.

A nice example.

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From aleax at aleax.it  Thu Apr 17 11:34:34 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 15:34:34 GMT
Subject: Complexity of standard Python data structures (Re: Speed of string += string)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi> <Oqwma.11524$LB6.319489@news1.tin.it> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <2DFma.19063$T34.605833@news2.tin.it> <slrnb9tfdn.lgn.marcus@infa.abo.fi>
Message-ID: <eQzna.42991$T34.1189411@news2.tin.it>

Marcus Alanen wrote:
   ...
> On the other hand, given the amount of questions regarding big-O
> notation & string += string at the moment on c.l.p, perhaps there
> should be a note -- a "warning" sounds a bit extreme?  Also in the docs
> of C functions, of course!  Sometimes, unix man pages have usage
> examples. Perhaps there should be "anti-usage" examples as well?  :)

Yes, I have changed my mind -- partly thanks to your arguments,
partly because of somebody else (a CS professor) who posted to 
c.l.py trying to understand Python's "lists" after reading the
tutorial -- it seems we wasted quite a bit of his time because
of his assumptions that "list" means "list"... a short footnote
in the tutorial might well help.  Darn -- now I wish I had put
just such a footnote in the Nutshell's Chapter 4 too, rather than 
just mentioning list performance behavior in chapter 17 -- oh well, 
perhaps I can send in a small errata...;-).

> people do seem to start explaining the python list as e.g. "a C++
> vector", or "a contiguous array of heterogeneous elements". I do
> understand that Python people don't want to make any guarantees,
> although describing some implementation might make it easier
> for people to learn (or migrate to) Python.

Yes, I now agree with you -- particularly in the tutorial, it
would help more people than it would hinder.  Thanks for helping
me see the light about this, btw.  Now we'll see if I can maybe
convince the docs' maintainer, and thus get that tiny fix into 
Python's docs in time for 2.3 beta...

An early _focus_ on implementation and performance vs functionality
issues would be too much, but a tiny little mint, oops I mean
footnote, surely _that_ couldn't hurt...


Alex




From jepler at unpythonic.net  Wed Apr  9 23:00:07 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Wed, 9 Apr 2003 22:00:07 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <yfs3ckrihgi.fsf@black132.ex.ac.uk>
References: <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk>
Message-ID: <20030410030005.GA3573@unpythonic.net>

On Wed, Apr 09, 2003 at 08:44:29PM +0100, Alexander Schmolck wrote:
> Beni Cherniavsky <cben at techunix.technion.ac.il> writes:
> 
> > It also describes some implementation details - not that I believe we
> > want dynamic scope...  
> 
> Why not? Amongst other things dynamically scoped global variables seems the
> only sane way to me to have global variables. Otherwise, everytime you screw
> around with some global variable temporarily you have to go through all sorts
> of pain making sure you clean it up after you're done (even if something went
> wrong) and you'd better not be multithreaded. Compare:
> 
> # simple, single threaded case
> try:
>    global precision
>    old = precision
>    precision = 10e-10
>    dosomething()
> finally:
>    precision = old
> 
> (let ((precision 10e-10)) (dosomething))

Ignoring the fact that "globals" are module-level, you can obtain
"stacking" of global variables with only a single level of name lookup,
including automatic removal on return, in Python.  All that you need is a
convenent way to write complex things in the 'body' argument.  Untested
code, of course

_let_undef = object()
def let(body, **kw):
	restore = {}
	g = globals()
	for k, v in kw.items():
		restore[k] = g.get(k, _let_undef)
		g[k] = v
	try:
		return body()
	finally:
		for k, v in restore.items():
			if v is _let_undef:
				del g[k]
			else:
				g[k] = v



Jeff




From exarkun at intarweb.us  Wed Apr  9 19:08:39 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 9 Apr 2003 19:08:39 -0400
Subject: Cross-reference 'import' in a class hierarchy
In-Reply-To: <b6vqhf$9n4qi$1@ID-169208.news.dfncis.de>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <mailman.1049778853.27420.python-list@python.org> <b6vqhf$9n4qi$1@ID-169208.news.dfncis.de>
Message-ID: <20030409230839.GD10833@meson.dyndns.org>

On Wed, Apr 09, 2003 at 12:56:18PM +1200, Greg Ewing (using news.cis.dfn.de) wrote:
> Jp Calderone wrote:
> >  It's simple.  Code should not depend directly on code higher than it in
> >the package hierarchy.  If it does (and it isn't one of a few extremely
> >rare cases), it is misdesigned.
> 
> But it's worse than that -- you can't even refer to things
> *deeper* in the hierarchy relative to yourself.
> 

  Yack, you're right.  I am actually surprised that people aren't bitten by
this more frequently.  I know I can't remember any time where I've had a bug
caused by it.

  It would be nice if it didn't happen at all, though :(

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 20 days, 19:01, 1 user, load average: 0.21, 0.15, 0.05




From aleax at aleax.it  Thu Apr 24 02:51:07 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 06:51:07 GMT
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com>
Message-ID: <vPLpa.51482$LB6.1161039@news1.tin.it>

Peter Hansen wrote:

> Aahz wrote:
>> 
>> Which brings up the question: do the unit-testers here write tests for
>> throwaway scripts?
    ...
> Basically, with a literally throwaway script, you do lose much of
> the business value of the tests, so as far as I can see you write
> them only if they truly help you write the code properly, which
> should once you're proficient at TDD and then only sometimes, since
> you're a very experienced programmer.  Otherwise, and certainly in some
> cases no matter who you are, they really would be a waste of time.
> IMHO.  Except maybe to help you develop the habits.   _Four_ reasons...

Oh come on -- we all know that some "throw-away" scripts happen
to NOT get thrown away, but rather lie around and get reused and
tweaked and grown and...  It's hard to quantify HOW MANY such
"throw-away" scripts are in fact thrown away, and how many end
up getting kept (and reused and grown) instead -- guess it depends
on how good one is at planning / forecasting -- but as for me,
I would paraphrase Brooks' MMM Chapter 11's famous title:

        Plan NOT to throw one away -- you won't, anyway.


Alex




From max at alcyone.com  Thu Apr 17 13:58:44 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 17 Apr 2003 10:58:44 -0700
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net> <3E9D35FC.FCAD17C9@alcyone.com> <b7mp2u$3k8$0@216.39.172.122>
Message-ID: <3E9EEB54.682751BC@alcyone.com>

Bengt Richter wrote:

> What other use is there (outside of strings and comments)?
> Have we purposely been set up for gaucheries?
> ;-)

In shell-like languages, the backquotes are used to mean "Execute what's
in between the backquotes as a shell command, and then substitute the
(stdout) output of that command and continue, e.g.:

max at oxygen:~% echo The date is: `date` 
The date is: Thu Apr 17 10:54:03 PDT 2003

In modern shells, the form $(...) is preferred to `...`, since the
former can be nested, whereas the latter cannot.

As for Python's use of the backquotes, I presume it's some historical
issue.  I don't know what Guido would say about its use now, but there
seems pretty near total consensus that repr(x) looks a lot better (and
clearer) than `x`.

I do occasionally use `x`, but only when inserting debug statements
since they're set off so obviously.  "Oh, I see backquotes, that
shouldn't stay in there in production code."

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Freedom is a road seldom travelled by the multitudes
\__/ Public Enemy
    The laws list / http://www.alcyone.com/max/physics/laws/
 Laws, rules, principles, effects, paradoxes, etc. in physics.



From duncan-news at grisby.org  Sun Apr  6 16:34:36 2003
From: duncan-news at grisby.org (Duncan Grisby)
Date: Sun, 06 Apr 2003 20:34:36 +0000
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
Message-ID: <7563a$3e908f5c$516049d2$26016@nf1.news-service.com>

In article <a626bbd7.0304020829.1e5cac2f at posting.google.com>,
robin <escalation746 at yahoo.com> wrote:

>I am attempting to summarise the distributed computing implementations
>available to Python programmers. My conclusions so far are as follows:
>
>If you want the biggest, boldest approach and don't care about
>overhead, use CORBA.

You might want to read the slides to the presentation I gave at the UK
Python conference last week. The title is "CORBA?  Isn't that
Obsolete?". You can get it here:

  http://www.grisby.org/presentations/accu2003.pdf

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --



From igor at tyumbit.ru  Mon Apr 21 11:19:47 2003
From: igor at tyumbit.ru (Igor Prischepoff)
Date: Mon, 21 Apr 2003 21:19:47 +0600
Subject: wxBoxSizers in window and two embedded panels.
Message-ID: <14641910474.20030421211947@tyumbit.ru>

Well, Donald, your code does what i've expected.
 :)
By the way i've changed hierarchy
 window1
  /    \
panel1  panel2

to

  panel0
  /    \
panel1  panel2

Just like in your code.
And everything worked just fine.
Very strange. Panel is like window, so what the difference ?
It's still a mistery for me.
Any way - workaround solution is still the solution.

Thank for your help!
  

-- 

                     mailto:igor at tyumbit.ru





From Kyler at news.Lairds.org  Tue Apr  8 13:23:59 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Tue, 08 Apr 2003 17:23:59 GMT
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com> <a8b7f07a.0304080539.465049c8@posting.google.com>
Message-ID: <1abbm-ccl.ln1@news.lairds.org>

alloydflanagan at attbi.com (A. Lloyd Flanagan) writes:

>I'm going to make a suggestion you may not hear from anybody else in
>this group.  If hiding the code is important to you, don't use Python.
> Use C++ or something similiar.

I suggest Perl.  I moved to Python in part because I could
rarely read Perl code generated by someone else or even my
own code after it sat for a month or so.

I consider Perl a write-only language.

>What about coding the sensitive algorithms in C and calling them from
>python?  There are of course decompilers for C as well, but it's at
>least harder.

Someday Microsoft will save us from this sorry state with
Palladium.  Then we can all relax.

--kyler



From bokr at oz.net  Tue Apr 29 16:06:59 2003
From: bokr at oz.net (Bengt Richter)
Date: 29 Apr 2003 20:06:59 GMT
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com> <zAara.22169$K35.739509@news2.tin.it> <roy-621090.18204128042003@reader1.panix.com> <fAvra.30824$K35.974064@news2.tin.it>
Message-ID: <b8mm13$p4e$0@216.39.172.122>

On Tue, 29 Apr 2003 14:00:43 GMT, Alex Martelli <aleax at aleax.it> wrote:

>Roy Smith wrote:
>
>> Alex Martelli <aleax at aleax.it> wrote:
>>> No statistically significant / measurable difference.  The tiny
>>> difference in conciseness is hardly compelling one way or another,
>>> either.
>> 
>> I disagree (respectfully).
>> 
>> I'm all about writing code that's easier to read.  I'll admit that the
>> += form is only marginally simplier to read, but marginal is better than
>> nothing.
>
>When what you're incrementing is a long name or an expression of
>any substantial complication, then "left += right" may indeed offer
>_significant_ advantages.
>
>When what you're incrementing is "i" (as in the original question),
>this is quite moot -- the alleged "increase in simplicity" is so
>small that experimental error may be < 0...;-).
>
>And I don't think that "one obvious way to do it" applies here --
>just as, when I'm adding two numbers a and b, I'm pretty happy to 
>get to choose between a+b and b+a depending on what's more readable
>in the given context, quite similarly the choice between "i=i+1"
>and "i+=1" is not necessarily preordained
>
>> I tend to pick-and-choose from the XP menu, but one of the items I
>> really like is "refactor mercilessly".  Why write a variable name twice
>> when you can write it once?  Easier to read, and less error prone.  Have
>
>When the variable name is short and sharp, writing it twice is no
>problem -- and sometimes it can be more readable e.g. by showing
>some parallelism:
>
>    if override_with_new:
>        current = incoming_new + increment
>    else:
>        current = current + increment
>
>The choice between this, the "refactored" forms such as:
>
>    if override_with_new:
>        current = incoming_new
>    current += increment
>
>etc, etc, is indeed strictly one of readability, and thus of style.
>I'm glad to have += as a CHOICE, but wouldn't like it as a MUST:-).
>
This is a bit contrived, but if there's a side effect buried in the access expression,
there might be a must if you want to avoid it:

 >>> class Foo:
 ...     def __init__(self, v): self.v = v
 ...     def __repr__(self): return '<Foo %s>' % self.v
 ...     def __iadd__(self, other):
 ...         self.v = self.v + other
 ...         return self
 ...     def __add__(self, other):
 ...         print 'side effect'
 ...         return Foo(self.v + other)
 ...
 >>> foo = Foo(10)
 >>> foo
 <Foo 10>
 >>> foo += 5
 >>> foo
 <Foo 15>
 >>> foo = foo + 5
 side effect
 >>> foo
 <Foo 20>

Kind of perverse design to make x += y differ from x = x + y,
but it's at least possible ;-)

Regards,
Bengt Richter



From marcus at infa.abo.fi  Mon Apr 14 06:40:38 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 14 Apr 2003 12:40:38 +0200
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com>
Message-ID: <slrnb9l0dm.96d.marcus@infa.abo.fi>

On Sat, 12 Apr 2003 20:14:51 -0400, Roy Smith <roy at panix.com> wrote:
>One of the very few things I like about C++ over Python is that the STL 
>is very careful to document the complexity of the various operations.  
>AFAICT, no such documentation exists for Python.

Indeed, that was the first thing I wondered about when I started
learning Python.

>As another example, I can guess that Python dictionaries work like STL 
>maps on the inside, but that would be just a guess.  Is inserting an 
>element logN?  Is keys() linear?  Is del on an element logN?  Is 
>has_key() logN?  Those would be my guesses.  But they're just guesses.

Others answered this very thoroughly, I'd just like to emphasize the
linearity of keys() --- most often you can manage with .iterkeys() or
.itervalues() just fine, and it's a lot faster.

-- 
Marcus Alanen
maalanen at abo.fi



From max at alcyone.com  Wed Apr 16 18:53:09 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 15:53:09 -0700
Subject: Big-O notation (was: Re: Python and Schools)
References: <qGqla.61$Tj7.91655@news.uswest.net>
		<pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
		<Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
		<v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
		<PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
		<3E9C6C62.8010808@rogers.com>
		<mailman.1050441947.6447.python-list@python.org>
		<3E9CBAFE.8016BFB2@alcyone.com> <mailman.1050506131.12672.python-list@python.org>
Message-ID: <3E9DDED5.726FAFB5@alcyone.com>

"Christopher A. Craig" wrote:

> Erik Max Francis <max at alcyone.com> writes:
> 
> > Similarly, O(2^n) ~ O(10^n) ~ O(e^n) too.
> 
> The second part of this isn't true.  If a process f(n) has a running
> time T(N) for in input of size N, said process is O(N) if there exist
> constants "k" and "j" such that O(N)*k <= T(N) for all N>=j
> 
> While there does exist a constant (log(A)/log(B)) for which
> log_A(N)*k <= log_B(N) for all sufficiently large N, there is
> no constant for which (A**N)*k <= B**N for all sufficiently large N
> when B>A.

[scribbles on paper]

Right you are, thanks for that important correction.  The ratio of two
logarithmic functions with different bases is a constant, but the ratio
of two exponential functions with different bases is a different
exponential function, e^(k n), where k is the difference in the natural
logarithms of the two original bases.  So indeed, O(a^n) !~ O(b^n) for a
!= b.

Good catch!

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The only source of knowledge is experience.
\__/ Albert Einstein
    Bosskey.net / http://www.bosskey.net/
 A personal guide to online multiplayer first person shooters.



From rganesan at myrealbox.com  Tue Apr 15 05:42:42 2003
From: rganesan at myrealbox.com (Ganesan R)
Date: 15 Apr 2003 15:12:42 +0530
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it> <ouadesfkn2.fsf@andlx-anamika.cisco.com> <7mOma.20244$T34.649927@news2.tin.it> <ouvfxgdv8n.fsf@andlx-anamika.cisco.com> <kqPma.20705$T34.658905@news2.tin.it>
Message-ID: <our884drl9.fsf@andlx-anamika.cisco.com>

>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:

> No!  the object that get_referrers is returning is the dictionary
> of the module -- that dictionary refers to all variables (that's
> where they live); it's NOT that variables refer to each other!

Duh! What was I thinking? Thanks again.

Ganesan

-- 
Ganesan R




From avner at skilldesign.com  Mon Apr 14 15:19:20 2003
From: avner at skilldesign.com (Avner Ben)
Date: Mon, 14 Apr 2003 21:19:20 +0200
Subject: Case-insensitive string comparison
Message-ID: <3e9b09bc@news.012.net.il>

Question:

    Does standard Python have case-insensitive string comparison and search?

        Avner.





From Paul_Rudin at scientia.com  Wed Apr  2 03:49:23 2003
From: Paul_Rudin at scientia.com (Paul Rudin)
Date: 02 Apr 2003 09:49:23 +0100
Subject: Questions on linked lists
References: <kl8ia.27315$OV.127379@rwcrnsc54> <5vonl-vm3.ln1@beastie.ix.netcom.com> <slrnb8ja7a.421.ramen@lackingtalent.com> <3E8A1189.3F806B4B@alcyone.com>
Message-ID: <uu1dh9td8.fsf@scientia.com>

>>>>> "EMF" == Erik Max Francis <max at alcyone.com> writes:

 EMF> Dave Benjamin wrote:
 >> That said, what would be an example of a legitimate need for a
 >> custom linked list data structure in Python? I can't say I've come
 >> across one yet.  Do linked lists outperform native Python lists at
 >> certain operations?

 EMF> Yes.  Python lists are (computer science) vectors, which means
 EMF> they're dynamically resizeable arrays.  It's inexpensive to
 EMF> index into the list [O(1)] and add to the list [amortized O(1)],
 EMF> but it's expensive to insert or delete from the list [O(n)].  A
 EMF> linked list, although it's expensive to index [O(n)], is very
 EMF> inexpensive to insert or delete objects at _any_ location [O(1)
 EMF> for either].

Another thing to bear in mind if that a lot of indexing is not random
- much code iterates over the elements of a list in sequence. In this
circumstance the you obviously don't have to traverse down from the
head of a linked list for each element access.


-- 
Sorry, wrong ZIP CODE!!



From a.schmolck at gmx.net  Mon Apr 14 11:09:17 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 14 Apr 2003 16:09:17 +0100
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10 <7h3vfxhpl23.fsf@pc150.maths.bris.ac.uk> <yfsu1d1cisj.fsf@black132.ex.ac.uk> <7h3adettapc.fsf@pc150.maths.bris.ac.uk>
Message-ID: <yfsptnpce02.fsf@black132.ex.ac.uk>

Michael Hudson <mwh at python.net> writes:

> Alexander Schmolck <a.schmolck at gmx.net> writes:
> 
> > Michael Hudson <mwh at python.net> writes:
> > 
> > > Paul Foley <see at below.invalid> writes:
> > > 
> > > > [It should do that for lexical variables, too...conflating binding and
> > > > assignment the way Python does is a major design error, but it's too
> > > > late to fix it now]
> > > 
> > > I'm unconvinced about the "major"... but anyway: how would *you* do
> > > it?  It seems unfeasible syntactically, for starters.
> > 
> > ... what's wrong with e.g:
> > 
> > def bar():
> >    # declares locals, similar to ``global``
> >    var a,b,c = 3
> >    x = 3 # error
> >    a = 4 # OK
> > 
> > or 
> > 
> > 
> > def bar():
> >    let x=3, y=4:
> >      print x, y # OK
> >    x, y # ERROR
> > 
> > ?
> 
> Requiring declarations of all variables would be a HUGE change to
> Python, IMHO.

Sure and I'm in no way arguing for such a change (and neither Paul Foley who,
unlike me, considers implicit declaration a major flaw). I just don't
understand why it would "unfeasible syntactically" -- *syntactically* it seems
quite unproblematic to me.

'as



From tim at bladerman.com  Fri Apr  4 07:40:11 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Fri, 04 Apr 2003 12:40:11 GMT
Subject: wxSplitterWindow to fill up its avalible space
References: <6I%ia.3871$Du.12407@newsc.telia.net> <Xns93529C2F73807notrealmail@216.168.3.44>
Message-ID: <L2fja.3970$Du.13011@newsc.telia.net>

"Max Erickson" <maxume at yahoo.com> wrote
> "Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote
>
> > Hi
> > I have put a wxSplitterWindow in a wxBoxSizer together with a panel.
> > My problem is that I must set the splitterWindowsize by hand when I
> > create it
> >
> >     wxSplitterWindow(self, -1, size=(300,300))
> >
> > Otherwise it gets no size at all. If I just put two panels there
>

> Are you adding any children to the splitter? It will take sizing cues from
> them...

Argh. I thought I had removed that post from usenet :-)
In Outlook Express it is possible to remove yout own posts from unsnet but I
guess that it wasnt possible after all. Annyone know if it is supposed to
work or supposed to not work?

Anyway, I had been sitting with the problem for a few hours but not found a
way to do it but once I posted the question I found the answer almos
imediatley. I had forgotten to add wxExpand when I added the spitterwindow
to the box.

middleBox.Add(edlOutput , 1, wxEXPAND)

Thanks anyway

Tim





From evan at 4-am.com  Thu Apr 17 12:38:17 2003
From: evan at 4-am.com (Evan Simpson)
Date: Thu, 17 Apr 2003 11:38:17 -0500
Subject: HTML form pre-populating functionality or hints on building my
 own
In-Reply-To: <200304161939.13004.scott_list@mischko.com>
References: <200304161939.13004.scott_list@mischko.com>
Message-ID: <3E9ED879.1030403@4-am.com>

Scott Chapman wrote:
> I'm wanting to find (or make) a module that will take a HTML file, locate the 
> input-related form tags (<INPUT, <TEXTAREA, <SELECT, <OPTION) and modify them 
> by putting in values from a dictionary (i.e. pre-populate the form). I want a 
> module that is implemented 100% in python and does _not_ use any special tags 
> to do this (such as webware, albatross, spyce, etc.)

This won't be of any immediate use to you (I haven't got working code 
yet), but I'm building a system called TERSE which will do what you 
want.  It's intended to be an enhancement to Zope Page Templates, 
allowing you to completely remove code from templates.

TERSE stands for TEmplate Rule Set Engine, where Rule Sets are lists of 
element-matching patterns and associated declarative logic.  Your 
application of it might look a bit like this:

@document {
     data = options/data;
}
textarea[name] {
     name = attrs/name;
     content: path(data/?name | nothing);
}
input[name][type='text'],
   input[name][type='password'],
   input[name]:not([type]) {
     name = attrs/name;
     [value] = path(data/?name | nothing);
}
select[name] {
     name = attrs/name;
     selected = path(data/?name | nothing);
}
select[name] option[value] {
     value = attrs/value;
     @if(value==selected) { [selected] = 'selected'; }
}

Cheers,

Evan @ 4-am






From heiko at hhenkelmann.de  Tue Apr  8 16:57:02 2003
From: heiko at hhenkelmann.de (Heiko Henkelmann)
Date: Tue, 8 Apr 2003 22:57:02 +0200
Subject: How to split string in equally long parts
Message-ID: <b6vd0e$i1b$01$1@news.t-online.com>

I need to split a string into an array of equally long substrings, like
e.g.:

'1234567890'   -> ['12', '34','56','78','90']

What is the most efficient way to do this?

Heiko





From hyperi at hyperi.org  Sat Apr 19 04:26:36 2003
From: hyperi at hyperi.org (Oliver Sakki)
Date: Sat, 19 Apr 2003 11:26:36 +0300 (EEST)
Subject: error in string module documentation?
In-Reply-To: <1050740555.624.4.camel@carpathia>
Message-ID: <Pine.LNX.4.33.0304191125230.11689-100000@ltv.cx>

Well lstrip takes leading whitespaces away, so another way to fix it would
be 'item = cmd.lstrip()' ? :o)

/-
/- Oliver Sakki (oliver.sakki at hyperi.org)
/- http://www.hyperi.org/
/-

On 19 Apr 2003, Michael Anckaert wrote:

> Hello all,
> In the string module documentation I read the following:
>
>
> lstrip(s[, chars])
>         Return a copy of the string with leading characters removed. If
>         chars is omitted or None, whitespace characters are removed. If
>         given and not None, chars must be a string; the characters in
>         the string will be stripped from the beginning of the string
>         this method is called on.
>
> But I get the following error from my program:
>
>   File "./game.py", line 35, in run_game
>     item = string.lstrip(cmd, "look")
> TypeError: lstrip() takes exactly 1 argument (2 given)
>
> Is there an error in the documentation or is it my error?
>
> --
> Greetings,
> Michael Anckaert, aka The XanTor
> OpenPGP key: F7A6C3AB
> Fingerprint: A329 43FC 3953 A944 5DDC  2E05 8E5D AD60 F7A6 C3AB
>





From theller at python.net  Thu Apr  3 08:41:07 2003
From: theller at python.net (Thomas Heller)
Date: 03 Apr 2003 15:41:07 +0200
Subject: writing to windows memory
References: <3E8C01CE.6070601@netscape.net>
Message-ID: <n0j7bswc.fsf@python.net>

vector <Vector180W at netscape.net> writes:

> newbie:
> pythonwin 2.2.2
> Id like to bind a button.widget that simply writes a one to windows
> address 440h. any ideas on where to start.(the button stuf is trivial)
> i need to know how to access the address space.

ctypes can do it http://starship.python.net/crew/theller/ctypes.html :

from ctypes import c_int

p = c_int.from_address(0x440)

p.value = 100

This would write the value '100' as integer to memory address 0x440.
Of course, for me, it only crashes.

Thomas



From leomasi at libero.it  Mon Apr 21 07:16:27 2003
From: leomasi at libero.it (QWE)
Date: Mon, 21 Apr 2003 11:16:27 GMT
Subject: Keyboard Commands
References: <Pine.SOL.4.51.0304201825001.6411@herald.cc.purdue.edu>
Message-ID: <81l7avg7r77ko9p80mggjlfdmb2bd98b51@4ax.com>

On Sun, 20 Apr 2003 18:27:21 -0500, "richard.a.charts.1"
<charts at purdue.edu> wrote:

>Hey,
>I'm working on a program and I want to use keyboard shortcuts for some of
>my menu items.  I can't seem to find the way to implement this.  Can
>anyone point me in the right direction?  ex.  Ctrl-x, ctrl-F1, etc.
>
>Thanks for any help.
>And please feel free to email ur answers to me directly.
>
>Charts
from Tkinter import *

root = Tk()

def hello():
    print "Hello"

frame = Frame(root)
root.bind("<Control-p>", lambda e: hello())
Label(text="Press Controll-p").pack()
frame.pack()

root.mainloop()




From canrocks at yahoo.ca  Wed Apr 23 18:44:05 2003
From: canrocks at yahoo.ca (Denis Agar)
Date: 23 Apr 2003 15:44:05 -0700
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <3ea434fd$0$180$a1866201@newsreader.visi.com> <25a672ee.0304211504.4f3cd067@posting.google.com> <mailman.1050968618.16005.python-list@python.org> <va956b925cj1df@corp.supernews.com>
Message-ID: <25a672ee.0304231444.1ce0dd1b@posting.google.com>

> >VB's GUI code is likely in the VB runtime. With "VC GUI" you likely mean 
> >the MFC classes, which are usually in the MFC42.DLL. It can be used from 
> >Python via the win32 extensions, which are a relatively slim download.
> >
> >But looking at the included examples, I wouldn't want to develop a GUI 
> >using the win32 extensions, as it's basically hardcore win32 API 
> >programming.
> 			.
> 			.
> 			.
> Let's note, though, that there are several contributors
> to c.l.p who speak with comfort about having delivered
> applications to customers coded in Python against MFC.
> Gerhard and I haven't, but apparently it's not infeasible.

Ok, try to ignore the thickness of my skull, but you're saying that it
is possible to create progams using mfc42.dll, a GUI that most people
have already? I have the win32 extensions and am in the process of
trying to understand them. (PS: Anyone got some (s)ample code they
could throw my way?)
But, as I understand it, this would be hard. 

This isn't a request nor a demand nor a manifesto of any kind, but ---
would it be possible to make a simple PythonCardish GUI shaping-type
languagey thing (I'm not yet up on the lingo) that interfaces with the
win32 extensions and dumbs it down to about my level? I'm just
wondering about it is all. . .



From stuffduff at cox.net  Sun Apr 27 10:49:43 2003
From: stuffduff at cox.net (stuffduff)
Date: Sun, 27 Apr 2003 14:49:43 GMT
Subject: Class problem: loosing my mind
References: <v0fmavkm9iknef7hit5n17d4m0qi6tpnia@4ax.com> <mailman.1051411241.22289.python-list@python.org> <diimavsd60audlnjm80m6dcan0e8b67hae@4ax.com> <3EAB5922.A4F8DDA5@alcyone.com>
Message-ID: <gaqnavcfl36jfa2m1k1gre429m5vts5od9@4ax.com>

I think I'm getting this now, and my zope example works.

Lessons learned:

Zope External Methods are not really python modules, so zope can read
a function from from a module without seeing a class definition
contained in the module.

Class pickles have a reference to the cass definition embedded in
them.  If you look at the pickle you can see the reference at the
start of the pickle.  Import errors that occur when unpickling a class
mean that pickle can't find the class definition that was used when
the object was pickled.

So when pickling a class, it's best for the class to be in a module
and the module in a package.  Then import the class definition into
both the pickling and unpickling modules, pickle sees class reference
the same way during both processes.

When package and module names are the same, import
constructs the reference with an unexpected behavior, 
which can be resolved by:

from {package} import {module}

Thanks to Andrew and Erik for their help!

Sean

'Onomonopia!' -- The Three Stooges





From roy at panix.com  Mon Apr 28 18:46:01 2003
From: roy at panix.com (Roy Smith)
Date: Mon, 28 Apr 2003 18:46:01 -0400
Subject: Arg names in class and function code - Newbie question
References: <64e13ba5.0304281424.52cb6b5d@posting.google.com>
Message-ID: <roy-F93739.18460128042003@reader1.panix.com>

newspost at lolmc.com (Lol) wrote:

> HI all,
> 	Could anyone let me know - do the names you give as arguments HAVE to be
> actual variable names?
> e.g def readFile(filename,dir):
> 	blah blah
> 	return blah
> or even class FileRead(filename,dir):
> 	def doSomething (self,this):
> 		yadda yadda
> 		return yadda
> 
> In the above examples do filename and dir have to be the name of the
> variables to pass as args to the class/functions OR are they descriptive
> so I can pass the real name of the file and directory?

I'm not completely sure I understand what you're asking, but if you're 
asking what I think you're asking, you're confused.  If not, them I'm 
confused.  Pressing on with the first assumption in mind....

When you say

def readFile (filename, dir):

the variables "filename" and "dir" are what's known in computer science 
circles as "formal parameters".  Basicly, you're saying that readFile is 
supposed to get called with two arguments.  This creates two variables 
for this function to use, "filename", which will get initialized to the 
value of the first argument, and "dir" which will get initialized with 
the value of the second.

Their names can be anything you like.  They need not (and usually don't) 
match the variable names used in the part of the program which calls the 
function.  You can call it with constants if you like:

def readFile (filename, dir):
   blah blah
   return blah

readFile ("eggs", "/spam/spam/spam/spam/spam/spam/spam")

As for the

class FileRead(filename,dir):
    def doSomething (self,this):
        yadda yadda
        return yadda

I'm really not sure what you're trying to do there.  As written, you've 
defined a class which inherits from a class who's name is in filename, 
and from another class who's name is in dir (at the time the class 
statement is executed).  This is legal, but I suspect it's not what you 
had in mind :-)

Perhaps you were thinking of an __init__() method which would take those 
two arguments?  In which case, you'd have something like:

class FileRead:
    def __init__ (self, filename, dir):
        self.file = open (dir + '/' + filename)

    def doSomething (self, this):
        yadda yadda
        return yadda

As a matter of style (and perhaps I'm really picking a nit here), I'd 
avoid "this" as a variable name.  In Python, "self" is the standard name 
for an object's handle to itself, and "this" plays that role in C++.  
There's nothing inherently wrong with using "this" as a Python varible 
name, but it just seems like a potential source of confusion.



From aleax at aleax.it  Tue Apr  1 03:53:02 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 08:53:02 GMT
Subject: Regular expression help
References: <2FGdnaUSmIhkghSjXTWc-g@speakeasy.net>
Message-ID: <Orcia.53206$i26.1631802@news2.tin.it>

<posted & mailed>

Stephen Boulet wrote:

> How do I use the sub function to replace my match with my match
> prepended by a newline?

Simplest is to enlose your RE pattern in parentheses (making it
a group) and then use a backreference \1 in the substitution:

resulting_string = re.sub(r'(x.y.x)', r'\n\1', original_string)


Alex




From BPettersen at NAREX.com  Thu Apr 24 11:35:26 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Thu, 24 Apr 2003 09:35:26 -0600
Subject: Question about accessing class-attributes.
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE0E9@admin56.narex.com>

> From: Alex Martelli [mailto:aleax at aleax.it] 
> 
> Duncan Booth wrote:
> 
> > Nikolai Kirsebom <nikolai.kirsebom.NOJUNK at siemens.no> wrote in
> > news:0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062 at 4ax.com:
> > 
> >> class A(object):
> >>      InstCount = 0
> >>      def __init__(self):
> >>           self.__class__.__dict__['InstCount'] += 1
> >> 
> >> class B(A):
> >>      InstCount = 0
> >> 
> >> class C(A):
> >>      InstCount = 0
> >> 
> >> Is the syntax used for incrementing the class-attribute for every
> >> instanciation 'the way' it should be done ?
> > 
> > It would be more normal just to write:
> > 
> > self.__class__.InstCount += 1
> 
> Yes, but do note that this may produce strange anomalies if some
> subclass erroneously forgets to rebind InstCount in its body...:
> 
> >>> class A(object):
> ...   InstCount = 0
> ...   def __init__(self):
> ...     self.__class__.InstCount += 1
> ...
> >>> class B(A): pass
> ...
> >>> a1=A()
> >>> a2=A()
> >>> b1=B()
> >>> b2=B()
> >>> B.InstCount
> 4
> >>> A.InstCount
> 2
> >>>
> 
> VERY unlikely to be what one wants, AND a rather subtle bug too.
> The OP's direct access to __class__.__dict__ would ensure an
> exception got raised upon instantiation of such an errant subclass.
> 
> 
> > Alternatively you could create a class method to do the 
> > increment, but
> > that is probably needlessly complex here.
> 
> Personally, I'd define a custom metaclass for this purpose.  Maybe
> I'm getting TOO familiar with those, but, it doesn't sound all that
> complex at all to me...:
[...]

I would too, but just for fun I thought "how about an abstract base
class?". Due to a 'thinko' I didn't consider that B wasn't calling its
superclass __init__, so I came up with:

  class AbstractBaseCounter(object):
      __InstCount = {}
      def InstCount(cls):
          return AbstractBaseCounter.__InstCount[cls]
      InstCount = classmethod(InstCount)
      
      def __init__(self):
          cls = self.__class__
          if cls == AbstractBaseCounter:
              raise TypeError("Attempt to instantiate abstract base
class")
          d = AbstractBaseCounter.__InstCount
          if cls in d:
              d[cls] += 1
          else:
              d[cls] = 1
          
  class A(AbstractBaseCounter):
      def __init__(self):
          AbstractBaseCounter.__init__(self)
          
  class B(A): pass
          
  a1=A()
  a2=A()
  b1=B()
  b2=B()
  print B.InstCount()

which happens to work...? I've verified that
AbstractBaseCounter.__init__ gets called for B(). Is this
correct/documented/guaranteed/works with MI/multiple args/old-style
classes, etc., etc.?

Of course, if you wanted to make InstCount a class atribute (so you
could say B.InstCount without the parentesis), you'd need a metaclass.

Python 2.3a2.

-- bjorn




From ark at research.att.com  Wed Apr  9 10:37:36 2003
From: ark at research.att.com (Andrew Koenig)
Date: Wed, 9 Apr 2003 14:37:36 GMT
Subject: Adding item in front of a list
References: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de>
Message-ID: <yu99k7e3wxcf.fsf@europa.research.att.com>

Thomas> Hi!
Thomas> I think it would be nice if
Thomas>  l=[2, 3, 4]
Thomas>  l.insert(1)

Thomas> would insert in front of the list, as opposed to append()
Thomas> which adds new elements at the end.

Thomas> l.insert(1) looks better than l.insert(-1,1)

So does l.insert(0, 1) :-)

I would have thought that after l.insert(-1, 1), l would be
[2, 3, 1, 4], but it doesn't work that way.

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From james at anchor.net.au  Wed Apr  9 14:47:32 2003
From: james at anchor.net.au (James Gregory)
Date: 09 Apr 2003 19:47:32 +0100
Subject: [Slightly OT]: More on ints and floats
In-Reply-To: <7h3wui4528k.fsf@pc150.maths.bris.ac.uk>
References: <fjss6b.1021.ln@boundary.tundraware.com>
	 <mailman.1049768113.16134.python-list@python.org>
	 <kauu6b.f2d1.ln@boundary.tundraware.com>
	 <7h3wui4528k.fsf@pc150.maths.bris.ac.uk>
Message-ID: <1049914052.3038.19.camel@pirate.bridge.anchor.net.au>

On Tue, 2003-04-08 at 18:27, Michael Hudson wrote:
> Tim Daneliuk <tundra at tundraware.com> writes:
> 
> > I am unclear, though, on whether a mathematician understands 1.0 to
> > mean 1.0000000... (a single number) or the _neighborhood_ around 1
> > with one digit precision.
> 
> I'm a mathematician, albeit not a numerical analyst, and I don't think
> considering a float as the interval of real numbers to which it is the
> closest floating point approximation is a particularly helpful point
> of view.

To my mind, 1.0 is a member of R, 1 is a member of Z. But I do crazy
things like think of "1/2" as the number that yields 1 when multiplied
by 2. Not 0.5. Nor the quantity of apple you get when sharing an apple
with one other person.

But it's context dependent, and that's one of the properties that having
all "numbers" as objects affords you. It means that my software can tell
if a particular number is a member of GF(5) (The Galois Field of 5
elements), and then it can infer that "1/2" actually means 3 (3 * 2 = 6,
6 % 5 = 1, as per the above definition).

What does 1/2 mean to you in Z? Does it mean the same thing in R?

(Z is the set of all integers, R is the set of all reals in case it
wasn't clear)

James.






From robin at jessikat.fsnet.co.uk  Fri Apr 11 13:58:08 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Fri, 11 Apr 2003 18:58:08 +0100
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
 <d116fbae.0304100704.68ef4833@posting.google.com>
 <c48$3e960655$516049d2$3498@nf1.news-service.com>
 <3e96efac$0$49107$e4fe514c@news.xs4all.nl>
 <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk>
 <3e96f481$0$49099$e4fe514c@news.xs4all.nl>
Message-ID: <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>

In article <3e96f481$0$49099$e4fe514c at news.xs4all.nl>, Irmen de Jong
<irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> writes
>Sending lots of ints to the server
>Time to send 5000 ints, 4.586772 (0.000917 per call)
>
is that 5000 connections with new sockets? Seems really fast to me.
>So you see, it's even faster when the server is running on another machine. 
>This can be explained because the pickling/unpickling is split across two 
>CPUs, when you run client+server on a single CPU they are fighting for cycles.
>
>--Irmen de Jong
>

-- 
Robin Becker



From tim.hochberg at ieee.org  Wed Apr  2 15:07:19 2003
From: tim.hochberg at ieee.org (Tim Hochberg)
Date: Wed, 02 Apr 2003 13:07:19 -0700
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n
 ...)
In-Reply-To: <3E8B4094.F61515B3@engcorp.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com> <mailman.1049298845.7655.python-list@python.org> <3E8B4094.F61515B3@engcorp.com>
Message-ID: <ZpHia.23$Sd.4@fed1read01>

Peter Hansen wrote:
> Jp Calderone wrote:
> 
>>  Now, in its current state, it is impossible for the Python interpreter to
>>know this.  What if range() is rebound?  So this can't be done now.  But, if
>>some of the talk about forbidding the shadowing of builtins bears fruit, it
>>might someday be possible for CPython to make this optimization, and all
>>without any ugly syntax hacks.
> 
> 
> Ouch!  *Forbidding* it?!  That would cripple some of my most sophisticated
> and effective unit and acceptance tests!  I dearly hope that "forbid" was
> an overly strong choice of verb.

I believe what was discussed was making shadowing of builtins from 
outside of a module forbidden. So:

    range = my_custom_range_function

would be OK. While,

    some_random_module.range = my_custom_range_function

would not.

> What would be wrong with "discouraging", or "making it clear to the 
> compiler when it happens" or something?

I think that was more or less the idea of forbidding rebinding the 
module -- the compiler can't find the rebinding when it occurs from 
outside of a module.

> Where is this discussion going on, and can I/need I participate to put
> forth my perspective on this.  

The discussion was going on on python-dev last month. It seemed to have 
died down last I lurked.

> I'm coming to *count* on this as one of
> the most effective solutions for certain kinds of testing.  (Specifically,
> by creating a mock file system by intercepting open() calls in my own
> method, to allow generating of various file system errors and situations
> and to increase performance of unit tests by avoiding the real file
> system.)
> 
> -Peter


-tim




From frithiof.jensen at removethis.ted.ericsson.dk  Fri Apr 11 06:47:04 2003
From: frithiof.jensen at removethis.ted.ericsson.dk (Frithiof Andreas Jensen)
Date: Fri, 11 Apr 2003 12:47:04 +0200
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com>
Message-ID: <b766fi$p3b$1@newstree.wise.edt.ericsson.se>

"kkennedy" <titanrebel at comcast.net> wrote in message
news:14bd1399.0304091426.1b08e005 at posting.google.com...
> I would like it to be smaller than
> the 7 MB download available from python.org, and be able to be
> installed transparent to the user.

On windows, Biig == Good!!

Users are so used to huge installs & downloads that they expect them and
they will be suspicious of anything that takes less than 30 minutes to
install - especially if it costs money. The "no-brain-install" is the right
idea, thoug.





From imbosol-1049850353 at aerojockey.com  Tue Apr  8 22:14:59 2003
From: imbosol-1049850353 at aerojockey.com (Carl Banks)
Date: Wed, 09 Apr 2003 02:14:59 GMT
Subject: usage of assert and -O
References: <mailman.1049844497.9845.python-list@python.org>
Message-ID: <DmLka.1720$d8.709@nwrdny02.gnilink.net>

Matthew Russell wrote:
> I would be interested to know what others think about usage of
> assert in context with optimised python code, in particular in
> releation to unittesting, as this is where i think it would be
> useful - if not critical to have one way of doing it (i.e not useing
> assert or -O [ one or the other ])



Well, theoretically, the point of assert is to test for conditions
that are supposed to be impossible.  (For example, suppose an argument
to a function must be positive.  Your program never calls it with a
negative value; therefore, it's impossible for the the argument to be
negative.)

The idea is that, since the conditions are impossible, you don't need
to waste time checking for them in production code.  However, when
debugging, you would like these checks so that you don't accidentally
do something that's supposed to be impossible.  You might think that
function never gets called with a negative value, but what if you
overlook something and suddenly it does get called with a negative
value?  Then your assertion catches it.

A lot of people think it's just a questionable optimization.  Just
checking if a value's positive isn't too expensive, and certainly not
worth removing in production code.

But I don't think it's that simple.  Sometimes you want to do a
not-so-cheap assertion (say, that a list is sorted).  In development,
you might want to make sure a list passed to a binary search function
is sorted, but you absolutely don't want to have this check enabled
production code, because checking that a list is sorted is O(n),
whereas a binary search is O(log n).  In this case, eliminating the
assertion test in production code is more than an optimization.

This is why I think assertions are useful in general.  I use them
frequently in C.

However, I don't use them as much in Python for a couple reasons.
First, Python catches a lot more errors than C, so it is not as
crucial to test for errors in Python.  Second, Python is an
interpreter, and it tests assertions by default.

This, IMO, is a mistake.  Python should only test assertions when a
debugging flag is present.  The reason is, I can't trust a user to
always run the optimized version.  Thus, I can't make expensive
assertions like ensuring a list is already sorted.  Yet this is when
assertions are most important.  Second, because assertions are tested
by default, a lot of people are tempted to use them in lieu of regular
error checking.


As for unit testing, I agree with Peter Hanson: it's a whole 'nother
realm of error testing.  I'm not an expert in unit testing, but it
seems to me unit testing serves a very different purpose than
assertions.  Unit tests try to enforce the robustness of a function by
throwing every possible sort of object at it, to make sure it fails
properly.  Assertions guard against unintended use of a function: they
make sure every possible sort of object doesn't get thrown at it.



-- 
CARL BANKS



From mwh at python.net  Wed Apr  9 12:55:10 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 9 Apr 2003 16:55:10 GMT
Subject: optimizing memory usage of large in memory structures
References: <aebfcc5a.0304090817.2c683c9e@posting.google.com>
Message-ID: <7h37ka31uhe.fsf@pc150.maths.bris.ac.uk>

theory79 at yahoo.com (Hugo Liu) writes:

> hi,
> 
> i'm hoping that someone out there has a solution to hack this problem.
> 
> working with corpus analysis and large knowledge bases, i often have
> to load large arrays into memory.  each array element is just a text
> string.  although the text is only 5mb, python consumes 40+ mb.  i
> would suspect that the overhead is from storing type info for each
> element.  if the strings were all fixed length, i'd just store them
> ordered and fixed width into a single string and use bisect, but they
> aren't fixed width.
> 
> does anyone know if there is a way to fake static typing for this
> situation

Don't think so.  You could try using arrays, but I doubt that would
save you much.

> or any other way to make this more efficient (without
> sacrificing operating system interoperability or having to write
> external C code?)  any suggestions or pointers would be much
> appreciated!

The only thing that leaps to mind is having one huge string and
instead of passing strings around, pass (start, end) tuples.  But
that may just be a crazy idea...

Cheers,
M.

-- 
  Hey, if I thought I was wrong, I'd change my mind.  :)
                                    -- Grant Edwards, comp.lang.python



From g.szolderits at gmx.at  Wed Apr 16 09:04:35 2003
From: g.szolderits at gmx.at (=?ISO-8859-15?Q?G=FCnter?= Szolderits)
Date: Wed, 16 Apr 2003 15:04:35 +0200
Subject: PyQT and Eric
References: <b7is7g$i1d$1@rex.ip-plus.net>
Message-ID: <3e9d548c$0$32618$91cee783@newsreader02.highway.telekom.at>

Daniele Trovato wrote:

> My problem is, when I'm trying to install Eric with the following command
> "python install.py", the shell returns me an error: "Sorry, Please install
> PyQT"  but I have PyQT already installed.
> Does anybody now where the problem is or how i can solve the problem ?

Install PyQt sources and recompile. Fixed the problem for me.

G?nter



From aahz at pythoncraft.com  Wed Apr 23 10:13:09 2003
From: aahz at pythoncraft.com (Aahz)
Date: 23 Apr 2003 10:13:09 -0400
Subject: ASCII database (was Re: Should I prefer an external database)
References: <u3ckbymri.fsf@mail.afm.dtu.dk>
Message-ID: <b8671l$d2s$1@panix2.panix.com>

In article <u3ckbymri.fsf at mail.afm.dtu.dk>,
Brian Elmegaard  <brian at rk-speed-rugby.dk> wrote:
>
>I am working on a script for organizing a program for a conference. I
>started out using a connection to mySQL, but then I thought that it
>would be easier to keep it within python. 

This seems like a good place to mention that I'm currently rolling my
own database.  I'm doing it because I want to keep the DB online in my
shell account, and it's a PITA using anything other than a simple text
editor to update bits within it.  So I'm using a simple RFC-822 layout
for the fields with '%%\n' as the record separator; fields are permitted
to have multiple instances.  There are no required fields in the design;
each record is free-form.

The fun part was implementing the search engine.  I wanted to implement
something usable, so I allow both includes and excludes.  Excludes are
always ORed; includes are ANDed; search terms are separated by commas.
An include can have an embedded pipe ('|'), which turns it into a list
of ORed terms.  Excludes are prepended with an exclamation point ('!').
(Yes, I don't like the visual confusion, but those are standard chars
and the location makes the intent clear.)

So here's my filter function:

class Excluded(StandardError):
    pass

class Found(StandardError):
    pass

class NotFound(StandardError):
    pass

# Includes by default are ANDed; excludes are always ORed
# Use "|" to OR includes; "!" indicates exclude
def filter(db, FilterString=None, fields=None, FuzzyMatch=True):
    includes, excludes = _buildFilters(FilterString)
    if fields is not None:
        fields = re_comma_field.split(fields)
        for i in range(len(fields)):
            fields[i] = fields[i].lower()
    if FuzzyMatch:
        match = _fuzzy_match
    else:
        match = _exact_match
    result = []
    for record in db:
        if fields is None:
            curr_fields = record.keys()
        else:
            curr_fields = []
            for field in fields:
                if field in record:
                    curr_fields.append(field)
        try:
            for field in curr_fields:
                for item in record[field]:
                    item = item.lower()
                    for filter in excludes:
                        if match(item, filter):
                            raise Excluded
        except Excluded:
            continue
        try:
            for ORlist in includes:
                try:
                    for filter in ORlist:
                        for field in curr_fields:
                            for item in record[field]:
                                if item.lower().find(filter) >= 0:
                                    raise Found
                    else:
                        raise NotFound
                except Found:
                    continue
        except NotFound:
            continue
        else:
            result.append(record)
    return result



def _fuzzy_match(string, filter):
    return string.find(filter) >= 0


def _exact_match(string, filter):
    return string == filter



re_filters = re.compile(r', *')
NOT = '!'
OR = '|'

# Each set of filter words is separated by comma
def _buildFilters(FilterString):
    if FilterString is None:
        return [], []
    else:
        FilterString = FilterString.lower()
        FilterList = re_filters.split(FilterString)
        includes, excludes = [], []
        for filter in FilterList:
            if filter[0] == NOT:
                excludes.append(filter[1:])
            else:
                includes.append(filter.split(OR))
        return includes, excludes
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From sdhyok at yahoo.com  Wed Apr 16 14:07:31 2003
From: sdhyok at yahoo.com (sdhyok)
Date: 16 Apr 2003 11:07:31 -0700
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net> <420ced91.0304152121.2af84a33@posting.google.com> <3E9D509F.3FC4073@engcorp.com>
Message-ID: <420ced91.0304161007.5688bf0d@posting.google.com>

Peter Hansen <peter at engcorp.com> wrote in message news:<3E9D509F.3FC4073 at engcorp.com>...
> How wide a range?  It is much easier to answer *specific*
> questions, than very broad ones as you are asking... narrow
> the scope a little and help us out.
> 
> -Peter

Thanks for your reply, Peter.
My area is hydrology, so the maximum range of sampling period 
is about several decades.
For instance, the period from 1930 to now.
In addition, not only year/month/day but also hour/min/sec should
be specified. So, integer is not enough, 
Real number expression is required.
As I mentioned at the first email,
the datenum() function in MATLAB is exactly what I need in Python.



From abelikov72 at hotmail.com  Wed Apr 30 11:26:29 2003
From: abelikov72 at hotmail.com (Afanasiy)
Date: Wed, 30 Apr 2003 15:26:29 GMT
Subject: fcntl in windows
Message-ID: <k2rvav06nhjgtr578fa5ostftfl69rethp@4ax.com>

Is there an fcntl in Windows? A stub, a wrapper, etc?

Such support is entirely platform-transparent in Perl.



From ckea25d02 at sneakemail.com  Tue Apr  1 18:59:42 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 11:59:42 +1200
Subject: When Python outruns C++ ...
In-Reply-To: <slrnb8j5og.o6a.chasm@galileo.rift>
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift> <3E890322.4B003C9C@engcorp.com> <slrnb8j5og.o6a.chasm@galileo.rift>
Message-ID: <b6d8m2$47nsc$1@ID-169208.news.dfncis.de>

Julian Tibble wrote:
> setdefault() actually makes the running time worse, however catching
> the KeyError did not.

This kind of thing has been observed before -- I believe
someone noticed that using get() on a dict was slower than
indexing and catching the exception.

I think it's because the index-and-catch method avoids
an attribute lookup and function call, which are somewhat
expensive in Python.

So the moral is that get() and setdefault() are for
conciseness and readability, not speed!

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ramen at lackingtalent.com  Tue Apr 29 02:34:30 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 29 Apr 2003 06:34:30 -0000
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it>
Message-ID: <slrnbas791.j6l.ramen@lackingtalent.com>

In article <UXara.22428$K35.745468 at news2.tin.it>, Alex Martelli wrote:
> "How many items does XX[a:b] have?"
> 
> With Python's design (following Koenig's principle), the normal
> answer (when a and b are both within XX'x boundaries) is:
>         b - a
> 
> If the last-bound had been deemed to be included, it would have been:
>         b - a + 1
> 
> Now, that innocuous-looking "+ 1" there is the cause of untold misery
> in programming -- the root of most "off-by-one" errors.  By removing
> the need for it, Koenig's principle makes you more productive.

Interesting! I've never heard that before. Thanks for the explanation.

I've definitely been bitten by this when trying to treat a 1D array as a 2D
array by multiplying rowcounts. It's even more confusing when it's not + 1
anymore, it's + something else... or is it minus? Ahhh!

In recovery,
Dave



From peter at engcorp.com  Wed Apr  9 14:40:11 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 09 Apr 2003 14:40:11 -0400
Subject: signals and system calls
References: <mailman.1049908218.16228.python-list@python.org>
Message-ID: <3E94690B.586BC8C8@engcorp.com>

Michael Stenner wrote:
> 
> Just to be clear, I'm not trying to "use" signals here.  They exist,
> they WILL be sent (when a system shuts down, for example), and I'm
> just trying to tolerate them in a way that doesn't lose data.
> 
> So, the bottom line is:  what do people recommend?

Don't.

That is, don't try to be incredibly tolerant even in the face of
system shutdown and signals that kill your program.

Instead, be tolerant of partially written data structures, half-completed
operations, and so forth.

There is *no* way you can protect against a power outage, and no
way you can protect against something like "kill -9" anyway, so
you are trying to achieve an unattainable level of perfection if
you insist that under other circumstances you *must not* ever
lose data.

At least with this view, you'll find it much easier to deal with
the particular nastiness of signals...  just my view.

-Peter



From duncan at NOSPAMrcp.co.uk  Tue Apr 29 04:18:03 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 29 Apr 2003 08:18:03 +0000 (UTC)
Subject: More rants on a bad sort
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com> <7ULqa.9804$3M4.264757@news1.tin.it> <g3qdncyq1Zx2FzGjXTWcrg@comcast.com> <mailman.1051497463.22158.python-list@python.org> <Xns936B6B217EB86duncanrcpcouk@127.0.0.1> <mailman.1051547264.16548.python-list@python.org>
Message-ID: <Xns936C5D9464E86duncanrcpcouk@127.0.0.1>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> wrote in 
news:mailman.1051547264.16548.python-list at python.org:

> Given all that, is it POSSIBLE to construct a list containing a string,
> int, and complex, that will not crash on a sort?  I don't know!  I
> couldn't find one--and I THINK the relative order of complex and strings
> precludes it... but actually proving this one way or another is
> decidedly non-obvious.

Given a, b, and c, if a < b < c then the sorting may assume that a < c. In 
the case with unicode and str comparisons you can have a < b < c but a and 
c cannot be compared. To do the same with str, int, complex values you 
would have to have the uncompareable values as a and c, but as was 
mentioned elsewhere in this thread, the numeric types are all fudged to 
have the same ordering against other types, so you can never have an 
ordering a < b < c where b is not a numeric type and a and c both are.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From staschuk at telusplanet.net  Sat Apr 26 15:02:59 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 26 Apr 2003 13:02:59 -0600
Subject: Some Python Code
In-Reply-To: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com>; from sross@connectmail.carleton.ca on Sat, Apr 26, 2003 at 01:16:39PM -0400
References: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com>
Message-ID: <20030426130259.D30093@tibia.amotlpaa.bogus>

Quoth Sean Ross:
> [...] I don't really want to do any
> maintenance on these projects, so I'll not be bothering to put them up in
> some place like sourceforge or pypi. However, the code may be useful to
> someone else, perhaps as a starting point for their own projects, so I'd
> like to make it available. [...]

You could submit them to Useless Python:
    <http://www.uselesspython.com/>

-- 
Steven Taschuk                               staschuk at telusplanet.net
"[T]rue greatness is when your name is like ampere, watt, and fourier
 -- when it's spelled with a lower case letter."      -- R.W. Hamming




From lafou at wanadoo.fr  Tue Apr  8 04:57:23 2003
From: lafou at wanadoo.fr (Philippe =?ISO-8859-1?Q?Lafoucri=E8re?=)
Date: Tue, 08 Apr 2003 10:57:23 +0200
Subject: ANN: SQLObject 0.3
References: <mailman.1049776336.2246.python-list@python.org>
Message-ID: <b6u2tj$9ea$1@news-reader12.wanadoo.fr>

Ian Bicking wrote:

> SQLObject 0.3 has been release:
>   http://sqlobject.org
> 
> Changes
> -------
> 
> * Automatic table creation based off class.
> * Create classes from database.
> * Dynamically add and remove columns to live classes and databases.
> * anydbm-based backend (queryable database without SQL).
> * Better caching.
> * Some bugs fixed.
> 
> What is it?
> -----------
> 
> SQLObject is an object-relational mapper.  It allows you to translate
> RDBMS table rows into Python objects, and manipulate those objects to
> transparently manipulate the database.
> 
> SQLObject emphasizes convenience.  It's easy to learn and easy to use,
> and doesn't require too much typing or use of any tools outside of your
> editor and your Python source.
> 
> 


It is quite similar to the Twisted Matrix Row object !?



From vlindberg at verio.net  Mon Apr  7 09:40:23 2003
From: vlindberg at verio.net (VanL)
Date: Mon, 07 Apr 2003 07:40:23 -0600
Subject: Calling python from a C function that is embedded in python?
Message-ID: <3E917FC7.40907@verio.net>

Hello,

I have a python framework that I am using, along with a couple C 
libraries that are wrapped using swig.

Is there a way to make the functions that I have written in python 
callable by one of the wrapped functions?

I don't have to stay with swig, pyrex would be fine too. (Or almost any 
other similar tool, those are the only two that I know.)

In short:  I want python to be able to call wrapped functions like they 
are python functions, and a wrapped C library to call wrapped python 
functions like they were in C.

Is there any way to do this?

Thanks,

VanL





From adalke at mindspring.com  Fri Apr 18 21:03:44 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Fri, 18 Apr 2003 19:03:44 -0600
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com>
Message-ID: <b7q72i$nvh$1@slb6.atl.mindspring.net>

[Aahz, on Dijkstra's BASIC quote]

Peter Hansen:
> Perhaps only once
> she advances beyond the student stage, and has some years of actual
> _practice_, will she discard the early mental mutilation of BASIC
> and find her way to Good Style.

I too learned programming starting with BASIC.  I started with
the manual which came with my TI 99-4/A some 20+ years ago,
and didn't give BASIC up until 12 years ago.  I think my last big
BASIC program was a X-ray crystallography program, with a
GUI to lay out and assign the lines, etc.

One thing to note, which Dijkstra didn't predict, is that I stopped
using a 'traditional' BASIC back in '86, when I got a copy of
QuickBasic for my birthday.  No line numbers, real functions,
better control structures, compiled code - while editing in the
IDE, so the compile-link-run cycle is really fast.  Quite nice,
and after that and Turbo Pascal and Turbo C, going to K&R
C on a unix machine was like a step back into darkness.

Anyway, so it is possible to get Good Style even under
BASIC.  (Shhh!  And I think my BASIC background is
one reason I like Python so much. ;)

                    Andrew
                    dalke at dalkescientific.com





From duncan at NOSPAMrcp.co.uk  Wed Apr 23 07:44:54 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 23 Apr 2003 11:44:54 +0000 (UTC)
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org>
Message-ID: <Xns936680CEC5FF4duncanrcpcouk@127.0.0.1>

Gr?goire Dooms <dooms at info.ucl.ac.be> wrote in 
news:mailman.1051093416.23728.python-list at python.org:

> I need a lower and higher integer for initializing lower and higher
> bounds in an algorithm.  I know integers have no limits(maximum and
> minimum) in python and infimum and supremum would do the job too.
> 
> I empirically discovered (Python 2.2.2) that None could be considered an
> infimum and any string could be considered a supremum.
> e.g "a">int(x) is always true and None<int(x) is always false
> 
> But this fact is not stated in the language reference. 
> Citing http://www.python.org/doc/current/ref/comparisons.html : 
> "The objects need not have the same type. If both are numbers, they are
> converted to a common type. Otherwise, objects of different types always
> compare unequal, and are ordered consistently but arbitrarily."
> 
> Could/should that order be hardcoded in the language ?

Perhaps the wording should be changed 'ordered consistently but arbitrarily 
within a single instance of the interpreter'. Sorry, there are no 
guarantees what relative ordering objects of different types will have on 
different systems or even on the same system but different processes.
Builtin types may tend to keep the same relative ordering, but even that 
has changed between Python versions. (On my copy of Python 2.2.2, I just 
got "a">1 true, None<1 true which is not what you said you got.)

Perhaps you could modify your algorithm to not require pre-initialising 
with upper and lower bounds? For example, if your algorithm generates a 
series of values and you need to know the highest and lowest that are 
generated, then keeping a list of the generated values and using the 
builtin min and max functions on the list is likely to be faster than doing 
the comparisons by hand until you get to a very large set of values.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From exarkun at intarweb.us  Tue Apr  1 15:27:49 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 1 Apr 2003 15:27:49 -0500
Subject: When Python outruns C++ ...
In-Reply-To: <698f09f8.0304011207.5552813c@posting.google.com>
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift> <mailman.1049181423.2644.python-list@python.org> <698f09f8.0304011207.5552813c@posting.google.com>
Message-ID: <20030401202749.GA950@meson.dyndns.org>

On Tue, Apr 01, 2003 at 12:07:57PM -0800, Jeremy Fincher wrote:
> Jp Calderone <exarkun at intarweb.us> wrote in message news:<mailman.1049181423.2644.python-list at python.org>...
> >   It looks like this could easily be improved with the application of a
> > slightly smarter searching algorithm, such as Rabin-Karp search or
> > Boyer-Moore find.
> 
> These searches depend on a fixed needle to search for in the haystack.
>  They can't be applied to a regular expression search.

  Quite true.  However, a regular expression is by no means the only way
to break a string up into "words".  For this simple task, I think either of
the above mentioned algorithms will work.  For more complex ones, involving
really complex regular expressions, perhaps not.

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 12 days, 16:00, 6 users, load average: 0.05, 0.14, 0.14




From donnal at donnal.net  Fri Apr 18 05:26:35 2003
From: donnal at donnal.net (Donnal Walter)
Date: Fri, 18 Apr 2003 04:26:35 -0500
Subject: wxPython: Intercepting keyboard events from multiple widgets?
In-Reply-To: <b7nu9u$2pbg0$1@ID-169208.news.dfncis.de>
References: <b7le9j$266le$1@ID-169208.news.dfncis.de> <mailman.1050572372.16930.python-list@python.org> <b7nu9u$2pbg0$1@ID-169208.news.dfncis.de>
Message-ID: <b7og9b$98j$1@main.gmane.org>

Greg Ewing (using news.cis.dfn.de) wrote:
> I have another problem now. I've set up an Edit
> menu with wxID_CUT etc. in the appropriate places.
> According to the docs wxTextCtrl is supposed to
> come with default event bindings for these, but
> they don't seem to do anything.
> 
> And if I try to
> use EVT_MENU to explicitly attach a handler for
> one of them to a wxTextCtrl, it never gets called,
> although if I do that for the parent window of the
> wxTextCtrl (which is a wxNotebook) it does get called.
> It's as though the menu event is not being delivered to the
> control itself, but to its parent.
> 
> Is there anything special I have to do to get menu
> events delivered to text controls?

I don't know the answer to this specific question, but if one is not 
forthcoming in this forum, you might try asking it on 
germane.comp.python.wxpython or the mailing list at:
wxPython-users at lists.wxwindows.org

In the meantime, try right-clicking on the wxTextCtrl and see if that 
gives you the functionality you are looking for. (I know you want the 
frame menubar to do this, but this might help for now.)

Cheers,
Donnal Walter
Arkansas Children's Hospital
http://mindwrapper.org







From robin at jessikat.fsnet.co.uk  Wed Apr  2 18:07:14 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Thu, 3 Apr 2003 00:07:14 +0100
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org>
 <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com>
 <wBIia.2000$y3.310439@news010.worldonline.dk> <b6fmcq$dh0$1@panix2.panix.com>
Message-ID: <miXIvJAi02i+Ew2O@jessikat.fsnet.co.uk>

In article <b6fmcq$dh0$1 at panix2.panix.com>, Roy Smith <roy at panix.com>
writes


>People have been commiting crimes against good design in the name of
>efficiency since punchcard days.  When I first learned C, it was on a
>pdp-11, where the compiler could put three variables per function in
>fast registers.  My major goal in refactoring (although we didn't call
>it that in those days) was to design functions that made extensive use
>of no more than three integers or pointers.
>
>Just a couple of weeks ago, I let the efficiency bug bite me and spent
>two days totally tearing apart a C++ class to make it use a different
>set of STL containers.  The savings?  I turned a rarely-used linear
>operation (where N <= 24 most of the time) into a constant-time
>operation.  Totally meaningless to overall performance, but once the
>bug bit, it was hard to stop myself :-)
>
>The dark side of the force is seductive.  It tries to make us
>efficient, even at the cost of clarity and maintainability.  In that
>way lies bit-twiddling and assembler language and soldering irons.  We
>must be diligent in our fight against it.

The Python community sometimes has the efficiency bug. As examples the
desire to optimize global access (ie forbidding 'from module import *'
in inner scopes) and more recently the proposal to inhibit shadowing of
builtin functions in certain circumstances. It seems people never learn
to make regularity a priority. Nobody uses those special cases anyhow
etc etc

Interestingly attempts to get an integer for loop (without range/xrange)
are met with extreme disapproval by the language gurus and the fastest
looping constructs of all map, filter et al are considered bad form.

-- 
Robin Becker



From usenet at rob.dqd.com  Thu Apr 10 13:51:46 2003
From: usenet at rob.dqd.com (Rob Mayoff)
Date: 10 Apr 2003 10:51:46 -0700
Subject: isearch, LEAP and patents  was: Re: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org> <yfsr88fkppr.fsf@black132.ex.ac.uk> <eaeka.1333$1h6.8933@news-server.bigpond.net.au> <yfs1y0el9xc.fsf_-_@black132.ex.ac.uk>
Message-ID: <e0caaad7.0304100951.57cb4572@posting.google.com>

Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfs1y0el9xc.fsf_-_ at black132.ex.ac.uk>...
> Maybe someone on comp.emacs can fill us in on isearch's origins and whether
> raskin holds any patents that are relevant to emacs (he certainly seems to
> either hold or have applied for patents concerning THE[1] and from what I gather
> from his webpage pretty much everything that would seem original enough to get
> patented is present in emacs)?

The LEAP patent (US patent 5,019,806) does not claim general
incremental search.  It claims an incremental search mechanism where
one key enables a forward search mode while the key is pressed, and a
second key enables a backward search mode while the second key is
pressed.  In other words, it claims quasi-modal incremental search. 
Emacs doesn't have that.



From fredrik at pythonware.com  Thu Apr 10 07:47:22 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Thu, 10 Apr 2003 13:47:22 +0200
Subject: A question about PIL attributes
References: <001401c2febb$5ec42ed0$a200000a@mattg>
Message-ID: <b73lje$nr1$1@main.gmane.org>

"Matt G" wrote:

> Hi.  I'm using the Python Image Library  to do some image manipulations.  I like that the PIL
> objects have attributes like format, size, info (dictionary) etc.  My problem is...when I
> apply some filter, say ImageChops.invert, or Image.transpose, it looses
> the addributes.  For example
>
> if DEBUG: print "Before transpose, format is: " + str(pil.format)
> pil = pil.transpose(Image.FLIP_LEFT_RIGHT)
> if DEBUG: print "After transpose, format is: " + str(pil.format)
>
> the first output prints "JPEG" or "GIF" or whatever.
>
> after the transpose, the output is "None"

the "format" attribute is only set on image objects that are
read directly from a file; once you start processing the image,
the format is no longer defined.

if you need to keep track of the original format, just copy the
attribute to a separate variable:

    format = pil.format
    pil = pil.transport(...)
    ...
    pil.save(outfile, format=format)

</F>







From nobody at nowhere.org  Sun Apr 20 10:48:45 2003
From: nobody at nowhere.org (.)
Date: Sun, 20 Apr 2003 14:48:45 GMT
Subject: telnet
Message-ID: <hryoa.600205$sj7.23974187@Flipper>

how can I assign a port with code like this?

import telnetlib

HOST = "localhost"

tn = telnetlib.Telnet(HOST)

print tn.read_all()




From dl_un-spam_rubin at yahoo.com  Tue Apr  8 18:51:22 2003
From: dl_un-spam_rubin at yahoo.com (Daniel)
Date: Tue, 8 Apr 2003 15:51:22 -0700
Subject: creating new clobs in DCOracle2
Message-ID: <b6vjpb$45h$1@news.Stanford.EDU>

I'm using DCOracle2 to access oracle and want to insert data into a clob
column.  The DCOracle2 docs say that you need to first insert an empty clob
using EMPTY_CLOB() and then query the table to grab it, and finally, write
to it using the write method.
 In JDBC you can do this:

tempClob = CLOB.createTemporary(conn, true, CLOB.DURATION_SESSION);

and then you can write data into the tempClob, and then insert that into the
table. Can't you do something like this in DCOracle2?

Thanks,
Daniel





From maxume at yahoo.com  Sat Apr  5 17:53:27 2003
From: maxume at yahoo.com (Max Erickson)
Date: Sat, 05 Apr 2003 22:53:27 -0000
Subject: Anyone know of a webmail(yahoo) <--> localhost mail sync app in Python
References: <6e07b825.0304051307.3b301dd7@posting.google.com>
Message-ID: <Xns9354B61B942DEnotrealmail@216.168.3.44>

yaipa at yahoo.com (yaipa h.) wrote in news:6e07b825.0304051307.3b301dd7
@posting.google.com:

> All,
> 
> I am looking for an application in Python that will allow me 
> to synchronize my free yahoo webmail account with a local mail 
> reader for off-line mail access. Using Outlook at this point is
> a don't care.
> 
> I have found some stuff in perl/php, but nothing in python.  
> Some apps that look like they will work are squirrelmail, 
> fetchyahoo.pl and Open WebMail.  Ultimately, I will need to 
> support this application company wide and would like it not to 
> be in perl/php.
> 
> -Alan Haffner
> 

It's not python, but you might want to take a look at yahoopops. It runs in 
the taskbar on windows, works pretty well. yahoopops.sourceforge.net



From ckea25d02 at sneakemail.com  Mon Apr 14 02:22:39 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 14 Apr 2003 18:22:39 +1200
Subject: Speed of string += string
In-Reply-To: <3E98D281.8C0CE87D@easystreet.com>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <3E98D281.8C0CE87D@easystreet.com>
Message-ID: <b7djfa$deo4s$3@ID-169208.news.dfncis.de>

achrist at easystreet.com wrote:
> Now, which is preferred:
> 
> 	if astring == "":
> 
> or 
> 
> 	if len(astring) == 0:

The first ought to be faster, since it avoids a
global name lookup and a Python function call,
both of which are fairly expensive.

But better still would be simply

   if astring:
     ...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From tim.one at comcast.net  Wed Apr 23 23:17:15 2003
From: tim.one at comcast.net (Tim Peters)
Date: Wed, 23 Apr 2003 23:17:15 -0400
Subject: Why functional Python matters
In-Reply-To: <b86lc3$tnd$1@plover.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCIECDEEAB.tim.one@comcast.net>

[attribution lost]
>>> But he regrets the inclusion of a crippled lambda, rather than the
>>> inclusion of support for anonymous functions (IIRC).

[Courageous]
>> Anyway, I guess what I'm saying is that I suspect that Guido
>> didn't have any good alternatives.

[Mark Jason Dominus]
> I don't believe that's correct.

In the Python world, you're missing the context:  lambda was in version 1.0
of Python.  They're "crippled" not so much because they didn't support
closures (they do in current Pythons, BTW), but because they're
syntactically limited to a single expression, and Python is not an
expression language (Python statements are not Python expressions, so no
matter how perverse you try to be, you can't write a Python lambda that
contains a Python statement -- what you can do inside a Python lambda is use
a crippled subset of the language).

That said, Guido regrets lambdas anyway <wink>.

> At an early Open Source Conference, perhaps around 1998 or so, I was
> present at a conversation between Guido and Tom Christiansen.  Tom asked
> Guido why Python didn't have closures, and Guido's reply (and I think
> this is an exact quote) was "Closures aren't important."
>
> I remember it because I was so astonished by the obtuseness of the
> answer.

I expect he still believes that.  I do too:  in Python, closures aren't
important.  Python eventually grew closures, and they're *nice* to have, but
it's always been possible to use Python's flexible classes to get the same
abilities, and it's even always been possible to get "real closures" via
abuse of Python's default-argument mechanism:

    def make_adder(n):
        def adder(i, n=n):
            return n+i
        return adder

    add5 = make_adder(5)
    add3 = make_adder(3)
    print add5(1), add3(1)

has printed 6 4 for eons.  This has too, replacing the function with a class
of the same name:

    class make_adder:
        def __init__(self, n):
            self.n = n

        def __call__(self, i):
            return self.n+i

It often proves much better to carry state around in explicit objects
instead of in hard to introspect, or to modify (unless you planned for that
from the start), anonymous closures.  In the above, I can easily introspect
to see what an instance of the make_adder class adds, and even change it if
I need to, and the class's author didn't need to anticipate either of those.
With the by-hand default-arg closure, introspection is still possible but
much harder, while mutation of what it does requires deep (and formally
undefined) wizardy.  With the current "real closure" version:

    def make_adder(n):
        return lambda i: n+i

introspection and mutation are both too painful to contemplate (of course
that inability to inspect or change is also a feature in some parts of some
apps -- usually brittle ones <0.9 wink>).

Closures can be much more convenient in the small (I think make_adder is a
good example of that); in the large, piles of nested closures are often more
a maintenance nightmare than a blessing, and especially when they're used to
maintain state more naturally modeled via classes.  Python doesn't need
closures -- they're just nice to have.

when-you-don't-have-a-hammer-only-nails-look-like-nails-ly y'rs  - tim





From rs55862 at NOSPAMyahoo.com  Tue Apr  8 06:15:14 2003
From: rs55862 at NOSPAMyahoo.com (Rami Saarinen)
Date: Tue, 08 Apr 2003 13:15:14 +0300
Subject: Making python module using Borland C++ 5
Message-ID: <b6u7fp$d2c$1@news.cc.tut.fi>

I am trying to build a simple module for python (2.1.1) using the 
Borland C++ Builder 5. I have managed to go as far as:

* used the coff2omf program to convert the python21.lib to omf format 
(guess how log it took to get past this..)

* the module is a crude and simple glue (for testing purposes) between 
python and a rule-engine (CLIPS) and looks like this:

[...]
PyMethodDef clipsMethods[] = {
   {"run",      pyclips_run,         METH_VARARGS},
   {"getMove",  pyclips_getMove,     METH_VARARGS},
   {"reset",    pyclips_reset,       METH_VARARGS},
   {NULL,       NULL,              NULL}
};

void _declspec(dllexport) initpyclips(void) {
   (void) Py_InitModule("pyclips", clipsMethods);
   InitializeEnvironment();
   Load("th-ng.clp");
   Reset();
}


(th-ng.clp is a rule-based version of tower-of-hanoi solver).

Everything goes fine (after I removed the _DEBUG flag) and the dll 
builds without errors. However, when I actually try to run the python 
code that uses the module I get the following error message:

File "py-tst.py", line 1, in ?
   import pyclips
ImportError: dynamic module does not define init function (initpyclips)

Any ideas? Or am I totally lost and in error here?

I have been able to build and use the module in Linux without any 
trouble at all. I also managed to make the module in Windows using VC++ 
(that required a bit of manual helping) and it worked fine. However, I 
may be forced to get it done using Borland so ...

Well, I have been learning a lot about making static and dynamic 
libraries on Windows (as well as using these graphical thingies), a 
topic I had no idea about few weeks ago.. 8)

Thanks in advance.

--
Rami Saarinen

Please do remove the obvious bit from the email
if you wish to reply to me personally.




From joze_storme at yahoo.com  Sun Apr 13 15:10:56 2003
From: joze_storme at yahoo.com (Joze)
Date: 13 Apr 2003 12:10:56 -0700
Subject: Creating a PyObject* from a C double*
References: <20cb355e.0304122315.c068128@posting.google.com> <YJ9ma.6928$LB6.196819@news1.tin.it>
Message-ID: <20cb355e.0304131110.55240129@posting.google.com>

Exactly what i was looking for. Thanks a lot !!

Is there any way to avoid copying the array element by element and
simply pass a pointer to it?

Thanks again.
Joe.

Alex Martelli <aleax at aleax.it> wrote in message news:<YJ9ma.6928$LB6.196819 at news1.tin.it>...
> Joze wrote:
> 
> > Hi,
> > 
> > As part of a Python extension, i have an array of doubles created in a
> > C function, which should return a corresponding PyObject* to Python.
> > The array is meant to be used by Numeric/numarray, but apparently, the
> > C function cannot directly return a PyArrayObject*.
> 
> It can _after casting the pointer_ -- is that what's blocking you?
> 
> 
> > I have tried PyArray_FromDims(), PyArray_FromDimsAndData() but i can't
> 
> Those should only be used for static arrays -- ones that never will
> be deallocated; it's unlikely that's what you want.
> 
> > seem to get it right. Can anyone tell me how this works? I'd be most
> > grateful.
> 
> Here's a tiny working example just hacked together:
> 
> 
> #include <Python.h>
> #include <Numeric/arrayobject.h>
> 
> static PyObject*
> tiny(PyObject* self, PyObject* args)
> {
>     PyArrayObject* parr;
>     int dims[1];
>     int i;
>     double xx[] = {1.0, 2.3, 4.5, 6.7, 8.9};
>     double *data;
> 
>     dims[0] = 5;
>     parr = (PyArrayObject*) PyArray_FromDims(1, dims, 'd');
>     if(!parr) return 0;
> 
>     data = (double*) parr->data;
>     for(i=0; i<5; ++i)
>         data[i] = xx[i];
> 
>     return PyArray_Return(parr);
> }
> 
> static PyMethodDef tinyMethods[] = {
>     {"tiny", tiny, METH_VARARGS, "example"},
>     {0}
> };
> 
> void
> inittiny(void)
> {
>     Py_InitModule("tiny", tinyMethods);
>     import_array();
> }
> 
> anf after a python setup.py install with the obvious setup.py,
> 
> [alex at lancelot pronu]$ python2.2 -c 'import tiny; print tiny.tiny()'
> array([1.0, 2.2999999999999998, 4.5, 6.7000000000000002, 
> 8.9000000000000004], 'd')
> [alex at lancelot pronu]$
> 
> 
> Alex



From aleax at aleax.it  Tue Apr 22 08:47:19 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 12:47:19 GMT
Subject: coverting list back to string
References: <3cf72d52.0304211132.799d8afb@posting.google.com> <878yu33a16.fsf@vercingetorix.vercingetorix>
Message-ID: <rRapa.44440$LB6.1000250@news1.tin.it>

Dialtone wrote:

> regnivon at netscape.net (scn) writes:
> 
>> my quest:  i know that output.write(remove) is not correct.  you
>> cannot use the 'write' method on a list.  is there a better way to
>> convert the list to a string after the first element of the list has
>> been removed and then output the strings to a file?  i'm a newbie
>> programmer and need guru help. thank you.
> 
> Sure:
> 
> input = file('c:/python22/programs/linetest.txt', 'r')
> output = file('c:/python22/programs/output.txt', 'w')
> 
> for line in input:
>     remove = line.split()
>     remove = remove[1:]
>     output.write(''.join(remove))

Even better might be to change the last two lines into:

    output.writelines(remove[1:])

(the writelines method accepts ANY sequence of strings, not
just lines).  If the lines are huge, a tiny but neat
optimization can further be (avoiding constructing the
list needed for the slicing):

for line in input:
    fields = iter(line.split())
    fields.next()
    output.writelines(fields)

Note however that all of these solutions remove whitespace
from the file -- including line-ends &c -- since split
splits on whitespace but the whitespace isn't reinserted
afterwards (neither with the explicit join, nor by the
implicit ones performed by the writelines method).  I am
not sure whether this behavior is what the original poster
wants, since I don't have his post on my feed -- it sure
does seem to be what he's asking for, given this partial
quote that I see here, but sometimes what people ask for is
not exactly what they want, so it may be worth double checking.

> input.close()
> output.close()

Leaving these in since they're helpful anyway;-).


Alex




From wilk-spamout at flibuste.net  Wed Apr 30 04:42:58 2003
From: wilk-spamout at flibuste.net (Wilk)
Date: 30 Apr 2003 10:42:58 +0200
Subject: POD vs. reST for standalone writing?
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <u3ck0696n.fsf@boost-consulting.com>
Message-ID: <87llxs74vh.fsf@flibuste.net>

David Abrahams <dave at boost-consulting.com> writes:

> > Now I start seeing mention of reStructuredText (especially in comments
> > about pyCon.)  reST files look very "ascii art" on the input side, but
> > after a little experiment, that seems to make them *harder* to write,
> > if somewhat (and I think only a little) easier to read. 
> >
> > Am I missing something?  Is there actually a reST clear short hand,
> > and a tool that expands it to full-blown ascii-art?  Or is at least
> > fuzzy in the counting of dashes and other punctuation for marking
> > different levels?
> 
> ReST rocks.  I'm using it to write a book, and think that it comes
> out very readably and is miles easier to write than any tweaky **ML
> or TeX.

sure, i agree.

It's not so difficult to use, even for people without
computer knowledge. And it's easy to extend with new "directives" for
python programers.

After study of latex, docbook and *ml technologies with a professional
editor, they finaly choose rst (to put the book online and keep
archive).

I think rst has the same phylosophy of python. It's simple when we need
to do simple things (a simple text need only spaces between paragraph,
italic and bold). But it can be powerfull when we need (with standards
directives), and more with possibility to create new ones.

I had the same feeling from docbook/html/latex to rst than from java to
python ! No extra markers, clear text !

Somes examples of non-computer rst text (a french newspaper):
http://cequilfautdetruire.org

bye

-- 
William Dode - http://flibuste.net



From mhammond at skippinet.com.au  Wed Apr 16 19:35:36 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Wed, 16 Apr 2003 23:35:36 GMT
Subject: win32net.NetUseAdd syntax, possible bug
In-Reply-To: <mailman.1050496946.25294.python-list@python.org>
References: <mailman.1050496946.25294.python-list@python.org>
Message-ID: <cNlna.9544$hF.75665@news-server.bigpond.net.au>

As I mentioned in email, "password" is correct.  What error do you get?

Mark.

Scott Prive wrote:
> Hello,
> 
> I'm attempting to use win32net.NetUseAdd, had a problem and I think there's
> a bug in the API, or the docs (or me! :-)
> 
> The ActiveState made it clear most of the properties need to be in a
> dictionary. It lacked an example use, but referred me to the MSDN site, and
> at MSDN could see there the data structure. In theory I should combine the
> two docs and know all the things I can pass into the call via dictionary.
> 
> Small problem: the MSDN data structure calls the PW key "password", and
> win32net.NetUseAdd does not understand this. It looks like the function
> needs a key called "passwd", and if not provided it falls upon some default
> value (or the current credentials), making it appear that your user/pass
> values are not valid on the server when in fact they were.
> 
> This is my first post to the list (cc: me, I'm not subscribed). Is this the
> appropriate place for possible bug reports? How do I determine a module's
> "owner", and should I mail them directly?
> 
> The code looks like:
> 
>     import win32net
>     data = {}
>     data['remote'] = r'\\mypdc.example.com\myshare'
>     data['local'] = 'Z:'
>     data['password'] = 'foo' #Hmm.. I should call this     data['passwd'] =
> 'foo'
>     data['user'] = 'foo'
>     data['asg_type'] = 0
>     win32net.NetUseAdd(None,1,data)
> 
> -Scott
> 
> 
> 
> 
> 




From martin at v.loewis.de  Fri Apr 18 04:45:28 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 18 Apr 2003 10:45:28 +0200
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com> <3e9daad2$0$49115$e4fe514c@news.xs4all.nl> <f7199550.0304161732.48c75c42@posting.google.com> <mailman.1050548247.26781.python-list@python.org> <f7199550.0304170555.77773f8a@posting.google.com> <mailman.1050606359.1431.python-list@python.org> <f7199550.0304171911.2ac60fa6@posting.google.com>
Message-ID: <m3lly89ot3.fsf@mira.informatik.hu-berlin.de>

pc451 at yahoo.com (Peter Clark) writes:

> > And scale is a Unicode string, right?
>     Only because the XML document has no specified encoding, so it
> defaults to UTF-8, yes. But all the text is straight ASCII, except of
> course for the inclusion of the degree symbol.

No. UTF-8 is *not* Unicode. In Python, there are two data types: <type
'string'>, and <type 'unicode'>. The type string represents bytes (8
bit per element), and the type unicode represents characters.

The string type can also be used to represent characters, but only if
you assume that you are using some encoding. UTF-8 is an encoding, and
so is Latin-1. A string encoded in UTF-8 is still a byte string, not a
character string. A Unicode object may contain characters that can be
encoded in ASCII, or it can contain characters that cannot be encoded
in ASCII.

If this is not the mental model that you have, you will have a hard
time understanding all the phenomenons you observe, and I suggest
reading

http://manatee.mojam.com/~skip/unicode/unicode/

Regards,
Martin



From tim.hochberg at ieee.org  Wed Apr  2 18:48:33 2003
From: tim.hochberg at ieee.org (Tim Hochberg)
Date: Wed, 02 Apr 2003 16:48:33 -0700
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
In-Reply-To: <mailman.1049321730.7572.python-list@python.org>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>        <b68jti$2mj66$1@ID-169208.news.dfncis.de>        <RbCia.6782$9I.555309@news20.bellglobal.com>        <mailman.1049298845.7655.python-list@python.org>        <3E8B4094.F61515B3@engcorp.com>        <ZpHia.23$Sd.4@fed1read01>        <3E8B48E0.C786D192@engcorp.com>        <mailman.1049316186.31356.python-list@python.org>        <3E8B55B2.D8D7D2F2@engcorp.com> <mailman.1049321730.7572.python-list@python.org>
Message-ID: <nFKia.141$Sd.25@fed1read01>

Skip Montanaro wrote:
>     >> How about calling new.module() to create a new module object from the
>     >> one you want to test which has suitably munged globals?
> 
>     Peter> Interesting idea... but ...
>     ...
>     Peter> # install mock open again
>     Peter> mungedMut.open = mockOpen
> 
>     Peter> What I don't understand is the difference between the final line
>     Peter> above and what I was originally trying to do, except that this is
>     Peter> a newly created module instead of one that was imported.
> 
>     Peter> Okay, I guess I must have missed a key point and I'm way off the 
>     Peter> track here.  Please clarify.
> 
> No, it turns out I was.  I was thinking you passed a bunch of stuff (like a
> dict to use as globals()) to new.module().  I guess I was thinking of
> new.code() and new.function().  You might be able to poke elements directly
> into the module's __dict__.  Perhaps new.module can grow an optional dict
> argument to handle this sort of thing.
> 
> I suggest you post your use case to python-dev as something for people there
> to think about.  In fact, new.module() would become pretty useless if you
> weren't able to do what you can do today, so I think it will probably need a
> dict argument if the "no modifying another module's globals" restriction is
> implemented.

I think that nearly this exact use case was discussed here: 
http://mail.python.org/pipermail/python-dev/2003-March/034272.html . If 
you go forward a few messages, GVR suggests adding "from __builtins__ 
import open" to modules that you might want to fiddle with.

This made me recall that the proposal was slightly more lenient than I 
initially remembered:

some_random_module.open = my_open

would be OK if open is already shadowed in some_random_module. The 
compiler then can figure out that optimizing access to open is not going 
to work.

-tim




From dkuhlman at rexx.com  Wed Apr  9 12:11:57 2003
From: dkuhlman at rexx.com (Dave Kuhlman)
Date: Wed, 09 Apr 2003 09:11:57 -0700
Subject: REST and directory servicies
References: <b717h2$t3c@newton.cc.rl.ac.uk>
Message-ID: <b71go4$a0f4o$1@ID-139865.news.dfncis.de>

Richard Brodie wrote:

[snip]
> 
> Granted, these are implementation issues but has anyone been down
> this way before? It seems to be that what I'm looking for is a
> lightweight, robust way of mapping a directory service into the
> http: namespace. Or maybe I have some fundamental architectural
> misunderstanding. (Did you say 'knives')?
> 
> Any views?

Take a look at Quixote.  It's one of the light-weight Python Web app 
frameworks.  It has reasonably convenient ways to map URLs to 
Python modules, classes, and functions.  See:

  http://www.mems-exchange.org/software/quixote/doc/

and, in particular:

  http://www.mems-exchange.org/software/quixote/doc/programming.html

Other Python Web app frameworks are likely to have solutions to this 
problem also.  So, you will want to look at:

  http://www.python.org/cgi-bin/moinmoin/WebProgramming

  - Dave

-- 
Dave Kuhlman
dkuhlman at rexx.com
http://www.rexx.com/~dkuhlman



From richie at entrian.com  Mon Apr 28 11:37:39 2003
From: richie at entrian.com (Richie Hindle)
Date: 28 Apr 2003 08:37:39 -0700
Subject: WML examples using mod_python?
References: <slrnbals94.2qp.grante@tuxtop.visi.com>
Message-ID: <8a6ba1da.0304280737.3b6692a3@posting.google.com>

Hi Grant,

> Any other python-related WML examples (normal CGI, etc.) would
> also be welcome. [...] an IMAP to WML gateway is the project I
> have in mind.

http://www.entrian.com/wapmail/ is a POP3 to WML gateway written in
Python as a CGI script.  That page includes an online demo that runs
in your web browser, and there's a link to the source (about 450 lines
plus documentation) at the bottom of the page.

The web demo was easy to add once the WML stuff was finished because
there's no WMLScript in the output, so it's just a matter of sending
out different preamble and, er, postamble.  The actual content is
identical between HTML and WML - nice for debugging and testing if you
can get away with it.

One day I'll integrate Spambayes into it.  8-)

-- 
Richie Hindle
richie at entrian.com



From mayhew at cis.ohio-state.edu  Wed Apr 23 16:28:07 2003
From: mayhew at cis.ohio-state.edu (Michael Mayhew)
Date: Wed, 23 Apr 2003 15:28:07 -0500
Subject: Using the MySQLdb module
References: <3EA6E75E.7070305@cis.ohio-state.edu> <mailman.1051131517.32137.python-list@python.org>
Message-ID: <3EA6F757.5080903@cis.ohio-state.edu>

Gerhard H?ring wrote:
> Also see the DB-API 2 specification and the documentation and examples 
> included in the MySQLdb source distribution.

Gerhard,

    Unfortunately, the MySQLdb documentation, both on the web and from 
the sourceforge location are a little lax for examples on using cursor 
objects. Believe me, I have referred to both on multiple occasions. If 
you know of another site where I could see an example of cursor object 
use, that would be of great help to me. Thanks.

Michael Mayhew






From skip at pobox.com  Thu Apr 10 17:35:32 2003
From: skip at pobox.com (Skip Montanaro)
Date: Thu, 10 Apr 2003 16:35:32 -0500
Subject: PEP305 csv package: from csv import csv?
In-Reply-To: <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>
References: <mailman.1049814434.25284.python-list@python.org>
        <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>
Message-ID: <16021.58276.70205.458931@montanaro.dyndns.org>

    Magnus> As far as I can see, csv.csv.__all__ contains the relevant
    Magnus> names, so simply adding

    Magnus>   from csv import *

    Magnus> to csv.__init__.py should do the trick nicely.

Done.  Thanks.

BTW, I also make the csv/cvs typo periodically.  I think we'll all just have
to grin and bear it...

Skip




From ckea25d02 at sneakemail.com  Tue Apr  8 20:56:18 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 09 Apr 2003 12:56:18 +1200
Subject: Cross-reference 'import' in a class hierarchy
In-Reply-To: <mailman.1049778853.27420.python-list@python.org>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <mailman.1049778853.27420.python-list@python.org>
Message-ID: <b6vqhf$9n4qi$1@ID-169208.news.dfncis.de>

Jp Calderone wrote:
>   It's simple.  Code should not depend directly on code higher than it in
> the package hierarchy.  If it does (and it isn't one of a few extremely rare
> cases), it is misdesigned.

But it's worse than that -- you can't even refer to things
*deeper* in the hierarchy relative to yourself.

It's like having a Unix where you can say "spam" or
"/bar/spam/foo" but not "spam/foo".

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From aahz at pythoncraft.com  Wed Apr 16 20:02:25 2003
From: aahz at pythoncraft.com (Aahz)
Date: 16 Apr 2003 20:02:25 -0400
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com> <gcjna.35728$T34.931514@news2.tin.it>
Message-ID: <b7kquh$k4c$1@panix2.panix.com>

In article <gcjna.35728$T34.931514 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>Aahz wrote:
>> In article <iCzma.16405$T34.518484 at news2.tin.it>,
>> Alex Martelli  <aleax at aleax.it> wrote:
>>>
>>>def test():
>>>    class Bunch: pass
>>>    loc = Bunch()
>>>    loc.count = 0
>>>    def inc_count():
>>>        loc.count += 1
>>>    inc_count()
>>>    inc_count()
>>>    print loc.count
>> 
>> Note that Guido deeply regrets permitting one-line blocks; this is
>> something he may try fixing for Python 3.0.
>
>Forcing linebreak after each & every colon?  Oh well -- he's probably
>getting bribes from large-screen and vertical-aspect-ratio monitor
>manufacturers, is my guess...;-)

The reason I mentioned it is because I did indeed miss your
"class Bunch: pass" the first time I read your code.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From gh at ghaering.de  Thu Apr 24 20:45:07 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 02:45:07 +0200
Subject: *Newbie Question* --> Is it viable to distribute Python apps
 over the Internet?
In-Reply-To: <Jq%pa.286$iG3.15927365@newssvr14.news.prodigy.com>
References: <25a672ee.0304210914.1cc2049c@posting.google.com>
	<mailman.1050946774.25112.python-list@python.org>
	<uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com>
	<mailman.1050978337.8048.python-list@python.org> <87wuhm2rsu.fsf@pobox.com>
	<Bkipa.1154$Hm6.104194064@newssvr14.news.prodigy.com>
	<87adefkbs9.fsf@pobox.com> <m7Ypa.156$Jr1.14453940@newssvr15.news.prodigy.com>
	<Jq%pa.286$iG3.15927365@newssvr14.news.prodigy.com>
Message-ID: <3EA88513.1030809@ghaering.de>

Dennis Reinhardt wrote:
>>sys.path = [".", ".\\bin", ".\\bin\\libstd", ".\\bin\\libxtra"]
>>Maintaining this "manual hack" is far less effort than you would guess.
> 
> 
> There is something I forgot to mention.  Suppose you start with a decent
> size app already?  You do not have to go through and add files to this
> structure 1-by-1.
> 
> To seed this structure initially, comment out the sys.path line so that you
> are using your system python path.  Now, delete *.pyc where the *.py exists.
> Run your program and exercise its functions.  Exit.  Now, where *.pyc files
> exist in the system python path, you need to copy the .py files to a
> subdirectory listed in sys.path.
> 
> In one step, you have now populated your subdirectories with the required
> libraries.  There may be other files that you will end up needing and
> restoring the sys.path statement will quickly find them.

Poor man's modulefinder, eh? Neat :-)

-- Gerhard




From amk at amk.ca  Thu Apr 24 14:11:14 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Thu, 24 Apr 2003 13:11:14 -0500
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com> <e9570f37.0304220625.11312687@posting.google.com> <K2qdnXJ_ZeZMVTujXTWcqg@speakeasy.net> <F3Epa.51023$LB6.1152908@news1.tin.it> <e9570f37.0304240556.50465c66@posting.google.com>
Message-ID: <k_2dnckpjsVftTWjXTWcpQ@speakeasy.net>

On 24 Apr 2003 06:56:51 -0700, 
	Graham Fawcett <graham__fawcett at hotmail.com> wrote:
> Of course we have no infrastructure for this (at least for the ones
> that cost money). But I'm sure we could create one if the need arose.

Two sites were created for this purpose, SourceExchange and Cosource.  Both
are now dead and gone, which should tell you something.  Such a tip-jar
system produces trivial amounts of money, $20 or $100. If you had $50K to
spend, you could hire a developer to work on something for 40 hours a week,
but that amount of money is not found in a tip jar.

The limited resource is not money, it's time, and tiny amounts of cash are
not sufficient incentive to make someone allocate time to a project.  Most
developers have lots and lots of projects they can spend time on, some
software-related, some not; their projects multiply over time.  The solution
is to add more people, expanding the amount of time available, instead of
using money to affect the allocation of a shrinking pool of available time.
Many projects will still never be completed, in that model.

--amk                                                    (www.amk.ca)
MACBETH: Throw physic to the dogs; I'll none of it.
      -- _Macbeth_, V, iii



From andrew-pythonlist at puzzling.org  Wed Apr 16 10:12:44 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Thu, 17 Apr 2003 00:12:44 +1000
Subject: Typing arguments in python
In-Reply-To: <fd1c664e.0304160524.295bbdc1@posting.google.com>
References: <fd1c664e.0304160524.295bbdc1@posting.google.com>
Message-ID: <20030416141244.GD27886@frobozz.local>

On Wed, Apr 16, 2003 at 06:24:15AM -0700, Danra wrote:
[...]
> 
> def f(fileobj):
>     if not hasattr(fileobj,'write'):
>         raise TypeError, "Given arg must have the 'write' attribute."
>     .
>     .
>     fileobj.write('...')
>     .
>     .
> 
> The advantages of the second definition are obvious - The fact that
> the argument passed doesn't have the write method would be detected
> immediately, rather than when trying to call the it.
> 
> Why is that so important? Well, for starters in many cases it isn't.
> In a simple script or application it really is sufficient to get the
> exception only when calling write.
> 
> But - what if:
> 
> 1) The stuff you do before calling write takes a lot of time, which
> could be saved if the error is detected when the function is called.

If this really is a concern, then this idiom takes care of it quite
succiently:

def f(fileobj):
    write = fileobj.write
    ...
    # Slow code goes here
    ...
    write('...')

> 2) The object gets passed around from function to function, and maybe
> changes its name on the way. In this case it would be more difficult
> to trace the problem from the errorneous call to the original function
> call.

1) Spaghetti code should be avoided
2) For geniunely complex code, you can resort to a debugger if the tracebacks
aren't sufficient.

I suggest learning to love and trust Python's dynamic typing, rather than
fighting it.

-Andrew.





From roy at panix.com  Tue Apr 22 18:28:37 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 22 Apr 2003 18:28:37 -0400
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org>
Message-ID: <roy-1971DE.18283722042003@reader1.panix.com>

In article <mailman.1051036237.30103.python-list at python.org>,
 Ian Bicking <ianb at colorstudy.com> wrote:

> import sys
> if sys.version_info < (2, 3):
>     True, False = 1==1, 0==1

If you insisted on doing something like that, you'd be a bit more 
portable making the test "'True' not in dir(sys)".  I think it would 
also be a lot easier to read if you broke the assignment up into two 
simplier statements:

import sys
if 'True' not in dir(sys):
   True = 1
   False = 0

I'm not 100% sure testing for the presense of 'True' in dir(sys) is the 
right test, but it's better than testing for an absolute version number.
  
A better test might be looking for 'BoolType' in dir(types), or testing 
for the presense of True in the global namespace by evaluating it inside 
a try block and catching NameError.  Or something.  The basic idea is to 
test for the functionality you're looking for, not a version number.



From grante at visi.com  Fri Apr 25 01:57:55 2003
From: grante at visi.com (Grant Edwards)
Date: 25 Apr 2003 05:57:55 GMT
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth>
Message-ID: <slrnbahk9n.ero.grante@localhost.localdomain>

In article <98imn-ust.ln1 at grendel.myth>, Jim Richardson wrote:

> I have come across a weirdness. I have a number of floats, (lat long
> pairs) which I stuff into a dict, but when I take them out, my nice %2f4
> number, (say, 47.5894) is something long and *almost* right, like
> 47.589333333333333339. What gives?

That's floats.  The best you can hope for is "almost right".  And that's
assuming your algorithm is right.  With the wrong algorithm you can easily
end up with "not at all right".

-- 
Grant Edwards                   grante             Yow!  I'm ANN LANDERS!! I
                                  at               can SHOPLIFT!!
                               visi.com            



From peter at engcorp.com  Thu Apr 17 09:20:54 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 17 Apr 2003 09:20:54 -0400
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl>
Message-ID: <3E9EAA36.BFB1018D@engcorp.com>

Anton Vredegoor wrote:
> 
> From the universe of all currently available webpages (where
> "currently available" means available on the world wide web at the
> runtime of the script) I want to select a page in such a way that each
> page has an equally likely chance of being selected. Any idea about
> how to accomplish this is equally likely to be welcome ;-)

Sounds rather silly, since there can't possibly be much value
in selecting from literally *all* pages with *equal* probability,
can there?  But anyone, here's an idea that goes part-way there.

1. Generate random IP addresses.  Easiest way is using an integer
   which is converted to a net address using tools in "socket".

2. Try connecting on port 80.

3. If there's something there, display it and you're done.  Else loop.

Now if you consider that a starting point, for discussion, what
part of it doesn't serve your purpose, and why?

-Peter



From python at dylanreinhardt.com  Sat Apr 19 10:53:01 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 19 Apr 2003 07:53:01 -0700
Subject: Parent references: hot or not?
Message-ID: <1050763980.14721.41.camel@ida.dylanreinhardt.com>

Hi all,

I'm working on a project where I have multiple instances of a container
class that each hold a large number of another class.

I need the "contained" object to be able to make use of methods and
attributes in its container. Inheritance won't do the trick since the
things I need to refer to will be pretty dynamic.

A very minimal example of what I'm doing:

---------------

class Container:
   def __init__(self):
      self.items = {}

   def add_item(self, item_name):
      self.items[item_name] = Item(self)

   some_attr = 'spam'


class Item:
   def __init__(self, parent):
      self.parent = parent

   def some_method(self):
      return self.parent.some_attr

---------------

This seems to be working well, but I want to check my assumptions.

If I'm understanding correctly:

1. Any Item's self.parent is only a *reference* to its parent.
2. Holding a reference to a parent isn't some kind of recursive paradox.
3. As I add Items, Container size will grow linearly, not exponentially
4. It will be impossible to fully delete a non-empty Container.

I think those are valid assumptions, but would appreciate any
confirmation, contradiction, or insight anyone cares to offer.

Also, if there's some other consideration or reason why holding a
reference to one's parent is a Bad Idea, I'd love to hear about it.

Thanks in advance,

Dylan





From abpillai at lycos.com  Tue Apr 22 02:50:06 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 21 Apr 2003 23:50:06 -0700
Subject: Python Thread Question
References: <bd993a2f.0304170534.245aa2d5@posting.google.com> <mailman.1050587609.6119.python-list@python.org> <b7meoo$mf6$1@panix2.panix.com> <mailman.1050593376.21366.python-list@python.org>
Message-ID: <bd993a2f.0304212250.41d04d@posting.google.com>

Thanks for your replies. It has helped
clear my ideas on threading. 

Anand

Jp Calderone <exarkun at intarweb.us> wrote in message news:<mailman.1050593376.21366.python-list at python.org>...
> On Thu, Apr 17, 2003 at 10:46:48AM -0400, Aahz wrote:
> > [Mostly good answers from Jp, but one correction is needed]
> > 
> > > [snip]
> > >
> > >  Context switches are usually relatively cheap compared to what your
> > >app is actually doing, though, so the difference between doing blocking
> > >IO in a multi-thread app and non-blocking IO in a single-thread app
> > >isn't always obvious.  (Other things associated with multi-threaded
> > >apps, such as deadlocks and race conditions are, though ;)
> > 
> > This isn't quite true on a multi-CPU box.  Because I/O releases the GIL,
> > the OS can schedule threads to run on all available CPUs.  So it's
> > certainly possible for a threaded app to outrun non-blocking I/O.
> 
>   Quite right.  Thanks for the correction.
> 
>   Jp
> 
> -- 
> Examinations are formidable even to the best prepared, for
> even the greatest fool may ask more the the wisest man can answer.
>                 -- C.C. Colton



From paul at prescod.net  Mon Apr  7 03:17:23 2003
From: paul at prescod.net (Paul Prescod)
Date: Mon, 07 Apr 2003 00:17:23 -0700
Subject: Problem setting COM property using win32com
In-Reply-To: <he9ck171.fsf@boeing.com>
References: <he9ck171.fsf@boeing.com>
Message-ID: <3E912603.3060803@prescod.net>

I don't claim to be a win32com expert but I notice two things in your 
message that made me suspicious.

Greg Green wrote:
> I have a commercial com api that I'm trying to use with python. I am
> trying to set a property of the com object. The property is called
> Value, and I know enough to use the SetValue function. But that
> doesn't work.
> 
> Here is the code:
>     attrs.SetValue("CN_ID", 0, "0000005")
> 
> Here is the error:
> Traceback (most recent call last):
>   File "example.py", line 49, in ?
>     prmidattrs.SetValue("CN_ID", 0, "0000005")
>   File "C:\PYTHON22\lib\site-packages\win32com\client\dynamic.py", line 438, in __getattr__
>     raise AttributeError, "%s.%s" % (self._username_, attr)
> AttributeError: SmRecList.SmRecordList.SetValue
> 
> If I look at the makepy generated module, I see the method.

Why are you getting an exception in "dynamic.py" if you are using "makepy"?

How are you creating the prmidattrs object?

  Paul Prescod





From hancock at anansispaceworks.com  Thu Apr 17 13:49:07 2003
From: hancock at anansispaceworks.com (Terry Hancock)
Date: Thu, 17 Apr 2003 10:49:07 -0700
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl> <3E9EAA36.BFB1018D@engcorp.com>
Message-ID: <IQmdnQg6xagDdQOjXTWcpA@august.net>

Peter Hansen wrote:
> 1. Generate random IP addresses.  Easiest way is using an integer
>    which is converted to a net address using tools in "socket".
> 
> 2. Try connecting on port 80.
> 
> 3. If there's something there, display it and you're done.  Else loop.
> 
> Now if you consider that a starting point, for discussion, what
> part of it doesn't serve your purpose, and why?

This easily sounds like the closest one can get to a real answer, but I do 
have one question -- what about virtual hosting?  Aren't a large number of 
websites set up to share IPs using name-based resolution nowadays.

I actually know little about how this is used on the web, but I have Apache 
configured on my testing server to deliver different sites for the same IP 
based on the domain name requested (this is on my LAN).  I figure it's 
probably used a lot, especially for personal web pages and the like.

Perhaps one can query the DNS registry in some way that would be more 
equal?  Can RARP be used to look up *all* domains associated with an IP, or 
only one canonical value?  If you could look up all of them, you could then 
select one at random -- but this will not generate evenly weighted results 
of course.

You *could* make an effort to record look-ups that resulted in large 
numbers of domains to give them more weight in the future.  But this would 
be expensive, and you'd probably have to run for a *long* time for it to 
become relevant.

Of course, the distribution is probably not too relevant -- there are so 
many web pages in the world now, that you are unlikely to retrieve the same 
one twice with any truly random system.  (Not true on Google, where 
connectivity highly biases the search).

I know there's an X screensaver that does something like this to retrieve 
random images.  I doubt they were too concerned about even distribution, 
though.

Cheers,
Terry

-- 
Terry Hancock
Anansi Spaceworks http://www.AnansiSpaceworks.com/




From oblivious at web.de  Sat Apr 12 05:48:19 2003
From: oblivious at web.de (Dominic)
Date: Sat, 12 Apr 2003 11:48:19 +0200
Subject: Macros in Python? (PEP 310)
In-Reply-To: <a8b7f07a.0304111546.3e5a5636@posting.google.com>
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org> <b770t8$2b8$00$1@news.t-online.com> <b771jb$ac6$03$1@news.t-online.com> <a8b7f07a.0304111546.3e5a5636@posting.google.com>
Message-ID: <b78nae$sft$06$1@news.t-online.com>

> If this isn't what you are trying to do, could you clarify what's
> missing?

Your approach is not flexible enough to wrap
arbitrary functions because _y is hardcoded.
(Or am I wrong?)
Using the generator-hack you don't even
need to wrap the code you want
to execute into a function (!)
But it is still guranteed that e.g. setuid
is set to it's old value.

I've been curious about how to
circumvent try: finally: in a general and
with minimum effort.

PEP305 and the generator based hack
have been very close to what I was
looking for.

Ciao,
  Dominic




From jkraska at san.rr.com  Thu Apr 17 11:39:44 2003
From: jkraska at san.rr.com (Courageous)
Date: Thu, 17 Apr 2003 15:39:44 GMT
Subject: "pythonic"
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
Message-ID: <qait9vsttagq0c8gk5f40ou991bitistfe@4ax.com>

>I do not know if this ignorance, laziness, or a meme, but I would like
>to know if there a real definition of "pythonic". Anyone?

First you have to understand the word "unpythonic".

This is kind of like the word "communist". You know? During the
50's, your life was over if you were labeled a "communist"? 
Guido, Tim, and team sit on a the Technical Committee for Un[ythonic
Activities.

Be afraid, be very afraid.

C//

p.s. we'll overlook this transgression of yours if report three
of your friends.




From LogiplexSoftware at earthlink.net  Mon Apr 21 06:05:13 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 21 Apr 2003 03:05:13 -0700
Subject: Python and Schools
In-Reply-To: <5.2.0.9.0.20030416092224.03b7abc0@mail.softhome.net>
References: <5.2.0.9.0.20030416092224.03b7abc0@mail.softhome.net>
Message-ID: <1050919513.1723.97.camel@software1.logiplex.internal>

On Wed, 2003-04-16 at 06:22, Andres Rosado wrote:
> On 05:06 AM 4/11/2003 -0400, python-list-request at python.org feed this fish 
> to the sharks:
> >- if you learn programming with Python and later have to use Java or
> >Visual Basic, it will be the most frustrating experience for the young
> >fellows.
> 
> Pardon my ignorance, but how could it be a frustating experience?

Like learning to drive in a BMW and then switching to a Yugo.

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From aleax at aleax.it  Tue Apr 15 04:30:08 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 08:30:08 GMT
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it> <ouadesfkn2.fsf@andlx-anamika.cisco.com> <7mOma.20244$T34.649927@news2.tin.it> <ouvfxgdv8n.fsf@andlx-anamika.cisco.com>
Message-ID: <kqPma.20705$T34.658905@news2.tin.it>

Ganesan R wrote:

> 
>>>>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:
>>> 
>>> Why am I getting a different list for the last call?
> 
>> Because you're asking about a distinct (even though equal) value,
>> AKA object.  Python MAY but doesn't HAVE TO reuse the same object
>> when you use several literals indicating equal immutables -- in
>> practice, currently, it does for small integers, but (in Python
>> 2.2) not for such strings as you used here.  Python can always
>> change this in future releases, since this is strictly an issue of
>> implementation -- optimizing one way or another, no more than that.
> 
> <andlx-anamika 1> ~ $ python
> Python 2.2.1 (#1, Jul 29 2002, 23:15:49)
> [GCC 2.95.4 20011002 (Debian prerelease)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import gc
>>>> a = 'hello world'
>>>> b = 'hello world'
>>>> c = a
>>>> gc.get_referrers(b)
> [{'a': 'hello world', 'c': 'hello world', 'b': 'hello world', 'gc':
> [{<module 'gc' (built-in)>, '__builtins__': <module '__builtin__'
> [{(built-in)>, '__name__': '__main__', '__doc__': None}]
> 
> The reason I was confused was because of the output from
> gc.get_referrers(b). I assumed that a and b held a reference to the same
> object. I should've checked the ids of a and b.
> 
> I now figured out that a is holding a reference to b because of some
> python implementation detail. In fact all module variables seem to hold a
> reference to other variables.

No!  the object that get_referrers is returning is the dictionary
of the module -- that dictionary refers to all variables (that's
where they live); it's NOT that variables refer to each other!


> Thanks for the clarifications.

You're welcome,

Alex




From uselpa at myrealbox.com  Fri Apr 25 17:18:03 2003
From: uselpa at myrealbox.com (Patrick Useldinger)
Date: Fri, 25 Apr 2003 23:18:03 +0200
Subject: zipfile with umlaut in filename
Message-ID: <3ea9a4bf$1_1@news.vo.lu>

Hello everybody,
I have a problem when creating an archive using zipfile. If the name of 
the file to be added contains a special character like an umlaut, this 
character gets replaced by another character.
What can I do against this?
I am using Python 2.2.2 on Win2K.
-Patrick




From sdhyok at yahoo.com  Mon Apr 28 22:25:31 2003
From: sdhyok at yahoo.com (sdhyok)
Date: 28 Apr 2003 19:25:31 -0700
Subject: [Q] How to plot arrays in KChart
Message-ID: <420ced91.0304281825.62df3995@posting.google.com>

I am considering to use KChart as my plotting program.
Would you tell me how to plot arrays in python with KChart?

For instance,

from Numeric import *
x = arange(10)
y = x*x
#How to plot x and y in KChart?

Thanks in advance.

Daehyok Shin



From rpluim at grandNOSPAMpied.com  Sat Apr  5 11:11:14 2003
From: rpluim at grandNOSPAMpied.com (Robert Pluim)
Date: 05 Apr 2003 18:11:14 +0200
Subject: POP3 and selective download/deletion
References: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com> <pan.2003.04.05.16.22.31.832230.878@thomas-guettler.de> <luCja.11925$It5.5735@news2.central.cox.net> <86d6k1ar0u.fsf@fibble.nospam.dev.null> <pan.2003.04.05.17.52.43.14616.1268@mindlessSPAM.com>
Message-ID: <868yuoc4bh.fsf@fibble.nospam.dev.null>

Dino Levy <dino_levy at mindlessSPAM.com> writes:

> On Sat, 05 Apr 2003 17:43:45 +0200, Robert Pluim wrote:
> 
> >> Unfortunately the POP3 protocol doesn't allow you download the headers
> >> of a message without downloading the whole thing. IMAP4 is much more
> >> capable, but it's also a very much more complex protocol.
> > 
> > Que?  You've never heard of the TOP command?
> 
> Hehe, I heard of it, but it doesn't help much in this case.

It doesn't?  TOP <n> 1 should give you all the headers of message <n>,
which you should then be able to match against your criteria, and the
you can issue DELE commands.

-- 
For mail, remove NOSPAM and translate from French.



From peter at engcorp.com  Tue Apr 29 13:37:20 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 29 Apr 2003 13:37:20 -0400
Subject: stupid q-n (wxPy)
References: <Ukxra.33100$7M5.3263531@twister.nyc.rr.com>
Message-ID: <3EAEB850.50D9DEB9@engcorp.com>

xam wrote:
> 
> I am trying to run some wxPy samples and I am getting an error here:
> if __name__ == '__main__':
>     import sys,os
>     import run
>     run.main(['', os.path.basename(sys.argv[0])])
> 
> - "import run" fails. I know this is something really stupid, but I could
> not find the answer on the net.
> thanks.

What is "run"?  Where is it?  Are you sure the script you are running
can find "run.py"?  If there's no run.py in the directory of the script
you show above, insert a "print sys.path" after the first import statement
and make sure the list it prints contains the directory where run.py is
actually installed.

A little more detail/context would be helpful next time.

-Peter



From ramen at lackingtalent.com  Tue Apr  1 12:50:08 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 17:50:08 -0000
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <d2c6a353-612a-4f45-8ddd-bddc04d59e6a@news.szn.dk> <slrnb8jfsq.9n8.philh@cabalamat.uklinux.net> <41d02ffe-a3d3-4bde-9466-d4b53309242a@news.szn.dk>
Message-ID: <slrnb8jheo.5u1.ramen@lackingtalent.com>

In article <41d02ffe-a3d3-4bde-9466-d4b53309242a at news.szn.dk>, Klaus
Alexander Seistrup wrote: Phil Hunt wrote:
> 
>>>#v+
>> 
>> What does this mean?
> 
> That's the beginning of a set of "verbatim marks" for the JED editor.

Really? How does that work?

Dave



From bokr at oz.net  Sat Apr 26 13:27:44 2003
From: bokr at oz.net (Bengt Richter)
Date: 26 Apr 2003 17:27:44 GMT
Subject: sys.argv problem in Win 2k
References: <bd993a2f.0304240121.7c795cf0@posting.google.com> <mailman.1051177117.30168.python-list@python.org> <bd993a2f.0304240506.5b206f3d@posting.google.com> <Xns936790B52180Eduncanrcpcouk@127.0.0.1> <bd993a2f.0304242310.586305df@posting.google.com> <Xns9368583183AC8duncanrcpcouk@127.0.0.1> <bd993a2f.0304260324.4e9a4f4e@posting.google.com>
Message-ID: <b8efig$o21$0@216.39.172.122>

On 26 Apr 2003 04:24:02 -0700, abpillai at lycos.com (Anand B Pillai) wrote:

>Well I was hoping that would fix it!
>I tried
>
>Python.File=D:\Python22\python.exe "%1" %*
>and
above should work IF assoc .py => .py=Python.File

If assoc .py => .py=someOtherName
check ftype someOtherName and see if the %* is left off.

>Python.File=D:\Python22\python.exe "%*"
above should fail to work

>and another 2 combinations,
>
>But still I cannot pass the cmd line arguments!
>Let me try and figure out myself :-O
Maybe you have a different association via .py -- e.g., I don't have the dotted name:

[10:28] C:\pywk\clp>assoc .py
.py=Pythonfile

[10:28] C:\pywk\clp>ftype Pythonfile
Pythonfile=d:\python22\python.exe "%1" %*


>
>ABP
>Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in message news:<Xns9368583183AC8duncanrcpcouk at 127.0.0.1>...
>> I wrote:
>> >> D:\temp>ftype Python.File
>> >> Python.File=D:\Python22\python.exe "%1" %*
>> 
>> abpillai at lycos.com (Anand B Pillai) wrote in
>> news:bd993a2f.0304242310.586305df at posting.google.com: 
>> 
>> > F:\users\anand\LocalWorkspaces\ResModWS>ftype Python.File
>> > Python.File=D:\Python22\python.exe "%1"
>> > 
>> 
>> Spot the difference.
>> 
>> The "%1" passes the name of the script to python.exe, the %* passes all the 
>> other arguments. Without the %* windows will simply drop the other 
>> arguments on the floor.

Regards,
Bengt Richter



From mertz at gnosis.cx  Tue Apr  8 14:09:29 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 14:09:29 -0400
Subject: [Slightly OT]: More on ints and floats
In-Reply-To: <7h3wui4528k.fsf@pc150.maths.bris.ac.uk>
References: <fjss6b.1021.ln@boundary.tundraware.com>
 <mailman.1049768113.16134.python-list@python.org>
 <kauu6b.f2d1.ln@boundary.tundraware.com>
 <7h3wui4528k.fsf@pc150.maths.bris.ac.uk>
Message-ID: <ZBxk+kKkXEtW092yn@gnosis.cx>         

|Floats are perfectly normal rational numbers; it's the *operations*
|which are weird.

Aren't you being slightly disingenuous here?

Any numbers are more-or-less only meaningful in terms of the operations
they enter into.  The Peano integers, or the Cauchy or Dedekind Reals,
are not really interesting because of the beauty of their construction,
but rather because it is relatively easy to define operations on them
that match our intuitions about how arithmetic is -supposed- to behave.
The operations are the whole point of the entities.

Likewise, a float isn't *really* a number at all.  It's really just a
pattern of bits, such as '01000001010000010100000101000001'.  Maybe
those bits are a way of representing the string "AAAA", or maybe they
are a way of representing a -thing- that enters into operations that act
(a little bit) like addition, multiplication, etc.  Just being a bit
pattern doesn't make it a *number* at all.

Yours, Lulu...

--
Keeping medicines from the bloodstreams of the sick; food from the bellies
of the hungry; books from the hands of the uneducated; technology from the
underdeveloped; and putting advocates of freedom in prisons.  Intellectual
property is to the 21st century what the slave trade was to the 16th.




From rstephens at vectron.com  Sun Apr 27 17:35:28 2003
From: rstephens at vectron.com (Ron Stephens)
Date: 27 Apr 2003 14:35:28 -0700
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <a5415304.0304271335.2f569879@posting.google.com>

Steven Hodgen <steven at twitch.net> wrote in message news:<oproar1sjtx3zh5c at news.mindspring.com>...
> Hello,
> 
> I'm currently teaching a high-school introduction to programming course.  
> I'm currently teaching in Java, but am seriously considering changing to 
> Python.  The problem is that I cannot seem to find any good textbooks, much 
> less ones for students with no prior programming experience.  This is the 
> one area in which Java excels. 

The only actual "Textbook" I am aware of is Deitel and Deitel's Python
book, which I found pretty good. However, Learning Python by Lutz and
Ascher would be excellent (and it includes exercises after every
chapter) as would Core Python by Chung. I have reviewed all these
books and more than 20 more and you can find the reviews at
www.awaretek.com/plf.html



From exarkun at intarweb.us  Thu Apr 10 18:58:00 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 10 Apr 2003 18:58:00 -0400
Subject: Can I define nested classes in Python?
In-Reply-To: <169a7283.0304101432.13f5786c@posting.google.com>
References: <169a7283.0304101432.13f5786c@posting.google.com>
Message-ID: <20030410225800.GA30405@meson.dyndns.org>

On Thu, Apr 10, 2003 at 03:32:01PM -0700, sdieselil wrote:
> Hi
> 
> Can I define nested classes in Python and what access do they have to outer
> classes?

  You can, though you might be surprised by their behavior:

    >>> class Foo:
    ...   class Bar:
    ...     pass
    ... 
    >>> print Foo.Bar.__name__
    Bar

  As well as:

    >>> class Foo: 
    ...   class Bar:
    ...     def __init__(self):
    ...       self.container = Foo()
    ... 
    >>> Foo.Bar()
    <__main__.Bar instance at 0x8165364>

  But not:

    >>> class Foo:
    ...   class Bar:
    ...     container = Foo()
    ... 
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
      File "<stdin>", line 2, in Foo
      File "<stdin>", line 3, in Bar
    NameError: name 'Foo' is not defined

  The reasons for this are perfectly reasonable, of course - the name is not
bound to the class object until the entire class suite has executed.

  It may be reasonable to need access to the outer class in the inner
class's suite (I haven't decided yet), and I recently wrote this to
demonstrate how one might go about achieving this:

  http://intarweb.us:8080/evil/inner_inheritance.py

  HTH,

  Jp

--
"If you find a neighbor in need, you're responsible for serving that
neighbor in need, you're responsible for loving a neighbor just like you'd
like to love yourself." -- George W. Bush, Sept. 16, 2002
-- 
 up 21 days, 19:01, 2 users, load average: 1.13, 1.16, 1.30




From news at snakefarm.org  Thu Apr 10 17:50:13 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 10 Apr 2003 21:50:13 GMT
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de>
Message-ID: <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de>

In article <3e95d63f.2635484 at news.muenster.de>, Martin Bless wrote:
> 
> To send mails from Python here's the example my provider provides:
> This http://faq.puretec.de/skripte/python/3.html
> 
> The relevant lines are these:
> 
> sendmail = os.popen("/usr/lib/sendmail -t", "w")
> sendmail.write(msg)
> sendmail.close()
> 
> This works just fine. But is it enough? How do I know the operation
> succeeded?

In an error condition python would either raise an exception or give you
sendmail's error status as the return value of the close() call.

If your code appears in a CGI script, you should catch the exception and
print it to stdout because otherwise the error would go to stderr which
will not be captured by the webserver. If it is a script run from a shell
or as a cron job, exception handling is not needed as long as you don't want 
to continue the program.

If errors should go to stdout I'd do something like the following:

import os, sys

try:
  sendmail = os.popen("/usr/lib/sendmail -t", "w")
  sendmail.write(msg)
  status = sendmail.close()
except:
  print "something went wrong:", sys.exc_type, sys.exc_value
else:
  exitcode = status >> 8
  signal = status & 0xFF
  print "sendmail exit code %d (killed by signal %d)" % (exitcode, signal)


If you really care about error messages from sendmail you should probably
go for the popen2 and select modules. But I've never used popen2.

cg.




From dextech at terrigal.net.au  Wed Apr  2 23:37:26 2003
From: dextech at terrigal.net.au (Paul Dexter)
Date: Thu, 3 Apr 2003 14:37:26 +1000
Subject: Python commands
Message-ID: <3e8bb9c5$1_1@news.iprimus.com.au>

    Hi all,
                I have just discovered Python and I was wondering if anyone
could help me find an equivalent command in Python for the "OUT" and "INP"
commands in Qbasic. I wish to convert a Qbasic program to Python.
    The following snippet is an example of the task I wish to perform;

base0 = &H378:

interface:
OUT base0, &HA0
OUT base0, &H80
OUT base0 + 2, &H8

V = (INP(base0 + 1) AND &H10) * &H8

    The base address is 378 (lpt) and the "interface" subroutine outputs
data to the port (378) and "V" is the input data statement I have used. I
also use the "GETKEY", "LOCATE" and "DRAW" commands in Qbasic

    Does anyone know how to perform the same tasks/commands in Python or are
they even available. I have come across other Basic computer language
variations that do not support the commands I require.

    Thankyou for any help,
                                        Paul







From tim_one at email.msn.com  Sat Apr 26 19:41:43 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sat, 26 Apr 2003 19:41:43 -0400
Subject: Reading a non-standard floating point format
In-Reply-To: <Xns9327331C4783Bfidtzxyzabc@195.8.68.209>
Message-ID: <LNBBLJKPBEHFEDALKOLCMEEKEIAB.tim_one@email.msn.com>

[Fidtz]
> ...
> The closest thing I have to a definition from the AlphaBasic programming
> manual: The reason for this is that floating point numbers occupy six
> bytes of storage. Of the 48 bits in use for each 6-byte variable, the
> high order bit is the sign of the mantissa. The next 8 bits represent
> the signed exponent in excess-128 notation, giving a range of
> approximately 2.9*10^-39 through 1.7*10^38. The remaining 39 bits
> contain the mantissa, which is normalized with an implied high-order bit
> of one. This gives an effective 40-bit mantissa which results in an
> accuracy of 11 significant digits.
>
> I had a go this afternoon and managed to read the sign and something
> approximating the exponent, though I will have to test the edge
> conditions carefully :)

That's helpful, but not enough to specify all the details you need to know.
Things it leaves unanswered:

+ Is an exponent field of all zeroes special?  Note that going strictly
  on what the quote said, the float 0.0 isn't representable!  (There's
  always an implied 1 bit in the mantissa, according to the quote.)

+ Is an exponent field of all ones special?

+ What's the base to which the exponent is to be raised?  2 and 16
  have both been used, even in hardware.

+ Where is the radix point in the mantissa?  Choices include "at the
  right end", between the implied 1 bit and the explicit bits (this
  is the most common choice when a 1 bit is implied), or to the left
  of the implied 1 bit.

+ Is the storage format big-endian or little-endian?

Note that 2.0**-128 ~= 2.9e-39 and 2.0**127 ~= 1.7e38.  Those are the bounds
in the quote.  From that we can deduce that the exponent base is 2.
However, if the exponent bias is really 128 (as the quote said), this is
pretty baffling, because an unbiased exponent of -128 is represented by a
biased exponent of -128 + 128 (the excess) == 0, and a biased exponent of 0
is almost always special-cased in schemes with an implied mantissa bit,
reserved to represent 0.0, and possibly also subnormals.  If this scheme
represented subnormals too, the lower bound would be smaller than 2.9e-39.
OTOH, if this scheme uses a biased exponent field of 0 to represent 0 and
the radix point is viewed as "between", the lower bound would be larger than
2.9e-39.  It's possible that the mantissa radix point is to be viewed as
being to the left of the implied 1 bit, and then we could get a lower bound
of 2.9e-39 via a mantissa of all zeroes and a biased exponent of 1.  (Then
the value is 0.5 * 2**(1-128) == 2**-128.)

The good news <wink> is that these kind of docs are often wrong, and usually
in the lower and/or upper bounds they claim.

There's no way to guess without examples, though.  If you can, show us the
48-bit byte strings for these numbers:

    0.0
    1.0
    1.25
    2.9e-39
    1.7e38

That should be enough to fill in the missing pieces.





From eppstein at ics.uci.edu  Wed Apr  2 10:42:37 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Wed, 02 Apr 2003 07:42:37 -0800
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to Python.Good idea? Bad Idea? PEP?
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com> <3E8AFB77.38D0EDDF@engcorp.com>
Message-ID: <eppstein-B80C10.07423702042003@news.service.uci.edu>

In article <3E8AFB77.38D0EDDF at engcorp.com>,
 Peter Hansen <peter at engcorp.com> wrote:

> I just scanned a few tens of thousands of lines of our code, and as near
> as I can tell only 13 cases out of over 160 instances of "for" loops are
> using range() or xrange().  This backs up my personal feeling that in
> Pythonic code, it's actually rare to iterate over an integer range 
> specified with simple upper and lower bounds like that, and certainly not
> common enough to justify special syntax.
> 
> I can't agree with the desire to make this a special case, when range()
> provides such a simple, clean answer.

On the other hand, when I'm covering dynamic programming in my 
algorithms classes, the lack of an obvious-to-the-uninitiated integer 
loop syntax (which I don't consider range or worse xrange to be) is the 
main obstacle for me from using actual Python instead of pseudocode in 
my examples.

Anyway, we've had enormous debates on this subject in this group for a 
long time, it's probably pointless to go on for too long on it again.  
For what it's worth, Guido seems to share your point of view.

And to answer the original poster's question in the subject, there 
already are several PEPs on the subject, which Guido commented 
negatively on in his "parade of PEPs" so unless he brings up the subject 
again himself I think they are dead.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From jason at tishler.net  Thu Apr  3 12:16:30 2003
From: jason at tishler.net (Jason Tishler)
Date: Thu, 03 Apr 2003 12:16:30 -0500
Subject: Cygwin Python regression testers wanted
In-Reply-To: <16012.24228.321113.363515@montanaro.dyndns.org>
References: <20030403160205.GA2896@tishler.net>
 <16012.24228.321113.363515@montanaro.dyndns.org>
Message-ID: <20030403171629.GC2896@tishler.net>

Skip.

Please post instead of sending private email.

On Thu, Apr 03, 2003 at 10:17:40AM -0600, Skip Montanaro wrote:
> I won't really be able to add anything to what you already know, since
> my Windows box runs Win2k, however, I'll work my way through the
> process just to stay in sync.  I do have a question, not being much of
> a Cygwin person.  Item 1a reads:
>     Jason> 1. rebase your system [1]
> 
>     Jason>    a. install the rebase package via Cygwin's setup.exe (if not
>     Jason>       installed)
> 
> The only setup.exe I can find in the cygwin tree is in
> /usr/local/bin/i686-pc-cygwin/setup.exe.  Is that the setup.exe you
> referred to above?

No, I was referring to Cygwin's setup.exe:

    http://www.cygwin.com/setup.exe

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From grante at visi.com  Sat Apr 26 16:46:07 2003
From: grante at visi.com (Grant Edwards)
Date: 26 Apr 2003 20:46:07 GMT
Subject: WML examples using mod_python?
Message-ID: <slrnbals94.2qp.grante@tuxtop.visi.com>

Can anybody point me towards any examples of using mod_python
for WML?

Any other python-related WML examples (normal CGI, etc.) would
also be welcome.

-- 
Grant Edwards                   grante             Yow!  ... A housewife
                                  at               is wearing a polypyrene
                               visi.com            jumpsuit!!



From mwh at python.net  Mon Apr 14 10:28:31 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 14 Apr 2003 14:28:31 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10
  <7h3vfxhpl23.fsf@pc150.maths.bris.ac.uk> <yfsu1d1cisj.fsf@black132.ex.ac.uk>
Message-ID: <7h3adettapc.fsf@pc150.maths.bris.ac.uk>

Alexander Schmolck <a.schmolck at gmx.net> writes:

> Michael Hudson <mwh at python.net> writes:
> 
> > Paul Foley <see at below.invalid> writes:
> > 
> > > [It should do that for lexical variables, too...conflating binding and
> > > assignment the way Python does is a major design error, but it's too
> > > late to fix it now]
> > 
> > I'm unconvinced about the "major"... but anyway: how would *you* do
> 
> So am I, but...
> 
> > it?  It seems unfeasible syntactically, for starters.
> > 
> 
> ... what's wrong with e.g:
> 
> def bar():
>    # declares locals, similar to ``global``
>    var a,b,c = 3
>    x = 3 # error
>    a = 4 # OK
> 
> or 
> 
> 
> def bar():
>    let x=3, y=4:
>      print x, y # OK
>    x, y # ERROR
> 
> ?

Requiring declarations of all variables would be a HUGE change to
Python, IMHO.

Cheers,
M.

-- 
  That one is easily explained away as massively intricate
  conspiracy, though.            -- Chris Klein, alt.sysadmin.recovery



From claird at lairds.com  Tue Apr  8 07:40:43 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 08 Apr 2003 11:40:43 -0000
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049722991.18682.python-list@python.org> <v938h26mdl9s8f@corp.supernews.com> <3E921B51.5030401@hotmail.com>
Message-ID: <v95d9rt698oo0c@corp.supernews.com>

In article <3E921B51.5030401 at hotmail.com>,
msw  <markscottwright at hotmail.com> wrote:
>Cameron Laird wrote:
>> I exaggerate a bit.  There are indeed legitimate roles
>> for obfuscating technologies.  Managers think they're
			.
			.
			.
>> Dave's right.
>
>Maybe, but since I'm the only one here who's read the requirements 
>document, I submit that I'm the only one qualified to suggest a software 
>design, no matter how limited my intelligence.
			.
			.
			.
>I'm right.  ;-)
>
>msw
>

Maybe both.

Perhaps I should have made it explicit that I already
have a couple of clients--hmmm; now that I think of it,
it's even more than that--who pay for obfuscated appli-
cations.  I sure didn't mean to be unhelpful, and I 
don't remember at all judging or commenting on your 
intelligence.  Here's where my experience has taught
me:
1.  I don't have anything more productive
    to say on the subject than others have
    already offered (I suspect that re-
    flects on *my* intelligence);
2.  I feel more useful relatings ways it
    can go wrong than suggesting local 
    hacks to patch up one aspect or another.

Yeah, with sufficient motivation, it's possible I'm
capable of better.  I'll repeat:  my experience in this
area has left me wary of saying more than, "you can
tart up import ...", in public.

It's a tough subject.  Good luck.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From danb_83 at yahoo.com  Sat Apr 12 14:03:18 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 12 Apr 2003 11:03:18 -0700
Subject: hex and unsigned longs
References: <xjvof3bu75x.fsf@margvisi.ifi.uio.no>
Message-ID: <ad052e5c.0304121003.7bea4848@posting.google.com>

igorr at ifi.uio.no (Igor V. Rafienko) wrote in message news:<xjvof3bu75x.fsf at margvisi.ifi.uio.no>...
> Hi,
> 
> The documentation for the built-in function hex() says that hex()
> always returns an unsigned literal. However, on my 2.2.1 python, I get
> the following results:
> 
> >>> hex( 1 )
>  '0x1'
> >>> hex( 1L )
>  '0x1L'
> >>> hex( -1 )
>  '0xffffffff'
> >>> hex( -1L )
>  '-0x1L'
> >>> 
> 
> The literal in the last case is obviously not unsigned. What am I
> missing? (I.e. why are longs treated differently than ints by hex?)

Because 0xffffffffL == 4294967295.

There's no wraparound for longs, so you can't represent a negative
long by an unsigned constant (unless it has an infinite number of
leading f's).



From gh at ghaering.de  Wed Apr 30 20:14:13 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Thu, 01 May 2003 02:14:13 +0200
Subject: Web graph producing library
In-Reply-To: <yDWra.11788$8Z5.69611@sea-read.news.verio.net>
References: <yDWra.11788$8Z5.69611@sea-read.news.verio.net>
Message-ID: <3EB066D5.50905@ghaering.de>

VanL wrote:
> Hello,
> 
> I have googled around for this, and I can't find it. [...]

Sometimes you'll find better results using a directory:

http://dmoz.org/Computers/Programming/Languages/Python/Modules/Graphics/

which not quite incidentally is the same as

http://directory.google.com/Top/Computers/Programming/Languages/Python/Modules/Graphics/

=>  PyGDChart

As a desperate googling strategy, you could have found out that the PERL 
and PHP solutions use gdchart internally, *then* search for "python 
gdchart". As a rule of thumb, you'll find the Python wrapper easily if 
you know the name of the wrapped C library :)

google-sucks-ly-yours,

-- Gerhard




From whisper at oz.net  Tue Apr 15 02:59:19 2003
From: whisper at oz.net (David LeBlanc)
Date: Mon, 14 Apr 2003 23:59:19 -0700
Subject: ANN: Capon 0.1 - a software build framework
In-Reply-To: <b7g2fk$g1u$1@panix3.panix.com>
Message-ID: <GCEDKONBLEFPPADDJCOEMEFFJIAA.whisper@oz.net>

> -----Original Message-----
> From: python-list-admin at python.org
> [mailto:python-list-admin at python.org]On Behalf Of Aahz
> Sent: Monday, April 14, 2003 21:40
> To: python-list at python.org
> Subject: Re: ANN: Capon 0.1 - a software build framework
>
>
> In article <87ptnuxryk.fsf at theasylum.dyndns.org>,
> Tim Lavoie  <tlavoie at acm.org> wrote:
> >>>>>> "aahz" == aahz  <aahz at pythoncraft.com> writes:
> >
> >    aahz> So how does Capon differ from SCons?
> >
> >Well, the name has funnier connotations, for one.   :)
>
> You deserve to be roasted for that comment.
> --
> Aahz (aahz at pythoncraft.com)           <*>
> http://www.pythoncraft.com/

A fate of many capons ;)

First they have "the" operation, then they grow up and get put on a spit and
thence over a slow fire. Life as a chicken is hell ;)

Dave LeBlanc
Seattle, WA USA





From tjreedy at udel.edu  Fri Apr 11 08:43:07 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Fri, 11 Apr 2003 08:43:07 -0400
Subject: pyorbit1.99.4  and Python2.3 problem
References: <slrnb9cuvj.49s.stehule@kix.fsv.cvut.cz>
Message-ID: <wFmdnSwh7MtWKwujXTWcpg@comcast.com>

"Pavel Stehule" <stehule at kix.fsv.cvut.cz> wrote in message
news:slrnb9cuvj.49s.stehule at kix.fsv.cvut.cz...
> I tried pyorbit with new version  of Python and when I try run test
I get
> error:
> [root at stehule pyorbit-1.99.4]# cd tests/
> [root at stehule tests]# python everything_server.py
> TypeError: MRO conflict among bases Object, B1, BaseServer, B2
> TypeError: MRO conflict among bases Object, B1, BaseServer, B2
>
> ** (process:13354): WARNING **: could not generate stub for base
> IDL:orbit/test/C1:1.0

MRO stands for 'method resolution order'  2.3 has improved the
definition and calculation of mro.  I believe the error message says
that there is a conflict or ambiguity in one of your object
inheritance trees (that was apparently not caught before).  The What's
New file should say something about this and maybe give a pointer.

> Is working orbit with 2.3 Python?

I believe that changing the order of your base classes may solve the
problem.  I hope someone else knows more about debugging this.

Terry J Reedy






From bokr at oz.net  Thu Apr 17 13:42:54 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 17:42:54 GMT
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net> <3E9D35FC.FCAD17C9@alcyone.com>
Message-ID: <b7mp2u$3k8$0@216.39.172.122>

On Wed, 16 Apr 2003 03:52:44 -0700, Erik Max Francis <max at alcyone.com> wrote:

>user at domain.invalid wrote:
>
>> What do the back ticks in method "__str__" mean?
>
>The backquotes are a way of saying "evaluate the repr of this
>expression."  That is, `x` is the same as repr(x).  Most consider the
>use of backquotes for this purpose gauche.
>
What other use is there (outside of strings and comments)?
Have we purposely been set up for gaucheries?
;-)

Regards,
Bengt Richter



From mirandacascade at yahoo.com  Fri Apr  4 12:17:00 2003
From: mirandacascade at yahoo.com (Miranda Evans)
Date: 4 Apr 2003 09:17:00 -0800
Subject: constructor inquiry
Message-ID: <59e5b87.0304040917.12c15907@posting.google.com>

When the demo() method below is run and one double clicks in a cell in
the Excel spreadsheet, the messages that get displayed are:

in OnSheetBeforeDoubleClick event
xl.event signalled
dbl click

It appears as though the constructor for the xlEv class never got
invoked, but the OnSheetBeforeDoubleClick method of the xlEv class got
invoked.  In addition, it appears as though the attribute 'type' gets
set in the OnSheetBeforeDoubleClick method and that the demo() method
has visibility to this attribute.  The fact that the print xl.type
statement results in the text

dbl click

being printed on the screen is why I believe the demo() method has
visibility to this attribute.

My questions are:
1) is it correct that the constructor for the xlEv class never got
invoked?  if so, is it correct to infer that in the sample code below,
at no time does an object of class xlEv get instantiated?
2) is it correct to infer that a class method (e.g.
OnSheetBeforeDoubleClick) can get invoked without an instance of the
class to which that method belongs getting instantiated?
3) is the 'xl' object in the demo() method an object of class xlEv? 
if so, how could the xl object come into existence without the
constructor getting invoked?  if not, how could the xl object have
visibility to the type attribute?

import time
import win32event
import win32com.client
import pythoncom

class xlEv:
	def __init__(self):
		print 'in constructor'
		self.type = ""
	def OnSheetBeforeDoubleClick (self, sh, target, cancel):
		print 'in OnSheetBeforeDoubleClick event'
		self.type = "dbl click"
		win32event.SetEvent(self.event)		
			
def demo():			
	# create an instance of an Excel app; tie it to the xlEv class	
	xl = win32com.client.DispatchWithEvents("Excel.Application", xlEv)
	xl.event = win32event.CreateEvent(None, 0, 0, None)	
	# for the purpose of this demo, make the worksheet visible and 
	# clickable; assumes user may double click in a cell
	xl.Visible=1
	xl.Workbooks.Add()	
	start_time = time.time()
	while 1:		
		rc = win32event.MsgWaitForMultipleObjects((xl.event,), 0, 1000,
win32event.QS_ALLEVENTS)
		if rc == win32event.WAIT_OBJECT_0:
			# Event signalled.
			print 'xl.event signalled'
			print xl.type
			break
		elif rc == win32event.WAIT_OBJECT_0+1:
			pythoncom.PumpWaitingMessages()			
		if time.time()-start_time > 10:
			print 'time out'		
			break



From lists at ghaering.de  Wed Apr  9 16:37:57 2003
From: lists at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 09 Apr 2003 22:37:57 +0200
Subject: string() - Is this some kind of sick joke?
In-Reply-To: <36dbf824.0304091142.527af834@posting.google.com>
References: <36dbf824.0304091142.527af834@posting.google.com>
Message-ID: <3E9484A5.1050706@ghaering.de>

Daniel Gowans wrote:
> There is some sick joke in the python 2.2.2 documentation I have. 
> This is located in the ActivePython documentation:
> 
> strip(s[, chars]) 
> [works for string methods, but not for the string module]

We've had this question asked on this list a few weeks ago, and I 
reported a bug. It should be fixed in CVS head and 2.2-maint now.

See:

http://sourceforge.net/tracker/index.php?func=detail&aid=697220&group_id=5470&atid=105470

Older related items:

http://sourceforge.net/tracker/?group_id=5470&atid=105470&func=detail&aid=656049
http://sourceforge.net/tracker/?group_id=5470&atid=105470&func=detail&aid=639170

-- Gerhard




From viszneki at telerama.com  Tue Apr  8 13:30:18 2003
From: viszneki at telerama.com (viszneki at telerama.com)
Date: Tue,  8 Apr 2003 13:30:18 -0400
Subject: atexit functionality
Message-ID: <1049823018.3e93072a10dce@webmail.telerama.com>

I am not a member of the Python mailing list, I dont' have the time to read any 
mailing lists at the moment. So if someone could please respond to me 
personally, that would be great.

Generally speaking I have a lot of objects that clean themselves up before 
python actually exits, so I was wondering if a deregister function could be 
provided.

Rather than a deregister function, I thought about making the following 
modification to atexit.py after _exithandlers.append((func,targs,kargs))

return _exithandlers[len(_exithandlers)]

My intention is to be able to store the returned value as a member variable in 
a class that registers its own cleanup functions. But if the class cleans 
itself up, the class could "clean up" its cleanup registration with atexit by 
saying something like del self.cleanup

I hope my ideas are well received! I think they will be because this 
modification is fully backward compatible, and doesn't introduce any negative 
side effects or anything like that.

Thanks for your time,
Donny Viszneki.

ps, could someone please personally respond to my email address 
viszneki at telerama.com about any decisions made about my idea? thanks a lot. 




From mcherm at mcherm.com  Thu Apr  3 09:18:32 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Thu,  3 Apr 2003 06:18:32 -0800
Subject: Comments
Message-ID: <1049379512.3e8c42b843139@mcherm.com>

Carsten Gaebler wrote:
> But multi-line strings are evil, too. At least when they occur in an 
> indented block and leading whitespace is not allowed in the string 
> your're building

Greg Ewing replied:
> Last time I pondered this issue, I decided that the
> correct way to fix this would be to have a statement
> like
> 
>    string my_string:
>      |  The string can contain any characters and
>      |have multiple lines and leading whitespace,
>      |and its indentation level is made clear by
>      |the leading | characters.
> 
> Teaching the tokenizer to understand this could
> be an interesting exercise, however...

Would this work for you?

     myString = stripLeading("""
       |  The string can contain any characters and
       |have multiple lines and leading whitespace,
       |and its indentation level is made clear by
       |the leading | characters.""")

If so, it has three advantages. First of all, no need for a
type declaration (been doing too much C lately?). Secondly,
it is clear where the construct ends (thus being a bit more 
readable if your string ended in a newline). Secondly, it 
can be implemented without having to teach the tokenizer 
anything.

     import re
     def stripLeading(s):
         prefix = re.match('\n[ \t]*.', s).group()
         return myString.replace(prefix, '\n')[1:]

I'm not sure this is the single BEST way to write indented
multi-line strings (personally I'd avoid the line of |'s),
but if you like it, it's pretty darn easy to get.

-- Michael Chermside




From jjl at pobox.com  Tue Apr 22 15:14:54 2003
From: jjl at pobox.com (John J. Lee)
Date: 22 Apr 2003 20:14:54 +0100
Subject: missing X11.h
References: <mailman.1051029574.24925.python-list@python.org>
Message-ID: <87u1cq4a4x.fsf@pobox.com>

Vince Buonassisi <vjb at vuit.com> writes:

> while compiling the source for python 2.2.2 the compilation was unable
> to locate X11/Xlib.h when compiling the file _tkinter.c.  is there a
> package that i need to install in order to get the headers so that i
> can compile X window programs?

Look for a package named something like xlibs-dev, or xlibs-devel.

On Debian, I did this to find the package name xlibs-dev:

$ locate Xlib.h
...
/usr/X11R6/include/X11/Xlib.h

...looks like it's already installed on my system, but which package?

$ dpkg -S Xlib.h
...
xlibs-dev: /usr/X11R6/include/X11/Xlib.h

(apt-file search Xlib.h should find the package even when it isn't
already installed, but my Perl installation is messed up ATM).


John



From skip at pobox.com  Wed Apr  9 13:57:50 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 9 Apr 2003 12:57:50 -0500
Subject: no Python mail archiving utility?
In-Reply-To: <joe-17FB5F.10302009042003@orngca-news02.socal.rr.com>
References: <joe-17FB5F.10302009042003@orngca-news02.socal.rr.com>
Message-ID: <16020.24350.605265.241987@montanaro.dyndns.org>

    Joe> Yesterday morning I posted asking where to find a web mail
    Joe> archive/search program written in Python.  I got nothing but the
    Joe> sound of chirping crickets in response.

    Joe> Does this mean that no such utility exists?  Are all the dozens of
    Joe> mailing lists about Python-related topics serving up their archives
    Joe> via programs written in C and Perl?

Mailman archives via Pipermail.  Since this makes the archives are available
via http, for most lists (those which are public) I suspect Google searching
is sufficient, which lowers the demand for an offline search facility.

Skip




From alloydflanagan at attbi.com  Tue Apr 22 14:41:40 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 22 Apr 2003 11:41:40 -0700
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <3EA494C2.5010701@removethis.free.fr>
Message-ID: <a8b7f07a.0304221041.9c3b97a@posting.google.com>

laotseu <bdesth at removethis.free.fr> wrote in message news:<3EA494C2.5010701 at removethis.free.fr>...
> (snip)
> > 
> <OT>
> Just for your information : your 92k VB program won't run without the 
> vbrun.dll, which is at least 1.x Mo... To be sure to have a correct 
> setup, you must package a VB program with the runtime (vbrun.dll) and 
> any 'external' ActiveX component, type libs etcs used by the application 
> (and be sure you won't write many real-life VB apps without ActiveX 
> components and type libs etc)
> </OT>
> 
> Laotseu
<OT><rant>
Yes, and it can be extremely difficult to ensure that you've got all
the ActiveX components and DLLs you need in your install package.  If
you're using database access, the client PC probably requires the
whole MS MDAC installed; they may well need to update various Windows
controls; any third-party controls need to be installed; etc.  God
help you if you've got version conflicts with another application, or
another version of VB.  All part of the wonders of component-based
development, or at least the MS implementation thereof.
</OT></rant>
I've had a lot fewer problems with distributed python apps.  Once your
target PC has python on it, and wxPython if appropriate, you can set
up everything with dist-utils.  3 easy installs is better than one
difficult one.



From sparky at gottachat.com  Fri Apr 25 17:43:20 2003
From: sparky at gottachat.com (sparky)
Date: 25 Apr 2003 21:43:20 GMT
Subject: The woes of event handling in wxPython
Message-ID: <b8ca5o$8fe@dispatch.concentric.net>

I recently took over a project from another programmer. His code is 
about 80% done and very buggy.

The one major bug that I have found in his code is the event handling.

He used wxPython's pubsub.py and evtmgr.py. Which means that I am kinda 
stuck with it.

In several places the software launches custom dialogs for the user to 
input data.

The first time any of these are launched they work fine.

However, on the second time that I launch any one of them the following
exception is raised from evtmgr.py:

wxPython.wx.wxPyDeadObjectError: The C++ part of the EditMemberDialog 
object has been deleted, attribute access no longer allowed.

Ok, I've dove into Python before, so I dove straight into the evtmgr.py 
file to find out what was going on.

It looks pretty straight forward. So I added a __del__() function to do 
clean up and unregister/unsubscribe the listeners and publishers.

This didn't work quite as planned. I continued to get the same exception.

So I started wrapping code in a try/except. It only took adding that 
twice to see that this was futile and not solving the problem.

Using the evtmgr interface to deal with cleaning up the old dead objects 
created by sending dialogs to the great bit bucket in the sky doesn't 
seem to work.

Here is some example code (sorry I cannot give you a complete class, 
they asked us not to post any code):

import pubsub, evtmgr
class EditMemberDialog(EditDialog):
     def __init__(self, parent, *args):
.
.
<snip a bunch of stuff that works>
.
.
.
	eventManager.Register(self.__saveMemberGoals, EVT_TEXT,
			      self.descriptionText)
         server.subscribe(topic    = CommandPattern.CommandHistory,
                          listener = self.__notifyObjectOfChange)

<snip a bunch of stuff that works>

     def __del__(self):

	eventManager.Deregister(self.__saveMemberGoals)
	server.unsubscribe(self.__notifyObjectOfChange)
	self.parent.__del__()



So, now when I launch the dialog the first time. It works just fine.
The second time it gets launched I get the:

wxPython.wx.wxPyDeadObjectError: The C++ part of the EditMemberDialog 
object has been deleted, attribute access no longer allowed.

What am I missing?




From cnetzer at mail.arc.nasa.gov  Tue Apr 15 22:14:21 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 15 Apr 2003 19:14:21 -0700
Subject: find non-empty value in a dictionary
In-Reply-To: <b7ib8b$3rk$1@lust.ihug.co.nz>
References: <b7ib8b$3rk$1@lust.ihug.co.nz>
Message-ID: <1050459260.11985.57.camel@sayge.arc.nasa.gov>

On Tue, 2003-04-15 at 18:26, matthew wrote:
> hi,
> 
> say you have a dictionay d =
> {'X': 0.0, 'Y'': 240.2, 'Z': 0.0, 'P': 13.3445}
> 
> where p divides pretty much evenly into Y; in this case approx 18
> 
> P will always be present and only one of X,Y,Z will have a non-zero value.

Hmmm, sounds homework-ish, but I'll byte...

> How can I easily find which X,Y, or Z is non-zero;

Well, if you can guarantee the 0.0 floating point values are exactly 0.0
(because you initialize them to 0.0, for example), then using direct
comparison is fine.

>  esp. given that 
> comparing floats is 'inaccurate' as in:

It is perfectly accurate.  It is just that floats can differ by very
small amounts, and comparision compares it they are EXACTLY equal.  If
you initialize to a specific value, you can do an exact comparision.

> 
> if X != 0:
> 	<stuff>
> elif Y != 0:
> 	<stuff>
> elif Z != 0:
> 	<stuff>
> else:
> 	<stuff-up>
> 
> Also, once the correct axis is found,  exactly the same maths will be 
> applied which means that in the above, the same statement will be 
> replicated excepting the axis variable. any ideas on the best pythonic 
> way to tackle this?

When you are repeating chunks of code, time to make a function. 
Presumably it will take an argument that represents the value of the
"axis variable", and return a result.


> any hints on best practice with comparing floats.

Well, comparing them works fine.  Floats can be large than, less than,
or equal to each other (basically).  The trick is that when people say
they want to know when a float is "equal" to zero, due to some
calculation, they often mean that they want to know when it is very
close to zero (since a calculation may return a very small result that
is not zero).

At that point, you have to decide how close is close enough to mean
"equality".

A common technique is to subtract the floats, and see if their
difference is less than some "tolerance" value. ie:

tol = 0.0000001   # this value is 1e-7

a = 0.000000000001  # A small value (1e-12)

if abs(0.0 - a) < tol:
    print "The value %s is very close to 0." % repr(a)


You can built a function to compare whether any two floats are
approximately equal:

def equal_floats(a, b, tol=1e-7):
    return abs(a-b) < tol


Knuth goes even further to allow defining tolerance as a proportion of
the largest of a or b (see Seminumerical Algorithms text)

Any further questions about floats, doubles, etc. may be answered by
Google search on "floating point FAQ".

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From http  Fri Apr 18 16:28:58 2003
From: http (Paul Rubin)
Date: 18 Apr 2003 13:28:58 -0700
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
Message-ID: <7x4r4vy2gl.fsf@ruckus.brouhaha.com>

guy at obstruction-no-spam.com (Guy Middleton) writes:
> What is the most Pythonic way to sum the n-th element of a list of lists (or
> of tuples)?
> 
> The obvious (to me) way is to use reduce, a list comprehension, and a lambda:
> 	... 
> 	>>> reduce(lambda a, b: a+b, [y[1] for y in x])
> 	6
> 	>>>
> 	$
> 
> But I don't really like lambda, is there a better way to do this?

import operator
reduce (operator.add, [y[1] for y in x])



From lyle at users.sourceforge.net  Wed Apr  2 10:47:47 2003
From: lyle at users.sourceforge.net (Lyle Johnson)
Date: Wed, 02 Apr 2003 09:47:47 -0600
Subject: swig help
In-Reply-To: <TGtia.6963$8Z5.50471@sea-read.news.verio.net>
References: <TGtia.6963$8Z5.50471@sea-read.news.verio.net>
Message-ID: <3E8B0623.8000401@users.sourceforge.net>

VanL wrote:

> I have a C file that has a function
> 
> void myfunc(int x, int y, int z, float *new_x, float *new_y, float *new_z);
> 
> I can create an interface to it fine, but the typemap stuff keeps 
> tripping me up -- I cannot seem to get it so that from python it looks like
> 
>  >>> result = myfunc(1,2,3)
>  >>> result
> (1.0, 2.0, 3.0)
> 
> Instead, it complains that it expects six arguments, the last three of 
> which need to be pointers to ints.
> 
> I have the following interface file:
> 
> %typemap(argout) float *OutValue {
>     PyObject *o, *o2, *o3;
>     o = PyFloat_FromFloat(*$1);
>     if ((!$result) || ($result == Py_None)) {
>         $result = o;
>     } else {
>         if (!PyTuple_Check($result)) {
>             PyObject *o2 = $result;
>             $result = PyTuple_New(1);
>             PyTuple_SetItem(target,0,o2);
>         }
>         o3 = PyTuple_New(1);
>         PyTuple_SetItem(o3,0,o);
>         o2 = $result;
>         $result = PySequence_Concat(o2,o3);
>         Py_DECREF(o2);
>         Py_DECREF(o3);
>     }
> }
> 
> %typemap(in,numinputs=0) double *OutValue(double temp) {
>     $1 = &temp;
> }
> 
> 
> void next(int x, int y, float z,
>           float *new_x, float *new_y, float *new_z);
> 
> 
> Could anyone tell me what I have wrong?

There are a number of problems ;)

Starting with the 'argout' typemap, the first problem is that there is 
no PyFloat_FromFloat() function, unless this is something new in Python 
2.3. The name of the function that converts from a floating-point value 
(either a float or a double) to a Python Float object is 
PyFloat_FromDouble().

The second problem is that you have a reference to 'target', which I 
think is a holdover from SWIG 1.1 days. That line of the typemap should 
instead read:

     PyTuple_SetItem($result, 0, o2);

Moving on to the 'in' typemap, the problem is that this typemap matches 
for doubles, whereas the first (the 'argout' typemap) matches for 
floats. The second typemap should presumably read:

     %typemap(in,numinputs=0) float *OutValue(float temp) {
         $1 = &temp;
     }

The final problem is that there's no way for SWIG to know that you want 
to apply those two typemaps to the "new_x", "new_y" and "new_z" 
arguments in the "next()" function declaration. Try adding these lines 
after the %typemap definitions, but before the declaration of next():

     %apply float *OutValue { float *new_x, float *new_y, float *new_z };

and then re-run SWIG, rebuild your extension and try again.

Now, having said all that, this is a pretty common case and is already 
handled by SWIG's built-in typemaps library. You could replace all of 
the above with:

     %include typemaps.i

     %apply float *OUTPUT { float *new_x, float *new_y, float *new_z };

     void next(int x, int y, float z,
               float *new_x, float *new_y, float *new_z);

and get the same result!

Hope this helps,

Lyle




From mwh at python.net  Sun Apr  6 13:05:45 2003
From: mwh at python.net (Michael Hudson)
Date: Sun, 6 Apr 2003 17:05:45 GMT
Subject: output buffering--am I missing something obvious?
References: <yu99istu9ds9.fsf@europa.research.att.com> <16013.63756.223973.484125@montanaro.dyndns.org> <mailman.1049492108.15272.python-list@python.org>
Message-ID: <7h34r5bk13q.fsf@pc150.maths.bris.ac.uk>

Andrew Koenig <ark at research.att.com> writes:

> Skip> I tried using 1.5.2 (!) on a Solaris 8 machine (it's what comes on the
> Skip> Software Companion CD for Sol8) I get the same results:
> 
> >>>> import sys
> >>>> sys.stdout.write("foo")
> foo> sys.stdout.flush()
> >>>> sys.stdout.write("bar\n")
> Skip> bar
> 
> Skip> Bit rot somewhere?  It certainly shouldn't do that.
> 
> In a follow-up message, I suggested that interactive input tossed
> unwritten output.  I'm now not so sure it's that simple, because if I
> do this:
> 
> >>> sys.stdout.write("foooooooooooooooooo"); sys.stdout.flush()
> 
> it appears to write the argument string *and then to backspace over it*,
> erasing the characters from view.
> 
> Very odd.

Now questions like the value of $TERM enter the picture... and
readline, like Paul said.

Do other programs exhibit similar behaviour?  Not sure what I'd try,
mind...

Cheers,
M.

-- 
  how am I expected to quit smoking if I have to deal with NT
  every day                                                -- Ben Raia



From heiland at ncsa.uiuc.edu  Fri Apr 25 11:03:09 2003
From: heiland at ncsa.uiuc.edu (Randy Heiland)
Date: Fri, 25 Apr 2003 10:03:09 -0500
Subject: install/import modules, /site-packages
Message-ID: <3EA94E2D.6010505@ncsa.uiuc.edu>

I need some basic Python module knowledge.  On my RH Linux installation, 
I've got the following:

/usr/local/lib/python2.2/xml
and
/usr/local/lib/python2.2/site-packages/_xmlplus

and when I >>> import xml.sax, I get:
'/usr/local/lib/python2.2/site-packages/_xmlplus/sax/__init__.pyc'

How does this module (in site-packages) take precedence over the other? 
  (I got _xmlplus by installing PyXML).  This relates to an earlier post 
of mine regarding:
AttributeError: SAX_expat instance has no attribute 'setContentHandler'


I also have:
 >>> print sys.path
['', '/home/heiland/xml_parse_test', '/usr/local/lib/python2.2', 
'/home/heiland/vtk_Dec17.02/VTK/bin', 
'/home/heiland/vtk_Dec17.02/VTK/Wrapping/Python', 
'/usr/local/lib/python2.2/plat-linux2', 
'/usr/local/lib/python2.2/lib-tk', 
'/usr/local/lib/python2.2/lib-dynload', 
'/usr/local/lib/python2.2/site-packages']

thanks for any insight and/or doc link,
--Randy




From gerritvd at decillion.net  Tue Apr 15 09:34:20 2003
From: gerritvd at decillion.net (Gerrit van Dyk)
Date: Tue, 15 Apr 2003 15:34:20 +0200
Subject: book on Python for Fortran programmers
References: <3064b51d.0304150458.595afc30@posting.google.com>
Message-ID: <3e9c0963$0$235@hades.is.co.za>

beliavsky at aol.com wrote:

> Python has been suggested by several posters to comp.lang.fortran as a
> scripting language worth learning. What are some good book(s) and web
> site(s) to get an experienced Fortran 95 programmer started in Python?

I recently found this site for python free technical books:

http://www.tcfb.com/freetechbooks/bookphyton.html

Maybe this link should also go onto the python web site.

The spelling is bookphyton and not bookpython.

Regards
Gerrit



From peter at engcorp.com  Tue Apr  1 06:06:20 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 06:06:20 -0500
Subject: [Stackless] Ann: Stackless Perl 4.0
References: <3E88EB84.7090708@tismer.com> <1b5c01c2f7f0$54712c70$0a06a8c0@FENX.COM> <mailman.1049164441.15390.python-list@python.org>
Message-ID: <3E8972AC.B4C82B05@engcorp.com>

Christian Tismer wrote:
> 
> Emile van Sebille wrote:
> > Oh god, is it April already?
> 
> Oh, is it, already?
> This was thought to be a real announce, and I carefully
> tried to get around that April issue. Whew, sorry.
> 
> I really thought it was time to inform the Python
> community that I'm now with Perl, jetzt und immerdar,
> and that I have found my final desper^H^H^Htination
> by this new paranoid^H^H^H^Hdigm, and I had a full bottle
> of white German whine, just to ensure (brrp) this.

Hmmm... "white German *whine*"?  Oh, that's just begging for it!

I think I'd better avoid pointing out your spelling error, Christian, 
or someone will invoke Godwin's Law on me!  :-)

-happy-march-thirty-second-ly y'rs,
 Peter



From giles_brown at hotmail.com  Thu Apr 10 07:47:28 2003
From: giles_brown at hotmail.com (Giles Brown)
Date: 10 Apr 2003 04:47:28 -0700
Subject: convert dictionary to string
References: <b72a7l$ifj$1@lust.ihug.co.nz>
Message-ID: <57de9986.0304100347.212d4dec@posting.google.com>

matthew <matthew at newsgroups.com> wrote in message news:<b72a7l$ifj$1 at lust.ihug.co.nz>...
> hi,
> 
> can anyone tell what is the fastest way to convert a dictionary to a string?
> 
> eg: d={'A':45.10, 'F':56.50,...}
> to 'A45.10F56.50...'
> 
> Thanks very much. matthew

Heres one option using the [c]StringIO module:

Python 2.2.1 (#34, Apr  9 2002, 19:34:33) [MSC 32 bit (Intel)] on win32
Type "copyright", "credits" or "license" for more information.
IDLE 0.8 -- press F1 for help
>>> d={'A':45.10, 'F':56.50,}
>>> from cStringIO import StringIO
>>> sio = StringIO()
>>> for item in d.iteritems():
	sio.write('%s%.2f' % item)

	
>>> sio.getvalue()
'A45.10F56.50'
>>>



From alloydflanagan at attbi.com  Fri Apr 25 16:49:46 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 25 Apr 2003 13:49:46 -0700
Subject: True/False
References: <mailman.1051110756.23235.python-list@python.org> <2c60a528.0304240517.25f2fe29@posting.google.com>
Message-ID: <a8b7f07a.0304251249.ad24819@posting.google.com>

and-google at doxdesk.com (Andrew Clover) wrote in message news:<2c60a528.0304240517.25f2fe29 at posting.google.com>...
<snip>
> 
> In this case I guess we're looking at something like:
> 
>   try:
>     True
>   except NameError:
>     exec 'False, True= 0, 1'
>     def bool(c):
>       if c:
>         return True
>       return False
> 
> Any objections?

Idlefork uses the following:

#"borrowed" from idlefork project, http://idlefork.sourceforge.net/
try:
    True
except NameError:
    import __builtin__
    __builtin__.True = 1
    __builtin__.False = 0
    from operator import truth
    __builtin__.bool = truth

This has the advantage of making True and False look more like
builtins.  The only difference is that str(True) is "1" not "True". 
Of course, if __builtin__ gets protected, this will die an ugly death.
 You could wrap it in "try: ... except Exception: pass", if you were
sufficiently paranoid.



From yuba at cyberback.com  Wed Apr  2 15:38:52 2003
From: yuba at cyberback.com (yuba at cyberback.com)
Date: Wed, 2 Apr 2003 14:38:52 -0600
Subject: (no subject)
Message-ID: <10da505acf564feba99d5ddcd154daed.yuba@cyberback.com>

I am attempting to find enough Python enthusiasts in Central Arkansas to form a user's group (we have folks running an open source, as well as a perl group).  If you're interested, drop me a note off-list.

Thanks!
--greg







From gh at ghaering.de  Tue Apr 22 13:47:16 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 22 Apr 2003 19:47:16 +0200
Subject: another python style question -- copying a data file
In-Reply-To: <zTepa.127177$Vzu.30273@news02.bloor.is.net.cable.rogers.com>
References: <zTepa.127177$Vzu.30273@news02.bloor.is.net.cable.rogers.com>
Message-ID: <3EA58024.1050909@ghaering.de>

Guy Middleton wrote:
> Ok, here's another style question.
> 
> I want to copy a data file, but if the file doesn't already exist, I don't
> care.  I do care about any other errors, such as wrong permissions.
> 
> I have the following code, is this a typical Python way to do this?
> 
>     try:
>         os.stat(DATAFILE)
>         os.system("cp %s %s.new" % (DATAFILE, DATAFILE))
>     except OSError, arg:
>         # ENOENT is ok, we didn't do the copy
>         # anything else is a problem
>         if arg.errno != errno.ENOENT:
>             raise

Here's what I'd do:

if os.path.exists(DATAFILE):
     shutil.copyfile(DATAFILE, "%s.new" % DATAFILE)

-- Gerhard




From roy at panix.com  Wed Apr 23 19:38:22 2003
From: roy at panix.com (Roy Smith)
Date: Wed, 23 Apr 2003 19:38:22 -0400
Subject: Creating dynamically named lists/dictionaries
References: <ydFpa.15$IO5.10324@newsfep2-gui.server.ntli.net>
Message-ID: <roy-9956AB.19382223042003@reader1.panix.com>

"Rogue 9" <rogue9 at ntlworld.com> wrote:
> Now what I've ended up with is a list containing one draw result
> for each week (768 at the last count)as individual items in that list e.g
> list =[[767,11,22,33,44,45,46]......[1,20,33,41,45,46,47]]

So far, this seems pretty reasonable.

> What I really want is a list or dictionary for each individual weeks
> result but I can't seem to work out how to dynamically create a list or
> dict with a unique name for each week.

The right way to think of both lists and dictionaries is that they are 
maps.  Not maps in the www.mapquest.com sense, but maps in the 
mathematical sense.  For example, let's say we were talking about the 
suits in a deck of cards.  I could make a list of them:

suits = ['clubs', 'diamonds', 'hearts', 'spades']

and now I can talk about mapping the integers 0, 1, 2, and 3 to the suit 
names 'clubs', 'diamonds', 'hearts', and 'spades'.  I give you a key, 
and you can give me back a value.

The difference between lists and dictionaries is that the keys in lists 
are restricted to be integers, while the keys in dictionaries can be 
either integers or strings (or other things, but for the sake of this 
discussion, integers and strings are all we need to worry about).

I don't know what you want to do with the data, but the list of lists 
you've already got seems like a pretty reasonable data structure to me.  
The keys seem like they should be integers, since you're talking about 
week #1, week #2, and so on.  You can already map weeks to winning 
numbers.  If I said, "what were the numbers for week 17?", you would 
just have to get list[17] (keeping in mind that lists are indexed 
starting from 0; so you might really want to get list[16]).

What is it that you want to do with the data that make you think you 
want anything more complex than what you've got?  You are right that 
creating hundreds of individually named lists or dictionaries would be a 
problem, but it's almost certainly not what you want to do.



From durdn at yahoo.it.oops!.invalid  Wed Apr  9 14:08:02 2003
From: durdn at yahoo.it.oops!.invalid (Nicola Paolucci)
Date: Wed, 09 Apr 2003 18:08:02 GMT
Subject: String splitting question
In-Reply-To: <9afd3f36.0304082245.33670341@posting.google.com>
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <6kZka.111728$sj7.4866848@Flipper>

Hello Jim,

Jim Shady wrote:
> I have a string:
> abcd/df/a/iiwk/abcdefghijkl/b/c
> I need to get the longest string between the /s of the string. For
> example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?

I didn't read all the other replies since I wanted to come up with a 
solution myself (so all excuse me if somebody suggested a similar thing):

 >>> p = 'abcd/df/a/iiwk/abcdefghijkl/b/c'
 >>> max([(len(e),e) for e in p.split('/')])[1]
'abcdefghijkl'

This did it for me.

Best regards,
	Nicola Paolucci

-- 
#Remove .oops!.invalid to email or feed to Python:
'Tmljb2xhIFBhb2x1Y2NpIDxuaWNrQG5vdGp1c3RjYy5jb20+'.decode('base64')




From eppstein at ics.uci.edu  Tue Apr 22 17:32:30 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Tue, 22 Apr 2003 14:32:30 -0700
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <3EA5B173.9420324C@alcyone.com>
Message-ID: <eppstein-D065BC.14323022042003@news.service.uci.edu>

In article <3EA5B173.9420324C at alcyone.com>,
 Erik Max Francis <max at alcyone.com> wrote:

> This would be okay (although the untyped version of the True and False
> builtins were introduced in 2.2, not 2.3).

2.2.2, I thought?  They are not in the 2.2 command-line python shipped 
with OS X, forex.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From theller at python.net  Wed Apr 30 05:27:41 2003
From: theller at python.net (Thomas Heller)
Date: 30 Apr 2003 11:27:41 +0200
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com> <b8lpof$q0f$1@pegasus.csx.cam.ac.uk> <Xns936C8AB67B739duncanrcpcouk@127.0.0.1> <vatlfeb16jlled@corp.supernews.com> <Xns936D5DA488705duncanrcpcouk@127.0.0.1>
Message-ID: <bryo1gj6.fsf@python.net>

Duncan Booth <duncan at NOSPAMrcp.co.uk> writes:

> claird at lairds.com (Cameron Laird) wrote in
> news:vatlfeb16jlled at corp.supernews.com: 
> 
> > In article <Xns936C8AB67B739duncanrcpcouk at 127.0.0.1>,
> > Duncan Booth  <duncan at rcp.co.uk> wrote:
> >                .
> >                .
> >                .
> >>The easiest solution is to leave the program in Python and just call
> >>the relevant functions from C. Possibly the easiest way to do this
> >>would be to use Pyrex.
> >                .
> >           [persuasive illustration]
> >                .
> >                .
> > ctypes <URL: http://starship.python.net/crew/theller/ctypes.html >
> > should be even easier to use.
> 
> Does ctypes support embedding? If so, it isn't obvious from that webpage.

ctypes allows to create C functions from Python callables, and pass the
pointers to them to C apis.

Normally this is used to implement "callback functions" in Python, to
call functions like Windows' EnumWindows(), or to implement COM object
method tables.

Does this count as embedding? It seems so, although I have never seen it
this way.

Thomas



From piotrus at sympatico.ca  Sun Apr 20 01:48:41 2003
From: piotrus at sympatico.ca (Peter)
Date: Sun, 20 Apr 2003 01:48:41 -0400
Subject: red hat 8.0 and python
References: <3EA21326.7070100@sympatico.ca> <mailman.1050811603.11948.python-list@python.org> <3EA2306C.9010808@sympatico.ca>
Message-ID: <3EA234B9.2080000@sympatico.ca>

Peter wrote:
> Dave Reed wrote:
> 
>> On Saturday 19 April 2003 23:25, Peter wrote:
>>
>>> I've just installed Red hat 8.0, and although python 2.2.1 comes with 
>>> it, i cannot use anything which uses the Tkinter library (this 
>>
>>
>> includes
>>
>>> IDLE). I a few other packages work, and the interpreter runs fine. I 
>>> suppose that this might be due to the setup, but I cannot figure it 
>>
>>
>> out.
>>
>>>  Could it be that Tkinter module does not come with the python 
>>> distribution which came with Red Hat 8.0? I read that it should, but 
>>> I searched my whole computer for anything with ?tk? in the name with no 
>>
>>
>> luck.
>>
>>
>> It's on one of the CDs - install the tkinter-2.2.1-17 rpm file.
>>
>>
>>
>>> I downloaded the rpm's for the python 2.2.2 (from www.python.org) but 
>>> simple "rpm -i" or "rpm -U" installing gives me errors; conflicts 
>>> with the currently installed version. I am not quite sure whether I 
>>> should try uninstalling what is there now because I've read that the 
>>> a few programs need it to run.
>>>
>>> Any suggestions? Also if anyone knows of a webpage listing the 
>>> step-by-step instructions on how to install/uninstall/upgrade the 
>>
>>
>> python
>>
>>> distribution, please reply. I have read the info provided on 
>>> www.python.org, but it is not detailed enough for me.
>>
>>
>>
>> If you just want Tkinter, install the above mentioned RPM rather than
>> going to the trouble of upgrading.
>>
>> Dave
>>
>>
> 
> Thanks Dave:
> 
> I found the file (it's on the third CD) but when I try to install it 
> (using rpm -i ) i get:
> 
> error:  Failed dependencies:
>      libtix.so.0 is needed by tkinter-2.2.1-17
> 
> I am not sure how to even begin looking where that library might be...
> 
> Any help would be appreciated..
> 
> Peter
> 

I found it. libtix.so.0 is on the same CD inside 
tix-8.2.0b1-74.i386.rpm. Now Tkinter works!

Cheers.






From mertz at gnosis.cx  Thu Apr  3 15:17:26 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Thu, 03 Apr 2003 15:17:26 -0500
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org>
 <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch>
 <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk>
 <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk>
 <x7u1dgiz1k.fsf@guru.mired.org>
 <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk>
 <x7el4jibzl.fsf@guru.mired.org>
Message-ID: <WbJj+kKkX0QX092yn@gnosis.cx>         

Mike Meyer <mwm at mired.org> wrote previously:
|As an existence proof, unless you believe that the compiler produces
|perfect code, dis-assembling (if needed) and hand-tuing the code
|produced by the compiler should get you faster results than you
|started with.

Hand-tuning only performs local optimizations.  I think it is not at all
unreasonable to suppose that an optimizing compiler produces locally
optimal machine code.  If there was a pattern that -could- be simply
rewritten, the compiler designer has probably looked for it.

Whether the program is Turing/Kolmogorov optimal globally is a different
question.  But there is no reason to think that a human can answer this
question either, in the general case.

Moreover, on the general point of whether a human CAN write better code
than a compiler, I think one needs a more contextual notion of CAN.  I
-can- perform, e.g. a simple decision process accurately, in principle.
But just because I can do a thing correctly in principle, it doesn't
mean I can do it without errors a million times in a row.  Machines come
much closer to error-free repetition of specifiable decisions.  In a
similar way, a lot of the things I *can* do manually--e.g. pound a nail,
loop a stitch, crack an egg--are nonetheless prone to an irremediable
error rate.... in some of those same cases, machines can perform the
action repeatedly with a substantially lower error rate.

Yours, Lulu...

--
Keeping medicines from the bloodstreams of the sick; food from the bellies
of the hungry; books from the hands of the uneducated; technology from the
underdeveloped; and putting advocates of freedom in prisons.  Intellectual
property is to the 21st century what the slave trade was to the 16th.




From aleax at aleax.it  Sat Apr 12 06:18:04 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 10:18:04 GMT
Subject: Macros in Python? (PEP 310)
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org> <b770t8$2b8$00$1@news.t-online.com> <b771jb$ac6$03$1@news.t-online.com> <a8b7f07a.0304111546.3e5a5636@posting.google.com>
Message-ID: <wJRla.3135$LB6.88227@news1.tin.it>

A. Lloyd Flanagan wrote:
   ...
> I think you are making this a lot harder than it needs to be.
> Remember that function definitions are first class objects in python:
> 
> def y(a, b):
>     " a simple function to do nothing significant
>     return a + b
> 
> _y = y  #save defn. of y
> 
> def x(a, b):
>     # function to wrap stuff around y
>     print "hi"
>     c = _y(a, b)
>     print "bye"
>     return c
> 
> y = x  #rebind y so it looks identical to caller

This is special-purpose code wrapping a function y callable
with two arguments, and no other -- to me, it's like answering
a question such as "can I write a Python function to sum two
numbers" with the answer "sure, 23+190".  It's not hard at all
to make a slightly umore general wrapper:

def xify(func):
    def xified(*args, **kwds):
        print "hi"
        result = func(*args, **kwds)
        print "bye"
        return result
    return xified

now,
y = xify(y)
gives you the special-case you hardcoded above, but the
(modest) generalization should be OK.

Why is this generalization still modest:
    -- func must be coded as a function (no way to pass
       xify anonymous blocks of inline code except with the
       feeble Python lambda which only allows an expression)
    -- the prefix and postfix that wrap func are hard-coded
    -- the postfix doesn't happen if func raises an exception
       rather than returning normally

Point 1 is just too hard to fix -- Python just doesn't like
treating anonymous blocks of code as first-class objects.
Point 2 is easy to ameliorate, and point 3 is trivial, e.g.:

def mkprint(something):
    def print_(): print something
    return print_

def wrapfunc(func, pre=mkprint("hi"), post=mkprint("bye")):
    def wrapped(*args, **kwds):
        pre()
        try: return func(*args, **kwds)
        finally: post()
    return wrapped

If you do this sort of thing a lot you'll no doubt want
lots of little helpers such as

def curry_to_0(func, *args, **kwds):
    def curried(): return func(*args, **kwds)
    return curried

def curry_to_anyp(func, *args, **kwds):
    def curried(**more): return func(*(args+more), **kwds)
    return curried

etc, etc (or, many of the little helpers could be grouped
into one larger helper, probably a class rather than a
closure).  But it's gonna be fussy, due to the inability
to pass anonymous blocks of inline code and the distinction
between expressions and statements.


Alex




From cpl.19.ghum at spamgourmet.com  Sat Apr 12 05:48:06 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Sat, 12 Apr 2003 11:48:06 +0200
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net>
Message-ID: <Xns935B779CE4612cpl19ghumspamgourmet@62.153.159.134>

Mads,

>Harald Massa wrote:
 
>> Yes, I love Python, and it is the most elegant programming language I
>> ever met.

> Mad wrote:
> all the self me here and self me there is - to me at last - exceedingly 
> clumsy and a constant source of frustration. 

Mad, I wrote that I love Python. Not: Python is perfect.

Especially if you look at OO: OO with Python is great. Adding/Removing 
Attributes on the fly, inspecting objects in the living system, getting 
documentation while running....

And then there is the self. It is not bad. It does not hurt your OO 
programming. It's like the nipples of a beautifull girl turning into a 
different color during intercourse. You think it's wired for the first 
time. You wonder, why is this necessary.

But sex with her is perfect. She is beautifull. She knows to dress and to 
behave with your geeky friends and with your business friends, handles 
situation in gym as well as in the opera. 
There is only that strange colour her nipples are turning when she's 
nearing orgasm.

And this strange colour of her nipples is like Pythons "self": it takes 
time to love it. 

It's not that great astonishing for var in iterator: do something. It's 
not the perfect ass of the girl, with that you fall in love immediately.

It's the strange colour the nipples turn during sex. But after some time 
you learn to love this colour ... you see this colour and know: she is 
hot.

Same happened to Python, self and me: Someday I recognized, that for 
nested classes self is a godsend: you can access the outerclass self with 
total ease, just name it differently:

class miriam:
    def __init__(self, numberofshoes):
        self.shoes=numberofshoes
    def goshopping(self):
        class necessaryitems:
            def __init__(innerself):
                if self.shoes > 10:
                    innerself.ni=["wine","grapes","cheese"]
                else:
                    innerself.ni=["shoes","wine","strawberrys"]
            def shoppinglist(innerself):
                for a in innerself.ni:
                    print a
        myneci=necessaryitems()
        myneci.shoppinglist()

a=miriam(8)
a.goshopping()

Doing similiar accesses without "SELF" is a real pain in the ass. It's 
like finding out what tearns a girl on without the colourchanging 
nipples. Once you learned to love it, you never will stop your loving.

best self-wishes

Harald
    	    	    	



From rogue9 at ntlworld.com  Tue Apr 29 00:06:50 2003
From: rogue9 at ntlworld.com (Rogue 9)
Date: Tue, 29 Apr 2003 05:06:50 +0100
Subject: Arg names in class and function code - Newbie question
References: <64e13ba5.0304281424.52cb6b5d@posting.google.com> <roy-F93739.18460128042003@reader1.panix.com>
Message-ID: <tunra.10$Ku2.3@newsfep3-gui.server.ntli.net>

On Mon, 28 Apr 2003 18:46:01 +0000, Roy Smith wrote:

>>snip snip
Hi Roy,
	I've not yet created any classes in programs yet so I got the
syntax wrong for the example but you answered me with what I wanted anyway
so thanks a lot.
	I use IDLE on Mandrake Linux 9.0 for my Python dabbling and prefer to
have the text tip pop up with a useful name when I'm calling a function
rather than the (a) or (x,y) that is appearing now.It's far more useful to
get a meaningful arg name in the text tip as I sometimes for get which arg
goes with which function and what it refers to i.e. a list,int etc.
	Once again,thanks for clearing it up, Lol McBride




From ckea25d02 at sneakemail.com  Wed Apr  9 22:30:02 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 10 Apr 2003 14:30:02 +1200
Subject: Import 'filename'? (was Re: Cross-reference 'import' in a classhierarchy)
In-Reply-To: <3E9428A8.9090005@mitretek.org>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <3e932e9f$1_2@corp-news.newsgroups.com> <b6vri8$9ob33$1@ID-169208.news.dfncis.de> <3E9428A8.9090005@mitretek.org>
Message-ID: <b72kcp$aejp6$1@ID-169208.news.dfncis.de>

Chris Reedy wrote:
> I don't like the idea of working in the file system namespace. I 
> observed a whole discussion on the GCC mailing list of when (in an os 
> independent fashion) two files are "the same".

In my opinion, it would be good enough to expand any
symbolic links (or aliases/shortcuts/whatever your platform
calls them) and check whether the pathnames are the same.

That won't always work (in the presence of mountings from
file servers, e.g.) but it can't be any worse than the
problems we already have with different names for the
same file.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From salvatore at wanadoo.fr  Sat Apr  5 12:20:42 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Sat, 05 Apr 2003 19:20:42 +0200
Subject: Stackless python
In-Reply-To: <mailman.1049558408.987.python-list@python.org>
References: <221d8dbe.0304050531.107bc603@posting.google.com> <mailman.1049558408.987.python-list@python.org>
Message-ID: <b6n3ce$pdp$1@news-reader12.wanadoo.fr>

Hello Cris,

When Stackless will be integrate in official Python ?

(I hope you'll have success with EveOnline ;-))




From ckea25d02 at sneakemail.com  Tue Apr  8 21:13:47 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 09 Apr 2003 13:13:47 +1200
Subject: Import 'filename'? (was Re: Cross-reference 'import' in a class
 hierarchy)
In-Reply-To: <3e932e9f$1_2@corp-news.newsgroups.com>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <3e932e9f$1_2@corp-news.newsgroups.com>
Message-ID: <b6vri8$9ob33$1@ID-169208.news.dfncis.de>

Chris Reedy wrote:
> I had to think about this one for a while. Here's the issue I see:
> 
> What exactly does import '../foo/ook.py' actual do?

You're being distracted by a red herring, I think. What
I took the original post to mean was doing the equivalent
thing in the Python module namespace, not the file
system namespace.

Some suitable syntax would need to be devised,
perhaps

   import ..foo.ook

although I'm not sure how you would refer to the
imported module after that... maybe an "as" clause
should be mandatory in this case?

If you want to work in the file system namespace, you
can already do that:

   ook = __import__("../foo/ook.py")

> Based on some reading in the email archives, there is clearly a desire 
> (maybe a mandate by the BDFL) that some sort of correspondence be 
> guaranteed to exist between the above four definitions (especially 1, 2, 
> and 3).

I sometimes wonder whether the Python module namespace and
the file system should be one and the same thing. I certainly
believe that PYTHONPATH is a mistake -- the environment required
for a program to function correctly is a property of the program
itself, not of the user's personal settings.

Something better is needed, I'm just not sure exactly what
yet.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From peter at engcorp.com  Thu Apr  3 21:55:53 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 03 Apr 2003 21:55:53 -0500
Subject: Beginner: How to copy a string?
References: <mailman.1048813246.30445.python-list@python.org> <d524bdb3.0304031822.2ed781be@posting.google.com>
Message-ID: <3E8CF439.D4D48B8A@engcorp.com>

dbrown2 at yahoo.com wrote:
> 
> But since everyone asked the same question "Horrors! Why in the world
> would you ever need to copy a string?" I took it as a challenge.  I
> started thinking about some not so esoteric possibilities,

Wow, it's nice to see someone trying to stand up to all that bullying
we gave him. :-)

> or hardware registers.  If the library is not written in Python (say
> it's compiled C or some other external code) then it seems the library
> could return that same pointer to the buffer or register every time,
> except it might be filled with different data. Perhaps this is not a
> correct assumption.  In any case, if I got a string buffer back from
> this kind of library I would want to copy the data to another Python
> string so I know it won't change later.

Excellent try, but I'm afraid it wouldn't work that way (unless someone
really devious deliberately tried to break how things should work).
The data returned by the external library, in order to be interfaced to 
Python, would have to create a *Python* string using calls into the 
Python interpreter (see http://www.python.org/doc/current/ext/ext.html
for more on this sort of thing).  Python would already have copied the
external data into a new area, which would then be preserved forever
in the face of everyone's futile :-) attempts to copy it.

So long as Python strings are not exposed via a pointer to some 
malicious external routine, I doubt you'll ever need to worry about
the copying thing, but I salute your attempt!

-Peter



From imbosol-1049731965 at aerojockey.com  Mon Apr  7 12:53:59 2003
From: imbosol-1049731965 at aerojockey.com (Carl Banks)
Date: Mon, 07 Apr 2003 16:53:59 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net> <mailman.1049721552.18851.python-list@python.org>
Message-ID: <H2ika.43170$wM6.23857@nwrdny01.gnilink.net>

Beni Cherniavsky wrote:
>> Next three functions are "save-current-buffer", "with-current-buffer",
>> and "with-temp-buffer".  These are actually not functions, but special
>> forms and macros.  Python has no analogue to these (barring some
>> obscure metaclass tricks).  Depending on the form, it might be useful
>> to have a corresponding method in Python, or it might be better to
>> think about a more Pythonic way to do it.  I haven't thought about it
>> too much, but off hand, I'd say I'd choose the latter in this case.
>> In fact, these seem to be mostly a convenience, so I wouldn't bother
>> to include them.
>>
> This is not a tiny issue trivial to give up.  This kind of macros is
> the heart of elisp programming.  The normal way to e.g. find another
> place in the buffer is to use functions that actually *move the point*
> and record the new position - all this inside `save-excursions'.
> Ditto for other tasks.  Elisp has a very imperative, plumbing style.
> It's very common for functions to actually change things.

This is kind of what Greg Ewing was trying to say.  I objected to the
idea that we should get rid of the notion of current buffer, but
certianly he was right on about having a "Pythonic" layer.

The idea is *not* to do excursions or switch current buffers; rather,
if you want to do something at another point or in another buffer, use
functions from the Pythonic layer that don't need to move the point or
switch the buffer.

In any case, with-style macros that don't bind new variables don't do
anything you can't do with try ... finally construct.
with_current_buffer could be replaced by:

    set_current_buffer(whatever)
    try:
        do_stuff_that_you_wanted_to_do_in_with_current_buffer()
    finally:
        set_current_buffer(original_buffer)


-- 
CARL BANKS



From fortepianissimo at yahoo.com.tw  Wed Apr  9 01:16:00 2003
From: fortepianissimo at yahoo.com.tw (Fortepianissimo)
Date: 8 Apr 2003 22:16:00 -0700
Subject: detecting modifications to an instance? (for debugging)
Message-ID: <ef08387c.0304082116.e658c31@posting.google.com>

I have a big chunk of Python code and I'd like to detect any run-time
modification to instances of a particular class.

I'd like to know how to do it with a minimal addition of code, i.e.,
it's much preferrable to make the code report any such instances
instead of using a third-party debugging tool for it.

Note I just want to detect the modification, not to prevent them
(which I think is impossible in Python).

Any suggestion is extremely welcome!



From francisgavila at yahoo.com  Sat Apr 12 02:28:09 2003
From: francisgavila at yahoo.com (Francis Avila)
Date: Sat, 12 Apr 2003 02:28:09 -0400
Subject: Unit testing "templates"?
Message-ID: <v9fcrcmbaa0eca@corp.supernews.com>

I have a class that is similar to the xdrlib.Unpacker class: you initialize
the class with some binary data, which is stuffed into an internal buffer
(with a pointer), and you then call various methods of the instance to
return chunks of interpreted data (via the struct module) until the buffer
is exhausted.

I'm now writing tests for these methods which instantiate the class with
known data and then call a method of that instance to get a known result.
But it's getting repetitive, and I'm wondering if there is a better way.

Here's a representative sample:

import unittest
import xdmcplib   #the module we're testing

class DatatypeUnpack(unittest.TestCase):
    """Verify xdmcplib.Unpacker conversions."""
    def testUnpackCard32(self):
        """unpack_card32() should return known result from known valid
input"""
        for value, data in ( (0,'\x00\x00\x00\x00'), (1,'\x00\x00\x00\x01'),
                             (2,'\x00\x00\x00\x02'),
(65535,'\x00\x00\xff\xff'),
                             (4294967295L,'\xff\xff\xff\xff') ):
            result = xdmcplib.Unpacker(data).unpack_card32()
            self.assertEqual(value, result)

    def testUnpackArray8(self):
        """unpack_array8() should return known result from known valid
input"""
        for value, data in ( ('','\x00\x00'), ('Test','\x00\x04Test') ):
            result = xdmcplib.Unpacker(data).unpack_array8()
            self.assertEqual(value, result)

Notice that the only difference between these functions is the (value, data)
pair and the function.  And there are lots of functions like this in the
DatatypeUnpack class.

Surely there's a better way?  I was thinking something like this:

class DatatypeUnpack2(unittest.TestCase):
    knownvals = { xdmcplib.Unpacker.unpack_card32:
              ( (0,'\x00\x00\x00\x00'), (1,'\x00\x00\x00\x01'),
                (2,'\x00\x00\x00\x02'), (65535,'\x00\x00\xff\xff'),
                (4294967295L,'\xff\xff\xff\xff') ) }

    def testUnpackCard32(self):
        """unpack_card32() should return known result from known valid
input"""
        funct = xdmcplib.Unpacker.unpack_card32
        for value, data in self.knownvals.get(funct):
            instance = funct.im_class(data)
            result = funct(instance)
            self.assertEqual(value, result)

Of course, this is more lines of code than before, but at least all the data
is in one place.

What I'd really like is never to have to write a test function at all: just
update the testcases dictionary and have the class "generate" one test
method for each key, docstrings and all.  Is this possible?  (It involves
metaclasses, doesn't it? *sigh*)

It seems like this would be a common sort of occurrence in a testing
framework, so is there some idiom or pattern I'm missing?  Is there any
other approach? Am I over-engineering?  Should I just forget about this, get
friendly with cut-and-paste, and get on with my life?
--
Francis Avila




From b.maryniuk at forbis.lt  Thu Apr 24 12:36:38 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Thu, 24 Apr 2003 18:36:38 +0200
Subject: What means exactly "Memory error"?
In-Reply-To: <16040.122.803504.236859@montanaro.dyndns.org>
References: <200304241406.20833.b.maryniuk@forbis.lt> <16040.122.803504.236859@montanaro.dyndns.org>
Message-ID: <200304241836.38456.b.maryniuk@forbis.lt>

On Thursday 24 April 2003 17:19, Skip Montanaro wrote:
>     Bo> Somebody knows what "Memory error" might appear and how to avoid
> (if Bo> possible) it?
> It means a memory allocation using malloc or realloc failed.  In short,
> virtual memory was exhausted.

Mmm... OK, but what else I can to do? Yes, Python process grows up to ~1.2G in 
the memory. The main code where it crashes it's like:

parsedBranches = {}
for onePart in aLotOfParts:
	parsedBranches[generateId()] = HtmlLib.FromHtml(onePart)

...and after 600-700 parts HtmlLib (4DOM) just crashes.

Any ideas?

-- 
Regards, Bogdan

Think different? I'd be happy if most people would just think...





From pobrien at orbtech.com  Sun Apr  6 10:50:46 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 09:50:46 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr>
Message-ID: <m2k7e7addl.fsf@localhost.localdomain>

laotseu <bdesth at removethis.free.fr> writes:

> Sounds like Real Fun !
> 
> Is there a way to get the current code ? (I was recently thinking of
> extending pycrust to make it a full programming environnement,
> something like a mix of Emacs and Smalltalk's programming
> environnement... Not sure if I'm able to do it, but...)

Sure.  The new editor, PyAlaMode, is in the PyCrust CVS repository on
SourceForge:

http://sourceforge.net/projects/pycrust

While it works, I plan to revise it quite a bit.  Right now it just
uses the standard Scintilla keybindings, and there is no way to change
them.  In fact, there is no way to customize it at all.  But before I
add all those kind of features I want to design the architecture and
make it based upon and semi-compatible with Emacs.  If we that and
make it easy to integrate with Bicycle Repair Man that should make it
closer to a Smalltalk type environment.  But first I need to get the
foundation in good shape before I pile on features.  I'll take all the
help I can get.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From imbosol-1050170059 at aerojockey.com  Sat Apr 12 13:55:18 2003
From: imbosol-1050170059 at aerojockey.com (Carl Banks)
Date: Sat, 12 Apr 2003 17:55:18 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
Message-ID: <aqYla.3041$883.245@nwrdny02.gnilink.net>

Mads Orbesen Troest wrote:
> Hi;
> 
> Given that strings are immutable, is it exceedingly slow (and memory 
> spiking) to do a lot of "string += string" operations in one's code? IOW, 
> does each and every += evaluation lead to a new concatenated copy of the 
> previous string (now freed for garbage collection) and the new string, so 
> that the longer the string to which stuff is appended is, the longer times 
> each += operation takes?

Yes, I think.  There are good ways to do this, though:

Build a list of strings, then join them together at the end.

Use cStringIO module.


-- 
CARL BANKS



From drs at ecp.cc  Fri Apr 11 01:34:44 2003
From: drs at ecp.cc (drs)
Date: Fri, 11 Apr 2003 05:34:44 GMT
Subject: Python & Remote Access Database
References: <a6b6f195.0304090408.21003c3a@posting.google.com> <kjYka.2914$kd1.2980829@newssrv26.news.prodigy.com> <a6b6f195.0304100351.54a1bef9@posting.google.com> <7enhm-os2.ln1@beastie.ix.netcom.com>
Message-ID: <Utsla.501$ug7.261464@newssrv26.news.prodigy.com>

"Dennis Lee Bieber" <wlfraed at ix.netcom.com> wrote in message
news:7enhm-os2.ln1 at beastie.ix.netcom.com...
> mat fed this fish to the penguins on Thursday 10 April 2003 04:51 am:
>
> >> Assuming that by access you mean MS Access, I think there are 3
> >> solutions. The real problem is that Access is not a db server, so
> >> using it remotely, even in an all windows environment, is not a
> >> simple thing (or at least not a good thing).
> >>
>
>         Access (actually JET -- Access is more of a GUI front-end to JET)
is
> described as something like "file server RDBM"; all the processing is
> normally done by running the JET engine on the local machine, and
> copying the .MDB file(s) across the network as needed. Of course, JET
> only runs on M$ Windows (as a DLL) so you can't get it from a Linux
> program.

Generally this is correct, but wrapping calls to Access/Jet in a remote
object keeps the processing remote.  Thus using pyro should allow a local
process on linux to manipulate an object on windows which would remotely
interacted with an access db.  As I said, it is neither a simple nor a good
way to proceed, but it should work in a pinch.

-d





From majbritt37 at hotmail.com  Sat Apr 12 08:17:13 2003
From: majbritt37 at hotmail.com (=?ISO-8859-1?Q?M=E5rten_Ask?=)
Date: 12 Apr 2003 05:17:13 -0700
Subject: PIL binary for windows?
Message-ID: <fe1cb5be.0304120417.5b756718@posting.google.com>

Does anyone know where I can find a PIL binary for windows and Python
2.3? Or maybe some help in building it myself?

When compiling, I get the error: 

LINK : fatal error LNK1181: cannot open input file "Imaging.lib"

How do I build the imaging.lib file? 

Everything else seems to be ok, it's just the imaging.lib file that's
missing. Help?



From danb_83 at yahoo.com  Thu Apr 10 19:24:31 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 10 Apr 2003 16:24:31 -0700
Subject: Python compared toJava
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
Message-ID: <ad052e5c.0304101524.62be38b@posting.google.com>

blaktyger at hotmail.com (Blaktyger) wrote in message news:<c9e6ac92.0304100824.eec0f5a at posting.google.com>...
> Hi!
> 
> I recently read that the Python interpreter is like the Java Virtual Machine.
> Is that true?

Depends on which Python implementation you were referring to.  With
Jython, the interpreter *is* the Java Virtual Machine.

If you meant CPython, the answer depends on your definition of "like"
(or "is" ;-).



From donn at drizzle.com  Sat Apr 19 10:34:18 2003
From: donn at drizzle.com (Donn Cave)
Date: Sat, 19 Apr 2003 14:34:18 -0000
Subject: make test fails on NetBSD 1.6
References: <ca3cce40.0304181809.5f704e9b@posting.google.com> <m3y9279dga.fsf@mira.informatik.hu-berlin.de>
Message-ID: <1050762858.161294@yasure>

Quoth martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=):
| carlggb at terc.edu (Carl Alexander) writes:
|> I'm new to Python.  Building 2.2.2 on NetBSD 1.6 seems to go fine, but I get
|> this from make test.  It's not quite clear to me how I should go about asking
|> "someone to teach regrtest.py about which tests are expected to get skipped 
|> on netbsd1".
|
| If you consider yourself a NetBSD expert, you should determine which
| of these tests should have worked: e.g. should it be expected that
| NetBSD supports the en_US locale (in which case test_locale should
| have worked); should it be expected that the system supports fork with
| threads (in which test_fork1 should be executed), and so on.
|
| If you have a list of tests that you would expect to be executed, find
| out why the tests that got skipped did so; for the rest, submit a
| patch to sf.net/projects/python to make an entry in
| Lib/test/regrtest.py to list the tests that are expected to be skipped.
|
| You should be concerned about the failure of test_threading; it means
| that your thread library doesn't work as Python expects it to work. In
| that case, it is best to either fix the problem, or compile Python
| with --without-threads.

Actually NetBSD is still getting around to supporting threads.
In the meantime, Python shouldn't expect the thread library to
be present.  When I run regrtest.py, I get "test_threading skipped --
No module named thread" (and of course several other tests skipped
for the same reason.)  But I see that's with 2.2.1.  Who knows,
anyway, one should not expect to be able to use the threading module
on NetBSD through 1.6.

	Donn Cave, donn at drizzle.com



From martin at v.loewis.de  Sun Apr 13 06:15:55 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 12:15:55 +0200
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <bJ2ma.5447$883.4511@nwrdny02.gnilink.net> <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net>
Message-ID: <m33ckmwvms.fsf@mira.informatik.hu-berlin.de>

Mads Orbesen Troest <mads at troest.NEVERMORE.dk> writes:

> But is that /guaranteed/ (as it is with STL), or does it just happen to be 
> that way with the current implementation (however unlikely it is to 
> change)?

The Python language "specification" makes no performance guarantees
whatsoever, primarily because it is a description of the current
implementation, and not an abstract prescription for implementations
of "the Python language". IOW, there is no Python language
specification.

Regards,
Martin




From shalehperry at attbi.com  Sat Apr  5 14:24:14 2003
From: shalehperry at attbi.com (Sean 'Shaleh' Perry)
Date: Sat, 5 Apr 2003 11:24:14 -0800
Subject: List of Numbers
In-Reply-To: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
Message-ID: <200304051124.14888.shalehperry@attbi.com>

On Saturday 05 April 2003 11:13, Simon Faulkner wrote:
> I have a list of about 5000 numbers in a text file - up to 14 digits
> each.
>
> I need to check for duplicates.
>
> What would people suggest as a good method?
>
> Simon

the canonical approach is is to place the numbers in a hash (dictionary in 
python) where the value is the key and stored data is the number of 
occurences.

for each number
   see if dict has number as a key
   if it does, increment the count
   else add the number to the dictionary and set the count to 1

then you just look for any dictionary elements with a count greater than 1.
Or if you just want uniques, you just ask the dictionary for its keys() list.




From pan-newsreader at thomas-guettler.de  Mon Apr  7 13:52:22 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Mon, 07 Apr 2003 19:52:22 +0200
Subject: Updating CVS
References: <5Nhka.139$N5.25540@mencken.net.nih.gov>
Message-ID: <pan.2003.04.07.19.52.22.351156.788@thomas-guettler.de>

On Mon, 07 Apr 2003 18:35:13 +0200, Mike wrote:

> Hello,
> 
> I wrote the following python web script to update a CVS directory and I
> don't receive any errors when executing it; however, the directory
> doesn't update.
> 
> #!/usr/bin/python
> 
> import cgi, os, urllib, sys

Since you import cgi, I assume you are writing a cgi script.

Which web server do you use?

Try running the script from the shell.

Have a look at the error log of you web server.

You can run the script in a wrapper like this. This
result in error pages which are displayed in the browser:

#!/usr/bin/python
import cgi


def main():
    print "Content-Type: text/html\n"
    try:
                import quizpy
                print quizpy.main()
    except:
                print "<html><body><pre>"
                cgi.print_exception()
                print "</pre></body></html>"
            
main()


 thomas

BTW: The subject of your post is not very good, since it is a cgi problem.

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From aleax at aleax.it  Sat Apr 12 13:19:53 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 17:19:53 GMT
Subject: Adding item in front of a list
References: <mailman.1050038237.25551.python-list@python.org>
Message-ID: <ZUXla.7029$T34.191442@news2.tin.it>

Tim Peters wrote:
   ...
> reverted).  Saving you from needing to type 0 when you mean 0 doesn't
> outweigh the ugliness of making the first argument mean different things
> depending on how many arguments there are.

Are we talking about the same language, Tim?  I.e., Python, where
range(7) and range(0,7) are synonyms...?  Why can't mylist.insert
use just the same "first argument means different things" convention
as range...?

I'm not necessarily arguing for this, but given the looming precedent
of range and xrange I do find your motivation confusing...


Alex




From gh at ghaering.de  Thu Apr 24 10:56:21 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Thu, 24 Apr 2003 16:56:21 +0200
Subject: serveur TFTP
In-Reply-To: <e4d8f90f.0304240622.64cf7f8e@posting.google.com>
References: <e4d8f90f.0304240622.64cf7f8e@posting.google.com>
Message-ID: <3EA7FB15.90500@ghaering.de>

alex wrote:
> j'ai recup?r? une librairie TFTP ecrit en python mais j'avoue avoir un
> peu de mal ? l'utiliser.comment faire un serveur TFTP avec? j'essai
> d'identifier les methodes que je pourrais appliquer a un objet
> mon_serveur par exemple :
> [...]

Ouf! C'est beaucoup de code et je ne sait rien de TFTP. Si tu lis ce 
group un peu (ce qui est recommand? avant d'y ?crire ;-) tu va trouver 
que la langue pr?fer?e est l'anglais. Est-ce que tu peut reposer ta 
question en anglais?

Pour la discussion de mati?res pythonesques en francais, je recommende 
le groupe fr.comp.lang.python.

En esp?rant que mon francais soit compr?hensible,
amicalement,

-- Gerhard





From tim_one at email.msn.com  Sat Apr  5 01:56:12 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sat, 5 Apr 2003 01:56:12 -0500
Subject: My Object pool isnt working
In-Reply-To: <3E8DD5B9.501@freemail.hu>
Message-ID: <LNBBLJKPBEHFEDALKOLCKEMOEEAB.tim_one@email.msn.com>

[Nagy L?szl? Zsolt]
> Whew! I always thought that I must use locks to access any object in
> Python before I modify it from a thread.

That belief won't get you in trouble, so if you're happy with it, stay there
<wink>.

> Is it possible to append a list or a dictionary from a thread without
> locking it?

Yes.

> What does thread-safe mean here?

Informally, that Python won't blow up -- the internal data structures remain
sane (self-consistent) whether or not you lock.  More formally, they supply
what's called "sequential consistency" in the literature:  the final result
is what you'd see if all operations were carried out one at a time, in order
within each thread's sequence of operations, but where the order across
threads isn't specified.  Internal sanity follows from that, but it's a weak
guarantee for most application-specific needs.

> For example, what would happen if I walk through a list in a for loop
> while another thread modifies the list?

In the thread that's walking the list, you would see an unspecified mixture
of the list's original state, with an unspecified subset of (zero or more
of) the modifications made by the other thread.  Whether that's a strong
enough guarantee for your app's needs depends entirely on what your app
needs, which isn't meant to be flippant.  Sometimes weak guarantees are
adequate, and then giving up stronger guarantees in return for avoiding
locking can be a good tradeoff.  When in any doubt, serialize.  Better
still, use the Queue module and don't play with locks at all yourself (e.g.,
the OP's problem (a pool of objects that expire over time when unused) is
much more easily solved with a Queue.Queue than via the hand-rolled
dicts+semaphores business posted).





From skip at pobox.com  Wed Apr  9 11:16:53 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 9 Apr 2003 10:16:53 -0500
Subject: Adding item in front of a list
In-Reply-To: <pan.2003.04.09.16.59.35.949871.811@thomas-guettler.de>
References: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de>
        <mailman.1049892916.24803.python-list@python.org>
        <pan.2003.04.09.16.59.35.949871.811@thomas-guettler.de>
Message-ID: <16020.14693.189073.740461@montanaro.dyndns.org>

    Thomas> insert(0, 1) behaves like insert(-1, 1). It is better, but
    Thomas> insert(1) would be best. 

.insert() is more general than what you want.  It can insert elements
anywhere in the list.  You are asking for something like .prepend().  For
that, just subclass list and add that method:

    class mylist(list):
        def prepend(self, obj):
            return self.insert(0, obj)

I hope you realize that the signature of .insert() can't be changed, beauty
or not.  *Way* too much code would have to be changed.

Skip




From mcfletch at rogers.com  Sun Apr  6 14:17:05 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 06 Apr 2003 14:17:05 -0400
Subject: properties and get/set methods
In-Reply-To: <v90of6avvh1leb@news.supernews.com>
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org> <v908s7echa191b@news.supernews.com> <mailman.1049643850.17962.python-list@python.org> <v90of6avvh1leb@news.supernews.com>
Message-ID: <3E906F21.5000802@rogers.com>

John Roth wrote:

>"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
>news:mailman.1049643850.17962.python-list at python.org...
>  
>
...

>>When people start using properties heavily, the definition migrates
>>closer to "describing a property/attribute slot of an object" with the
>>access mediation of getter/setters being a side-effect of that
>>description.  So after a while, the people using them start seeing
>>    
>>
>them
>  
>
>>in a different light than those for whom they are still just for
>>wrapping getter/setter methods.
>>
del Mike's continued blathering...

>I agree, that's a better solution although there is one big difficulty.
>The approach works quite well in Java, where you've got the ability
>to declare a variable as "final." You don't have that ability in Python,
>so you run the risk of someone accidentally rebinding the field rather
>than
>using the extra level of indirection needed to access it correctly.
>
I haven't had a problem with needing final yet.  I can certainly imagine 
cases where it would cause a problem, but I'm one of those old-style 
Python programmers.  I figure if you're sub-classing a class you better 
know what it's fields are before you start sub-classing.  You could do 
something about it by defining properties on your meta-classes, but as I 
say, I haven't felt the need to do it yet.

>Following on to this idea, would you say that the property syntax
>introduced in Python 2.2 was a mistake? It should have been more
>general?
>
Fixes I can easily imagine being useful:

    * add a parameter to get/set/del methods telling them which
      attribute name is being used to access them
          o possibly do this as another hook level to leave current code
            as-is, but it really does seem like something to which the
            methods should have access
    * provide C-level (i.e. fast) standard get/set/del implementation
      functions that support both __dict__ and __slots__ based objects
      for use when you just want to say "do what you'd normally do"
      (i.e. when you only want to override set and del, not get (which
      is normally far more common and speed-sensitive))

I'm not, however, saying the 2.2 properties were a mistake.  If I recall 
correctly, they were introduced as a "provide something now, worry about 
details (such as potentially a syntactic wrapper) later" feature.  
They're a very useful feature even with a few little warts, and I'd much 
rather have them as-is than not have them (which would likely have been 
the other option).  The API could have been slightly more general, true, 
but oh well.

Enjoy,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From eppstein at ics.uci.edu  Thu Apr 17 22:41:09 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Thu, 17 Apr 2003 19:41:09 -0700
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <3E9DDC81.932475AF@alcyone.com> <3ckgewva.fsf@morpheus.demon.co.uk>
Message-ID: <eppstein-BB4C70.19410917042003@news.service.uci.edu>

In article <3ckgewva.fsf at morpheus.demon.co.uk>,
 Paul Moore <gustav at morpheus.demon.co.uk> wrote:

> But for quicksort you partition, and compare each element against the
> partitioning element. That's n comparisons against log(n) partitioning
> elements (each partitioning step halves the size of the chunks, so you
> do this log2(n) times). So the complexity is O(n*log(n)).
> 
> Disclaimer: This is all *very* rough, and from memory as well. But I
> hope it's enough that you get the idea.

This is not completely accurate, since each partition is not usually 
exactly in half.  My favorite version of the quicksort analysis is in 
the new Cormen-Leiserson-Rivest-Stein version of Intro to Algorithms:
number the items according to their eventual sorted positions, then item 
i is compared to item j exactly when the first pivot in the range i..j 
is either i or j, which occurs with probability 2/|i-j+1|. Sum up these 
probabilities for all i<j to get the expected number of comparisons and 
you get roughly 2 n H_n (harmonic numbers) = O(n log n).

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From DennisR at dair.com  Mon Apr 21 16:49:30 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Mon, 21 Apr 2003 20:49:30 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org>
Message-ID: <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com>

"Jp Calderone" wrote
>  It is unlikely you'll be able to get your app down to 800kB, though 3MB
or
> 4MB may be likely.  If size is your main concern, I'd recommend simply
*not*
> bundling the Python runtime with your application.  This will give you a
> "Hello world" that is 1kB or 2kB.

On Windows, which is where Visual Basic runs after all,  a complete Python
hello world is 459 KB.   3-4 MB is *not* required.  See the online builder
for constructing Python hello-world class executables at
http://www.dair.com/autorun/arpyexe.cgi

A serious app (SpamAI below) is built differently and has much more code
than hello world.  Current size is 797 KB.  800 KB *is* likely.

The python DLL is bundled as are all required PY files in both of the
executables above.
--

Dennis Reinhardt

DennisR at dair.com   http://www.spamai.com?ng_sig





From ajsiegel at optonline.net  Mon Apr 21 18:34:30 2003
From: ajsiegel at optonline.net (Arthur)
Date: Mon, 21 Apr 2003 18:34:30 -0400
Subject: instance creatiion as an "event"
Message-ID: <000701c30856$2e056120$0c02a8c0@Arts>

Alex writes-

>I'm somewhat confused, I think -- is your problem:

>[A] how do I distinguish whether an operation is being performed "at   the
interactive prompt" or elsewhere?

>and/or is it:

>[B] how do I find out when an instance of a class has been created (quite
apart from the issue of whether it's created at the interactive prompt or
within some module),

>and/or is it:

>[C] how do I find out all the existing instances of a given class at any
given point (during an interactive session), whether that is the time  of
the creation of a new >instance,or not

>Each of these tasks is quite feasible -- some trivially easy (B), some
devilishly hard and tricky (A), some not too bad (C).  But unless you
clarify which one (or more) of >these tasks you want to perform, it's quite
hard to offer you appropriate suggestions

Let me try to specify.

First, it is necessary to understand that the VPython (visual) module is
capable of immediate mode rendering from the intereactive prompt.

>>> from visual import *
Visual-2003-03-15
>>> s=sphere(pos=(1,2))   # a display window is opened and a sphere at
coords (1,2) is rendered
>>> s.pos=(-2,1)                #the sphere moves to coords (-2,1)

My app is using VPython as a rendering backend, with the rendering code
wrapped is classes.

So I have a module, say,  "test_point.py"

"""

from visual import sphere

points=[]
class Point:
   def __init__(self,x,y,z=0):
      points.append(self)
      self.x=x
      self.y=y
      self.z=z
      self.sphere=sphere()

def  find_coords(self):
    pass

def redraw(self):
     self.find_coords()
     self.sphere.pos=(self.x,self.y)

"""

And a module "test_draw.py"


"""

# a gui module, which includes, in essense:

from test_point import *

def draw_all():
   for point in points:
      point.redraw()

if __name__=='__main__':
   p1=TestPoint(1,2)
   p2=TestPoint(-2,-1)
   draw_all()

"""

"python test_draw.py"

and I have 2 points, one at (1,2) and one at (-2,-1).

In the "real world", a point might have 3 planes as its construction
arguments, so that there is some work to be done in coming to the (x,y,z)
coords as the point of intersection of the planes.And there is a factory
function to analyze the given arguments and call the child of an  abstract
Point class that accepts the given arguments and processes them with the
appropriate "find_coords" method.

All is well, until I decide I want to recover some of VPython's immediacy at
the interactive prompt.

>>> from test_draw import *
Visual-2003-03-15
>>> p1=TestPoint(1,2)
>>> p2=TestPoint(-2,1)
# what I have is 2 spheres at default location (0,0,0) - the result of
self.sphere=sphere()
>>> draw_all()
# which now puts the points where they belong.

All I am trying to accomplish is eliminate the need for the explicit call to
draw_all at the command prompt after creating each instance.

I could, of course, add self.redraw() to TestPoint.__init__.  But it seems,
because of the overall "real world" design where TestPoint is somewhere down
an inheritance chain, I need to call this explicitly on the top level
class - and need to do some restructuring of the inheritance scheme, because
if self.redraw is in a base class's __init__ and that base class's
find_coords methods deals with attributes specific to that class - then
super.__init__ ends up with attribute errors, despite the fact that the
find_coords method is overriden in the child.  This was somehow surprising
to me - and where I might be missing something easy and fundamental.

Being stumped there, and lazy, I was looking for some interactive magic.

so that p1=TestPoint(1,2) would somehow register as an event calling
draw_all().

I on one hand want an easy answer.  On the other hand hoping it is not *so*
easy that I will have embarrassed myself by having made a mountain of a
molehill.

Art








From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Mon Apr 14 13:42:57 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Mon, 14 Apr 2003 19:42:57 +0200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <3E9A4E32.F179FDDC@alcyone.com>
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com>
Message-ID: <3e9af32e$0$49104$e4fe514c@news.xs4all.nl>

[offtopic]

Erik Max Francis wrote:

>>>>Course(lowEarthOrbit, lowMarsOrbit).transfers()
> 
> [<ExtractionTransfer @ 0x82893ac (out of `low orbit around `Earth'')>,
> <HohmannTransfer @ 0x812cef4 (Hohmann from `Earth' to `Mars')>,
> <InsertionTransfer @ 0x82adb64 (into `low orbit around `Mars'')>]
> 

Is this actual code? What does this do? I'm interested.

--Irmen




From mark at metaconcepts.com  Tue Apr  1 05:32:35 2003
From: mark at metaconcepts.com (Mark Brady)
Date: 1 Apr 2003 02:32:35 -0800
Subject: [Stackless] Ann: Stackless Perl 4.0
References: <3E88EB84.7090708@tismer.com> <mailman.1049161621.6631.python-list@python.org>
Message-ID: <56443b6.0304010232.557ceadf@posting.google.com>

I can't believe I nearly fell for this!!! 
After 30 seconds of shocked disbelief I remembered the date and that
I swore I wouldn't fall for anything this year ;-)

Mark.

"Emile van Sebille" <emile at fcfw.fenx.com> wrote in message news:<mailman.1049161621.6631.python-list at python.org>...
> Oh god, is it April already?
> 
> --
> 
> Emile van Sebille
> emile at fenx.com
> 
> ---------
> ----- Original Message -----
> From: "Christian Tismer" <tismer at tismer.com>
> To: "Pythonistas" <python-list at python.org>; <stackless at tismer.com>
> Sent: Monday, March 31, 2003 5:29 PM
> Subject: [Stackless] Ann: Stackless Perl 4.0
> 
> 
> > Dear Python/Perl Community,
> >
> > Insiders have known this from the beginning, and they were waiting
> > for this to happen for much too long time:
> >
> > Stackless 3.0 has been dead-born from the beginning.
> > The new live language is Stsckless Perl 5.0!
> >
> > This is not due to the nature of Stackless, it is due to the
> > nature of Python. There is a better language than Python,
> > and its name is PERL!



From oblivious at web.de  Fri Apr 11 14:31:25 2003
From: oblivious at web.de (Dominic)
Date: Fri, 11 Apr 2003 20:31:25 +0200
Subject: Macros in Python? (PEP 310)
In-Reply-To: <b770t8$2b8$00$1@news.t-online.com>
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org> <b770t8$2b8$00$1@news.t-online.com>
Message-ID: <b771jb$ac6$03$1@news.t-online.com>

Dominic wrote:
> Beni Cherniavsky wrote:
> 
>> Dominic wrote on 2003-04-09:
>>
>>
>>> (defmacro with-root-privs (() &body body)
>>>   (let ((oldidsym (gensym)))
>>>     `(let ((,oldidsym (geteuid)))
>>>        (seteuid 0)
>>>        (unwind-protect
>>>        (progn , at body)
>>>      (seteuid ,oldidsym)))))
>>>
>>> Is there an easy way to produce similar code
>>> in Python which gurantees that certain things
>>> happen around some code? (Maybe by using
>>> generators?)
>>>
>>
>> To complete the list of solutions, there is an evil way to abstract
>> try..finally that is not recommended due to it's unreadability to
>> anybody that doesn't know the trick.  It also only works in CPython
>> with reference counting that invokes __del__ immediately.
>>
>> The basic idea was to use a for loop and a generator:
>>
>> def with_root_privs():
>>     oldid = geteuid()
>>     seteuid(0)
>>     yield None # dummy
>>     seteuid(oldid)
PEP 310 is actually what I asked for in the first.
Well probably it is better to introduce some more basic
constructs which allow it to build such constructs.
Otherwise Python may end up with too much keywords etc.
Continuations and/or macros may be a possibility.
But I do not have enough experience and knowledge
to actually judge about such decisions. So I let
it up to the Python developers.
So far they've done a great job and I'm very happy with Python though
I wish it would have macros&continuations like
e.g. scheme/lisp ;-) . (Continuations would also allow
it to build low-overhead concurrency control flow,
instead of os-level threads)
Nevertheless those things are nice to have but
not necessities.

Ciao,
  Dominic




>>
>> for _ in with_root_privs:  # `_` is dummy
>>     # Dohere whatever you wanted.  The generator will cause this to be
>>     # executed exactly once, with pre- and post- code excuted...
>>
>> The problem, of course is that the for loop's body can raise an
>> exception, or exit the loop in other evil ways.  Luckily in CPython,
>> exiting the loop immediately drops the only reference to the
>> generator, so hooking it's __del__ would do the trick.  Writing it as
>> an iterator class instead of a generator would be be ugly but we can
>> only do it once, by a universal wrapper taking a naive generator and
>> returning a safe one.  Google for "abstracting try..finally with
>> generators" in c.l.py if you are interested.  Again, it's hard to
>> recommend this hack.  Instead, consider pushing PEP 310 :-).
> 
> 
> Actually I also considered using generators but I would never have
> come by this idea to use ref-counting side effects to make
> it work. ;-) This one is really great though it shouldn't be used...
> Now I'll see what PEP 310 is all about.
> 
> Ciao,
>  Dominic
> 
> 
> 
> 
> 




From mcfletch at rogers.com  Tue Apr  8 15:19:29 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 08 Apr 2003 15:19:29 -0400
Subject: Is there any way to make repr(aString) use double quotes?
In-Reply-To: <+6xk+kKkXQeJ092yn@gnosis.cx>
References: <698f09f8.0304081057.1ba7041f@posting.google.com> <+6xk+kKkXQeJ092yn@gnosis.cx>
Message-ID: <3E9320C1.1000903@rogers.com>

Lulu of the Lotus-Eaters wrote:

>tweedgeezer at hotmail.com (Jeremy Fincher) wrote previously:
>|If it's not, then is it possible to achieve the same thing (everything
>|escaped properly, etc.) through other means?
>
>It's not beautiful, but this does what you want:
>
>    '"%s"' % repr(aString)[1:-1]
>  
>

Except when there's a " character in the string ;) .

def SFString_vrmlstr( value, lineariser=None):
    """Convert the given value to a VRML97 representation"""
    return '"%s"'%(
        '\\"'.join(
            '\\\\'.join(
                value.split('\\')
            ).split('"')
        )
    )

But that doesn't do any escaping of characters (e.g. \xff), so there's 
more to be done :) ,
Mike






From fperez528 at yahoo.com  Wed Apr  2 18:23:16 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 02 Apr 2003 16:23:16 -0700
Subject: A few errors in "Python in a Nutshell" - Numeric related
Message-ID: <b6frd5$uq$1@peabody.colorado.edu>

Hi,

I recently got Python in a Nutshell, and as I expected, I've found it to be
extremely informative and well written.  I haven't read it in detail, but
I'm glad I got it.

However, I'd like to point out a few problems I found in Chapter 15, Numeric
processing.  I did not read it in excruciating detail, so please don't
consider this an exhaustive bug hunt of that chapter.  And by the way, that
chapter is a fabulous reference to NumPy, far better for most things than
the standard manual which is organized in a 'cute' (read random) way and
for some things is nearly useless.  Ch. 15 actually explains many things
which the NumPy docs gloss over, and the function reference is
*alphabetized*!! (how hard would it be to do that in the normal manual??)

Anyway, my few nits for future reference:

1) p 306 - a[i][j] vs a[i,j].  This one isn't so much an error but rather a
possibly misleading subtlety.  In p. 306 it says

"... if a's rank is at least 2, you can write a[i][j] as a[i,j] for any
valid i and j..." 

This is, of course, perfectly true and valid.  However, I've seen many
people bitten by thinking that [foo][bar] == [foo,bar] in all cases, and it
is not.  Here's a simple example:

In [1]: z=zeros((3,4,5))

In [2]: z[:][0].shape
Out[2]: (4, 5)

In [3]: z[:,0].shape
Out[3]: (3, 5)

Obviously, in the first case the [:] simply makes a dummy reference to all
of z and the [0] picks its first element along its first axis, giving an
array of shape (4,5).  In the second case, we get the first element along
the _second_ axis, hence an array of shape (3,5).  It might have been
worthwhile pointing this out, as it's a trap that newcomers fall into
repeatedly, especially those with a C/C++ background (a significant chunk
of NumPy's audience).



2) p 312 - flat

In page 312, it says that

"a.flat is an attribute that contains an array with rank of one less than
a..."

This is actually only true for rank 2 arrays.  In general, a.flat is
_always_ a rank 1 array, as long as a is contiguous:

In [4]: z.flat.shape
Out[4]: (60,)

a.flat is truly a 'flat' view of a's data, not just a 'rank(a)-1' view.


3) p 320 - ravel

In p. 320, it says 

"ravel (a) returns the flat form o f a, just like array(a,copy=False).flat"

This is true _only_ if a is contiguous.  The whole point of using ravel is
that it works even if a is non-contiguous:

In [8]: z1=z[::2,::2,::2]

In [9]: z1.iscontiguous()
Out[9]: 0

In [10]: ravel(z1)
Out[10]: array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])

In [11]: array(z1,copy=0).flat
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)

?

ValueError: flattened indexing only available for contiguous array


So the above would be true if it said "... just like
array(a,copy=True).flat".  In that case it would be correct, as a fresh
copy will always be contiguous, hence with a valid .flat attribute.


Anyway, these are just some small nits which readers which use Numeric may
want to be aware of.  If there is an errata page for the book, it may be
worth adding them to it.

Cheers,

Fernando



From aleax at aleax.it  Sat Apr 12 15:25:44 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 19:25:44 GMT
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <b73o7m$lci$1@panix1.panix.com>
Message-ID: <YKZla.7613$T34.208546@news2.tin.it>

Aahz wrote:

> In article <b7340h0r7m at enews1.newsguy.com>,
> Alex Martelli  <aleaxit at yahoo.com> wrote:
>>
>>__slots__ is, essentially, an optimization (in terms of memory
>>consumption) intended for classes that are likely to be created
>>in very large numbers: you give up the normal flexibility of
>>being able to add arbitrary attributes to an instance, and what
>>you gain in return is that you save one dictionary per instance
>>(dozens and dozens of bytes per instance -- that will, of course,
>>not be all that relevant unless the number of instances that are
>>around is at the very least in the tens of thousands).
> 
> More than that, because of the way modern computer architectures do
> caching, using __slots__ tends to speed up instance attribute access.
> This also is an optimization, but it's effective with even hundreds of
> instances, if you're rapidly switching between instances.

Yes, some TINY such effect can be observed:

[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s' def 
__init__(self): self.x=23' -s'xs=[X() for i in range(999)]' 'z=[x.x for x 
in xs]'
1000 loops, best of 3: 769 usec per loop
[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s' def 
__init__(self): self.x=23' -s'xs=[X() for i in range(999)]' 'z=[x.x for x 
in xs]'
1000 loops, best of 3: 778 usec per loop
[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s' 
__slots__="x"' -s' def __init__(self): self.x=23' -s'xs=[X() for i in 
range(999)]' 'z=[x.x for x in xs]'
1000 loops, best of 3: 682 usec per loop
[alex at lancelot python2.3]$ python -O timeit.py -s 'class X(object):' -s' 
__slots__="x"' -s' def __init__(self): self.x=23' -s'xs=[X() for i in 
range(999)]' 'z=[x.x for x in xs]'
1000 loops, best of 3: 684 usec per loop
[alex at lancelot python2.3]$

But such speedups of 10% to 20% or thereabouts are really "lost in the
noise" in just about any practical case, IMHO (differently from the 2 times
to 3 times speedups I've shown for creation of tens or hundreds of
thousands of instances -- that, I think, is roughly the level at which
optimization starts to be worth paying something substantial for -- and
in fact optimization of memory footprint has even higher potential gain,
though it's devilishly hard to explore it given current profiling tools...).


Alex




From jkraska at san.rr.com  Tue Apr 22 00:13:19 2003
From: jkraska at san.rr.com (Courageous)
Date: Tue, 22 Apr 2003 04:13:19 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <g8n5avgooim13l90bv5t9gu6ilg1oar101@4ax.com> <mailman.1050982240.24284.python-list@python.org>
Message-ID: <t6g9av8aeurhfis19opg84ov34rvr69lpk@4ax.com>

>Except that a collection of threads isn't as powerful as continuations.  As
>soon as "the program counter" changes in one of those threads, the
>continuation it's standing in for has been changed.  A true continuation can
>be invoked any number of times, starting at exactly the same state each
>time.  You'd have to augment those threads with some platform-specific
>grunge to clone the thread control state top-to-bottom to repair that
>weakness.

Actually, you're right. The metaphor sort of broke down, didn't it? What
you can do with park threads is indeed just a subset of what continuations
can do. I suppose this is troubled by the fact that I've *demonstrated*
that particularly whacky feature of continuations to friends to impress,
but never actually made use of it in a production system. Continuations
allow you to do whacky upside down things like make a function return,
say, TWICE. :)

>That's the rub.  Continuations are simple from a very abstract view, in
>function spaces.  Implementing them efficiently on real hardware with finite
>resources is a challenge.

Eh, YEAH. It's even more of a challenge to create a maintainable
implementation. :)

C//




From llothar at web.de  Sun Apr  6 21:22:15 2003
From: llothar at web.de (Lothar Scholz)
Date: 6 Apr 2003 18:22:15 -0700
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain>
Message-ID: <6ee58e07.0304061722.2472b3d0@posting.google.com>

> "comparable in power."  The reason is that there is a ton of elisp
> code out there, including most of Emacs itself, that could be
> translated to work with PyAlaMode.  My goal isn't to be able to
> convert in a completely automated fashion, but close enough to be
> attractive to both an elisp programmer and a Python programmer.

Sorry, have you ever written a elisp program, something that is larger
then
100 lines. I can't think about a way to provide this in python without
writting a lisp like interpreter. Remember that the scope of elisp
variables for example are not lexical. This is something very hard to
emulate in python.
I would really suggest to look at the structure of keyhandling,
keymaps and buffers. And then forget about lisp and design a good
python editor.



From volvox-ml at sferacarta.com  Tue Apr  8 11:27:06 2003
From: volvox-ml at sferacarta.com (Marco Mariani)
Date: Tue, 8 Apr 2003 17:27:06 +0200
Subject: How to protect python code ?
In-Reply-To: <u8bka.651$001.591065@monger.newsread.com>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com>
Message-ID: <20030408152706.GA17737@sferacarta.com>

On Mon, Apr 07, 2003 at 09:02:18AM +0000, msw wrote:

> I'm really interested in this as well, and I think that it's an 
> important issue for python.  Unfortunately, as I'm sure you've 
> discovered by now, it's a question that comp.lang.python is 
> constitionally incapable of answering.  The only answers you will get in 
> this group are:
> 
> 1. intellectual property law is sufficient
> 2. no protection scheme is perfect

3. write in perl and sell the source





From Kyler at news.Lairds.org  Fri Apr 11 15:23:18 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Fri, 11 Apr 2003 19:23:18 GMT
Subject: urlretrieve() or urlopen() ??
References: <mailman.1050064926.20624.python-list@python.org>
Message-ID: <chfjm-khk.ln1@news.lairds.org>

MCollins at co.seminole.fl.us writes:

>still working on the utility to download log files via http.

>my first solution (which worked) was using urlopen(), reading the logfile
>line by line and writing it to a new file line by line.  this felt a little
>slow when working with 200mb files across an http connection.

>so, i switched to urlretrieve(), thinking that would be faster.  however,
>with urlretrieve() i get a "no buffer space available" error.

Did you consider the urlretrieve(url, filename) form?  Am I
missing some processing that you're doing as you receive the
data?

--kyler



From nospam at here  Sat Apr 19 14:51:17 2003
From: nospam at here (Richard Townsend)
Date: Sat, 19 Apr 2003 19:51:17 +0100
Subject: widget like listview
References: <mailman.1050751832.30200.python-list@python.org>
Message-ID: <1050778120.45599.0@damia.uk.clara.net>

> Is there any widget/mega widget available in Tk like ListView in windows
(in
> Visual Studio). Actually i want to show a list of names with their
addresses
> etc from a table and the user is allowed to choose a name from that list
and
> the application will do processing for this name.
>
> I have tried ScrolledList widget of PMW but it only shows the table, i can
> not select an entry from it. Moreover if the name etc in the table are not
> of same length then the format of the list is disturbed ( in windows , on
> linux it works fine ).
>

If what you require is a multi-column list widget, then try
MCScrolledListBox widget from the PmwContribD package.

See http://pmwcontribd.sourceforge.net/
and http://sourceforge.net/projects/pmwcontribd/


You need to have Pmw installed to be able to use it.

There is a bug when used with Pmw 1.1 due to a name clash. A simple
workaround is to replace the 2 occurrences of 'vertscrollbar' in
MCScrollListBox.py with a different name e.g. 'mcvertscrollbar'.










From aleax at aleax.it  Fri Apr 25 19:44:53 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 23:44:53 GMT
Subject: Permanent Clock
References: <d613c14.0304251200.33f1b741@posting.google.com>
Message-ID: <VLjqa.8428$K35.254939@news2.tin.it>

<posted & mailed>

Mauricio wrote:

> Is there some way of accessing a permanent clock, i.e., a clock that
> works even if the system is rebooted or poewered off?
> I would like to increment a variable each day, independent of the
> system date (once this one users can change!), but not leaving my
> script running all the time. Better explaining: every time tha script
> is run, it must now if the current day is the same of tha last time
> the script was started.If not, increment tha variable!

If your script can access the internet, it can get an accurate
date from several servers there.  (Of course, if your users are
REALLY out to screw you, they could find out which IP address
you're trying to get time from, intercept these requests, and
respond with a lie... but then, you could counter that, to some
extent, by putting up your own time server and using a challenge-
response protocol to try and validate it's the real thing... but
then, users could decompile your program, and...;-).


Alex




From donn at drizzle.com  Sat Apr 19 12:28:48 2003
From: donn at drizzle.com (Donn Cave)
Date: Sat, 19 Apr 2003 16:28:48 -0000
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com>
Message-ID: <1050769725.516604@yasure>

Quoth Courageous <jkraska at san.rr.com>:
[apparently quoting Christian Tismer ...]
|> No, he was not. He believes continuations are an insane feature,
|
| Continuations are a _Cool_ feature, Christian. But try not to worry
| about the rejection. If you pause and think for a while, really
| think, they actually don't fit well with Python's mandate of being
| a "simple, approachable language for everyone."

It's an interesting question, I'm on both sides.  I fiddled around
a little with continuations and came up with an interesting way to
fold an event driven, message based system up, with continuations,
to make it procedural on a higher level.  I thought that was kind
of exciting, as computer programming exercises go, but of course it
would seriously complicate any effort to reason about the code.

In a way that's a good thing about GvR's rejection of Stackless,
it's a barrier that still allows serious applications but prevents
the language we all use from being infested with these features
whether we need them or not.  I suspect once it stabilizes, the
Stackless interpreter will do very well on its own.

	Donn Cave, donn at drizzle.com



From bkc at Murkworks.com  Wed Apr  2 13:21:55 2003
From: bkc at Murkworks.com (Brad Clements)
Date: Wed, 2 Apr 2003 13:21:55 -0500
Subject: slurping yahoogroups postings with Python, like yahoo2mbox?
References: <8351bb33.0304020617.61007d5f@posting.google.com>
Message-ID: <3e8b2af6$1_1@spamkiller.newsgroups.com>

I would like to get this too. Something that de-evolves the html marked-up
back to an rfc822 message work-alike. I don't care about reconstructing the
obfuscated email addresses.


--
Novell DeveloperNet Sysop #5

_
"Jon Nicoll" <jkn at nicorp.f9.co.uk> wrote in message
news:8351bb33.0304020617.61007d5f at posting.google.com...
> hi there
>     Does anyone know of a python equivalent (or near-equivalent, that
> I can fiddle with) of the Perl module yahoo2mbox? This 'slurps'
> Yahoogroups postings into an mbox format.
>
> Thanks for any pointers
>
>     Jon N





From Kyler at news.Lairds.org  Wed Apr 30 17:22:14 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Wed, 30 Apr 2003 21:22:14 GMT
Subject: Python / Zope Flash Remoting
References: <MPG.1919beb84d09f671989680@netnews.attbi.com>
Message-ID: <hso5o-m7i.ln1@news.lairds.org>

Joel Jensen <catharsis at nospam.com> writes:

>I am trying to convert some perl/php code to python for use in zope.
>I have never done anything like this. 

Ug.  Conversion is likely to be ugly.  If you don't start with
a Zope mindset, Zope seems horribly limiting.

>Flash Remoting is a way to share datastructures between Cold Fusion and 
>flash,make a dataset, varible, or array in cold fusion/java, retreve it 
>as a map,variable,array in flash via flashs' OO language actionscript.

Sounds like what XML-RPC does for the rest of the world.
Zope supports it nicely.

>have hacked the protocol and written scripts to send datastructures from 
>perl and php.

I hate the thought of having to crack software I want to use
and depend on.

Still, if you're set on using proprietary software, you can
apparently weld on a standard interface.
	http://sourceforge.net/projects/xmlrpcflash/
I haven't tried it.

	http://www.macromedia.com/devnet/server_archive/articles/macromedia_and_xml2.html

--kyler



From theller at python.net  Fri Apr 25 09:16:37 2003
From: theller at python.net (Thomas Heller)
Date: 25 Apr 2003 15:16:37 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <Kq9qa.9083$h%2.957726@read1.cgocable.net> <mailman.1051273778.5621.python-list@python.org>
Message-ID: <fzo6emyy.fsf@python.net>

Gerhard Haering <gh at ghaering.de> writes:

> Mark Nenadov wrote:
> > Alessio Pace wrote:
> >
> 
> >>Hi, I'd like to make a GUI for a program written in python, meant to run
> >>on mac, win and linux: which is the preferable solution? I saw around
> >>tkinter, wxpython and pyqt....
> >
> > I'd strongly recommend wxPython. wxPython is quite a rich GUI
> > library and I have found it to be more advanced/modern/feature-rich
> > than tkinter.
> 
> 
> You'll get a Python interface with the error behaviour of the wrapped
> C++ library. The default is to crash. In my limited experience, anyway.

Not for me, but I can only speak for Windows.

If you go the wxPython route, you should consider pythoncard, which is a
layer on top of wxPython.

Thomas



From rmunn at pobox.com  Wed Apr  2 14:24:30 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 19:24:30 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org> <3e8afeb9.1125578@news-server> <3E8B0060.D9C449A3@engcorp.com> <3e8b1b0f.8380222@news-server>
Message-ID: <slrnb8lp7n.d7q.rmunn@localhost.localdomain>

BBFoto <bbfoto2002 at netscape.net> wrote:
> I apologize for top posting.  I clicked on "Followup Usenet Message".
> My mistake was unintentional.  I will try to find out more about it as
> was suggested.  And as to why I tried to execute a .pyc file:  If I
> knew what I was doing, then I would do it right.  I do appreciate all
> attempts to help.
> 
> I did a search on the PySol directory.  Nowhere is there a file called
> "PySol.py".  There is a series of files in sequence named
> "PySol_15.pyc"  "PySol_16.pyc"  "PySol_20..pyc"  "PySol_21.pyc" and
> "PySol_22.pyc".
> 
> 
> I tried to execute this line,
>  
> execfile('c:/program files/pysol/pysol-4.80/data/PySol_22.pyc') 
> 
> The path is correct.  After it failed I copied this error message from
> the Dos screen:
> 
>   File "c:/program files/pysol/pysol-4.80/data/PySol_22.pyc", line 1
>     -_
>      ^
> SyntaxError: invalid syntax

What the execfile() function does is read a Python source file and run
it. You're getting a SyntaxError because .pyc files are not Python
source files. Python source files end in .py; when the Python
interpreter first runs a .py file, it will byte-compile the file, then
save the .pyc for later reference. That's what those .pyc files are:
byte-compiled files created with various versions of Python.

What icon do you show for those .pyc files? What happens when you
double-click on PySol_22.pyc?

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From mark at freelance-developer.com  Sat Apr 26 05:15:38 2003
From: mark at freelance-developer.com (Mark Nenadov)
Date: Sat, 26 Apr 2003 09:15:38 +0000
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com>
Message-ID: <7xvqa.4236$b67.1131392@read2.cgocable.net>

Anna wrote:

> My daughter just asked me about this "python stuph". So I pulled up
> livewires.org.uk and showed her some basics.
> 
> She just coded her first while loop and for loop... she thought it was
> kewl! In fact, we had to kick her off the computer and send her to bed cuz
> she's got so much to do tomorrow...
> 
> FWIW, she's 12...

Wow! That is excellent!

I introduced by nephew to the Python language when he was about 8 or so. We 
never get much further than the 'print' statement, though :>

-- 
~Mark



From eamonreyn at clear.net.nz  Sun Apr 27 05:41:58 2003
From: eamonreyn at clear.net.nz (Eamon Reyn)
Date: Sun, 27 Apr 2003 21:41:58 +1200
Subject: newbie question about forms in tkinter
Message-ID: <3eaba626@clear.net.nz>

Hi All,

Usually I can find the data I want doing a google group search (kudos to you
guys for having such good answers to most peoples questions) but in this
case I have run into a bit of a wall.

I am creating a little app which needs to have multiple forms pop up but
whenever I try to create a second form in my application the content I want
to appear on the new form just gets placed into my existing one. Here is som
sample code which looks vaguely sane to me but can someone please tell me
what is wrong with it.

from Tkinter import *

class Interface2(Frame):
    def __init__(self, master = None):
        Frame.__init__(self, master)

        # Put the objects on the window
        self.createObjects()
        self.pack()

    def createObjects(self):
        self.lbltest=Label(self)
        self.lbltest.grid(row=0, col=0)
        self.lbltest["text"]="On an new form"


class Interface(Frame):
    def __init__(self, master = None):
        Frame.__init__(self, master)

        # Put the objects on the window
        self.createObjects()
        self.pack()

    def createObjects(self):
        self.b=Button(self, text="create form")
        self.b.grid(row=0, col=0)
        self.b["command"]=self.newForm

    def newForm(self):
        myInter2 = Interface2()
        myInter2.mainloop();

if __name__ == "__main__":
    myInter = Interface()
    myInter.master.title("Test app")
    myInter.mainloop()

Thanks in advance
Eamon.






From peter at engcorp.com  Fri Apr 25 09:17:30 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 09:17:30 -0400
Subject: OT: losing memories (was Re: How many of you are Extreme 
 Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <oafgavg8e0pf512712g4aenldqktvmb0dr@4ax.com> <I78qa.30798$y3.2515733@news010.worldonline.dk> <3EA91334.A7A1D8F2@alcyone.com>
Message-ID: <3EA9356A.DB117F94@engcorp.com>

Erik Max Francis wrote:
> 
> In all seriousness, there are cases of severe brain injury (volume wise)
> which result in surprisingly little actual impairment of abilities.
> There's obviously the Phineas Gage case (who, after having been involved
> in a railroad construction accident had a very large piece of metal fire
> through his brain, destroying most of the left frontal part of his
> brain) retained most of his mental acuities but had some emotional
> behaviorial changes).  

Okay, let me get this straight.  The very large piece of metal
fired through his brain *after* he was involved in an accident?

He must consider himself lucky that the accident itself didn't
cause any damage!

;-)

-Peter



From usenet2002 at krap.dk  Fri Apr  4 12:31:23 2003
From: usenet2002 at krap.dk (Svenne Krap)
Date: Fri, 04 Apr 2003 19:31:23 +0200
Subject: Python and web programming .. my way ?
Message-ID: <odfr8v8uuaf9vcgdom0vfmue7p5h4dv6or@4ax.com>

Hi.

First of all, despite being a complete Python-newbie, I consider
myself a fairly competent PHP-programmer (amongst other things) with
more than 5 years of experience. 

My question is, how does you handle page generation nicely in python ?
In PHP i have adopted a way of generating a gui.php file, which holds
a gui-object, that takes care of drawing the "framework" php-page
(that is top, menues and bottom with colors, semi-dynamic content
(like username for the user) and so on....) 

Meaning that, a simple website, using my system could be : 

=== example.php ===
<?php
include "basics.inc";  
// includes gui.inc amongst other and gives a $gui object
$gui->top("frontpage.aboutme.hobbies") ; 
// the in the above memberfunction the argument 
// is usually a logical path used for menues 
?>
Hello wolrd.
<?php
$gui->bottom();
?>
=== example.php ===

The above code could produce litterally hundreds of lines of html
(with database-powered dhtml-menues and so on...)...

While haveing some drawbacks (ie. its not easy to make major
design-changes, it proves to be a very effective way to write a whole
website, and the resulting pages are very homogene and the sources are
very simple and unbelievable easy to maintain afterwards (even for
people who had prevously never seen the aproach before). It also has
the benefit that every logical webpage has a physical php-file and
that deeplinking and such stuff is tremendously easy.

The mod_python way with a single .py file to handle the whole
directory seems ackward and clumsey to me. By the way.. is there a way
to make the PythonHandler to handle the whole website (ie. from the
document-root and downwards) and if so, is it easy to get the
requested url ? Then I think, i can figure out a likewise system
myself.

Is there any way to do something likewise in Python ? Or is there any
good reason why my way (tm) is not easily doable in python ? And how
should it be made instead....

Svenne



From cartermark46 at ukmail.com  Tue Apr 15 07:44:25 2003
From: cartermark46 at ukmail.com (Mark Carter)
Date: 15 Apr 2003 04:44:25 -0700
Subject: newbie: raise socket.error, msg in smtplib
Message-ID: <d3c9c04.0304150344.cf0f33c@posting.google.com>

I'm getting

Traceback (most recent call last):
  File "C:\My Documents\myemail.py", line 26, in ?
    s.connect()
  File "C:\PYTHON22\lib\smtplib.py", line 283, in connect
    raise socket.error, msg
error: (10061, 'Connection refused')


from running a slightly modified version of the email example in the docs using
Python 2.2.1 (#34, Apr  9 2002, 19:34:33) [MSC 32 bit (Intel)] on win32

What are the likely causes of this problem?





source code begin:

# Import smtplib for the actual sending function
import smtplib

# Here are the email pacakge modules we'll need
from email import Encoders
from email.MIMEText import MIMEText
# Open a plain text file for reading
textfile = 'msg.txt'
fp = open(textfile)
# Create a text/plain message, using Quoted-Printable encoding for non-ASCII
# characters.
msg = MIMEText(fp.read(), _encoder=Encoders.encode_quopri)
fp.close()

# me == the sender's email address
# you == the recipient's email address
msg['Subject'] = 'The contents of %s' % textfile
me = "cartermark46 at ukmail.com"
you = me
msg['From'] = me
msg['To'] = you

# Send the message via our own SMTP server.  Use msg.as_string() with
# unixfrom=0 so as not to confuse SMTP.
s = smtplib.SMTP()
s.connect()  ######## LINE 26
s.sendmail(me, [you], msg.as_string(0))
s.close()



From aleax at aleax.it  Sun Apr 13 17:57:56 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 21:57:56 GMT
Subject: Question regarding indirect function calls
References: <pan.2003.04.13.20.13.21.504295@maxwell.com>
Message-ID: <E3lma.9999$LB6.267395@news1.tin.it>

Maxwell Hammer wrote:

> I am just starting to learn python but I love it already..
> Ok the question..
> In the program I'm trying to write I have a module called "plugins"
> containing a number of functions. I only know at runtime the name of a
> function to call from "plugins", i.e. need to call a function indirectly.
> The problem is that assigning the name of one of plugin's functions to a
> variable, I cannot then call the function. IOW:
>  func=function1
>  plugins.func() ..........will fail
>  plugins.function1() .......will pass

getattr(plugins, func)()

should do what you desire.


Alex




From teddrain at earthlink.net  Tue Apr 22 12:10:42 2003
From: teddrain at earthlink.net (Ted Drain)
Date: 22 Apr 2003 09:10:42 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>
Message-ID: <f6a31cb8.0304220810.547f5acd@posting.google.com>

My $.02:

We use Python as a scripting interface to an underlying C++ system. 
The users write Python code to do what they want.  We ended up writing
__repr__ and __str__ in our classes to be the same thing because our
users were very annoyed when they typed:

>>> s
>>> print s

and got different results.

This doesn't necessarily mean that I'm not in favor of having the
distinction.  I think I'd rather see the behavior change when you do:
>>> s
To call __str__ instead of __repr__.  It seems to me that repr is a
specialized functionality that should be explicitly asked for when
needed.  __str__ is the more normal output and should be the default
in 'print s' or 's'.

Ted   teddrain at earthlink.net



From gh at ghaering.de  Wed Apr 30 03:26:58 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 30 Apr 2003 09:26:58 +0200
Subject: running a PyUnit TestSuite
In-Reply-To: <b8nrlv$dvn$07$1@news.t-online.com>
References: <b8nrlv$dvn$07$1@news.t-online.com>
Message-ID: <3EAF7AC2.7040503@ghaering.de>

Achim Domma wrote:
> Hi,
> 
> I have a TestSuite which depends on the execution order of the tests. If I
> use the 'main' object provided by unittest.py I can not specify the
> execution order, so I set up a TestSuite by hand. But how should I run it? I
> found no easy way to pass a TestSuite for unittest.main. Any hint?

I found it rewarding to split unit tests into modules, each one having a 
suite() function. This function returns a test suite, using the fact 
that you can pass a sequence TestSuites to the TestSuite constructor:

[pysqlite/test/all_tests.py]:
#v+
import api_tests, lowlevel_tests, pgresultset_tests, type_tests, 
userfunction_tests, transaction_tests

# [...]

def suite():
     suite = unittest.TestSuite((lowlevel_tests.suite(), api_tests.suite(),
             type_tests.suite(), userfunction_tests.suite(),
             transaction_tests.suite(), pgresultset_tests.suite()))

     return suite

def main():
     warnings.filterwarnings("ignore", "DB-API extension")
     runner = unittest.TextTestRunner()
     runner.run(suite())
#v-

HTH,

-- Gerhard




From peter at engcorp.com  Tue Apr  8 20:29:22 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 08 Apr 2003 20:29:22 -0400
Subject: usage of assert and -O
References: <mailman.1049844497.9845.python-list@python.org>
Message-ID: <3E936962.27F9C388@engcorp.com>

> Matthew Russell wrote:
> 
> I understand that when you invoke the interpreter with the -O flag assert 
> statments are not executed - is this 100% correct? (as i belive it to be)
> and if so, shouldn't the use of assert (if not assert itself) be depricated?

I guess you could deprecate them, but why bother?  Some people seem to
find them useful.

> I note that in unittest.py assert is not use,  we have failIf or assert_ 
> instead - thus if you avoided using assert in your own tests you can 
> saftly run all your unittests optimised....

That surely would be the idea.  Besides, assert is part of the language,
while failIf/assert_ and so on are part of the unit test framework.  There
is no particular relationship between the two.

> ..or can you...?

Yes.

> As some code in the standard library use assert,
> and since most python users' applications use library modules.....

No problem.  Asserts should not change the functionality of the 
code (except to abort it), so whether present or absent they 
shouldn't occur under normal use.  Code which causes any library
code's asserts to fire is effectively broken, since the whole point
of those asserts is to make sure those routines are never required
to handle really bad data.

> I would be interested to know what others think about usage of 
> assert in context with optimised python code, in particular in 
> releation to unittesting, as this is where i think it would be 
> useful - if not critical to have one way of doing it (i.e not
> useing assert or -O [ one or the other ])

1. I never use -O or -OO anyway, so the issue of asserts being
removed from certain code has never come up for me.

2. I never use asserts either.

3. Unit testing should be done with a unit test framework, such
as doctest or unittest, not with asserts.  Asserts are for 
run-time protection, basically, while unit tests are run during
development.  I don't see the connection between the two.

4. Code which is dependent on the presence or absence of 
optimization done by the compiler is broken, but this should
be something checked by unit tests perhaps, not by asserts
(even if they worked when using -O).

In summary: don't bother with asserts if you're trying to do
unit testing.

-Peter



From DennisR at dair.com  Thu Apr 24 20:36:57 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Fri, 25 Apr 2003 00:36:57 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com> <mailman.1050978337.8048.python-list@python.org> <87wuhm2rsu.fsf@pobox.com> <Bkipa.1154$Hm6.104194064@newssvr14.news.prodigy.com> <87adefkbs9.fsf@pobox.com> <m7Ypa.156$Jr1.14453940@newssvr15.news.prodigy.com>
Message-ID: <Jq%pa.286$iG3.15927365@newssvr14.news.prodigy.com>

> sys.path = [".", ".\\bin", ".\\bin\\libstd", ".\\bin\\libxtra"]
> Maintaining this "manual hack" is far less effort than you would guess.

There is something I forgot to mention.  Suppose you start with a decent
size app already?  You do not have to go through and add files to this
structure 1-by-1.

To seed this structure initially, comment out the sys.path line so that you
are using your system python path.  Now, delete *.pyc where the *.py exists.
Run your program and exercise its functions.  Exit.  Now, where *.pyc files
exist in the system python path, you need to copy the .py files to a
subdirectory listed in sys.path.

In one step, you have now populated your subdirectories with the required
libraries.  There may be other files that you will end up needing and
restoring the sys.path statement will quickly find them.
--

Dennis Reinhardt

DennisR at dair.com   http://www.spamai.com?ng_clp





From skywalkerpackage at hotmail.com  Tue Apr 15 06:28:27 2003
From: skywalkerpackage at hotmail.com (Johannes Eble)
Date: Tue, 15 Apr 2003 10:28:27 GMT
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de> <Xns9358723026D8Cduncanrcpcouk@127.0.0.1>
Message-ID: <3e9bdd3d.100422968@news.online.de>

Hi Duncan,

On Wed, 9 Apr 2003 10:20:01 +0000 (UTC), Duncan Booth
<duncan at NOSPAMrcp.co.uk> wrote:

>skywalkerpackage at hotmail.com (Johannes Eble) wrote in
>news:3e93e50f.676072541 at news.online.de: 
>
>> is there a library available for Python that behaves similar to MS
>> Visio? For example, I would like to create vector graphic objects and
>> connect the with a line. If I move, flip or rotate a connected object,
>> the line should follow the connection.
>> 
>> I would like to draw UML diagrams.
>
>Maybe not quite what you asked but you might want to look at
>ObjectDomain which is a UML diagram editor that is fully scriptable in
>Python (or rather Jython).

I would prefer an Open Source product.

>
>> But the program is in c and it is tightly coupled with GTK+.
>> It seems to be an impossible mission to control dia from
>> another program, at least in Windows.
>
>A quick look at the Dia website seems to indicate that DiaPython exists.

Not really. There is a way to use Python plug-ins, but it is
undocumented. It is impossible to write a plug-in without a knowlede
of the c code. Also, I don't like GTK+ because this pseudo object
oriented c style is very ugly IMHO let apart the problems you face
with Windows.


Regards


Johannes




From glc at well.com  Sun Apr  6 12:31:33 2003
From: glc at well.com (Greg Chapman)
Date: Sun, 06 Apr 2003 16:31:33 GMT
Subject: properties and get/set methods
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org>
Message-ID: <5kl09vses40u5v7jqrecgu5co8epaqbidh@4ax.com>

On Sat, 05 Apr 2003 23:48:41 -0500, "Mike C. Fletcher" <mcfletch at rogers.com>
wrote:

>As for how the problem ((a deficiency in the property API IMO) of not 
>knowing the name being set/get within the get/set methods) gets worked 
>around.  I (as seen above) do some comparatively minor (but still 
>annoying) duplication of code by passing an explicit string version of 
>the name to the property sub-class's constructor and then use that name 
>when the property instances are called.  Technically, the string value 
>can be anything, and certain sub-classes of property just ignore it, 
>while in any case, it can be something different than the "public" name 
>through which the property API calls were made :( .
>
>i.e. this works, storing the value as instance.__dict__['_myProperty']:
>
>    myProperty = StringProperty(
>       "_myProperty", """Some documentation""",
>       defaultValue = "this value",
>    )
>
>but the property doesn't know what name it's _really_ being accessed as 
>in this case :( .

You could always use somthing like this:

import sys
import re
import inspect
from compiler.misc import mangle

# seems like this ought to be built in someplace, but I couldn't find it
def isident(name):
    return bool(re.match('[A-Za-z_][A-Za-z_0-9]*$', name))

def DeclareProperty(name, doc, defvalue):
    frame = sys._getframe(1)
    framecode = frame.f_code
    codename = framecode.co_name
    inclassdef = (                 #not a perfect test by any means
        not (framecode.co_flags & inspect.CO_OPTIMIZED) and
        isident(codename)
    )
    if inclassdef and name.startswith('__'):
        name = mangle(name, codename)
    def getter(obj):
        return obj.__dict__.get(name, defvalue)
    def setter(obj, value):
        obj.__dict__[name] = value
    prop = property(getter, setter, doc=doc)
    if inclassdef:    
        frame.f_locals[name] = prop
    return prop                  #in case you want to use this outside
                                 #a class def

class Test(object):
    DeclareProperty('myProperty', 'Some documentation', 'default value')


---
Greg Chapman




From max at alcyone.com  Tue Apr  1 17:31:25 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 01 Apr 2003 14:31:25 -0800
Subject: (no subject)
References: <mailman.1049234524.11247.python-list@python.org>
Message-ID: <3E8A133D.27B3286B@alcyone.com>

[proceeding carefully given the date ...]

Santoshi Reddy wrote:

> I am new to python.can Someone tell me what _int_ is?

It's a function/type that corresponds to basic integers.  You can pass
it an object and it will try to convert it to an integer.  If the
argument is a string, an optional second argument can be used to specify
the radix.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ A good indignation brings out all one's powers.
\__/ Ralph Waldo Emerson
    Esperanto reference / http://www.alcyone.com/max/lang/esperanto/
 An Esperanto reference for English speakers.



From staschuk at telusplanet.net  Sun Apr 20 02:52:51 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 20 Apr 2003 00:52:51 -0600
Subject: Sorting without transitivity
In-Reply-To: <20030419213405.GA588@obstler>; from niessink@serc.nl on Sat, Apr 19, 2003 at 11:34:05PM +0200
References: <20030419213405.GA588@obstler>
Message-ID: <20030420005251.D8441@tibia.amotlpaa.bogus>

Quoth Frank Niessink:
> [It's been while since I've dealt with transitivity and stuff, so
> I hope I got the terminology right]
> 
> I have objects that are not transitivily ordered, e.g.:
> 
> A > B
> A == C
> B == C
  [...]

Martelli's already given the magic words "topological sort"; this
post is just about the terminology.  (All from memory; corrections
welcomed.)

Usually == is supposed to be an "equivalence relation", which
means that
    1. x == x (reflexivity)
    2. x == y if and only if y == x (symmetry)
    3. if x == y and y == z then x == z (transitivity)
Above we see a violation of these assumptions.  Since b == c, by
symmetry we expect c == b.  Then, since a == c, by transitivity we
expect a == b; but actually a > b.

So your == above seems not to be an equivalence relation; it seems
to mean "not ordered with respect to" instead of "equal to" or
"equivalent to".  Some authors use <> for "not ordered with
respect to".  (This is not its meaning in Python, where it is a
synonym of !=, "not equal to".)  In this notation, your statements
above would be
    a > b
    a <> c
    b <> c
Formally, x <> y means just that neither x > y nor y > x.

Usually the symbol > denotes a "total order" or a "partial order".
A total order is a relation > such that
    1. not x > x (irreflexivity?)
    2. if x > y then not y > x (antisymmetry)
    3. if x > y and y > z then x > z (transitivity)
    4. for any x and y, either x == y, x > y, or y > x (trichotomy?)
The difference between a total order and a partial order is that
(4) need not hold for the latter; thus we can have unequal
elements which have no determinate order relationship.

So your > looks at first glance like what they call a partial
order.

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From eprice at tjhsst.edu  Mon Apr  7 12:28:29 2003
From: eprice at tjhsst.edu (Eric Price)
Date: 7 Apr 2003 09:28:29 -0700
Subject: Customize rlcompleter
References: <b6risd$giq$1@biggoron.nerim.net>
Message-ID: <80ab95a6.0304070828.4697e4b9@posting.google.com>

Gilles Lenfan wrote:
> I'd like to use "rlcompleter" with a custom vocabulary for a personal
> command line interpreter based on the "cmd" package. But the doc for
> rlcompleter is quite sparse, and I can't see how to tailor it to my own
> usage.

In rlcompleter.py, the Completer class has a method global_matches:

    def global_matches(self, text):
        """documentation"""
        import keyword
        matches = []
        n = len(text)
        for list in [keyword.kwlist,
                     __builtin__.__dict__.keys(),
                     __main__.__dict__.keys()]:
            for word in list:
                if word[:n] == text and word != "__builtins__":
                    matches.append(word)
        return matches

To edit the vocabulary for global matches, change the lists in the 
"for list in ..." loop to include the wanted vocabulary.

This will only change global matching, though.  Global matching is
called when there is no '.' in the string you are completing.  If you 
always want to complete using those lists, edit the complete method:

    def complete(self, text, state):
        """Docstring"""
        if state == 0:
            if "." in text:
                self.matches = self.attr_matches(text)
            else:
                self.matches = self.global_matches(text)
        [...]

and change it to

    def complete(self, text, state):
        """Docstring"""
        if state == 0:
            self.matches = self.global_matches(text)
        [...]

as well as editing global_matches as before.

Then, if tab-completion is enabled, the prompt will tab-complete to
any words in the lists in the global_matches function.

Hope this helps,
Eric



From theller at python.net  Tue Apr  1 10:09:55 2003
From: theller at python.net (Thomas Heller)
Date: 01 Apr 2003 17:09:55 +0200
Subject: Implementing a "classmethod" in an extension type
References: <mailman.1049205604.27996.python-list@python.org>
Message-ID: <smt2ntj0.fsf@python.net>

"Mark Franklin" <mfranklin at tcsi.com> writes:

> Hi,
> 
> I've got an built-in extension type, implemented in C. I want to be able to implement the equivalent of a "classmethod" in that type. 
> 
> I thought I might be able to do something like this, which fairly closely mimics what you do in Python:
> 
> 	PyTypeObject myType = {
> 		// initialised in the usual way. It's tp_methods field includes a method definition for "MyMethod".......
> 	};
> 
> 	PyObject *method = PyObject_GetAttrString(myType, "MyMethod");
> 	// method is a PyMethodDescrObject
> 	PyObject *classMethod = PyClassMethod_New(method);
> 	PyObject_SetAttrString(myType, "MyMethod", classMethod);
> 	Py_DECREF(method);
> 
> 	PyType_Ready(myType);
> 
> 
> However, when I do this I get a TypeError exception from the PyObject_SetAttrString call: "can't set attributes of built-in/extension type 'MyType' ".
> 
> So, how do I get the class method into the type object?
> 
> 
> Thanks,
> Mark

In Python 2.3, there is a METH_CLASS flag. Unfortunately not in 2.2.
Here are the relevant snippets that I use - it is from
http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/ctypes/ctypes/source/_ctypes.c
and it compiles with Python 2.2 as well as 2.3.

HTH,

Thomas


#include "Python.h"

#ifndef METH_CLASS
#define METH_CLASS 0x0010 /* from Python 2.3 */
#define NO_METH_CLASS
#endif

...

/*
 * If this is a real C classmethod (Python 2.3 and later),
 * it has METH_O style. So the type is in the first argument,
 * and the arg in the second.
 *
 * We fake this in Python 2.2, where METH_CLASS is not present,
 * and 'args' is a tuple containing the type as first member.
 */
static PyObject *
CString_from_param(PyObject *cls, PyObject *args)
{
	PyObject *value;

#ifdef NO_METH_CLASS
	PyObject *ignore;

	if (!PyArg_ParseTuple(args, "OO", &ignore, &value))
		return NULL;
#else
	value = args;
#endif
...
}


static PyMethodDef CString_methods[] = {
#ifdef NO_METH_CLASS /* Python 2.2 */
	{ "from_param", CString_from_param, METH_VARARGS | METH_CLASS,
#else
	{ "from_param", CString_from_param, METH_O | METH_CLASS,
#endif
	  from_param_doc },
	{ NULL, NULL },
};


...

	if (PyType_Ready(&CString_Type) < 0)
		return;
#ifdef NO_METH_CLASS
	DoClassMethods(&CString_Type);
#endif
	PyModule_AddObject(m, "c_string", (PyObject *)&CString_Type);




From staschuk at telusplanet.net  Wed Apr 16 17:54:21 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 15:54:21 -0600
Subject: Python and Schools
In-Reply-To: <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>; from greg@brondo.com on Tue, Apr 15, 2003 at 07:07:27PM +0000
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
Message-ID: <20030416155421.G4266@tibia.amotlpaa.bogus>

Quoth Greg Brondo:
  [...]
> I understand why the code is bad (copying the string each time in
> memory as it grows).  However, I've never understood the notation's
> O(N) and O(N2)?

Others have given good explanations of the usual use of big-O
notation, namely, to describe roughly how the running time of an
algorithm relates to the size of its input.  Imho this is actually
a convenient but slightly sloppy way of speaking; strictly
speaking one counts operations performed.  Roy Smith's excellent
post, for example, focusses on memory copy operations,
essentially; that's a good choice for that problem, while counting
string concatenations would not be, if we're interested in
predicting runtime.

And you can use big-O notation for counting things other than
operations.  For example, it is correct usage to describe a
trade-off as being between, say, an algorithm which runs in O(n)
time but needs O(n^2) memory to do it, and an alternative which
runs in O(n^2) time but only needs O(n) memory.

-- 
Steven Taschuk                  staschuk at telusplanet.net
"Telekinesis would be worth patenting."  -- James Gleick




From andrew-pythonlist at puzzling.org  Wed Apr 16 09:22:54 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 16 Apr 2003 23:22:54 +1000
Subject: FTP
In-Reply-To: <20030416021556.GA8693@meson.dyndns.org>
References: <e62b26d3.0304151727.6de2e2e4@posting.google.com> <20030416021556.GA8693@meson.dyndns.org>
Message-ID: <20030416132254.GA27886@frobozz.local>

On Tue, Apr 15, 2003 at 10:15:56PM -0400, Jp Calderone wrote:
> On Tue, Apr 15, 2003 at 06:27:11PM -0700, Matt Singer wrote:
> > I have an application with a PYQT interface that uploads large files. 
> > Is there an async method to get how much data has been sent with
> > storlines to update a progress bar?  Or is there an alternate ftp
> > interface?
> 
>   Twisted might help.
> 
>     http://twistedmatrix.com/documents/TwistedDocs/current/api/public/twisted.protocols.ftp.html
> 
>   and doc/examples/ftpclient.py in the source tarball.

Note that upload support has only been added recently, so you'll need to get
1.0.4rc1.

-Andrew.





From salvatore at wanadoo.fr  Sat Apr  5 04:18:20 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Sat, 05 Apr 2003 11:18:20 +0200
Subject: In memory File
Message-ID: <b6m741$dqo$1@news-reader12.wanadoo.fr>

Hello,

How can I create file-like objects ?
I want to work in memory only.

Thanks for your help

Salvatore




From robin at execulink.com  Sat Apr  5 15:33:03 2003
From: robin at execulink.com (robin)
Date: Sat, 05 Apr 2003 15:33:03 -0500
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <b6f66v$28k$1@slb3.atl.mindspring.net> <v8m9a4nk4dpjf7@corp.supernews.com> <a626bbd7.0304031149.7abe84c0@posting.google.com> <v8toehsbm8bob5@corp.supernews.com>
Message-ID: <c8fu8v0q3k04prtgamlqt8nnsjg2jo93sr@4ax.com>

claird at lairds.com (Cameron Laird) wrote:

>I'll do this in an abbreviated form.

Thank you. That was very informative. At least it confirms my
suspicions.

I suppose now I'll be looking closer at Dopy, Pyro, and Twisted
Spread, so anything that will help me distinguish between them is
welcome.

-- robin



From jack at performancedrivers.com  Mon Apr 28 13:02:49 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Mon, 28 Apr 2003 13:02:49 -0400
Subject: metaclasses, how you might actually use 'em
In-Reply-To: <2259b0e2.0304280514.5ca68d50@posting.google.com>; from mis6@pitt.edu on Mon, Apr 28, 2003 at 06:14:37AM -0700
References: <mailman.1051458882.30400.python-list@python.org> <2259b0e2.0304280514.5ca68d50@posting.google.com>
Message-ID: <20030428130248.N15881@localhost.localdomain>

On Mon, Apr 28, 2003 at 06:14:37AM -0700, Michele Simionato wrote:
> Jack Diederich <jack at performancedrivers.com> wrote in message news:<mailman.1051458882.30400.python-list at python.org>...
> 
> > <snip evangelism to metaclasses>
> 
> All your usages of metaclasses seems to me quite typical and "good"
> (to see a "bad" usage, google for my "black magic trick" ;)
> 
> Do you know that the standard 'type'
> metaclass (and therefore all metaclasses derived from it, which means
> ALL metaclasses) contains a __subclasses__ method ?
>
> I think it is quite similar in spirit to your Register, isn't it ?

Very similar for class trees with one common parent.
The Register pattern stores the info in an arbitrary location so different
class trees that have a Register metaclass can aggregate their info in a common
place.

Also, the Register pattern builds the list once when classes are defined.
Using the list afterwards is cheap.  Instead of using the list generated
by Register you could recurse down __subclassess__() each time you wanted
to iterate on the list or check for membership, but it would be more expensive.

instead of
 if (ob.__class__ in Factory.registered_classes):
  # do stuff

 if (ob.__class__ in func_to_gather_subclasses(Base)):
  # do stuff

So the Factory has to know about the Base, instead of the Base's register
method knowing about the Facotry.  The special DO_NOT_REGISTER attribute
for intermediate virtual classes also works well in the Register pattern
because we can inspect it then delete it so it isn't inherited by subclasses
and won't outlive its brief purpose (As a flag saying dont register me).

> 
> > metaclasses just allow you to tweak a class definition on the fly.  Mainly 
> > adding or removing attributes just before the class becomes 'real'. 
> 
> is correct when you redefine the __new__ method of the metaclass. However,
> in your examples, you redefine the __init__ method. This means that the
> class has been *already* created, i.e. it is 'real', and it is modified
> *after* creation. This is just nitpicking, but since you asked for pedantry...

I could never get the __new__ to do what I wanted.  I would always end up
recursively calling __new__.  Doing everything in __init__ works, so I just
kept with that.

-jackdied
http://jackdied.com/code/




From rvprasad at cis.ksu.edu  Tue Apr 15 23:39:59 2003
From: rvprasad at cis.ksu.edu (Venkatesh Prasad Ranganath)
Date: Wed, 16 Apr 2003 03:39:59 GMT
Subject: Protected Methods and Python
In-Reply-To: <mailman.1050275130.5456.python-list@python.org>
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <18lb3o2cs1se2$.isoixa5igmj6$.dlg@40tude.net> <12aeh974i1kdh.dfbwdn5e18px.dlg@40tude.net> <mailman.1050275130.5456.python-list@python.org>
Message-ID: <jg4na.146561$S14.33908@news1.central.cox.net>

Jp Calderone wrote:
> On Sun, Apr 13, 2003 at 09:26:37PM +0200, Mads Orbesen Troest wrote:
> 
>>On Sun, 13 Apr 2003 21:24:24 +0200, Mads Orbesen Troest wrote:
>>
>>
>>>>Or may be it is possible to emulated Protected Methods in Python.
>>>
>>>It can be "emulated" by prepending the method/attribute name with 2 
>>>underscores. 
>>
>>Whoops, I was too quick and didn't see that you wrote "protected" and not 
>>"private".
>>
>>At any rate: the mangling scheme can be used for "protected" methods, since 
>>you know the mangled name is _classname__attribute/method. You can then be 
>>one of the "evil" persons I talked about in the previous post, and access 
>>the "private" member that way.
>>
> 
> 
>   Double underscore prefix is for avoiding name collisions in inheritance,
> not preventing people from touching things you consider private to the
> implementation.
> 
>   Language enforced private attributes are a Wrong idea because they
> needlessly increase the difficulty of people trying to use code which
> employs this feature.
> 
>   There are a few possible scenarios:
> 
>   1) I am a skillful coder and I am using a library written by a skillful
> coder.  The library is bug-free (for the purpose which I use it) and I have
> no need to touch the implementation's internal variables.  In this scenario,
> runtime protected private attributes are not useful.
> 
>   2) I am not a skillful coder and I am using a library written by a
> skullful coder.  The library is bug-free (for the purpose which I use it)
> but I feel the need to touch the implementation's internal variables anyway. 
> Whether or not they are protected by the runtime environment, I will
> probably find a way to screw up my application, because I am not skilled. 
> In this scenario, runtime protected private attributes are not useful.
> 
>   3) I am a skillful coder and I am using a library written by someone who
> is not skillful.  The library contains bugs which disrupt my application. 
> In this scenario, runtime protected private attributes make it impossible
> for me to work around bugs in the library, and so are worse than useless.
> 

All said, how do you ensure "design by interface" in Python?   If every feature of a class is open to the user then he/she 
will use them as he/she sees fit.  However, the writer of the class could have coded some methods for ancillary purposes and 
later decides to do away with them in the next version of the class.  How should the user deal with this situation?  In my 
opinion, protected and private as in C++ or Java empowers the library developer to expose just the interface that he/she 
intends the user to see.  Nothing more and nothing less.  (It's an issue dealing with abstraction.)  As a result, the user can 
be more care-free while using the library code as he cannot screw up the internals if the library code doesn't do it on it's 
own.  On the other hand, if all members of a class are visible to the user then the user needs to be more diligent while using 
the library.  I prefer the first option.

On the issue of working around bugs in a library via runtime protected attributes, it would be better if the skillful coder 
fixed the bug ;-) or use another library.  This is just what I would do.

Well, that's my opinion.

>   Since Python provides no real runtime protection of private attributes,
> the best you can hope to do is make it excessively painful for people using
> your code to work around your bugs.  I can't imagine why anyone would want
> to do this.
> 
>   If an attribute is private to your implementation, document it this way
> and prefix it with a single underscore.  Nothing else is necessary.
> 
>   high-and-mightily y'rs,
> 
>   Jp
> 


-- 

Venkatesh Prasad Ranganath,
Dept. Computing and Information Science,
Kansas State University, US.
web: http://www.cis.ksu.edu/~rvprasad




From mwm at mired.org  Thu Apr  3 15:04:27 2003
From: mwm at mired.org (Mike Meyer)
Date: Thu, 03 Apr 2003 20:04:27 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org> <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk>
Message-ID: <x7el4jibzl.fsf@guru.mired.org>

"Anders J. Munch" <andersjm at dancontrol.dk> writes:
> "Mike Meyer" <mwm at mired.org> wrote:
> > I disagree with that statement. Clearly, *someone* knows all the
> > coding tricks used by the compiler - because they put it into the
> > compiler. So they can achieve the same level of performance. 
> And chess program authors can beat their own inventions?  Not so.

Go program authors can beat their own inventions, so clearly this
depends on the problem domain. In the chess (and go) problem domains,
the time the human spends on the problem is artificially limited by
the rules of the game. For programming, limiting the programmer time
is a goal, not an axiom.

As an existence proof, unless you believe that the compiler produces
perfect code, dis-assembling (if needed) and hand-tuing the code
produced by the compiler should get you faster results than you
started with.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From aleax at aleax.it  Sat Apr 12 14:45:26 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 18:45:26 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
Message-ID: <a9Zla.7472$T34.203357@news2.tin.it>

Mads Orbesen Troest wrote:


> Given that strings are immutable, is it exceedingly slow (and memory
> spiking) to do a lot of "string += string" operations in one's code? IOW,

Yes -- it's THE one major performance trap in naively coded Python,
since, when "a lot" is a sufficiently large N, a lot of += on short
strings ends up being O(N squared) where O(N) approaches would be
available.  And turning something potentially O(N) into something
O(N squared) is the kind of performance hit that profiling won't
necessarily catch, because it shows up (as unboundedly bad) only
for N sufficiently large...

So, it's basically the one performance trap I find it most necessary
to teach even to total beginners!


> does each and every += evaluation lead to a new concatenated copy of the
> previous string (now freed for garbage collection) and the new string, so

Exactly (it's not necessarily the case that the previous string is freed
for garbage collection, but whether it is, or not, a new concatenated copy
is indeed unavoidable).


> that the longer the string to which stuff is appended is, the longer times
> each += operation takes?

Again, exactly -- s += l takes roughly O(len(s)+len(l)) time (as it
necessarily has to copy the len(s) previous chars followed by the len(l)
new ones).

Building up lists of strings with mylist.append(piece), then joining
them together at the end with ''.join(mylist), is the canonical solution;
others, which I've seen already pointed out to you, include using cStringIO.


Alex




From reply.in at the.newsgroup  Thu Apr 10 18:40:55 2003
From: reply.in at the.newsgroup (Rene Pijlman)
Date: Fri, 11 Apr 2003 00:40:55 +0200
Subject: Can I define "static" (like in C++) methods in Python?
References: <169a7283.0304101430.2bec922b@posting.google.com>
Message-ID: <fnsb9vcog73uj94f3ncq2tuqcf07v8s2ku@4ax.com>

sdieselil:
>Can I define methods which are attributes of class and not attributes of
>instance?

http://www.python.org/doc/FAQ.html#4.84

-- 
Ren? Pijlman



From nomad*** at ***absamail.co.za  Mon Apr 21 16:07:50 2003
From: nomad*** at ***absamail.co.za (Nomad)
Date: Mon, 21 Apr 2003 22:07:50 +0200
Subject: Mail and text extraction
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com> <mailman.1050882393.32651.python-list@python.org>
Message-ID: <8gj8av8e9uq2blldocm28s11t8qdveu9cr@4ax.com>

On Mon, 21 Apr 2003 01:45:55 +0200, Gerhard H?ring <gh at ghaering.de>
wrote:

>* Dino Levy <dino_levy at mindlessSPAM.com> [2003-04-21 01:19 +0200]:
>> I have the following problem:
>> I'm trying to make a script which will connect to POP3 server, print
>> number of mails, print their size and subject, and then will ask the user
>> which mail should be downloaded...
>> So far so good, but I have problem: 
>> [...]
>> The
>> mail looks like this:
>> Agnell company
>> 
>> -Worker		Payment
>> Joe Doe		500$
>> Mary Smith	550$
>> Karen Bay	495$
>> 
>> -New Employee: Peter Maley
>> 
>> some text...
>> 
>> -Raises:
>> some text
>> 
>> And all you want to do is to connect to your POP3 server, and download
>> the section called Raises ( for example, two or three lines after the
>> title ), so all you want to do is to download those few lines containing your
>> text... ( ONLY those lines, not the whole e-mail ) [...]
>
>AFAIK, the POP3 protocol doesn't support downloading parts of emails. To
>my knowledge, you can either download the headers (via TOP) or the body
>(via RETR).

TOP takes two arguments, the message number (from LIST) and the number
of lines of the body to retrieve (eg TOP 1 1 gives you the headers and
one line of message).

RETR returns the whole rfc*822 message, both headers and body.

Just my 2p.

-- 
Nomad

Wondering of the vast emptyness of the 'net
in search of something cool.



From s.thuriez at laposte.net  Wed Apr  9 10:01:09 2003
From: s.thuriez at laposte.net (sebastien)
Date: 9 Apr 2003 07:01:09 -0700
Subject: PCA principal component analysis
Message-ID: <b1ab591.0304090601.302cc4b6@posting.google.com>

Hi,

Is there any PCA analysis tools for python ?
If it does, do you have any idea on how well it would scale ?

I have already seen PyClimate (but it is not available for Windows
which will be one of the target). Is there some LAPACK like packages ?

Thanks for reading this mail.

Sebastien.



From ark at research.att.com  Tue Apr 15 18:03:16 2003
From: ark at research.att.com (Andrew Koenig)
Date: Tue, 15 Apr 2003 22:03:16 GMT
Subject: Big-O notation (was: Re: Python and Schools)
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org>
Message-ID: <yu99y92bh10b.fsf@europa.research.att.com>

Mike>     Characterisation of an algorithm's (as distinct from an
Mike>     implementation's) performance in relation to the size of the input
Mike>     set.  Normally of interest in determining the
Mike>     scalability/suitability of a particular algorithm.  The static (e.g.
Mike>     setup) and proportional (e.g. implementation) factors in the
Mike>     algorithm are factored out under the assumption that they are
Mike>     swamped by the relative effects of the size of the input set as that
Mike>     goes to infinity.

Also, big-O notation always refers to an upper bound, not an estimate
that might be high or low.  So, for example, every quantity that is
O(N) is also O(N log N), but the reverse is not true.

Mike> Common examples of big-O performance types:

Mike>     * O(1) -- time is linear, that is, it takes the same amount of time
Mike>       to do something regardless of how large the input set is.

O(1) doesn't mean linear, it means bounded by a constant.  That is,
if the running time of an algorithm is O(1), it means that there exists
a time interval t such that the algorithm always finishes in time < t.

Mike>     * O(log N) -- time grows as the log of N, (for those who don't
Mike>       remember high-school math, this means that it grows very slowly
Mike>       compared to N).

You can also think of O(log N) as being no worse than proportional to
the number of digits in N.

Mike>     * O(N) -- time grows linearly as the size of the set grows (e.g.
Mike>       process every item in the set in turn)
Mike>     * O(N log N) -- see O(N) and O(log N), basically, you do an amount
Mike>       of work tied to the size of the whole set for every stage in an O(
Mike>       log N ) operation (this is comparatively rarely discussed in my
Mike>       experience).

Many sorting  algorithms are O(N log N).

Mike>     * O(N**2) O(N**3) etceteras -- time grows as an exponent (e.g.
Mike>       2,3,3.4,10) of the size of the set, so you get geometric explosion
Mike>       of running time as set size increases.  These algorithms tend to
Mike>       fail on even small N values.

Mike> You can have all sorts of fun learning how to combine and
Mike> re-factor the equations to get a final big-O notation, but for
Mike> the most part, practical coders can get away with recognising
Mike> and avoiding O(N**X) where X > 1 situations.  Recognising the
Mike> underlying patterns in O(log N) (binary searches and the like),
Mike> and O(1)-ish situations is useful as well, as they provide very
Mike> good general scalability.  The calculus of big-O notations is
Mike> just a convenient shorthand for people doing rigorous analysis
Mike> AFAICS.

O(N**1.5) is often bearable.

Mike> Keep in mind that the setup and constant factors of the
Mike> algorithm are factored out of the equations.  Something which is
Mike> O(1) may be slower than O(N**3) for all practical values of N if
Mike> the setup time for the O(1) algorithm approaches the age of the
Mike> universe.  Similarly, two algorithms which are O(N), but with
Mike> vastly different constant factors (implementations) may have
Mike> dramatically different running times.

Right.


-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From buzzard at urubu.freeserve.co.uk  Thu Apr 10 14:19:12 2003
From: buzzard at urubu.freeserve.co.uk (Duncan Smith)
Date: Thu, 10 Apr 2003 19:19:12 +0100
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com> <b73vov$d92$00$1@news.t-online.com> <HzGdnQNo1Z2-AAijXTWcog@comcast.com>
Message-ID: <b74ckl$itb$1@newsg4.svr.pol.co.uk>

"Terry Reedy" <tjreedy at udel.edu> wrote in message
news:HzGdnQNo1Z2-AAijXTWcog at comcast.com...
>
> "Stephan Diehl"
> > You are probably looking for "linear regression".
>
> Linear regression is designed for continuous, not binary outcome
> variables.
>
> "Duncan Smith"
> > But linear discriminant analysis and logistic regression would give
> you
>
> Logistic regression use 0-1 outcomes to fit a very particular model: p
> = exp(s)/(1+exp(s)) where p is a probability of 1 versus 0 and s is a
> linear sum of 'independent' variables.  It is typically used for
> modelling probability of response within one group.  It is less likely
> to be useful for two-group classification.
>
> Terry
>

Yes, you do have the link function as well as the linear predictor, and the
cut-off would be 0.5 rather than 1.  But in terms of absolute efficiency
there is (according to the literature) little to choose between linear
discriminant analysis and logistic regression.  Changing the cut-off would
be a simple way of attempting to minimise Alex's 'error-cost'.

Duncan





From skip at pobox.com  Sun Apr 13 20:05:30 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sun, 13 Apr 2003 19:05:30 -0500
Subject: isearch,
        LEAP and patents  was: Re: Could Emacs be rewritten in Python?
In-Reply-To: <yfsadeudlud.fsf@black132.ex.ac.uk>
References: <m2u1dcw22s.fsf@localhost.localdomain>
        <6ee58e07.0304061144.35c5ab92@posting.google.com>
        <m2k7e78j5s.fsf@localhost.localdomain>
        <mailman.1049662570.29853.python-list@python.org>
        <yfsr88fkppr.fsf@black132.ex.ac.uk>
        <eaeka.1333$1h6.8933@news-server.bigpond.net.au>
        <yfs1y0el9xc.fsf_-_@black132.ex.ac.uk>
        <e0caaad7.0304100951.57cb4572@posting.google.com>
        <yfsadeudlud.fsf@black132.ex.ac.uk>
Message-ID: <16025.64330.59727.700595@montanaro.dyndns.org>

    >> The LEAP patent (US patent 5,019,806) ...

    as> Have you got a publically available reference at hand by any chance?
    as> I'd be really curious to learn how the patent claim is justified.

Straight from the horse's mouth:

    http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=/netahtml/search-bool.html&r=4&f=G&l=50&co1=AND&d=ptxt&s1=5019806&OS=5019806&RS=5019806

Skip




From nikolai.kirsebom.NOJUNK at siemens.no  Fri Apr 25 10:50:15 2003
From: nikolai.kirsebom.NOJUNK at siemens.no (Nikolai Kirsebom)
Date: Fri, 25 Apr 2003 16:50:15 +0200
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com>
Message-ID: <ikiiavkejtem0ij2s6m1ukt67fe8ai6h1f@4ax.com>

On Thu, 24 Apr 2003 13:29:44 +0200, Nikolai Kirsebom
<nikolai.kirsebom.NOJUNK at siemens.no> wrote:

>Given the following class hierarchy:
>	
>class A(object):
>	InstCount = 0
>	def __init__(self):
>		self.__class__.__dict__['InstCount'] += 1
>
>class B(A):
>	InstCount = 0
>
>class C(A):
>	InstCount = 0
>
>Is the syntax used for incrementing the class-attribute for every
>instanciation 'the way' it should be done ?
>
>Nikolai Kirsebom
>
>

Thanks for all the replies.  Have to go back and start learning using
metaclasses....
Nikolai




From abelikov72 at hotmail.com  Thu Apr 17 12:13:55 2003
From: abelikov72 at hotmail.com (Afanasiy)
Date: Thu, 17 Apr 2003 16:13:55 GMT
Subject: Web Hosting
References: <2rjna.171$hT2.182473@news2.news.adelphia.net> <6Bjna.111823$It5.95226@news2.central.cox.net> <e5os9vg7aj3gfek0abt05s1bjh9pfvfvb3@4ax.com> <0Puna.38530$T34.1097507@news2.tin.it>
Message-ID: <htkt9v85i5i7f8q3bi7unf0ulitlc9ktbd@4ax.com>

On Thu, 17 Apr 2003 09:51:56 GMT, Alex Martelli <aleax at aleax.it> wrote:

>Afanasiy wrote:
>   ...
>> I was under the impression mod_python was not safe for shared
>> environments. If this is true, I don't think you will find a
>
>I'm not aware of such concerns; could you please provide an
>URL discussing the issue, or what gave you this impression?
>
>> reasonably price host which is also a good host.
>> 
>> If it is not true, I wonder why pair.com will not run mod_python.
>
>No idea, but inertia is often a helpful explanation.

No URL, but you should be able to find one. I just trust pair.

>> CGI, on the other hand, is CGI. :p
>
>Sure, but that's often not too bad anyway (particularly if you
>can run Webware in an external process and just use the CGI
>adapter to connect to Webware...).

Yes, CGI is bad. That's what I said. ;p

>Alex

This thread seems to have strayed from being helpful, as is common.



From nick at foobar#delete2email#.org  Sat Apr  5 06:42:20 2003
From: nick at foobar#delete2email#.org (Nick Hilliard)
Date: Sat, 05 Apr 2003 12:42:20 +0100
Subject: Why deleting large variables doesn't free up any memory?
In-Reply-To: <169a7283.0304040936.69e2b9a1@posting.google.com>
References: <169a7283.0304040936.69e2b9a1@posting.google.com>
Message-ID: <xizja.6994$pK2.7213@news.indigo.ie>

sdieselil wrote:

> I'm using Python 2.2.2 under FreeBSD. When I create a large variable
> (using for example "a = range(1000000)") 'top' command indicates that
> Python ate 12MB of swap space. But when I execute "del a" nothing
> happens, it doesn't release these 12MB! Using "gc.collect()" also
> doesn't help. What's the problem? What happened to Python's garbage
> collector?

Regardless of the other answers in this thread, I suspect that this problem is 
probably less important than you think.  If FreeBSD thinks that the memory isn't 
being used, it will be swapped out if there's a memory shortfall at any stage. 
The important factor is amount of memory which is being actively used at a 
particular time.

Nick




From itamar at itamarst.org  Mon Apr 21 11:04:38 2003
From: itamar at itamarst.org (Itamar Shtull-Trauring)
Date: Mon, 21 Apr 2003 11:04:38 -0400
Subject: ANN: Twisted 1.0.4
Message-ID: <20030421110438.4991f713.itamar@itamarst.org>

Twisted is an event-driven networking framework for server and client
applications. More detailed information can be found after the "What's
New" section.

For more information, visit http://www.twistedmatrix.com, join the list
http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python or
visit us on #twisted at irc.freenode.net.


What's New in 1.0.4
===================

This release is mainly a bug fix release. 1.0.4 includes large numbers
of bug fixes, small feature enhancements, additional tests and so on.
Additional reliability is assured by our deployment of Brian Warner's
Buildbot (http://buildbot.sf.net) for real-time testing
(http://twistedmatrix.com/~warner.twistd/ -- note that certain
functionality is experimental and therefore the tests will be failing,
but this does not affect the stability or performance of supported,
stable packages).


What is Twisted?
================

Twisted is an event-driven framework for building networked clients and
servers.  It contains a powerful and simple networking core, a
full-featured suite of interoperable protocols, among them a powerful
web server and applications framework.

Twisted supports many event loops for both server apps and GUI
integration on the client side, including:

 - Win32 events, including GUI support
 - GTK+
 - GTK+ 2
 - Qt
 - wxPython
 - Tkinter
 - Java

Twisted can run protocols over TCP, SSL, UDP, multicast, Unix sockets
and subprocesses. It also includes scheduling support, threading
integration, RDBMS event loop integration and other basic requirements
for networked applications.

Also included are implementations of many protocols. In some cases this
includes complete frameworks providing facilities on top of the base
protocol:

 - SSH
 - FTP
 - HTTP, including a complete web framework
 - XML-RPC
 - SOAP server framework
 - NNTP and complete NNTP server framework
 - SOCKSv4 (server only)
 - SMTP
 - IRC
 - telnet
 - POP3
 - AOL's instant messaging TOC
 - OSCAR, used by AOL-IM as well as ICQ (client only)
 - DNS
 - MouseMan serial mice, and GPS devices
 - Twisted Perspective Broker, a remote object protocol




From antonmuhin at sendmail.ru  Wed Apr 23 04:45:10 2003
From: antonmuhin at sendmail.ru (Anton Muhin)
Date: Wed, 23 Apr 2003 12:45:10 +0400
Subject: why not say just what wants an integer?
In-Reply-To: <87d6jeeyo3.fsf@jidanni.org>
References: <87d6jeeyo3.fsf@jidanni.org>
Message-ID: <b85jqo$mi3$1@news.peterlink.ru>

Dan Jacobson wrote:
> Say, does lack of ability to figure out how to make this work 
> $ python -c 'print range(map(lambda x:x+1,(11,131,11)))'
> Traceback (most recent call last):
>   File "<string>", line 1, in ?
> TypeError: an integer is required
> 
> indicate 1. newbie, 2. lazy newbie, 3. can't really blame me?
> 
> BTW, couldn't python be improved by saying just what wants an integer?
> 
> Yes, it happened after I added range(), so range wants the integer, I
> can tell that.  So list() won't help etc.
> 
> P.S. "pydoc print" gives nothing. Should it give some clue to the user
> wanting to know about "print", or is that some violation of principles?

Did you mean something like this:

print range(*map(lambda x:x+1,(11,131,11)))

anton.




From aahz at pythoncraft.com  Mon Apr  7 16:50:39 2003
From: aahz at pythoncraft.com (Aahz)
Date: 7 Apr 2003 16:50:39 -0400
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com>
Message-ID: <b6soav$7p2$1@panix1.panix.com>

In article <fd1c664e.0304071056.31a0d377 at posting.google.com>,
Danra <danra at walla.co.il> wrote:
>
>Here's my experience with 'super':
>
>>>> class b:
>	pass

No time to test, but try:

    class b(object):
        pass
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From claird at lairds.com  Fri Apr 25 11:12:06 2003
From: claird at lairds.com (Cameron Laird)
Date: Fri, 25 Apr 2003 15:12:06 -0000
Subject: ftp problem on solaris machines with python
References: <88bc63c6.0304240407.674220cb@posting.google.com> <b88rae$etu$1@heckle.cc.uregina.ca> <88bc63c6.0304250651.5ec01610@posting.google.com>
Message-ID: <vaik267b6da20b@corp.supernews.com>

In article <88bc63c6.0304250651.5ec01610 at posting.google.com>,
Doug Farrell <writeson at earthlink.net> wrote:
>Hi,
>
>We're not gzipping the files first because we're really short on space
>on the Solaris machine and gzipping takes quite a bit of temp space to
>work, unfortunately. By croaking I mean the the python program croaks
>in mid transmission while ftping a large file. However, ftping the
>file from the shell command line works fine.
>
>One of our guys found a module pyExpect (I think) that will let our
>program ftp using the command line rather than the ftplib module.
>We're looking into that.
			.
			.
			.
Something's fishy.  ftplib ought to work.  If it
doesn't, it ought to be fixed.

I can understand, though, that you might be in a
situation where you can ill-afford to research such
matters.  HOWEVER, do not think you need to resort 
to pyExpect; the standard Unix ftp client can, in
general, be driven purely from the command line.  I
recommend you use it so.  Does <URL: http://
phaseit.net/claird/comp.unix.programmer/ftp_automation.html#re-directed >
give you enough of a hint about how to do so?

-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From oblivious at web.de  Sun Apr  6 16:46:39 2003
From: oblivious at web.de (Dominic)
Date: Sun, 06 Apr 2003 22:46:39 +0200
Subject: Programming Erlang style in Python :-)
Message-ID: <b6q3pj$mp3$03$1@news.t-online.com>

Python's dynamic capabilities allow it to
program finite state machines and asynchronous
message passing similar to the programming
language Erlang which is used in soft-realtime
telecommunication systems.

The first demonstration implements a finite state
machine in a recursive manner. By using
generators recursion is actually removed :)

The 2nd demonstration shows two objects which
send messages to each other asynchronously
Real argument pattern matching like in Erlang
has not been implemented though.

Enjoy :)

Ciao,
  Dominic

P.S. This 'pattern' could be used for keeping state
      in a server for example.

Output:

>>>>>>>>>>>>>>>>>>>>>>>>>
>>> Finite State Demo >>>
>>>>>>>>>>>>>>>>>>>>>>>>>
access denied to Schmitt
access denied to Mayer
2 failed attempts so far
access denied to Schmitt
access denied to Mayer
4 failed attempts so far
access denied to Schmitt
access denied to Mayer
6 failed attempts so far
access granted
connected Dominic, previous failed attempts 6

>>>>>>>>>>>>>>>>>>>>>>>>
>>>  Ping-Pong Demo  >>>
>>>>>>>>>>>>>>>>>>>>>>>>
ping got pong
pong got ping
ping got pong
pong got ping
ping got pong
pong got ping
ping got pong
pong got ping
ping got pong
pong got ping
ping got pong
pong got ping
ping got pong
pong got ping

Source:


from __future__ import generators
import sys

class Queue(list):
	
	
	def send(self,cmd, *args):
		self.insert(0,(cmd,args))
		self.items={}

	def __init__(self):
		self.items={}
		self.queue=self
		
	def __getattr__(self,name):
		return self.items[name]

	def receive(self,expect):
		"""Lot's of interesting stuff could be done here"""
		expect=expect.split(',')
		msg=self[-1]
		if msg[0]==expect[0]:
			args=list(msg[1])
			expect=expect[1:]
			for i in xrange(0,len(expect)):
				self.items[expect[i]]=args[i]
			self.pop()
			return 1
		else:
			return 0

############################################################

def ping(_q):
	if _q.receive('pong,pid'):
		print 'ping got pong'
		_q.pid.send('ping',_q)
		
	yield ping(_q)

def pong(_q):

	if _q.receive('ping,pid'):
		print 'pong got ping'
		_q.pid.send('pong',_q)
		
	yield pong(_q)

############################################################

def client(_q, _name, _password, _failed=0):
	
	if _q.receive('login,name,password'):
		if _q.name==_name and _q.password==_password:
			print 'access granted'
			yield connected(_q,_name,_failed)
		else:
			
			print 'access denied to %s' % _q.name
			if _failed>20:
				print 'too many bad login attempts'
				raise 'bad login'

			yield client(_q,_name,_password,_failed+1)
			
	elif _q.receive('info'):
		print '%d failed attempts so far' % _failed
		yield client(_q,_name,_password,_failed)
	else:
		_q.pop()

def connected(_q,_name,_failed):

	if _q.receive('info'):
		print 'connected %s, previous failed attempts %d' % (_name,_failed)

	yield connected(_q,_name,_failed)
	
############################################################

print '>>>>>>>>>>>>>>>>>>>>>>>>>'
print '>>> Finite State Demo >>>'
print '>>>>>>>>>>>>>>>>>>>>>>>>>'

sys.stdin.readline()

queue=Queue()
c=client(queue,'Dominic','1234',_failed=0)

for x in xrange(1,4):
	
	queue.send('login','Schmitt','798')
	queue.send('login','Mayer','123')
	queue.send('info')

	# Scheduler
	while 1:
		c=c.next()
		if len(queue)==0:
			break

queue.send('login','Dominic','1234')
c=c.next()

queue.send('info')
c=c.next()




print '\n>>>>>>>>>>>>>>>>>>>>>>>>'
print '>>>  Ping-Pong Demo  >>>'
print '>>>>>>>>>>>>>>>>>>>>>>>>'

sys.stdin.readline()


# Two "tasks" which exchange messages

queue_pong=Queue()
queue_ping=Queue()

# initiate
queue_ping.send('pong',queue_pong)

pg=pong(queue_pong)
pi=ping(queue_ping)

# 'Scheduler'
for tmp in xrange(1,8):
   pi=pi.next()
   pg=pg.next()








From aleax at aleax.it  Thu Apr 17 02:09:03 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 08:09:03 +0200
Subject: Pythonic way of web-programming
In-Reply-To: <1050526692.30714.586.camel@lothlorien>
References: <cQ_la.5597$LB6.158861@news1.tin.it> <dl8na.20637$LB6.515421@news1.tin.it> <1050526692.30714.586.camel@lothlorien>
Message-ID: <200304170809.03363.aleax@aleax.it>

On Wednesday 16 April 2003 10:58 pm, Ian Bicking wrote:
   ...
> For a difference in interface, Cheetah takes a list of namespaces that
> get folded together.  I haven't looked at preppy, but let's say it takes
> a dictionary or a single namespace.  That's a different interface, but
> it can be turned into one interface, and then both templating languages
> will benefit.

Ah, I see your point, now -- you're talking about the way a templating
operation is initiated _from the Python side_ -- Cheetah's class Template
constructor (returning a class instance x that's then normally expanded
by str(x)), preppy's getModule function (returning a module object m that's
then normally expanded by calling m.run with a dict and output specs), &c.

Yes, I agree, it would definitely be possible to merge these interfaces
into a single one, so that Python code that needs templating can most
easily switch between using different templating packages -- at least
as long as the templating model is somewhat abstractly compatible,
enough commonality should be easily identifiable for this purpose,
while leaving each templating package the choice of what templating
little-language to implement, which is clearly the major differentiatior
(while in the DB-API analogy, the difference between SQL dialects is
hopefully minor in comparison).


Thanks for clarifying,

Alex





From sholden at holdenweb.com  Sun Apr 13 09:58:56 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sun, 13 Apr 2003 13:58:56 GMT
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <d116fbae.0304100704.68ef4833@posting.google.com> <c48$3e960655$516049d2$3498@nf1.news-service.com> <3e96efac$0$49107$e4fe514c@news.xs4all.nl> <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk> <3e96f481$0$49099$e4fe514c@news.xs4all.nl> <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk> <3e970abc$0$49101$e4fe514c@news.xs4all.nl> <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk> <3e972e43$0$49109$e4fe514c@news.xs4all.nl> <b77hpo$59a$0@216.39.172.122> <6ZY$RKABo9l+Ew+N@jessikat.fsnet.co.uk>
Message-ID: <A2ema.65154$It5.1935@news2.central.cox.net>

"Robin Becker" <robin at jessikat.fsnet.co.uk> wrote in message
news:6ZY$RKABo9l+Ew+N at jessikat.fsnet.co.uk...
> In article <b77hpo$59a$0 at 216.39.172.122>, Bengt Richter <bokr at oz.net>
> writes
> .....
> >>I really can't explain these numbers either? Localhost is 10 times
slower
> >>than over the network???
> >
> >IIRC, there was some discussion some time ago about this being a
particular
> >FreeBSD pessimality problem that wasn't true of Linux, but someone was
fixing
> >it?
> >But FreeBSD has good networking reputation I thought, so this needs
double
> >check,
> >for sure. Maybe it was for an old version?
> >
> ....this is version 4.5. I guess things would be easier if one could be
> sure that it's not a dns type thing, but I see from the dev list that
> there's already some discussion about how/where dns lookups can be
> avoided/speeded up with sockets which have a supposedly known address.
> Naively I thought it would be relatively easy to just get a 32 / 64 bit
> int back from some socket module function and then shove that into the
> address part of the socket address.
>
> Given the above perhaps it's not dns at all, but some other things that
> get in the way. Do OSes impose limits on number of ports or
> connections/second etc?
>

Seems that the easiest way to get insight into the network traffic you're
generating is to install Ethereal or some similar software and simply look
to see what's going over the wire.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From wweexxsseessssaa at telusplanet.net  Mon Apr 21 18:36:37 2003
From: wweexxsseessssaa at telusplanet.net (John Hall)
Date: Mon, 21 Apr 2003 22:36:37 GMT
Subject: Regarding Dictionaries in python
References: <mailman.1050957134.26691.python-list@python.org>
Message-ID: <cgs8avgk927ufqdf72mnhe0thon1i4bpv2@4ax.com>

On Mon, 21 Apr 2003 22:34:35 +0200, "Mehta, Anish"
<Anish.Mehta at enst-bretagne.fr> wrote:

>why it is not showing the items in order in which i have inserted. 
>Please tell me why it is happing like this.

Because the order of items in dictionaries is inherently undefined, or
whimsical if you prefer. Your book should tell you that, but of course
you need to read the relevant sentence.

There are cookbook recipes for sorting them if you really need to.

-- 
John W Hall <wweexxsseessssaa at telusplanet.net>
Calgary, Alberta, Canada.
"Helping People Prosper in the Information Age"



From mads at troest.NEVERMORE.dk  Tue Apr 15 13:09:08 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Tue, 15 Apr 2003 19:09:08 +0200
Subject: Using __repr__ or __str__ for own printable class?
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org> <b7di0n$dd6gj$1@ID-169208.news.dfncis.de> <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net> <1050390062.742825@yasure>
Message-ID: <qz134np9z11d$.vzaddqyidl9x.dlg@40tude.net>

On Tue, 15 Apr 2003 07:01:04 -0000, Donn Cave wrote:

Hi Donn, thanks for your feedback;

> I personally think the simplest and most useful way to look at it is
>   __str__: your object's data as a string
>   __repr__: your object as a string.

That's a good rule of thumb, I think; I will remember that. :)

Regards,
/\/\\ads Orbesen Troest



From staschuk at telusplanet.net  Tue Apr 22 22:24:17 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 22 Apr 2003 20:24:17 -0600
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <3EA4A843.2050507@tismer.com>; from tismer@tismer.com on Tue, Apr 22, 2003 at 04:26:11AM +0200
References: <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <3EA2C43F.2080802@tismer.com> <20030421121843.A596@tibia.amotlpaa.bogus> <3EA4A843.2050507@tismer.com>
Message-ID: <20030422202417.A1462@tibia.amotlpaa.bogus>

Quoth Christian Tismer:
> Steven Taschuk wrote:
  [...]
> > Before studying continuations, function calls are function calls
> > and loops are loops.  While studying continuations, function calls
> > are not function calls and loops are not loops.  After studying
> > continuations, once again function calls are function calls and
> > loops are loops.
> 
> But then, you should have understood how to build function
> calls and loops from continuations, and that in fact the is no
> "call" or "return" in the first place, but these are artificial,
> from a POV where continuations are the basic building block.

Indeed; after studying continuations you are enlightened.

(My blurb above is a well-known quote about Zen, after the
substitutions
    s/Zen/continuations/g
    s/men/function calls/g
    s/mountains/loops/g
A wee joke.)

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From rmunn at pobox.com  Mon Apr 14 12:50:40 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 14 Apr 2003 16:50:40 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz>
Message-ID: <slrnb9l507.1q1.rmunn@localhost.localdomain>

Paul Foley <see at below.invalid> wrote:
> On Mon, 14 Apr 2003 05:52:31 GMT, Carl Banks wrote:
> 
>> Ok, let me back off a bit.  That you might want to change a program's
>> state in a dynamic scope is reasonable.
> 
>> Typically, however, that kind of thing should be handled by passing
>> objects around as arguments.  For example, consider the use of current
>> buffer in Emacs.  (We are still leaving Emacs issues out of it; I'm
>> just using it as an example of bad programming.)  Most functions in
>> Emacs operate on the current buffer, a piece of global state.  In
>> order to operate on a non-current buffer, you have to temporarily
>> change the current buffer.  That is bad programming; functions should
>> take the buffer to operate on as an argument.
> 
> Don't forget the ~20000 other variables that you'd have to pass around
> as explicit arguments to *every single function*.
> 
> Good luck trying to maintain that code!
> 
>> However, I think changing a global value within a dynamic scope, in
>> situations where it wouldn't be better to pass around an argument, is
>> a very uncommon concern. 
> 
> Well, you think wrong.
> 
>> So if you need that kind of thing, please don't destroy the language
>> by adding dynamic scoping.  Just live with a try ... finally.
> 
> I.e., write broken and fragile code just to avoid a perfectly good
> construct that you happen to dislike for religious reasons.
> Brilliant!

How is try ... finally "broken and fragile code"? How is this:

    try:
        old_stdout = sys.stdout
        new_stdout = cStringIO.StringIO()
        sys.stdout = new_stdout
        function_whose_output_I_want_to_capture()
        do_something_with(new_stdout)
    finally:
        sys.stdout = old_stdout

in any way inferior to this (making up a semi-Pythonic syntax):

    with sys.stdout = cStringIO.StringIO():
        function_whose_output_I_want_to_capture()
        do_something_with(sys.stdout)

Yes, sample #2 is shorter. But it would require adding a new keyword to
the language, which is a pretty heavy price to pay. And sample #2
doesn't need a temporary variable -- but sample #1's temporary variable
is a function-local variable, so speed and namespace-clutter costs are
minimal.

What I really want to know is why you would consider sample #1 "broken
and fragile code". I can't think of any circumstances where sample #1
would fail to achieve its desired effect. Can you?

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From cmkleffner at gmx.de  Wed Apr 30 06:10:34 2003
From: cmkleffner at gmx.de (cmkl)
Date: 30 Apr 2003 03:10:34 -0700
Subject: os.spawn
References: <Xns936C9D9E91B1Bthewolffman@194.109.133.29>
Message-ID: <3b091a1c.0304300210.28c90901@posting.google.com>

Jeroen Wolff <no at mail.com> wrote in message news:<Xns936C9D9E91B1Bthewolffman at 194.109.133.29>...
> Hi,
> 
> I want to run some shell commands from within python. 
> Wat i normal do at the shell prompt is:
> 'flow-cat <dirname> | flow-stat <param> > outputfile.txt'
> 
> I like to do this from within python and wait for it to finish and the 
> process the outputfile.txt
> Is this possible? The problem is with '|' and '>" piping....
> 
> Can somebody please help?
> 
> Thanks,
> 
> Jeroen Wolff

Hi,

I didn't tried this right now, but you may find Trent Mick's
process 0.5.0 interesting. See his talk at pycon:
http://starship.python.net/~tmick/downloads/process/pycon_lightning_talk.ppt
and his Website: http://starship.python.net/crew/tmick/

Basically process.py supports three different types of controlling
processes; Unix and Windows is supported as far as I understand;
Windows support needs Hammonds win32 module.

1) Process(): launch it and forget about the output (like os.system)
2) ProcessOpen(): interact with it (like os.popen)
3) ProcessOpenProxy(): eventlike interface to the launched process,
   needs Python with thread support (default on most platforms)

BTW: pexpect (on sourceforge) is another way to interact with external 
processes. pexpect does not need thread support but runs processes 
within pty's (pseodoterminal), so it is Unix only.

For your problem you can give it a try with ProcessOpen() 
without piping the output to a temporary file but use the 
stdout directly:

>>> import process
>>> CMD = 'flow-cat <dirname> | flow-stat <param>'
>>> p = process.ProcessOpen(cmd=CMD)
>>> p.stdout.read()
.......
>>> p.wait()   # returns the child's exit status

See Mick's presentation ...

Carl



From ckea25d02 at sneakemail.com  Tue Apr  1 19:40:38 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 12:40:38 +1200
Subject: converting from shell script to python
In-Reply-To: <b6bccj$3fsej$1@ID-15245.news.dfncis.de>
References: <kfn0jaoig1.fsf@shaoz.activesky.com.au> <b6bccj$3fsej$1@ID-15245.news.dfncis.de>
Message-ID: <b6db2p$48k88$1@ID-169208.news.dfncis.de>

Mathias Waack wrote:
> import os
> 
> ROOTDIR='/usr/lib'
> CLASSPATH="%s/a/a.jar:%s/b/b.jar:%s/c/c.jar" % \
>           (ROOTDIR,ROOTDIR,ROOTDIR)
> os.system('java -classpath %s foo' % (CLASSPATH))

Alternatively, you can substitute things from a
dictionary instead of a tuple:

   vars = {'ROOTDIR': '/usr/lib'}
   CLASSPATH = \
     "%{ROOTDIR}s/a/a.jar:%{ROOTDIR}s/b/b.jar:%{ROOTDIR}s/c/c.jar" % vars

That technique can be more readable if you have
more variables that you want to substitute into
the string, since you can name them where they're
used instead of relying on positional matching.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From missive at frontiernet.net  Fri Apr 11 17:17:25 2003
From: missive at frontiernet.net (Lee Harr)
Date: Fri, 11 Apr 2003 21:17:25 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <boejm-khk.ln1@news.lairds.org>
Message-ID: <FhGla.401$Zw2.21@news02.roc.ny.frontiernet.net>

In article <boejm-khk.ln1 at news.lairds.org>, Kyler Laird wrote:
> OTOH, if we're teaching programming (I have a Scheme course
> going now.) it's a whole different ballgame.  There I feel it
> is vital to understand the basics and be able to build them
> from scratch.
> 

> So...I think you probably all understand the difference
> between Python as a tool and Python as a base language for
> teaching programming.  I'd just like to see a more deliberate
> distinction so that we don't confuse the goal of "Python and
> Schools".
> 


I have a feeling that going beyond using Python as a great tool
for getting work done with computers is probably out of scope
for most schools, but I can certainly imagine a small percentage
of students signing up for an essentially college-level comp-sci
course in their last year or two of school.

Ideally, the students would already be quite proficient at using
the tools provided by Python to get things done, and could use
those tools to prototype the lower-level details of the tools
themselves as a learning exercise.

Then they might go in to the "comparative anatomy" of learning to
do the same things with C or C++ or Java. And, of course the best
part is that they have all of the CPython and Jython source code
so they can see how experienced programmers have solved these
problems. They might even study how the implementations have
changed historically. (Open source software... wow.)




From just at xs4all.nl  Tue Apr  1 14:24:38 2003
From: just at xs4all.nl (Just)
Date: Tue, 01 Apr 2003 21:24:38 +0200
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <just-0075D9.21243801042003@news1.news.xs4all.nl>

In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer <mwm at mired.org> 
wrote:

>     Roman numeral literals will consist of the characters M, D, C, L,
>     X, V and I, and only those characters.  They must be in upper
>     case, and represent an integer with the following rules:

I *strongly* object to the upper case requirement. Also, when using 
lower case, a j should be allowed to use instead of an i, if it's 
preceded by an i. Here's some info:

  http://www.unc.edu/~rowlett/units/roman.html
  http://www.wilkiecollins.demon.co.uk/roman/calco2.htm

Google found much more.

Just



From alloydflanagan at attbi.com  Wed Apr 16 15:58:30 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 16 Apr 2003 12:58:30 -0700
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com>
Message-ID: <a8b7f07a.0304161158.29c0ef6b@posting.google.com>

Peter van Kampen <news at datatailors.xs4all.nl> wrote in message news:<slrnb9q90a.mhg.news at datatailors.com>...
> In article <mailman.1050457621.2105.python-list at python.org>, Mike C.
> Fletcher wrote:

> I think I almost 'get it'. Except who or what decides what Big-O a
> certain algorithm has. Is that by 'hunch' or by running benchmarks or
> are there still other ways to recognize 'Big-O's (except for the
> formal computations that I understand are not very common)? 
> 
Technically the only way to determine the O() for a particular
algorithm is to do a complete mathematical analysis.  One of the other
posts has an excellent simple example.  You can do comparative times
for particular data sets, and fit a curve to the times, but that
doesn't guarantee that the algorithm will behave the same way for some
other set of data sets.

This mathematical analysis of an algorithm's behavior (and best-case,
average, and worst-case performance) can be trivial.  For example, x +
1 is O(1).  Or it can be so complex that proving the exact behavior is
a Ph.D. thesis, or even an unsolved problem.

> But these are not really 'algorithms'. Most algorithms are a lot more
> complex even when reduced to it's essentials. I understand why
> quicksort is better than bubblesort but I don't see an obvious way to
> label quicksort O(??). 
> 
Everything's an algorithm :).  Quicksort analysis isn't terribly
complicated as these things go.  There's got to be an analysis up on
the web somewhere.



From nas-usenet at arctrix.com  Sun Apr  6 15:01:35 2003
From: nas-usenet at arctrix.com (Neil Schemenauer)
Date: Sun, 06 Apr 2003 19:01:35 GMT
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <e9570f37.0304060944.79564f88@posting.google.com>
Message-ID: <jQ_ja.12641$7w2.9369@nwrddc01.gnilink.net>

Graham Fawcett <graham__fawcett at hotmail.com> wrote:
> [...] what kinds of applications?  We are talking only about
> Web application frameworks, which are request handlers and are
> therefore event-driven by definition.

For example, we have a web application that interacts with
FedEx's servers in order to create new shipping labels, track
packages, etc.  With an event driven system, waiting for a
response for FedEx would freeze the entire server.  Obviously
that can be fixed the hooking into the event loop.  It's just a
matter of programming, right?

Another example: we have an application that does some CPU
intensive operations.  We have multiple CPUs in our server.  An
event driven system would not (normally) take full use of the
processors.  Also, clients who have IO bound requests would be
blocked by one CPU bound request.

I have nothing against the event driven model.  However, it is
not the solution to every network server problem.

  Neil



From dave at pythonapocrypha.com  Fri Apr 18 16:12:22 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Fri, 18 Apr 2003 13:12:22 -0700 (PDT)
Subject: Classes in Python
In-Reply-To: <20030418110750.D709@tibia.amotlpaa.bogus>
References: <3EA01A9B.4030209@vnet.net> <20030418110750.D709@tibia.amotlpaa.bogus>
Message-ID: <Pine.LNX.4.51.0304181302450.23611@ns0.tbye.com>

On Fri, 18 Apr 2003, Steven Taschuk wrote:

> Quoth Bill Martin:
> > I'm wondering about the value of allowing a class definition like this:
> >
> >              class C:
> >                pass
[snip]
> One slightly different form which I find very useful:
>
>     class HullBreachError(Exception):
>         pass

Another use is if you don't like the 'global' keyword much:

class GV: pass
GV.someCounter = 0
GV.someOtherVar = 'foo'

def Biff():
  GV.someCounter += 1

That way it's perfectly obvious that something is global.

(I hope this doesn't start the "why do I have to type 'self.'" thread yet
again, although it _has_ been almost a week so it's overdue <0.5 wink>).

-Dave




From duane at franz.com  Thu Apr 17 12:22:10 2003
From: duane at franz.com (Duane Rettig)
Date: 17 Apr 2003 09:22:10 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr>
Message-ID: <4znmprt59.fsf@beta.franz.com>

laotseu <bdesth at removethis.free.fr> writes:

> Ok, cross-posted and fu2 positionned.

[re-cross-posted]

> Lispers friends, let's have you're opinion on that point : is Common
> Lisp a functionnal language ?

Other CL-ers have already correctly answered you.  And the answers
assume an understanding of the dimensions of this question.  There are
in fact at least a couple of dimensions to your question - it can be taken
as either a question of capability or as a question of focus.  The "is",
on the other hand, implies a question of definition (e.g. what _is_ it?)


Definition: CL is a powerful, general-purpose programming language.

Capability:  Can CL users program functionally in CL?  Absolutely.
             Can CL users write non-functional programs?  Absolutely.

Focus:  Does CL focus on or force a functional programming style? No.
        Might some CL users force a functional programming style on
        their programs?  Yes, of course.  Are such users still programming
        in CL?  Yes.  Must other programmers modifying the original code
        stick to the same functional model of programming?  Not by CL
        standards; any such agreement to stay within FP boundaries would
        be external to CL (and would probably involve some kind of monetary
        transaction :-)

-- 
Duane Rettig    duane at franz.com    Franz Inc.  http://www.franz.com/
555 12th St., Suite 1450               http://www.555citycenter.com/
Oakland, Ca. 94607        Phone: (510) 452-2000; Fax: (510) 452-0182   



From peter at engcorp.com  Fri Apr 11 13:58:25 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 11 Apr 2003 13:58:25 -0400
Subject: Confusion: assignments create object references.
References: <mailman.1050080058.23175.python-list@python.org>
Message-ID: <3E970241.483698C7@engcorp.com>

"Havens, Peter" wrote:
> 
> I'm a silly rabbit. I hand copied the test incorrectly. The real test was:
> 
> >>> L = 1
> >>> M = L
> >>> L
> 1
> >>> M = 0
> >>> L
> 1
> >>>
> 
> Those are the real results (I double checked), but now I'm really confused. 
> Why does the former test (the one I lied about) above give expected results, 
> and the later doesn't?

Try this on for size:

The thing on the left side of the equal sign is "bound" (made to reference) 
the objects on the right side of the equal sign.  This is a universal, 
whether you are referring to simple names such as M and L, or *entries in
a list*, such as when you use M[0] and L[0].
 
The behaviour you show above is simple to explain, of course, as there is
never any relationship between M and L in this case.  Initially you
set L to be bound to the integer object containing the value 1.  Then
you copy the *reference* for that object to M, or to put it better, you
rebind the name M to the same object to which L is bound.

Later on, you rebind M to the different integer object containing the 
value 0.  L is of course not affected by this, so it stays bound to
the object 1.

Now take the case of lists.  The fundamental fact you are overlooking or
missing is that lists are *lists of references to things*.  Now you have
a list L, bound to a list object which contains three references, to the
objects 0, 1, and 2:

L = [0, 1, 2]

You bind the name M to the same object as L references: the *list* itself
(i.e. this has nothing to do with the objects referenced by the list).
So M and L "point" to the same object.

M = L

Now you go into that object, the list itself, via one of the *two* names
which are bound to it, and you index into it to find the first reference,
which is currently "pointing" to the integer object 0.  You then change
*that* reference to point to the integer object 9.  You have not changed
anything else in the list, and you have not in any way affected the names
which are bound to the *list*, only one of the references in the list.

M[0] = 9

When you then try to look at the list using the *other name* which is 
bound to it, you of course will get exactly the same list as you saw
using the other name:

>>> L[0]
9

In Python, think always in terms of objects and references to them, 
stored in names or as items in, for example, the list above.  If you
know about pointers, you won't be far off if you think of every name,
or every element in containers such as lists and tuples, as being
a pointer to some object.

(A lengthy explanation for what others have explained better and more
succinctly in the past.  Search the archives for them, and check the 
FAQ, if you are still confused.)

-Peter



From piotrus at sympatico.ca  Sat Apr 19 23:45:00 2003
From: piotrus at sympatico.ca (Peter)
Date: Sat, 19 Apr 2003 23:45:00 -0400
Subject: red hat 8.0 and python
References: <3EA21326.7070100@sympatico.ca>
Message-ID: <3EA217BC.9090902@sympatico.ca>

Aghh sorry for spamming but perhaps I was not clear enough in what I am 
asking:

- I would like to use the Tkinter library.

- Would like to know whether it is safe to unistall the current version 
of python which came with RedHat (and possibly how to go about doint it)

- Detailed instructions on how to install a newer version (a link to a 
webpage would be great)

Thanks again
Peter



Peter wrote:
> I've just installed Red hat 8.0, and although python 2.2.1 comes with 
> it, i cannot use anything which uses the Tkinter library (this includes 
> IDLE). I a few other packages work, and the interpreter runs fine. I 
> suppose that this might be due to the setup, but I cannot figure it out. 
>  Could it be that Tkinter module does not come with the python 
> distribution which came with Red Hat 8.0? I read that it should, but I 
> searched my whole computer for anything with ?tk? in the name with no luck.
> 
> I downloaded the rpm's for the python 2.2.2 (from www.python.org) but 
> simple "rpm -i" or "rpm -U" installing gives me errors; conflicts with 
> the currently installed version. I am not quite sure whether I should 
> try uninstalling what is there now because I've read that the a few 
> programs need it to run.
> 
> Any suggestions? Also if anyone knows of a webpage listing the 
> step-by-step instructions on how to install/uninstall/upgrade the python 
> distribution, please reply. I have read the info provided on 
> www.python.org, but it is not detailed enough for me.
> 
> Thanks
> Peter
> 





From gtalvola at nameconnector.com  Thu Apr 10 11:13:26 2003
From: gtalvola at nameconnector.com (Geoffrey Talvola)
Date: Thu, 10 Apr 2003 11:13:26 -0400
Subject: ssl error "PRNG not seeded"
Message-ID: <61957B071FF421419E567A28A45C7FE59AF3AD@mailbox.nameconnector.com>

I have a Python program that uses SSL to make https requests every few
minutes.  The program happens to be a Windows NT Service.  I have multiple
machines running this service and they have been working fine for months.
Overall, there have been hundreds of restarts of the service without
problems.

Yesterday, one of these machines was rebooted.  When my service started, it
couldn't make any SSL requests.  My logfile reveals that every time it tried
to call socket.ssl(sock) the following exception was raised:

sslerror: (1, 'error:24064064:random number generator:SSLEAY_RAND_BYTES:PRNG
not seeded')

Once I stopped and started the service, everything was fine.

My question is, why would this program work fine through hundreds of runs,
then just this one time flake out and claim that the PRNG wasn't seeded?  (I
hate intermittent problems like this!)  Could it have anything to do with
the fact that the machine was rebooted?

Is this likely to be a problem in the Python SSL implementation or an
OpenSSL problem?

Details:

Python 2.2.1
_socket.pyd compiled with OpenSSL 0.9.7 using VC++ 6.0


- Geoff




From theller at python.net  Fri Apr 25 09:14:09 2003
From: theller at python.net (Thomas Heller)
Date: 25 Apr 2003 15:14:09 +0200
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk>
Message-ID: <of2uen32.fsf@python.net>

Michael Hudson <mwh at python.net> writes:

> mis6 at pitt.edu (Michele Simionato) writes:
> 
> > The meta-type conflict is rather annoying, indeed. According to the
> > metaclass book, the language should generate the correct metaclass
> > that solves the conflict automatically, without burden for the
> > programmer. Python is not that magic (yet) therefore you have to
> > solve the conflict yourself.
> 
> I guess a potential problem is that if you do
> 
> class C(object):
>     __metaclass__ = A
> 
> class D(object):
>     __metaclass__ = B
> 
> class E(C, D):
>     pass
> 
> class F(C, D):
>     pass
> 
> you'd like both E and F to have the *same* (rather than merely
> equivalent) metaclasses.  This would seem to be tricky to arrange.

Wouldn't a caching class factory be able to do this?

Thomas



From aleax at aleax.it  Sat Apr 12 13:16:06 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 17:16:06 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <qRXla.7011$T34.190925@news2.tin.it>

Harald Massa wrote:

>> [cut]
>> I think python is a very good language for education. You could get in
>> contact with local schools.
> 
> Thomas, I am not so sure about this.
> 
> Yes, I love Python, and it is the most elegant programming language I
> ever met.
> 
> Python gives you objectoriented, functionorientet and imperative
> programming paradigm in one fitting package.
> 
> Python does not bother beginners with a million typecasts. Very great.
> 
> For education there are small drawbacks:
> 
> - 50% of all exercises "to teach programming" are already solved in the
> standard-library or are a method of a buildin object / are allready build
> in the core language
> 
> - 45% of these exercises are solved within "the python cookbook",
> available also online

Why are these "drawbacks"?  I see them as opportunities to:

-- have users develop/learn simple solutions to see how some things
   could be done then compare them with "professionally developed"
   approaches

-- then use the professional stuff as building-blocks for higher level
   problems (closer to applications of interest to the student)

When I taught programming I did show, for example, how to build an
integer multiplication function with a loop on top of addition and
shift-left -- even though the programming language in use DID have a
built-in multiplication operator, I thought that was instructive
(for my target audience -- engineering students in their 3rd year
of university).  Then, of course, we moved on to using multiplication
as a "built-in".  Similarly, later in the course, I showed how to
implement saxpy in bare Fortran -- then matrix multiplication on
top of saxpy -- etc, etc -- and as the course progressed we then
used excellent library implementations of those BLAS tools, not the
one we had put together "naively" (the subtext: I thought, and still
think, that engineers should have a feel for how things work inside,
BUT also resist the temptation to reinvent the wheel when good tools
are available in libraries).

Isn't the situation here quite similar?


> - if you learn programming with Python and later have to use Java or
> Visual Basic, it will be the most frustrating experience for the young
> fellows.

If you learn arithmetic with arabic notation and later have to use
roman numerals (try long division with them one of these days... but
even multiplication is cumbersome enough!-), it will be at least as
frustrating.  So, should we teach arithmetic using roman numerals
rather than arabic ones?-)

Any time you learn a good tool and later are forced to use inferior
tools it can be frustrating.  If the prospect of such forced use of
inferior tools is realistic, then the course that teaches the better
tool might choose to spend some time explaining how one can manage
to get things done anyway even with the inferior ones; it's not a
sound motivation to avoid teaching the better tool, anyway!


> Anyway, I regret not having met Python earlier in my programming career.

Sure -- me too.  Though "from pain we learn" (Pindarus), and maybe I
wouldn't have learned quite as much without the pain of dealing with
lower-level languages throughout my career, still I'd probably have
gotten more done in my life;-).


Alex




From machowsk1 at attbi.com  Sun Apr 27 19:40:17 2003
From: machowsk1 at attbi.com (Zipper114)
Date: Sun, 27 Apr 2003 23:40:17 GMT
Subject: Sorting files
Message-ID: <ATZqa.667651$S_4.718911@rwcrnsc53>

I'm a bit of a noob when it comes to Python, so this might seem a bit
trivial, but.....

How can I get the listing of files of a particular type in a directory,
sorted by date? For example, I using the following call to get all the
JPEG's in a dir:

Images = glob.glob1(MyPath+"\\images", "*.jpg")


I want to do just this, but have them sorted by date. Any ideas?





From glenfant at NOSPAM.bigfoot.com  Tue Apr 22 08:26:31 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Tue, 22 Apr 2003 14:26:31 +0200
Subject: Zope and FreeBSD
References: <e786d63d.0304201806.4c5293a6@posting.google.com>
Message-ID: <b83bkt$2v7h$1@biggoron.nerim.net>

"Wenshan Du" <python at dohao.org> a ?crit dans le message de news:
e786d63d.0304201806.4c5293a6 at posting.google.com...
> Hi,
>
> I've built a website with Zope 2.5.1 and FreeBSD 4.3, it's better than
> other solutions.
>
> The server got so slow recently that I have to reboot every 3 days,
> and Zope stop sometimes.
>
> I don't know why, is zope not stable on FreeBSD? or FreeBSD got virus?

It depends on what you call *slow*.
It's a good practice to restart Zope servers that have high loads.
My Zope instances are restarted automatically every night.
Be warned that Zope requires a big amount of memory.
You should not have less than 256Mb on a production server. That's perhaps
your problem.
Just make a "top" when you notice this problem to see what's going on.

HTH

--Gilles




From artyprog at wanadoo.fr  Thu Apr 17 06:59:09 2003
From: artyprog at wanadoo.fr (Salvatore)
Date: Thu, 17 Apr 2003 12:59:09 +0200
Subject: variable in Python
Message-ID: <b7m1g4$qqb$2@news-reader11.wanadoo.fr>

Hello,

I have a little doubt concerning variables.
When one declare x = 3
does x contains the value 3 or
does x acts like pointer in C ?

Regards

Salvatore





From ma at Bell.com  Tue Apr 29 12:00:52 2003
From: ma at Bell.com (xam)
Date: Tue, 29 Apr 2003 16:00:52 GMT
Subject: stupid q-n (wxPy)
Message-ID: <Ukxra.33100$7M5.3263531@twister.nyc.rr.com>

I am trying to run some wxPy samples and I am getting an error here:
if __name__ == '__main__':
    import sys,os
    import run
    run.main(['', os.path.basename(sys.argv[0])])

- "import run" fails. I know this is something really stupid, but I could
not find the answer on the net.
thanks.





From martin at v.loewis.de  Mon Apr 21 12:25:17 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Mon, 21 Apr 2003 18:25:17 +0200
Subject: looping and deleting
In-Reply-To: <3cf72d52.0304210801.2128eca2@posting.google.com>
References: <3cf72d52.0304210801.2128eca2@posting.google.com>
Message-ID: <b81617$2km$01$2@news.t-online.com>

scn wrote:

> what is a good method for looping over a file line-by-line and
> deleting the first 3 characters?

What methods did you try already?

Regards,
Martin




From peter at engcorp.com  Wed Apr  2 10:20:23 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 10:20:23 -0500
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com> <slrnb8lbo0.ccp.rmunn@localhost.localdomain>
Message-ID: <3E8AFFB7.9516A49F@engcorp.com>

Robin Munn wrote:
> 
> sdieselil <sdieselil at yahoo.com> wrote:
> > Hi
> >
> > How can I implement double-linked lists and other complex tree-like
> > structures in Python?
> 
> Do you *need* doubly-linked lists? Is there something that Python's
> builtin list type can't do for you? Consider whether it's worth taking
> the time to code up a data structure, with all the attendant edge cases,
> when you've already got such nice useful builtin types...

I can think of at least one example where one would *need* doubly-linked
lists... a homework assignment.  ;-)

-Peter



From aleax at aleax.it  Sat Apr 12 12:58:23 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 16:58:23 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net>
Message-ID: <PAXla.4934$LB6.139412@news1.tin.it>

Mads Orbesen Troest wrote:

> On Fri, 11 Apr 2003 10:48:49 +0200, Harald Massa wrote:
> 
>> Yes, I love Python, and it is the most elegant programming language I
>> ever met.
> 
> I agree it's mostly elegant, but the OO features are clumsy in my oppinion
> (coming as a professional developer from other languages). Not as bad as

Interesting, because I also came to Python as a professional developer
from other languages (C++ above all) and the utter *ELEGANCE* of Python's
OO features was one of the things that hooked me to it irretrievably.

> Perl's bless-a-dictionary-into-being-a-class retro-fitting hack, of
> course, but really; all the self me here and self me there is - to me at
> last - exceedingly clumsy and a constant source of frustration.

Let's put it this way: in C++ you have to use this->foo SOME of the
time (to avoid running afoul of extremely subtle and deuced bugs
in some templates) but it's quite hard to predict exactly when you'll
need to do it.  With Python, you ALWAYS use self.foo (one character
less than C++'s this->foo!-) and thus your code is more regular and
uniform, there is no risk of ugly conventions such as naming member
variables m_this and m_that (popular in several C++ shops), no bugs
due to accidental local "hiding" of a member variable by a local one
(I've taught and consulted on C++ long enough to have seen THAT
happen many, MANY times -- and not to idiots, either... rather, to
very good professional programmers!) -- a HUGE amount of conceptual
AND pragmatical simplification of your life.

If that's the worst aspect of Python, then how must the rest be?-)

You can gain most (not quite all) of these advantages in C++ by 
using the this->foo way of accessing instance variables -- this 
also makes your C++ and Python code more similar, and helps you get
into the good habit.  And if any C++ bigot ever challenges you,
explain that this ensures optimal behavior of your code should it
ever be turned a template under ANY circumstances (it's true...!-).


Alex




From martin at v.loewis.de  Sat Apr 26 20:14:54 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 27 Apr 2003 02:14:54 +0200
Subject: zipfile with umlaut in filename
References: <3ea9a4bf$1_1@news.vo.lu> <m3ist1ruee.fsf@telia.com>
Message-ID: <m3k7dg23up.fsf@mira.informatik.hu-berlin.de>

Andrew Markebo <andrew.markebo at telia.com> writes:

> A quick note maybe is, which format is the string sent to the
> zip-functions? I mean, does the zipfile classes handle wide
> (unicode/utf-8) strings and so on, these can't be used in the
> zip-file.. yet..?

I think currently zipfile.py uses the encoding of the file name as the
system reports it, whereas Winzip uses CP437 or CP850. Neither is a
reasonable solution to the problem.

Regards,
Martin




From tyler at cg1.org  Mon Apr 14 18:30:31 2003
From: tyler at cg1.org (Tyler Eaves)
Date: Mon, 14 Apr 2003 22:30:31 GMT
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org>
Message-ID: <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net>

Nicola Mingotti wrote:

> import time
> t = time.time()
> # t is the number you asked for
> 
> If you want to know more about this look at
> the "time" module .
> 
> 
> bye.

The problem is that you can only represent dates from 1970 - 2030-someting
that way.




From staschuk at telusplanet.net  Sat Apr 26 12:36:37 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 26 Apr 2003 10:36:37 -0600
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
In-Reply-To: <4o3lav4k04m9g38lm0k6e1mlau6lgma81m@4ax.com>; from intentionally@blank.co.uk on Sat, Apr 26, 2003 at 03:13:04PM +0100
References: <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <slrnbajav2.e5n.ramen@lackingtalent.com> <pan.2003.04.25.17.49.27.407432@jerf.org> <4o3lav4k04m9g38lm0k6e1mlau6lgma81m@4ax.com>
Message-ID: <20030426103637.D29986@tibia.amotlpaa.bogus>

Quoth Stephen Horne:
  [...]
> Yeah, but my brain object must have been written in C++
> [...] it's so unreliable I
> get hours of downtime almost every single day.

Yes, most versions of the brain object require several hours of
downtime each day for routine maintenance tasks, notably to move
data from cache to permanent storage.

Unfortunately the storage format is proprietary and the details
are very well-protected.  (This is one of the reasons that
interoperability between brains and other systems is so poor.)
Unfortunately, the vendor has what seems to be an unassailable
monopoly position in this market; open source alternatives are
extremely primitive, and the effort to reverse engineer the brain
-- which has already been underway for decades -- is not expected
to be complete any time soon.

(Wishing for a SQL interface to my memories... it's *my*
intellectual property, damn it!)

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From mcfletch at rogers.com  Tue Apr 15 10:46:25 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 15 Apr 2003 10:46:25 -0400
Subject: Case-insensitive string comparison
In-Reply-To: <3e9c1413$1@news.012.net.il>
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1> <3e9c1413$1@news.012.net.il>
Message-ID: <3E9C1B41.8030602@rogers.com>

Avner Ben wrote:
...

>    I am searching for one or more members of a list of names (that may get
>near 1000 pieces in a normal application and may average some 40 characters)
>in a string (that may average 100 characters). For each name found in the
>string, I do something to the string, using the original name (in its proper
>case). At present, I am looking for the exact name as stated, with agreeable
>performance. However, this makes me miss many cases where the author of the
>string did not care for the casing (which happens frequently). I have little
>problem with either lowering or uppering the entire list once, because it is
>not expected to change often and when that happens once in a while, a
>subscription mechanism can take care of that. However, this requires holding
>two lists (I still need the original names as are) and lowering or uppering
>each string for comparison (and keep its original version for
>restructuring).
>
>    Obviously, the efficient algorithm here would be to skip the standard
>comparisons altogether and write an algorithm that advances on the string
>character by character and - for each - attempts to parse using the list.
>
...

Sure, but as you discovered, it's not the best way in Python.  I'd guess 
that if you analysed it it's a pretty bad algorithm for C too.  Unless 
your comparison function is such that it needs to adapt to *both* words 
in a pair, it's far more time-efficient to do the translation of each 
lexicon word once, cache the results (trading space) and use them.  
Upper/lower casing is a pretty trivial task, so that might be swamped by 
the setup/teardown cost, but the more involved your fuzziness gets the 
more the iterative approach becomes a liability without some pre-setup.

lexicon = { # translated:[ matching, words]
}

def lookup( word ):
    """Return the lexicon words which match word"""
    key = translate( word )
    return lexicon.get( key )
def addWord( word ):
    """Add a new word to the lexicon"""
    key = translate( word )
    lexicon.setdefault( key, []).append( word )
    return key

That approach has the neat benefit of allowing phonetic compression (or 
the like) at some later stage, which lets you get even more "fuzzy" in 
your matching.  Using the bisect module with (translated,original) pairs 
is a useful modification if you eventually need ordering of the 
word-list (or scanning of an entire range of words (useful in 
larger-distance fuzzy searching)).

Anyway, enjoy yourself,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From gk_2345 at yahoo.com  Fri Apr 25 22:13:35 2003
From: gk_2345 at yahoo.com (George Kinney)
Date: Fri, 25 Apr 2003 22:13:35 -0400
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.29.48.204375@soul.net>
Message-ID: <RZlqa.4182$lf3.712485@cletus.bright.net>

"lost" <lost at soul.net> wrote in message
news:pan.2003.04.25.23.29.48.204375 at soul.net...
> On Fri, 25 Apr 2003 22:51:12 +0000, Alex Martelli wrote:
> Hi Alex,
> sorry i wasn't a bit more specific in what i needed, but what you posted
> was exactly what i was looking for!  Thanks a lot.
>

FWIW, theres also:

from glob import glob
file_list = glob.glob(SomePath)
outfile = open(OutputFilename, 'w')

for file in file_list :
    print file
    outfile.write(file + '\n')

outfile.close()







From cartermark46 at ukmail.com  Sat Apr  5 14:57:15 2003
From: cartermark46 at ukmail.com (Mark Carter)
Date: 5 Apr 2003 11:57:15 -0800
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com> <b6l0eq$tf0$1@main.gmane.org> <mailman.1049539029.20263.python-list@python.org> <roy-C473C1.08401305042003@reader2.panix.com>
Message-ID: <d3c9c04.0304051157.2812dd3f@posting.google.com>

> There was a recent item on slashdot lamenting how much effort is wasted 
> in the free software community on reinventing the wheel, 

It would also help if we all wrote in the same programming language -
then people would only need to write libraries once. I intimated as
much in a casual post I made here:

http://groups.google.com/groups?dq=&hl=en&lr=&ie=UTF-8&safe=off&threadm=pan.2003.04.05.12.33.23.954023%40knm.org.pl&prev=/groups%3Fhl%3Den%26lr%3D%26ie%3DUTF-8%26safe%3Doff%26group%3Dcomp.lang.misc

According to a respondent, we should all be writing in Lisp. 

Who knows, he might even be right.



From adalke at mindspring.com  Tue Apr  1 22:34:44 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Tue, 1 Apr 2003 20:34:44 -0700
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift> <3E890322.4B003C9C@engcorp.com> <slrnb8j5og.o6a.chasm@galileo.rift> <b6d8m2$47nsc$1@ID-169208.news.dfncis.de>
Message-ID: <b6dlhh$uif$1@slb0.atl.mindspring.net>

Greg Ewing:
> This kind of thing has been observed before -- I believe
> someone noticed that using get() on a dict was slower than
> indexing and catching the exception.

Barry Warsaw, IPC...6?  The one in San Jose, I think.

Yep, here it is.
  http://barry.warsaw.us/papers/ecshort.pdf

It shows that try/except is faster than setdefault when
the number of exceptions is low.

                    Andrew
                    dalke at dalkescientific.com





From vlokk at hot.ee  Tue Apr 15 05:30:31 2003
From: vlokk at hot.ee (Vahur Lokk)
Date: Tue, 15 Apr 2003 12:30:31 +0300
Subject: Catching Save - newbie
Message-ID: <3e9bd039_2@news.estpak.ee>

Hi!
What I want to do is write a small daemon that can catch Save File type of 
events in the system and manipulate files being saved.
Having mostly sorted out the rest of the program I cannot think of a way to 
catch these saves.
I must warn that I am complete newb and doing my first steps in Python (and 
programming overall).

Wahur



From martin at v.loewis.de  Wed Apr 23 15:22:05 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 23 Apr 2003 21:22:05 +0200
Subject: how to link python against static sys libs?
References: <pan.2003.04.23.12.53.47.253065@nospam-freenet.org.nz>
Message-ID: <m3znmhf28y.fsf@mira.informatik.hu-berlin.de>

"David McNab" <david at nospam-freenet.org.nz> writes:

> I'm having a bit of a nightmare trying to build python to link the sys
> libs (eg glibc) statically.

I haven't check recently, but I think static libraries are not really
supported on Linux anymore. The C libraries relies on dynamic loading
for certain functions (such has netdb), and there is no way to build a
statically-linked executable that uses these functions.

Regards,
Martin




From LogiplexSoftware at earthlink.net  Tue Apr  8 13:10:06 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 08 Apr 2003 10:10:06 -0700
Subject: How to protect python code ?
In-Reply-To: <1abbm-ccl.ln1@news.lairds.org>
References: <3E8BE455.4000904@novagrid.com>
	 <mailman.1049389032.4904.python-list@python.org>
	 <u8bka.651$001.591065@monger.newsread.com>
	 <mailman.1049722991.18682.python-list@python.org>
	 <3E921A4C.1020908@hotmail.com>
	 <a8b7f07a.0304080539.465049c8@posting.google.com>
	 <1abbm-ccl.ln1@news.lairds.org>
Message-ID: <1049821806.8219.15.camel@software1.logiplex.internal>

On Tue, 2003-04-08 at 10:23, Kyler Laird wrote:

> Someday Microsoft will save us from this sorry state with
> Palladium.  Then we can all relax.

Although having someone else choose my software and content for me *is*
convenient, I just can't get comfortable wearing those
electroencephalograph leads all the time.

Now if only McDonald's would get in the game, I wouldn't have to decide
what to eat, either...


-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308

"What! They've got chairs with wheels! And here I am using my legs like 
a sucker!"  - Homer Simpson








From peter at engcorp.com  Wed Apr 16 08:46:23 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 08:46:23 -0400
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net> <420ced91.0304152121.2af84a33@posting.google.com>
Message-ID: <3E9D509F.3FC4073@engcorp.com>

sdhyok wrote:
> 
> Tyler Eaves <tyler at cg1.org> wrote in message news:<bEGma.462$8c4.82 at tornadotest1.news.pas.earthlink.net>...
> > Nicola Mingotti wrote:
> >
> > > import time
> > > t = time.time()
> > > # t is the number you asked for
> > >
> > > If you want to know more about this look at
> > > the "time" module .
> > >
> > >
> > > bye.
> >
> > The problem is that you can only represent dates from 1970 - 2030-someting
> > that way.
> 
> So, this is not the solution I am seeking.
> Better module to handle wider range of time?

How wide a range?  It is much easier to answer *specific*
questions, than very broad ones as you are asking... narrow
the scope a little and help us out.

-Peter



From martin at v.loewis.de  Tue Apr 29 17:42:46 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Tue, 29 Apr 2003 23:42:46 +0200
Subject: Some oddities in 2.3b1, esp. with Tkinter
In-Reply-To: <b8mpp1$25hm$1@nntp6.u.washington.edu>
References: <b8mpp1$25hm$1@nntp6.u.washington.edu>
Message-ID: <3EAEF1D6.30900@v.loewis.de>

Russell E. Owen wrote:
> 1) A filter to remove fonts with non-ascii names (used to filter out 
> stuff that I could not display in a pop-up menu) broke.
>             try:
>                 unicode(fontName, 'ascii', 'strict')
>                 return True
>             except UnicodeError:
>                 return False
> 
> Python 2.2 throws UnicodeError. Python 2.3b1 throws TypeError. The fix 
> is trivial (e.g. "except (UnicodeError, TypeError)". The filter is 
> probably not needed anymore anyway, since Tk 8.4.2 can apparently 
> display menu items with non-ascii characters.

I don't understand. Why do you ever need to convert the *font name*
to a Unicode string, or why would Python ever fail to do so? Font
names are ASCII only, all the time, so you can't possibly ever get
an exception here.

> 2) Tk file events are no longer supported:
> 
> RuntimeError: _tkinter.createfilehandler not supported for threaded Tcl
> 
> This is painful since it's hard to see how to write a proper networking 
> GUI with Tkinter without file events for reading data. Perhaps I can get 
> them back again by disabling the new threaded tcl support.

Are you using Debian by any chance? Otherwise, you should not see
a threaded Tcl.

If your Tcl installation is threaded, you should use 
tkapp.createfilehandler instead.

> 3) Tk returning things as objects is nice, but I had this break at least 
> one thing of mine and it will break others (see 4 below):
> 
> I was assuming that the values in the dictionary returned by 
> tkFont.Font.config() were strings. Fortunately this was easy to fix. A 
> bit disconcerting -- I wonder how many any other little pitfalls like 
> this still buried in my code -- but well worth the cost of the change.

What is the nature of that change?

> Example:
> v = StringVar()
> v is a <Tkinter.StringVar instance at 0x2c98a0>
> l = Label(textvariable=v)
> x = l["textvariable"]
> x is reported as a <namespaceVarName object at 0x0168b048>
> but is a _tkinter.Tcl_Obj based on the following traceback:
> root.getvar(x)
> Traceback...
> TypeError: getvar() argument 1 must be string, not _tkinter.Tcl_Obj

Please submit a bug report. This is easy to fix (just invoke
str() of the Tcl_Obj)

> 5) Note that Tkinter's handling of font objects does not seem to have 
> changed. I was hoping it would start returning font objects instead of 
> strings, or offer some way to convert the string to a font object, but 
> apparently not:

Does Tk have the notion of "font objects"?

Regards,
Martin




From peter at engcorp.com  Wed Apr  2 17:41:26 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 17:41:26 -0500
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org> <3e8afeb9.1125578@news-server> <3E8B0060.D9C449A3@engcorp.com> <3e8b1b0f.8380222@news-server> <slrnb8lq1h.d7q.rmunn@localhost.localdomain> <3e8b5589.10546880@news-server>
Message-ID: <3E8B6716.11FDDC7F@engcorp.com>

BBFoto wrote:
> 
> I copied your line and pasted it.
> 
> File "<stdin>", line 1
>     C:\Python22\Python.exe "c:/program
> files/pysol/pysol-4.80/data/PySol_22.pyc"
> 
>      ^
> SyntaxError: invalid syntax

This is a message from the Python interpreter.  You should
not be *in* the interactive interpreter when you type that.
It is an MS-DOS command, to execute the bytecode in the .pyc
file using the python.exe executable.

May I suggest you read some parts of the FAQ at www.python.org, 
and also go through the Python tutorial available there, before
trying to do much more?  You are missing a substantial amount of
background knowledge necessary to understand the advice you are
being given, both in Python and Windows/DOS.

-Peter



From martin at v.loewis.de  Sun Apr  6 13:43:52 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 06 Apr 2003 19:43:52 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain>
Message-ID: <m3y92nmsh3.fsf@mira.informatik.hu-berlin.de>

pobrien at orbtech.com (Patrick K. O'Brien) writes:

> No, I really want something very similar to Emacs, not just
> "comparable in power."  The reason is that there is a ton of elisp
> code out there, including most of Emacs itself, that could be
> translated to work with PyAlaMode.  

In that case, you really should target a completely automated
translation process. Nobody will port ten thousands of lines of elisp
to Python manually.

Regards,
Martin



From aahz at pythoncraft.com  Thu Apr 24 10:14:09 2003
From: aahz at pythoncraft.com (Aahz)
Date: 24 Apr 2003 10:14:09 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <X8Mpa.30197$y3.2398527@news010.worldonline.dk>
Message-ID: <b88rfh$t5l$1@panix1.panix.com>

In article <X8Mpa.30197$y3.2398527 at news010.worldonline.dk>,
Max M  <maxm at mxm.dk> wrote:
>
>While I do understand the principles of tdd, testing first and all
>that, I have a hard seing why it should be an advantage to do it in the
>early stages of a project.
>
>Normaly if a am entering a new domain, where I don't have a clear
>picture of the requirements, I try out different approaches. Each
>approach using a different combination of ideology, objects and
>functions.
>
>This is all to get a feel of the best approach in this particular
>domain. Some times I try out a *lot* of approaches before I end up with
>a particular elegant design.
>
>At which point I am ready to "solve the problem". It is at this point I
>can see the point of unit testing.
>
>In tdd do you really use testing for the first "creative" fase also?

That's another point I was getting at, yes.  But let me pretend I
actually know something about TDD (test-driven development, for
Googlers), and I'll try to answer your question:

When you're exploring solutions, how do you know that your solutions are
correct?  If your exploration phase keeps getting checked against your
evolving design document (as represented by unit tests), you know that
the solution you create actually works!

One thing I was particularly smug about was figuring out how to
implement my query "language" without a counter for tracking how many
search terms matched.  I'm not sure whether I would have figured that
out slower or faster with TDD.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From zdys at zdys.de  Tue Apr 15 03:46:04 2003
From: zdys at zdys.de (luigi monaco)
Date: Tue, 15 Apr 2003 09:46:04 +0200
Subject: syntax highlighting file python 2.1
Message-ID: <b7gd6g$200$04$1@news.t-online.com>

Hi clp,

how would I get all the relevant data I need to build a syntax coloring
definition file for python 2.1?

I am using my own format in a self written src2tex converter. 
(java src: http://www.beanstalk.de/z/download/Src2TeX4Lyx.jar)

Basically I require a full list of methods, modules or a way how to
extract this from a python release. Keywords etc. should
suffice from e.g.: /usr/share/vim/vim61/syntax/python.vim. 

Thanks in advance

Luigi



From johnochiltree at blueyonder.co.uk  Sun Apr 27 04:48:05 2003
From: johnochiltree at blueyonder.co.uk (John Ochiltree)
Date: Sun, 27 Apr 2003 09:48:05 +0100
Subject: Problems making Python-2.3b1
References: <HuLqa.2812$BL6.1867@news-binary.blueyonder.co.uk> <m3he8ktkzy.fsf@mira.informatik.hu-berlin.de>
Message-ID: <kPMqa.3577$BL6.2708@news-binary.blueyonder.co.uk>

Martin v. L?wis wrote:
<snip>
> Please do report bugs at sf.net/projects/python in the future; a bug
> report in comp.lang.python will get lost.
> 

Thanks Martin this worked fine. I posted it as a bug along with your fix on
sf.net/projects/python

Regards John Ochiltree



From rmunn at pobox.com  Wed Apr 23 11:55:16 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 23 Apr 2003 15:55:16 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com> <e9570f37.0304220625.11312687@posting.google.com>
Message-ID: <slrnbacp3b.22r.rmunn@localhost.localdomain>

Graham Fawcett <graham__fawcett at hotmail.com> wrote:
> I guess the moral of the story is, "contribute some time to
> PyPI/Distutils development."  Or, befriend a PyPI/Distutils developer
> and send him food and get-out-of-work-free tickets. ;-)

I've heard of those. Those are those little pieces of paper that say
"Pay to the order of ________", right? :-)

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From cejlai at lafn.org  Wed Apr 16 09:10:03 2003
From: cejlai at lafn.org (Jason Lai)
Date: Wed, 16 Apr 2003 06:10:03 -0700
Subject: Subclassing a C type
Message-ID: <b7jkni$1v7d$1@zook.lafn.org>

Hello,

I wrote a Python type in C, Actor, and created a subclass Tester 
(new-style classes). Actor has a getattro function that gets information 
from elsewhere in the C program.

Apparently, the Tester class inherits tp_getattro from Actor. What I'd 
like it to do is have it use the regular lookup so that it checks its 
own dict before deferring to its parent classes. Without having to 
define a __getattr__ function on Tester. Maybe I could change the 
__getattr__ on every tp_new, but that seems a bit messy.

I thought about having Actor_getattro look up the value in 
Tester-instances' dictionaries, but I don't think it would work if there 
were more subclasses between Actor and Tester in the hierarchy.

And I do need to intercept the getattro calls, so I can't just use the 
standard object one and define tp_members/tp_methods.

Am I doing something wrong?

Thanks in advance for any advice you can offer,

  - Jason Lai




From nirina at mail.blueline.mg  Tue Apr  1 15:05:10 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Wed, 2 Apr 2003 00:05:10 +0400
Subject: need help with game
References: <3e88b3d3.30570106@news-server>
Message-ID: <024501c2f88a$52419440$f3a2383e@raseliar>

"BBFoto" <bbfoto2002 at netscape.net> wrote:

> I run the Python executable and type in "PySol" at the prompt on the
> black screen and get "name not defined."

What happens if you type:
import PySol

There is a colorful screen at
C:\Python23\Tools\idle\idle.py

Hope this help

nirinA





From mike at bindkey.com  Tue Apr  8 13:52:22 2003
From: mike at bindkey.com (Mike Rovner)
Date: Tue, 8 Apr 2003 10:52:22 -0700
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org> <XSXja.346441$3D1.186570@sccrnsc01> <mailman.1049741353.6320.python-list@python.org> <tntka.364035$F1.57656@sccrnsc04>
Message-ID: <b6v27l$9g0$1@main.gmane.org>

"Bryan" <belred1 at yahoo.com> wrote in message
news:tntka.364035$F1.57656 at sccrnsc04...
> how is pyrex at handling a c varargs function?

Never tried myself, but ToDo.txt says:
'''
DONE - Varargs C func defs and calls.

Varargs argument traversal.
'''

Mike







From tismer at tismer.com  Sun Apr 20 12:01:03 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sun, 20 Apr 2003 18:01:03 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <b7ufg4$o1d$1@news.hccnet.nl>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl>
Message-ID: <3EA2C43F.2080802@tismer.com>

Anton Vredegoor wrote:
> "A.M. Kuchling" <amk at amk.ca> wrote:
> 
> 
>>http://mail.python.org/pipermail/python-dev/1999-July/000467.html
> 
> 
> Thanks. It seems Tim is making a heroic effort to explain the new in
> terms of the old. This necessarily must fail, but what other option is
> there instead of rebuilding all terminology out of gotos?

Quoting Bjorn Pettersen about continuations:

"""
They also seem to have the property that before you understand them, no
explanation will improve your understanding; and after you understand
them you lose the ability to explain it <wink>.
"""

happy easter -- chris
-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From dave at pythonapocrypha.com  Tue Apr  1 17:12:40 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Tue, 1 Apr 2003 14:12:40 -0800 (PST)
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <20030401153111.B936@localhost.localdomain>
References: <x7y92uhzpw.fsf@guru.mired.org> <Pine.LNX.4.51.0304011257400.15952@ns0.tbye.com>
 <20030401153111.B936@localhost.localdomain>
Message-ID: <Pine.LNX.4.51.0304011410550.16080@ns0.tbye.com>

On Tue, 1 Apr 2003, Jack Diederich wrote:

> On Tue, Apr 01, 2003 at 01:01:14PM -0800, Dave Brueck wrote:
> > On Tue, 1 Apr 2003, Mike Meyer wrote:
> >
> > > PEP: 313
> > > Title: Adding Roman Numeral Literals to Python
> >
> > Before we submit this to the BDFL, I propose that we wait a short while to
> > give the community an opportunity to debate its merits, followed by a
> > vote, just so Guido can get idea of where the community stands. Further, I
> > propose that Aahz be the official vote counter. Does anybody have any
> > preferences as to the voting method we should use?
>
> Yes, I propose we reuse an idea from the ternary vote.
> In order to demostrate understanding of the proposal your preferences should
> be listed in order, with the enumeration done in valid roman numerals.

Brilliant! +I for PEP CCCXIII!

-Dave




From gansevle at ewi.utwente.nl  Wed Apr  9 03:01:53 2003
From: gansevle at ewi.utwente.nl (Fred Gansevles)
Date: Wed, 09 Apr 2003 09:01:53 +0200
Subject: Wrapper for _winreg
References: <b6slau$c0l$1@netlx020.civ.utwente.nl>
Message-ID: <b70ggj$ecq$2@netlx020.civ.utwente.nl>

In my previous post the module didn't contain a license
(some people obviously care about this, thanks Bjorn) so I added
a licens to 'reg.py'

The module is still available at
http://wwwhome.cs.utwente.nl/~gansevle/python/reg.py

If anyone finds this module usefull, please drop me a line at
gansevle at ewi.utwente.nl




From nick_bower at yahoo.com.au  Wed Apr  2 08:50:44 2003
From: nick_bower at yahoo.com.au (nick)
Date: 2 Apr 2003 05:50:44 -0800
Subject: Socket timeouts for HTTP connections (httplib)?
Message-ID: <9df3d134.0304020550.7e32bf6b@posting.google.com>

I'm using httplib in a threaded HTTP load imparting test suite for our
portal.  A problem is that HTTP GETs often hang, blocking my
requesting threads.  I was highly disapointed to find, after writing
my test suite, that no high-level HTTP timeout functionality exists in
Python.

I have neither an expert understanding of socket programming,
threading or asyncronous IO, yet can get by using threading.Thread and
httplib.  I am therefore looking for a timeout at a fairly high level
(ie the httplib api).

There were rumblings about Guido implementing this at a socket level
in June 2002, but I've not been able to find anything since. 
Presumably this has not yet risen up into httplib either because the
python.org documentation doesn't mention it.

Has/is someone involved with this?  timeoutsocket simply doesn't work
for me and throws continuous "connection reset by peer" exceptions. 
So I'm completely stuck for options.

Thanks, Nick



From ponderor at lycos.com  Wed Apr  2 16:48:35 2003
From: ponderor at lycos.com (Dean Goodmanson)
Date: 2 Apr 2003 13:48:35 -0800
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com> <slrnb8lbo0.ccp.rmunn@localhost.localdomain> <3E8AFFB7.9516A49F@engcorp.com>
Message-ID: <e81be8b2.0304021348.c0f4c9a@posting.google.com>

> 
> I can think of at least one example where one would *need* doubly-linked
> lists... a homework assignment.  ;-)
>
> -Peter

And what better language to experiment, analyze and learn the facets
of Computer Science than with Python.

For even more help and questions, stop by the Python tutor list.
http://mail.python.org/mailman/listinfo/tutor

For discussion educating with and for Python:
http://mail.python.org/mailman/listinfo/edu-sig

Welcome to the Python list(s), where homework assignments are
questions too.

-- 
-Dean

When sending someone to Google, remember to support page ranking and
send `em with one of them URL's that are apparrantly so easy to
generate.



From antonmuhin at sendmail.ru  Sat Apr 19 14:00:09 2003
From: antonmuhin at sendmail.ru (Anton Muhin)
Date: Sat, 19 Apr 2003 22:00:09 +0400
Subject: Testing methods
In-Reply-To: <9d8ac130.0304190938.755d4b4c@posting.google.com>
References: <9d8ac130.0304190938.755d4b4c@posting.google.com>
Message-ID: <b7s2r9$11ka$1@news.peterlink.ru>

Ovid wrote:
> Hello,
> 
> Relatively new to Python and I thought I would teach myself the
> language by building a test harness.  After I'm done, I will probably
> be throwing it away in favor of Pyunit, but it seemed like a good way
> to learn some of the fiddly bits.
> 
> Fiddly bits such as "can this object do that?"  In my test code, I
> want to test whether or not a given class *or* instance of that class
> can execute a particular method (i.e., if it has or inherits the
> method).  For example, one way (in Perl), to test if object "$foo" can
> execute an arbitrary method would be:
> 
>   if ( $foo->can($some_method) )
> 
> For the life of me, I can't figure out the equivalent in Python.  I
> suppose I could do:
> 
>   if some_method in Foo.__dict__:
> 
> But that doesn't tell me whether or not "some_method" is a method. 
> Further, if I have an instance of a class, that doesn't seem to work. 
> I've been searching through the docs, but I can't quite figure this
> out.
> 
> I'm a clueless newbie here, so please be kind :)
> 
> Cheers,
> Ovid

There are several ways to do it.

1. More Pythonic IMHO:

class Foo:
     def bar(self):
         print "Foo.bar"

foo = Foo()

try:
     foo.absent_method()
except (AttributeError, TypeError), e:
     print "Failed. Message @%s@" % e, "Type: %s" % e.__class__.__name__
else:
     print "OK"

2. Look at hasattr and callable functions

foo = Foo()
foo.xxx = 239

print hasattr(foo, "absent_method") # false
print hasattr(foo, "xxx") # true
print callable(foo.xxx) # false
print callable(foo.bar) # true

Sure, there are circumstances, then one of methods is better.

BTW, look for a thread about typing in Python---you might find 
interesting information there.

hth,
Anton.




From aleax at aleax.it  Fri Apr 18 17:34:24 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 21:34:24 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
Message-ID: <Ab_na.34355$LB6.777939@news1.tin.it>

Guy Middleton wrote:

> What is the most Pythonic way to sum the n-th element of a list of lists
> (or of tuples)?

_most Pythonic_ is probably:

total = 0
for sublist in listoflists:
    total = total + sublist[n]

I think that's what Guido would do if asked -- it's simple, it's entirely
obvious, it's totally maintainable and clear.

Saving two lines and some bananoseconds (if that) by using reduce and
operator.add may be more to the taste of many regular readers of this
newsgroup, but I wouldn't characterize it as "most Pythonic"!-)


Alex




From abcdebl2nonspammy at verizon.net  Fri Apr 11 16:25:28 2003
From: abcdebl2nonspammy at verizon.net (David Lees)
Date: Fri, 11 Apr 2003 20:25:28 GMT
Subject: Loss of precision / data corruption with Tkinter.DoubleVar ?!
In-Reply-To: <f70e3538.0304111007.b0a72c3@posting.google.com>
References: <f70e3538.0304111007.b0a72c3@posting.google.com>
Message-ID: <YwFla.1596$U34.457@nwrddc02.gnilink.net>

Lonnie Princehouse wrote:
> Greetings, 
> 
> I stumbled across this curiosity yesterday:
> 
> 
>>>>x = 0.33333333333333098
>>>>x
> 
> 0.33333333333333098     
> 
>>>>import Tkinter
>>>>y = Tkinter.DoubleVar(Tkinter.Tk())
>>>>y.set(x)
>>>>y.get()
> 
> 0.33333333333300003  
> 
> What the....?
> 
> Sure, it's only on the order of 10e-13 delta, 
> but why is there any difference at all??
> 
> 
> I hit this because at some point I was trying to do:
> 
> y.set(x)
> .
> .
> .
> if y.get() == x:
>     ...
> 
> and of course it wasn't working properly!
> 
> I have tried this on Python 2.2 for Win32 and Linux/x86
> 
> 
> 
> 
> -Lonnie Princehouse

See:  http://www.python.org/doc/current/tut/node14.html
I do not know what goes on inside Tkinter when you do a 'set', but it is 
bad practice to do the kind of exact comparison in your 'if' statement 
on floating point numbers.

David Lees





From skvamme at as4-2-8.an.g.bonet.se  Wed Apr 30 09:18:03 2003
From: skvamme at as4-2-8.an.g.bonet.se (Sten Kvamme)
Date: Wed, 30 Apr 2003 15:18:03 +0200
Subject: Self modifying systems question
In-Reply-To: <33803989.0304300426.402cd1ef@posting.google.com>
Message-ID: <2CA35DF0-7B0E-11D7-93CD-003065DC4D52@as4-2-8.an.g.bonet.se>

Hello,
Thanks to all of you who answered my question.
Is it also possible to save the changes made to an instance in a 
persistant way?

Thanks again,
Sten

On Wednesday, April 30, 2003, at 02:26 PM, Miki Tebeka wrote:

> Hello Sten,
>
>> Self modifying code is a well known technology. Has anyone done this 
>> in
>> Python? What I'm seeking for is examples of software that adds to
>> itself (methods and variables) in runtime. I'd like an instantiated
>> object to
>> be able to add a method to itself.
>>>> class A:
> 	pass
>
>>>> a = A()
>>>> a.f = lambda x: x + 1
>>>> a.f(100)
> 101
>
> HTH.
> Miki
> -- 
> http://mail.python.org/mailman/listinfo/python-list





From oblivious at web.de  Tue Apr  8 04:48:04 2003
From: oblivious at web.de (Dominic)
Date: Tue, 08 Apr 2003 10:48:04 +0200
Subject: Comms Protocol design dilemma
In-Reply-To: <mailman.1049738831.30630.python-list@python.org>
References: <mailman.1049738831.30630.python-list@python.org>
Message-ID: <3E928CC4.6090403@web.de>

> number generation involved, I can pretty much rule out asynchronous
> sockets. Is there any better strategy to tackle this problem?

Well asynchronous sockets are fine. I've written a load-test
client in Python which handles 1000 tcp connections simultaneously.
(1000 sockets is the per process limit for non root users)
I've used the python asyncore module. If you use asynchronous
sockets you need to keep your session state explicitly since
you do not have threads. You could use a finite-state-machine
or Python's generators. (See my previous mail on comp.lang.python,
Programming Erlang style in Python. You should certainly
check out the Erlang programming language homepage (for the concepts)
from Erricson since that language has been designed for servers which 
you intend to write.)

Ciao,
  Dominic




From SBrunning at trisystems.co.uk  Thu Apr 17 04:29:24 2003
From: SBrunning at trisystems.co.uk (Simon Brunning)
Date: Thu, 17 Apr 2003 09:29:24 +0100
Subject: Teaching : Python, Scheme, Java...
Message-ID: <31575A892FF6D1118F5800600846864D0120082A@intrepid>

> From:	Erik Max Francis [SMTP:max at alcyone.com]
> Jean-Paul Roy wrote:
> 
> > - how do you ask if a list is empty ? Is the following ok ?
> >       if L == [] :
> 
> Even simpler:
> 
> 	if L: ...
 
Surely:

if not L:

;-)

Cheers,
Simon Brunning
TriSystems Ltd.
sbrunning at trisystems.co.uk
--LongSig




-----------------------------------------------------------------------
The information in this email is confidential and may be legally privileged.
It is intended solely for the addressee. Access to this email by anyone else
is unauthorised. If you are not the intended recipient, any disclosure,
copying, distribution, or any action taken or omitted to be taken in
reliance on it, is prohibited and may be unlawful. TriSystems Ltd. cannot
accept liability for statements made which are clearly the senders own.




From jdhunter at ace.bsd.uchicago.edu  Sun Apr 13 13:33:06 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Sun, 13 Apr 2003 12:33:06 -0500
Subject: dll load failed: cannot find file specified
Message-ID: <m2brzajoa5.fsf@mother.paradise.lost>

I am using python 2.2.2 from the win32 installer from python.org.  I
am trying to import the gtk.gtkgl package, which is the python
extension to gtkglext.  I get the error on the gdkgl import line

  ImportError: DLL load failed.  The system cannot find the file
  specified

Normally, I would think this meant the DLLs from the C lib weren't in
the PATH environment variable, but they are: I can run the C examples
and I can load all the pygtkglext *.pyd files in Dependency Walker
without errors.  I have installed the GTK runtime, those libs are in
my PATH, and pygtk 1.99.16 is running fine.  But something is going
wrong when I try and import pygtkglext into python.

Any ideas what can be causing this message, or suggestions for how to
debug what is going wrong, or other advice?

Thanks,
John Hunter




From Padraig at Linux.ie  Tue Apr 15 05:34:26 2003
From: Padraig at Linux.ie (Padraig at Linux.ie)
Date: Tue, 15 Apr 2003 10:34:26 +0100
Subject: vim configuration help
In-Reply-To: <P4Oma.78974$rd4.2975295@twister.austin.rr.com>
References: <P4Oma.78974$rd4.2975295@twister.austin.rr.com>
Message-ID: <3E9BD222.5040005@Linux.ie>

Randall Smith wrote:
> I'm learning Python and I've been looking for a good IDE to use with it. 
>    I haven't found an IDE that has text editing capabilities close to 
> Vim/Emacs.
> 
> Much of my previous programming is with Emacs/PHP and I love the 
> auto-indenting, tab-indenting, well thought-out key strokes, and the 
> ability to ssh into a server and run Emacs.

When learning vim myself I wrote this:
http://www.pixelbeat.org/vim.tips

Here's my vim settings
http://www.pixelbeat.org/settings/.vimrc

Note you might want to try some of the scintilla
based editors (SciTE/anjuta/...)

P?draig.




From steve.cassidy at mq.edu.au  Tue Apr  1 07:18:25 2003
From: steve.cassidy at mq.edu.au (Steve Cassidy)
Date: Tue, 01 Apr 2003 22:18:25 +1000
Subject: Finding Python help...
References: <3E88310F.6090709@mq.edu.au> <b6b39a$1gd$0@216.39.172.122>
Message-ID: <3E898391.9070205@mq.edu.au>

Bengt Richter wrote:
> On Mon, 31 Mar 2003 22:14:07 +1000, Steve Cassidy <steve.cassidy at mq.edu.au> wrote:
> 
> 
>>One problem I'm having learning (and teaching) Python is finding help on 
>> some parts of the language. I'm wondering if I'm just looking the 
>>wrong way...

Thanks for the responses to this query. I've learned a little more about 
working in the Python world.

One thing that struck me learning Python having been a long time Tcl 
coder was the difference between built in and library functions in terms
of level and location of documentation. I'm a Linux user and I'm used to 
either man pages or web based help. I did find the full text search 
available in the windows ActivePython distributiton the most useful but 
since that's not my main platform it doesn't help day to day.

Cheers,

Steve







From mwh at python.net  Wed Apr 23 07:55:26 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 23 Apr 2003 11:55:26 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <g8n5avgooim13l90bv5t9gu6ilg1oar101@4ax.com> <mailman.1050982240.24284.python-list@python.org> <t6g9av8aeurhfis19opg84ov34rvr69lpk@4ax.com>
Message-ID: <7h3he8pwhqq.fsf@pc150.maths.bris.ac.uk>

Courageous <jkraska at san.rr.com> writes:

> Continuations allow you to do whacky upside down things like make a
> function return, say, TWICE. :)

Which is part of the attraction.  I've done this, sort of: it was a
simulator for a non-deterministic system, and so sometimes you didn't
know which of two things was going to happen.  The "obvious" solution
was to return from that function twice.

The irritating bit was that this was written in C++ and so involved
packing and unpacking state into data structures in a tedious way.
There wasn't so much state that this was too annoying, but scheme
would have made things easier, I suspect.

Cheers,
M.

-- 
  ARTHUR:  But which is probably incapable of drinking the coffee.
                    -- The Hitch-Hikers Guide to the Galaxy, Episode 6



From lists at ghaering.de  Wed Apr 23 18:14:07 2003
From: lists at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Thu, 24 Apr 2003 00:14:07 +0200
Subject: Using the MySQLdb module
In-Reply-To: <3EA6F757.5080903@cis.ohio-state.edu>
References: <3EA6E75E.7070305@cis.ohio-state.edu>
	<mailman.1051131517.32137.python-list@python.org>
	<3EA6F757.5080903@cis.ohio-state.edu>
Message-ID: <3EA7102F.6000201@ghaering.de>

Michael Mayhew wrote:
> Gerhard H?ring wrote:
> 
>> Also see the DB-API 2 specification and the documentation and examples 
>> included in the MySQLdb source distribution.
> 
> Gerhard,
> 
>    Unfortunately, the MySQLdb documentation, both on the web and from 
> the sourceforge location are a little lax for examples on using cursor 
> objects. [...]

Ok, let's repeat for the pedagocical effect:

   """
   ... and the documentation and examples
   included in the MySQLdb *source distribution*.
   """

Let's see what we have here:

gerhard at lilith:/usr/ports/databases/py-MySQLdb/work/MySQL-python-0.9.2 > 
find doc/
doc/
doc/MySQLdb-1.html
doc/MySQLdb-2.html
doc/MySQLdb-3.html
doc/MySQLdb-4.html
doc/MySQLdb-FAQ-1.html
doc/MySQLdb-FAQ-2.html
doc/MySQLdb-FAQ-3.html
doc/MySQLdb-FAQ.html
doc/MySQLdb-FAQ.sgml
doc/MySQLdb.html
doc/MySQLdb.sgml

The documentation isn't very broad, but from what I can see it 
exhaustively explains how to use cursor.execute. IMO it explains 
everything necessary without wasting many words. I like that.

-- Gerhard




From dreed at capital.edu  Sun Apr 27 19:53:32 2003
From: dreed at capital.edu (Dave Reed)
Date: Sun, 27 Apr 2003 19:53:32 -0400
Subject: Textbooks?
In-Reply-To: <a5415304.0304271335.2f569879@posting.google.com>
References: <oproar1sjtx3zh5c@news.mindspring.com> <a5415304.0304271335.2f569879@posting.google.com>
Message-ID: <200304271953.32645.dreed@capital.edu>

On Sunday 27 April 2003 17:35, Ron Stephens wrote:
> Steven Hodgen <steven at twitch.net> wrote in message 
news:<oproar1sjtx3zh5c at news.mindspring.com>...
> > Hello,
> > 
> > I'm currently teaching a high-school introduction to programming 
course.  
> > I'm currently teaching in Java, but am seriously considering 
changing to 
> > Python.  The problem is that I cannot seem to find any good 
textbooks, much 
> > less ones for students with no prior programming experience.  This 
is the 
> > one area in which Java excels. 
> 
> The only actual "Textbook" I am aware of is Deitel and Deitel's Python
> book, which I found pretty good. However, Learning Python by Lutz and
> Ascher would be excellent (and it includes exercises after every
> chapter) as would Core Python by Chung. I have reviewed all these
> books and more than 20 more and you can find the reviews at
> www.awaretek.com/plf.html
> -- 


John Zelle at Wartburg College just finished a traditional college CS1
Python textbook that is very unintimidating for students so it should
also be apropriate for high school. It is not a huge book like
Deitels' books (i.e., it isn't a Python reference, but is an excellent
beginning programming book).

We used a prepublication copy of it this past fall with excellent
results. The students liked the book and found it very
understandable. John has indicated the book should be officially
available this fall. If you're interesting in using it at the college
or high school level this fall, you may want to contact him at:
john.zelle at wartburg.edu

Dave







From claird at lairds.com  Mon Apr  7 12:06:58 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 07 Apr 2003 16:06:58 -0000
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org>
Message-ID: <v938h26mdl9s8f@corp.supernews.com>

In article <mailman.1049722991.18682.python-list at python.org>,
Dave Brueck  <dave at pythonapocrypha.com> wrote:
			.
			.
			.
>adding too much complexity or maintenance costs. Beyond that, the law of
>diminishing returns starts to win because pretty quickly you get into
>guarding against people who are willing to invest large amounts of time
>cracking it or, worse, people who crack for fun.
>
>Even once you have the source code, there's still enormous work involved
>before you can use it to generate revenue (to create your own knock-off
>product). In the case of protecting the code to protect the revenue stream
>(anti-piracy), you have to realize things like many/most pirates would not
>have purchased your program anyway, your selling price may be too high,
>having your customer relationship depend completely on the product itself
>is a weak position and tough to defend, etc.
			.
			.
			.
Gratuitous factual tangent:  'might surprise folks
how often acquirers of software companies essentially
discard source code.  They want customer lists and
relations, and occasionally the engineering staff,
but existing programs?  Those are nearly valueless.

I exaggerate a bit.  There are indeed legitimate roles
for obfuscating technologies.  Managers think they're
far more common than is true, though, and programmers
far, FAR overestimate the objective value of source
code.  It's one of our occupational hazards.

Dave's right.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From imbosol-1050441179 at aerojockey.com  Tue Apr 15 18:44:26 2003
From: imbosol-1050441179 at aerojockey.com (Carl Banks)
Date: Tue, 15 Apr 2003 22:44:26 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com>   <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it> <slrnb9o9ob.493.rmunn@localhost.localdomain>
Message-ID: <eX%ma.18570$Gb1.6139@nwrdny01.gnilink.net>

Robin Munn wrote:
> I suppose I just illustrated how a try ... finally block might be
> considered "broken and fragile". I tried to use it, and missed some
> subtleties. I'm still not convinced that a "with"-like statement is
> necessary in Python, but I think I just went from -1 to -0 on it.

I wouldn't be so hasty.

First, you're basing it on a bad example.  Rebinding sys.stdout is bad
programming, period.  Wanting to save and restore it is even worse.
You might need this if you can't change the code you're calling, but
ordinarily, you should run away screaming from this kind of thing.
So, do you really want to go to -0 based on an example that really has
no business existing?

Second, that the "rebinding" statement should come before the "try:"
is *not* a subtlety.  It's the idiomatic way to do it, as familiar to
those who use it as the "while 1: break" idiom.


-- 
CARL BANKS



From duncan at NOSPAMrcp.co.uk  Tue Apr 29 08:41:02 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 29 Apr 2003 12:41:02 +0000 (UTC)
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com> <b8lpof$q0f$1@pegasus.csx.cam.ac.uk>
Message-ID: <Xns936C8AB67B739duncanrcpcouk@127.0.0.1>

"Garrison Tsang" <hygt2 at cam.ac.uk> wrote in
news:b8lpof$q0f$1 at pegasus.csx.cam.ac.uk: 

> I know nothing about Python but know a fair bit about C.
> 
> There is a program written in Python by a friend.  It takes in a file
> and decrypts it , which I would like to integrate into a C program I
> am writing. I want to see how the python code was implemented and
> whether I can implement it is C.
> 
> Any solution for this?
> 

The easiest solution is to leave the program in Python and just call the 
relevant functions from C. Possibly the easiest way to do this would be to 
use Pyrex.

Here is a simple example of calling Python from C taken from the Pyrex 
distribution. The Pyrex code in the .pyx file can import and call any 
existing Python modules. You can also pass arguments and return results 
with automatic conversion from simple C types (more complex types you have 
to convert yourself).

---- main.c ----
#include "Python.h"
#include "embedded.h"

int main(int argc, char *argv) {
  Py_Initialize();
  initembedded();
  spam();
  Py_Finalize();
}
---- embedded.pyx ----
cdef public void spam():
  praise()

def praise():
  print "Spam, glorious spam!"
---- end ----

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From mwh at python.net  Wed Apr 16 15:12:36 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 19:12:36 GMT
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com> <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk> <3E9D8B1B.24050D34@engcorp.com>
Message-ID: <7h3brz6qmsc.fsf@pc150.maths.bris.ac.uk>

Peter Hansen <peter at engcorp.com> writes:

> Michael Hudson wrote:
> > 
> > Peter Hansen <peter at engcorp.com> writes:
> > > The last time I used Google (of which I'm well aware of course), it
> > > told me that Knuth was in fact merely paraphrasing someone else
> > > (google can tell you who it was, if you wish ;-),
> > 
> > Yes, you can find lots of people claiming it was Hoare that said this.
> > However, their plurality doesn't necessarily make them right.
> 
> True... but neither does the fact that Knuth (also?) said it make
> them wrong.

This is true.

> On the other hand, this is the first time I've seen a claim, found
> via google or not :-), that Hoare did not in fact say that.  Is your
> view authoritative, or a supposition?

This all came up in a thread on comp.lang.lisp (I have the quote in my
sigfile), which you can probably google-group for and where people
(eventually) sounded quite certain.

And unless someone went to the bother of faking a poorly-scanned pdf
of Knuth's paper, the quotation and attribution I gave is correct.

It is entirely possible that Hoare said something like this *as well*,
of course.  I don't have a reference, however, and have never seen
one.

Cheers,
M.

-- 
  Worryingly, DEFUN appears to be a function that removes all the
  fun from something: after using it all your code is converted 
  to C++.                              -- Tim Bradshaw, comp.lang.lisp



From andrew at acooke.org  Sat Apr 26 18:52:14 2003
From: andrew at acooke.org (andrew cooke)
Date: Sat, 26 Apr 2003 18:52:14 -0400 (CLT)
Subject: please comment on technologies
In-Reply-To: <slrnbalv50.254.luc@trillian.dont-panic.info>
References: <slrnbalplq.254.luc@trillian.dont-panic.info> 
     <HqCqa.8931$3M4.241359@news1.tin.it> 
     <slrnbalv50.254.luc@trillian.dont-panic.info>
Message-ID: <49920.127.0.0.1.1051397534.squirrel@127.0.0.1>

luc wastiaux said:
> In article <HqCqa.8931$3M4.241359 at news1.tin.it>, Alex Martelli wrote:
>>> The project would be developped by a group of 4 to 6 CS students
>>> (including me), over a three months timespan (but not full time, we
>>> have
>>> to go to classes and other stuff)
>>
>> The most valuable thing you'll learn will be how to structure your
>> work.  I heartily recommend incremental development (most crucial
>> features first), test-driven design, pair programming, and a healthy
>> diffidence towards "big design up front" - you may get the temptation
>> of designing in some big chunks of infrastructure preparing for future
>> enhancements... that may never come.
>
> In short, you recommend following extreme programming guidelines, did I
> understand your advice correctly ?

[i'm adding unsolicited comments here!]

there's no one right answer in my opinion - everything depends on
circumstance.  if you're a tight, cohesive group that know each other and
have lots of experience then incremental development works (can't comment
on pair programming).  a lot of planning might be better if you're new
together, haven't  had much experience, or have different opinions on
what's happening(!).  most likely you'll find your own compromise -
planning at a high level, followed by implementing the main structures,
followed by iterations of planning and implementation at lower levels,
with some "jumping outside the box" to ensure that you're not duplicating
work as you work down the tree of tasks.  alternatively, you can try going
bottom up and try writing "one to throw away" in the first third of your
time available.  might work with python, could be risky...

the one thing that does seem to be (almost!) always worth doing is writing
plenty of tests, when it's easy to do so (and consider modifying your
design to make testing easier).

>>> Spyce today and it looks like I could use it for the web interface.
>>
>> Sure, that's one possibility.  Personally I'd choose Webware, which
>> offers an amount of flexibility I appreciate, but Spyce's OK too.
>
> I will have a look a webware. I was looking forward to using spyce because
> it features <?php style tags to embed code into HTML templates, does
> webware have that ?
> I just know webware from this url:
> http://colorstudy.com/docs/shootout.html
[...]
>>> Then I have this idea about using XSL for templating, the idea is not
>>> only
>>> to separate the data from the representation, but also allow
>>> programmers
>>> to "leech off" the search engine (it's easier to do this if the output
>>> is
>>> XML).
>>
>> The ability for other programmers to send "queries" to your search
>> engine and get XML output is good.  However, personally, I would avoid
>> XSL (unless you guys are already well familiar with it -- or NEED to
>> get very familiar with it anyway to pursue your studies).  It's simpler
>> to have separate URLs, or a form field requesting XML or HTML output,
>> and just use different templates with the same contents -- doing your
>> templating in Cheetah (or preppy/Spyce/whatever).
>
> Ok, I thought XSL was the definitive answer to all templating needs, but
> you don't seem very excited about it ?

in my personal experience, xsl is a very neat solution that's very hard to
self in practice.  if you're a bunch of smart programmers that are capable
with functional programming then i'd say go with it - you'll save time. 
but it's a difficult job getting people used to more traditional
approaches to take it seriously for long enough to get past the learning
curve.

that's from working with java - i don't know what python's xsl support is
like (when i last looked i was less than impressed with python compared to
java in this regared, but i suspect it's improved since then.  you should
check this out as there's nothing worse than having to work round bugs in
external libraries that are critical to the project).

have fun,
andrew

-- 
http://www.acooke.org/andrew




From grante at visi.com  Wed Apr 16 14:06:49 2003
From: grante at visi.com (Grant Edwards)
Date: 16 Apr 2003 18:06:49 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com>
Message-ID: <3e9d9bb9$0$151$a1866201@newsreader.visi.com>

In article <a8b7f07a.0304160946.73760d3a at posting.google.com>, A. Lloyd Flanagan wrote:
> Dan Jacobson <jidanni at dman.ddts.net> wrote in message news:<87he90b8rp.fsf at jidanni.org>...
>> I need to use one liners every day, especially in Makefiles.  I thus
>> need to know how to make any kind of python program into a big long one
>> liner. However, right out of the starting gates I run into
>> $ python -c 'print 2;for i in (1,4):print i'
>>   File "<string>", line 1
>>     print 2;for i in (1,4):print i
>>               ^
>> SyntaxError: invalid syntax
>>
> ... 
>> 
>> This makes no output and returns no error value to the shell:
>> $ python -c 'import math' -c 'for i in range(200,500,50): print
>>   i,360/2/math.pi*math.asin(i/1238.4)'
> ...
> 
> You don't have to do these as one liners.
> 
> You can use a shell trick called a 'here document'.  It lets you put a
> set of lines directly in the shell script, and pass them as input to a
> program.  Here's an example:
> 
> #!/bin/ksh
> python << EOF
> y = 0
> for i in range(10):
>         y += i
>         print i, y
> EOF
> 
> The '<< EOF' tells the script to send everything to python until the
> line that contains just EOF  (and you can use any label, not just
> EOF).  So for your example:
> 
> python << DONE
> print 2
> for i in (1,4):
>      print i
> DONE

But how do you do it in a Makefile?

-- 
Grant Edwards                   grante             Yow!  I want to dress you
                                  at               up as TALLULAH BANKHEAD and
                               visi.com            cover you with VASELINE and
                                                   WHEAT THINS...



From fperez528 at yahoo.com  Wed Apr  9 17:03:17 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 09 Apr 2003 15:03:17 -0600
Subject: [Resend] A few errors in "Python in a Nutshell" - Numeric related
Message-ID: <b721ql$akr$1@peabody.colorado.edu>

This is just a resend since it seems A Martelli is back, and I'm pretty sure
he did a massive delete of c.l.py, considering the volume this group
generates :)  I did drop a note on the O'reilly site about this stuff, so
this is just a heads up.

--------------------------
Hi,

I recently got Python in a Nutshell, and as I expected, I've found it to be
extremely informative and well written.  I haven't read it in detail, but
I'm glad I got it.

However, I'd like to point out a few problems I found in Chapter 15, Numeric
processing.  I did not read it in excruciating detail, so please don't
consider this an exhaustive bug hunt of that chapter.  And by the way, that
chapter is a fabulous reference to NumPy, far better for most things than
the standard manual which is organized in a 'cute' (read random) way and
for some things is nearly useless.  Ch. 15 actually explains many things
which the NumPy docs gloss over, and the function reference is
*alphabetized*!! (how hard would it be to do that in the normal manual??)

Anyway, my few nits for future reference:

1) p 306 - a[i][j] vs a[i,j].  This one isn't so much an error but rather a
possibly misleading subtlety.  In p. 306 it says

"... if a's rank is at least 2, you can write a[i][j] as a[i,j] for any
valid i and j..." 

This is, of course, perfectly true and valid.  However, I've seen many
people bitten by thinking that [foo][bar] == [foo,bar] in all cases, and it
is not.  Here's a simple example:

In [1]: z=zeros((3,4,5))

In [2]: z[:][0].shape
Out[2]: (4, 5)

In [3]: z[:,0].shape
Out[3]: (3, 5)

Obviously, in the first case the [:] simply makes a dummy reference to all
of z and the [0] picks its first element along its first axis, giving an
array of shape (4,5).  In the second case, we get the first element along
the second axis, hence an array of shape (3,5).  It might have been
worthwhile pointing this out, as it's a trap that newcomers fall into
repeatedly, especially those with a C/C++ background (a significant chunk
of NumPy's audience).



2) p 312 - flat

In page 312, it says that

"a.flat is an attribute that contains an array with rank of one less than
a..."

This is actually only true for rank 2 arrays.  In general, a.flat is
always a rank 1 array, as long as a is contiguous:

In [4]: z.flat.shape
Out[4]: (60,)

a.flat is truly a 'flat' view of a's data, not just a 'rank(a)-1' view.


3) p 320 - ravel

In p. 320, it says 

"ravel (a) returns the flat form o f a, just like array(a,copy=False).flat"

This is true only if a is contiguous.  The whole point of using ravel is
that it works even if a is non-contiguous:

In [8]: z1=z[::2,::2,::2]

In [9]: z1.iscontiguous()
Out[9]: 0

In [10]: ravel(z1)
Out[10]: array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])

In [11]: array(z1,copy=0).flat
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)

?

ValueError: flattened indexing only available for contiguous array


So the above would be true if it said "... just like
array(a,copy=True).flat".  In that case it would be correct, as a fresh
copy will always be contiguous, hence with a valid .flat attribute.


Anyway, these are just some small nits which readers which use Numeric may
want to be aware of.  If there is an errata page for the book, it may be
worth adding them to it.

Cheers,

Fernando



From inyeol.lee at siimage.com  Fri Apr 11 14:55:37 2003
From: inyeol.lee at siimage.com (Inyeol Lee)
Date: Fri, 11 Apr 2003 11:55:37 -0700
Subject: split encloser
In-Reply-To: <435d40e0.0304101653.1eb11a6c@posting.google.com>
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com>
Message-ID: <20030411185537.GF14255@siliconimage.com>

On Thu, Apr 10, 2003 at 05:53:53PM -0700, Chris wrote:
> Jason Tiller <jtiller at sjm.com> wrote in message news:<mailman.1049417184.1536.python-list at python.org>...
> 
> > 
> > Good luck!  I hope this pattern at least gives you a starting point
> > for implementing the split() in Python.
> > 
> > ---Jason
> > Sonos Handbell Ensemble
> 
> Jason, thanks a lot for the example and the explanation. I fired up
> Perl and it worked like a charm. Now to try and understand it!  I
> looked at stuff about Python regular expressions and there is
> something called finditer and findall that might be useful for parsing
> a string, but can't seem to find any documentation on them.

It's in Library Reference section 4.2.3. (I've checked this in 2.3a2
document, but It'll be there in 2.2 also.)

Brief summary for findall and finditer is;

findall returns list of matched string.
finditer returns iterator of match object. You can do a lot more than
findall since it supports all match object attributes.

Inyeol...




From abpillai at lycos.com  Thu Apr 17 09:42:14 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 17 Apr 2003 06:42:14 -0700
Subject: PDFLIB dll
References: <mailman.1050570689.8988.python-list@python.org>
Message-ID: <bd993a2f.0304170542.65966e40@posting.google.com>

Stupid answer

Find if your python dlls folder is there
in your PYTHONPATH. Probably you need to
copy it as lidpdf.pyd 

Anand

"romain chabot" <s p a w k at hotmail.com> wrote in message news:<mailman.1050570689.8988.python-list at python.org>...
> Hi,
> 
> I'm trying to insert an existing pdf document into a new one, created 
> with PDFLIB.
> I tried the quickreference.py file given with the library, but an error 
> occurs when interpreting : 
>  
> File "E:\PDFlib-5.0.0-Windows\bind\python\quickreference.py", line 8, in 
> ?
> from pdflib py import *
> ImportError: DLL load failed
> 
> I have put the libpdf py.dll in the python dlls folder. I don't 
> understand the situation
> Sorry for this stupid question.
> 
> Thanks a lot
> 
> --



From sean at activeprime.com  Wed Apr 30 14:43:38 2003
From: sean at activeprime.com (Sean)
Date: 30 Apr 2003 11:43:38 -0700
Subject: mxODBC: Getting source table for columns in the cursor.description
Message-ID: <ef7a350b.0304301043.529674f@posting.google.com>

Quick mxODBC question.  I have a db with two tables, 'accounts' and
'contacts', where both tables have a column named 'phonenumber'.  I'm
pushing an SQL select statement using Egenix's mxODBC along the lines
of "select * from accounts a1 INNER JOIN contacts c1 on a1.<some
column> = c1.<some column>".

Everything works great.  The only problem is when I get the
description from the cursor I have a list of column names without
knowing what table they came from.  Using the above example, I would
have two "phonenumber" columns listed in the description.  Is there a
way to determine which listing came from which table?  I would like to
avoid having to rewrite my SQL statement to specifically list each and
every column selected with "select x as table.x", but I need a way to
know what columns came from which table.  Is there functionality built
into mxODBC to do this that I'm missing?  I've been going through the
API here:

http://www.egenix.com/files/python/mxODBC.html

but without any luck.

-Sean



From yaipa at yahoo.com  Thu Apr 17 02:00:48 2003
From: yaipa at yahoo.com (yaipa h.)
Date: 16 Apr 2003 23:00:48 -0700
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com>
Message-ID: <6e07b825.0304162200.34884d36@posting.google.com>

All,

Here in the SF Bay Area I have only met one VB programmer willing to say
that he uses VB.net. The other VB programmers I find are talking about
either sticking with VB6 (most), taking Java courses (some) or playing 
around with the C# idea (few).  Most share my feelings about the .net
thing. That it will not be stable until release +3.x and there are 
little to no existing libraries. You know, batteries. ;-) So yes there 
is a window and it is wide open.

With that said, without a slick, stable GUI builder with all the 
look and feel of the VB GUI builder it's a long way from here
to there. I mean what do you expect them to do, pass in command-line 
arguments or hand assemble TK widgets. :^) 

People have said it in the past, the Python community needs to
pick ONE of the Python wxWindow GUI builders currently under
development and really promote the hell out of it.

Cheers,

Alan


aahz at pythoncraft.com (Aahz) wrote in message news:<b7kqr8$jsq$1 at panix2.panix.com>...
> In article <Wrena.33096$T34.876856 at news2.tin.it>,
> Alex Martelli  <aleax at aleax.it> wrote:
> >Steve Holden wrote:
> >>
> >> You'd do better to call it "Python for VB programmers", expecially since
> >> the change to VB6 is so radical that migrating to a new and more sensible
> >> language might be an attractive option to some!
> >
> >Changes from VB5 to VB6 weren't all that radical.  It was the change
> >from VB6 to VB.NET, aka VB7, that was the shocker (dozens of backwards
> >incompatible changes -- mostly taking VB's semantics closer to Python's,
> >btw, and therefore actually for the better, but still...).  I thought
> >the window of opportunity -- millions of VB6 programmers reluctant to
> >move to dotNET -- had more or less closed by now, but Aahz apparently
> >has identified that it's still open.  If that perception is correct,
> >then the book would indeed have quite a target-market, potentially.
> 
> Correct.  I was at the Waterside Conference last week (put on by my
> agent's company, aimed at writers and publishers).  One presentation
> claimed that a recent survey found 50% of VB6 programmers refusing to
> upgrade to VB.net.  While that's likely overstated, that's a good
> indication of where the market currently is, and even a tiny fraction of
> that market would be a big boost for us.



From anton at vredegoor.doge.nl  Mon Apr 21 13:35:06 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 21 Apr 2003 19:35:06 +0200
Subject: __setitem__ puzzle
References: <mailman.1050940844.16977.python-list@python.org>
Message-ID: <b81bh5$jca$1@news.hccnet.nl>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote:

>          root = y         # save 1st level mutable items

Thanks a lot!

>Is equivalent to L[1,1,1], and you can also do fun things like L[1:2,
>4:9]. Advice: special case one tuples (L[n]).

Hmm, I hadn't thought about slicing. I think raising an error for
arguments without len() defined is a feature. One could always use
(0,) if necessary. This way it's clear that it's not a list.

By the way, see below for the reason this came up. Using different
filter expressions on the result of the indices function it's possible
to make a list of all leaf nodes or do other tree-like operations on a
list (not shown here).

Anton

#By Anton Vredegoor anton at vredegoor.doge.nl Apr 21, 2003
# with some help by Bjorn Pettersen

from __future__ import generators
False,True = 0,1

class indexer:

    def __init__(self,maxima = []):
        self.maxima = maxima
        self.indices = [0 for i in range(len(maxima))]
    
    def __iter__(self):
        maxima,indices = self.maxima, self.indices
        while maxima:            
            yield tuple(indices)
            indices[-1]+=1
            while indices[-1] == maxima[-1]:
                maxima.pop()
                indices.pop()
                indices[-1]+=1

    def addlevel(self,n,start=0):
        self.maxima.append(n)
        self.indices.append(start)

class indexable(list):
    
    def __getitem__(self,x):
        y = self[:]
        for z in x:
            y = y[z]
        return y

    def __setitem__(self,x,val):
        root = y         # BP: save 1st level mutable items
        for z in x[:-1]:
            y = y[z]
        y[x[-1]] = val
        self[:] = root   # BP: update 1st level mutable items
       
def issequence(item):
    try:
        item + ''
    except TypeError:
        try:
            for x in item:
                break
        except TypeError:
            return False
        else:
            return True
    else:
        return False

def indices(L):
    res = []
    d = indexer([len(L)])
    g = iter(d)
    R = L[:]
    while R:
        res.append(g.next())
        x,R = R[0],R[1:]
        if issequence(x):
            R = x + R
            d.addlevel(len(x),-1)
    return res

def test():
    L = indexable(['aa',['bb',['cc','dd'],'ee'],'ff',['gg']])
    print L
    X = indices(L)
    for x in X:
        print "%-15s%-30s" %(x,L[x])

if __name__=='__main__':
    test()





From ianb at colorstudy.com  Mon Apr  7 01:02:54 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 07 Apr 2003 00:02:54 -0500
Subject: cgi form handler question
In-Reply-To: <3e90ff67$0$13167$5a62ac22@freenews.iinet.net.au>
References: <3e90ff67$0$13167$5a62ac22@freenews.iinet.net.au>
Message-ID: <1049691774.660.213.camel@lothlorien>

On Sun, 2003-04-06 at 23:32, DJ_3000 wrote:
> eg I have the fields:
> 
> name
> email
> address
> age
> salary
> 
> 
> the user forgets to input his email and age
> how would i write a for loop to output:
> 
> email
> age
> 
> for name in form.keys():
>         print name

You'd have to define the fields you required, like:

requiredFields = ['name', 'email', 'address', 'age', 'salary']
for field in requiredFields:
    if not form.has_key(field):
        print field


Or you might be able to do:

for name in form.keys():
    if not form[name].value:
        print name






From garryknight at gmx.net  Wed Apr 16 21:46:57 2003
From: garryknight at gmx.net (Garry Knight)
Date: Thu, 17 Apr 2003 02:46:57 +0100
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org>
Message-ID: <1050544015.99640.1@iris.uk.clara.net>

Gerhard H?ring wrote:

> python -c """
> print 2
> for i in (1, 4):
>     print i
> """

Maybe I didn't completely understand what the OP meant, but this isn't a
one-liner. It's a five-liner.

-- 
Garry Knight
garryknight at gmx.net   ICQ: 126351135
Linux registered user 182025



From aussie2010 at yahoo.com  Tue Apr  1 17:38:14 2003
From: aussie2010 at yahoo.com (Chris)
Date: 1 Apr 2003 14:38:14 -0800
Subject: Problem with Idle
Message-ID: <435d40e0.0304011438.5b7b0595@posting.google.com>

I have a module and a main program that imports it. When I make
changes to the module in one IDLE window, and run the program in the
main window, the changes are not picked up (I did do ctrl-s in the
module window to save it). But if I go to the MS-DOS prompt and run
the main program it does run the module with the latest changes to
it??????



From sjmachin at lexicon.net  Sat Apr 19 18:59:04 2003
From: sjmachin at lexicon.net (John Machin)
Date: 19 Apr 2003 15:59:04 -0700
Subject: When not to use an RE -- an example
Message-ID: <c76ff6fc.0304191459.2adf73a@posting.google.com>

I needed a check for strings consisting of repeated characters -- like
when users type "ZZZZZZZ" instead of "UNKNOWN" into a database field.
After implementing the obvious overlapping-substring comparison, I got
to thinking how this could be done with REs. The following resulted:

import re
repeats1 = re.compile(r"^(?:(.)(?=\1))+\1\Z", re.DOTALL).match
def repeats2(s):
   return len(s) > 1 and s[1:] == s[:-1]
for testvalue, expected in zip(
   ['','x','xx','xxx','xxxxxx','xy','xxy','xyy','\n\n\n','aaa\n'],
   [0,  0,  1,   1,    1,       0,   0,    0,    1,       0     ]):
   print repr(testvalue), not not repeats1(testvalue),
repeats2(testvalue), expected

Pergly/phugly, eh? Note the effort required to ensure the newline
cases worked.



From skip at pobox.com  Thu Apr 24 11:19:22 2003
From: skip at pobox.com (Skip Montanaro)
Date: Thu, 24 Apr 2003 10:19:22 -0500
Subject: What means exactly "Memory error"?
In-Reply-To: <200304241406.20833.b.maryniuk@forbis.lt>
References: <200304241406.20833.b.maryniuk@forbis.lt>
Message-ID: <16040.122.803504.236859@montanaro.dyndns.org>

    Bo> Somebody knows what "Memory error" might appear and how to avoid (if
    Bo> possible) it?

It means a memory allocation using malloc or realloc failed.  In short,
virtual memory was exhausted.

Skip




From jhaveri at usc.edu  Fri Apr 11 03:59:51 2003
From: jhaveri at usc.edu (jinal jhaveri)
Date: Fri, 11 Apr 2003 00:59:51 -0700
Subject: Help with text editing!
Message-ID: <8f53798f7995.8f79958f5379@usc.edu>

Dear All


I have written following code for the file open wizard of my editor

which sets the text.
where text arg is one read from a file.


 def setAllText(self,text):
                self.text.delete('1.0',END)
                self.text.insert(END,text)
                self.text.mark_set(INSERT,'1.0')
                self.text.see(INSERT)

Now I want two things to be donw

1) I want to display this text such that alternate rows have same color and alternate columns have same color

for eg. 
assume columsn are alternatively in blue and green.

BGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGBGB (ASSUMING THIS IS BLACK)
WGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGWGW (ASSUMING THIS IS WHITE)

Can anyone give me an idea how do I modify this text to do that?

2) Instead of allowing the edit for the whole text I have entered, I want editing possible only line per line basis.
for eg. 

name1 JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
name2 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
name3 KKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKK

so when I select this text , it should allow me to edit only the line  (only line not even name1,name2 etc) on which I do double click and it shouldn't allow me to edit the whole text.


I will highly appreciate if someone can give me hints on how to go about





From steve.cassidy at mq.edu.au  Tue Apr  1 17:37:49 2003
From: steve.cassidy at mq.edu.au (Steve Cassidy)
Date: Wed, 02 Apr 2003 08:37:49 +1000
Subject: HTMLView widget for Tkinter
References: <mailman.1049197502.20990.python-list@python.org> <b6c0rm$1o95$1@biggoron.nerim.net>
Message-ID: <3E8A14BD.2070601@mq.edu.au>

Gilles Lenfant wrote:
> "Egor Bolonev" <ebolonev at front.ru> a ?crit dans le message de news:
> mailman.1049197502.20990.python-list at python.org...
> 
>>Hello all,
>>
>>I need to show html e-mail in my Tkinter program. Is it possible?
>>
> 
> 
> There's no HTML panel in Tkinter like wxPython (and perhaps others)
> provides. 

I'm not 100% on how tkinter organises itself but there is a Tk html 
display widget:

http://www.hwaci.com/sw/tkhtml/

you'd obviously need to load this into tkinter's tcl interpreter and 
figure out how to call it etc. This shouldn't be impossible.

There's a web browser (browsex) based on this widget as well as the five 
line web browser at http://mini.net/tcl/2993

Steve




From dlarnold at west.com  Fri Apr 11 09:39:19 2003
From: dlarnold at west.com (Don Arnold)
Date: 11 Apr 2003 06:39:19 -0700
Subject: collecting data from file
References: <3E9671A3.2050004@sympatico.ca> <slrnb9d2sh.1oq.jeremy@saturn.cps.co.uk>
Message-ID: <609dd410.0304110539.57abe2d0@posting.google.com>

Jeremy Yallop <jeremy at jdyallop.freeserve.co.uk> wrote in message news:<slrnb9d2sh.1oq.jeremy at saturn.cps.co.uk>...
> Mustafa Celik wrote:
> > I want to scroll thru a file;
> >    * find lines that match a string (e.g. HELLO) on the 2nd column
> >    * add up the 4th column (an integer) on each matching line, say the 
> > variable is TOTAL
> >    * subtract the 4th column from TOTAL if another is string (e.g. BYE) 
> > is hit on 2nd column of a line
> 
> I'd use awk for this.
> 
>   { 
>     if ($2 == "HELLO")    { total += $4 }
>     else if ($2 == "BYE") { total -= $4 }
>   }
>   END { print total }
> 
> Jeremy.

Last time I checked, this was the Python newsgroup, so I think a
solution _in_ Python is more appropriate:

total = 0
infile = open('c:/temp2/input.txt')

for line in infile.readlines():
    line_items = line.split()
    if len(line_items) >= 4:
        print '[%s]' % line_items
        tag = line_items[1]
        amt = float(line_items[3])
            
        if tag == 'HELLO':
            print 'adding', amt, '...'
            total += amt
        elif tag == 'BYE':
            print 'subtracting', amt, '...'
            total -= amt
        print 'total:', total

---- output: ----

[['0', 'HELLO', 'stuff', '5']]
adding 5.0 ...
total: 5.0
[['1', 'HELLO', 'stuff', '3']]
adding 3.0 ...
total: 8.0
[['2', 'BYE', 'stuff', '2']]
subtracting 2.0 ...
total: 6.0
[['3', 'HELLO', 'stuff', '4']]
adding 4.0 ...
total: 10.0
[['4', 'BYE', 'stuff', '5']]
subtracting 5.0 ...
total: 5.0
[['5', 'HELLO', 'stuff', '2']]
adding 2.0 ...
total: 7.0
[['6', 'BYE', 'stuff', '1']]
subtracting 1.0 ...
total: 6.0
[['7', 'HELLO', 'stuff', '5']]
adding 5.0 ...
total: 11.0
[['8', 'BYE', 'stuff', '10']]
subtracting 10.0 ...
total: 1.0

HTH,
Don



From oktaysafak at ixir.com  Wed Apr  2 05:36:14 2003
From: oktaysafak at ixir.com (Oktay Safak)
Date: Wed, 2 Apr 2003 13:36:14 +0300
Subject: Python equivalent of _tempnam?
References: <3e8ab45b$1@mail.hmgcc.gov.uk>
Message-ID: <007201c2f903$b19d7660$0200a8c0@b>

I use two functions like the ones below (slightly modified for general use):

def get_range(folder):
    import os
    folderlist = os.listdir(folder)

    if len(folderlist) == 0:
        return ([], []) # Empty folder
    else:
        existing_list = []
        for item in folderlist:
            # get the file name without path and ext.,
            # extract the number from filename and append to existing_list

existing_list.append(extract_number(os.path.split(os.path.splitext(item)[0])
[1]))

            check_list = [f for f in range(min(existing_list),
max(existing_list)+1)]
            missing_list = [f for f in check_list if f not in existing_list]

            return existing_list, missing_list

def extract_number(filename):
    """Very simple minded solution, assumes that filename
       is in *****123.*** form, * being non-numeric."""
    import string
    num = ""
    for char in filename:
        if char in string.digits:
            num += char
    return int(num) # required for number comparisons!

The first function gives you the list of existing files and the missing ones
in that range. All you have to do then is to use
int(max(get_range(folder)[0]))+1 for your next file number.

Note that these are quick hacks just to get the work done, not works of art.
Be careful about exceptions and make sure you handle them properly. Also,
when you have too much files in your folder (I mean thousands!) this
approach gets verrry slow! Then maybe you should start considering keeping
an index about the files in the folder.

Hope this helps.


----- Original Message -----
From: "Richard" <richardd at hmgcc.gov.uk>
Newsgroups: comp.lang.python
To: <python-list at python.org>
Sent: Wednesday, April 02, 2003 12:58 PM
Subject: Python equivalent of _tempnam?


> Hi,
>
> I am writing an application which produces file output. Files are
outputted
> to a directory and I want to ensure that if an output filename is given
that
> already exists in that directory a number is affixed to the filename to
> ensure that the existing file is not appended to or overwritten. i.e.
> output_0.txt, then output_1.txt etc...
>
> I have been told that a way of doing this in C is by using _tempnam( ).
> However I want to write this in Python, so is there an equivalent
function?
> My application will only be executed occasionally so I am unable to keep
an
> internal record of file numbering. I assume that I need to find the
highest
> number file (if exists) in the directory and then increment that by one
> before creating the new file.
>
> Any ideas?
>
> Cheers
>
> Richard
>
>
> --
> http://mail.python.org/mailman/listinfo/python-list
>





From revanna at mn.rr.com  Mon Apr 21 23:08:58 2003
From: revanna at mn.rr.com (Anna)
Date: Mon, 21 Apr 2003 22:08:58 -0500
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net> <nCUoa.41676$LB6.941312@news1.tin.it>
Message-ID: <pan.2003.04.22.03.08.58.526973@mn.rr.com>

On Mon, 21 Apr 2003 16:02:27 +0000, Alex Martelli wrote:


> As a bonus -- sum([]) now returns 0, while in my original design it raised
> a ValueError, just like e.g. max([]) does.  I thought raising an error was
> best in the name of "errors shouldn't pass silently" and "in the face of
> ambiguity, refuse the temptation to guess".  But -- once we decree that
> "sum is about numbers", there's no real ambiguity, so we can have the
> little extra handiness of having sum([]) not be an error;-).

Hooray!

Sum() is a goodness! And sum([]) returns 0 is a great goodness.
Thanks Alex.

Just my $.03 worth,
Anna



From tj_scarlet at yahoo.com  Wed Apr 16 08:50:11 2003
From: tj_scarlet at yahoo.com (Tj)
Date: 16 Apr 2003 05:50:11 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk>
Message-ID: <ccc7084.0304160450.680bf7ef@posting.google.com>

Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfs7k9vcu2p.fsf at black132.ex.ac.uk>...
> I think your life might be much happier if you just went to
> http://www.drscheme.org/, dowloaded Dr Scheme and used *that* for
> webdevelopment. 

I don't know what it is, but while I learn a lot from those lispish
languages, I never use them in practice.  So much community
fragmentation, and Dr. Scheme seems to want you to differ from the
standard.  I find it more effective to work around Python's functional
inexactness.

I can't really live without lambda since it's a keyword.  But
map/filter/reduce can be done without since I can just express them as
libs.  In fact, "reduce" is one of those terrible lisp names (for
fold-left) that explain nothing.  Lambda could also stand renaming;
Pythonistas probably think it's an arrogant name.  (Or exotic?  I
can't tell.)

The main advantage of functional constructs is that they're a readable
way to get things done in less lines.  Less lines is more important
because Python is a whitespace language, and for that you want your
functions to be concise.  Otherwise you can start to lose your way. 
As far as readability, I think it's more readable than all those for-
and while-loops.  Not having to name simple functions is more
readable.  The only thing is that people used to C programming find it
disorienting to deal with lists as a central thing rather than some
annoyance to package up a bunch of elements.

In my code, I'm almost always handling lists of things.  It's like, if
what I'm doing is useful for one thing, it's probably useful for a
lot.

Tj



From hs2l2ATyahoodotcom at yahoo.com  Sat Apr 19 01:32:53 2003
From: hs2l2ATyahoodotcom at yahoo.com (Haran Shivanan)
Date: 19 Apr 2003 05:32:53 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <b7q2n1$24d2$1@nntp6.u.washington.edu>
Message-ID: <b7qn24$3hmc1$1@ID-148859.news.dfncis.de>

module "Donn Cave" core dumped with the following output:
> So you see the functional programming answers as "higher level"
> than the imperative, procedural approach taken above?  I think
> many functional programming enthusiasts would see it that way,
> but it's more interesting to hear it from a Python programmer.
For me personally, yes, I prefer the functional approach to the
imperative one. I can't speak for the Python programming community
as a whole, though.
> 
> I personally don't see it - I mean, I think there are some good
> things about the functional approach, but I guess I have higher
> standards for higher level.

Are you saying the imperative programming style is higher-level
than functional?
I never really thought about that.





From oliphant at ieee.org  Fri Apr 18 11:53:17 2003
From: oliphant at ieee.org (Travis E. Oliphant)
Date: Fri, 18 Apr 2003 09:53:17 -0600
Subject: Sparse matrix data structures?
References: <20cb355e.0304152053.5d5caaec@posting.google.com>
Message-ID: <3EA01F6D.8010905@ieee.org>


Joze wrote:
> Hi,
> 
> Are there any Python packages implementing some of the most usual data
> structures for sparse matrices (coordinates, compressed rows,
> compressed columns, etc.)? I found SparsePy on the web, but it doesn't
> seem to be maintained any more, and the link to the source is broken.
> How well do these integrate with numarray and/or SciPy?

The CVS version of SciPy has a sparse matrix package.  Last I checked it 
built and worked on my Linux system.

You can get the CVS version of SciPy as per instructions at www.scipy.org

To build the sparse package be sure to modify the setup.py file and add 
'sparse' to the list of packages to be built.


The sparse support is a pretty good start, but it's been a while since I 
last played with it.  It is based on UMFPACK (whose license is another 
reason Sparse is not turned on by default in scipy) and SuperLU.

-Travis Oliphant




From brian at rk-speed-rugby.dk  Thu Apr 24 05:36:17 2003
From: brian at rk-speed-rugby.dk (Brian Elmegaard)
Date: 24 Apr 2003 11:36:17 +0200
Subject: pickling a class definition
References: <uy923x83o.fsf@mail.afm.dtu.dk> <3EA50896.CBA88543@alcyone.com>
Message-ID: <uwuhkp78u.fsf@mail.afm.dtu.dk>

Erik Max Francis <max at alcyone.com> writes:

<snipped a detailed explanation>

Thanks, Eric. I am able to this now.

-- 
Brian (remove the sport for mail)
http://www.et.dtu.dk/staff/be



From ark at research.att.com  Fri Apr  4 16:10:32 2003
From: ark at research.att.com (Andrew Koenig)
Date: Fri, 4 Apr 2003 21:10:32 GMT
Subject: output buffering--am I missing something obvious?
References: <yu99istu9ds9.fsf@europa.research.att.com>
Message-ID: <yu99el4i9dfb.fsf@europa.research.att.com>

ark Can someone please explain the following behavior?  Under Python 2.2.2
ark and Solaris 2.8:

>>> import sys
>>> sys.stdout.write("foo")
>>> sys.stdout.flush()
>>> sys.stdout.write("bar\n")
>>> bar

Never mind -- I figured it out.  Apparently the interactive loop
discards all pending output before reading the next input statement.

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From aleax at aleax.it  Tue Apr 15 08:08:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 12:08:39 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
Message-ID: <bDSma.22753$T34.687158@news2.tin.it>

Aahz wrote:

> In article <v9e21mhemm8c2 at corp.supernews.com>,
> Paul Watson <pwatson at redlinec.com> wrote:
>>
>>We need to teach students correct design priciples to build something
>>greater than what already exists.  We will never get very far if we
>>require everyone to start with a quark or atom.  Yes, of course we need
>>some people who design silicon and create microcode.  They will learn
>>the low-level details what they need to know as they need it.  Knowing
>>great design and organization principles will enable them to make the
>>most of it.
> 
> While there's some truth to that, try explaining why the following code
> is a Bad Idea to someone who has no basic understanding of CS:
> 
>     s = ''
>     for i in range(1000000):
>         s += str(i)

I will gladly accept the challenge, if we can agree on the details to
determine whether I've won or lost the bet.  What I require is somebody
with real, earnest *interest* in *understanding* this stuff -- I'll
gladly undertake to supply all necessary *background*, but I have no
idea on how to awake real interest in somebody whose motivation for
learning is just "serving time" or some vague hope of getting a good
paying job (and that would be just as much of a problem, whether the
subject was "CS", "programming", or "artichoke farming techniques").

If the interest is there, and enough Python background to understand
exactly what this program snippet DOES, then in about one or two
hours of face to face instruction I can give a model that will allow
the student to predict this snippet's disastrous performance and
compare it to the expected performance for, e.g.,

    s = ''.join([str(i) for in in range(1000000)])

It's not rocket science (that's _another_ thread...;-) -- at this
level, it's really simple arithmetic, a decent mental model of "an
abstract underlying machine", AND an interest in understanding this.


> Knowing the difference between O(N) and O(N^2) is critical to writing
> even the simplest programs.  OTOH, I do agree with you that focusing on
> programming as a craft is more important to being a programmer than
> learning CS.

I'm not dissing CS (though I could well have some fun dissing it on
some other thread...!-), I'm claiming that one can become a rather
effective programmer in Python with only those modest, approximate CS
concepts that it's feasible to explain within a first programming
course (to _interested_ students...).  For example, while we keep
mentioning big-O ("worst-case"...), what we're really USING most of
the time is really big-Theta ("_typical_ case", _expected_ performance).
Except in hard real-time programming, hardly anybody ever does more
sophisticated analysis than "_expected_ performance" during the design
of actual programs...


Alex




From afriere at yahoo.co.uk  Tue Apr 15 02:03:41 2003
From: afriere at yahoo.co.uk (Asun Friere)
Date: 14 Apr 2003 23:03:41 -0700
Subject: Reading/writing files
References: <mailman.1050342562.4533.python-list@python.org>
Message-ID: <38ec68a6.0304142203.17e426f0@posting.google.com>

Brian Quinlan <brian at sweetapp.com> 
> How do you want the list represented in the file? One element per line?
Maybe he wants them represented as lists (ie as a string representation of the list)

You _could_ do this:


#!/usr/bin/python
#silly_list_writer.py

#create some sample listage
a=5;b=3;c=1
x=[a,b,c];y=[a,b];z=[c,b]

#open a file to write these to, convert em to strings and write em.
outfile = file('lists.txt', 'w') #note the write flag
for lizt in x, y, z :
    outfile.write(str(lizt) + '\n')
outfile.close()

#now read them back into 'collection'
collection = []
infile = file('lists.txt', 'r') #read flag is optional
for line in infile :
    collection.append(eval(line))
infile.close()

#ok lets test that the collection really is [x, y, z]
print collection == [x,y,z] and "It worked!" or "Darn!"
#Note: don't use this last idiom, unless you understand why
#(and when) it is unsafe.  Gee, I really wish 308 would fly
#so people would stop writing stuff like this! :p

As I said you _could_ do this, but you should probably try this code instead:

import pickle; help(pickle)



From i at mindlace.net  Thu Apr 24 22:59:45 2003
From: i at mindlace.net (emf)
Date: Thu, 24 Apr 2003 22:59:45 -0400
Subject: What means exactly "Memory error"?
In-Reply-To: <200304241843.04939.b.maryniuk@forbis.lt>
References: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net> <200304241843.04939.b.maryniuk@forbis.lt>
Message-ID: <3EA8A4A1.4010405@mindlace.net>

Bo M. Maryniuck wrote:

>>The most likely cause is that you've
>>simply run out of memory (#1 above) -- but that's more a symptom than an
>>explanation.  Don't expect it to be easy to resolve.
> 
> Yes, you're right: I'm just out of memory, though server has 4G RAM... :/ Have 
> you at least any ideas how to decrease/pack collected data to avoid fat 
> growing?
> 
> P.S. Or at least which library I should use to parse HTML with DOM using 
> something else, instead HtmlLib...?

You might try minidom or cDomlette instead, if you need something 
DOMlike. If you can deal with an even more lightweight structure I might 
suggest pyRXP or elementtree (my favorite).

Your application really sounds like it could work with a SAX based 
approach, as Skip suggested.

~mindlace





From duncan at NOSPAMrcp.co.uk  Thu Apr 17 11:24:00 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 17 Apr 2003 15:24:00 +0000 (UTC)
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <d6jmqr2q.fsf@python.net> <Xns936072232AA84duncanrcpcouk@127.0.0.1> <he8xgr3g.fsf@python.net> <Xns9360A249561CBduncanrcpcouk@127.0.0.1>
Message-ID: <Xns9360A663C7BA5duncanrcpcouk@127.0.0.1>

Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in
news:Xns9360A249561CBduncanrcpcouk at 127.0.0.1: 

> The difference code could actually be made much faster quite easily.
> If I added a search for the first difference, then I could do the
> ellipsis and truncating first and only pass what is left to difflib.

Old version took 0.18 seconds on Python 2.3, gave up waiting on Python 2.2. 
This version takes 0.18 on Python 2.2, 0.08 on Python 2.3.

---- unittest.diff ---
*** unittest.py.orig	Thu Apr 17 10:08:03 2003
--- unittest.py	Thu Apr 17 16:18:54 2003
***************
*** 146,152 ****
  
  class TestFailed(Exception):
      pass
!   
  class TestCase:
      """A class whose instances are single test cases.
  
--- 146,202 ----
  
  class TestFailed(Exception):
      pass
! 
! def shortdiff(x,y):
!     '''shortdiff(x,y)
! 
!     Compare strings x and y and display differences.
!     If the strings are too long, shorten them to fit
!     in one line, while still keeping at least some difference.
!     '''
!     import difflib
!     LINELEN = 79
!     def limit(s):
!         if len(s) > LINELEN:
!             return s[:LINELEN-3] + '...'
!         return s
! 
!     def firstdiff(s, t):
!         for pos in range(0, max(len(s), len(t)), 100):
!             if s[pos:pos+100] != t[pos:pos+100]:
!                 for index in range(pos, pos+100):
!                     if s[index:index+1] != t[index:index+1]:
!                         return index
! 
!     left = LINELEN/4
!     index = firstdiff(x, y)
!     if index > left + 7:
!         x = x[:left] + '...' + x[index-4:index+LINELEN]
!         y = y[:left] + '...' + y[index-4:index+LINELEN]
!     else:
!         x, y = s[:LINELEN+1], t[:LINELEN+1]
! 
!     d = difflib.Differ()
!     diffs = list(d._fancy_replace([x], 0, 1, [y], 0, 1))
!     if len(diffs) == 3:
!         if diffs[1].startswith('+ '):
!             diffs.insert(1, '')
!         else:
!             diffs.append('')
!     if len(diffs) != 4:
!         # String were the same!
!         #return '\n'.join(diffs)
!         return limit(diffs[0][2:])
! 
!     # Remove '- ', '+ ', '? ' markers from start of lines,
!     # and newlines from end.
!     diffs = [ s[2:].rstrip() for s in diffs]
!     if max([len(s) for s in diffs]) < LINELEN:
!         return '\n'.join(diffs)
! 
!     diffs = [ limit(s) for s in diffs ]
!     return '\n'.join(diffs)
! 
  class TestCase:
      """A class whose instances are single test cases.
  
***************
*** 317,324 ****
             operator.
          """
          if first != second:
              raise self.failureException, \
!                   (msg or '%s != %s' % (`first`, `second`))
  
      def failIfEqual(self, first, second, msg=None):
          """Fail if the two objects are equal as determined by the '=='
--- 367,377 ----
             operator.
          """
          if first != second:
+             reprfirst, reprsecond = repr(first), repr(second)
+             if not msg and len(reprfirst) + len(reprsecond) > 60:
+                 msg = "failUnlessEqual\n" + shortdiff(reprfirst, 
reprsecond)
              raise self.failureException, \
!                   (msg or '%s != %s' % (reprfirst, reprsecond))
  
      def failIfEqual(self, first, second, msg=None):
          """Fail if the two objects are equal as determined by the '=='
----------------------

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From skywalkerpackage at hotmail.com  Tue Apr 15 06:45:38 2003
From: skywalkerpackage at hotmail.com (Johannes Eble)
Date: Tue, 15 Apr 2003 10:45:38 GMT
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de> <3E93FBE3.2050301@Linux.ie>
Message-ID: <3e9be022.101163638@news.online.de>

Hello Padraig,

thank for your answer.

Gaphor is an option too, since it is written in Python. However, it is
based on GTK+. I work in Windows and I prefer the WxPython toolkit.
See my other answers.


Regards

Johannes






From tismer at tismer.com  Wed Apr 23 09:42:10 2003
From: tismer at tismer.com (Christian Tismer)
Date: Wed, 23 Apr 2003 15:42:10 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <27cc4d4a.0304222142.199ba7c9@posting.google.com>
References: <mailman.1050569128.32513.python-list@python.org> <27cc4d4a.0304222142.199ba7c9@posting.google.com>
Message-ID: <3EA69832.5070505@tismer.com>

Patrick Logan wrote:
> Christian Tismer <tismer at tismer.com> wrote:
> 
> 
>>Let me simply end this pamphlete with some simple sentences:
>>Stackless Python is more capable of tasklets switching than any
>>other light-weight threading software package.
>>If anyone disagrees, please give me a runnable counter-example.
> 
> 
> This would match the results seen in other implements of lightweight
> threads, e.g. in Erlang, PLT Scheme, Gambit Scheme, etc.

Can you give me some pointers to timings nd maybe comparisons?
I'm not familiar with those langauge's web sites.

> So I can believe a similar implementation of Python would achieve
> similar results.

How fast the fastest can those languages switch tasks on a 1.67 GHz
Pentium IV?

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Sat Apr 26 11:21:45 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Sat, 26 Apr 2003 17:21:45 +0200
Subject: RuntimeError at end of test.autotest in Python 2.3b on windows,
 tmpfile also fails
In-Reply-To: <mailman.1051367863.3218.python-list@python.org>
References: <mailman.1051367863.3218.python-list@python.org>
Message-ID: <3eaaa408$0$49109$e4fe514c@news.xs4all.nl>

Tim Peters wrote:
> [Irmen de Jong]
> 
>>when running "import test.autotest" in Python2.3b on windows xp,
>>I get a RuntimeError at the end:
> 
> 
> As the download page says, please report bugs on SourceForge.

You're right ofcourse but I first wanted to be sure it is a problem
indeed and not some kind of artifact...

I'll report what I know on SF.

--Irmen




From achrist at easystreet.com  Sat Apr 26 01:06:27 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Fri, 25 Apr 2003 22:06:27 -0700
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <14bd1399.0304252032.71752cb5@posting.google.com>
Message-ID: <3EAA13D3.F7B4B9A6@easystreet.com>

kkennedy wrote:
> 
> RTF format (MS open format???) is a moving target.  RTF files created
> by MS Word 97 will look odd or worse in MS Word 95, and be completely
> unreadable in WordPad. Either find out what version of RTF WordPad
> uses and use that, or require your users to have a specific version
> of MS Word (or the free download - MS Word viewer).
> 

That's one reason why I asked if there were any solutions already
available.  If someone had already identified the forms of rtf 
markup that worked well in all those versions, I would not want 
to rediscover the same set of tires.  

> why do your users have to open the document in a word processor if
> it isn't going to be edited?  

OK.  I'll give you an A for deduction.  It is going to be edited.  
I'm generating boilerplate.  User gets to customize.  I should hope
that I don't need to dictate specific word-processor to users. MS
Word may be the most common, but there are many versions, and I have
no reason to favor any particular vendor as they have never
particularly favored me.  I would recommend Abiword as a free 90%
solution for someone who didn't have a word-processor, but I'd like
to make a file that looks better than Ascii that the people with
MS-Works or Corel are who-knows-what will also be able to deal with.

I suppose that if I could use COM to grab the RTF control that comes
with windows, that might be the best way to create an RTF file that
would work on the same machine, but IDK.  And I have no idea if or
where I can get documentation of the API for that RTF COM object, and
I do remember that there are two versions of the MS RTF  dll, too.
My objectives are limited, so now I'm just starting with output from
a simple RTF editor and trying to pick it apart into  templates for
my Python code to fill in.  Maybe that will work, but I think that
there are a few tricks, I'm not sure.  I did something like this to
generate RTF from C++ a few years back, but what we did was very
simple and didn't generate much markup, mostly fixed-font text. I'd
figure that there's a better way someone has worked out.


Al



From max at alcyone.com  Tue Apr 22 18:50:22 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 15:50:22 -0700
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <roy-1971DE.18283722042003@reader1.panix.com>
Message-ID: <3EA5C72E.8C5426CA@alcyone.com>

Roy Smith wrote:

> I'm not 100% sure testing for the presense of 'True' in dir(sys) is
> the
> right test, but it's better than testing for an absolute version
> number.

I don't think it's the right test at all, since True (and False) aren't
in sys at all, they're in the builtins:

Python 2.3a2 (#1, Feb 19 2003, 20:06:49) 
[GCC egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.True
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AttributeError: 'module' object has no attribute 'True'
>>> __builtins__.True
True

Note that your general approach (correcting __builtins__ for sys) is the
same as those others that have been suggested, it's just that you're
doing a manual check in a module rather than just trying to use the name
and see if you get a NameError.

So there's (at least) three approaches of trying to get safe True/False
compatibility:

1.  Unconditionally defined True/False; this might generate warnings
("Hey, you're rebinding a constant!") at some unspecified point in the
future.

2.  Do an explicit test for the version number.

3.  Check for the existence of the names True and/or False (either by
looking in a module or testing by name and catching an error if it's not
there).

Of these, I think the third is the most elegant and least likely to muck
up.

> A better test might be looking for 'BoolType' in dir(types), ...

It's actually called BooleanType (which I only just noticed now; I've
never had the occasion to use it).  I'm not sure why.

[same session, still 2.3a2]
>>> import types
>>> types.BoolType
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AttributeError: 'module' object has no attribute 'BoolType'
>>> dir(types)
['BooleanType', 'BufferType', 'BuiltinFunctionType',
'BuiltinMethodType', 'ClassType', 'CodeType', 'ComplexType',
'DictProxyType', 'DictType', 'DictionaryType', 'EllipsisType',
'FileType', 'FloatType', 'FrameType', 'FunctionType', 'GeneratorType',
'InstanceType', 'IntType', 'LambdaType', 'ListType', 'LongType',
'MethodType', 'ModuleType', 'NoneType', 'NotImplementedType',
'ObjectType', 'SliceType', 'StringType', 'StringTypes', 'TracebackType',
'TupleType', 'TypeType', 'UnboundMethodType', 'UnicodeType',
'XRangeType', '__builtins__', '__doc__', '__file__', '__name__']

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From pinard at iro.umontreal.ca  Sat Apr  5 17:18:35 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 05 Apr 2003 17:18:35 -0500
Subject: Calling C libraries from a Python script
In-Reply-To: <qmCja.4148$Du.14013@newsc.telia.net>
References: <e81e0f3d.0304050639.37b59655@posting.google.com>
	<qmCja.4148$Du.14013@newsc.telia.net>
Message-ID: <oqbrzk7flw.fsf@titan.progiciels-bpi.ca>

[Tim Gahnstr?m /Bladerman]

> "Joshua Goodlett" <joshua at goodish.org> wrote in message
> > I'm trying to find examples of calling C library functions from within
> > a Python script.  Any advice or pointers on where to look for examples
> > would be appreciated.


> There is a whole chapter in the docs dedicated to the subject, Have a look
> at http://www.python.org/doc/current/ext/ext.html


Yet, this whole chapter does not hint at all (so far that I remember) to the
existing tools which are making this whole adventure all simpler.  Nowadays,
I doubt this is the best reference one should offer to a newcomer.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From bokr at oz.net  Thu Apr 17 14:48:47 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 18:48:47 GMT
Subject: More
References: <teBna.9109$pK2.11130@news.indigo.ie>
Message-ID: <b7msuf$aa7$0@216.39.172.122>

On Thu, 17 Apr 2003 16:52:27 +0100, "kpop" <snag at jool.po> wrote:

>Hi
>I usualy print strings the normal way
>print x, but if the string is too big it will
>go by the screen too fast and i wont be
>able to see anything. I notice the python
>internel help system hasa thing like the
>unix prgram more, where you press enter
>to go down. How can i use that in my own
>programs? Is there a function or a module.
>I am using windows so i dont have more.

Do you want to do it for your program with code
in or used by the program, or do you
actually want a "more" program to pipe your
program's output to? I.e., you say you don't
have "more" on your windows, but from the rest
of what you say, having it wouldn't solve your problem.

So you could write a quick and dirty more.cmd (that would
run a simple more.py to do the work) that you could pipe to,
or you could rebind sys.stdout in your program to control
the output stream (assuming your program isn't bypassing
that with low level io somehow).

Regards,
Bengt Richter



From antonmuhin at sendmail.ru  Thu Apr 17 07:25:45 2003
From: antonmuhin at sendmail.ru (Anton Muhin)
Date: Thu, 17 Apr 2003 15:25:45 +0400
Subject: Parsing
In-Reply-To: <Xns93607C5492A1Eduncanrcpcouk@127.0.0.1>
References: <9afd3f36.0304170247.2a082573@posting.google.com> <Xns93607C5492A1Eduncanrcpcouk@127.0.0.1>
Message-ID: <b7m2vt$29kr$1@news.peterlink.ru>

Or something like this:

from __future__ import generators

def aggr(seq, n):
     t = []
     for e in seq:
         t.append(e)
         if len(t) == n:
             yield t
             t = []
     if t:
         yield t + [None * (4 - len(t))]

f = file("e:/1.txt")

for section, num0, num1, _ in aggr(f, 4):
     print section.rstrip(), num0.rstrip(), num1.rstrip()

f.close()

BTW, if I'm correct modern itertools module has analog (much more 
elaborated, I suppose) for aggr.

hth,
anton.




From lx1978 at 263.net  Sun Apr  6 16:38:48 2003
From: lx1978 at 263.net (Xin Liu)
Date: Mon, 7 Apr 2003 04:38:48 +0800
Subject: Is there swap function in Python?
References: <b6q0n6$1c55$1@mail.cn99.com> <3r119vosrdukh6til7bdmgredlhhehqg64@4ax.com> <b6q1so$1cm2$1@mail.cn99.com> <just-25C845.22234606042003@news1.news.xs4all.nl>
Message-ID: <b6q379$1dfv$1@mail.cn99.com>

Just:
    You are right. Thank you very much.
> Try
>
>    b, a = a, b
>
> Just





From aleax at aleax.it  Thu Apr 17 06:44:55 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 10:44:55 GMT
Subject: get pid from popen3 or get redirection in execl
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <bDSma.22753$T34.687158@news2.tin.it> <slrnb9o8s8.iea.news@datatailors.com> <mailman.1050427771.31620.python-list@python.org> <3E9E0FBA.6050503@indy.rr.com> <mailman.1050559712.17758.python-list@python.org>
Message-ID: <HAvna.26564$LB6.621392@news1.tin.it>

Yu Wang wrote:

> Hi,all
>    Does "execl", or whatever this family, support redirection of output?

exec...* are the system calls on top of which redirection of output
is implemented by the shell (there's also a fork in between of course).

>    Say, I have a script: ( on my Linux7.2,C shell environment,with
> python2.2.1)
> 
> command="somecommand"
> arg="-i somefile > output file"
> execl(command,arg)
> 
> what's wrong with this?

You have not involved the shell, so the "> whatever" syntax ain't
gonna fly (who would be parsing it?).  You can explicitly invoke
/bin/sh -c "whatever >theoutputfile" if you wish; or else, do it
at a lower level (with os.dup2 and friends, fork maybe, then exec).


> And,
>     how could I get child pid from "popen3(somecommand)".

http://www.python.org/doc/current/lib/popen3-objects.html


Alex




From claird at lairds.com  Wed Apr  2 13:00:36 2003
From: claird at lairds.com (Cameron Laird)
Date: Wed, 02 Apr 2003 18:00:36 -0000
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <b6f66v$28k$1@slb3.atl.mindspring.net>
Message-ID: <v8m9a4nk4dpjf7@corp.supernews.com>

In article <b6f66v$28k$1 at slb3.atl.mindspring.net>,
Andrew Dalke <adalke at mindspring.com> wrote:
>robin:
>> If you want the biggest, boldest approach and don't care about
>> overhead, use CORBA.
>
>What overhead would this be?  From what I see of omniORB,
>there isn't really that much.  Also, CORBA is the most complete,
>eg, it allows callbacks and passing around object references
>the others don't have.
At a technical level, CORBA's as easy for an application
developer as, say, SOAP.  It has a lot of overhead, though,
in the muddy thoughts of managers.
			.
			.
			.
>I wouldn't put SOAP as simple, and I've had problems with
>interoperability between various packages.  If I needed to pass
>simple data around, I would use XML-RPC.
My first instinct is to write my own tiny protocol, and
pass strings around.  Often enough, that's all I need--
and look!  we're back to language-neutrality.
>
>> If you want the ultimate in simplicity and are willing to foresake
>> multi-language support, use Dopy, Pyro, or Twisted Spread.
>
>One reason I'm looking at Twisted is because it handles other
>interfaces as well.  I need to talk to SQL databases, SOAP and
>XML-RPC servers, straight HTTP, and spawned off external
>processes.
>
>There is also older interfaces, like PVM and MPI from the
>high-performance computing world, and Linda, and more,
>but I haven't looked into them for years.
Tuple space and parallel architectures can be FAR more 
"efficient" for scientific and engineering problems.
SOAP and such are just concessions to commercial misunder-
standings about what business needs.

I write this as a SOAP specialist and advocate.
			.
			.
			.

-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From stigwood4 at yahoo.com  Wed Apr  2 14:32:30 2003
From: stigwood4 at yahoo.com (Jason C.)
Date: 2 Apr 2003 11:32:30 -0800
Subject: Some newbie questions the FAQs didn't answer!
Message-ID: <6ffd97db.0304021132.62f86c7b@posting.google.com>

I was thinking about taking Python up this summer.  I was initially
drawn towards it because I had heard how easy it was to program GUIs
using the various packages.  But then I read that in order to run a
python program, the machine had to have the python VM installed, and
the user had to go to a command line and type "python <program
name>.py".

What I had planned to do was write Win32 GUIs, but I would guess that
most common users don't have Python installed.  Is there some way to
write the program in C++ and then have it bring up the GUI written in
Python without Python being installed?  Wouldn't that require the code
being compiled into machine code?



From rmunn at pobox.com  Wed Apr  2 10:58:22 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 15:58:22 GMT
Subject: else: w/o except: - why not?
References: <mailman.1049211247.3694.python-list@python.org> <pan.2003.04.01.16.47.16.340267@mn.rr.com>
Message-ID: <slrnb8ld57.ccp.rmunn@localhost.localdomain>

Anna <revanna at mn.rr.com> wrote:
> On Tue, 01 Apr 2003 07:33:57 +0000, Michael Chermside wrote:
> 
>> I can't speak for Tim, but my OWN opinion is precisely in alignment with
>> what you have written.
>> 
>> I think that a "bare except" is useful in only two ways. The first is like
>> a "finally" clause that's skipped if everything works. In other words, the
>> following bits of code is (almost[1]) equivalent to your second example:
>> 
>>     everythingWorked = False
>>     try:
>>         f1()
>>         everythingWorked = True
>>     finally:
>>         if not everythingWorked:
>>             f2()
>> 
>> This doesn't come up much, but it's occasionally useful, and your version
>> (using bare except with a raise) is a little prettier.
> 
> I thought that finally clauses are *always* run, whether an exception was
> raised or not. 
> 
> Looks at code closely... So, in this case the finally is run (in terms of
> evaluating the "if" statement), but function f2 may not happen depending
> on the results of the "if". So the finally isn't really skipped, it just
> doesn't do much if everything worked properly in the try clause. Right?
> 
> I'm not being picky - just seeing if I understand this correctly. Thanks.

Yes, you've got it right. If an exception is raised during f1(), then
the "everythingWorked = True" line is never reached, because program
execution flow jumps to the finally: clause, which would then call f2()
because everythingWorked is still False. Note that after any finally:
clause is executed, the exception then continues to propagate upwards.
If f1() raised no exceptions, then the finally: clause is still
executed, but this time everythingWorked is True, so f2() is not called.

The way to express that using try...except is:

    try:
        f1()
    except:
        f2()
        raise

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From otttr440 at student.liu.se  Tue Apr  1 13:04:59 2003
From: otttr440 at student.liu.se (Otto Tronarp)
Date: 01 Apr 2003 20:04:59 +0200
Subject: [Pysqlite-users] Re: How to use dates with pysqlite
In-Reply-To: <20030401013809.GA2127@mephisto.ghaering.test>
References: <1049147540.14660.6.camel@localhost>
	 <20030331222442.GA1336@mephisto.ghaering.test>
	 <20030401013809.GA2127@mephisto.ghaering.test>
Message-ID: <1049220299.8710.3.camel@localhost>

On Tue, 2003-04-01 at 03:38, Gerhard H?ring wrote:
> * Gerhard H?ring <gh at ghaering.de> [2003-04-01 00:24 +0200]:
> > * Otto Tronarp <otttr440 at student.liu.se> [2003-03-31 23:52 +0200]:
> > > Hi,
> > > 
> > > I can't get dates to work with pysqlite. [...]
> > > >>> cur.execute('create table test (d date)')
> > [...]
> > > >>> cur.execute('select * from test')
> >
> > I'll look into getting more date/time support into CVS soon, [...]
> 
> It's in CVS now. Commit message:
> 
> - Dropped DATETIME type code, added support for INTERVAL type
> - Added support for automatic conversion from
>   DATE/TIME/TIMESTAMP/INTERVAL columns, without the need to use the "--
>   types" magic SQL comment
> - Added tests for this automatic conversion and made the tests for the
>   SQLite type support actually work


Hey, youre quick! 
I just compiled sqlite cvs and it seems to work like I charm, well at
least for my simple example.

Thanks a bunch!

Otto







From skip at pobox.com  Wed Apr  2 10:26:45 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 2 Apr 2003 09:26:45 -0600
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to 
        Python.Good idea? Bad Idea? PEP?
In-Reply-To: <3E8AFB77.38D0EDDF@engcorp.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
        <b68jti$2mj66$1@ID-169208.news.dfncis.de>
        <RbCia.6782$9I.555309@news20.bellglobal.com>
        <3E8AFB77.38D0EDDF@engcorp.com>
Message-ID: <16011.309.861144.167177@montanaro.dyndns.org>

    Peter> I just scanned a few tens of thousands of lines of our code, and
    Peter> as near as I can tell only 13 cases out of over 160 instances of
    Peter> "for" loops are using range() or xrange().  

Similar numbers here: 146 out of 898.

It appears the builtins namespace (or at least foreign modules' namespaces)
will get somewhat less dynamic -- just enough less that the byte compiler
can tell if an instance of x?range() is calling the usual builtin instead of
a shadow copy -- so tools that care will be able to optimize "for i in
range(x)" just as well as the pyrex syntax.  There have been discussions
recently about this topic in python-dev.  Also, one of the sprints at PyCon
was aimed at a new version of the bytecode compiler, probably one which
generates code from an abstract syntax tree, allowing easier analysis (and
thus optimization) before generation of bytecode.  So instead of

    LOAD_GLOBAL range

the bytecode might be

    LOAD_BUILTIN range

which tools like Psyco can detect can use to their advantage.

Skip




From martin at v.loewis.de  Tue Apr 29 16:07:23 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 29 Apr 2003 22:07:23 +0200
Subject: Python 2.3b1: RuntimeError using rexec
References: <b8mh4l$2oo0$1@nntp6.u.washington.edu>
Message-ID: <m31xzl6pac.fsf@mira.informatik.hu-berlin.de>

"Russell E. Owen" <no at spam.invalid> writes:

> I've looked at the what's new and the standard documentation for rexec 
> and it seems normal -- no hint that the module has been deprecated or 
> that one needs to jump through some "yes I know it's dangerous" hoop to 
> use it.

See

http://www.python.org/2.3/highlights.html

The rexec module has been disabled.

> I stumbled across this when testing some existing code. This code uses 
> rexec to convert a string representation of a dictionary to a dictionary 
> as part of reading in font preferences.

There simply is no "safe" way to do that in Python.

> 
> Any suggestion on how to do that without rexec would also be 
> appreciated. (I was quite disappointed that "dict" didn't do it!)

eval should work, but it won't be safe if you cannot trust the string.

Regards,
Martin



From pyth at devel.trillke.net  Thu Apr 17 13:38:53 2003
From: pyth at devel.trillke.net (holger krekel)
Date: Thu, 17 Apr 2003 19:38:53 +0200
Subject: Customize rlcompleter
In-Reply-To: <b6sfn0$1aav$1@biggoron.nerim.net>; from glenfant@NOSPAM.bigfoot.com.trillke.net on Mon, Apr 07, 2003 at 08:36:01PM +0200
References: <b6risd$giq$1@biggoron.nerim.net> <80ab95a6.0304070828.4697e4b9@posting.google.com> <b6sfn0$1aav$1@biggoron.nerim.net>
Message-ID: <20030417193853.H22015@prim.han.de>

Gilles Lenfant wrote:
> Eric,
> 
> Many thanks for the enlightenment. Gonna make my own Completer class.

Maybe you are interested in my rlcompleter replacement which has
some nice completion features.

    http://codespeak.net/rlcompleter2/

ciao holger




From drew at astro.pas.rochester.edu  Thu Apr  3 14:02:47 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 3 Apr 2003 11:02:47 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <b6hjle$b4m$1@news1.wdf.sap-ag.de>
Message-ID: <9383481e.0304031102.2db9b048@posting.google.com>

"Daniel Dittmar" <daniel.dittmar at sap.com> wrote in message news:<b6hjle$b4m$1 at news1.wdf.sap-ag.de>...
> Drew Moore wrote:
> > some say in how this takes place. I'm at a loss to explain why
> > the "most trivial case of augmented assignment" was denied this power.
> 
> Performance comes to mind. Assignment is a frequent statement and having to
> search for an __assign__ method will slow things down a lot.
> 
> And if an object has an __assign__ method, how do you perform a real rebind?
> Do we add a 'becomes' operator (similar to the ==/is operators)?
> 
> Daniel

Good points! (and ones I considered)

If the interpreter could tell quickly that the __assign__ method
was in the dictionary (ie, a cached bool flag) the performance
hit would be pretty insignificant, I suspect..

I considered ":=" for the "force name rebind" operator, but don't think
there would be a pressing need for this, would there?

You could always delete the name and reassign it if that is 
what you really needed. The code doing the assign in most
cases would not care about what the assignment did, as
long as the result acted sensibly. Adding a "force rebind" operator
would cause paranoid programmers to use it all the time..
and I don't like the sounds of that. Too messy.

polite __assign__ methods might return the reference to other
rather than self, if they "detected" the need..

Drew



From puccio_13 at yahoo.it  Wed Apr 16 10:49:36 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Wed, 16 Apr 2003 16:49:36 +0200
Subject: reading Unix mailbox
References: <hZbna.17434$DN.480144@tornado.fastwebnet.it> <Z5dna.110590$It5.98167@news2.central.cox.net>
Message-ID: <xdena.17614$DN.485153@tornado.fastwebnet.it>

> 
> I seem to rememeber that a recent version (probably 2.2) modified the
> mailbox objects so you could provide an optional (callable) factory
> argument when creating them, and this would then be called when the
> mailbox instance needed to create a message. The default is
> rfc822.Message, hence the behavior you are seeing.
> 
> Does this help any?

Yes, almost..I tried with this simple:

import email.Message
import mailbox

mbox_file = file('../spam1.txt')
mbox_obj = mailbox.PortableUnixMailbox(mbox_file, email.Message)
msg = mbox_obj.next()           # <-----what's wrong with this??

But I got:

Traceback (most recent call last):
  File "testMailBox.py", line 6, in ?
    msg = mbox_obj.next()
  File "/usr/lib/python2.3/mailbox.py", line 35, in next
    return self.factory(_Subfile(self.fp, start, stop))
TypeError: 'module' object is not callable

I am not a Python guru as you see...
thanks


> 
> regards
> --
> Steve Holden                                  http://www.holdenweb.com/
> How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
> Python Web Programming                 http://pydish.holdenweb.com/pwp/

-- 
bye
Alessio Pace



From mathias-usenet at valpo.de  Tue Apr  1 04:32:44 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 1 Apr 2003 09:32:44 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift>
Message-ID: <b6bmbs$3depo$1@ID-15245.news.dfncis.de>

Julian Tibble <chasm at rift.sytes.net> wrote:
> I would also be interested to see the code, because I recently
> wrote a similar program in C.  My program counts the number of
> occurences of each word in a file (where a word is a positive
> number of characters from a-z --- uppercase letters are
> converted to lowercase). 

As usual the more a program is IO bound the lesser is the difference 
between the used programming language or the abstraction level. And 
I'm afraid Alex didn't used IO redirection as you did.  

And for IO the C++ implementation used in the g++ has an important 
"feature": to be able to mixup any sequence of C and C++ io (ie. cout 
<< and printf) it calls a write for each single character. Disabling 
this usually gives a significant performance boost (look at 
std::ios::sync_with_stdio(false)). 

Mathias



From jjl at pobox.com  Fri Apr  4 18:58:35 2003
From: jjl at pobox.com (John J Lee)
Date: Sat, 5 Apr 2003 00:58:35 +0100
Subject: cgi.py -- anybody using it?
Message-ID: <Pine.LNX.4.44.0304050051160.1193-100000@alice>

I just tried testing some HTML form upload code using cgi.py, and
discovered that it wants lower-case HTTP headers :-/

This is broken (unless I missed something in RFC 2616).  So, does anybody
actually use this module?

Maybe MSIE etc. always send lower-case headers, I don't remember, but it
seems odd that cgi.py gets this wrong after all this time.  I remember
filing a similar bug about another standard library web module, don't
remember which one now.

I'm in two minds about the 'batteries included' principle, but I suppose
there is no obvious best trade-off point between convenience and the
problems of maintainability, obsolescence, etc...


John




From aleax at aleax.it  Thu Apr 24 05:50:23 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 09:50:23 GMT
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <NKxpa.85054$T34.2112670@news2.tin.it> <698f09f8.0304231141.5d458eb0@posting.google.com> <vYDpa.50997$LB6.1152215@news1.tin.it> <b88akc$31dm$1@ns.felk.cvut.cz>
Message-ID: <zrOpa.90172$T34.2242021@news2.tin.it>

Frantisek Fuka wrote:

> Many thanks to everyone. Can I get the "new" string functions
> explanation somewhere? I have Python 2.2.2 but the current documentation
> only seems to mention the old-style STRING module...

http://www.python.org/doc/current/lib/string-methods.html


Alex




From pwatson at redlinec.com  Mon Apr 28 14:47:28 2003
From: pwatson at redlinec.com (Paul Watson)
Date: Mon, 28 Apr 2003 13:47:28 -0500
Subject: Class attribute or instance attribute?
Message-ID: <vaqtlsqbi52j93@corp.supernews.com>

When is an attribute a class attribute or an instance attribute?

In the following code, is the attribute (variable/member) a class attribute
or an instance attribute?  I thought that it was an instance attribute and
that the class attribute was created in __init__().

My biggest question is why self.v has a value of "Consequences" in the
_init__() of f2?  Is self.v not an instance variable?

pwatson [ /home/pwatson/src/python/c ] 270
$ cat d.py
#! /usr/bin/env python

class whatisit:
  v = "Ambiguous"

  def __init__(self):
    print '=== __init__', self.v, whatisit.v
    self.v = "Truth"
    whatisit.v = "Consequences"

  def showem(self):
    print '=== showem  ', self.v, whatisit.v

if (__name__ == "__main__"):
  f1 = whatisit()
  f1.showem()
  f2 = whatisit()
  f2.showem()
pwatson [ /home/pwatson/src/python/c ] 271
$ ./d.py
=== __init__ Ambiguous Ambiguous
=== showem   Truth Consequences
=== __init__ Consequences Consequences
=== showem   Truth Consequences





From marcus at infa.abo.fi  Wed Apr  9 06:16:03 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 9 Apr 2003 12:16:03 +0200
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <mailman.1049778853.27420.python-list@python.org> <slrnb9580p.7og.marcus@infa.abo.fi> <mailman.1049825355.20979.python-list@python.org>
Message-ID: <slrnb97p3l.bap.marcus@infa.abo.fi>

On Tue, 8 Apr 2003 14:08:04 -0400, Jp Calderone <exarkun at intarweb.us> wrote:
>  This line:
>
>    "from foobar import t2 as t1"
>
>  causes "foobar" to be imported, adding "foobar" to sys.modules, then
>causes "foobar.t2" to be imported, adding "foobar.t2" to sys.modules, then
>binds the local "t1" to that module object.
>
>  This line:
>
>    "import t2"
>
>  causes "t2" to be imported.  Before it is -actually- imported, sys.modules
>is checked for the string "t2" to see if the module has already been loaded. 
>If it has, that module object is returned, but if it hasn't (and as far as
>Python knows, it hasn't, only "foobar.t2" has been imported, and that is
>clearly a different module), "t2" is added to sys.modules and the local "t2"
>is bound to the new module object.

OK, this certainly explains very well "how" things happen. Thank you.

>  Hopefully you see now that there are two copies of A, because there are
>two copies of the module.

Yes. Although I find it non-intuitive; python could have determined
it's the same file by checking major/minor and inode fields of the
file (on unix).  Hmm.. perhaps this (or something similar) isn't
possible on all operating systems?

>> This leads (IMHO) to the fact that one _never_ should use "import X",
>> (witness isinstance(t1.A(), t2.A) returning 0...) everything should go via
>> PYTHONPATH. Which in turn means that you have to set PYTHONPATH for all
>> programs. ?
>  One could look at it that way.  One could look at it another way, too:
>don't mix imports in such a way that the same module may be referred to by
>more than one name.

I fail to see how this feature would help anybody. Where's the benefit
of being able to import the same module into two different names
with "import Y" and "from X import Y" statements, and _relying_ on them
being different copies? Brittle at best, and most likely error-prone.
I would understand it if a local "__import__" would create a separate
namespace.

Regards,
Marcus

-- 
Marcus Alanen
maalanen at abo.fi



From staschuk at telusplanet.net  Fri Apr 18 12:46:50 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 18 Apr 2003 10:46:50 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <1050640161.271511@yasure>; from donn@drizzle.com on Fri, Apr 18, 2003 at 04:29:28AM -0000
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <mailman.1050634170.24420.python-list@python.org> <1050640161.271511@yasure>
Message-ID: <20030418104650.B709@tibia.amotlpaa.bogus>

Quoth Donn Cave:
> Quoth Steven Taschuk <staschuk at telusplanet.net>:
  [str : repr :: data : object ?]
> [...] But we're definitely on the
> same page here.  Your objection is probably reasonable, though
> arguable, but the thing is, we're on ground here that we can
> reason about.  It may be underspecified, but it isn't the kind
> of thing I call fuzzy.  I agree that it's worth a brief sentence
> or two, but at some point one has to leave it to the programmer's
> judgement what's actually going to be desirable.

Fair enough.  (I won't insist on "fuzzy"'s applicability in this
case.  Whether we call it fuzzy or not is, in the end, of merely
ritual significance.)

I certainly agree that it is highly desirable to have grounds on
which to reason about the quality of a given implementation of
__str__ or __repr__.  I will stipulate that the documentation as
it stands doesn't provide such grounds.

I also agree (as does the timbot, to judge by his posts in this
thread) that the matter must largely be left to the judgement of
the programmer; the best the documentation can do is state the
general principles and give examples of them at work.

  [draft revision of __str__ docs]
> |     This differs from __repr__() in that its primary audience is not a
> |     programmer inspecting the object during debugging.  What the
> |     primary audience is depends on the object: the __str__
> |     representation might, for example, be designed for display to a
> |     user of the application, or for processing by other code which
> |     requires a string.  Thus this representation might be less
> |     detailed, less precise, or in some way more convenient than that
> |     computed by __repr__.
  [...]
> The __str__ section could use more work.  [...]

I agree.  I've tried to describe it in terms of intended audience,
but am coming to doubt that that's a good way of describing
__str__.  (Though it works quite nicely for __repr__, imho.)

I wonder whether __str__ allows any description less vague than
"stringification as appropriate to the object".  The appropriate
stringification varies a great deal from object to object: numbers
and datetimes call for end-user-friendly representations;
rfc822.Message calls for the textual representation as per the
RFC; a class representing a blurb of bytecode might str simply
into the string of bytes.  And so forth.  They're all more
convenient than the corresponding reprs, but more convenient for
very different purposes.

> [...] I really think the last sentence is as
> wrong as it is right - sure it might be, or it might be more detailed
> etc.  Cf. rfc822.Message.__str__ - complete contents of the header.
> In what way is that less detailed or precise, or more convenient?
> It is however exactly the ``data''.

Yes, I suppose I have to agree that __str__ is sometimes more
detailed than __repr__.  (There's another nice example in
compiler.pyassem.Block, incidentally.)  I'll go along with
dropping that last sentence.

-- 
Steven Taschuk                                                   w_w
staschuk at telusplanet.net                                      ,-= U
                                                               1 1




From ramen at lackingtalent.com  Thu Apr 24 04:55:03 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Thu, 24 Apr 2003 08:55:03 -0000
Subject: python -c "help('while')"
References: <878yu0h60i.fsf@jidanni.org>
Message-ID: <slrnbaf9k9.gb9.ramen@lackingtalent.com>

In article <878yu0h60i.fsf at jidanni.org>, Dan Jacobson wrote:
> Is this the easiest way to get help on things like "while" in one
> shell command, outside the interpreter: 
> $ python -c "help('while')"

Well, maybe someone has a better answer, but you could always make yourself
a little helper script (or use aliases in your shell, even):

$ cat > pyhelp
#!/usr/bin/env python
import sys
help(sys.argv[1])
^D
$ chmod +x pyhelp
$ ./pyhelp while

   7.2 The while statement...
   
Dave



From s_p_a_w_k at hotmail.com  Thu Apr 17 05:09:58 2003
From: s_p_a_w_k at hotmail.com (romain chabot)
Date: Thu, 17 Apr 2003 11:09:58 +0200
Subject: PDFLIB dll
Message-ID: <Sea1-DAV17QYrVjYKtW00000dbb@hotmail.com>

Hi,

I'm trying to insert an existing pdf document into a new one, created with PDFLIB.
I tried the quickreference.py file given with the library, but an error occurs when interpreting : 
 
File "E:\PDFlib-5.0.0-Windows\bind\python\quickreference.py", line 8, in ?
from pdflib_py import *
ImportError: DLL load failed

I have put the libpdf_py.dll in the python dlls folder. I don't understand the situation
Sorry for this stupid question.

Thanks a lot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030417/04b4c0af/attachment.html>

From martin at v.loewis.de  Tue Apr 29 16:00:29 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 29 Apr 2003 22:00:29 +0200
Subject: zipfile with umlaut in filename
References: <3ea9a4bf$1_1@news.vo.lu> <b8caue$g65$05$3@news.t-online.com> <3eaa52b1_1@news.vo.lu> <m3fzo28ip9.fsf@mira.informatik.hu-berlin.de> <3eaecb38_2@news.vo.lu>
Message-ID: <m365ox6plu.fsf@mira.informatik.hu-berlin.de>

Patrick Useldinger <uselpa at myrealbox.com> writes:

> I don't, WinZip does ;-)
> And being compatible to WinZip doesn't seem so bad to me.

I'm strongly opposed to hard-coding this into zipfile.py. This
supports latin characters only, and relies on undocumented Winzip
features.

I think WinZip Computing should make an official statement on how they
want to support Unicode in WinZip; if they do so, I'll happily
incorporate their approach into zipfile.py.

[IOW: I trust their authority if it is paired with expertise]

Regards,
Martin




From pc451 at yahoo.com  Thu Apr 17 23:11:57 2003
From: pc451 at yahoo.com (Peter Clark)
Date: 17 Apr 2003 20:11:57 -0700
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com> <3e9daad2$0$49115$e4fe514c@news.xs4all.nl> <f7199550.0304161732.48c75c42@posting.google.com> <mailman.1050548247.26781.python-list@python.org> <f7199550.0304170555.77773f8a@posting.google.com> <mailman.1050606359.1431.python-list@python.org>
Message-ID: <f7199550.0304171911.2ac60fa6@posting.google.com>

Steven Taschuk <staschuk at telusplanet.net> wrote in message news:<mailman.1050606359.1431.python-list at python.org>...
> Quoth Peter Clark:
> > Steven Taschuk <staschuk at telusplanet.net> wrote in message news:<mailman.1050548247.26781.python-list at python.org>...

> And scale is a Unicode string, right?
    Only because the XML document has no specified encoding, so it
defaults to UTF-8, yes. But all the text is straight ASCII, except of
course for the inclusion of the degree symbol.

> I'd suggest doing all your string assembly with Unicode strings,
> and only encoding to ISO-8859-1 at the output stage.  Mixing the
> two kinds of strings is asking for trouble, as you have seen.
    You're right, of course. The list is sent to a simple function
that prints it to stdout; the following worked fine, and handles any
integers or floats that might otherwise cause an attribute error:

def printlist(w):
  for item in w:
    try:
      print item.encode('latin-1')
    except:
      print item

    Which is much nicer than having to mess with the strings as they
are added to the list. Thanks.



From staschuk at telusplanet.net  Fri Apr 25 10:28:41 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 08:28:41 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
In-Reply-To: <b8afri$oqt$1@slb4.atl.mindspring.net>; from adalke@mindspring.com on Thu, Apr 24, 2003 at 11:11:26PM -0600
References: <mailman.1051023275.29338.python-list@python.org> <b8420j$629$1@slb3.atl.mindspring.net> <mailman.1051043856.10118.python-list@python.org> <b86gim$60r$1@slb5.atl.mindspring.net> <mailman.1051127503.25664.python-list@python.org> <b87il2$3rf$1@slb6.atl.mindspring.net> <mailman.1051211247.5278.python-list@python.org> <b8afri$oqt$1@slb4.atl.mindspring.net>
Message-ID: <20030425082841.A261@tibia.amotlpaa.bogus>

Quoth Andrew Dalke:
> Steven Taschuk:
> > Hm.  I'm not sure I understand what would make such a language not
> > Python.
> 
> I believe apologies are due.  It appears I misread flippancy into
> your response, so this one of mine is now more serious.

On the contrary!  No apology necessary at all, and you correctly
detected flippancy.  (Also see sig.)  I'm just valiantly trying to
bat the nit back.

  [my claim that generators were a bigger change than silly __and__]
> I don't think generators are that large a change.  Pre-generators I
> wrote a lot of code which looked like this
  [iterator example]
> Conceptionally then, I see yield (on top of the iterator protocol)
> as a great way to simplify building an existing idiom.  [...]
> But Python the language did let me get that feel, even in the
> 1.4 days, and adding generators to Python has not affected
> that feel.
  [comparison of silly __and__'s arguments and code blocks]
> Were Python to have code blocks, then it should be applied
> more generally to Python code.  Consider the if/else PEP of
> a couple months ago.  With code blocks, the following would
> be possible
> 
> ifelse(a != 0, 1/a, None)
> 
> which would pass the 1/a and None as independent code
> blocks, to be executed based on the evaluation of a != 0.

Ah, you argue well.

I agree that the canonical use of generators -- to write iterators
-- is merely an improvement of a previously existing idiom, and
does not in itself change the feeling of the language.

But generators' suspend/resume powers let them be used far more
generally.  PEP 288 alludes to some such uses; see also David
Mertz's developerworks articles on using generators to implement
coroutines.

(I have occasionally thought of using Mertz's technique for
*every* kind of call, not merely coroutine calls.  If (say, by
bytecode translation) we turned every call
    foo(*args, **kwargs)
into
    yield foo, args, kwargs
every function would become a generator.  This would allow the
call semantics to be implemented in Python, in what Mertz calls a
scheduler.  You could, for example, build microthreads this way;
if generators could be copied, you could even build continuations!
At least I think so; this is all just a vague idea I have, not
supported by any detailed knowledge of Python internals.  There
are many details to be worked out, e.g., how return values are
provided to the "calling" function, if the call "returns".  (PEP
288 would provide one mechanism.)  There are also efficiency
questions, since e.g. every use of + is potentially a call.)

In other words, the suspend/resume powers of generators open up
all sorts of interesting possibilities which would be difficult to
implement in pre-generator Pythons, and imho these capabilities
can make for significant changes to the feel of the language.

I feel that generators, conceived this broadly, are just as
significant a change as code blocks would be.  (I'm backing down
from my previous claim that they are *more* of a change than the
code blocks implied by the arguments to my silly __and__; now I
claim only that they're of about the same magnitude.)

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From naoaqualis at aol.com  Thu Apr 24 10:22:26 2003
From: naoaqualis at aol.com (alex)
Date: 24 Apr 2003 07:22:26 -0700
Subject: serveur TFTP
Message-ID: <e4d8f90f.0304240622.64cf7f8e@posting.google.com>

j'ai recup?r? une librairie TFTP ecrit en python mais j'avoue avoir un
peu de mal ? l'utiliser.comment faire un serveur TFTP avec? j'essai
d'identifier les methodes que je pourrais appliquer a un objet
mon_serveur par exemple :

import tftplib
mon_serveur = tftplib.TFTPServer()

...et apr?s je suis un peu perdu.

voici la librairie tftplib.py:

#!/usr/bin/env python
#
#
# Standard imports
#
import string
import struct, socket, select
import thread

#
# Default variables
#
TFTP_PORT = 69


#
# TFTP Errors
#
class TFTPError(Exception):
    pass


#
# A class for a TFTP Connection
#
class TFTPConnection:

    RRQ  = 1
    WRQ  = 2
    DATA = 3
    ACK  = 4
    ERR  = 5
    HDRSIZE = 4  # number of bytes for OPCODE and BLOCK in header

    def __init__(self, host="", port=0,blocksize=512, timeout=2.0,
retry=5 ):
        self.host = host
        self.port = port
        self.blocksize = blocksize
        self.timeout   = timeout
        self.retry     = retry

        self.client_addr = None
        self.sock        = None
        self.active      = 0
        self.blockNumber = 0
        self.lastpkt     = ""

        self.mode        = ""
        self.filename    = ""
        self.file        = None

        self.bind(host, port)
    # end __init__

    def bind(self, host="", port=0):
        sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
        self.sock = sock
        if host or port:
                sock.bind(host, port)
    # end start

    def send(self, pkt=""):
        self.sock.sendto(pkt, self.client_addr)
        self.lastpkt = pkt
    # end send

    def recv(self):
        sock        = self.sock
        F           = sock.fileno()
        client_addr = self.client_addr
        timeout     = self.timeout
        retry       = self.retry

        while retry:
            r,w,e = select.select( [F], [], [F], timeout)
            if not r:
                # We timed out -- retransmit
                retry = retry - 1
                self.retransmit()
            else:
                # Read data packet
                pktsize = self.blocksize + self.HDRSIZE
                data, addr = sock.recvfrom(pktsize)
                if addr == client_addr:
                    break
                else:
                        raise TFTPError(4, "Transfer timed out")
        # end while

        return self.parse(data)
    # end recv

    def parse(self, data, unpack=struct.unpack):
        buf = buffer(data)
        pkt = {}
        opcode = pkt["opcode"] = unpack("!h", buf[:2])[0]
        if ( opcode == self.RRQ ) or ( opcode == self.WRQ ):
            filename, mode, junk  = string.split(data[2:], "\000")
            pkt["filename"] = filename
            pkt["mode"]     = mode
        elif opcode == self.DATA:
            block  = pkt["block"] = unpack("!h", buf[2:4])[0]
            data   = pkt["data"]  = buf[4:]
        elif opcode == self.ACK:
            block  = pkt["block"] = unpack("!h", buf[2:4])[0]
        elif opcode == self.ERR:
            errnum = pkt["errnum"] = unpack("!h", buf[2:4])[0]
            errtxt = pkt["errtxt"] = buf[4:-1]
        else:
            raise TFTPError(4, "Unknown packet type")

        return pkt
    # end recv

    def retransmit(self):
        self.sock.sendto(self.lastpkt, self.client_addr)
        return
    # end retransmit

    def connect(self, addr, data):
        self.client_addr = addr
        RRQ  = self.RRQ
        WRQ  = self.WRQ
        DATA = self.DATA
        ACK  = self.ACK
        ERR  = self.ERR

        try:
                pkt    = self.parse(data)
                opcode = pkt["opcode"]
                if opcode not in (RRQ, WRQ):
                        raise TFTPError(4, "Bad request")

            # Start lock-step transfer
                self.active = 1
                if opcode == RRQ:
                        self.handleRRQ(pkt)
                else:
                        self.handleWRQ(pkt)

            # Loop until done
                while self.active:
                        pkt = self.recv()
                        opcode = pkt["opcode"]
                        if opcode == DATA:
                                self.recvData(pkt)
                        elif opcode == ACK:
                                self.recvAck(pkt)
                        elif opcode == ERR:
                                self.recvErr(pkt)
                        else:
                                raise TFTPError(5, "Invalid opcode")
        except TFTPError, detail:
                self.sendError( detail[0], detail[1] )

        return
    # end connection


    def recvAck(self, pkt):
        if pkt["block"] == self.blockNumber:
            # We received the correct ACK
            self.handleACK(pkt)
        return
    # end recvAck

    def recvData(self, pkt):
        if pkt["block"] == self.blockNumber:
            # We received the correct DATA packet
            self.active = ( self.blocksize == len(pkt["data"]) )
            self.handleDATA(pkt)
        return
    # end recvError

    def recvErr(self, pkt):
        self.handleERR(pkt)
        self.retransmit()
    # end recvErr

    def sendData(self, data, pack=struct.pack):
        blocksize = self.blocksize
        block     = self.blockNumber = self.blockNumber + 1
        lendata   = len(data)
        format = "!hh%ds" % lendata
        pkt = pack(format, self.DATA, block, data)
        self.send(pkt)
        self.active  = (len(data) == blocksize)
    # end sendData

    def sendAck(self, pack=struct.pack):
        block            = self.blockNumber
        self.blockNumber = self.blockNumber + 1
        format = "!hh"
        pkt = pack(format, self.ACK, block)
        self.send(pkt)
    # end sendAck

    def sendError(self, errnum, errtext, pack=struct.pack):
        errtext = errtext + "\000"
        format = "!hh%ds" % len(errtext)
        outdata = pack(format, self.ERR, errnum, errtext)
        self.sock.sendto(outdata, self.client_addr)
        return
    # end sendError


    #
    #
    # Override these handle* methods as needed
    #
    #
    def handleRRQ(self, pkt):
        filename  = pkt["filename"]
        mode      = pkt["mode"]
        self.file = self.readRequest(filename, mode)
        self.sendData( self.file.read(self.blocksize) )
        return
    # end readFile

    def handleWRQ(self, pkt):
        filename  = pkt["filename"]
        mode      = pkt["mode"]
        self.file = self.writeRequest(filename, mode)
        self.sendAck()
        return
    # end writeFile

    def handleACK(self, pkt):
        if self.active:
            self.sendData( self.file.read(self.blocksize) )
        return
    # end handle ACK

    def handleDATA(self, pkt):
        self.sendAck()
        data = pkt["data"]
        self.file.write( data )
    # end handleDATA

    def handleERR(self, pkt):
        print pkt["errtxt"]
        return
    # end handleERR

    #
    # You should definitely override these
    #
    def readRequest(self, filename, mode):
        from StringIO import StringIO
        return StringIO("")
    # end readRequest

    def writeRequest(self, filename, mode):
        from StringIO import StringIO
        return StringIO()
    # end writeRequest

# end class TFTPConnection


#
# Simple TFTP Server
#    Each connection is handled in its own thread.
#
class TFTPServer:

    """TFTP Server
    Implements a threaded TFTP Server.  Each request is handled
    in its own thread
    """

    def __init__(self, host="", port=TFTP_PORT,conn=TFTPConnection,
srcports=[] ):
        self.host = host
        self.port = port
        self.conn = conn
        self.srcports = srcports

        self.sock = None
        self.bind(host, port)
    # end __init__

    def bind(self, host, port):
        sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
        self.sock = sock
        sock.bind(host, port)
    # end start

    def forever(self):
        while 1:
                data, addr = self.sock.recvfrom(516)
                self.handle(addr, data)
    # end serve_forever

    def handle(self, addr, data):
        if self.srcports:
                nextport = self.srcports.pop(0)
                self.srcports.append( nextport )
                T = self.conn( self.host, nextport )
        else:
            T = self.conn( self.host )
        thread.start_new_thread( T.connect, (addr, data) )
        return
    # end handle

# end class TFTPServer


if __name__ == "__main__":
    import sys
    from StringIO import StringIO

    #
    # Subclass to create our own TFTP Connection object
    #
    class MyTFTP( TFTPConnection ):

        def readRequest(self, filename, mode):
            randomstring = "Here is a sample string"
            return StringIO( randomstring )

        def writeRequest(self, filename, mode):
            return StringIO()
    # end class

    port = 3000 + TFTP_PORT
    if sys.argv[1:]:
        port = string.atoi(sys.argv[1])

    try:
        serv = TFTPServer( "", port, conn=MyTFTP )
        serv.forever()
    except KeyboardInterrupt, SystemExit:
        pass

d'avance merci.
Alex



From cben at techunix.technion.ac.il  Fri Apr 18 04:34:14 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Fri, 18 Apr 2003 11:34:14 +0300 (IDT)
Subject: Python and Schools
In-Reply-To: <mailman.1050604354.15055.python-list@python.org>
Message-ID: <20030418083414.0D3CD222E0@techunix.technion.ac.il>

In comp.lang.python, you wrote:
>[David Eppstein]
>> ...
>> A simple example would be an algorithm that takes as input a sequence of
>> values that are all (for whatever reason) guaranteed to be integers in
>> range(n**2), and tests whether any two of them are the same:
>>
>> def anysame(seq):
>>     allocate array A, of dimension n**2, WITHOUT INITIALIZING IT
>>         (is this possible in pure python?)
>
>It shouldn't be possible:  if you find a way to do it, it would be
>considered a security bug.  No memory is really uninitialized -- if Python
>gave you read access to n**2 bytes in constant time, it would be revealing
>to you the state left behind by some other function (or user, or process).
>If you created a file with at least n**2 bytes beforehand, then you could
>mmap the file in constant time, and that acts like an array object
>initialized to whatever bytes happen to be sitting in the file.  But the O()
>behavior of access to an mmap'ed region depends on the platform
>implementation of mmap'ed regions.
>
It's impossible in plain portable C but an operating system can do this using
memory mapping hardware.  It's possible to mark the required range of memory
pages to cause an exception and be initialized [to zeros] only when they are
accessed.  Since the algorithm only accesses a few pages, this should work
fine.  This reveals a little clarification to the original claim: the
algorithm doesn't use O(N**2) memory, it eats O(N**2) of your *address space*.

In fact, that's what mmaping /dev/zero probably does on most platforms... 
There is a catch: for large enough N, you need to mark O(N**2) pages in this
way, so it's O(N**2) time to allocate again :-(.  This can be overcome by
hierarchical page tables.  To be precise, you would need an arbitrarily deep
hierarchy rather than one of fixed depth that real computers tend to have
<wink>.  Note that such a hierarchy is actually a way of implementing a
dictionary from digital trees :-).

You would actually need infinite memory to have an unbounded hierarchy.  Which
exposes a small catch with application of asymptotic analysis once you
approach limits of address space, integer size, etc.  No algorithm takes
O(f(N)) when N approaches infinity - instead it just runs out of resources
:-)!  So many algorithms can be pervertly claimed to be O(1) on a given
architecture.  For example one can sort any array of 16 bit integers in O(1)
by just allocating a O(2**16) array counting how many times each number
appeared and then scanning the array.  Of course this is worse than any
sensible algorithm one would write - because of the huge factor.  In truth
this is not O(1) - it's O(M) where M is your address space or "integer space"
size.  But omitting such considerations can lead to funny "paradoxes"...

P.S. Cute algorithm indeed!

Beni Cherniavsky <cben at tx.technion.ac.il>




From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr 11 06:20:52 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 11 Apr 2003 10:20:52 GMT
Subject: Python & Remote Access Database
References: <a6b6f195.0304090408.21003c3a@posting.google.com> <kjYka.2914$kd1.2980829@newssrv26.news.prodigy.com> <a6b6f195.0304100351.54a1bef9@posting.google.com> <7enhm-os2.ln1@beastie.ix.netcom.com>
Message-ID: <slrnb9d5o2.1m0.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <7enhm-os2.ln1 at beastie.ix.netcom.com>, Dennis Lee Bieber wrote:
<snip>
>         What type of network?
> 
>         TCP/IP? You could run FTP/Web servers and make the files available
> via them. You could run NFS mounts, and share directories.
> 
>         NetBUIE... Install Samba on Linux and share Windows drives... (and 
> printers)

That's SMB you're thinking of, which is a whole different layer.  SMB used
to require NetBIOS, and NetBEUI used to be part of NetBIOS.  However,
NetBIOS was long ago split so that it could be used with TCP/IP or Novell's
IPX/SPX.  (In fact, SMB can be separated from NetBIOS too, starting with
Windows 2000.)  Some layered protocol diagrams for you:

 Before         After
+---+---+ +-------+-------+
|SMB|NMB| |  SMB  |  NMB  |
+---+---+ +-------+-------+
|       | |    NetBIOS    |
|       | +-------+---+---+
|NetBIOS| |       |TCP|SPX|
|       | |NetBEUI+---+---+
|       | |       |IP |IPX|
+-------+ +-------+---+---+

(NMB is the NetBIOS Master Browser service that will, if you're lucky,
produce a list of computers in the Network Neighbourhood view, but may
fail in ways that are hard to debug.  NetBIOS is basically a crappy
peer-to-peer system.)



From gh at ghaering.de  Tue Apr 15 07:51:27 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Tue, 15 Apr 2003 13:51:27 +0200
Subject: newbie: raise socket.error, msg in smtplib
In-Reply-To: <d3c9c04.0304150344.cf0f33c@posting.google.com>
References: <d3c9c04.0304150344.cf0f33c@posting.google.com>
Message-ID: <3E9BF23F.1010303@ghaering.de>

Mark Carter wrote:
> I'm getting
> 
> Traceback (most recent call last):
>   File "C:\My Documents\myemail.py", line 26, in ?
>     s.connect()
>   File "C:\PYTHON22\lib\smtplib.py", line 283, in connect
>     raise socket.error, msg
> error: (10061, 'Connection refused')
> [...]
> s = smtplib.SMTP()

smtplib.SMTP is defined similar to:

class SMTP:
	def __init__(self, host="localhost", port=25)
		[...]

So if you omit host and port, it tries to connect to the computer you're 
running the program from (127.0.0.1, aka localhost on the default SMTP 
port - 25).

It seems you're not running an SMTP server on your box, so just specify 
a SMTP server you have access to.

-- Gerhard





From mwh at python.net  Wed Apr 30 04:59:10 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 30 Apr 2003 08:59:10 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304251030.30fd1f94@posting.google.com> <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk> <mailman.1051635832.26364.python-list@python.org>
Message-ID: <7h3el3kgy4j.fsf@pc150.maths.bris.ac.uk>

Steven Taschuk <staschuk at telusplanet.net> writes:

> Quoth Michael Hudson:
> > mis6 at pitt.edu (Michele Simionato) writes:
>   [...]
> > > 2) why you say this makes classes unhashable ?
> > 
> > Well, I thought it would.  I was wrong:
>   [...]
> > Then you get this kind of weirdness:
> > 
> > />> class C(type):
> > |..     def __eq__(self, other):
> > |..         return 1
> > \__ 
> 
> Can't the instances of this metaclass be made sensible for use in
> dicts simply with
>     def __hash__(self):
>         return hash(C)
> for example?  (To ensure that if x == y then hash(x) == hash(y),
> as desired.)  

Yes!  The point I was making was that *not* doing that *doesn't* lead
to error messages.

Hang on...

/>> class C(object):
|..  def __eq__(self, other):
|..   return 1
\__ 
->> hash(C())
137007476
/>> class C:
|..  def __eq__(self, other):
|..   return 1
\__ 
->> hash(C())
Traceback (most recent call last):
  File "<input>", line 1, in ?
TypeError: unhashable instance

Hmm.  I wonder if this is deliberate...

Cheers,
M.

-- 
  MARVIN:  Do you want me to sit in a corner and rust, or just fall
           apart where I'm standing?
                    -- The Hitch-Hikers Guide to the Galaxy, Episode 2



From scott_list at mischko.com  Tue Apr 15 16:10:11 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Tue, 15 Apr 2003 13:10:11 -0700
Subject: O'Reilly Python Books
In-Reply-To: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <200304151310.11490.scott_list@mischko.com>

On Tuesday 15 April 2003 12:37, W S wrote:
> Hello,
>
> I find the material, as presented, in the book Programming Python
> frustrating and extremely difficult to understand.
>
> Are the nutshell and cookbook O'Reilly Python books easier to read?
>
> I'm fairly new to programming but trying to learn. I have read Learning
> Python, Core Python, Python Bible and a bunch of perl books

I read http://www.ibiblio.org/obp/thinkCSpy/ and went to Python Essential 
Reference (New Riders).

I'm still working on OOP so I read the Tao of Objects.  It's dated and 
examples are in C++ and TurboPascal but I found the information useful.

Anyone else have good recommends regarding OOP books, especially focusing on 
Python?

Scott





From gildemere at hotmail.com  Mon Apr 14 05:43:54 2003
From: gildemere at hotmail.com (Marc Floessel)
Date: 14 Apr 2003 02:43:54 -0700
Subject: Calling Python Code Snippets within Python Application
Message-ID: <236da4ed.0304140143.25abc577@posting.google.com>

This is very likely a newbie question, but I'd still appreciate a
pointer. I need to call Python Code snippets (saved in a file/database
etc) within my main Python Application. I would like to do the
following things:

- Code Snippet should not have access to or knowledge of Main
Application variables and methods - ie. each code snippet has its own
private scope
- main application exposes a number of specific methods and variables
("environment") to the snippet that it can use
- a way for the code snippet to return some variables (ie. a success
bool) - possibly using (re-importing back from the snippet) the
variables exposed by the main application

In which direction should I look?

Thanks a lot,
Marc Floessel



From oren-py-l at hishome.net  Wed Apr 23 11:18:45 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Wed, 23 Apr 2003 11:18:45 -0400
Subject: Testing if string contains a substring
In-Reply-To: <b865dq$28gi$1@ns.felk.cvut.cz>
References: <b865dq$28gi$1@ns.felk.cvut.cz>
Message-ID: <20030423151845.GA3000@hishome.net>

On Wed, Apr 23, 2003 at 03:44:57PM +0200, Frantisek Fuka wrote:
> I want a simple test to see if str contains any occurences of substr. I 
> am currently using this:
> 
> import string
> if string.find(str,substr) + 1:
>     ...do something...
> 
> But it somehow looks "ugly" to me. Is there more "standard" way to test 
> this?

if haystack.count(needle):
   ...

"If the count of needles in haystack is nonzero then..."

    Oren




From niessink at serc.nl  Sat Apr 19 17:34:05 2003
From: niessink at serc.nl (Frank Niessink)
Date: Sat, 19 Apr 2003 23:34:05 +0200
Subject: Sorting without transitivity
Message-ID: <20030419213405.GA588@obstler>

Hi all,

[It's been while since I've dealt with transitivity and stuff, so
I hope I got the terminology right]

I have objects that are not transitivily ordered, e.g.:

A > B
A == C
B == C

Now I have a list [A, B, C] and I want to sort it.
The resulting list should have B before A, I don't care where
C is. However, the Python list.sort() method (Python 2.2.2 on cygwin)
doesn't put A and B in the right order.

How can I get the desired behavior? Should I implement my own
sort algorithm? Any pointers?

Thanks in advance, Frank

PS: Below follows a demonstration of my problem:


import unittest

class Point:
    def __init__(self, x, y):
        self.x = x
        self.y = y

    def __repr__(self):
        return 'Point(%d, %d)'%(self.x, self.y)

    def __cmp__(self, other):
        if self.x == other.x and self.y == other.y:
            return 0
        elif self.x <= other.x and self.y <= other.y:
            return -1
        elif self.x >= other.x and self.y >= other.y:
            return 1
        else:
            return 0


class PointTest(unittest.TestCase):
    def setUp(self):
        self.origin = Point(0, 0)

    def testOrigin(self):
        self.assertEqual(self.origin, self.origin)

    def testReallySmaller(self):
        oneOne = Point(1, 1)
        self.failUnless(self.origin < oneOne)

    def testOnlyXSmaller(self):
        oneZero = Point(1, 0)
        self.failUnless(self.origin < oneZero)

    def testOnlyYSmaller(self):
        zeroOne = Point(0, 1)
        self.failUnless(self.origin < zeroOne)

    def testNotSmallerNotBigger(self):
        a = Point(1, 2)
        b = Point(2, 1)
        self.assertEqual(a, b)


class SortTest(unittest.TestCase):

    def testSortPoints(self):
        """ This test fails """
        a = Point(0, 1) # a < b
        b = Point(0, 2) 
        c = Point(1, 0)
        pointList = [b, c, a]
        pointList.sort()
        self.failUnless(pointList.index(a) < pointList.index(b))


if __name__ == '__main__':
    unittest.main()


-- 
The moment became a longer moment, and suddenly it was a very long moment,
so long one could hardly tell where all the time was coming from.
	-- Douglas Adams, 'So long, and thanks for all the fish'




From aleax at aleax.it  Sat Apr 12 14:53:55 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 18:53:55 GMT
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com>
Message-ID: <7hZla.5370$LB6.150543@news1.tin.it>

Chris wrote:

> looked at stuff about Python regular expressions and there is
> something called finditer and findall that might be useful for parsing
> a string, but can't seem to find any documentation on them.

http://www.python.org/doc/current/lib/node99.html

Actually I find sub (passing a function as "repl") more useful than 
findall (finditer is roughly as useful as sub), but sub's documented 
on the same URL, anyway.  Not everybody likes working "by callbacks",
which is what sub + a function lets you do -- findall and finditer
are more direct, and finditer is basically just as powerful as sub,
as it returns match-objects one after the other, while findall just
returns a list of substrings, not quite as powerful -- though surely,
some of the time (when that's all you need), it may be even handier.


Alex




From roy at unice.fr  Fri Apr 25 05:40:06 2003
From: roy at unice.fr (Jean-Paul Roy)
Date: Fri, 25 Apr 2003 11:40:06 +0200
Subject: Char to ASCII and back ?
Message-ID: <roy-CF32FA.11400525042003@malibu.unice.fr>

How do you transform a char (or a one-char string) into the integer 
ascii code and back ? Say to code a message.
The string type is not mutable in Python. Is there something like the 
StringBuffer class in Python whose instances would be mutable ?

Sorry, I'm just beginning in Python...

Thanks,

   -jpr



From aussie2010 at yahoo.com  Wed Apr 16 02:11:51 2003
From: aussie2010 at yahoo.com (Chris)
Date: 15 Apr 2003 23:11:51 -0700
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com> <mailman.1050087439.11600.python-list@python.org> <435d40e0.0304130212.5659bd4d@posting.google.com> <7Bgma.8749$LB6.237526@news1.tin.it>
Message-ID: <435d40e0.0304152211.5a9511d4@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<7Bgma.8749$LB6.237526 at news1.tin.it>...
> Chris wrote:
>    ...
> > Thanks Inyeol. Be nice if they gave a small example of using finditer.
> 
> How small?  E.g., here's how you could rewrite (a simpler version of)
> findall using finditer (warning, untested):
> 
> def findall(are, instring):
>     return [mo.group(0) for mo in are.finditer(instring)]
> 
> Too small?  Then perhaps you could mention a larger one...?
> 
> 
> Alex

I had some small exposure to "iterators" in C++, and I remember them
being used in loops - iterating over the items in a container. The
example you have given appears to create a list from an "iterator"
with no loop. Something called an "iterator" creating a list with no
loop is very perplexing to me. But maybe my problem is I don't
recognize the syntax [ mo.group(0) for ... ]. It looks like a method
call on an object, followed by an incomplete for loop.



From martin at v.loewis.de  Sat Apr  5 03:10:33 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 05 Apr 2003 10:10:33 +0200
Subject: cgi.py -- anybody using it?
References: <mailman.1049502008.18357.python-list@python.org>
Message-ID: <m3vfxt9xfq.fsf@mira.informatik.hu-berlin.de>

Brian Quinlan <brian at sweetapp.com> writes:

> Probably this feature is rarely used and most user agents use lowercase
> headers.

No. The spelling of the header is irrelevant. rfc822.Message
transparently converts them to lower case.

Regards,
Martin




From carroll at tjc.com  Sat Apr  5 14:04:39 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Sat, 05 Apr 2003 11:04:39 -0800
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <m3d6k0rfqp.fsf@telia.com>
Message-ID: <co9u8vch87uimkl36jmrv8kdsliia3u17o@4ax.com>

On Sat, 05 Apr 2003 17:55:25 GMT, Andrew Markebo
<andrew.markebo at telia.com> wrote:

>/ carroll at tjc.com wrote:
>|
>| Looking at nntplib.py, most of the logic to do this is already there:
>| xover calls longcmd to actually issue the command, and longcmd can
>| take an optional File parameter to which getlongresp (which actually
>| retrieves the response) will write its output, if present.
>
>Which format should the xover output be in the file? 

The file contains a series of lines, corresponding to the lines
returned by the NNTP server in response to the XOVER command, as
documented in RFC 2980.  This is the same data that (in the present
implementation) is stored in the "list" tuple, separated by tabs:
"subject, author, date, message-id, references, byte count, and line
count.  Other optional fields may follow line count." (from the RFC).

>Or have I missed something? You seem to have gotten it working?

Yes, it turned out to be a very easy change, primarily due to the
flexible way it was implemented already for body().



From riccardo_togli_questo at _e_questo_sideralis.net  Sat Apr 19 09:32:55 2003
From: riccardo_togli_questo at _e_questo_sideralis.net (Riccardo Galli)
Date: Sat, 19 Apr 2003 15:32:55 +0200
Subject: Simple wxPython Application
References: <7d22d8d6.0304190355.6098714c@posting.google.com>
Message-ID: <pan.2003.04.19.13.32.53.11272@_e_questo_sideralis.net>

On Sat, 19 Apr 2003 04:55:41 -0700, ekranawetter_steyr wrote:

> Hi all,
> 
> 1.Problem
> ---------
> 
> I was not able to run this simple example from the tutorial:
> 
> from wxPython.wx import wxPySimpleApp, wxFrame
>   app = wxPySimpleApp()
>   frame = wxFrame(None, -1, "Hello World")
>   frame.Show(1)
>   app.MainLoop()
> 
> I got a syntax error in line 2
...
> Thanks, Erwin

Indentation is not correct.
Delete first spaces from every line, like this:

from wxPython.wx import wxPySimpleApp, wxFrame
app = wxPySimpleApp()
frame = wxFrame(None, -1, "Hello World")
frame.Show(1)
app.MainLoop()

It works.

Ciao,
Riccardo

-- 
-=Riccardo Galli=-

 _,e.
s~  ``
 ~@.   ideralis Programs
.   ol 
 `**~  http://www.sideralis.net



From gh at ghaering.de  Wed Apr 16 11:59:16 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 16 Apr 2003 17:59:16 +0200
Subject: Python for VB6 programmers (was Re: Complexity of standard Python
 data structures)
In-Reply-To: <Wrena.33096$T34.876856@news2.tin.it>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
	<slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it>
	<slrnb9lsbe.4pa.marcus@infa.abo.fi> <b7ih78$fb2$1@panix1.panix.com>
	<Ajdna.110595$It5.16632@news2.central.cox.net>
	<Wrena.33096$T34.876856@news2.tin.it>
Message-ID: <3E9D7DD4.3040403@ghaering.de>

Alex Martelli wrote:
> Steve Holden wrote:
>    ...
> 
>>You'd do better to call it "Python for VB programmers", expecially since
>>the change to VB6 is so radical that migrating to a new and more sensible
>>language might be an attractive option to some!
> 
> 
> Changes from VB5 to VB6 weren't all that radical.

To put it sarcastically, VB6 is Service Pack 4 to VB5 that you have to 
pay for.

> It was the change
> from VB6 to VB.NET, aka VB7, that was the shocker (dozens of backwards
> incompatible changes -- mostly taking VB's semantics closer to Python's,
> btw, and therefore actually for the better, but still...).

It is technically equivalent to C#, but with a much more clumsier syntax 
and advantages over C# that I would know of.

And btw., just forget about the VB6-to-VB.NET source code converter 
thingie that's included. We tested it for a project once, and after it 
was finished after several hours you'd still have to fix half the code 
manually ;-)

> I thought
> the window of opportunity -- millions of VB6 programmers reluctant to
> move to dotNET -- had more or less closed by now, but Aahz apparently
> has identified that it's still open.

I really have no idea. All I can say is that the world turns slower when 
Big Corporations [tm] are involved: Our customer ($German_Chip_Producer) 
still hasn't .NET as approved technology and some people there still 
prefer VB6 as implementation language for new software (mostly boring 
Windoze GUI frontends to databases).

For databases, Oracle 8i is the latest and greatest, with still some 
Oracle 8 instances and I even had the pleasure of encountering Oracle 7 
clients, which pretty much restricted me to using DCOracle 1.x :-/

Oh, and they're now accelerating their Windows 2000 rollout :-)

You wouldn't even want to know what hardware you'll get as Windows 
terminal servers ...


Personally, I don't see much overlap in the area of application of 
Python and VB6.

VB6 isn't suitable for much more than as an environment to assemble 
pre-made COM components with an easy-to-use GUI designer. I don't see 
what Python solution would come closer than Qt Designer.

 From the Big Corporation viewpoint that I know a little about, one 
argument of some managers there is that they can squeeze hourly rates if 
VB6 is involved, in comparison to, say C++, Java, Centura or PL/SQL.

Of course then there are some stupid companies *cough* who demand the 
same hourly rate no matter how much qualification is needed :-/

-- Gerhard




From achrist at easystreet.com  Tue Apr  8 15:55:09 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Tue, 08 Apr 2003 12:55:09 -0700
Subject: [Q] pychecker memory usage
Message-ID: <3E93291D.D28D20B4@easystreet.com>

I've been using the pychecker program (pychecker.sourceforge.net) to
find problems in code, and it works very nicely.  However, its memory
consumption is breathtaking -- more than 80 Mb to check a little
10k line program I've been writing for a month or so.  This program
uses 17 MB when run with python from the Windows NT command line, but
80 MB to run through pychecker.  Given that I've got a 128 MB machine,
this will likely soon be a problem, as it appears that the memory usage
has increased from about 55 MB a week or so ago as my program has 
grown, and Windows is now giving me messages about low memory as I
run it.  
  
Has anyone else noted any similar problem and found a good way to 
set pychecker options so that the memory usage is reduced much and the
value of the checks not much?  Any other advice?

Thanks for any info,


Al



From lale at fotonation.com  Wed Apr 30 06:57:48 2003
From: lale at fotonation.com (Ilariu Raducan)
Date: Wed, 30 Apr 2003 11:57:48 +0100
Subject: looping through a file
In-Reply-To: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au>
References: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au>
Message-ID: <3EAFAC2C.9020109@fotonation.com>

 From Python Doc:
readline([size])
Read one entire line from the file. A trailing newline character is kept 
in the string2.11 (but may be absent when a file ends with an incomplete 
line). If the size argument is present and non-negative, it is a maximum 
byte count (including the trailing newline) and an incomplete line may 
be returned. An empty string is returned when EOF is hit immediately. 
Note: Unlike stdio's fgets(), the returned string contains null 
characters ('\0') if they occurred in the input.


If your file has no empty lines (not even at  the end) this shoul work
users= {}
username =InFile.readline()
while(len(line)!=0):
	password=InFile.readline()
	users[username]=password
	username=InFile.readline()

Regards,
Lale
Psybar Phreak wrote:
> hi all - ive just started with python (after having done java and c) and am
> having a little bit of trouble with the script.  Im using as a login script.
> 
> There's a users.dat file in the format
> user1
> user1password
> user2
> user2password
> ... etc
> 
> at the moment - the code i work, only checks the first and second lines (ie.
> details of user1).  but it appears that python doesn't have a do... while...
> until loop.
> 
> 
> can anyone help?
> 
> thanks!!
> 
> Ive attached the appropriate part of the code i have working
> 
> -------------------------------
> 
> #!/usr/local/bin/python
> 
> import cgi, string
> 
> form = cgi.FieldStorage()
> 
> if form.has_key('usernameField') and form.has_key('passwordField'):
>          users_username = string.strip(form['usernameField'].value)
>          users_password = string.strip(form['passwordField'].value)
> 
> InFile = open('users.dat', 'r')
> 
> file_username = string.strip(InFile.readline())
> file_password = string.strip(InFile.readline())
> 
> 
> 
> 
> authorised = 0
> 
> while file_username and file_password:
>        if file_username == users_username:
>               if file_password == users_password: # correct
>              authorised = 1
>       break
>               else:                               # login matched but not
> password
>                     file_username = InFile.readline()
>              file_password = InFile.readline()
>        else: # neither match
>               file_username = InFile.readline()
>               file_password = InFile.readline()
> 
> 




From peter at engcorp.com  Sun Apr 20 12:50:58 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 20 Apr 2003 12:50:58 -0400
Subject: printing email parts of type message/external-body??
References: <ZKzoa.602320$sj7.24103068@Flipper>
Message-ID: <3EA2CFF2.1440294E@engcorp.com>

Magnus Therning wrote:
> 
> I am having problems with the following piece of code, would appreciate
> any insight that can be offered!
> 
> ---- test.py ------------
> from email.Message import Message
> 
> meb_message = Message()
> meb_message.add_header('Content-Type', 'message/external-body',
>     access_type='local-file')
> meb_message.set_payload('name="test"\n')
> print meb_message
> -------------------------
> magnus at die:[test].$ python test.py
> 
> Traceback (most recent call last):
>    g.flatten(msg.get_payload(0), unixfrom=False)
>  File "/usr/lib/python2.2/email/Message.py", line 179, in get_payload
>    raise TypeError, i
> TypeError: 0

Almost looks like a bug.  According to email.Message's documentation,
a TypeError will be raised by get_payload() if an index parameter is
provided but the payload is a string, as in this case.

I would think email.Message.as_string() should be catching that case
and handling it differently.

-Peter



From tjreedy at udel.edu  Mon Apr  7 21:08:52 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Mon, 7 Apr 2003 21:08:52 -0400
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com>
Message-ID: <rB2dnW1Ih5rzvA-jXTWcpg@comcast.com>

"Tim Daneliuk" <tundra at tundraware.com> wrote in message
news:fjss6b.1021.ln at boundary.tundraware.com...
> OK, I don't want to resurrect another interminable discourse in
ints,
> floats, and how they are/ought/can be handled in the language. But,
I
> have a sort of theoretical mathematical question which the whole
> business brought to mind.

> As I understand it, integers and floats are distinct mathematical
> entities. A colleague of mine claims, that insfar as we use them in
> computing, ints are merely a proper subset of floats.

These are two different metaphysical (and quasi-religious) viewpoints.
The entanglement of this issue with that of Python's representation of
int division contributed somewhat unnecessary 'heat' to the int
division debate.  (And its removal help lead to its resolution.)

In my view, counts (0, 1, 2, ...) are the basic numbers -- whether
considered to be primitive (axiomatic) or derived (from functions,
sets, or logic values taken as protonumbers).  They are the possible
answers to the question "How many (discrete units)?"  Integers are
pairs (differences) of counts, just as rationals are pairs (ratios) of
integers, and complex Xs are pairs of Xs (with yet a third set of
operation definitions).  To me, the fact that the parent set can be
matched with a subset of the daughter set does not make it identical
to or 'merely' that subset.

I am aware that in set theory, isomorphism between sets is usually
considered to mean identity.  That is why there is (considered to be)
only 'one' empty set.  On the other hand, a) number theory does not
have to be governed by the same rules as set theory and b)
'isomorphism' is partly a matter of definition.  I could claim that
the fact, for instance, that count 2 does not have a square root while
real 2.0 does makes them not 'fully' isomorphic.  This gets to the CS
view that types are defined by the operations allowed on the instances
(at least roughly put).

There is also a third (non-Platonic) view: numbers are purely human
invention.  Thus, decisions about identity are arbitrary, and hence
arguments about the 'truth' of such is at least somewhat nonesensical.
Thus, my argument above would only make sense if recast as one about
pragmatics.

Take your pick.

Terry J. Reedy







From tjreedy at udel.edu  Sat Apr 26 00:21:48 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Sat, 26 Apr 2003 00:21:48 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <8CydnSQn2I0ulDejXTWcpA@comcast.com>

"Frantisek Fuka" <fuka at fuxoft.cz> wrote in message
news:b8cadp$1fj9$1 at ns.felk.cvut.cz...
> When I try to call methods, I sometimes forget to include the
> parentheses. Instead of:

In Python, '()' is, in the proper context, the 'call me with the
contained (possible none) arguments' operator.  So the above is much
like saying, "When I try to calculate products, I sometimes forget to
include the asterisk' -- like mathemeticians are in the habit of
doing -- and then asking that interpreter start guession that missing
binary operators are implied multiplication.

Python code in not math notation.

> if object.isGreen():
> do something...
>
> i sometimes write:
>
> if object.isGreen:
> do something...
>
> If I understand it correctly,

which you do not, quite yet ;-)

> the if statement in this case tests if  pointer to hasParent method
is non-zero,

In Python, every expression evaluates to a Python object.  In
expressions, bare names evaluate to the object they are bound to in
the the appropriate namespace.  By default, absent user written
behind-the-scenes trickery, dotted (attribute) names evaluate to the
object they are bound to in the preceeding objects private namespace.
In any case, if statements branch according to the boolean value of
the object resulting from the conditional expression.  The boolean
value of function objects is always True.

Pointers (and memory addresses) are not part of the definition of
Python.  Python is not C.

> You can say to me "Don't forget to always include the parenteses"

Just as you should not forget the asterisk in multiplications.

> I'm still curious if this cannot be somehow configured,

The consistency of Python expression evaluation is one of its basic
features.  Fighting it is the path to frustration.

> the possibility of accessing the method pointers but I usually don't
> need it and for a beginner like me this makes the applications very
hard
> to debug.

Have you never yet typed in 'obj.meth.__doc__' at the interactive
interpreter prompt?  That is possible precisely because 'obj.meth' is
precisely what it claims to be.  Also, it should not be too long
before you want to start abbreviating obj.meth and do something like

stack = []
push = stack.append
pop  = stack.pop
#code using these three names, simple example being
push(1); print stack
# [1]
print pop(), stack
# 1, []

Terry J. Reedy





From anton at vredegoor.doge.nl  Mon Apr 28 18:18:36 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Tue, 29 Apr 2003 00:18:36 +0200
Subject: Python-2.3b1.exe install error on W98se
References: <mailman.1051538624.9122.python-list@python.org>
Message-ID: <b8k9rp$9fn$1@news.hccnet.nl>

Tim Peters <tim.one at comcast.net> wrote:

>IE, or vice versa).  Since I wasn't looking over your shoulder, I'm afraid I
>can't point at anything insane you may have done <wink>.

You want insanity? I've got a whacky modem that consistently refuses
to agree with itself over a small range of Python-2.3b1.exe's data, no
matter which way I try to download it. Please run this program and
show me the output and I will be able to complete Python-2.3b1.exe
myself. 

Anton

def main():
    data = file("Python-2.3b1.exe",'rb').read()
    R = xrange(15800,16200)
    for i in R:
        print '%4i' %ord(data[i]),
        if  (i+1) % 10 == 0:
            print
    print

if __name__=='__main__':
    main()




From news at snakefarm.org  Thu Apr 10 17:59:13 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 10 Apr 2003 21:59:13 GMT
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de> <pan.2003.04.10.23.50.37.62608.750@thomas-guettler.de>
Message-ID: <slrnb9bq9h.113.news@cipw08.physik.uni-bonn.de>

In article <pan.2003.04.10.23.50.37.62608.750 at thomas-guettler.de>,
Thomas Guettler wrote:

> popen can you give an stdout and stderr (Standard out/error) stream. Read
> them after writing to stdin.

No, you can't read the stdout and stderr of a process spawned by popen()
because its stdout and stderr are connected to the stdout and stderr of
the calling process.

cg.



From tanzer at swing.co.at  Mon Apr 14 03:08:02 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Mon, 14 Apr 2003 09:08:02 +0200
Subject: an ugly file-reading pattern
In-Reply-To: Your message of "Sat, 12 Apr 2003 21:34:31 CDT."
             <16024.52407.405254.424039@montanaro.dyndns.org>
Message-ID: <E194y4Q-0002W1-00@tswings.swing.cluster>

Skip Montanaro <skip at pobox.com> wrote:

> Recent versions of Python (2.2 and later) allow this:
>
>     file = file("somefile")
>     for line in file:
>         process(line)
>
> It does the right thing from an efficiency standpoint, neither slurping the
> file in all at once nor reading it line-by-line or char-by-char.
>
> Note also that in 2.2 and later there is a file object in builtins ("open"
> and "file" are the same object).  "open" is discouraged.  "file" is
> preferred.

Maybe you should have written it as :

    open = file("somefile")
    for line in open:
        process(line)

<2.2 wink>

Hiding-builtins-often-bytes-the-unwary-ly yr's,

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From jdhunter at ace.bsd.uchicago.edu  Wed Apr 23 09:52:09 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Wed, 23 Apr 2003 08:52:09 -0500
Subject: ANN: matplotlib 0.1
Message-ID: <m28yu1ianq.fsf@mother.paradise.lost>

Matplotlib 

  Matplotlib is a pure python plotting package for python and pygtk.
  My goal is to make high quality, publication quality plotting easy
  in python, with a syntax familiar to matlab users.  matplotlib is
  young, and several things need to be done for this goal is achieved.
  But it works well enough to make nice, simple plots.

Requirements

   python 2.2, GTK2, pygtk-1.99.x, and Numeric.

Download

   See the homepage - http://nitace.bsd.uchicago.edu:8080/matplotlib

Here are some of the things that matplotlib tries to do well

  * Allow easy navigation of large data sets.  Right click on figure
    window to bring up navigation tool bar for pan and zoom of x and y
    axes.  This requires a wheel mouse.  Place the wheel mouse over
    the navitation buttons and scroll away. 

  * Handle very large data sets efficiently by making use of Numeric
    clipping.  I have used matplotlib in an EEG plotting application
    with 128 channels and several minutes of data sampled at 400Hz,
    eg, plotting matrices with dimensions 120,000 x 128.

  * Choose tick marks and labels intelligently

  * make easy things easy (subplots, lines styles, colors)
 
  * make hard things possible (OO interface for full control)

Matplotlib is a class library that can be used to make plots in pygtk
applications.  But there is a matlab functional compatibility
interface that you can get with, eg::

  from matplotlib.matlab import plot, subplot, show, gca

Example scripts and screenshots available at
http://nitace.bsd.uchicago.edu:8080/matplotlib

John Hunter






From spam at magnetic-ink.dk  Wed Apr 23 10:23:24 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Wed, 23 Apr 2003 14:23:24 +0000 (UTC)
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz>
Message-ID: <70846faf-b437-47ee-8a2c-7f685b00425b@news.szn.dk>

Frantisek Fuka wrote:

> I want a simple test to see if str contains any occurences of substr.
> I am currently using this:
> 
> import string
> if string.find(str,substr) + 1:
>      ...do something...

If you're using a newer version of Python (2.2+?), you should be able
to do something like this:

#v+

>>> contains = lambda haystack, needle: haystack.find(needle) > 0
>>> contains('foo bar fly', 'foo')
1
>>> contains('foo bar fly', 'xyz')
0
>>> 

#v-


  // Klaus

-- 
 ><> 	http://pnx.dk/



From marcus at infa.abo.fi  Thu Apr 10 09:28:59 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 10 Apr 2003 15:28:59 +0200
Subject: Import 'filename'? (was Re: Cross-reference 'import' in a classhierarchy)
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <3e932e9f$1_2@corp-news.newsgroups.com> <b6vri8$9ob33$1@ID-169208.news.dfncis.de> <3E9428A8.9090005@mitretek.org> <b72kcp$aejp6$1@ID-169208.news.dfncis.de>
Message-ID: <slrnb9aopc.ln1.marcus@infa.abo.fi>

On Thu, 10 Apr 2003 14:30:02 +1200, Greg Ewing (using news.cis.dfn.de)
<ckea25d02 at sneakemail.com> wrote:
>In my opinion, it would be good enough to expand any
>symbolic links (or aliases/shortcuts/whatever your platform
>calls them) and check whether the pathnames are the same.
>
>That won't always work (in the presence of mountings from
>file servers, e.g.) but it can't be any worse than the
>problems we already have with different names for the
>same file.

I'm not sure what your "good enough" algorithm actually does.
Consider also the fact that e.g. on Linux you can mount the same
filesystem several times, in several places. There is no _one unique_
canonical name for a file.

Personally I'd just like a simple stat() for each (non-symlink) file
and remembering their st_dev and st_ino values. I don't know how this
would be emulated on other operating systems, though.

-- 
Marcus Alanen
maalanen at abo.fi



From jack at performancedrivers.com  Thu Apr  3 08:20:06 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Thu, 3 Apr 2003 08:20:06 -0500
Subject: Overloadable Assignment PEP
In-Reply-To: <qp5o8voto8li6p854klt7vo6dhon69hk2o@4ax.com>; from richie@entrian.com on Thu, Apr 03, 2003 at 01:42:28PM +0100
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <qp5o8voto8li6p854klt7vo6dhon69hk2o@4ax.com>
Message-ID: <20030403082006.D936@localhost.localdomain>

On Thu, Apr 03, 2003 at 01:42:28PM +0100, Richie Hindle wrote:
> So the line `page.message = "New message"` changes the content of the tag
> with id 'message'.  In this context, it makes sense that assigning to an
> object *changes* that object, rather than *replacing* it with another.
> 
> What you can't do with PyMeld is this:
> 
> >>> message = page.message
> >>> message = "Another new message"
> >>> print page
> <html><body>
> <textarea id="message">Another new message.</textarea>
> </body></html>
> 

As others have pointed out, this isn't pythonic, but you could make a string 
wrapper class so you could do:
  messgae = page.message
  message.txt = "Another New Message"

you still couldn't do
  foo = message.txt
  foo = "another new message"

You might want to rethink using the magic of page.message finding the element
in page with the 'id' attribute equal to 'message'.  People will undoubtably
start naming their element 'id' attributes to '__init__' or something else
to make your life difficult.

(psuedo python debugging below)

>>> msg = page.byId('message')
>>> print msg['id']
message
>>> print msg.txt
another new message
>>> print page.elements
[<class textarea>,]

I'm not sure what happens when two elements at the same level have the same
'id' attribute (that applies to any way you do it).

-jackdied




From bens at replytothegroupplease.com  Thu Apr 10 16:29:23 2003
From: bens at replytothegroupplease.com (Ben S)
Date: Thu, 10 Apr 2003 21:29:23 +0100
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <b74k25$qlo$1@news6.svr.pol.co.uk>

Alex Martelli wrote:
> I _know_ there's a decent algorithm to solve the following problem,
> but I don't recall its name and thus can't rapidly google for the
> details... can somebody help?
>
> The problem: I need to design a "decision criterion" to classify
> "observations".  For each observation I measure the values of a
> number N of features, x1, x2, ... xN; the desired design criterion
> is a set of weights w1, w2, ... wN such that for any observation I
> will then just compute a weighted sum
>   S = w1*x1 + w2*x2 + ... + wN*xN
> and classify the observation as Black if S<=1, White if S>1.

This accurately describes a neural network (in addition to the other
concepts that other (more learned) people have already posted about). It
starts off with random weights, you feed in the values, see if you got
the right answer, and adjust the weights accordingly. Eventually you
have your set of weights.

--
Ben Sizer
http://pages.eidosnet.co.uk/kylotan






From aleax at aleax.it  Sat Apr 19 12:15:48 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 16:15:48 GMT
Subject: cgi parser question ?
References: <pan.2003.04.19.17.59.39.393264@sdf-eu.org>
Message-ID: <UCeoa.57982$T34.1547349@news2.tin.it>

Geirgat Jonas wrote:

> Hello,
> maybe this isn't the right place to ask this question I know....
> I have written a basic HTTP server in python now I would like it
> to have python CGI support only I haven't found how CGI works

One way to find out is to study the module CGIHTTPServer.py you
get as part of the Python standard library -- compare and contrast
it with the BaseHTTPServer.py that you also get there, and you'll
start seeing what you need to add to support CGI in addition to
basic HTTP serving.

> is it like this a user requests a page you just parse it "python page.py"
> and send to output to the user ?

Normally, what the server does is, when the user requests a page
that the server can somehow identify as connected to a CGI script,
run the CGI script with the environment appropriately prepared,
and send the script's output as the page to the user.  The way a
server identifies a page request as connected to a CGI script, and
which one, is up to the server, of course -- e.g., looking for
"cgi-bin" in the path to the page is one common convention.


> any help, links would also help on this subject

http://hoohoo.ncsa.uiuc.edu/cgi/interface.html


Alex




From cben at techunix.technion.ac.il  Mon Apr 28 16:32:18 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Mon, 28 Apr 2003 23:32:18 +0300 (IDT)
Subject: In which versions is list.sort stable?
In-Reply-To: <mailman.1051479463.462.python-list@python.org>
References: <mailman.1051479463.462.python-list@python.org>
Message-ID: <Pine.GSO.4.51.0304282316240.8857@techunix.technion.ac.il>

Tim Peters wrote on 2003-04-27:

> [Beni Cherniavsky]
> > Why don't we just grow a `list.stablesort` method that can be an alias
> > to `list.sort` currently but is guaranteed to be some stable sort
> > forever?  Explicit is better than implicit, so in the rare case that
> > you do rely on it, why not spell it out?
>
> Because Guido has shown no interest in saying that Python-the-language
> guarantees to provide a stable sorting method.  Now that 2.3 has a stable
> sort, I expect many apps will find that it's still more efficient to force
> stability via injecting sequence indices into sort tuples.  That is,
>
>     temp = [(obj.key(), i, obj) for i, obj in enumerate(list_of_objects)]
>     temp.sort()
>     result = [obj for key, i, obj in temp]
>
Neat trick, never occurred to me to force stability with the
{,un}decorate pattern (although I'm well familiar with it), perhaps
because I never needed a stable sort so far <wink>.  I made this
suggestion simply because I thought of this as a cleaner way
immediately after I saw the announcement that "it's now stable but
without promises"...

> will often be faster than
>
>     temp = [(obj.key(), obj) for obj in list_of_objects]
>     temp.sort()
>     result = [obj for key, obj in temp]
>
> The hangup is that tuple comparison breaks ties in the first element
> position by going on to compare more element positions.  In the latter
> spelling, that can invoke arbitrarily expensive object comparison, but in
> the former spelling is always resolved no latter than by comparison of
> unequal ints.
>
> This became painfully clear during timing tests run while the 2.3 sort was
> being developed, on a database-sorting example contributed by an alpha
> tester, where primary keys where often equal.  Forcing stability in that
> test ran twice as fast as when leaving the sequence indices out of the
> tuples.
>
Even more neat; I think I can use this trick right now.  Perhaps the
docs should mention this right after saying that "code that intends to
be portable across implementations and versions must not rely on
stability".  Hmm, this "perhaps the docs should say X" theme is
recurring lately.  I wander what would happen if the docs would be
turned into a Wiki (probably - a mess - but a pseudo-wiki that
automagically turns edits into submitted patches or something like
that could be convenient.

[s.split("-ly y'rs\n")[::-1] for s in signatures].sort()-ly y'rs
Beni Cherniavsky <cben at tx.technion.ac.il>




From rainer.klueting at epost.de  Mon Apr  7 13:17:31 2003
From: rainer.klueting at epost.de (Rainer Klueting)
Date: Mon, 07 Apr 2003 19:17:31 +0200
Subject: Datagrid-Widget for Python?
Message-ID: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com>

Hi,

can anybody point me to a source where I can find a datagrid-widget to
use with Python? It should basically show a flat table in a
spreadsheet-like way. I tried for a while to do it by myself using
Tkinter, but I'm lacking experience (just trying to change that...)

A search on several Python-sites wasn't successful. Thank you for any
hint!

Rainer Klueting, Stuttgart, Germany



From SBrunning at trisystems.co.uk  Tue Apr 15 04:41:40 2003
From: SBrunning at trisystems.co.uk (Simon Brunning)
Date: Tue, 15 Apr 2003 09:41:40 +0100
Subject: DB API for ADO/DAO?
Message-ID: <31575A892FF6D1118F5800600846864D012007FB@intrepid>

> From:	john.burton at jbmail.com [SMTP:john.burton at jbmail.com]
> I want to be use an .MDB database from a python application on
> windows.
> 
> The problem with odbc is that you need to manually set up data sources
> using the control panel which isn't something I want the users to have
> to do.
> 
> So I'm using DAO via com to access the database and it's working fine
> except that I have to use the DAO com api instead of the standard
> python api. So I was wondering if there is a suitable DB API wrapper
> for DAO or ADO I can use? Or a better solution to my problem?
 
http://www.google.com/search?q=python+ado+db%2Dapi

;-)

Cheers,
Simon Brunning
TriSystems Ltd.
sbrunning at trisystems.co.uk




-----------------------------------------------------------------------
The information in this email is confidential and may be legally privileged.
It is intended solely for the addressee. Access to this email by anyone else
is unauthorised. If you are not the intended recipient, any disclosure,
copying, distribution, or any action taken or omitted to be taken in
reliance on it, is prohibited and may be unlawful. TriSystems Ltd. cannot
accept liability for statements made which are clearly the senders own.




From ramen at lackingtalent.com  Tue Apr 15 16:40:08 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 15 Apr 2003 20:40:08 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <v9oq71itajkh39@news.supernews.com>
Message-ID: <slrnb9orhr.toh.ramen@lackingtalent.com>

In article <v9oq71itajkh39 at news.supernews.com>, John Roth wrote:
> I agree completely. I got turned on to the existing functional
> constructs while I was doing a TDD challenge. I managed
> to do that with exactly one if statement and no loops because
> of the functional constructs.

TDD as in "test-driven development"? I'm not really familiar with this.
Got any good URLs to look at?

> However, losing apply() isn't a big deal because there is
> an exactly equivalent construct. I'd hate to see the rest
> of them go away, especially since the (possible) implementation
> of PEP 308 will remove one of the restrictions on lambdas.

You're right, it's not a big deal. I never use apply anymore, since
Python provides a very convenient syntactic-sugaring as a replacement.

But, I still want to point out the fact that this construct is Python-only.
On the other hand, apply can be implemented with nearly identical behavior
in many other languages, if it doesn't exist already.

> Also notice that there are a set of new functional constructs
> in 2.3 alpha.

I assume you're talking about the "itertools" module. I'm very interested in
this module, though I haven't fully grokked it. I like the "starmap" function.

I think it'd be good to have a module called "functional" where the more
unusual functional tools can live, and I know that it's been brought up before.
Polluting the built-ins further is obviously a bad thing for backward
compatability, not to mention the learning curve. But it would at least make
a statement that functional programming is possible and supported in Python.

Take care,
Dave



From oblivious at web.de  Fri Apr 18 10:46:59 2003
From: oblivious at web.de (Dominic)
Date: Fri, 18 Apr 2003 16:46:59 +0200
Subject: Toy Pyrex - XMMS interface
Message-ID: <b7p31u$jpv$02$1@news.t-online.com>

I've programmed a simple toy-python
interface to XMMS (MP3-player).

I use it to pipe files into the playlist:

find /mnt/music/ | grep -i latin | xmms-pase.py

and store playlists e.g.:

xmms-list.py latin

stores the playlist including the current title
and it's position. On reload (when the playlist is
empty)

xmms-list.py latin

restores the old settings and continues to
play from where you left.

If you want to use it, you can download it from:

http://www.rhrk.uni-kl.de/~hillbr/xmms.tar.gz

Ciao,
  Dominic




From anton at vredegoor.doge.nl  Thu Apr 17 19:50:24 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Fri, 18 Apr 2003 01:50:24 +0200
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl> <3E9EAA36.BFB1018D@engcorp.com> <IQmdnQg6xagDdQOjXTWcpA@august.net>
Message-ID: <b7neqo$knk$1@news.hccnet.nl>

Terry Hancock <hancock at anansispaceworks.com> wrote:

<snip>

>Perhaps one can query the DNS registry in some way that would be more 
>equal?  Can RARP be used to look up *all* domains associated with an IP, or 
>only one canonical value?  If you could look up all of them, you could then 
>select one at random -- but this will not generate evenly weighted results 
>of course.

But why do the DNS-lookup if it's possible to load a page using an ip?
In fact the reason I asked this question here is that I *have* a
script (partly based on some code I found in clp) that generates
random ip's, then looks up their names using a DNS-server. The problem
I have with this script is that it necessarily blasts my providers DNS
server since a lot of addresses need to be looked up, most of which
will result in an error, and the remaining successfully resolved
addresses very seldomly have a HTTP page.

If I want to pursue the matter any further there's probably no other
choice than to follow Peters advice, but it feels like some kind of
strange contortion of a portscan (which is generally frowned upon)
where the port is held constant and the ip-address is varied. I would
welcome some advice about the level of politeness such actions are
generally associated with since I would not like to evoke unfriendly
reactions.

<snip>

>Of course, the distribution is probably not too relevant -- there are so 
>many web pages in the world now, that you are unlikely to retrieve the same 
>one twice with any truly random system.  (Not true on Google, where 
>connectivity highly biases the search).

Well, to me it is. I could easily generate lots of legitimate reasons
for wanting an even distribution (for example  statistical analysis of
net content and such) but that would be lying since reality is, I just
want such a tool for my personal pleasure to play with :-)

Anton.

PS. here's my unsatisfying DNS-code (use at your own risk, I am not
sure providers would like this script being executed regularly) The
threading code seems to be necessary because the script would be very
slow without it. I haven't got a lot of experience with threading so I
can't get rid of the global variables yet. I would welcome
improvements for this code as well as a script to generate a random
webpage.

import threading, socket
import time
import random

def rAddress():
    return '.'.join([str(random.randint(0,255)) for i in range(4)])

def LookupThread():
  while 1:
    try:
        ip = ipList.pop()
        try:
            ipAndHostList.append((ip, socket.gethostbyaddr(ip)[0]))
        except socket.herror:
            pass
    except IndexError:
      break

ipList = [rAddress() for i in range(1000)]
ipAndHostList = []

for i in range(500):
  threading.Thread(target=LookupThread).start()

while ipList:
  time.sleep(0.1)

fn = 't4.txt'
f = file(fn,'a+')

for x in ipAndHostList:
    print x
    f.write('%s %s\n' %x)
f.close()







From cnetzer at mail.arc.nasa.gov  Fri Apr 18 14:37:13 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 18 Apr 2003 11:37:13 -0700
Subject: Classes in Python
In-Reply-To: <3EA01A9B.4030209@vnet.net>
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <1050691033.587.5.camel@sayge.arc.nasa.gov>

On Fri, 2003-04-18 at 08:32, Bill Martin wrote:
> I'm wondering about the value of allowing a class definition like this:
> 
>              class C:
>                pass
> 
> Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or 
> some such. If I try to print a.x2 or b.x1, I get an exception message 
> basically saying those member variables don't exist.

You were assigning to created instances of class c, not the class
itself.  By assigning to the class object (the object that, when called,
creates instaces a and b), you can create new class attributes that ARE
shared by all instances.  (Note that you can also define the class with
these instances)

$ python
Python 2.2.2 (#1, Mar 21 2003, 23:01:54)

>>> class c:
...     pass
...
>>> a = c()   # These are instances of class c
>>> b = c()
>>> dir(a)
['__doc__', '__module__']
>>> c.foo = 4
>>> dir(a)
['__doc__', '__module__', 'foo']
>>> a.foo
4
>>> b.foo
4

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From jdavis at empires.org  Sat Apr  5 06:30:27 2003
From: jdavis at empires.org (Jeff Davis)
Date: Sat, 05 Apr 2003 11:30:27 GMT
Subject: handling SIGINT
Message-ID: <n7zja.27949$Hx.18737997@twister.socal.rr.com>

I've got a python module that I wrote in C and it spends a good portion of
the time sleep()ing. When I run the python program that calls the C module
than a SIGINT only interrupts the current sleep and the main program can't
intercept the signal.

This may not be python-specific, but how do I handle a SIGINT while the
program is sleep()ing?

Thanks, 
        Jeff



From sholden at holdenweb.com  Sat Apr  5 15:49:31 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sat, 05 Apr 2003 20:49:31 GMT
Subject: POP3 and selective download/deletion
References: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com> <pan.2003.04.05.16.22.31.832230.878@thomas-guettler.de> <luCja.11925$It5.5735@news2.central.cox.net> <mailman.1049568069.20218.python-list@python.org>
Message-ID: <vjHja.15799$It5.12275@news2.central.cox.net>

"David Mertz" <mertz at gnosis.cx> wrote in message
news:mailman.1049568069.20218.python-list at python.org...
> "Steve Holden" <sholden at holdenweb.com> wrote previously:
> |Unfortunately the POP3 protocol doesn't allow you download the headers of
a
> |message without downloading the whole thing.
>
> Nonsense.  This is part of the script I use to filter SPAM based only
> on headers (so I don't download the large virus/spam message bodies).
There is
> a bit of cruft in that I display some progress information and log some
> actions, but it shows a realistic program.
>
[...]

Indeed. Unfortunately TOP isn't something you can rely on being available on
all servers, and its implementations are inconsistent. The 2.2 docs for
poplib objects say:

"""The POP3 TOP command this method uses, unlike the RETR command, doesn't
set the message's seen flag; unfortunately, TOP is poorly specified in the
RFCs and is frequently broken in off-brand servers. Test this method by hand
against the POP3 servers you will use before trusting it."""

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From tzahij at hotmail.com  Tue Apr 22 05:14:34 2003
From: tzahij at hotmail.com (Tzahi)
Date: 22 Apr 2003 02:14:34 -0700
Subject: Python implementation for a web spider \ site mirroring
Message-ID: <662cede5.0304220114.54e4dd10@posting.google.com>

Hi

I need to copy complete web sites form the web.

I think Python is a natural fit for this task, with the intensive HTML
and HTTP support. Being lazy, I would rather have someone else's
implementation, so that I do not start from scratch.

Anyone knows of such a free implementation?

Thanks in advance,

Tzahi Matar



From hawkeye.parker at autodesk.com  Tue Apr 29 17:52:56 2003
From: hawkeye.parker at autodesk.com (haughki)
Date: 29 Apr 2003 14:52:56 -0700
Subject: windows-cmd shell doesn't reflect _winreg changes
References: <b6e56f7a.0304021609.521700e6@posting.google.com> <slrnb8p06l.174.do-not-spam-ben.hutchings@tin.bwsint.com> <b6e56f7a.0304091108.6030e453@posting.google.com>
Message-ID: <b6e56f7a.0304291352.7ff33a86@posting.google.com>

follow up:

the SendMessage command as suggested below didn't actually work.  I
did some poking around on the win32 newsgroup and eventually found MS
Knowledge Base's Q104011, "HOWTO: Propagate Environment Variables to
the System".  it suggests:

SendMessageTimeout(HWND_BROADCAST, WM_SETTINGCHANGE, 0, (LPARAM)
"Environment", SMTO_ABORTIFHUNG, 5000, &dwReturnValue);

which is pythonified to:

win32gui.SendMessageTimeout(win32con.HWND_BROADCAST,
win32con.WM_SETTINGCHANGE, 0, "Environment",
win32con.SMTO_ABORTIFHUNG, 5000)

and actually works.  you still have to close the cmd shell and reopen
it, but this is much faster than going through the whole process
manually.


hawkeye.parker at autodesk.com (haughki) wrote in message news:<b6e56f7a.0304091108.6030e453 at posting.google.com>...
> Many thanks Ben!!
> 
> 
> 
> 
> Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> wrote in message news:<slrnb8p06l.174.do-not-spam-ben.hutchings at tin.bwsint.com>...
> > In article <b6e56f7a.0304021609.521700e6 at posting.google.com>, haughki wrote:
> > > hi
> > > 
> > > i'm using _winreg to 'hard-change' and environmental variable in
> > > windows.  the change is 'successful': regedit correctly reflects the
> > > changes, and crtl panel > system > properties envronmental vars have
> > > been correctly changed.  but, when i start a cmd shell (dos), the
> > > shell evironment still reflects the old values until i manually open
> > > the env vars cntrl panel and click 'ok'.  anyone?
> > 
> > This is really a Win32 programming issue rather than a Python one.
> > However, I do vaguely remember this one.
> > 
> > Environmental variables are normally passed to a new process by the
> > process that creates it, and they can never be updated from outside
> > a process.  The exception to this is that Windows Explorer (and any
> > other program that wishes to) will read those initial environment
> > variables out of the registry when notified of a change in settings.
> > The changes will then be passed on to any program started by Explorer
> > afterwards.
> > 
> > Here's the Python code:
> > 
> >     import win32con, win32gui
> >     win32gui.SendMessage(
> >         win32con.HWND_BROADCAST, win32con.WM_SETTINGCHANGE, 0, 0)
> > 
> > You might want to use SendMessageTimeout instead of SendMessage.



From bfordham at socialistsushi.com  Fri Apr 11 10:18:18 2003
From: bfordham at socialistsushi.com (Bryan L. Fordham)
Date: Fri, 11 Apr 2003 10:18:18 -0400
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <3e96b088_2@127.0.0.1> <xPzla.367$xm2.20@news02.roc.ny.frontiernet.net>
Message-ID: <3E96CEAA.1090008@socialistsushi.com>

Lee Harr wrote:

> What I mean is that students should first learn how to use the tools before
> they are required to know how to make them.

Absolutely.  This is how my calculus teach in HS taught us: he'd show us 
the really complicated way, then show us the shortcut.

It would make perfect sense for a teacher to require you to use 
something more than list.sort() if you're studying sorting... but after 
that, why make you write your own sorting routines if what's provided 
works just fine?

--B






From aleax at aleax.it  Mon Apr 28 10:32:20 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 14:32:20 GMT
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
References: <mailman.1051532982.25352.python-list@python.org>
Message-ID: <UXara.22428$K35.745468@news2.tin.it>

<posted & mailed>

nguyen khanh wrote:

> a=range(10)
> b=a[3:7]
> b=[3,4,5,6]
> Why the design exclude the last element ?

It's an important design principle, which I first saw proposed in print
by Andrew Koenig in his landmark book "C Traps and Pitfalls": *all bounded
loops should be first-bound-included, last-bound-excluded* -- and which I
will therefore call "Koenig's Principle" although I have never seen the
principle *named* in the literature (I'mm CC'ing Andrew because of this --
I know there are already other things named after him, such as the "Koenig
Lookup" in C++, that he might like to get his name AWAY from, so it seems
only fair to allow him to disclaim THIS one too ASAP if he dislikes it!-).


Koenig wasn't talking about "slices", of course (there are no slices in C!),
but it works just as well for a C for-loop, "for(i=first; i<last; i++)" i.e.
using < and NOT <= for the terminating-condition.

It's an important design principle because of human-factors considerations,
which Koenig expresses better than I can (I really wish he would one day
write a "Python Traps and Pitfalls" book, but he seems too busy writing
great C++ books instead... perhaps the Python one would be too thin?-).  
But, for example, consider a question such as:

"How many items does XX[a:b] have?"

With Python's design (following Koenig's principle), the normal
answer (when a and b are both within XX'x boundaries) is:
        b - a

If the last-bound had been deemed to be included, it would have been:
        b - a + 1

Now, that innocuous-looking "+ 1" there is the cause of untold misery
in programming -- the root of most "off-by-one" errors.  By removing
the need for it, Koenig's principle makes you more productive.

Thanks to the principle, all sorts of pleasant invariants follows.

For example, for all XX, x, y, z, XX[x:y]+XX[y:z] == XX[x:z].  This
kind of regularity is exactly what facilitates reasoning about
programs, and in particular modifying programs.

> I's because the distinction of "element and list-element" ?
> a[3]=3        # element
> a[3:4]=[3]   # list-element
> 
> Why the designer d'nt use a[3:3]=[3] instead of a[p:p]=[] ?

Not sure at all what you mean by ``the distinction of "element and 
list-element" (the operations are called INDEXING and SLICING in
Python).  But I hope I started clarifying the design decision for you.

> The result a[p:p]=[] for all p d'nt add new information ,

Wrong!!!  Are you kidding...?

>>> a = list('ciao')
>>> a[3:3] = list('pe')
>>> a
['c', 'i', 'a', 'p', 'e', 'o']

vs

>>> a = list('ciao')
>>> a[2:2] = list('pe')
>>> a
['c', 'i', 'p', 'e', 'a', 'o']

Of COURSE it does "add new information", WHICH exact empty list you
are slicing -- i.e., exactly which p you're using for a[p:p] -- even
though you can "notice" this information only when you ASSIGN to the
particular slice rather than USE it (but of course, for reasons of
regularity, a[x:y] must indicate the same slice, for all x and y,
whether it's being assigned-to, or used).


> The design a[n:p] clarify the mind to include the last element p .

I deeply believe, on the contrary, that a systematic "last element
included" design would seriously damage Python, while its current
systematic "last element excluded" design enhances it.

> and a[p:p]=[p] is very clear and dinctint from a[p]=p .

In the current design, it is -- the first assignment INSERTS p
at the p-th position, the second assignment REPLACES the p-th
position with p, so the two are indeed "very clear AND DISTINCT".
If Python were to use the "last-element-included" design, then
the two assignments wouldn't be really distinct any more (except
perhaps in that slicing is more tolerant than indexing -- a[p]
raises unless -len(a) <= p < len(a), a[p:p] doesn't).


Alex




From martin at v.loewis.de  Mon Apr 14 15:57:39 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 14 Apr 2003 21:57:39 +0200
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com>
Message-ID: <m3y92cvolo.fsf@mira.informatik.hu-berlin.de>

paul.moore at atosorigin.com (Paul Moore) writes:

> To use a concrete example, I'd like to print the Euro symbol. A
> Unicode string for this is u'\20a0'
> 
> >>> unicodedata.name(u'\u20a0')
> 'EURO-CURRENCY SIGN'

As Irmen points out, U+20AC is the euro sign; U+20A0 predates the
European monetary union, and denotes the European Currency Unit (ECU).

> >>> print u'\u20a0'.encode("latin-15")
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> LookupError: unknown encoding: latin-15

There is no encoding latin-15. You probably mean "latin-9", which is
more formally known as "iso-8859-15" (in ISO 8859, not all encodings
are Latin: some are Arabic, Hebrew, Cyrillic, and Greek).

> 
> >>> print u'\u20a0'.encode("latin-1")
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> UnicodeError: Latin-1 encoding error: ordinal not in range(256)

Right. Latin-1 supports neither the ECU nor the EUR.

> 
> >>> print u'\u20a0'.encode("iso8859_15")
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "C:\Python22\lib\encodings\iso8859_15.py", line 18, in encode
>     return codecs.charmap_encode(input,errors,encoding_map)
> UnicodeError: charmap encoding error: character maps to <undefined>

Right. Only the EUR is in Latin-9, not the ECU.

> >>> print u'\u20a0'.encode("cp1258")
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "C:\Python22\lib\encodings\cp1258.py", line 18, in encode
>     return codecs.charmap_encode(input,errors,encoding_map)
> UnicodeError: charmap encoding error: character maps to <undefined>

Likewise.

> As I say, I *know* this is subtle. I'm happy to work out the details,
> once I can get a simple example working. But how do I get started? How
> do I get u'\u20a0' to display on my screen as a Euro character???

You won't ever get this. The best you can hope for is to get a symbol
which looks like a C on the top-left of an E.

> PS While I can live with having to know details of how the console is
> configured in order to get this working, ideally I'd like to know a
> way of getting anything I need out of Windows (even if it requires API
> calls...), so that I can write something generic which, when given a
> Unicode string, can display it on the console without needing any
> extra information from the user...

Microsoft has added API to support output of arbitrary Unicode
characters on the console, regardless of the code page
(WriteConsoleW). This only works if the console uses a Unicode font
(Lucida Sans Unicode).

Regards,
Martin



From staschuk at telusplanet.net  Tue Apr 29 12:22:20 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 10:22:20 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <NTjra.23897$WQ4.21608@news2.central.cox.net>; from sholden@holdenweb.com on Tue, Apr 29, 2003 at 12:42:21AM +0000
References: <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <20030417205405.A10672@tibia.amotlpaa.bogus> <mailman.1051310385.1948.python-list@python.org> <b8cgst$20ho$1@nntp6.u.washington.edu> <mailman.1051490038.11587.python-list@python.org> <NTjra.23897$WQ4.21608@news2.central.cox.net>
Message-ID: <20030429102220.C338@tibia.amotlpaa.bogus>

Quoth Steve Holden:
> "Steven Taschuk" <staschuk at telusplanet.net> wrote in message
> news:mailman.1051490038.11587.python-list at python.org...
  [...]
> Finally, the murk lifts and a dim light becomes perceptible. "str() is what
> the users will recognize. repr() is what the programmers will thank you
> for."

For repr, I'm happy with that statement.

But for str: Donn Cave pointed out rfc822.Message.__str__ (which
returns the complete headers) a while back; I feel that's a good
__str__, yet is not necessarily interesting or useful to a user
of, say, a mail client written in Python.  Indeed, it is primarily
useful programmatically, at the recipient's mail client.

> > *chuckle*  True.  (Not that I'm in any special position vis a vis
> > changing the documentation.  Quite the contrary.)
>
> Everybody is in the *same* position where changing the documentation is
> concerned: [...]

Yes.  Therefore nobody has a special position in this respect, and
in particular, I do not.  QED!

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From michael at stroeder.com  Fri Apr 11 12:00:27 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Fri, 11 Apr 2003 18:00:27 +0200
Subject: popen(), sendmail: Success?
In-Reply-To: <slrnb9dma7.1vh.news@cipw08.physik.uni-bonn.de>
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de> <v29im-ia2.ln1@nb2.stroeder.com> <slrnb9dma7.1vh.news@cipw08.physik.uni-bonn.de>
Message-ID: <rd5jm-5aq.ln1@nb2.stroeder.com>

Carsten Gaebler wrote:
> In article <v29im-ia2.ln1 at nb2.stroeder.com>, Michael Str?der wrote:
> 
>>For security reasons I'd be scared of printing an arbitrary exception 
>>message to stdout in a web application since it is sent to the user's browser!
> 
> Of course I wouldn't do that in a real world situation. It was just a simple
> example of how to handle an exception.

The problem is that people pick up those examples without further thinking. 
I already saw public web applications sending the full Python traceback to 
the browser... :-(

Ciao, Michael.




From aleax at aleax.it  Sun Apr 27 03:27:57 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 07:27:57 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <b8fsvk$kkg$1@panix2.panix.com>
Message-ID: <1ELqa.9765$3M4.264114@news1.tin.it>

Aahz wrote:
   ...
> Anyway, all objects that do not implement a __nonzero__ method are true:

...unless they implement __len__, which is used as a fallback if it's
present but __nonzero__ is absent (I'm sure Aahz know this, I'm just
dotting t's and crossing i's for other readers who might not).


Alex




From adamv at nwlink.com  Wed Apr 23 19:36:02 2003
From: adamv at nwlink.com (Adam Vandenberg)
Date: Wed, 23 Apr 2003 16:36:02 -0700
Subject: Curses on Win32
References: <b86r8p$27n$1@gateway.northgrum.com>
Message-ID: <vae8s19g0aug3f@corp.supernews.com>

You can take a look at:
http://flangy.com/dev/python/curses/

So far I have a binary distribution for Python 2.2.2, but I'll be releasing
the source soonish.
Note that in no way do I recommend this module for production work, as it is
buggy and incomplete, but it might be enough to get you going, and once the
source is out there you can of course modify it to your heart's content.

-Adam Vandenberg






From andrew-pythonlist at puzzling.org  Tue Apr 29 10:13:10 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 30 Apr 2003 00:13:10 +1000
Subject: POD vs. reST for standalone writing?
In-Reply-To: <3EAE801D.4010005@python.org>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <mailman.1051584944.30800.python-list@python.org> <b8kuet$ftn$1@panix1.panix.com> <20030429043815.GB13813@frobozz.local> <3EAE801D.4010005@python.org>
Message-ID: <20030429141310.GA14632@frobozz.local>

On Tue, Apr 29, 2003 at 09:37:33AM -0400, David Goodger wrote:
> Andrew Bennetts wrote:
> >Hmm, interesting.  I thought the :Foo: syntax was for definition lists, not
> >inline in text?
> 
> As a block-level device, it's used for field lists:
> 
>     :Author: Me
>     :Date: Today

Right, that's what I meant by "definition lists".

> It's also used in inline text, to extend interpreted text roles 
> (explicit markup):
> 
>     Here is some `interpreted text`:role:.
> 
> Both uses attach a label or classifier to something.

Ah, ok.  I didn't know that form.

[snip discussion of how my example might be supported with future
enhancements to reStructuredText]

> However, the above is not typical reStructuredText usage.  When a 
> document gets to that level of complexity, it may be beyond what 
> reStructuredText aims to support.  ReStructuredText doesn't aim to be a 
> solution to every problem, just to the most common 80%-90%.

Exactly :)

I don't see Lore and reStructuredText as being competitors -- they're aimed
at different "markets" (although there may be a slight overlap?).

> >And how would I make the ReST tools link that to my API documentation
> >generated with pydoc, epydoc, or whatever?
> 
> Currently, with URL links.  Again, there's been some discussion about 
> using other methods (perhaps XLink/XPath-based), but no implementation 
> yet.  How does Lore link to outside docs?

By default, it knows how to link to epydoc-generated API docs[1].  A typical
invocation might be:

    lore --config baseurl=http://mysite/doc/api/%s.html mydocument.html

The "%s" will be substituted for the API name.  For other URL schemes, you
could write your own output plugin that overrides the processing of that
tag.  Ideally, I'd like to include at least vanilla pydoc/happydoc support
that you could access with --config apitype=pydoc or similar.  Potentially,
this sort of information could be embedded in the template rather than
passed as --config options.

I should note that Lore is still fairly young, and while it's settled down
quite alot, it is still undergoing a fair bit of development, so there are
some rough edges.  If this bothers you, wait a few months :)

> >(Btw, don't get me wrong; I *like* ReST for some uses.  I just think it's
> >better suited to wikis than to reference manuals and the like.  I don't
> >really like ReST's liberal use of punctuation in large doses.)
> 
> The "liberal use of punctuation in large doses" only really kicks in for 
> advanced uses.  Personally, I find occasional punctuation easier to 
> ignore than <ubiquitous> <angle> <brackets>.  To each his own :-).

Yep.  For Twisted's documentation (now over 200 pages as a PDF), I think
Lore was the right choice (even if we did have to invent it first).  Other
users will have to make their own decisions about what's right for them.
Much of the time, I'm sure that reStructuredText will be the choice -- I
hope that sometimes, someone will find Lore useful.

I must admit I'm not a great fan of <lots> <of> <angle> <brackets> either,
but Lore isn't too bad most of the time.  And while it's a little more
verbose, there's also a significant benefit to using tags rather than syntax
for markup: you can look up a tag name or an attribute value in
documentation far more easily than you can arbitrary syntax.  You'll never
look at Lore source and go, "now what the hell does ::..__"xyzzy"==::
mean?!" ;)

-Andrew.

[1] Except we use a small hack to epydoc to make it generate easier to
    predict filenames.  By default it uses "x.y.z-class.html" or
    "x.y.z-function.html", and so on, but it's more convenient if all Lore
    needs to link to is "x.y.z.html", so we make it do that.





From BPettersen at NAREX.com  Wed Apr 23 15:12:05 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 23 Apr 2003 13:12:05 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE0C5@admin56.narex.com>

> From: Andrew Dalke [mailto:adalke at mindspring.com] 
> 
> Steven Taschuk
> > __and__ could be passed other as a callable which computes and
> > returns the second operand.  Then the standard __and__ could be
> >     def __and__(self, other):
> >         if bool(self):
> >             return self
> >         return other()
> 
> So a new idiom and new code (to generate a function from a set of
> unevaluated code), just to allow overriding booleans?  Seems
> rather excessive to me.

It occurred to me that C++ allows overloading of && (non c-s), while the
regular && is c-s.

Could it be that c-s is only useful in domains with exactly two elements
(true, false). I.e. don't worry about c-s for overloaded __and__ since
(a) the class you're overloading it on will probably not be limited to
two instances, and (b) even if it was, how would you know c-s is
desireable or even appropriate?

-- bjorn




From duncan at NOSPAMrcp.co.uk  Wed Apr  9 06:09:50 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 9 Apr 2003 10:09:50 +0000 (UTC)
Subject: atexit functionality
References: <mailman.1049821814.20582.python-list@python.org>
Message-ID: <Xns935871190AFC8duncanrcpcouk@127.0.0.1>

[posted and mailed]

viszneki at telerama.com wrote in
news:mailman.1049821814.20582.python-list at python.org: 

> return _exithandlers[len(_exithandlers)]

> I hope my ideas are well received! I think they will be because this 
> modification is fully backward compatible, and doesn't introduce any
> negative side effects or anything like that.
> 
In what way is asking for atexit.register to throw an IndexError exception 
every time it is called NOT introducing a negative side effect?

If you want to muck about with atexit functions the way to do it is to 
register your own function once, and have that maintain its own list that 
you can modify to your heart's content.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From max at alcyone.com  Sun Apr 27 04:22:03 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sun, 27 Apr 2003 01:22:03 -0700
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <U0Mqa.13860$K35.457281@news2.tin.it>
Message-ID: <3EAB932B.A5B42BE0@alcyone.com>

Alex Martelli wrote:

> I think "if f is None:" and even more "if callable(f):" are
> better ways to express this test -- explicit is better than
> implicit.

I agree wholeheartedly.  I was just giving an real-world example of use
where the "Booleanness" of a function-or-maybe-not-a-function would be
useful.

I personally explicitly do "x is None" tests for anything when I'm
checking for Noneness, this was just as an example.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The tremor of awe is the best in man.
\__/ Goethe
    Erik Max Francis' bookmarks / http://www.alcyone.com/max/links/
 A highly categorized list of Web links.



From mcfletch at rogers.com  Tue Apr 15 21:31:35 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 15 Apr 2003 21:31:35 -0400
Subject: Big-O notation
In-Reply-To: <1050441790.11992.18.camel@sayge.arc.nasa.gov>
References: <qGqla.61$Tj7.91655@news.uswest.net>	 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>	 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>	 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>	 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>	 <3E9C6C62.8010808@rogers.com> <1050441790.11992.18.camel@sayge.arc.nasa.gov>
Message-ID: <3E9CB277.8040409@rogers.com>

Chad Netzer wrote:

>On Tue, 2003-04-15 at 13:32, Mike C. Fletcher wrote:
>  
>
>>Greg Brondo wrote:
>>    
>>
...

>>    * O(1) -- time is linear
>>    
>>
>
>You meant "time is constant".  The algorithm requires the same number of
>operations regardless of input size.
>  
>
Yup, absolutely correct, though in my defense, a horizontal line is 
linear too ;) . 

Don't post when tired (I just took a nap :) )!

>>    * O(N log N) -- see O(N) and O(log N), basically, you do an amount
>>      of work tied to the size of the whole set for every stage in an O(
>>      log N ) operation (this is comparatively rarely discussed in my
>>      experience).
>>    
>>
>
>Really?  Some of the most important classes of algorithms are O( N log N
>), specifically, many common sort algorithms.  
>
Ah, but I never took comp-sci, so I never got into those cool "what's 
the best sort algorithm" debates ;) .  Most of the debates I've gotten 
into this way have been around choices such as binary trees vs. hash 
tables, so that's probably skewed my experience.

...

>Also, one further class that is import to know is O( 2**N ), such as the
>brute-force solution to the "travelling salesman problem".  These are
>much, MUCH worse than O( N**2 ) algorithms, and even today can rarely be
>computed for problems sets much larger than 100 (whereas O( N**2 )
>problem sets can grow to thousands or perhaps even millions and still be
>practically solved).  Rather than computing "solutions" with these types
>of algorithms, it is required to instead come up with "close-enough" or
>"best guess" type algorithms that CAN be computed practically on large
>data sets.
>  
>
Cool.  Haven't come across an algorithm described as such, though I'd 
guess there've been a few that matched it over the years. 

Thanks for all the clarifications and corrections,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From aahz at pythoncraft.com  Wed Apr 30 10:43:52 2003
From: aahz at pythoncraft.com (Aahz)
Date: 30 Apr 2003 10:43:52 -0400
Subject: How to call built in function 'open' from inside object that has 'open' method?
References: <db81c492.0304300526.7b8d3c86@posting.google.com> <mailman.1051711257.436.python-list@python.org>
Message-ID: <b8onf8$b7d$1@panix2.panix.com>

In article <mailman.1051711257.436.python-list at python.org>,
Fredrik Lundh <fredrik at pythonware.com> wrote:
>"vadim" <vf at interna.ru> wrote:
>>
>> I am new to Python and do not know the answer to very simple how to
>> 'open' from inside another object that has 'open' method. I know that
>> in C++ I should use following syntax ::open.
>>
>> More details on reasons. I have met this problem trying to write a
>> simple script use ActiveScripting inside IE.  I am trying "file =
>> open(...) but because context window object has 'open' method that
>> just opens window it is called first.
>
>Unless you're typing "file = self.open(...)", the class you're in has
>nothing to do with it.

Not quite true; what happens with

    class C:
        def open(self, name):
            f = open(name)

(Which is how I interpreted vadim, possibly incorrectly.)
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Sat Apr 19 16:21:13 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Sat, 19 Apr 2003 22:21:13 +0200
Subject: arg-passing style question
In-Reply-To: <A4ioa.87812$Vzu.36931@news02.bloor.is.net.cable.rogers.com>
References: <A4ioa.87812$Vzu.36931@news02.bloor.is.net.cable.rogers.com>
Message-ID: <3ea1afb8$0$49117$e4fe514c@news.xs4all.nl>

Guy Middleton wrote:
> I want to pass two values to a function that expects a single argument.
> 
> So, I could wrap them in either a list or a tuple.  Is either way preferred
> over the other?
> 
> 
> queue = Queue.Queue()
> a = "foo"
> b = {"key": "value"}
> queue.put((a,b))	# this works
> queue.put([a,b])	# this works too


This depends on what you do with the argument.
Keep in mind that a tuple is immutable.
A list object is mutable.

So if you want to change the objects that you put in the queue
(your example above), you cannot use a tuple.

--Irmen




From sismex01 at hebmex.com  Tue Apr 29 15:27:25 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Tue, 29 Apr 2003 14:27:25 -0500
Subject: Problem subclassing tuple
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F062330489D915@mail.hebmex.com>

> From: John Wilson [mailto:tug at wilson.co.uk]
> Sent: Tuesday, April 29, 2003 2:22 PM
> 
> Success!!!
> 
> class Holder(tuple):
>     def __new__(cls, *args, **kargs):
>         return tuple.__new__(cls, args)
> 
> 
> is the way to do it for a tuple :))))))
> 
> John Wilson
>

Heh, excellent :-)

You might want to make it "cooperative", that is,
instead of using "tuple.__new__", use
"super(Holder,cls).__new__"

This delegates the search for the superclass to the
function super(); and that gives you the ability
to have multiple inheritance, since super() will
"do the right thing" when getting your current
class' superclass; so every class which defines
a __new__ method has it called.

:-)

Like this:

Python 2.2 (#28, Dec 21 2001, 12:21:22) [MSC 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> class Holder(tuple):
...   def __new__(klass, *args, **kwargs):
...     return super(Holder,klass).__new__(klass, args)
...
>>> h = Holder(1,2,3)
>>>
>>> h
(1, 2, 3)
>>> type(h)
<class '__main__.Holder'>
>>>


-gus




From abpillai at lycos.com  Thu Apr 24 05:21:33 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 24 Apr 2003 02:21:33 -0700
Subject: sys.argv problem in Win 2k
Message-ID: <bd993a2f.0304240121.7c795cf0@posting.google.com>

Hi pythonistas

       All my python scripts on Win 2k does not take command line arguments.
i.e, len(sys.argv) is always 1. Because of this I cannot use any scripts which
takes cmd line arguments.

 Any ideas ?

Regards

Anand Pillai



From yorklee70 at yahoo.com  Tue Apr 29 12:16:31 2003
From: yorklee70 at yahoo.com (York)
Date: Tue, 29 Apr 2003 17:16:31 +0100
Subject: The behavior of List.insert
Message-ID: <b8m8dn$b03$1@pegasus.csx.cam.ac.uk>

Hi, 

given

mylist = [1,2,3,4,5], 

mylist.insert(-1,99) will make mylist to be [99,1,2,3,4,5]

just wonder why not let it to be [1,2,3,4,99,5]

Thanks,

York



From maxm at mxm.dk  Wed Apr  9 04:03:11 2003
From: maxm at mxm.dk (Max M)
Date: Wed, 09 Apr 2003 10:03:11 +0200
Subject: String splitting question
In-Reply-To: <b70icc$9n0ff$1@ID-99293.news.dfncis.de>
References: <9afd3f36.0304082245.33670341@posting.google.com> <b70icc$9n0ff$1@ID-99293.news.dfncis.de>
Message-ID: <LqQka.8341$y3.1005660@news010.worldonline.dk>

William Park wrote:
> Jim Shady <iamshady at rediffmail.com> wrote:

>>abcd/df/a/iiwk/abcdefghijkl/b/c
>>
>>I need to get the longest string between the /s of the string. For
>>example, longest_str() for the above line should return
>>'abcdefghijkl'. How do I go about doing this?

> out = []
> for i in 'abcd/df/a/iiwk/abcdefghijkl/b/c'.split('/'):
>     out.append( (len(i), i) )
> print max(out)[1]

Or the ultra terse version, which basically does the same thing:

s = 'abcd/df/a/iiwk/abcdefghijkl/b/c'
print max([(len(sub), sub) for sub in s.split('/')])[-1]

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From abhisek_hazarika at yahoo.com  Fri Apr  4 08:01:42 2003
From: abhisek_hazarika at yahoo.com (Abhi Haz)
Date: 4 Apr 2003 05:01:42 -0800
Subject: tkMessageBox returns 0 errorneously the first time -URGENT-Plz help ....
Message-ID: <6bf71308.0304040501.58c4bdee@posting.google.com>

I am using Tkinter interface for Python1.5 ...
On a particular quit routine...I'm doing the following-

******************************************************
...all imports go here...

...some app...from where onQuit is invoked if user presses a ceertain "quit"
button on the toolbar of the app....

def onQuit(self):
ans = tkMessageBox.askokcancel('Verify','Are you sure you want to quit?')
if ans:
    self.quit()


*******************************************************
The strange thing about this is that the first time this is invoked 
no matter what the user chooses-ok or cancel, ans = 0
The next invocation onwards picks the value correctly....

Python,Tk gurus...your help reqd....

Rgds
Abhi



From alloydflanagan at attbi.com  Wed Apr 16 17:41:28 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 16 Apr 2003 14:41:28 -0700
Subject: Using property in classic class may not work
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it> <a8b7f07a.0304150527.5b0b1187@posting.google.com> <hRUma.23848$T34.709378@news2.tin.it> <b7hies$ol8$0@216.39.172.122>
Message-ID: <a8b7f07a.0304161341.c84eefb@posting.google.com>

bokr at oz.net (Bengt Richter) wrote in message news:<b7hies$ol8$0 at 216.39.172.122>...
[lots of missing stuff]
> I may be fooling myself here, but a quick experiment suggests being
> able to define a new class that inherits from both a classic class
> that you can't touch and a new-style one that you can give a property, e.g.:
> 
>  Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
>  Type "help", "copyright", "credits" or "license" for more information.
>  >>> class Old3rdParty:
>  ...     def oldMethod(self): print 'oldMethod was called'
>  ...
>  >>> class NewPropertyCarrier(object):
>  ...     def __init__(self, v): self.prop = v
>  ...     def _get_p(self): return self._v
>  ...     def _set_p(self, v): self._v = v
>  ...     prop = property(_get_p, _set_p)
>  ...
>  >>> class Derived(Old3rdParty, NewPropertyCarrier):
>  ...     pass
>  ...
>  >>> d = Derived(123)
>  >>> d
>  <__main__.Derived object at 0x007D68E0>
>  >>> d.prop
>  123
>  >>> d.oldMethod()
>  oldMethod was called
>  >>> d.prop = 456
>  >>> d.prop
>  456
> 
> It seems like prop is surviving as a property, not getting clobbered into an attribute:
> 
>  >>> dir(d)
>  ['__class__', '__delattr__', '__dict__', '__doc__', '__getattribute__', '__hash__', '__init__',
>  '__module__', '__new__', '__reduce__', '__repr__', '__setattr__', '__str__', '__weakref__', '_ge
>  t_p', '_set_p', '_v', 'oldMethod', 'prop']
>  >>> vars(d)
>  {'_v': 456}
>  >>> d._v
>  456
>  >>> d.prop
>  456
>  >>> d._v = 'sneaky'
>  >>> d.prop
>  'sneaky'
>  >>> d.prop = 'normal'
>  >>> d.prop
>  'normal'
>  >>> d._v
>  'normal'
> 
> So could wxPython play the role of Old3rdParty above and allow building a Derived with
> the properties desired? Or is mixed bases a no-no?
> 
> Regards,
> Bengt Richter

That's a very clever idea.  I have no idea if it will work or not, but
definitely worth a try :)



From zephyr01 at alltel.net  Fri Apr 25 13:06:24 2003
From: zephyr01 at alltel.net (Hans Nowak)
Date: Fri, 25 Apr 2003 13:06:24 -0400
Subject: Vim Newbie ? Python Edit/Compile/Run  cycle and integrating py
 help questions
In-Reply-To: <cb035744.0304250832.71aef2d9@posting.google.com>
References: <a791ab19.0304240532.5a025b81@posting.google.com> <3EA7EBD3.3010605@ghaering.de> <mailman.1051243975.12570.python-list@python.org> <cb035744.0304250832.71aef2d9@posting.google.com>
Message-ID: <b8bpf6$sgv$1@main.gmane.org>

Will Stuyvesant wrote:
> [Hans Nowak]
> 
>>I have been trying to edit Python source code with vim on Mac OS X.  Most of 
>>the settings in my .vimrc file (copied from Windows) work just fine, except for 
>>one.  I use Tab to indent and Shift-Tab to dedent (a line, not a block).  Tab 
>>behaves as usual, but Shift-Tab doesn't seem to work; it just behaves like I 
>>pressed Tab.
> 
> 
> In gvim on windows and linux (untested on Mac, since I have no access
> to a Mac anywhere):
> 
> Start block selection with the v key, or with <SHIFT>v.  And then of
> course the j key for line down and the l key for going right etc.
> 
> To indent a line, or a selected block, use <SHIFT>>>.  To dedent it do
> <SHIFT><<.

Thanks for your reply; however, I know how to do indents and dedents in vim, 
the problem is more how to map Shift-Tab to do the dedent.  The reason is 
simply that I have been using it for years and don't want to unlearn it and get 
used to a different key combination. :(

I'll see if I can do something with the key combinations you mentioned.

Cheers,






From mwh at python.net  Sun Apr 20 06:03:19 2003
From: mwh at python.net (Michael Hudson)
Date: Sun, 20 Apr 2003 10:03:19 GMT
Subject: Sorting without transitivity
References: <20030419213405.GA588@obstler> <mailman.1050821403.13707.python-list@python.org>
Message-ID: <7h3y925o596.fsf@pc150.maths.bris.ac.uk>

Steven Taschuk <staschuk at telusplanet.net> writes:

> Martelli's already given the magic words "topological sort"; this
> post is just about the terminology.  (All from memory; corrections
> welcomed.)

Do you (or anyone else) know why it's called a *topological* sort?
Don't see no topology here, off hand.

It could be that the open subsets of a topological space are a poset
under inclusion, I guess.

Cheers,
M.

-- 
 "Well, the old ones go Mmmmmbbbbzzzzttteeeeeep as they start up and
  the new ones go whupwhupwhupwhooopwhooooopwhooooooommmmmmmmmm."
                         -- Graham Reed explains subway engines on asr



From ramen at lackingtalent.com  Tue Apr 15 19:17:26 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 15 Apr 2003 23:17:26 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk>
Message-ID: <slrnb9p4oo.vg8.ramen@lackingtalent.com>

In article <yfs7k9vcu2p.fsf at black132.ex.ac.uk>, Alexander Schmolck wrote:
> I think your life might be much happier if you just went to
> http://www.drscheme.org/, dowloaded Dr Scheme and used *that* for
> webdevelopment. Then you can use a real functional programming language and
> read books by people who actually have a real clue about functional
> programming (there is lots of *excellent* material available on-line and for
> free, like SICP and HTDP, the latter uses Dr Scheme -- also have a look at
> http://www.cs.brown.edu/courses/cs173/2001/Lectures/). Why settle for some
> second-rate immitation? [1]

I may just take your advice and investigate that. Thanks for the reference.
But keep in mind that I'm not saying "I don't like Python, it's not
functional enough". I'm saying "I love Python, it lets me work in both an FP
and OOP style, and this makes me happy. I just don't want to see FP support
drop out of the picture."

> (Dr Scheme *does* have web libraries, I've never used them but I'd suspect
> they shouldn't be too bad, because they seem to have hit on trick of using
> webdevelopment as application domain to bring things like continuations to the
> unwashed masses (viz. their sophomores) :).

Interesting.

> Ain't gonna happen. Anyway you can always replace 
> 
>   foo(lambda x: x+3)
> with: 
>   def bar(x): return x + 3
>   foo(bar)
> 

That's true, but then I have to name "bar". And part of the joy of FP is not
having to name *everything*. Naming things is important if you need to call
them up later, but it's a hassle to name intermediate values, just as it
would be a hassle to have to say:

result = list()
result.append(x)
result.append(y)
result.append(z)
return result

When Python lets you instead say:

return [x, y, z]

>> keyword-argument trick wasn't really that bad, once you got used to it. In
>> any case, with dynamic scoping, this problem has completely disappeared.
>                  ^^^^^^^^^^^^^^^
> not quite.

Am I using the wrong terminology here? I'm specifically talking about the
named-argument hack to get around Python's previous lack of nested scopes.
This specific problem, as far as I know, is gone now. Am I missing something?

> Only that python won't let you use rather helpful things like ``if`` as
> an expression, but never mind.

Well, this leads us to the ternary operator thing, which I'll let other
people argue about. ;) I'm happy with my short-circuiting and/or for now.

>> In fact, I believe that it is advantageous to design functions specifically
>> for use with map and filter. If you design this way, you will benefit from
>> cleaner syntax than list comprehensions can offer. For example:
> 
> I wonder how a function written for map would look different from one written
> for a list comprehension...

Well, I don't think people usually write functions for list comprehensions
at all. I've never done it. I've never made it an explicit part of my design
for an API that someone use my functions in list comprehensions. But this
could just be because I'm still learning where they fit in best.

However, I have written an API that consisted of functions that curried
their own arguments and could be used in a map, like this:

>>> addx = lambda x: lambda y: x + y
>>> map(addx(5), range(5))
[5, 6, 7, 8, 9]

>> Rather than remove these important functional built-ins, I would like to see
>> two last built-ins be added: unzip (to reverse the behavior of zip), and
> 
> There is already an unzip function and it's written ``zip(*...)``

Thanks, I didn't know this.

>> curry (see the example by Alex Martelli, et. al. in the Python Cookbook,
> 
> Why not just use lambda? Python's argument passing is fairly complex, so I'm
> not sure whether one (or rather two) curry function(s) would be worth the pain
> (at least not if you tell your emacs to display lambda as greek character).

Because it's a common operation in some contexts, and by naming it, its
behavior can be formalized. You can of course write curried functions, and
even self-currying functions (like my "addx" example above) with lambdas.

Perhaps it should go in the "functional" module, along with "compose" and
all that other fun stuff. I still like map/filter being global, though...

Thanks,
Dave




From skip at pobox.com  Wed Apr  2 13:00:32 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 2 Apr 2003 12:00:32 -0600
Subject: Socket timeouts for HTTP connections (httplib)?
In-Reply-To: <s18m8v4tkt4jgp02ntu85qe886g86j93oe@4ax.com>
References: <9df3d134.0304020550.7e32bf6b@posting.google.com>
        <s18m8v4tkt4jgp02ntu85qe886g86j93oe@4ax.com>
Message-ID: <16011.9536.757953.437769@montanaro.dyndns.org>

    Rene> It's in 2.3:
    Rene> http://www.python.org/2.3/highlights.html

    Rene> But I don't know if and how this ripples up to higher level APIs.

Turns out it did.  The socket's makefile() method dup()s the file
descriptor.  Timeouts during select were only being set on the original
socket file descriptor, so any protocols (most text-based ones) which called
conn.makefile() then proceeded use the resulting file object would bomb.

The fix turned out to be surprisingly easy.  Some systems (Windows among
them), didn't have dup(), so makefile() couldn't be implemented in the
straightforward fashion.  A workaround was added to socket.py for those
platforms.  I simply extended the workaround to all platforms.  Now
makefile() returns a file-like object (not a file) which chit-chats directly
through the un-dup()'d socket, and timeouts work as expected.

Skip





From mgerrans at mindspring.com  Thu Apr 10 01:40:55 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Wed, 9 Apr 2003 22:40:55 -0700
Subject: BEGINNER: converting integer to short month name?
References: <OF47A94545.30289BFA-ON85256D03.006E4C80-85256D03.006FA823@seminole.fl.us> <mailman.1049921355.15625.python-list@python.org>
Message-ID: <b73066$evv$1@slb9.atl.mindspring.net>

>     [ 'Dec','Jan','Feb','Mar', ... ] [ the_month ]
>
> or
>
>     import calendar
>     calendar.month_abbr[ the_month ]

Looks like it would be
   calendar.month_abbr[ the_month + 1 ]

Interestingly, calendar.month_abbr[ 0 ] is 'Saturday'.   Hmm...






From paul at boddie.net  Wed Apr 23 07:02:02 2003
From: paul at boddie.net (Paul Boddie)
Date: 23 Apr 2003 04:02:02 -0700
Subject: pyxml minidom: I can remove and append, but not replace
References: <3ea5562f$0$22541$b45e6eb0@senator-bedfellow.mit.edu>
Message-ID: <23891c90.0304230302.62a89bf4@posting.google.com>

reagle at mit.edu (Joseph M . Reagle Jr.) wrote in message news:<3ea5562f$0$22541$b45e6eb0 at senator-bedfellow.mit.edu>...
> I have a snippet that replacing a title element with a new element with a
> modified one (base on an amazon query) -- if that title element already
> exists. However, I can't do an insertBefore or replace for some odd reason:

[...]

>                     if feature:
>                         book.appendChild(r_title)
>                         book.removeChild(feature)
> #                         book.replaceChild(feature,r_title)

[...]

> xml.dom.NotFoundErr: Node does not exist in this context

That's because replaceChild expects the replacement node first in the
argument list, followed by the node to be replaced. What your
replaceChild call is attempting to do is to replace the uninserted
node with the existing node, and obviously, the uninserted node is not
yet present in the document.

> The node is using the same namespace, and I've used the replaceChild in other
> cases (where I'm actually using a node from a different XML instance!)

Strictly, you should use the document node's importNode method on
nodes from other instances before introducing those nodes into the
document.

Paul



From matt at teaphoo.fsnet.co.uk  Fri Apr  4 17:32:51 2003
From: matt at teaphoo.fsnet.co.uk (Matt Russell)
Date: 4 Apr 2003 14:32:51 -0800
Subject: FTP : get  dir name with space caracters embedded  + is ther something quicker than LIST
References: <b1ab591.0304030143.52422a3e@posting.google.com>
Message-ID: <459f9b68.0304041432.3451f57b@posting.google.com>

I think FTP.nlst([ directory ])
solves your problem.
you may be able to use this method in conjunctio n with os.path.walk
to achieve your goal.
Matt

s.thuriez at laposte.net (sebastien) wrote in message news:<b1ab591.0304030143.52422a3e at posting.google.com>...
> Hi ,
> 
> I have 2 problems regarding FTP methods.
> 
> 
> My GOAL :
> ---------
> 
> I am looking for a specific file on a huge ftp server. The structure
> of the ftp server is not known. I have to determine which are
> directories and which are files so that it can recursively go into the
> subdirectories to scan the files (it is multi threaded).
> 
> It is working well but I have 2 problems.
> 
> 
>  My trouble :
> -----------------
> 
> It am using the ftp.retrlines('LIST',liste.append)
> and analyse which one are directories since they have the letter d at
> the beginning of the line.
> 
> 1) The LIST METHOD IS EXTREMLY SLOW !!!
> 
> In certain directories there is thousand of file and the LIST method
> is extremely slow.
> It works but is there other method to get the directories without
> having to do a LIST ?
> 
> 2) DIRECTORY NAME WITH SPACES  
> 
> LIST is slow but it works. One other problem is the directory name in
> which there is spaces : like
> "data file today".
>  I can do a :
> elements= line.split(" "), 
> 
> but the data before my directory name has not the same format : for
> example :
> 
> .....
> may 22   2001     name of directory
> mar 15    02:00   other directory name
> .....
> 
> I do not know how to get the directory name simply.
> 
> One method would be to get a len(elements) and if len(elements) > 9
> for example on my server. All the last elements would be part of
> directory name. It is not so clean since it may depend on the FTP
> server configuration.
> 
> Do you have any idea ?
> 
> Thanks for having taken the time to read my mail.
> 
> Sebastien.



From theller at python.net  Wed Apr 30 05:18:52 2003
From: theller at python.net (Thomas Heller)
Date: 30 Apr 2003 11:18:52 +0200
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304251030.30fd1f94@posting.google.com> <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk> <mailman.1051635832.26364.python-list@python.org> <7h3el3kgy4j.fsf@pc150.maths.bris.ac.uk>
Message-ID: <k7dc1gxv.fsf@python.net>

Michael Hudson <mwh at python.net> writes:

> Steven Taschuk <staschuk at telusplanet.net> writes:
> 
> > Quoth Michael Hudson:
> > > mis6 at pitt.edu (Michele Simionato) writes:
> >   [...]
> > > > 2) why you say this makes classes unhashable ?
> > > 
> > > Well, I thought it would.  I was wrong:
> >   [...]
> > > Then you get this kind of weirdness:
> > > 
> > > />> class C(type):
> > > |..     def __eq__(self, other):
> > > |..         return 1
> > > \__ 
> > 
> > Can't the instances of this metaclass be made sensible for use in
> > dicts simply with
> >     def __hash__(self):
> >         return hash(C)
> > for example?  (To ensure that if x == y then hash(x) == hash(y),
> > as desired.)  
> 
> Yes!  The point I was making was that *not* doing that *doesn't* lead
> to error messages.
> 
> Hang on...
> 
> />> class C(object):
> |..  def __eq__(self, other):
> |..   return 1
> \__ 
> ->> hash(C())
> 137007476
> />> class C:
> |..  def __eq__(self, other):
> |..   return 1
> \__ 
> ->> hash(C())
> Traceback (most recent call last):
>   File "<input>", line 1, in ?
> TypeError: unhashable instance
> 
> Hmm.  I wonder if this is deliberate...
> 

See also http://www.python.org/sf/660098

Thomas



From rmunn at pobox.com  Fri Apr 25 12:21:29 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 25 Apr 2003 16:21:29 GMT
Subject: ActivePython Scripts help with ?
References: <b89bu8$5bp$1@newsg2.svr.pol.co.uk>
Message-ID: <slrnbai3cd.84b.rmunn@localhost.localdomain>

Hyperion <Sorry at ButNo.com> wrote:
[snip]
> 
> Wll I maniged to install it fine, but now I haven't got the slightest clue
> what to put at the top of my scripts so the interpreter or what ever can
> find them.
>  You no the little bit at the top that starts with,  #!
> 
> The ActivePython Install is setup here as follows:
> 
> ---------------
> /home/me/ActivePython-2.2/bin
> ---------------

In that case the Python interpreter should be:

    /home/me/ActivePython-2.2/bin/python

So at the start of your scripts you would put:

    #!/home/me/ActivePython-2.2/bin/python

Also don't forget to make your scripts executable ("chmod +x script.py")
or you won't be able to run them even with the right #! line.

> . I added that line to me PATH also in my /etc/profile  file as follows:
> 
> --------------------
> # Set the default system $PATH:
> PATH="/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/usr/games:/home/me/ActivePython-2.2/bin"
[snip]
> ------------------

Another alternative, since the python interpreter is in your PATH, is to
use /usr/bin/env to find it. This trick can come be handy if you want
other people to be able to use your python scripts. Most people put
python in /usr/bin or /usr/local/bin, so they would have to change the
#! line in your scripts if you use "#!/home/me/...". But if you put this
line at the start of your scripts:

    #!/usr/bin/env python

Then people who didn't install Python in the same place as you will
still be able to use your scripts verbatim.

> So anyways can someone please help me with what I put in the top of my
> scripts after the , #! ,  so I can run them.
> And if there are any linux guru's out there that could check that PATH thing
> out and see if its ok, I would be greatfull.

Looks to me like you included all the needed directories. One caveat,
though: your system path (used by all users on your system) includes a
part of your home directory (used by only you). Are you planning on
making your home directory world-readable (i.e., readable by other users
on your machine)? If you're the only user on your system, that's not a
big security concern; but you should be aware of the potential security
implications of what you're doing. That's why most people will install
Python in a standard place like /usr/local or /opt. (If you've never
heard of /opt, basically it's designed to work a bit like the Program
Files directory in Windows. Go read section 3.12 of the Filesystem
Hierarchy Standard -- http://www.pathname.com/fhs/2.2/fhs-3.12.html --
if you want all the details).

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From jimmy at retzlaff.com  Tue Apr 15 14:34:32 2003
From: jimmy at retzlaff.com (Jimmy Retzlaff)
Date: Tue, 15 Apr 2003 11:34:32 -0700
Subject: Printing Columns
Message-ID: <A7B5C22F20EE3C4E95744E2F03F1887F0F1D38@MAYOR.retzlaff.com>

Lindstrom Greg - glinds (Greg.Lindstrom at acxiom.com) wrote:
>Given a list of length N with 0<=N<=infinity and a number of columns, C
>(C>0 and may be less than N), how do we print out the elements of the
list
>in the above (column major?) form?

Here's a relatively straightforward approach:

import math

def printInColumns(items, numberOfColumns):
    numberOfRows = int(math.ceil(1.0*len(items)/numberOfColumns))
    for row in range(numberOfRows):
        for column in range(numberOfColumns):
            index = column*numberOfRows + row
            if index < len(items):
                print items[index], '\t',
        print

printInColumns(['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H'], 3)


If you test this a bit you can find some trouble spots (for the problem
in general). For example how do you print 8 items in 7 columns? The
above approach would result in:

A 	C 	E 	G 	
B 	D 	F 	H 	

Is there actually a way to use all 7 columns?

Jimmy




From dwelch91 at no.spam.attbi.com  Fri Apr  4 23:04:13 2003
From: dwelch91 at no.spam.attbi.com (djw)
Date: Sat, 05 Apr 2003 04:04:13 GMT
Subject: pyCUPS still alive?
Message-ID: <1Bsja.72740$OV.186811@rwcrnsc54>

Is the pyCUPS project sill alive? The website says source code coming 
"soon" and its dated April 2000. (Maybe I just answered my own 
question!). Does anybody have the code from this project so that I could 
carry it forward?

Thanks,

Don




From edsZZZ_2 at yahoo.com  Tue Apr 15 20:45:26 2003
From: edsZZZ_2 at yahoo.com (Eric Scharff)
Date: 15 Apr 2003 18:45:26 -0600
Subject: encapsulation checking features
References: <3e9b1298@cs.colorado.edu> <U8Gma.14362$LB6.388431@news1.tin.it>
Message-ID: <3e9ca7a5@cs.colorado.edu>

Thanks for the reference to PyChecker, I'll have to check it out.

I've not read enough python code to know how people traditionally treat
"instance variables that ought to be public" and "instance variables that
ought to be private" (other than prepending underscores, which I never
liked).

> But I really don't see the problem here.  If you want myVar.x to
> call an accessor, you make x into a property of myVar's class,
> period.  Absolutely no need to design trivial do-nothing accessors
> either -- you can perfectly well have all client code access all
> instance variables you want, and if and then such "variables"
> need to be "computed on the fly" or have special effects when
> set, then you make them into properties, hey presto.

I come from a Smalltalk background, which implicitly treats all direct
variable references as private to a class, requiring do-nothing accessors.
The advantage of these in Smalltalk is that I find it makes refactoring
easier - you can ask Smalltalk to tell you everywhere where the "myVar:"
setter is called.  I could grep all of my python code for ".myVar =" but
that technique is slightly less accurate.

>> These are great features to have, but it'd be nice to have warnings to
>> discourage their use at times.

> I can see that for the "ensure all instance variables exist at
> the end of __init__".  I can't see that, given the existence of
> properties, for "no access to instance attributes from outside
> of methods".  Perhaps you can clarify what advantage the latter
> would provide, that properties just can't...?

Here's an example:  I implemented a simulation of something using python
objects that represent objects in the simulation.  In my rapid prototyping,
the objects had two "aspects", the simulation and the visualization.  Every
object had a tkItem instance variable that the object used to configure
using Tkinter.  Now, I want to create an easy way to enable and disable the
views, or have more flexible views.  Turning off the visualization means
finding all the references to tkItems and changing their structure.

Yes, it was badly written to start, but it made sense at the time. :)  Now
I want to get as much built-in help for refactoring as possible.

-Eric
 



From pc451 at yahoo.com  Wed Apr 16 21:32:06 2003
From: pc451 at yahoo.com (Peter Clark)
Date: 16 Apr 2003 18:32:06 -0700
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com> <3e9daad2$0$49115$e4fe514c@news.xs4all.nl>
Message-ID: <f7199550.0304161732.48c75c42@posting.google.com>

Irmen de Jong <irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> wrote in message news:<3e9daad2$0$49115$e4fe514c at news.xs4all.nl>...

> The "degree" symbol is chr(176) in what character encoding?
> Certainly not in UTF-8. Perhaps in windows cp-1252 or ISO-8859-1.
    Yes, latin-1.
 
> (if you look what is written you'll see that the character
> is encoded in two bytes when using UTF-8: 0xc2,0xb0)
    This works in the shell, expect that 'print
degreeChar.encode('UTF-8')' returns two characters, '??' Well,
obviously, since it's Unicode being displayed in a latin-1
environment. (Strictly speaking, sys.getdefaultencoding() returns
'ascii', but the interactive prompt still prints characters greater
than 128.) However, when I put it into the code, I get this error:

    w += [deg.encode('UTF-8') + scale.strip()]
UnicodeError: ASCII decoding error: ordinal not in range(128)

    Since the output is meant to be read to be displayed by a font
which is in essentially latin-1 encoding, I need to restrict the
manner in which the degree symbol is displayed to one byte. Yet I
cannot get it to behave, even though 'print chr(176) works perfectly
fine at the prompt. My suspicion is that the default encoding of the
system is messing python up somewhere along the way--is there any way
to tell it to just print the stupid character and not be concerned
with the output?



From staschuk at telusplanet.net  Wed Apr 30 01:04:47 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 23:04:47 -0600
Subject: Raising objects
In-Reply-To: <tpltav0bbeehseajdo3osampi7flsvhk6g@4ax.com>; from jolderdissen@ixiacom.com on Tue, Apr 29, 2003 at 12:55:03PM -0700
References: <453d10b2.0304290750.222fb453@posting.google.com> <tpltav0bbeehseajdo3osampi7flsvhk6g@4ax.com>
Message-ID: <20030429230447.C3128@tibia.amotlpaa.bogus>

Quoth Jan Olderdissen:
  [...]
> That said, I question the point of deriving from object as classes
> (and in fact all objects as far as I can tell) are derived from object
> as it is. [...]

Depends what you mean by "derived".  On the one hand, both old-
and new-style instances are instances of object according to
isinstance():

    >>> class OldStyle: pass
    ... 
    >>> class NewStyle(object): pass
    ... 
    >>> o = OldStyle()
    >>> n = NewStyle()
    >>> isinstance(o, object)
    1
    >>> isinstance(n, object)
    1

But on the other hand, old-style classes are not subclasses of
object:

    >>> issubclass(OldStyle, object)
    0
    >>> issubclass(NewStyle, object)
    1

For example, they do not inherit attributes from object:

    >>> dir(object)
    ['__class__', '__delattr__', '__doc__', '__getattribute__', '__hash__', '__init__', '__new__', '__reduce__', '__repr__', '__setattr__', '__str__']
    >>> n.__reduce__
    <built-in method __reduce__ of NewStyle object at 0x815045c>
    >>> o.__reduce__
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    AttributeError: OldStyle instance has no attribute '__reduce__'

> [...] Since object isn't a class, deriving from it probably is
> bogus to begin with. [...]

Not at all; it's exactly the right way to make a new-style class,
which is (among other things) needed to get properties to work:

Right:

    >>> class NewStyle(object):
    ...     x = property(lambda self: 3)
    ... 
    >>> n = NewStyle()
    >>> n.x
    3
    >>> n.x = 2
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    AttributeError: can't set attribute

Wrong:

    >>> class OldStyle:
    ...     x = property(lambda self: 3)
    ... 
    >>> o = OldStyle()
    >>> o.x
    3
    >>> o.x = 2 # shouldn't be able to do this!
    >>> o.x
    2

What *is* bogus imho is the distinction between old- and new-style
classes.  (It exists for historical reasons; eventually old-style
classes will disappear, I presume and hope.)

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From mwh at python.net  Mon Apr 14 03:58:29 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 14 Apr 2003 07:58:29 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10
Message-ID: <7h3vfxhpl23.fsf@pc150.maths.bris.ac.uk>

Paul Foley <see at below.invalid> writes:

> [It should do that for lexical variables, too...conflating binding and
> assignment the way Python does is a major design error, but it's too
> late to fix it now]

I'm unconvinced about the "major"... but anyway: how would *you* do
it?  It seems unfeasible syntactically, for starters.

Cheers,
M.

-- 
  The Programmer's Quick Guide To Python (Time Machine version):
    You try to shoot yourself in the foot, only to realize that
    there's no need, since Guido thoughtfully shot you in the foot 
    years ago.                     -- Nick Mathewson, comp.lang.python



From bignose-hates-spam at and-zip-does-too.com.au  Tue Apr  8 00:55:09 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Tue, 08 Apr 2003 04:55:09 GMT
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de>
Message-ID: <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake>

On Tue, 08 Apr 2003 15:44:10 +1200, Greg Ewing (using news.cis.dfn.de) wrote:
> It's been suggested before (and I happen to agree) that Python should
> provide a general way of making *relative* references to modules or
> packages.

I continue to be amazed that there's simply no way to say

  import "../foo/ook.py"

as should be possible to implement any well-architected class hierarchy.

Anyone who can show me otherwise gets a jelly baby.

-- 
 \        "If it ain't bust don't fix it is a very sound principle and |
  `\      remains so despite the fact that I have slavishly ignored it |
_o__)                                  all my life."  -- Douglas Adams |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From ialbert at mailblocks.com  Sun Apr 13 16:59:11 2003
From: ialbert at mailblocks.com (Istvan Albert)
Date: Sun, 13 Apr 2003 20:59:11 GMT
Subject: an ugly file-reading pattern
In-Reply-To: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
Message-ID: <zckma.472677$sf5.796082@rwcrnsc52.ops.asp.att.net>

thanks all for the solutions

cheers,

Istvan.




From aleax at aleax.it  Wed Apr 23 04:34:36 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 08:34:36 GMT
Subject: capturing output from os.system() in Windows?
References: <f405d9e7.0304230003.63baab05@posting.google.com>
Message-ID: <wespa.47515$LB6.1087137@news1.tin.it>

David wrote:

> os.system('NET VIEW') returns 0 under windows, although the DOS box
> pops up and the NET VIEW executes correctly.
> 
> Is there some other command I should be using instead? I guess an

Yes, os.popen('NET VIEW') should serve your needs -- it returns a
filelike object, open for reading, with the output of 'NET VIEW'.


Alex




From sholden at holdenweb.com  Tue Apr 29 08:36:57 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 29 Apr 2003 12:36:57 GMT
Subject: Mail and text extraction
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com> <3EA34DAE.EC67C1E3@engcorp.com> <pan.2003.04.21.13.28.15.303279.2294@mindlessSPAM.com> <3EA3DB6B.DAC620BD@engcorp.com> <pan.2003.04.21.22.56.47.757048.1444@mindlessSPAM.com>
Message-ID: <Jlura.27216$WQ4.24329@news2.central.cox.net>

"Dino Levy" <dino_levy at mindlessSPAM.com> wrote in message
news:pan.2003.04.21.22.56.47.757048.1444 at mindlessSPAM.com...
> On Mon, 21 Apr 2003 13:52:11 +0200, Peter Hansen wrote:
> > Oh, it's quite achievable that way.  It's just the wrong way to do it.
> >
> > Mixing in the various operations like that is a form of optimization,
> > and since nothing's actually working yet it's clearly premature, and
> > therefore not good. :-)
>
> Hehe, I certainly agree with this one... But I like to do such things,
> like polishing the code while it does not yet exist, or make hundreds of
> new options DURING the programming. That's why many of my programs never
> see the daylight:)
>
> > MIME and POP3 are orthogonal.  MIME is an encoding, POP3 is a transport
> > protocol (or something).  You can have one, or the other, or both, or
> > neither...
> >
> > By the way, the "quite achievable" part above simply means you could
> > easily build this capability into custom software.  The "wrong way" just
> > means that would likely be a large waste of time, given that the
> > individual pieces already exist in standard modules, quite well tested
> > and debugged.  Just combine them in appropriate ways and keep life
> > simple.
> >
> > -Peter
>
> Sooorry, Mea culpa. I overlooked your sentence and wrote it
> automatically. I was actually thinking about IMAP. I don't want IMAP, I
> want POP3. Sorry.
> I'm going to experiment with this downloading the message first and then
> parsing it method...

Also don't forget that any solution that breaks the socket link with the
POP3 server in an uncontrolled manner will likely leave that mailbox locked
for some unspecified period, which might not be what you want.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From duncan at NOSPAMrcp.co.uk  Wed Apr 30 07:06:03 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 30 Apr 2003 11:06:03 +0000 (UTC)
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com> <b8lpof$q0f$1@pegasus.csx.cam.ac.uk> <Xns936C8AB67B739duncanrcpcouk@127.0.0.1> <vatlfeb16jlled@corp.supernews.com> <Xns936D5DA488705duncanrcpcouk@127.0.0.1> <bryo1gj6.fsf@python.net>
Message-ID: <Xns936D7A5A717Aduncanrcpcouk@127.0.0.1>

Thomas Heller <theller at python.net> wrote in news:bryo1gj6.fsf at python.net:

> Normally this is used to implement "callback functions" in Python, to
> call functions like Windows' EnumWindows(), or to implement COM object
> method tables.
> 
> Does this count as embedding? It seems so, although I have never seen it
> this way.

I think it counts as half of embedding. It lets you call Python from C, 
provided the C was called originally from Python. I don't see that it helps 
at all with the situation in my original posting which was calling Python 
from an otherwise purely C program.

I'm impressed by the example wrapping qsort. I can't think offhand of a way 
to achieve this easily in Pyrex. That is, Pyrex could easily wrap qsort to 
call a specific Python function for the comparison, but there isn't 
anywhere in the qsort callback to store state, and I can't see how you 
manage to convert a Python function into a C level function without some 
additional state.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From sjmachin at lexicon.net  Thu Apr 10 18:10:58 2003
From: sjmachin at lexicon.net (John Machin)
Date: 10 Apr 2003 15:10:58 -0700
Subject: class too big for one file! How can I split it?
References: <e956b5f1.0304100512.597654d3@posting.google.com>
Message-ID: <c76ff6fc.0304101410.b280d5c@posting.google.com>

meinmailfilter at gmx.de (HHaegele) wrote in message news:<e956b5f1.0304100512.597654d3 at posting.google.com>...
> OTLib.py :
> ---------------
> 
> XX = None
> 
> class Ot(a, e):
> 
>     def __init__(self, a = None, e = None):
>         global XX
>         XX = e    
>        
>     def Att(self):
>         global XX
>         print XX
>         ...
>         
> ---------------
> 
> Now, OTLib.py has more then 32000 lines. And that doesn?t work. Python
> 1.5 can?t execute then. So, how can I add the method Btt()? How can I
> make another file, which knows the global XX and who can be handled
> from the ot_starter, like it is in one file? Good would also be, if
> the Btt() can access XX the same way like Att() can. Is that possible?
> Or how must I change the structure?
> 
 
It does seem to be the number of lines that's a problem, not the size
of the information content. For example, a file huge.py created from:

   "# a biggie\n" + 40000 * "\n" + "def foo():\n   print 'bar'\n"

caused this:

Python 1.5.2 (#0, Apr 13 1999, 10:51:12) [MSC 32 bit (Intel)] on win32
Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam
>>> import huge
Traceback (innermost last):
  File "<stdin>", line 1, in ?
SystemError: com_addbyte: byte out of range

but changing the 40000 to 20000 gave the expected result.

So, why not delete some blank lines? Just enough to squeeze in the new
methods, not so many as to impair the legibility of your 32K line
source file :-)



From eppstein at ics.uci.edu  Thu Apr 17 15:42:45 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Thu, 17 Apr 2003 12:42:45 -0700
Subject: Python and Schools
References: <mailman.1050604354.15055.python-list@python.org>
Message-ID: <eppstein-B6AB90.12424517042003@news.service.uci.edu>

In article <mailman.1050604354.15055.python-list at python.org>,
 Tim Peters <tim.one at comcast.net> wrote:

> >     for i,x in enumerate(seq):
> >         if A[x] is an integer and 0 <= A[x] < i and A[A[x]] == x:
> >             return True
> >         A[x] = i
> >     return False
> 
> Something must be missing here, right?  For example, suppose A just happens
> to be uninitialized to all zeroes, and we pass in seq=[8, 8].  Then A[8] is
> set to 0 in the first loop trip, and in the second
> 
>     A[8] is 0 so 0 <= A[x] < i is true
>     A[A[x]] == A[A[8]] == A[0] == 0 != 8 so the if guard is false
>     A[8] is set to 1
> 
> Then the loop ends and the function returns False.

Sorry, bug in untested code.  Instead of A[A[x]]==x the test should be 
seq[A[x]]==x.

Anyway, this is not code you would want to actually use, just an 
illustration of how random access can sometimes make it useful to have 
much more memory than you actually touch during an algorithm.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From ianb at colorstudy.com  Tue Apr  8 13:15:01 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 08 Apr 2003 12:15:01 -0500
Subject: ANN: SQLObject 0.3
In-Reply-To: <Pine.LNX.4.44.0304081013030.16340-100000@bagend.odc.vzwcorp.com>
References: 	 <Pine.LNX.4.44.0304081013030.16340-100000@bagend.odc.vzwcorp.com>
Message-ID: <1049822101.734.40.camel@lothlorien>

On Tue, 2003-04-08 at 09:13, Scherer, Bill wrote:
> On 7 Apr 2003, Ian Bicking wrote:
> 
> > SQLObject 0.3 has been release:
> >   http://sqlobject.org
> 
> Is support for Oracle planned?

Sure, someone just has to add it (I don't have Oracle).  It should be
quite easy.

  Ian






From jjl at pobox.com  Fri Apr 25 10:58:35 2003
From: jjl at pobox.com (John J. Lee)
Date: 25 Apr 2003 15:58:35 +0100
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3ea7b12a@news.012.net.il> <7h3k7dkkwko.fsf@pc150.maths.bris.ac.uk> <Rpu197AqP9p+Ew5w@jessikat.demon.co.uk>
Message-ID: <87vfx28vz8.fsf@pobox.com>

Robin Becker <robin at jessikat.fsnet.co.uk> writes:
[...]
> I think that key board layout is also a pain, but I disagree that things
> can't be unlearned. I got used to having the esc key on the right hand
> side of the keyboard and then switched, it was painful, but eventually
> my viing got better again. Those reflex things are hard, but certainly
> seem to be conquerable. I'm having the same sort of experience moving
> back and forth to a Mac wrt ctrl-C vs splat-C etc 

The worst thing about Mac vs. PC keyboards is that Mac keyboards have
the home-key bumps on d and k, whereas PC keyboards have them on f and
j.  This really messes up my touch-typing.  I have no doubt I'd get
used to it eventually, though.

Sorry, utterly off-topic now...

The same goes for BASIC: I've thoroughly unlearned that, and Dijkstra
was wrong (if he was ever serious, which I guess he wasn't, really).


John



From kdahlhaus at yahoo.com  Tue Apr  1 14:43:14 2003
From: kdahlhaus at yahoo.com (Kevin Dahlhausen)
Date: 1 Apr 2003 11:43:14 -0800
Subject: The Whitespace Language
Message-ID: <283adf56.0304011143.4f4a6934@posting.google.com>

I'm getting so much entertrainment out of the the joke on the jokers
that created the Whitespace Language page -since we use indentation to
delimit block strucutre, white-space has been part of Python's syntax
since the beginning.  Ha!



From johnochiltree at blueyonder.co.uk  Sun Apr 27 12:41:08 2003
From: johnochiltree at blueyonder.co.uk (John Ochiltree)
Date: Sun, 27 Apr 2003 17:41:08 +0100
Subject: metaclasses, how you might actually use 'em
References: <mailman.1051458882.30400.python-list@python.org>
Message-ID: <PKTqa.1656$TR3.904@news-binary.blueyonder.co.uk>

Pedantic follow up:

Aren't you reinventing the wheel? Don't the concepts of reference counting
and the interpreter suggest that 'counting' on the fly has already been
invented. "Python thinks, therefore it is the metaclass"

John Ochiltree



From cappy2112 at yahoo.com  Thu Apr  3 00:31:47 2003
From: cappy2112 at yahoo.com (Tony C)
Date: 2 Apr 2003 21:31:47 -0800
Subject: Looking for ATA/IDE driver/wrapper for Python/Win32
Message-ID: <8d3e714e.0304022131.6439a188@posting.google.com>

Can anyone recommend a Python  ATA driver/wrapper for Win32 
I realize this will require a kernel mode driver for Win2000/NT, but
I'm sure someone must have done one by now.

I know of the SCSI one by Sam S.



From peter at engcorp.com  Fri Apr 25 09:05:51 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 09:05:51 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com> <b8b7kj$g74$1@news.hccnet.nl>
Message-ID: <3EA932AF.3BAD2FBF@engcorp.com>

Anton Vredegoor wrote:
> 
> The problem that comes up is that often, while playing, pieces of code
> are produced that could become important later. After all this is the
> source of almost all coding creativity isn't it?
> 
> So now we have a respectable programmer with all kinds of crazy code
> snippets, some of which could become of considerable value to himself,
> to the employer, and to other coders. However, there's no way to share
> them without losing credibility among the other "professional" coders.

You make an excellent point.  I believe that in some throwaway scripts,
perhaps most, there is somewhere a nugget or two of real value.  It
is *those* pieces for which I would write tests, even in a throwaway
script, perhaps because I instinctively believe the chance of that
*portion* of the script needing to be re-used is somewhat higher.

I wrote elsewhere that code without tests is a liability, while code
with good tests is an asset.  I think that goes directly to the heart
of this matter: the throwaway script *can* be made at least partially
useful to others, later, if it has good tests for the important parts.  
If it does not, you are quite right that you can't share it without
losing some credibility, with co-workers or other programmers.

Thanks for finding a reason to continue the discussion, and perhaps
a point where Alex and I will find our agreement. :-)

-Peter



From ajsiegel at optonline.net  Mon Apr 21 21:44:34 2003
From: ajsiegel at optonline.net (Arthur)
Date: Mon, 21 Apr 2003 21:44:34 -0400
Subject: "pythonic"
Message-ID: <000501c30870$ba44b450$0c02a8c0@Arts>

Cliff writes -

>How can you protect the implementation of something (in this case, the
>government) by destroying its foundations (the Constitution and Bill of
>Rights)?

>Confused'ly,

In the days leading up to WWII pro-Nazi elements in the US were active in
leading an "anti-war" movement. Once the war began such activities were
deemed Sedition and actively prosecuted as such. By the Roosevelt
administration, with FDR quite involved in policy-making on the issue.

Practicality beats purity.

Brecht, the Hollywood 12th, was probably in fact in Hollywood as a conscious
agent of Stalin. Though history chooses to gloss the fact.

Art





From fcslka at yahoo.com  Tue Apr 29 10:27:11 2003
From: fcslka at yahoo.com (Kent)
Date: 29 Apr 2003 07:27:11 -0700
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org>
Message-ID: <1bf4033f.0304290627.40f7d831@posting.google.com>

"Roman Yakovenko" <romany at actimize.com> wrote in message news:<mailman.1051603275.12037.python-list at python.org>...
> ). I decided to propose 
> python instead of VB. For this purpose I wrote small VB-Python compare.

Managers always fear new development tools because of the productivity
hit while we get up to speed on them.  Here's how I'm starting to
convince them that Python is a good tool (and therefore worth the
investment of time):

1.  Show that Python can be used to implement our requirements.  This
shows that Python is right for _that_ job.

2.  Show that Python can be used to implement anything that can be
implemented in the other environments.  This shows that it's not going
to be a long-term hinderance, and that it's a real language, not just
a toy.  The available GUI toolkits, and the immense standard library
(with all its Internet protocol support) really show this point.

By this time, you'll have shown that Python is a viable development
tool, on par with all the others.  Now that you've leveled the playing
field, show off some of its unique benefits:

3.  For many reasons, Python is vastly more productive than other
environments.  It's refreshing to use.  It's rich syntax and data
types turn 10 lines of code into 3.  It's standard library (mentioned
above, but now it really stands out) provides access to 95% of system
services you'll ever need in developing systems and applications.

4.  Python is cross-platform.  And, unlike that other popular
(snobbish) cross-platform language,  it really supports each platform
it's on.  The Win32 extensions make Python so much better a
development environment on Windows than Java.  I can actually use
those rich COM interfaces, and Windows system services.  Byte-compiled
files are still binary-compatible across platforms.

5.  Python's agility is one of its biggest values.  You can write huge
full-blown, complex applications, that integrate services from all
over the local machine, and from the network.  You can then write the
automated test scripts for your system (using the cool standard
library).

6.  Oh, yeah.  Python is free and open.

7.  In the interest of being honest, Python support in development
tools is weak.  There are many IDEs for Python, most of which are
immature, or are Python-only.  What needs to happen is for full
support of Python to be built into existing IDEs.  There's Visual
Studio .NET support, but boy, it's expensive.  NetBeans and Eclipse
could really help us out here if we could get more than just cursory
support for Python in those otherwise great environments.



From tebeka at cs.bgu.ac.il  Wed Apr  2 03:35:47 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 2 Apr 2003 00:35:47 -0800
Subject: [Q]How to append paths in a platform-independent way?
References: <420ced91.0304011946.56a22d2a@posting.google.com>
Message-ID: <33803989.0304020035.3d298d9b@posting.google.com>

Hello Daehyok,
> How to do the following jobs in a platform-independent way?
> 
> path = /home/sdhyok
> file = data.txt
> 
> 1. Append the file name to the path => /home/sdhyok/data.txt
> 2. Extract only path from /home/sdhyok/data.txt => /home/sdhyok
http://www.python.org/doc/current/lib/module-os.path.html

HTH.
Miki



From mwh at python.net  Tue Apr 15 05:47:03 2003
From: mwh at python.net (Michael Hudson)
Date: Tue, 15 Apr 2003 09:47:03 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10
  <7h3vfxhpl23.fsf@pc150.maths.bris.ac.uk> <yfsu1d1cisj.fsf@black132.ex.ac.uk> <7h3adettapc.fsf@pc150.maths.bris.ac.uk> <yfsptnpce02.fsf@black132.ex.ac.uk>
Message-ID: <7h3k7dwrt2h.fsf@pc150.maths.bris.ac.uk>

Alexander Schmolck <a.schmolck at gmx.net> writes:

> > Requiring declarations of all variables would be a HUGE change to
> > Python, IMHO.
> 
> Sure and I'm in no way arguing for such a change (and neither Paul
> Foley who, unlike me, considers implicit declaration a major
> flaw). I just don't understand why it would "unfeasible
> syntactically" -- *syntactically* it seems quite unproblematic to
> me.

Well, maybe I should have been more clear than just "syntactically" --
but you do understand what I mean, no?  I don't think you can
distingusish establishment and mutation of bindings in Python without
significant damage to the "feel" of the language.

Cheers,
M.

-- 
  The use of COBOL cripples the mind; its teaching should, therefore,
  be regarded as a criminal offence.
           -- Edsger W. Dijkstra, SIGPLAN Notices, Volume 17, Number 5



From tjreedy at udel.edu  Mon Apr 28 13:04:45 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Mon, 28 Apr 2003 13:04:45 -0400
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
References: <mailman.1051532982.25352.python-list@python.org>
Message-ID: <WhqdnS7YzrOVwjCjXTWcpQ@comcast.com>

"nguyen khanh" <phuoc-khanh.nguyen at bull.net> wrote in message
news:mailman.1051532982.25352.python-list at python.org...
>a=range(10)
>b=a[3:7]
>b=[3,4,5,6]
>Why the design exclude the last element ?

Think of slice coordinates as vertical bar cursor coordinates that
indicate positions before, between, and after the items of the
sequence.  n items gives n+1 cursor positions.  A slice indicates the
items between the cursor positions.

Terry







From aleax at aleax.it  Sun Apr 13 15:52:27 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 19:52:27 GMT
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com>
Message-ID: <%djma.11763$T34.381354@news2.tin.it>

<posted & mailed>

srijit at yahoo.com wrote:

> I do not see a  concept of Protected Methods in Python 2.2 like C++ and

Right: you don't see it because it isn't there.

> Ruby. Is this a serious limitation for Object Oriented Programing with

Judge for yourself: the introduction of "protected" is one of the few
aspects of C++ which Stroustrup *regrets*, according to his excellent book 
on the design and evolution of the C++ programming language.

> Python? Or may be it is possible to emulated Protected Methods in Python.

To quote St. Paul out of context, everything is permissible, but not
everything is opportune.  And that is very much Python's philosophy:
rather than focusing on trying to make some things impossible (and
generally failing -- a simple cast in most implementations of C++ lets
you blast away any "protected" that a silly library designer may have
tried to impose on you;-), Python empowers and trusts the programmer.

At the recent ACCU (Association of C and C++ Users) conference in Oxford,
Greg Colvin gave a wonderful keynote speech titled "in the spirit of C".

The FIRST items on the list are to trust programmers, and not stand in 
their way.  I found it interesting that, while Greg doesn't even KNOW
Python (at least that's what he said more than once at the conference),
he was making a substantial part of the case for Python so well:-).

If you don't trust programmers' skills and good intentions, you should
not give them powerful tools at all -- then, C, C++, Java, Ruby, Python,
would ALL be totally inappropriate, as each of them is easily powerful
enough to let an unskilled or evil programmer make a horrid mess of
things.  If you DO trust programmers, then "not standing in their way",
letting them get the job done, is very important.

In practice, what happens in languages such as C++ or Java is a very
interesting, almost schizophrenic split -- the programmer is "trusted"
when he's WRITING a library or framework, but NOT trusted when he's
USING that library or framework.  The library designer must be so
omniscient as to KNOW all uses to which his code will be put, and in
particular what pieces of code, that he found useful and put in his
library, will NOT be useful to other programmers in slightly different
contexts.  The library user is presumed to be such an idiot that a
simple INDICATION of "this code is intended for internal library use
only" would not suffice -- nay, the using programmer must be "coerced"
to keep his grubby hands away from said code.  In practice, there is
no such split in the population of programmers -- authors and users
of libraries are drawn from the same pool.  The sociology implicit in
the language design just isn't a good match for reality.

In Python, the amount of trust vested in you, programmer, is just
about the same whether you're using an existing library or coding
up a new one (most of the time, you're doing both -- using some
existing libraries AND also producing code that you or somebody
else will want to reuse in the future in slightly different ways).

You indicate "this is meant for internal use only" by following the
convention of starting a name with a single underscore; you respect
that convention in libraries you use, unless you've got deucedly
good reasons not to (and can live with the consequences, such as
your code not working any more with a newer release of the library).

The latter possibility is basically the same thing as putting
#define private public
#define protected public
at the top of your .cpp file, before any #include, in most C++
implementations -- except that instead of this one lump reminder
that you're breaking conventions, you get reminded of the fact
throughout your code, whenever you refer to a name starting with
an underscore.

As for the specific indication "this name may be used by
derived classes but not by other code", there's never really any
particularly good reason for that (whence Stroustrup's regrets) --
public and private have reasons for existing (public is about
a library's interface to the world, private about its current
implementation), protected doesn't (no real need for a library
to distinguish SEVERAL "interfaces to the world" depending on
where inheritance is being used -- inheritance is just a tool
like many others and it's wrong to put it upon some altar...).


Alex




From mwh at python.net  Sun Apr  6 13:03:16 2003
From: mwh at python.net (Michael Hudson)
Date: Sun, 6 Apr 2003 17:03:16 GMT
Subject: negative values in profile output
References: <mailman.1049479748.24160.python-list@python.org>
Message-ID: <7h38yunk17w.fsf@pc150.maths.bris.ac.uk>

"Matthew M. Boedicker" <mboedick at mboedick.org> writes:

> When using the profile module to profile a long-running program,
> I am getting some negative values:
> 
> 6753628 -4036.027   -0.001 -4036.027   -0.001 gobject_sql.py:36(<lambda>)
> 
> Am I correct to assume these don't have special meaning and are the
> result of something wrapping?

They may be a result of mis-calibration of the profiler -- see the
docs.

Last time i used the profiler, I more-or-less ignored the values
output and just assumed I was doing the right thing when they got
smaller...

Cheers,
M.

-- 
  Haha! You had a *really* weak argument! <wink>
                                      -- Moshe Zadka, comp.lang.python



From max at alcyone.com  Thu Apr 17 19:05:50 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 17 Apr 2003 16:05:50 -0700
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
	 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
	 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
	 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
	 <mailman.1050438927.20366.python-list@python.org>
	 <yu99y92bh10b.fsf@europa.research.att.com>
	 <mailman.1050457621.2105.python-list@python.org>
	 <slrnb9q90a.mhg.news@datatailors.com> <3E9DDC81.932475AF@alcyone.com> <3ckgewva.fsf@morpheus.demon.co.uk>
Message-ID: <3E9F334E.2B03D5@alcyone.com>

Paul Moore wrote:

> Just to clarify this point a bit, the measure usually used in sorting
> is the number of swaps of two elements (or the number of comparisons -
> these are usually equivalent in terms of O() calculations).

That touches on a point I meant to address but forgot, which is that the
big-O notation can really refer to anything at all.  Usually it refers
to algorithm complexity (how much "stuff" needs to get done -- as you
rightly point out in the case of a sorting algorithm it would be the
number of compares), but it could apply to anything like numeric
multiplications, or even amount of memory used either by a static data
structure or an algorithm.

It's really just a method for expressing how complex "something" gets
for very large n (mathematically, as n -> oo).  What that something is
can vary from usage to usage.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Covenants without the sword are but words.
\__/ Camden
    Bosskey.net: Aliens vs. Predator 2 / http://www.bosskey.net/avp2/
 A personal guide to Aliens vs. Predator 2.



From writeson at earthlink.net  Fri Apr 25 10:51:47 2003
From: writeson at earthlink.net (Doug Farrell)
Date: 25 Apr 2003 07:51:47 -0700
Subject: ftp problem on solaris machines with python
References: <88bc63c6.0304240407.674220cb@posting.google.com> <b88rae$etu$1@heckle.cc.uregina.ca>
Message-ID: <88bc63c6.0304250651.5ec01610@posting.google.com>

Hi,

We're not gzipping the files first because we're really short on space
on the Solaris machine and gzipping takes quite a bit of temp space to
work, unfortunately. By croaking I mean the the python program croaks
in mid transmission while ftping a large file. However, ftping the
file from the shell command line works fine.

One of our guys found a module pyExpect (I think) that will let our
program ftp using the command line rather than the ftplib module.
We're looking into that.

Thanks for your response,
Doug



From alan_salmoni at yahoo.com  Tue Apr  8 07:23:24 2003
From: alan_salmoni at yahoo.com (Alan James Salmoni)
Date: 8 Apr 2003 04:23:24 -0700
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com> <urnka.4658$Du.16154@newsc.telia.net>
Message-ID: <b68f71a7.0304080323.75b5ed1a@posting.google.com>

As Tim said, wxPython has an excellent wxGrid widget capable of doing
what you want (see http://salstat.sunsite.dk/screenshots.html for some
examples), but I also understand that there is a similar kind of thing
which can be used with the Tkinter toolkit known as the tkTable. I
haven't used it myself and therefore cannot comment on it, but there
is a spreadsheet written in it called Novagrid which you might like to
try:

http://sourceforge.net/projects/novagrid

Hope this helps!

Alan James Salmoni
SalStat Statistics
http://salstat.sunsite.dk

"Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote in message news:<urnka.4658$Du.16154 at newsc.telia.net>...
> "Rainer Klueting" <rainer.klueting at epost.de> wrote
> > can anybody point me to a source where I can find a datagrid-widget to
> > use with Python? It should basically show a flat table in a
> > spreadsheet-like way. I tried for a while to do it by myself using
> > Tkinter, but I'm lacking experience (just trying to change that...)
> 
> wxPython is a guitoolkit with a good datagrid widget.
> First you must install wPython, http://www.wxpython.org/ but once that is
> done you could write a small app like this if I am not mistaken
> 
> 
> from wxPython.wx import *
> from wxPython.grid import *
> 
> class MyFrame(wxFrame):
>     def __init__(self, parent, title):
>         wxFrame.__init__(self, parent, -1, title, size=(350, 200))
>         grid=wxGrid(self, -1, size=(150,150))
>         grid.CreateGrid(4,4)
> 
> app = wxPySimpleApp()
> frame = MyFrame(None, "Simple wxPython Grid")
> frame.Show(True)
> app.MainLoop()
> 
> Then you an have a look at the wxGrid widget wich is qiute extensive.
> 
> Good luck
> 
> Tim



From pc451 at yahoo.com  Thu Apr 17 09:55:48 2003
From: pc451 at yahoo.com (Peter Clark)
Date: 17 Apr 2003 06:55:48 -0700
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com> <3e9daad2$0$49115$e4fe514c@news.xs4all.nl> <f7199550.0304161732.48c75c42@posting.google.com> <mailman.1050548247.26781.python-list@python.org>
Message-ID: <f7199550.0304170555.77773f8a@posting.google.com>

Steven Taschuk <staschuk at telusplanet.net> wrote in message news:<mailman.1050548247.26781.python-list at python.org>...
> Quoth Peter Clark:
> > [...] However, when I put it into the code, I get this error:
> > 
> >     w += [deg.encode('UTF-8') + scale.strip()]
> > UnicodeError: ASCII decoding error: ordinal not in range(128)
> 
> Aha!  Yes, it's a *decoding* error.  Your 'deg' variable is a
> normal string, right?
    That's correct.

> >     Since the output is meant to be read to be displayed by a font
> > which is in essentially latin-1 encoding, I need to restrict the
> > manner in which the degree symbol is displayed to one byte. [...]
> 
> So just do something like
>     print >>fileobject, chr(176)
>     print >>fileobject, u'\N{DEGREE SIGN}'.encode('latin-1')

    It worked fine when printing to a file (i.e., from the prompt),
but I still got an error when I try to append it to a string destined
to be included in a list:

    w += [u'\N{DEGREE SIGN}'.encode('latin-1') + scale.strip()]
UnicodeError: ASCII decoding error: ordinal not in range(128)

    In this case, 'scale' is either 'C' or 'F' (hence the need for a
degree character). But then I realized that what you said above would
also apply to 'scale' as well. So when I changed the line to:

w += [u'\N{DEGREE SIGN}'.encode('latin-1') +
scale.strip().encode('latin-1')]

it worked perfectly. Problem solved, thank you very much! So in
retrospect, my mistake was in not paying attention to the encoding of
the rest of the string, which was why the error messages kept popping
up. Duely noted.



From skip at pobox.com  Sat Apr 12 22:34:31 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sat, 12 Apr 2003 21:34:31 -0500
Subject: an ugly file-reading pattern
In-Reply-To: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
Message-ID: <16024.52407.405254.424039@montanaro.dyndns.org>

    Istvan> if I choose to process the file line by line the "Learning
    Istvan> Python" book advises me to use the following atrocity:

    Istvan> while 1:
    Istvan>     line = file.readline()
    Istvan>     if not line: break

    Istvan> Maybe I'm picky here but having to start an infinite loop then
    Istvan> needing a conditional to break out of it is anything but an
    Istvan> elegant pattern.

    Istvan> This has left me wondering about python,

Recent versions of Python (2.2 and later) allow this:

    file = file("somefile")
    for line in file:
        process(line)

It does the right thing from an efficiency standpoint, neither slurping the
file in all at once nor reading it line-by-line or char-by-char.

Note also that in 2.2 and later there is a file object in builtins ("open"
and "file" are the same object).  "open" is discouraged.  "file" is
preferred.

Skip





From mis6+ at pitt.edu  Wed Apr 30 16:23:44 2003
From: mis6+ at pitt.edu (Michele Simionato)
Date: Wed, 30 Apr 2003 16:23:44 -0400
Subject: Question about accessing class-attributes.
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE192@admin56.narex.com>
Message-ID: <3EB030D0.300D84FD@pitt.edu>

Bjorn Pettersen wrote:
> > Michele Simionato wrote:
> > I must say, however, that len(S) is not *literally*
> > type(S).__len__(S):
> 
> Yes, that was one of my problems. Where did you find this rule? I can't
> find a real reference to it anywhere. For classic classes lookup is
> o.__class__.__dict__['attr'].__get__(o.__class__, o) for regular
> methods, which sort of looks like the above, but...

Never found any official reference except postings from Alex Martelli,
and
maybe I have read that in the Nutshell, too (Alex, correct me if I am
wrong!)
I do agree with you that the rule does not work in the case you pointed
out;
not sure if it is a bug or intended behaviour.

> [...]
> > I did not understood your point in the other thread, I didn't imagine
> > you where thinking about meta-metaclasses. As you see, there are
> > situations where they may be useful, i.e. in the customization of
> > metaclasses, but most people don't think at this level yet ;)
> 
> I wasn't (or at least wasn't trying to :-). They're facinating beasts
> though -- a semantic system with recursion in its axioms ;-)
> 
> -- bjorn

With meta-metaclasses you may implement type(S).__len__(S) but 
I see now that your point is more with len(S) that does not 
recognizes '__len__' methods provided indirectly via __getattr__ .
You can always do that

class Super(super):
    def __getattr__(self,somename):
        return somename
      
or even overrides __getattribute__ and implement an entirely
different version of super. Anyway, this does not solve your
problem. 

It seems to me more a problem of len that of super,
in the sense that len(x) is not *literally* type(x).__len__(x)
(idem for __getitem__ in the other thread; I would expect the
same for str and repr, I should check ...)

-- 
Michele Simionato - Dept. of Physics and Astronomy
210 Allen Hall Pittsburgh PA 15260 U.S.A.
Phone: 001-412-624-9041 Fax: 001-412-624-9163
Home-page: http://www.phyast.pitt.edu/~micheles/




From erwin at andreasen.com  Mon Apr 28 14:48:36 2003
From: erwin at andreasen.com (Erwin S. Andreasen)
Date: 28 Apr 2003 20:48:36 +0200
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com>
Message-ID: <878ytuo3uj.fsf@andreasen.org>

achrist at easystreet.com writes:

> Any suggestions of the shortest path from a Python program to RTF
> output?  RTF seems to be the best (reasonable combination of 
> features and compatability) common denominator for exporting
> data so that (windows) users can bring it into their own word
> processing program, whatever that may be.  (Any problem with
> that?)

I have written some code for exporting RTF reports from a web-based
survey/reporting system. It runs on UNIX, so I could not use any COM
objects or similar; the RTF code was generated by hand. Fortunately
the system was designed templatable, so exporting RTF was mostly just
a matter of selecting a set of different style descriptions that
contain RTF instead of HTML.

Others have already linked to the RTF specs. I wrote some basic output
based on that, then we made it nicer using Word. Word's RTF export
exports a lot of unneeded information, so I just cut and pasted the
interesting parts until my output looked right.

Unfortunately Words tends to crash very easily when given an invalid
RTF file. For verifying correctness I used the "ted" editor which is
an ancient RTF-capable graphical UNIX editor. Especially getting
tables displayed right was difficult, the RTF syntax for specifying
them is annoying. ted dumps some debug info when it sees tables it
cannot handle, so that was quite useful.

The RTFDoc.py file from the ptommi project looks like it holds very
useful information; I wish I had found it a few months ago :)

Next step is generating Power Point documents (I found a spec on
wotsit.org and I guess I'll have to wrap some OLE library (cole or
gnome's) and see if I can generate something useful).


-- 
===============================================================
<erwin at andreasen.org>                           Herlev, Denmark     
<URL:http://www.andreasen.org/>                             <*>   
===============================================================




From martin at v.loewis.de  Fri Apr 18 04:38:54 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 18 Apr 2003 10:38:54 +0200
Subject: [Que]Do any SGML lib good for Python?
References: <2778697.1050595278@dbforums.com> <m3of35ni8d.fsf@mira.informatik.hu-berlin.de> <2781052.1050630152@dbforums.com>
Message-ID: <m3ptnk9p41.fsf@mira.informatik.hu-berlin.de>

dolephi9080 <member27604 at dbforums.com> writes:

> What is your mean parsing the output of nsgmls?

Invoke nsgmls, and see for yourself.

Regards,
Martin



From maxm at mxm.dk  Wed Apr 23 02:39:08 2003
From: maxm at mxm.dk (Max M)
Date: Wed, 23 Apr 2003 08:39:08 +0200
Subject: pattern match
In-Reply-To: <220420031949394828%greg.bland@valpo.edu>
References: <220420031949394828%greg.bland@valpo.edu>
Message-ID: <Pwqpa.29249$y3.2305712@news010.worldonline.dk>

Mitch MF Cummstain wrote:
>                I just started using python and I need help with a
> pattern match.
>       
>    m = re.match("(.{30})(word)(.{30})",paragraph);
> 
> I need to match 30 characters on each side of word.  The problem word
> is getting treated as a string not a variable like I need it to.  How
> can I make it treat word as a variable ?

Are you sure this is how you want to do it? Perhaps a simple slice can 
do the trick. It's hard to know without an example string.

word = paragraph[31:word.rfind(')')]

Also you say you want word to be a variable. But in which way, and what 
value should it hold?

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From joe at strout.net  Tue Apr  8 11:39:54 2003
From: joe at strout.net (Joe Strout)
Date: Tue, 08 Apr 2003 15:39:54 GMT
Subject: Python searchable mail archive utility?
Message-ID: <joe-D88C59.08403208042003@orngca-news02.socal.rr.com>

I'm looking for a web-based mailing list archive/search utility, 
preferably written in Python.  It'll be running under Mac OS X 
(command-line).  It needs to be able to handle several very high-volume 
lists (the largest gets up to a thousand messages a day).

I'm sure I saw something like this in the past, but now I can't find it.  
The only tools I've found are MHonArc, which is Perl (ick) and requires 
Namazu (which we're having trouble building) to search multiple lists; 
and Hypermail, but hypermail.org seems to be down and it looks like 
that's written in C anyway.

We looked briefly at Mailman, but it seems to be mailing list software; 
the docs we could find say nothing about its archiving capabilities 
(except to hint that it has them).

We're getting so desperate, we're about to install a Perl tool -- please 
help us avoid that fate!

Thanks,
- Joe

,------------------------------------------------------------------.
|    Joseph J. Strout         Check out the Mac Web Directory:     |
|    joe at strout.net           http://www.macwebdir.com             |
`------------------------------------------------------------------'



From yorklee70 at yahoo.com  Tue Apr 29 13:19:47 2003
From: yorklee70 at yahoo.com (York)
Date: Tue, 29 Apr 2003 18:19:47 +0100
Subject: The behavior of List.insert
References: <b8m8dn$b03$1@pegasus.csx.cam.ac.uk> <TZOdnU3XYIdTMzOjXTWcpQ@comcast.com>
Message-ID: <b8mc4b$e80$1@pegasus.csx.cam.ac.uk>

>> mylist = [1,2,3,4,5],
>> mylist.insert(-1,99) will make mylist to be [99,1,2,3,4,5]
>>
>> just wonder why not let it to be [1,2,3,4,99,5]
> 
> Guido decided otherwise for whatever reason.  What you observed is
> intended: from LibRef 2.2.6.4 Mutable Sequence Types
> "(4)
> When a negative index is passed as the first parameter to the insert()
> method, the new element is prepended to the sequence. "
> so this is not an implementation bug.

Thanks. I don't know what Guido's reason is, but I still believe that it is 
nice to allow list.insert distinguish different negative indice rather than 
treat all them as zero.

Cheers,

York



From hwlgw at hotmail.com  Thu Apr 24 05:18:03 2003
From: hwlgw at hotmail.com (Will Stuyvesant)
Date: 24 Apr 2003 02:18:03 -0700
Subject: Why functional Python matters
References: <mailman.1051154381.32098.python-list@python.org>
Message-ID: <cb035744.0304240118.56baa849@posting.google.com>

[Tim Peters]
> ...
>     class make_adder:
>         def __init__(self, n):
>             self.n = n
> 
>         def __call__(self, i):
>             return self.n+i
> 
> It often proves much better to carry state around in explicit objects
> instead of in hard to introspect, or to modify (unless you planned for that
> from the start), anonymous closures.  
> ...
> when-you-don't-have-a-hammer-only-nails-look-like-nails-ly y'rs  - tim

Another great posting Tim!  Who much time do you take to write
something like that?
Indeed, using a class it would be possible to *extend* a hammer, for
example such that it counts how much hits it did.  With *pure*
closures this would be a pain.

closure-classes-for-hammers-that-count-ly y'rs



From piotrus at sympatico.ca  Sun Apr 20 01:30:20 2003
From: piotrus at sympatico.ca (Peter)
Date: Sun, 20 Apr 2003 01:30:20 -0400
Subject: red hat 8.0 and python
References: <3EA21326.7070100@sympatico.ca> <mailman.1050811603.11948.python-list@python.org>
Message-ID: <3EA2306C.9010808@sympatico.ca>

Dave Reed wrote:
> On Saturday 19 April 2003 23:25, Peter wrote:
> 
>>I've just installed Red hat 8.0, and although python 2.2.1 comes with 
>>it, i cannot use anything which uses the Tkinter library (this 
> 
> includes 
> 
>>IDLE). I a few other packages work, and the interpreter runs fine. I 
>>suppose that this might be due to the setup, but I cannot figure it 
> 
> out. 
> 
>>  Could it be that Tkinter module does not come with the python 
>>distribution which came with Red Hat 8.0? I read that it should, but I 
>>searched my whole computer for anything with ?tk? in the name with no 
> 
> luck.
> 
> 
> It's on one of the CDs - install the tkinter-2.2.1-17 rpm file.
> 
> 
> 
>>I downloaded the rpm's for the python 2.2.2 (from www.python.org) but 
>>simple "rpm -i" or "rpm -U" installing gives me errors; conflicts with 
>>the currently installed version. I am not quite sure whether I should 
>>try uninstalling what is there now because I've read that the a few 
>>programs need it to run.
>>
>>Any suggestions? Also if anyone knows of a webpage listing the 
>>step-by-step instructions on how to install/uninstall/upgrade the 
> 
> python 
> 
>>distribution, please reply. I have read the info provided on 
>>www.python.org, but it is not detailed enough for me.
> 
> 
> If you just want Tkinter, install the above mentioned RPM rather than
> going to the trouble of upgrading.
> 
> Dave
> 
> 

Thanks Dave:

I found the file (it's on the third CD) but when I try to install it 
(using rpm -i ) i get:

error:  Failed dependencies:
      libtix.so.0 is needed by tkinter-2.2.1-17

I am not sure how to even begin looking where that library might be...

Any help would be appreciated..

Peter




From bnelson at appservers.us  Thu Apr 10 19:20:41 2003
From: bnelson at appservers.us (Brian Nelson)
Date: Thu, 10 Apr 2003 16:20:41 -0700
Subject: popen4 hangs when calling rsync
Message-ID: <3E95FC49.2080807@appservers.us>

I'm having problems when using popen from within threads. The problem 
only appears when I'm calling rsync. It seems the process never exits. 
It runs fine when run from the main thread of the application.

The code is as follows....


     def syncPortage(self):
         SYNC_CMD = "/usr/bin/rsync -rlptDvz --no-detach --progress 
--stats --delete --delete-after --delete-excluded --timeout=180 
--exclude='distfiles/*' --exclude='packages/*' 
rsync://mirror.appservers.us/appserver /usr/portage "
         self._runCmdAndLog(SYNC_CMD)


     def _runCmdAndLogPopen4(self, cmd):
         logger.info("runCmdAndLog")
         re, w = popen2.popen4(cmd)
         while 1:
             logger.info("read next line")
             char = re.readline()
             if not char: break
             logger.info(char)
         logger.info("after command executed")
         re.close()
         w.close()
         logger.info("end runCmdAndLog")



Is this a general problem with using popen from within threads? Or is it 
something to do with rsync.

Thanks for your help!








From tjreedy at udel.edu  Thu Apr 10 12:35:34 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Thu, 10 Apr 2003 12:35:34 -0400
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com> <b73vov$d92$00$1@news.t-online.com>
Message-ID: <HzGdnQNo1Z2-AAijXTWcog@comcast.com>

"Stephan Diehl"
> You are probably looking for "linear regression".

Linear regression is designed for continuous, not binary outcome
variables.

"Duncan Smith"
> But linear discriminant analysis and logistic regression would give
you

Logistic regression use 0-1 outcomes to fit a very particular model: p
= exp(s)/(1+exp(s)) where p is a probability of 1 versus 0 and s is a
linear sum of 'independent' variables.  It is typically used for
modelling probability of response within one group.  It is less likely
to be useful for two-group classification.

Terry





From bbeck at NOSPAM.austin.rr.com  Thu Apr 17 02:29:29 2003
From: bbeck at NOSPAM.austin.rr.com (Brandon Beck)
Date: Thu, 17 Apr 2003 06:29:29 GMT
Subject: Memory leak while looping
References: <1020cbf8.0304162100.6fc49e6f@posting.google.com>
Message-ID: <n8t1n-ms4.ln1@isomorphism.org>

It appears your "leak" is with the call to range().  range(1, 10000000)
causes a list of 10000000 to be created in memory.  This has nothing to
do with item being assigned in the body of the loop, and you should see
the same behavior in a script without the for loop and just the call to
range.  Consider using the generator (or maybe it's just generator like)
version of range, xrange.  That should run in constant memory.

Brandon


"Roger Hancock" <roger_hancock at hotmail.com> wrote in message
news:1020cbf8.0304162100.6fc49e6f at posting.google.com...
> I'm just starting to play with writing C extensions for Python.  In
> the process of doing so, I called one of my extension functions
> thousands of times only to discover that a leak was eating up all the
> memory in my system.  Upon further investigation, I found I could
> produce the same result with the following python code with my
> extension completely out of the picture:
>
> #!/usr/bin/env python2
>
> for i in range(1,10000000):
>     item = i
>
>
> I would have thought that the memory for "item" would be freed as new
> "item"s were created.  However, this doesn't seem to be the case.  I
> would appreciate any clarification on this issue.
>
>
> Thanks,
>
> Roger





From hwlgw at hotmail.com  Sun Apr 27 15:13:14 2003
From: hwlgw at hotmail.com (Will Stuyvesant)
Date: 27 Apr 2003 12:13:14 -0700
Subject: OO + Python Developers Wanted
References: <mailman.1051183297.29110.python-list@python.org>
Message-ID: <cb035744.0304271113.c0880dd@posting.google.com>

[Neeta Ramchandani]
> I am looking for an OO Scriptor with Python Scripting experience.
> You must have at least 2 years Java experience with Unix and Proper Pyhton 
> development expertise.
> This is a role for an Investment Bank in the City, for duration of 3-6 mont
> hs.
> Rates are OPEN!!

This looks like shit.

Amateur!



s-h-a-d-d-u-p-ly-'yrs



From aleax at aleax.it  Sat Apr 12 15:04:04 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 19:04:04 GMT
Subject: Can I define nested classes in Python?
References: <169a7283.0304101432.13f5786c@posting.google.com> <1JGdnc952fcaaAijXTWcrg@comcast.com> <6dub9v0rjv0s7r87hhjsjqb0oqjuu9m8ko@4ax.com>
Message-ID: <EqZla.5404$LB6.151197@news1.tin.it>

Rene Pijlman wrote:

> Terry Reedy:
>>sdieselil:
>>> Can I define nested classes in Python
>>
>>Try it and see what happens.
> 
> This may have been how Thomas Edison constructed the light bulb,
> but this is not the Right Way to learn a programming language.

Different people learn best in different ways.  There's no One
Right Way.  Reading manuals that are identified as being "for
Language Lawyers", such as the one you mention:

> A better approach is to read the manual.
> http://www.python.org/doc/current/ref/nested-scopes.html

may perhaps be a Right Way for you, or me, but I think that "try
and see" will probably work for a vaster proportion of users.

Though most users will probably be best served by another approach
yet, one between those two extremes -- some experimentation, some
reasoning and explanation, then back to some more experimentation...


Alex




From fuka at fuxoft.cz  Thu Apr 24 05:26:02 2003
From: fuka at fuxoft.cz (Frantisek Fuka)
Date: Thu, 24 Apr 2003 11:26:02 +0200
Subject: Testing if string contains a substring
In-Reply-To: <vYDpa.50997$LB6.1152215@news1.tin.it>
References: <b865dq$28gi$1@ns.felk.cvut.cz> <NKxpa.85054$T34.2112670@news2.tin.it> <698f09f8.0304231141.5d458eb0@posting.google.com> <vYDpa.50997$LB6.1152215@news1.tin.it>
Message-ID: <b88akc$31dm$1@ns.felk.cvut.cz>

Many thanks to everyone. Can I get the "new" string functions 
explanation somewhere? I have Python 2.2.2 but the current documentation 
only seems to mention the old-style STRING module...

Alex Martelli wrote:
> Jeremy Fincher wrote:
> 
> 
>>Alex Martelli <aleax at aleax.it> wrote in message
>>news:<NKxpa.85054$T34.2112670 at news2.tin.it>...
>>
>>>However, such new functionality doesn't get back-ported to previous
>>>releases of Python, such as 2.2.*.  In all Python releases from 1,6
>>>included to 2.3 excluded, "needle.find(haystack) > 0" is the idiom.
>>>Module string is needed if you want to use antediluvian versions of
>>>Python, such as 1.5.2, where strings didn't have methods.
>>
>>Of course, you  meant "haystack.find(needle) >= 0", right?
> 
> 
> Yes, sorry.  See my response to Bengt as to why I was confused
> between needle-on-left and haystack-on-left;-).
> 
> 
> 
>>I always use (and have seen used) "haystack.find(needle) != -1", myself.
> 
> 
> I never saw any added value in that one extra character -- if I did
> want to use -1 I would also use > rather than != ...;-)
> 
> 
> Alex
> 




From andrew-pythonlist at puzzling.org  Wed Apr 23 04:05:19 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 23 Apr 2003 18:05:19 +1000
Subject: why not say just what wants an integer?
In-Reply-To: <20030423075316.GA13945@upl.cs.wisc.edu>
References: <87d6jeeyo3.fsf@jidanni.org> <20030423075316.GA13945@upl.cs.wisc.edu>
Message-ID: <20030423080519.GG10754@frobozz.local>

On Wed, Apr 23, 2003 at 02:53:16AM -0500, Adam Hupp wrote:
> > 
> > P.S. "pydoc print" gives nothing. Should it give some clue to the user
> > wanting to know about "print", or is that some violation of principles?
> 
> Sort of a violation.  Print is a keyword just like 'if' and 'def'.  so
> pydoc would have to special-case all of those.  Not a bad idea but I
> doubt it comes up very often.  

Try help('print') at the interactive prompt.

-Andrew.





From logistix at zworg.com  Mon Apr  7 09:28:48 2003
From: logistix at zworg.com (logistix)
Date: 7 Apr 2003 06:28:48 -0700
Subject: COM and MS Word
References: <2eb85cfb.0304070115.6aecb644@posting.google.com>
Message-ID: <bdb5b58f.0304070528.4541df3b@posting.google.com>

anthony.evershed at npl.co.uk (CTony) wrote in message news:<2eb85cfb.0304070115.6aecb644 at posting.google.com>...
> It appears that some data files I need are in WordPerfect (.dat)
> format, and I need to get them into Python for processing and later
> entry into a database. I haven't been able to find a way of
> interfacing to WordPerfect directly, but MS Word 2000 can read these
> files...
> 
> Here's the snag, though. All the examples of Python/COM/Word usage
> I've seen have involved pushing data into Word, rather than extracting
> data *from* it. I know that the data field I need is x lines below a
> known heading, so getting to the field is no problem:
> 
> finder = wd.Selection.Find
> finder.Text = 'foo'
> finder.Execute()
> for n in range(x):
>     wd.Selection.MoveDown
> wd.Selection.HomeKey
> 
> get you to the beginning of the field, no problems. How, however, can
> I select (block mark) and extract that line of text?
> 
> If it helps, I'm using Python 2.2, with win32all build 148, on Windows
> NT 4 (though the resulting program will be used on Windows 2000 as
> well)...
> 
> Many thanks,
> 
> Tony Evershed

Set Selection.Start and Selection.End manually.  Sorry I can't tell
you what to set them to (I think that'll be document specific)



From reinhard at proceryon.at  Wed Apr  9 05:58:05 2003
From: reinhard at proceryon.at (reinhard nadrchal)
Date: Wed, 09 Apr 2003 11:58:05 +0200
Subject: killing child processes
References: <39ee6fa9.0304090148.745fdc46@posting.google.com> <3E93ECC2.9080207@Linux.ie>
Message-ID: <3E93EEAD.6030908@proceryon.at>

Sorry, but fork() is not available under win32 (which i'm working on).

reinhard

Padraig at Linux.ie wrote:
> reinhard wrote:
> 
>> I've created a process (some batchfile) which in turn creates some
>> long running child processes using the win32api CreateProcess call
>> (declaring them as a group).
>>
>> Is there a way to terminate (kill) all the child processes in the
>> group?
> 
> 
> yep. just fork() and setpgrp()
> Then all processes will have the
> same groupid which you can kill(-groupid)
> You can look at the code in FSlint:
> http://www.pixelbeat.org/fslint/FSlint-2.01.tar.gz
> 
> Padraig.
> 




From Kyler at news.Lairds.org  Fri Apr 18 11:22:09 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Fri, 18 Apr 2003 15:22:09 GMT
Subject: Zope authentication from python scripts
References: <HJNna.3123$AB5.402420@news02.tsnz.net>
Message-ID: <93h5n-vpi.ln1@news.lairds.org>

Mike McGavin <jester at NOSPAM.mcsnospam.vuw.acNOSPAM.nz> writes:

>The specific problem I'm having is that I can't figure out how to get 
>the code that handles the submitted web form to authenticate and log the 
>user in so that Zope recognises an Authenticated User in the session.

It's left to a user folder of some ilk.  Take a look at
AccessControl/User.py.  To use your own authentication
scheme, you probably want something like Extensible User
Folder.
	http://exuserfolder.sourceforge.net/

I've only used standard HTTP authentication, but I think
these typically have hooks for cookie-based (ick) 
authentication.

--kyler



From dave at boost-consulting.com  Tue Apr 29 21:55:12 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Tue, 29 Apr 2003 21:55:12 -0400
Subject: POD vs. reST for standalone writing?
In-Reply-To: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> (eichin@metacarta.com's
 message of "28 Apr 2003 22:21:59 -0400")
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
Message-ID: <u3ck0696n.fsf@boost-consulting.com>

> Now I start seeing mention of reStructuredText (especially in comments
> about pyCon.)  reST files look very "ascii art" on the input side, but
> after a little experiment, that seems to make them *harder* to write,
> if somewhat (and I think only a little) easier to read. 
>
> Am I missing something?  Is there actually a reST clear short hand,
> and a tool that expands it to full-blown ascii-art?  Or is at least
> fuzzy in the counting of dashes and other punctuation for marking
> different levels?

ReST rocks.  I'm using it to write a book, and think that it comes
out very readably and is miles easier to write than any tweaky **ML
or TeX.

An example of something recent (not from the book):
http://www.boost-consulting.com/writing/new-iter-concepts.html

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From ckea25d02 at sneakemail.com  Mon Apr 14 01:57:46 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 14 Apr 2003 17:57:46 +1200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <mailman.1050253975.31483.python-list@python.org>
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org>
Message-ID: <b7di0n$dd6gj$1@ID-169208.news.dfncis.de>

Beni Cherniavsky wrote:
> Well, it is true that things will get confusing with strings.  But
> when you do an str() of a list, how much do you care for it?  Perhaps
> it's worth to apply str() in list.__str__, keeping list.__repr__ for
> the cases when you need an unambiguos representation?  I.e. leve it up
> to the user to decide whether he will be confused.

Python *does* leave it up to the user! Its containers
don't have a default __str__ at all, because there's
no way of knowing what the programmer wants it to
do.

If you care about the precise format of the string
representation you're trying to get for a container,
you shouldn't be using either str() or repr() --
you should be writing your own function that does
exactly what you want.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From revanna at mn.rr.com  Tue Apr  1 10:37:44 2003
From: revanna at mn.rr.com (Anna)
Date: Tue, 01 Apr 2003 15:37:44 GMT
Subject: list index question
References: <e62b26d3.0304010613.39f98991@posting.google.com>
Message-ID: <pan.2003.04.01.15.38.54.181811@mn.rr.com>

On Tue, 01 Apr 2003 07:06:28 +0000, Matt Singer wrote:

> I'm sure this is probably a simple question but... I'm a python rookie.
> 
> I'm reading an index table from a database of two columns using a
> fetchall() call.
> 
> If I print this, I get something like
> 
> (100, 1)     #results[0]
> (100, 10)    #results[1]
> (101, 1)
> (101, 5)
> 
> I want to find the index for (100, 10) without looping.
> 
> Any suggestions gracefully accepted.
> 
> Thanks

You want index. 

s.index(x)
return smallest i such that s[i] == x

You can find it here:
http://www.python.org/doc/current/lib/typesseq-mutable.html

>>> list1=[(100,0), (10,5), (100,10), (100,0)]
>>> print list1
[(100, 0), (10, 5), (100, 10), (100, 0)]
>>> list1.index((100,10))
2
>>> 

Warning - look at this example on the same list:
>>> list1.index((100,0))
0
>>> 

Note that index will not return the offset for *every* instance of x in
your list, just the smallest offset (the first one it comes to). Since you
might have multiple instances of (100,10) in your list, be sure that this
is the behavior you actually want.

HTH,
Anna



From paul at prescod.net  Tue Apr  1 13:55:14 2003
From: paul at prescod.net (Paul Prescod)
Date: Tue, 01 Apr 2003 10:55:14 -0800
Subject: Vancouver Python XML User's Group
Message-ID: <3E89E092.9000907@prescod.net>

Smoke: an automated build, regression and performance tracking toolset
Presented by Trent Mick

Tuesday April 1, 2003 7:00 PM
ActiveState, 580 Granville St.

Smoke is a SQL-backed client-server system used to manage information 
about builds of software packages as the source code base evolves, which 
builds pass, which regression tests pass, and how performance metrics 
evolve with the software. Smoke is open source, and built on top of 
Quixote and Chaco.

http://www.vanpyz.org/news





From martin at v.loewis.de  Fri Apr 18 04:46:55 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 18 Apr 2003 10:46:55 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org>
Message-ID: <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de>

Steven Taschuk <staschuk at telusplanet.net> writes:

> > At the same time, I dismissed all of my Stackless 1.0 code, which was
> > continuation-based, an absolute no-no in Guido's eyes. [...]
> 
> Is there a short way to say why Guido didn't want continuations?

He didn't like the implementation strategy; I believe he was neutral
about the feature per se.

> Or should I go digging in the archives?

Yes.

Martin



From claird at lairds.com  Tue Apr  1 09:50:19 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 01 Apr 2003 14:50:19 -0000
Subject: else: w/o except: - why not?
References: <mailman.1049156943.12744.python-list@python.org>
Message-ID: <v8j9pb3v2cnra6@corp.supernews.com>

In article <mailman.1049156943.12744.python-list at python.org>,
Tim Peters  <tim.one at comcast.net> wrote:
			.
			.
			.
>> try:
>>     # some code that may except
>> except:
>>     # and if it does, then, okay, just ignore it
>>     pass
>> else:
>>     # but if it didn't except, I want to do this
>
>Now I know what you intended, assuming you wrote what you meant.  Reason
>enough for me.  If you want the absence of an except clause to imply the
>presence of a catch-everything except clause, there's really no chance:  a
>bare except is bad practice.  For example, it catches SystemExit and
>KeyboardInterrupt too, and bare excepts rarely *intend* to catch those.
>Guido isn't going to make it easier to write poor code.
>
>

I'm leaving aside Mr. Hand's goals; I'm still unsure about them.

I want to be clear about the judgment that "a bare except is bad
practice".  Are you talking about any unqualified except?  Is
  try:
      f1()
  except ZeroDivisionError:
      f2()
  except:
      f3()
an instance of "a bare except" in the sense that you intend here?
And, for you, does
  try:
      f1()
  except:
      f2()
      raise
become *not* bad practice?
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From logistix at cathoderaymission.net  Thu Apr 17 01:28:43 2003
From: logistix at cathoderaymission.net (logistix)
Date: Thu, 17 Apr 2003 01:28:43 -0400
Subject: Memory leak while looping
In-Reply-To: <1020cbf8.0304162100.6fc49e6f@posting.google.com>
Message-ID: <000001c304a2$368b7cb0$20bba8c0@XP>


> -----Original Message-----
> From: python-list-admin at python.org 
> [mailto:python-list-admin at python.org] On Behalf Of Roger Hancock
> Sent: Thursday, April 17, 2003 1:00 AM
> To: python-list at python.org
> Subject: Memory leak while looping
> 
> 
> I'm just starting to play with writing C extensions for 
> Python.  In the process of doing so, I called one of my 
> extension functions thousands of times only to discover that 
> a leak was eating up all the memory in my system.  Upon 
> further investigation, I found I could produce the same 
> result with the following python code with my extension 
> completely out of the picture:
> 
> #!/usr/bin/env python2
> 
> for i in range(1,10000000):
>     item = i
> 
> 
> I would have thought that the memory for "item" would be 
> freed as new "item"s were created.  However, this doesn't 
> seem to be the case.  I would appreciate any clarification on 
> this issue.
> 
> 

The range function returns a list.  So in this case you're creating a
list of 10 million items before you even enter the body of the for loop.
Try using xrange() instead.  This is an iterator so it doesn't need to
create an entire list in memory.

Were you calling range() in your extension code or doing something else?





From amk at amk.ca  Tue Apr 15 08:01:36 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Tue, 15 Apr 2003 07:01:36 -0500
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <mailman.1049653811.11491.python-list@python.org> <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
Message-ID: <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net>

On Tue, 15 Apr 2003 02:30:49 +0100, 
	Andy Robinson <andy at reportlab.com> wrote:
> Second, it has always seemed to me that there is a very well
> understood model of a web app consisting of a controller, Request,
> Response and a few other things, handled in very similar ways by Java
> servlets, ASP and WebKit.  

I've speculated about adding a standard HTTP{Request,Response} type to the
standard library, but even this seems unlikely because I doubt we can agree
on what should be in it, or what naming convention to use.  (Quixote uses
underscore_names, WebWare uses camelCase; whose users have to change all
their code.)

There's also the inflexibility that comes with putting something in the
standard library; the interfaces can't be changed without breaking existing
users, so it's impossible to fix things.  cgi.py has lots of cruft in it,
but I've never tried to clean it up because I have no idea what applications
out there might be broken, so cgi.py stays crufty.  It's noteworthy that
Quixote's upload handling only got really reliable when Greg Ward wrote his
own upload parsing code and jettisoned use of cgi.py's code.

All this talk of unifying Web frameworks strikes me as pointless.  Different
toolkits have different design principles and different goals; hoping for a
unified one is like hoping Python and Perl will settle their differences and
merge.  The most to hope for is that some of the systems will die out,
leaving two or three more major ones in control of the space. The best we
can do now is to make it easier for users to choose which toolkit to use by
documenting them and comparing their differences.

--amk                                                    (www.amk.ca)
LEAR: I am a man more sinn'd against than sinning.
      -- _King Lear_, III, ii



From ark at research.att.com  Mon Apr 28 12:17:16 2003
From: ark at research.att.com (Andrew Koenig)
Date: Mon, 28 Apr 2003 16:17:16 GMT
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it>
Message-ID: <yu99sms2k35f.fsf@europa.research.att.com>

Alex> It's an important design principle, which I first saw proposed
Alex> in print by Andrew Koenig in his landmark book "C Traps and
Alex> Pitfalls": *all bounded loops should be first-bound-included,
Alex> last-bound-excluded* -- and which I will therefore call
Alex> "Koenig's Principle" although I have never seen the principle
Alex> *named* in the literature (I'mm CC'ing Andrew because of this --
Alex> I know there are already other things named after him, such as
Alex> the "Koenig Lookup" in C++, that he might like to get his name
Alex> AWAY from, so it seems only fair to allow him to disclaim THIS
Alex> one too ASAP if he dislikes it!-).

Nope, I'm not going to disclaim this one :-)

First, a bit of history: What is sometimes called ``Koenig lookup'' in
C++, and officially called ``argument-dependent lookup,'' is not my
invention.  My main association with it was to point out to the
standards committee that there was a problem that had to be solved,
that an existing proposal--I don't know the original author--would
solve the problem, and I knew of no other proposal that would solve
the problem.  Therefore, even though the proposal wasn't perfect,
adopting it would be an improvement.  I still feel that way, even
though the problems are greater than the committee realized at the
time they adopted it.  

--- Readers uninterested in C++ should skip the following ---

Here is an example that shows the magnitude of the problem:

    #include <iostream>

    int main()
    {
        std::cout << "Hello, world!\n";
        return 0;
    }

How does the compiler know where to look for << ?  The answer is that
<< is a member of the std::cout object, and it is overloaded with the
right-argument type of ``const char *,'' which is the type of the
literal "Hello, world!\n".

Now, let's change the program a tiny bit:

    #include <iostream>
    #include <string>

    int main()
    {
        std::string hello = "Hello, world!\n";
        std::cout << hello;
        return 0;
    }

Now, how does the compiler know where to look for <<?  The version of
<< with a right-argument type of ``std::string'' is not a member of
the std::cout object, because making it a member would create an
unnecessary dependency between the I/O and string classes.  But if <<
isn't a member, how does the compiler find it?  Why, in other words,
do we not have to write

        std::operator<<(std::cout, hello);

instead?

This question has to be answered somehow, and I still think that the
best answer available is to say that because the right argument of <<
has a type that is defined in namespace ``std,'' the compiler should
include namespace ``std'' in the list of places it searches for the
declaration of <<.

--- Readers uninterested in C++ can resume here ---

Now let's talk about how to express ranges.

It is certainly true that if you are expressing a range of (signed)
integers, you can use either the last element of the range or one past
the last element as your ending specification.  Similarly, but less
obviously, you can also use the first element or one before the first
element as your starting point.

So, for example, to express the integers 4, 5, 6, 7, you can write [4,
7], (3, 7], [4, 8), or (3, 8).  Here, the [ ] mean that the range
*includes* the number next to the bracket, and the ( ) mean that the
range *excludes* the number next to the parenthesis.  The question,
then, is which kind of range to prefer.

Alex has already pointed out that if you use asymmetric bounds, there
is the great convenience that the number of elements in the range is
equal to the difference between the bounds.  So, for example, if you
use (3, 7] or [4, 8), the difference between the bounds is 4 in either
case, and that's how many elements there are.  However, the real
convenience of using asymmetric bounds doesn't show up until you start
talking about ranges of values that don't have ordering comparisons
available.

Suppose, for example, that you have a singly-linked list of objects.
In Python, you might do that with a class with a ``next'' method that
yields either the next object in the list, or None if you've run off
the end.

Let's suppose that ``listhead'' refers to the first object in the
list (or is None if the list is empty).  Then you might count the
objects in the list this way:

        item = listhead
        count = 0
        while item is not None:
            count += 1
            item = item.next()

Now, suppose that you want to be able to represent an arbitrary
subsequence of the elements in this list.  How would you do so?
Presumably you're going to want to store a pair of references--but
references to what?  You *cannot* use a reference to the first and
last elements of the sequence, because if you do, you will find that
there is no way to represent an empty sequence.

What you can do is use a reference to the first element of the
subsequence and the first element beyond the subsequence.  Then
it all works out.  For example, you can use (listhead, None) to
represent the whole sequence, (listhead, listhead) to represent
the null subsequence at the beginning, (listhead, listhead.next())
to represent only the first element (assuming that there is a
first element), and so on.

Then, presuming that ``begin'' and ``end'' represent the (asymmetric)
bounds, you can process all of the elements in the range by writing

        item = begin
        while item is not end:
            # process item
            item = item.next()

Try writing this loop with symmetric bounds and you'll see why
I don't like them.

Now, of course, there is no requirement to use the same strategy
for integers and iterator-like values.  But what's the advantage
of two different strategies?  I don't see any.

>From the foregoing argument, I conclude that it's easiest
overall to represent *all* ranges asymmetrically: the first
element of the range and the first element beyond the range.


-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From andersjm at dancontrol.dk  Tue Apr  8 10:47:49 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Tue, 8 Apr 2003 16:47:49 +0200
Subject: Making python module using Borland C++ 5
References: <b6u7fp$d2c$1@news.cc.tut.fi> <3e92b527$0$32016$edfadb0f@dread12.news.tele.dk> <b6ue3b$hfb$1@news.cc.tut.fi>
Message-ID: <3e92e0f5$0$10064$edfadb0f@dread11.news.tele.dk>

"Rami Saarinen" <rs55862 at NOSPAMyahoo.com> wrote:
> > Just write a setup.py and run it with --compiler=bcpp.
> 
> Ah, the bcpp option is good to know. 

Also, bcppcompiler.py from distutils is a good source for BCB compiler
options that are compatible with the standard build, for alignment and
such. If your two libs are compiled with different options for
alignment or enum size or such, you'll need to handle that.

- Anders






From cnetzer at mail.arc.nasa.gov  Sat Apr 19 18:09:36 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 19 Apr 2003 15:09:36 -0700
Subject: Tkinter help
In-Reply-To: <48BEA671.1EA9CE07.0CCD53D2@netscape.net>
References: <48BEA671.1EA9CE07.0CCD53D2@netscape.net>
Message-ID: <1050790175.589.13.camel@sayge.arc.nasa.gov>

On Sat, 2003-04-19 at 14:57, SocLcrBoy2 at netscape.net wrote:
> im having trouble learning Tkinter, can anyone point me too
>  a good guide, book, etc.

http://home.att.net/~stephen_ferg/thinking_in_tkinter/
http://www.pythonware.com/library/tkinter/introduction/
http://www.python.org/doc/life-preserver/

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From mwh at python.net  Wed Apr  2 12:35:04 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 2 Apr 2003 17:35:04 GMT
Subject: Pycon DC was terrific
References: <5b4785ee.0303310948.26afe16b@posting.google.com> <slrnb8lgns.ccp.rmunn@localhost.localdomain>
Message-ID: <7h3d6k4ls53.fsf@pc150.maths.bris.ac.uk>

Robin Munn <rmunn at pobox.com> writes:

> Wade Leftwich <wade at lightlink.com> wrote:
> [snip words of praise about PyCon DC 2003]
> > Many thanks to the PSF, the sponsors, the organizers, Guido, and
> > conference chairman / master of ceremonies Steve Holden.
> > 
> > Wade Leftwich
> > Ithaca, NY
> 
> Hear, hear! *applauds*

(with-competive-streak :transatlantic
  EuroPython will be better, dammit!
)

Cheers,
M.
:-)

-- 
  If you don't use emacs, you're a pathetic, mewling, masochistic
  weakling and I can't be bothered to convert you.    -- Ron Echeverri



From max at alcyone.com  Fri Apr  4 17:48:54 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 04 Apr 2003 14:48:54 -0800
Subject: output buffering--am I missing something obvious?
References: <yu99istu9ds9.fsf@europa.research.att.com> <yu99el4i9dfb.fsf@europa.research.att.com>
Message-ID: <3E8E0BD6.EFC3C7FE@alcyone.com>

Andrew Koenig wrote:

> ark Can someone please explain the following behavior?  Under Python
> 2.2.2
> ark and Solaris 2.8:
> 
> >>> import sys
> >>> sys.stdout.write("foo")
> >>> sys.stdout.flush()
> >>> sys.stdout.write("bar\n")
> >>> bar
> 
> Never mind -- I figured it out.  Apparently the interactive loop
> discards all pending output before reading the next input statement.

Weird, here's what I get on my Python 2.2.2 on Solaris 8 (which I built
from source):

>>> import sys
>>> sys.stdout.write('foo')
foo>>> sys.stdout.flush()
>>> sys.stdout.write('bar\n')
bar

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If the sky should fall, hold up your hands.
\__/ (a Spanish proverb)
    The laws list / http://www.alcyone.com/max/physics/laws/
 Laws, rules, principles, effects, paradoxes, etc. in physics.



From duncan at NOSPAMrcp.co.uk  Thu Apr 17 05:02:54 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 17 Apr 2003 09:02:54 +0000 (UTC)
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com> <3e9d9bb9$0$151$a1866201@newsreader.visi.com>
Message-ID: <Xns936064FC1331duncanrcpcouk@127.0.0.1>

grante at visi.com (Grant Edwards) wrote in 
news:3e9d9bb9$0$151$a1866201 at newsreader.visi.com:

>> python << DONE
>> print 2
>> for i in (1,4):
>>      print i
>> DONE
> 
> But how do you do it in a Makefile?

Switch to Windows? Microsoft's nmake happily handles this :^)

Microsoft's syntax isn't, of course, quite identical, they use the << to 
end the document, and create a (optionally named) temporary file which is 
deleted afterwards.
Here's an example I used recently:

.SUFFIXES: .exe .dll .obj .c .cpp .pyx
.pyx.c:
        <<pyrexc.py $<
import Pyrex.Compiler.Main
Pyrex.Compiler.Main.main(command_line = 1, c_only = 1)
<<

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From anthony at interlink.com.au  Wed Apr 16 02:22:05 2003
From: anthony at interlink.com.au (Anthony Baxter)
Date: Wed, 16 Apr 2003 16:22:05 +1000
Subject: Why functional Python matters 
In-Reply-To: <slrnb9pt59.iin.ramen@lackingtalent.com> 
References: <slrnb9pt59.iin.ramen@lackingtalent.com>  <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <mailman.1050453385.17091.python-list@python.org> <b7iigt$1d3jt$1@ID-169208.news.df
Message-ID: <200304160622.h3G6M5u13119@localhost.localdomain>

>>> Dave Benjamin wrote
> Three? I thought it was two! Locals, globals, and...?

builtins.





From ramen at lackingtalent.com  Tue Apr  1 16:41:20 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 21:41:20 -0000
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <3E89EC01.F0CF6D31@engcorp.com> <b6d0jm$n7j$1@mailgate2.lexis-nexis.com>
Message-ID: <slrnb8jv09.5u1.ramen@lackingtalent.com>

In article <b6d0jm$n7j$1 at mailgate2.lexis-nexis.com>, rzed wrote:
> I'm all for this proposal, but I'm afraid their may be some confusion
> with proto-PEP 7334(Romansch Numerals) and PEP 2/14 (Romantic
> numerals), not to mention PEPS 3.95/# (Romaine Minerals) and 5/$1
> (Ramen noodles).

Hey! I resemble that remark! ;)
http://ramenfest.com/



From guido at python.org  Fri Apr 25 20:12:41 2003
From: guido at python.org (Guido van Rossum)
Date: Fri, 25 Apr 2003 20:12:41 -0400
Subject: RELEASED: Python 2.3b1
Message-ID: <200304260012.h3Q0CgE03802@pcp02138704pcs.reston01.va.comcast.net>

Python 2.3b1 is the first beta release of Python 2.3.  Much improved
since the last alpha, chockfull of things you'd like to check out:

  http://www.python.org/2.3/

Some highlights of what's new since 2.3a2:

- sum() builtin, adds a sequence of numbers, beats reduce().

- csv module, reads comma-separated-value files (and more).

- timeit module, times code snippets.

- os.walk(), a generator slated to replace os.path.walk().

- platform module, by Marc-Andre Lemburg, returns detailed platform
  information.

For more highlights, see http://www.python.org/2.3/highlights.html

New since Python 2.2:

- Many new and improved library modules, e.g. sets, heapq, datetime,
  textwrap, optparse, logging, bsddb, bz2, tarfile,
  ossaudiodev, and a new random number generator based on the highly
  acclaimed Mersenne Twister algorithm (with a period of 2**19937-1!).

- New builtin enumerate(): an iterator yielding (index, item) pairs.

- Extended slices, e.g. "hello"[::-1] returns "olleh".

- Universal newlines mode for reading files (converts \r, \n and \r\n
  all into \n).

- Source code encoding declarations.  (PEP 263)

- Import from zip files.  (PEP 273 and PEP 302)

- FutureWarning issued for "unsigned" operations on ints.  (PEP 237)

- Faster list.sort() is now stable.

- Unicode filenames on Windows.

- Karatsuba long multiplication (running time O(N**1.58) instead of
  O(N**2)).

See also http://www.python.org/doc/2.3b1/whatsnew/ - Andrew Kuchling's
description of all important changes since 2.2.

We request widespread testing of this release but don't recommend
using it for production situations yet.  Beta releases contain bugs.
New APIs are expected to be stable, and may be changed only if serious
deficiencies are found.  No new APIs or modules will be added after
the first beta release.  If you have an important Python application,
we strongly recommend that you try it out with a beta release and
report any incompatibilities or other problems you may encounter, so
that they can be fixed before the final release.  To report problems,
use the SourceForge bug tracker:

  http://sourceforge.net/tracker/?group_id=5470&atid=105470

Enjoy!

--Guido van Rossum (home page: http://www.python.org/~guido/)




From staschuk at telusplanet.net  Wed Apr  9 06:36:23 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 09 Apr 2003 10:36:23 -0000
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr  9)
Message-ID: <v97tt7736tvu2e@corp.supernews.com>

QOTW: "Figuring out the simplest thing is often very hard."  Bruce Eckel

"People are so full of context."  Bengt Richter


Discussion:
    A proposal for overloadable assignment is met with general
    disapproval, but provokes the suggestion of a new ':=' operator
    with no default semantics.
        http://groups.google.com/groups?threadm=9383481e.0304012011.78f05548%40posting.google.com

    Informal comparisons of distributed computing technologies,
    including CORBA, XML-RPC, SOAP, ...
        http://groups.google.com/groups?threadm=a626bbd7.0304020829.1e5cac2f at posting.google.com

    More descriptions than you want of the theory of ints and floats.
        http://groups.google.com/groups?threadm=fjss6b.1021.ln at boundary.tundraware.com

Announcements:
    apm.py 0.1.0: The goal of apm.py is providing of an uniform and
    platform independent interface to APM.
        <http://www.iapp.de/~riemer/projects/apm.py/>

    Novelwriting 0.3: Generates random text based on a user-defined
    grammar.
        <http://unpythonic.net/~jepler/novelwriting/index.html>

    pyGDChart2 alpha: A Python interface to the popular GDChart
    graphing library.
        <http://www.nullcube.com/software/pygdchart2.html>

    PyKDE-3.5-1: A set of Python bindings for the basic KDE libraries.
    It allows you to build KDE applications in Python and use KDE
    widgets, dialogs and other features as well.
        <http://www.riverbankcomputing.co.uk/pykde>

    python-ldap-2.0.0pre07: An object-oriented API to access LDAP
    directory servers from Python programs.
        <http://python-ldap.sourceforge.net/>

    PythonMagick 0.4: PythonMagick is a wrapper around GraphicsMagick.
        <http://www.procoders.net/pythonmagick>

    SCons 0.13: SCons is a software construction tool (build tool, or
    make tool) written in Python.
        <http://www.scons.org/>

    SQLObject 0.3: SQLObject is an object-relational mapper.  It
    allows you to translate RDBMS table rows into Python objects, and
    manipulate those objects to transparently manipulate the database.
        <http://sqlobject.org>


========================================================================

Everything you want is probably one or two clicks away in these pages:

    Python.org's Python Language Website is the traditional
    center of Pythonia
        http://www.python.org
    Notice especially the master FAQ
        http://www.python.org/doc/FAQ.html

    PythonWare complements the digest you're reading with the
    daily python url
         http://www.pythonware.com/daily  
    Mygale is a news-gathering webcrawler that specializes in (new)
    World-Wide Web articles related to Python.
         http://www.awaretek.com/nowak/mygale.html 
    While cosmetically similar, Mygale and the Daily Python-URL
    are utterly different in their technologies and generally in
    their results.

    comp.lang.python.announce announces new Python software.  Be
    sure to scan this newly-revitalized newsgroup at least weekly.
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python.announce

    Brett Cannon continues the marvelous tradition established by 
    Andrew Kuchling and Michael Hudson of summarizing action on the
    python-dev mailing list once every other week.
	http://www.python.org/dev/summary/

    The Python Package Index catalogues packages.
        http://www.python.org/pypi/

    The somewhat older Vaults of Parnassus ambitiously collects references
    to all sorts of Python resources.
        http://www.vex.net/~x/parnassus/   

    Much of Python's real work takes place on Special-Interest Group
    mailing lists
        http://www.python.org/sigs/

    The Python Business Forum "further[s] the interests of companies
    that base their business on ... Python."
	http://www.python-in-business.org

    The Python Software Foundation has replaced the Python Consortium
    as an independent nexus of activity
        http://www.python.org/psf/

    Cetus does much of the same
        http://www.cetus-links.org/oo_python.html

    Python FAQTS
        http://python.faqts.com/

    The old Python "To-Do List" now lives principally in a
    SourceForge reincarnation.
        http://sourceforge.net/tracker/?atid=355470&group_id=5470&func=browse
        http://python.sourceforge.net/peps/pep-0042.html
     
    The online Python Journal is posted at pythonjournal.cognizor.com.
    editor at pythonjournal.com and editor at pythonjournal.cognizor.com
    welcome submission of material that helps people's understanding
    of Python use, and offer Web presentation of your work.

    *Py: the Journal of the Python Language*
        http://www.pyzine.com

    Tenth International Python Conference 
        http://www.python10.org            

    Archive probing tricks of the trade:
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python&num=100
        http://groups.google.com/groups?meta=site%3Dgroups%26group%3Dcomp.lang.python.*

Previous - (U)se the (R)esource, (L)uke! - messages are listed here:
  http://www.ddj.com/topics/pythonurl/         
  http://purl.org/thecliff/python/url.html (dormant)
or
  http://groups.google.com/groups?oi=djq&as_q=+Python-URL!&as_ugroup=comp.lang.python


Suggestions/corrections for next week's posting are always welcome.
E-mail to <Python-URL at phaseit.net> should get through.

To receive a new issue of this posting in e-mail each Monday morning
(approximately), ask <claird at phaseit.net> to subscribe.  Mention
"Python-URL!".


-- The Python-URL! Team--

Dr. Dobb's Journal (http://www.ddj.com) is pleased to participate in and
sponsor the "Python-URL!" project.




From bokr at oz.net  Sat Apr 26 14:07:21 2003
From: bokr at oz.net (Bengt Richter)
Date: 26 Apr 2003 18:07:21 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <vuiqa.3522$%_3.2810696@newssrv26.news.prodigy.com> <3EA9B7A4.59F9CE04@easystreet.com> <b8eas6$c14$0@216.39.172.122>
Message-ID: <b8ehsp$upi$0@216.39.172.122>

On 26 Apr 2003 16:07:34 GMT, bokr at oz.net (Bengt Richter) wrote:
[...]
>====< rtfexample.py >========================
>import sys, os
>if len(sys.argv)<3:
>    print 'Usage: rtfexample.py arg1 arg2'
>    raise SystemExit
>template = file('rtfexample.rtf').read().rstrip('\x00')
                                         ^^^^^^^^^^^^^^^-[1]
>places = {'redtext':sys.argv[1], 'bluetext':sys.argv[2]}
>file('rtfout.rtf','w').write(template % places)
>os.system('start write rtfout.rtf')
>=============================================
[...]
[1] You don't need this rstrip -- sorry. But if you isert the
whole file verbatim to make a string in a python file, like
template = """<insert here>"""
The null character will not be legal python source and has to
be edited out or changed to '\x00'. I ran into that and mistakenly
transferred the concept. A null as a final character in the string
*data* is fine, just not in the representation source text.

Regards,
Bengt Richter



From duncan-news at grisby.org  Sun Apr 13 13:48:38 2003
From: duncan-news at grisby.org (Duncan Grisby)
Date: Sun, 13 Apr 2003 17:48:38 +0000
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk> <3e970abc$0$49101$e4fe514c@news.xs4all.nl> <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>
Message-ID: <1e740$3e99a2f6$516049d2$15395@nf1.news-service.com>

In article <71M$aHCTKxl+Ew8l at jessikat.demon.co.uk>,
 Robin Becker  <robin at jessikat.fsnet.co.uk> wrote:

>sigh :(
>for real distributed computing I wouldn't want to hold the sockets open.

Why ever not!?  Most protocols people use for distributed computing do
hold sockets open. As long as connections are seen as a cache, and can
be closed if necessary, it's a good thing. TCP start-up overhead is
huge, especially if the network has high latency.

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --



From martin at v.loewis.de  Sun Apr 13 16:46:51 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 22:46:51 +0200
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk> <m38yuepdva.fsf@mira.informatik.hu-berlin.de> <LtFzMJA4ZZm+EwDh@jessikat.fsnet.co.uk> <3kjma.11793$T34.382463@news2.tin.it>
Message-ID: <m3vfxi15xg.fsf@mira.informatik.hu-berlin.de>

Alex Martelli <aleax at aleax.it> writes:

> The simplest solution is to have pyRXP return a tree of *LISTS*,

Why is that a solution?

>>> x=[]
>>> import weakref
>>> weakref.ref(x)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: cannot create weak reference to 'list' object

Regards,
Martin



From j4nff5m02 at sneakemail.com  Wed Apr 16 09:10:45 2003
From: j4nff5m02 at sneakemail.com (Roel Schroeven)
Date: Wed, 16 Apr 2003 13:10:45 GMT
Subject: wxPython & PyOpenGL
Message-ID: <pDcna.48562$t_2.4267@afrodite.telenet-ops.be>

Hi,

I'm trying to make a small application using wxPython to provide a GUI 
and PyOpenGL hardware acceleration. I know I have to create a 
wxGLCanvas, which will be used by PyOpenGL functions. I wrote some GUI 
code and adapted the code from the first example in the Red Book to fit 
in, and it works. Sorta. I can set the background color, but nothing 
else. No exceptions are thrown, the canvas just doesn't display anything 
(besides the background color).

In case it has anything to do with it, I'm using ActivePython on Windows 
2000 (though I'm planning to do most of the development on Linux).

Am I doing something wrong? Is there a problem with 
ActivePyton/Windows/wxPython/PyOpenGL? Any toughts are more than welcome.

The code I have so far:


from wxPython.wx import *
from wxPython.glcanvas import *
from OpenGL.GL import *

def DrawGL():
     glClearColor(0.8, 0.1, 0.1, 0.0)
     glClear(GL_COLOR_BUFFER_BIT)
     # (so far it works)

     glColor3f(1.0, 1.0, 1.0)

     glMatrixMode(GL_PROJECTION)
     glLoadIdentity()
     glOrtho(-1.0, 1.0, -1.0, 1.0, -1.0, 1.0)

     glBegin(GL_POLYGON)
     glVertex2f(-0.5, -0.5)
     glVertex2f(-0.5, 0.5)
     glVertex2f(0.5, 0.5)
     glVertex2f(0.5, -0.5)
     glEnd()
     # (but all this has no effect whatsoever)

     glFlush()


class MyGL(wxGLCanvas):
     def __init__(self, parent):
         wxGLCanvas.__init__(self, parent)
         EVT_PAINT(self, self.OnPaint)

     def OnPaint(self, event):
         dc = wxPaintDC(self)
         self.SetCurrent()
         DrawGL()
         self.SwapBuffers()

class MyFrame(wxFrame):
     def __init__(self, parent, ID, title):
         wxFrame.__init__(self, parent, ID, title, wxDefaultPosition, 
wxSize(700, 500))
         self.GLCanvas = MyGL(self)

         MainSizer = wxBoxSizer(wxVERTICAL)
         MainSizer.Add(wxButton(self, -1, "top"), 0, wxEXPAND)

         MiddleSizer = wxBoxSizer(wxHORIZONTAL)
         MiddleSizer.Add(wxButton(self, -1, "left"), 1, wxEXPAND)
         MiddleSizer.Add(self.GLCanvas, 2, wxEXPAND)
         MiddleSizer.Add(wxButton(self, -1, "right"), 1, wxEXPAND)
         MainSizer.Add(MiddleSizer, 1, wxEXPAND)

         MainSizer.Add(wxButton(self, -1, "bottom"), 0, wxEXPAND)

         self.SetSizer(MainSizer)
         self.SetAutoLayout(True)
         #MainSizer.Fit(self)

class MyApp(wxApp):
	def OnInit(self):
		frame = MyFrame(NULL, -1, "Python/OpenGL")
		frame.Show(True)
		self.SetTopWindow(frame)
		return True

app = MyApp(0)
app.MainLoop()



-- 
"Codito ergo sum"
Roel Schroeven




From max at cNOvSisiPonAtecMh.com  Thu Apr 10 18:44:10 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Thu, 10 Apr 2003 22:44:10 GMT
Subject: Can I define nested classes in Python?
References: <169a7283.0304101432.13f5786c@posting.google.com>
Message-ID: <_smla.1698$mC1.731296@twister.nyc.rr.com>

The answer to both of your questions is 'yes'. It appears that you are in
need of some python beginner book or tutorial, asking such basic questions
via the newsgroup is certainly the long way to learning a language ;).
Here are some free books to start with:

http://www.tcfb.com/freetechbooks/bookphyton.html


--
========================================
Max Khesin, software developer -
max at cNvOiSsPiAoMntech.com
[check out our image compression software at www.cvisiontech.com, JBIG2-PDF
compression @
www.cvisiontech.com/cvistapdf.html]


"sdieselil" <sdieselil at yahoo.com> wrote in message
news:169a7283.0304101432.13f5786c at posting.google.com...
> Hi
>
> Can I define nested classes in Python and what access do they have to
outer
> classes?





From regnivon at netscape.net  Mon Apr 21 12:01:38 2003
From: regnivon at netscape.net (scn)
Date: 21 Apr 2003 09:01:38 -0700
Subject: looping and deleting
Message-ID: <3cf72d52.0304210801.2128eca2@posting.google.com>

what is a good method for looping over a file line-by-line and
deleting the first 3 characters?

thank you.



From nav at adams.patriot.net  Fri Apr 11 09:29:19 2003
From: nav at adams.patriot.net (Nick Vargish)
Date: 11 Apr 2003 09:29:19 -0400
Subject: Promoting Python
References: <qGqla.61$Tj7.91655@news.uswest.net>
Message-ID: <yyyistlf9hs.fsf@adams.patriot.net>

Shame about python.com... Doesn't look respectable at all. :^(

Nick

-- 
#include<stdio.h> /* SigMask 0.3 (sig.c) 19990429 PUBLIC DOMAIN "Compile Me" */
int main(c,v)char *v;{return !c?putchar(*v-1)&&main(0,v+ /* Tweaks welcomed. */
1):main(0,"Ojdl!Wbshjti!=obwAqbusjpu/ofu?\v\1");}  /* build: cc -o sig sig.c */



From santoshi_reddy at hotmail.com  Tue Apr  1 17:00:54 2003
From: santoshi_reddy at hotmail.com (Santoshi Reddy)
Date: Tue, 01 Apr 2003 17:00:54 -0500
Subject: (no subject)
Message-ID: <F12rbu5ZkmH5cEXalfz000262bd@hotmail.com>

Hi,

I am new to python.can Someone tell me what _int_ is?

Thanks,
S.




_________________________________________________________________
STOP MORE SPAM with the new MSN 8 and get 2 months FREE*  
http://join.msn.com/?page=features/junkmail





From mcherm at mcherm.com  Thu Apr  3 09:18:27 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Thu,  3 Apr 2003 06:18:27 -0800
Subject: Comments
Message-ID: <1049379507.3e8c42b33acae@mcherm.com>

Carsten Gaebler wrote:
> But multi-line strings are evil, too. At least when they occur in an 
> indented block and leading whitespace is not allowed in the string 
> your're building

Greg Ewing replied:
> Last time I pondered this issue, I decided that the
> correct way to fix this would be to have a statement
> like
> 
>    string my_string:
>      |  The string can contain any characters and
>      |have multiple lines and leading whitespace,
>      |and its indentation level is made clear by
>      |the leading | characters.
> 
> Teaching the tokenizer to understand this could
> be an interesting exercise, however...

Would this work for you?

     myString = stripLeading("""
       |  The string can contain any characters and
       |have multiple lines and leading whitespace,
       |and its indentation level is made clear by
       |the leading | characters.""")

If so, it has three advantages. First of all, no need for a
type declaration (been doing too much C lately?). Secondly,
it is clear where the construct ends (thus being a bit more 
readable if your string ended in a newline). Secondly, it 
can be implemented without having to teach the tokenizer 
anything.

     import re
     def stripLeading(s):
         prefix = re.match('\n[ \t]*.', s).group()
         return myString.replace(prefix, '\n')[1:]

I'm not sure this is the single BEST way to write indented
multi-line strings (personally I'd avoid the line of |'s),
but if you like it, it's pretty darn easy to get.

-- Michael Chermside




From max at alcyone.com  Mon Apr 14 20:26:24 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 17:26:24 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl>
Message-ID: <3E9B51B0.BDE36C0C@alcyone.com>

Irmen de Jong wrote:

> Erik Max Francis wrote:
> 
> >>>>Course(lowEarthOrbit, lowMarsOrbit).transfers()
> >
> > [<ExtractionTransfer @ 0x82893ac (out of `low orbit around
> > `Earth'')>,
> > <HohmannTransfer @ 0x812cef4 (Hohmann from `Earth' to `Mars')>,
> > <InsertionTransfer @ 0x82adb64 (into `low orbit around `Mars'')>]
> 
> Is this actual code? What does this do? I'm interested.

One of my side projects is a computer roleplaying game set in a real
Solar System with realistic orbital mechanics, drives, and fuel
requirements.  As a precursor to that (and whether or not anything ever
comes of the game), I've been working on a "calculator" that will allow
one to compute all the data tables required for the game.

It requires a lot of physical data, and it can't really be used for
industrial strength calculations, since it employs a lot of
approximations to make the calculations manageable (coplanar, prograde,
circular orbits), and some of the calculations are placeholders, but it
can still do a fair amount of stuff (all figures are in SI units):

max at oxygen:~/projects/adrift% python -i stella.py 
>>> info(lowEarthOrbit, lowMarsOrbit)
Source: low orbit around `Earth'
Destination: low orbit around `Mars'
Opportunities:
 Opportunity: Earth/Mars
  Angle: 0.773813615969 rad (44.3362543248 deg)
  Period: 67410366.2965 s
Transfers:
 Transfer: out of `low orbit around `Earth''
  Duration: 0.0 s
  Burns: 3250.27120897 m/s
 Transfer: Hohmann from `Earth' to `Mars'
  Duration: 22367816.9538 s
  Burns: 2943.59381931, 2647.9758885 m/s
 Transfer: into `low orbit around `Mars''
  Duration: 0.0 s
  Burns: 1450.74569639 m/s
Course: from `low orbit around `Earth'' to `low orbit around `Mars''
Duration: 22367816.9538 s
Deltavee: 10292.5866132 m/s

And it knows all sorts of other things, too:

>>> lowEarthOrbit.orbitalPeriodAroundPrimary()
5181.5057206353176 # s
>>> earth.rocheLimit()
556362389.03351605 # m
>>> pluto.insolation()/earth.insolation()
0.00064958695692894674 # Sun is that much dimmer from Pluto than 
                       # from Earth
>>> venus.temperature()/venus.idealTemperature()
2.249475368547301 # that's thermodynamic temperature ratio -- now 
                  # THAT'S a greenhouse effect!
>>> triton.angularRadiusOfPrimary()/earth.angularRadiusOfSecondary(moon)
15.335201130577547 # from Triton, Neptune looks 15 times bigger than
                   # the Moon from Earth
>>> pluto.atmosphereMass()
9220763709116.8496 # kg, computed from surface gravity and pressure
>>> io.tidalStrengthFromPrimary()/earth.tidalStrengthFromSecondary(moon)
19581.842788149675 # Jupiter's tides on Io are this much strong than
                   # the Moon's on Earth

And since I've recently integrated information on all named objects in
the Solar System, you can even ask about obscure asteroids:

>>> SYSTEM['Yamatotakeru'].distance()
408470876880.0 # m, distance from the Sun
>>> Course(earth, SYSTEM['Yamatotakeru']).deltavee()
11077.130833824534 # m/s

and on and on.

Since the calculator is distinct from the game itself, whatever happens
(or doesn't happen :-) with the game won't affect the availability of
the calculator when the time comes.  And yes, I'll be happy to release
it as open source ... :-).

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Society attacks early when the individual is helpless.
\__/ B.F. Skinner
    CSBuddy / http://www.alcyone.com/pyos/csbuddy/
 A Counter-Strike server log file monitor in Python.



From ramen at lackingtalent.com  Fri Apr 25 17:57:30 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Fri, 25 Apr 2003 21:57:30 -0000
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <slrnbajav2.e5n.ramen@lackingtalent.com> <pan.2003.04.25.17.49.27.407432@jerf.org>
Message-ID: <slrnbajbre.e5n.ramen@lackingtalent.com>

In article <pan.2003.04.25.17.49.27.407432 at jerf.org>, Jeremy Bowers wrote:
> for memory in brain.iter_memories():
> 	print memory

Of course! I knew iterators were useful for something... it would take an
awful lot of memory to construct a list first.

> Implementation of the brain object left as an exercise for the reader. I
> know it can be done, as I am in possession of such an object right now.

It'd make for one hell of a base class.

Dave



From lost at soul.net  Sat Apr 26 10:19:35 2003
From: lost at soul.net (lost)
Date: Sat, 26 Apr 2003 15:19:35 +0100
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.29.48.204375@soul.net> <RZlqa.4182$lf3.712485@cletus.bright.net>
Message-ID: <pan.2003.04.26.14.19.35.262492@soul.net>

On Fri, 25 Apr 2003 22:13:35 -0400, George Kinney wrote:
> from glob import glob
> file_list = glob.glob(SomePath)
> outfile = open(OutputFilename, 'w')
> 
> for file in file_list :
>     print file
>     outfile.write(file + '\n')

thanks for the info, however that doesn't seem to work either? here's the
whole program that I'm trying:
#!/usr/bin/python
from glob import glob
file_list = glob.glob('/var/www/html/') for each_file in file_list:
        print each_file
print "done"

Running that gives me:
Traceback (most recent call last):
  File "files.py", line 6, in ?
    file_list = glob.glob('/var/www/html/')
AttributeError: 'function' object has no attribute 'glob'

Any ideas where I'm going wrong?
Thanks again...



From oktaysafak at ixir.com  Fri Apr 25 09:28:23 2003
From: oktaysafak at ixir.com (Oktay Safak)
Date: Fri, 25 Apr 2003 16:28:23 +0300
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <Kq9qa.9083$h%2.957726@read1.cgocable.net> <mailman.1051273778.5621.python-list@python.org> <fzo6emyy.fsf@python.net>
Message-ID: <004501c30b2e$8ed60e20$0200a8c0@b>

And what about effbot's Tkinter 3000? Has anybody got first hand experience
with it, especially with its ability to make brand new widgets from scratch,
entirely in python?


----- Original Message -----
From: "Thomas Heller" <theller at python.net>
Newsgroups: comp.lang.python
To: <python-list at python.org>
Sent: Friday, April 25, 2003 4:16 PM
Subject: Re: GUI libs


> Gerhard Haering <gh at ghaering.de> writes:
>
> > Mark Nenadov wrote:
> > > Alessio Pace wrote:
> > >
> >
> > >>Hi, I'd like to make a GUI for a program written in python, meant to
run
> > >>on mac, win and linux: which is the preferable solution? I saw around
> > >>tkinter, wxpython and pyqt....
> > >
> > > I'd strongly recommend wxPython. wxPython is quite a rich GUI
> > > library and I have found it to be more advanced/modern/feature-rich
> > > than tkinter.
> >
> >
> > You'll get a Python interface with the error behaviour of the wrapped
> > C++ library. The default is to crash. In my limited experience, anyway.
>
> Not for me, but I can only speak for Windows.
>
> If you go the wxPython route, you should consider pythoncard, which is a
> layer on top of wxPython.
>
> Thomas
> --
> http://mail.python.org/mailman/listinfo/python-list
>





From ckea25d02 at sneakemail.com  Wed Apr 16 23:02:27 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 17 Apr 2003 15:02:27 +1200
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
In-Reply-To: <gcjna.35728$T34.931514@news2.tin.it>
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com> <gcjna.35728$T34.931514@news2.tin.it>
Message-ID: <b7l4qt$23f39$1@ID-169208.news.dfncis.de>

Alex Martelli wrote:
> Forcing linebreak after each & every colon?  Oh well -- he's probably
> getting bribes from large-screen and vertical-aspect-ratio monitor
> manufacturers, is my guess...;-)

Rumour has it the PSU has a secret plan to quadruple
the number of programmer's cubicles that can be fitted
into the average office space, by making everyone use
monitors 5 feet high and 6 inches wide.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From max at alcyone.com  Tue Apr 22 20:28:14 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 17:28:14 -0700
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org>
		 <3EA5B173.9420324C@alcyone.com> <mailman.1051056037.8077.python-list@python.org>
Message-ID: <3EA5DE1E.22704437@alcyone.com>

Ian Bicking wrote:

> Syntax-related errors/warnings aren't runtime.

That's highly poopy.  I guess you could do gymanstics like:

Python 2.3a2 (#1, Feb 19 2003, 20:06:49) 
[GCC egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> None = 'asdf'
<stdin>:1: SyntaxWarning: assignment to None
>>> __builtins__.__dict__['None'] = 'asdf'
>>> None
'asdf'

but at this point it's looking awfully ugly.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Men play the game; women know the score.
\__/ Roger Woddis
    ZOE / http://www.alcyone.com/pyos/zoe/
 A simple Python OpenGL rendering engine.



From max at alcyone.com  Thu Apr 10 23:12:24 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 10 Apr 2003 20:12:24 -0700
Subject: Can I define "static" (like in C++) methods in Python?
References: <169a7283.0304101430.2bec922b@posting.google.com>
Message-ID: <3E963298.21891CAD@alcyone.com>

sdieselil wrote:

> Can I define methods which are attributes of class and not attributes
> of
> instance?

Oops, I missed the part about methods in there.  Sorry about that.  As
other people have pointed out, the right way to do that (with modern
versions of Python) is staticmethod.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The basis of optimism is sheer terror.
\__/ Oscar Wilde
    Rules for Buh / http://www.alcyone.com/max/projects/cards/buh.html
 The official rules to the betting card game, Buh.



From tismer at tismer.com  Thu Apr 17 00:40:32 2003
From: tismer at tismer.com (Christian Tismer)
Date: Thu, 17 Apr 2003 06:40:32 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
Message-ID: <3E9E3040.5040409@tismer.com>

Dear community, dear Stackless addicts, dear friends,

Ich habe Euch wirklich was zu erz?hlen, liebe Freunde,

I really have to tell you a story!

During the last four months, I have been struggling with
Stackless Python, and especially with myself and how to
get re-focused on my major project which you know very well.
Some of you might know quite well too how hard this was for me,
especially in the context of my parent's endangeroured health.
This particular problem seems to be solved,
for the moment, so let's celebrate the moment, celebrate the moment!

Without going into details, I would like to tell you about the
current status of Stackless Python.
For short, like an abstract, Stackless 3.0 is something like an
or-merge of Stackless 1.0 and 2.0 technology.

Guido, Tim, you both will probably remember my lengthy approaches
to introduce those continuations, years ago, you both convinced
me to drop them, and I did what I was supposed to do. I'm hopefully
a proper citizen, right now. Anyway, you know I'll never really be...

After a long period of depression, I re-invented Stackless in early
2002, with a version number of 2.0, denoting that I had dropped all the
1.0 paradigms (as there are: (1) try to keep compatible, (2) do minimal
changes only, (3) absolutely avoid assembly code at all)

At the same time, I dismissed all of my Stackless 1.0 code, which was
continuation-based, an absolute no-no in Guido's eyes. I still do think
that TimP wasn't that conformant to this "nono"-statement, after I read
a lot of his comments, especially side-notes on the thread-sig,
but this time Guido's veto was clearly stronger than Tim's arguing,
a thing that doesn't happen so often, but I'm proactively respecting
this, positively.

Now, after all that rubbish, let's go into facts, which are quite
interesting.

-------------------------------------------------------------

Today, I finished Stackless Python 3.0, alpha 3.0.1!

First of all, I would like to talk about the new principles.
Yes, no, there are no longer continuations in that sense.
I'm meanwhile convinced that we don't want to support them,
any longer, although I'm happy that Stackless allowed me to
learn *all* any much more about them that that is avalable
on the wor(th|ld) w/h)i(d|l)e net!!

Q&A:

Q: What is it about that Stackless 3.0, will this guy never shut up???

A: No, he most probably never will, unless he's dead, and this is
another 40 or more years in advance, for heaven's sake.

Q: So, what is it about that Stackless 3.0 hype around since months?

A: Simple! Stackless 3.0 has all the hardware switching stuff in it
that Stackless 2.0 had. Stackless 3.0 also incorporates 80% of the
soft switching protocol that Stackless 1.0 had.
But there are a lot of new features:
Stackless has again shown how to marry the impossible with the
imbelievable, and this is the new concept of Stackless 3.0:
There is a maerge between (1.0) soft context switching and (2.0)
hard context switching, which always does the most reasonable thing.

There are a lot of benefits which stem from this hybrid solution,
which will appear in one of my most recent papers, pretty soon.

--------------------------------------------------------------

BLURB

Let me simply end this pamphlete with some simple sentences:
Stackless Python is more capable of tasklets switching than any
other light-weight threading software package.
If anyone disagrees, please give me a runnable counter-example.

Here are some impressive site-specific time measurements, which
especially show, that 20.000.000 cframe tasklet switches per
second are really, really hard to beat.

Pythonon Win32:

D:\slpdev\src\2.2\src\Stackless\test>..\..\pcbuild\python taskspeed.py
10000000 frame switches      took 3.83061 seconds, rate =    2610551/s
10000000 frame softswitches  took 2.40112 seconds, rate =    4164718/s
10000000 cfunction calls     took 2.13033 seconds, rate =    4694098/s
10000000 cframe softswitches took 0.49296 seconds, rate =   20285627/s
10000000 cframe switches     took 1.98907 seconds, rate =    5027486/s
10000000 cframe 100 words    took 3.93737 seconds, rate =    2539768/s
The penalty per stack word is about 0.980 percent of raw switching.
Stack size of initial stub   = 14
Stack size of frame tasklet  = 58
Stack size of cframe tasklet = 35

D:\slpdev\src\2.2\src\Stackless\test>

Python on Debian

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From afriere at yahoo.co.uk  Sun Apr 27 23:54:49 2003
From: afriere at yahoo.co.uk (Asun Friere)
Date: 27 Apr 2003 20:54:49 -0700
Subject: Vim Newbie ? Python Edit/Compile/Run  cycle and integrating py help questions
References: <a791ab19.0304240532.5a025b81@posting.google.com> <3EA7EBD3.3010605@ghaering.de> <mailman.1051243975.12570.python-list@python.org>
Message-ID: <38ec68a6.0304271954.340c28f0@posting.google.com>

Hans Nowak <zephyr01 at alltel.net> wrote in message news:<mailman.1051243975.12570.python-list at python.org>...
> Gerhard Haering wrote:
> I use Tab to indent and Shift-Tab to dedent (a line, not a block). 
 
IMHO it is much preferable to use ^T (insert mode) to indent and ^D
(insert mode) to dedent.  To control the size of the indent set the
shiftwidth property (eg set sw=4).  The reason I prefer this way of
indenting is that it does not introduce Tabs into the source, which is
an especial advantage in Python code.  The underlying ex commands
corresponding to '^T' and '^D' are '>' and '<'.  (eg ':237,348 >>' 
will indent lines 237 to 384 by 2 (>>) shifts, ':5,12 <' will dedent
lines 5 to 12 by one).

Note that ^T/^D work more nicely under vim than they do under trad vi,
as it makes no difference (in vim) where in the line your cursor is,
vim will automagically (in|de)dent the entire line, and not split the
line like vi does.

You personally (assuming you are already working with tab studded
python code) should probably stick to using tabs (for fear of invoking
precisely the sort of indent-hell that not using tabs is supposed to
avoid), but any other vi(m)? newbie, or vi using python newbie, really
ought to use vi's built-in indentation mechanism and not stuff up your
source with tabs.



From rqaox at yandex.ru  Tue Apr  8 09:25:57 2003
From: rqaox at yandex.ru (=?windows-1251?b?0/fl4e376SDk5e/g8PLg7OXt8g==?=)
Date: Tue, 08 Apr 2003 13:25:57 +0000
Subject: =?windows-1251?b?xOv/IOTo8OXq8u7w4CDv7iDs4PDq5fLo7ePzLiAgcw==?=
Message-ID: <200304081325.K5HTFC6K@hvluiyzacs.com>

      ? ????? ???????????? ?????? ?? ????? ??????????? ?? ?????? ???????????
????????????? ????????????, ?? ? ??????? ????????????????????? ???????????,
???????????? ???????? ????, ??????????? ?????, ???????????? ??????????? ??
??????????? ????????. 
      ????? ?? ????? ??????????? ??????????? ??????? ???????????, ??????
?????????? ??????? ???????? ?????????? ?????? ??????, ????????? ??????
????????. ??????????, ?????? ??, ??? ??????????, ??????????? ????????????
????????? ???????? ???????.
      ? ????? ?????????? ???????????? ???????? ????? ??? ???????? ?????
????????????? ????????????? ?????, ??????? ???????? ??????? ????? ??? ????
??????? ? ???????????. ??? ??????? ?????????? ?????, ??????? ????? ??????
?????????? ?? ??????????, ?? ??????? ?????????? ?????????? ? ???????????,
????????? ?? ? ?????????


          ?????????? ??????????????????? ????????-??????????:

                           ???????????? ???? 
                      ??? ????????? ?? ??????????

                       14 - 20  ?????? 2003 ????


 ? ????????? ?????:

- ????? ?????????? ???????????
- ???????? ????????????? ???????????????? ????????????? ??????
- ?????????? ??????? ????????????? ?????????????
- ?????????? ?????????????? ?????
- ?????? ????????????????????? ??????????? 
- ?????????? ?????, ??????? ????????
- ??????????? ???????? ????
- ?????????????? ? ?????? ??????????
- ????????????? ????????????
- ????????????? ???????????
- ???????????????? (???????? ?????????? ??????? ?? ?????) 

      ?????? ???? ????? ????????????? ???????????? ??????????????. ? ????????
???????? ????????? ??????? ?????????? ???????, ?????????? ? ???????? ????????-
??? ????????????? ????????????.
      ?????? ????????????? ??????? ??? ?????? ?? ???????? ? ??????? ???????
????????? ???????: ????? ???? "?????????? ???????????", ???????????? ??
????????????? ???????.

      ????? ???????? ????????? ?????????? ??? ????????? ?????, ????????? ??
???. (095) 258-51-85


malvpvt fvjsl




From dwelch91 at attbi.com  Sat Apr 12 23:01:11 2003
From: dwelch91 at attbi.com (djw)
Date: Sun, 13 Apr 2003 03:01:11 GMT
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
Message-ID: <Xp4ma.109353$ug3.212688@rwcrnsc51.ops.asp.att.net>

Istvan Albert wrote:

> 
> I have been working with perl and java for quite a
> few years now and I am currently actively looking
> to replace perl wit something better, a language that
> promotes software engineering principles. I was
> directed toward python and ruby, I have been
> working with both of them for a couple of weeks now.
> 
> And now for the python question:
> 
> Suppose that I'm looking for a way to
> read lines from a text file:
> 
> now there is the readlines() method
> that slurps everything up (no good for me
> since I don't like to build into my systems
> implicit assumptions about filesizes)
> 
> if I choose to process the file line by line
> the "Learning Python" book advises me to use the
> following atrocity:
> 
> while 1:
> line = file.readline()
> if not line: break
> 
> Maybe I'm picky here but having to start an infinite loop
> then needing a conditional to  break out of it is
> anything but an elegant pattern.
> 
> This has left me wondering about python,
> 
> cheers,
> 
> Istvan.

I recommend that you put down the (outdated) "Learning Python" book and pick 
up the (new) "Python Cookbook". 

A more recent (Python 2.2+ish) way of doing this task is:

for line in open('filename'):
        print line 

See section 4.1 in "Python Cookbook" for more info.

/djw/



From max at cNOvSisiPonAtecMh.com  Tue Apr 15 15:19:01 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Tue, 15 Apr 2003 19:19:01 GMT
Subject: super-problem
References: <J_Xma.4475$Up3.2766033@twister.nyc.rr.com> <Pine.WNT.4.53.0304152114130.326@fagnt.zentrale.feldmann-mg.com>
Message-ID: <FWYma.10297$TR5.4201@twister.nyc.rr.com>

thanks!





From mwh at python.net  Thu Apr 17 04:17:14 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 17 Apr 2003 08:17:14 GMT
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <mailman.1050529828.3295.python-list@python.org>
Message-ID: <7h3ptnlpmgm.fsf@pc150.maths.bris.ac.uk>

Steven Taschuk <staschuk at telusplanet.net> writes:

> Quoth Terry Reedy:
>   [...]
> > I like having the actual values too.  It should make debugging an
> > unexpected failure easier.  But what if each value is more that a few
> > bytes?  Example (from my future plans): two lists of all permutations
> > of range(8), one produced by an obvious recursion and the other by a
> > much less obvious iteration.
> 
> Now and then I replace
>     self.assertEquals(longlist1, longlist2)
> with
>     map(self.assertEquals, longlist1, longlist2)
> 
> It's not quite ideal; in some cases it would be nice to see the
> index of the mismatch too, for example.

I've used ndiff (or rather difflib) for this.

Cheers,
M.

-- 
  > It might get my attention if you'd spin around in your chair,
  > spoke in tongues, and puked jets of green goblin goo.
  I can arrange for this.  ;-)            -- Barry Warsaw & Fred Drake



From rob at zoism.org  Sat Apr  5 04:22:51 2003
From: rob at zoism.org (Rob Brown-Bayliss)
Date: 05 Apr 2003 21:22:51 +1200
Subject: segfault with imp.load_module()
Message-ID: <1049534591.20683.22.camel@musicbox>

Hi, I can do this in interactive python:

>>> f,n,d = imp.find_module("oda",["./items/"])
>>> print f,n,d
<open file './items/oda.py', mode 'r' at 0x82c6db8> ./items/oda.py
('.py', 'r', 1)
>>> it = imp.load_module("oda",f,n,d)
>>> print it
<module 'oda' from './items/oda.pyc'>

but this code:
file, filename, description = imp.find_module(item, ["./items/"])
print file, filename, description
item_module = imp.load_module(item,file, filename, description)
file.close()

gets this:

[rob at musicbox BHH-V]$ python utils.py
<open file './items/oda.py', mode 'r' at 0x81c2af8> ./items/oda.py
('.py', 'r', 1)
Segmentation fault


-- 

*  Rob Brown-Bayliss
*  =================	
*      zoism.org






From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Sun Apr 27 18:23:38 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Mon, 28 Apr 2003 00:23:38 +0200
Subject: RuntimeError at end of test.autotest in Python 2.3b on windows,
 tmpfile also fails
In-Reply-To: <3eaaa408$0$49109$e4fe514c@news.xs4all.nl>
References: <mailman.1051367863.3218.python-list@python.org> <3eaaa408$0$49109$e4fe514c@news.xs4all.nl>
Message-ID: <3eac5942$0$49098$e4fe514c@news.xs4all.nl>

Irmen de Jong wrote:
> Tim Peters wrote:
> 
>> [Irmen de Jong]
>>
>>> when running "import test.autotest" in Python2.3b on windows xp,
>>> I get a RuntimeError at the end:
>>
>>
>>
>> As the download page says, please report bugs on SourceForge.
> 
> 
> You're right ofcourse but I first wanted to be sure it is a problem
> indeed and not some kind of artifact...
> 
> I'll report what I know on SF.
> 
> --Irmen
> 

This issue has been resolved;
https://sourceforge.net/tracker/?func=detail&aid=728101&group_id=5470&atid=105470

Thanks, Tim!

--Irmen


PS if developers don't run "import test.autotest" what *do* they run?




From LogiplexSoftware at earthlink.net  Sat Apr 19 03:24:33 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 19 Apr 2003 00:24:33 -0700
Subject: Why functional Python matters
In-Reply-To: <KeAna.115813$It5.26090@news2.central.cox.net>
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
	 <3E9C9A91.8090209@removethis.free.fr>
	 <m27k9vfcz3.fsf@mycroft.actrix.gen.nz>
	 <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk>
	 <3E9EB3E9.1040208@removethis.free.fr>
	 <87k7dt5m3j.fsf@interim.henrik-motakef.de>
	 <1azna.115402$It5.22896@news2.central.cox.net>
	 <20030417114316.M13181@lain.cheme.cmu.edu>
	 <KeAna.115813$It5.26090@news2.central.cox.net>
Message-ID: <1050737073.1723.30.camel@software1.logiplex.internal>

On Thu, 2003-04-17 at 09:02, Steve Holden wrote:

> Did I mention that you only need nine commands to write any program? ;-)

Yes, but what many fail to realize is that the nine commands can never
be referred to directly, not even in casual conversation.  This is
evident from the ruebot's reluctance to explicitly name them or write
code that describes them in any fashion.  This being the case, even
writing something as seemingly simple as hello_world.vic would require a
seemingly endless entanglement of pointers.  Only at the moment when
equilibrium is achieved between utter incomprehensibility and endless
self-references does a vic program actually begin to function.  In fact,
being either too smart or too dumb will prevent one from grasping the
tao of vic.  The exact IQ requirements aren't certain at this point, but
apparently it's somewhere between being able to put up a web page and
being able to write code.  Also, if your website seems coherent, then
you're probably too smart as well.  It's also known that people too
smart to understand the vic often write working code out of spite and
jealousy.

That definitely makes anyone who writes books like, hm... say _Python
Web Programming_ a loser.  Not to name any names <wink>

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From paul.moore at atosorigin.com  Mon Apr 14 11:44:05 2003
From: paul.moore at atosorigin.com (Paul Moore)
Date: 14 Apr 2003 08:44:05 -0700
Subject: Displaying Unicode on the console (Windows)
Message-ID: <182bcf76.0304140744.1b3d43ee@posting.google.com>

OK. I know this is a common question, and I know the answer is
basically "it's not as easy as you think" :-) But I'm confused as to
what I need to do to get a Unicode string to print on the console in
Windows.

To use a concrete example, I'd like to print the Euro symbol. A
Unicode string for this is u'\20a0'

>>> unicodedata.name(u'\u20a0')
'EURO-CURRENCY SIGN'

Now, I know I can't just print directly:

>>> print u'\u20a0'

Traceback (most recent call last):
  File "<stdin>", line 1, in ?
UnicodeError: ASCII encoding error: ordinal not in range(128)

But what *do* I need to do? None of the obvious encodings work:

>>> print u'\u20a0'.encode("latin-15")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
LookupError: unknown encoding: latin-15

>>> print u'\u20a0'.encode("latin-1")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
UnicodeError: Latin-1 encoding error: ordinal not in range(256)

>>> print u'\u20a0'.encode("iso8859_15")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "C:\Python22\lib\encodings\iso8859_15.py", line 18, in encode
    return codecs.charmap_encode(input,errors,encoding_map)
UnicodeError: charmap encoding error: character maps to <undefined>

>>> print u'\u20a0'.encode("cp1258")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "C:\Python22\lib\encodings\cp1258.py", line 18, in encode
    return codecs.charmap_encode(input,errors,encoding_map)
UnicodeError: charmap encoding error: character maps to <undefined>

>>> print u'\u20a0'.encode("mbcs")
?

I *know* I can display a Euro character on the console. Heck, just
hitting AltGr-4 or Alt-0128 on my keyboard, at the Python prompt,
displays as a Euro character. If I do

>>> print ord('?')
128

Which doesn't look much like 0x20a0, so there's something odd going
on...

As I say, I *know* this is subtle. I'm happy to work out the details,
once I can get a simple example working. But how do I get started? How
do I get u'\u20a0' to display on my screen as a Euro character???

[BTW, in case it's relevant - I know it probably is - the output from
"chcp" at the console prompt is "Active code page: 1252"]

Thanks for any help,
Paul.

PS While I can live with having to know details of how the console is
configured in order to get this working, ideally I'd like to know a
way of getting anything I need out of Windows (even if it requires API
calls...), so that I can write something generic which, when given a
Unicode string, can display it on the console without needing any
extra information from the user...



From martin at v.loewis.de  Thu Apr  3 17:57:25 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 04 Apr 2003 00:57:25 +0200
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <a626bbd7.0304031146.6f98fc28@posting.google.com>
Message-ID: <m3of3n6vfu.fsf@mira.informatik.hu-berlin.de>

escalation746 at yahoo.com (robin) writes:

> martin at v.loewis.de (Martin v. L?wis) wrote:
> > I can't agree with that analysis: CORBA has, of all your alternatives,
> > the least network bandwidth requirements.
> 
> I was referring to overhead like:
> a) writing the interface definition and dealing with ID strings

It is true that you have to write interface definitions. However,
considering that you are doing distributed computing, I can't really
see this as "overhead". Overhead compared to what?

What are ID strings, and why do you need them in CORBA?

> b) more code 

Code written by yourself? Compared to what? A CORBA client in Python
is really short. A CORBA server is larger, but then, the servers
for other distributed computing infrastructures are also larger.

> c) size of the "library"

That depends on the implementation you use. For a client, it is
certainly true that XML-RPC libraries are significantly (factor 10)
smaller than the Fnorb libraries.

Regards,
Martin



From tim.one at comcast.net  Wed Apr  9 13:49:34 2003
From: tim.one at comcast.net (Tim Peters)
Date: Wed, 09 Apr 2003 13:49:34 -0400
Subject: python 2.3a2 Win32 expected failures in regression tests?
In-Reply-To: <e2d925a8.0304090914.2fd8c154@posting.google.com>
Message-ID: <BIEJKCLHCIOIHAGOKOLHOEJFFEAA.tim.one@comcast.net>

[Harri Pasanen]
> I've built python 2.3a2 myself on Win32, using Visual C++ 6.0.

Which flavor of Windows ("Win32" is ambiguous -- they're not identical)?
2.3a2 passed all tests on Win98SE and Win2K before it was released.

> When running the regression tests, the following 3 fail:
>
> test_tarfile, test_timeout, test_zipimport
>
> (test_zip passes)
> All in all, I have 201 tests that pass, but those 3 fail.
>
> I wonder if this is normal state of the affairs for Python 2.3a2 on
> win32, or something in my build that does not click.

See Just's msg about zipimport vs DST.  The other failures aren't expected,
although test_timeout *can* fail sporadically, depending on how busy the
machine is at the time the test is run.  I made that test less sensitive
after 2.3a2 was released.  2.3a2 is so old now that it would be better to
test current CVS Python.





From ckea25d02 at sneakemail.com  Mon Apr 14 02:05:13 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 14 Apr 2003 18:05:13 +1200
Subject: hex and unsigned longs
In-Reply-To: <3n5mm-pg4.ln1@beastie.ix.netcom.com>
References: <xjvof3bu75x.fsf@margvisi.ifi.uio.no> <3n5mm-pg4.ln1@beastie.ix.netcom.com>
Message-ID: <b7diem$deo4s$1@ID-169208.news.dfncis.de>

Dennis Lee Bieber wrote:
>         Since longs have an undefined length, -1L would, to be "unsigned", be 
> represented by an infinite number of FFFFs.

When ints and longs become fully unified, presumably
hex() will have to always return signed results?

Unless one of the earlier suggestions for representing
negative longs as unsigned is adopted, such as 1x...
or Fx.... :-)

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From tim at bladerman.com  Mon Apr  7 19:01:46 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Mon, 07 Apr 2003 23:01:46 GMT
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com>
Message-ID: <urnka.4658$Du.16154@newsc.telia.net>

"Rainer Klueting" <rainer.klueting at epost.de> wrote
> can anybody point me to a source where I can find a datagrid-widget to
> use with Python? It should basically show a flat table in a
> spreadsheet-like way. I tried for a while to do it by myself using
> Tkinter, but I'm lacking experience (just trying to change that...)

wxPython is a guitoolkit with a good datagrid widget.
First you must install wPython, http://www.wxpython.org/ but once that is
done you could write a small app like this if I am not mistaken


from wxPython.wx import *
from wxPython.grid import *

class MyFrame(wxFrame):
    def __init__(self, parent, title):
        wxFrame.__init__(self, parent, -1, title, size=(350, 200))
        grid=wxGrid(self, -1, size=(150,150))
        grid.CreateGrid(4,4)

app = wxPySimpleApp()
frame = MyFrame(None, "Simple wxPython Grid")
frame.Show(True)
app.MainLoop()

Then you an have a look at the wxGrid widget wich is qiute extensive.

Good luck

Tim





From member27604 at dbforums.com  Thu Apr 17 12:01:18 2003
From: member27604 at dbforums.com (dolephi9080)
Date: Thu, 17 Apr 2003 16:01:18 +0000
Subject: [Que]Do any SGML lib good for Python?
Message-ID: <2778697.1050595278@dbforums.com>

I search  the SGML Lib for python, but it's hard to find. even the
python contain a sgml parser itself. But that's a limite parser.
Does some one know it? pls tell me. Thx

--
Posted via http://dbforums.com



From exarkun at intarweb.us  Wed Apr 16 03:52:54 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 16 Apr 2003 03:52:54 -0400
Subject: Protected Methods and Python
In-Reply-To: <jg4na.146561$S14.33908@news1.central.cox.net>
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <18lb3o2cs1se2$.isoixa5igmj6$.dlg@40tude.net> <12aeh974i1kdh.dfbwdn5e18px.dlg@40tude.net> <mailman.1050275130.5456.python-list@python.org> <jg4na.146561$S14.33908@news1.central.cox.net>
Message-ID: <20030416075253.GA10905@meson.dyndns.org>

On Wed, Apr 16, 2003 at 03:39:59AM +0000, Venkatesh Prasad Ranganath wrote:
> [snip]
> 
> All said, how do you ensure "design by interface" in Python?   If every 
> feature of a class is open to the user then he/she will use them as he/she 
> sees fit.  However, the writer of the class could have coded some methods 
> for ancillary purposes and later decides to do away with them in the next 
> version of the class.  How should the user deal with this situation?  In my 
> opinion, protected and private as in C++ or Java empowers the library 
> developer to expose just the interface that he/she intends the user to see. 
> Nothing more and nothing less.  (It's an issue dealing with abstraction.)  
> As a result, the user can be more care-free while using the library code as 
> he cannot screw up the internals if the library code doesn't do it on it's 
> own.  On the other hand, if all members of a class are visible to the user 
> then the user needs to be more diligent while using the library.  I prefer 
> the first option.

  If it isn't documented, don't use it :)  You have a point though.  Some
libraries don't make it clear what is public and what is an implementation
detail.  What's the worst case, though?  You upgrade, your software breaks,
you downgrade, it works again, and maybe you fire off an unhappy email to
the dev list for the library.

> 
> On the issue of working around bugs in a library via runtime protected 
> attributes, it would be better if the skillful coder fixed the bug ;-) or 
> use another library.  This is just what I would do.

  Good point.  I omitted this possibility in hopes of keeping the point
clear, but it's what I try to do as well :)

  Jp

--
 up 27 days, 3:02, 3 users, load average: 1.45, 1.42, 1.43




From vjb at vuit.com  Tue Apr 22 12:38:09 2003
From: vjb at vuit.com (Vince Buonassisi)
Date: Tue, 22 Apr 2003 09:38:09 -0700
Subject: missing X11.h
Message-ID: <CD5787B6-74E0-11D7-ADF9-000A95682CD2@vuit.com>

hi,

while compiling the source for python 2.2.2 the compilation was unable 
to locate X11/Xlib.h when compiling the file _tkinter.c.  is there a 
package that i need to install in order to get the headers so that i 
can compile X window programs?

thanks,





From brian at sweetapp.com  Tue Apr  1 03:28:03 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Tue, 01 Apr 2003 00:28:03 -0800
Subject: Comments
In-Reply-To: <435d40e0.0304010014.503f9f8f@posting.google.com>
Message-ID: <001401c2f828$9cdb8e50$21795418@dell1700>

> I would like to request that Python support multi-line comments (e.g.
/*
> */ in C).

Just insert a multiline string literal.

Cheers,
Brian





From abpillai at lycos.com  Fri Apr 25 03:10:58 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 25 Apr 2003 00:10:58 -0700
Subject: sys.argv problem in Win 2k
References: <bd993a2f.0304240121.7c795cf0@posting.google.com> <mailman.1051177117.30168.python-list@python.org> <bd993a2f.0304240506.5b206f3d@posting.google.com> <Xns936790B52180Eduncanrcpcouk@127.0.0.1>
Message-ID: <bd993a2f.0304242310.586305df@posting.google.com>

Hi Duncan

 There was a problem with this but I fixed it in the
registry. Now the associations are fine, but still the
cmd line problem remains!

#test+
F:\users\anand\LocalWorkspaces\ResModWS>ftype Python.File
Python.File=D:\Python22\pythonw.exe "%1"

F:\users\anand\LocalWorkspaces\ResModWS>ftype Python.File
Python.File=D:\Python22\python.exe "%1"

F:\users\anand\LocalWorkspaces\ResModWS>assoc .py
.py=Python.File

F:\users\anand\LocalWorkspaces\ResModWS>assoc .pyw
.pyw=Pythonw.File

F:\users\anand\LocalWorkspaces\ResModWS>ftype Pythonw.File
Pythonw.File=D:\Python22\pythonw.exe "%1"
#test-

#example+
#cmdtest.py
import sys

print sys.argv
for x in sys.argv:
    print x

D:\OpensourceApps>cmdtest.py
['D:\\OpensourceApps\\cmdtest.py']
D:\OpensourceApps\cmdtest.py
#example-

 Huh ?  ---

Thanks

ABP
Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in message news:<Xns936790B52180Eduncanrcpcouk at 127.0.0.1>...
> abpillai at lycos.com (Anand B Pillai) wrote in 
> news:bd993a2f.0304240506.5b206f3d at posting.google.com:
> 
> >  As you see the arguments I pass are not being printed!
> > This is the case with all my scripts now. They were working
> > well when I was on a WinNT workstation :-?
> 
> What does the command 'assoc .py' print?
> What does the command 'ftype Python.File' print?
> 
> Running a .py file depends on the file association which the assoc command 
> will show, and if it is Python.File the ftype command will show how the 
> command line is passed.
> 
> They should look something like this:
> 
> D:\temp>ftype Python.File
> Python.File=D:\Python22\python.exe "%1" %*
> 
> D:\temp>assoc .py
> .py=Python.File



From peter.maas at mplusr.de  Thu Apr 10 12:14:00 2003
From: peter.maas at mplusr.de (Peter Maas)
Date: Thu, 10 Apr 2003 18:14:00 +0200
Subject: ASTVisitor
In-Reply-To: <mailman.1049983637.20728.python-list@python.org>
References: <b73pks$t0t$1@swifty.westend.com> <mailman.1049983637.20728.python-list@python.org>
Message-ID: <b7453c$7df$1@swifty.westend.com>

frank at niessink.com wrote:
> I agree the docs aren't very clear. Apparently, you have to restart 
> the traversal after each match. 

Frank,

first, thanks for your help. That means the visit<class> functions have
to handle the traversal. Then it would be sufficient to have one function
that does traversal *and* visit. For example, define a root level visit
function that is called by ASTVisitor.preorder and does a postorder. :)

Mit freundlichen Gruessen,

Peter Maas

-- 
-------------------------------------------------------------------
Peter Maas, M+R Infosysteme, D-52070 Aachen, Hubert-Wienen-Str. 24
Tel +49-241-93878-0 Fax +49-241-93878-20 eMail peter.maas at mplusr.de
-------------------------------------------------------------------




From warlock at eskimo.com  Sat Apr 12 22:44:37 2003
From: warlock at eskimo.com (Jim Richardson)
Date: Sat, 12 Apr 2003 19:44:37 -0700
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com> <ugham-t2f.ln1@grendel.myth> <fi%la.7870$T34.222571@news2.tin.it>
Message-ID: <lhvmm-7c7.ln1@grendel.myth>

On Sat, 12 Apr 2003 21:11:39 GMT,
 Alex Martelli <aleax at aleax.it> wrote:
> Jim Richardson wrote:
> 
>> On Sat, 05 Apr 2003 20:13:45 +0100,
>>  Simon Faulkner <news at titanic.co.uk> wrote:
>>> I have a list of about 5000 numbers in a text file - up to 14 digits
>>> each.
>>> 
>>> I need to check for duplicates.
>>> 
>>> What would people suggest as a good method?
>>> 
>>> Simon
>> 
>> I'd use sort|uniq, but I don't know if that's available for MS type osen.
> 
> You can get implementations of sort and uniq for MS, but a short
> Python script is better IMHO.

yeah, but that's usually the case :) 

> 
> 
>> In python, just stuff them all in a dictionary, any repeats, will be
>> eliminated. But this is rather crude and probably slow. But it would
>> work.
> 
> Anything but slow!  Python dictionaries are quite fast.  But removing
> duplicates is not the same as 'checking for duplicates' -- Simon
> might rather want (e.g.) a list of all numbers that WERE in fact
> duplicate.  A script that plays with a Python dict is still no doubt
> the right solution, but it's hard to write one without more precise
> specifications regarding what is desired.
> 
> 

yeah, I didn't look at the check for part, I just parsed it as get rid
of... <sigh> must need a brain upgrade. 

I don't know how fast/slow the dict would be to tell the truth, it just
doesn't seem that "elegant" and elegance, is often (wrongly I know)
associated with speed.

Having said that, I have found that usually, the simpler the script, and
the closer to the pythonic "metal" it is, the faster it is. For some
reason, the folks who wrote python, are a lot better at programming that
I am :) 

-- 
Jim Richardson         http://www.eskimo.com/~warlock

Linux, because eventually, you grow up enough to be trusted with a fork()



From jxlu at fudan.edu.cn  Sun Apr  6 09:45:28 2003
From: jxlu at fudan.edu.cn (Lu Jian Xiong)
Date: Sun, 06 Apr 2003 21:45:28 +0800
Subject: Help: "undefined reference to `PyArg_ParseTuple'" when execising
 'Extending Python with C or C++ '
Message-ID: <0HCX00EK7CRVDS@mail.fudan.edu.cn>

Hello,
I am a newbie of python.
When I was trying the example read from the following URL,
http://www.python.org/doc/current/ext/simpleExample.html,
it showed the following error message:

[root at localhost user0]# gcc -I /usr/include/python2.2/ -L /usr/lib/python2.2/c
onfig/ -l python2.2 spammodule.c
/tmp/ccIl3dpq.o: In function `spam_system':
/tmp/ccIl3dpq.o(.text+0x16): undefined reference to `PyArg_ParseTuple'
/tmp/ccIl3dpq.o(.text+0x47): undefined reference to `Py_BuildValue'
collect2: ld returned 1 exit status
#
#I tried to find out the problem
[root at localhost user0]# slocate *.a|grep ython
/usr/lib/python2.2/config/libpython2.2.a
[root at localhost user0]# rpm -qa|grep ^python
python-docs-2.2.2-7
python-optik-1.3-2
python-tools-2.2.2-7
python-devel-2.2.2-7
python-2.2.2-7
[root at localhost user0]# cat spammodule.c
#include <Python.h>
static PyObject *
spam_system(self, args)
    PyObject *self;
    PyObject *args;
{
    char *command;
    int sts;

    if (!PyArg_ParseTuple(args, "s", &command))
        return NULL;
    sts = system(command);
    return Py_BuildValue("i", sts);
}
int main()
{
    return (0);
}
[root at localhost user0]#

What is the matter? What should I do?
Can you help me?

Excuse for my poor English.






From cpl.19.ghum at spamgourmet.com  Sat Apr 12 05:54:43 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Sat, 12 Apr 2003 11:54:43 +0200
Subject: module for doing unix "find"-like things?
References: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com>
Message-ID: <Xns935B78BC3EC4Fcpl19ghumspamgourmet@62.153.159.134>

Mark,

also take a look at

http://www.jorendorff.com/articles/python/path/

It's quite helpfull.


Harald



From mads at troest.NEVERMORE.dk  Mon Apr 14 14:05:42 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Mon, 14 Apr 2003 20:05:42 +0200
Subject: Using __repr__ or __str__ for own printable class?
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org> <b7di0n$dd6gj$1@ID-169208.news.dfncis.de>
Message-ID: <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net>

On Mon, 14 Apr 2003 17:57:46 +1200, Greg Ewing (using news.cis.dfn.de) 
wrote:

> If you care about the precise format of the string
> representation you're trying to get for a container,
> you shouldn't be using either str() or repr() --
> you should be writing your own function that does
> exactly what you want.

But then one wouldn't be able to do a standard "print" of the class, would 
one? That was what I was after in the first place, and it seemed like the 
way to do it is to override the __str__ method. I would like to be able to 
do this ...
	print someInstanceOfMyClass

... and then determine what should be printed. This works with __str__ but 
I was/am afraid if this overriding of __str__ will conflict with other uses 
of __str__ and/or __repr__.

Regards,
   /\/\\ads Orbesen Troest



From harri.pasanen at trema.com  Wed Apr  9 13:14:39 2003
From: harri.pasanen at trema.com (Harri Pasanen)
Date: 9 Apr 2003 10:14:39 -0700
Subject: python 2.3a2 Win32 expected failures in regression tests?
Message-ID: <e2d925a8.0304090914.2fd8c154@posting.google.com>

I've built python 2.3a2 myself on Win32, using Visual C++ 6.0.

When running the regression tests, the following 3 fail:

test_tarfile, test_timeout, test_zipimport

(test_zip passes)
All in all, I have 201 tests that pass, but those 3 fail.

I wonder if this is normal state of the affairs for Python 2.3a2 on
win32, or something in my build that does not click.

-Harri



From yaipa at yahoo.com  Sat Apr  5 16:07:20 2003
From: yaipa at yahoo.com (yaipa h.)
Date: 5 Apr 2003 13:07:20 -0800
Subject: Anyone know of a webmail(yahoo) <--> localhost mail sync app in Python
Message-ID: <6e07b825.0304051307.3b301dd7@posting.google.com>

All,

I am looking for an application in Python that will allow me 
to synchronize my free yahoo webmail account with a local mail 
reader for off-line mail access. Using Outlook at this point is
a don't care.

I have found some stuff in perl/php, but nothing in python.  
Some apps that look like they will work are squirrelmail, 
fetchyahoo.pl and Open WebMail.  Ultimately, I will need to 
support this application company wide and would like it not to 
be in perl/php.

-Alan Haffner



From aleax at aleax.it  Sat Apr 12 05:58:00 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 09:58:00 GMT
Subject: mirroring text window to file
References: <3E9783E3.6010108@videotron.ca>
Message-ID: <IqRla.4499$T34.110153@news2.tin.it>

Marcel Achim wrote:

> I was wondering if there's an existing solution other than writing
> to both a file and an output text window (Tk text widget), as in :
> 
>    text_window.insert('end', output)
>    log_file.write(output)

Depends to what "problem" you want an "existing solution" for.  For
the purpose of having some string be appended to both a file and a
text window, you do need these two calls to happen _some_where, but
you can arrange for that to happen in several different ways, mostly
by exploiting the concept of "filelike objects" -- it all depends
on what, exactly, you're trying to accomplish.  E.g.:

class LogAndText:
    def __init__(self, thewindow, thefile):
        self.thewindow = thewindow
        self.thefile = thefile
    def write(self, stuff):
        self.thewindow.insert('end', stuff)
        self.thefile.write(stuff)

now you can instantiate

logansho = LogAndText(text_window, log_file)

and just usr logansho as if it was a file open for output (e.g.
you can print>>logansho,whatever -- you can set logansho as
sys.stdout so that all normal print statements will use it -- &c).


Alex




From winhelp at columbus.rr.com  Sat Apr 19 00:53:55 2003
From: winhelp at columbus.rr.com (berg_)
Date: Sat, 19 Apr 2003 04:53:55 GMT
Subject: variable in Python
References: <b7lvbh$an3$1@news-reader13.wanadoo.fr> <mailman.1050586367.28668.python-list@python.org>
Message-ID: <DD4oa.113758$0X.26571104@twister.columbus.rr.com>

> object it refers to; additionally, there is no concept of a "null"
> reference.

Python does have the concept of null, it is represented by None.

cursor = None

shain

>
>   Hope this helps,
>
>   Jp
>
> --
> It is practically impossible to teach good programming style to
> students that have had prior exposure to BASIC: as potential
> programmers they are mentally mutilated beyond hope of
> regeneration.        -- Dijkstra
> --
>  up 28 days, 9:02, 5 users, load average: 0.04, 0.03, 0.00
>





From cejlai at lafn.org  Thu Apr 17 02:52:01 2003
From: cejlai at lafn.org (Jason Lai)
Date: Wed, 16 Apr 2003 23:52:01 -0700
Subject: Subclassing a C type
In-Reply-To: <b7jkni$1v7d$1@zook.lafn.org>
References: <b7jkni$1v7d$1@zook.lafn.org>
Message-ID: <b7liuk$2l7t$1@zook.lafn.org>

Actually, nevermind, I found another part of the program that was 
tampering with the inheritance. Works well enough for me now.

Jason Lai wrote:
> Hello,
> 
> I wrote a Python type in C, Actor, and created a subclass Tester 
> (new-style classes). Actor has a getattro function that gets information 
> from elsewhere in the C program.
> 
> Apparently, the Tester class inherits tp_getattro from Actor. What I'd 
> like it to do is have it use the regular lookup so that it checks its 
> own dict before deferring to its parent classes. Without having to 
> define a __getattr__ function on Tester. Maybe I could change the 
> __getattr__ on every tp_new, but that seems a bit messy.
> 
> I thought about having Actor_getattro look up the value in 
> Tester-instances' dictionaries, but I don't think it would work if there 
> were more subclasses between Actor and Tester in the hierarchy.
> 
> And I do need to intercept the getattro calls, so I can't just use the 
> standard object one and define tp_members/tp_methods.
> 
> Am I doing something wrong?
> 
> Thanks in advance for any advice you can offer,
> 
>  - Jason Lai
> 




From vlindberg at verio.net  Wed Apr 30 16:47:26 2003
From: vlindberg at verio.net (VanL)
Date: Wed, 30 Apr 2003 20:47:26 GMT
Subject: Web graph producing library
Message-ID: <yDWra.11788$8Z5.69611@sea-read.news.verio.net>

Hello,

I have googled around for this, and I can't find it.

I have a web app that produces some reports.  I would like to produce 
some reports in graph form, as well.  I have found a couple interfaces 
to various image libraries, but none that are particularly specialized 
for graphs.

For example I would love to have something like this:

import webgraph

data = {'adam': (30, 'yellow'),
         'jeff': (40, 'green'),
         'rob': (10, 'red'),
         'bill': (2, 'purple'),
         'bob': (15, 'black'),
         'uneaten': (3, 'white')
        }
title = "Apples eaten by team member"

piechart = webgraph.PieChart(data, title=title)

piechart.dump() # Dumps pie graph in png (or other) format

bargraph = webgraph.LineChart(data, title=title)

bargraph.dump()

Is there anything out there like this?  I know there is for perl and 
php, but I haven't found a python version.

VanL




From duncan at NOSPAMrcp.co.uk  Wed Apr 30 04:13:54 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 30 Apr 2003 08:13:54 +0000 (UTC)
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com> <b8lpof$q0f$1@pegasus.csx.cam.ac.uk> <Xns936C8AB67B739duncanrcpcouk@127.0.0.1> <vatlfeb16jlled@corp.supernews.com>
Message-ID: <Xns936D5DA488705duncanrcpcouk@127.0.0.1>

claird at lairds.com (Cameron Laird) wrote in
news:vatlfeb16jlled at corp.supernews.com: 

> In article <Xns936C8AB67B739duncanrcpcouk at 127.0.0.1>,
> Duncan Booth  <duncan at rcp.co.uk> wrote:
>                .
>                .
>                .
>>The easiest solution is to leave the program in Python and just call
>>the relevant functions from C. Possibly the easiest way to do this
>>would be to use Pyrex.
>                .
>           [persuasive illustration]
>                .
>                .
> ctypes <URL: http://starship.python.net/crew/theller/ctypes.html >
> should be even easier to use.

Does ctypes support embedding? If so, it isn't obvious from that webpage.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From mcfletch at rogers.com  Fri Apr 18 19:26:28 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Fri, 18 Apr 2003 19:26:28 -0400
Subject: hexadecimal calculations
In-Reply-To: <b7q0uv$hco$1@aphrodite.grec.isp.9tel.net>
References: <b7q0uv$hco$1@aphrodite.grec.isp.9tel.net>
Message-ID: <3EA089A4.8070103@rogers.com>

Did you try it at the command line?  Hexidecimal numbers are prefixed 
with 0x, XOR is spelled ^, and you can see the hex representation of a 
number with the built-in hex function:

    hex(0xA0 ^ 0x79)

I'll leave the other as an exercise for the reader.
Mike

Andros wrote:

>Hello,
>how to make these hexadecimal calculations in python?
>A0 XOR 79 = D9
>D9 - A8   = 31
>
>Thanks ;)
>
>
>
>  
>

-- 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From martin at v.loewis.de  Fri Apr 25 06:11:19 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 12:11:19 +0200
Subject: GUI libs
In-Reply-To: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
Message-ID: <b8b1k3$t5f$00$1@news.t-online.com>

Alessio Pace wrote:

> Hi, I'd like to make a GUI for a program written in python, meant to run on
> mac, win and linux: which is the preferable solution? I saw around tkinter,
> wxpython and pyqt....

I recommend Tkinter. It comes with the standard Python installation, and 
it works quite well atleast on Windows and X11 (I've never used it on 
the Mac).

Regards,
Martin




From martin at v.loewis.de  Sat Apr 19 05:40:20 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 19 Apr 2003 11:40:20 +0200
Subject: error in string module documentation?
References: <mailman.1050740611.30741.python-list@python.org>
Message-ID: <m33ckeakqj.fsf@mira.informatik.hu-berlin.de>

Michael Anckaert <xantor at linux.be> writes:

> Is there an error in the documentation or is it my error?

It is your error: you are either looking at the wrong version of the
documentation, or are using the wrong version of Python.

Regards,
Martin




From ta-meyer at ihug.co.nz  Wed Apr 30 01:57:11 2003
From: ta-meyer at ihug.co.nz (Tony Meyer)
Date: Wed, 30 Apr 2003 17:57:11 +1200
Subject: Temporary Files
In-Reply-To: <1ED4ECF91CDED24C8D012BCF2B034F13015F252F@its-xchg4.massey.ac.nz>
Message-ID: <1ED4ECF91CDED24C8D012BCF2B034F13C8CAE8@its-xchg4.massey.ac.nz>

[me]
> Is there perhaps a different method of copying?  One that 
> doesn't take filenames, but file objects?

Ack.  Yes, there is shutil.copyfileobj().

*sigh*  Rephrased in light of this rather belated discovery; is this
what I should be doing?

>>> import shutil
>>> import tempfile
>>> out = tempfile.TemporaryFile("w")
>>> out.write(data)
>>> out.flush()
>>> # shutil.copyfile(out.name, permanent_name)  # not this!
>>> permanent_file = file(permanent_name)
>>> shutil.copyfileobj(out, permanent_file)
>>> out.close()
>>> permanent_file.close()

TIA.

=Tony Meyer





From avner at skilldesign.com  Mon Apr 14 15:37:35 2003
From: avner at skilldesign.com (Avner Ben)
Date: Mon, 14 Apr 2003 21:37:35 +0200
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it>
Message-ID: <3e9b0e03$1@news.012.net.il>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:Y%Dma.13754$LB6.370780 at news1.tin.it...
> Avner Ben wrote:
>
> > Question:
> >
> >     Does standard Python have case-insensitive string comparison and
> >     search?
>
> No.  You normally accomplish them either with regular expressions (which
> CAN be compiled into case-insensitive mode) or by coercing strings to
> one fixed case (e.g. with the .lower method of string objects).
>

    I have tried the lower them all approach, but at the rate of processing
I need, it has turned out to reduce application speed by a factor of four.
What I need is an efficient built-in way of doing this.

    I will try the regular expression approach and see if it meets my
performance criteria.

    If that does not help, I guess I will have to delegate the job to a C
routine.

        Avner.





From aleaxit at yahoo.com  Thu Apr 10 09:13:11 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: Thu, 10 Apr 2003 15:13:11 +0200
Subject: looking for "optimal weighting" algorithm
Message-ID: <b73ql8011te@enews4.newsguy.com>

I _know_ there's a decent algorithm to solve the following problem,
but I don't recall its name and thus can't rapidly google for the 
details... can somebody help?


The problem: I need to design a "decision criterion" to classify
"observations".  For each observation I measure the values of a
number N of features, x1, x2, ... xN; the desired design criterion
is a set of weights w1, w2, ... wN such that for any observation I 
will then just compute a weighted sum
  S = w1*x1 + w2*x2 + ... + wN*xN
and classify the observation as Black if S<=1, White if S>1.  To
train my classifier I have a large corpus of observations already
made and annotated with "ground-truth" data about whether each
given observation should have been classified as B or W, and an
error-cost value for each kind of classification (Ebw is the cost
of erroneously classifying a feature as W when it should be B,
Ewb is that of classifying it as B when it should be W).  So,
what's the algorithm to estimate the weights given the corpus of
observation and ground-truth data, and the error-costs?

Responses welcome either here or directly to my mailbox -- if I
do get responses to my mailbox, I'll eventually summarize here --
as a bonus, I promise to make available a Python implementation of 
whatever algorithm I end up with, if one's not already around...;-).


TIA,

Alex




From brian at sweetapp.com  Mon Apr 14 13:53:20 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Mon, 14 Apr 2003 10:53:20 -0700
Subject: Reading/writing files
In-Reply-To: <pan.2003.04.14.17.32.01.209427@soul.net>
Message-ID: <00ec01c302ae$bc7b7b40$21795418@dell1700>

> for line in a file
> 	print line
> print done

Close:

# file is a really bad name for a variable!
for line in file:
    print line
print 'done'

> and for writing, I've tried doing file.write(blah) (blah is a list),
but
> it wont write.

How do you want the list represented in the file? One element per line?

Cheers,
Brian





From jack at performancedrivers.com  Sat Apr 12 12:48:56 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Sat, 12 Apr 2003 12:48:56 -0400
Subject: PEP 754: IEEE 754 Floating Point Special Values
In-Reply-To: <D7A3CFD7825BD6119B880002A58F06C202F2C7F9@groexmb02.pfizer.com>; from gregory_r_warnes@groton.pfizer.com on Sat, Apr 12, 2003 at 11:46:08AM -0400
References: <D7A3CFD7825BD6119B880002A58F06C202F2C7F9@groexmb02.pfizer.com>
Message-ID: <20030412124856.G1039@localhost.localdomain>

On Sat, Apr 12, 2003 at 11:46:08AM -0400, Warnes, Gregory R wrote:
> Copyright
> =========
> This document has been placed in the public domain.

...

> LEGAL NOTICE
> Unless expressly stated otherwise, this message is confidential and may be privileged. It is intended for the addressee(s) only. Access to this E-mail by anyone else is unauthorized. If you are not an addressee, any disclosure or copying of the contents of this E-mail or any action taken (or not taken) in reliance on it is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately.


It's funny, laugh.




From Anish.Mehta at enst-bretagne.fr  Tue Apr 29 05:44:26 2003
From: Anish.Mehta at enst-bretagne.fr (Mehta, Anish)
Date: Tue, 29 Apr 2003 11:44:26 +0200
Subject: ttList-module problem
Message-ID: <3EAE497A.1080103@antares.enst-bretagne.fr>

Hello !

I am having some problem in understanding this code. Firstly i am not 
finding TTFont which is being imported with the command

from fontTools.ttLib import TTFont

I have searched in the ttLib directory but i am unable to find TTFont so 
that i can read about it. As far as i am getting is that this module is 
taking input and then some information about      tags( ? ) and then 
printing out the format in the form of checksum, length and offset. 
Please elaborate this a bit more.

def ttList(input, output, options):
        ttf = TTFont(input)
        reader = ttf.reader
        tags = reader.keys()
        tags.sort()
        print 'Listing table info for "%s":' % input
        format = "    %4s  %10s  %7s  %7s"
        print format % ("tag ", "  checksum", " length", " offset")
        print format % ("----", "----------", "-------", "-------")
        for tag in tags:
                entry = reader.tables[tag]
                checksum = "0x" + hex(entry.checkSum)[2:].zfill(8)
                print format % (tag, checksum, entry.length, entry.offset)
        print
        ttf.close()


I'll be waiting for your response.

Thanks in advance.

Regards,

Anish MEHTA





From mc at mclaveau.com  Sat Apr 19 12:02:39 2003
From: mc at mclaveau.com (Michel Claveau)
Date: Sat, 19 Apr 2003 18:02:39 +0200
Subject: Python COM Object
References: <78088784.0304181010.91e3641@posting.google.com>
Message-ID: <b7rrvo$6up$1@news-reader13.wanadoo.fr>

Hi !

Python COM is not Active-X. I don't know Jscript's syntax,.but,
with.Paradox, I use a.Python-COM-server, all days.

@-salutations
--
Michel Claveau
site : http://mclaveau.com
news://news.zoo-logique.org/programmation.Paradox





From FBatista at uniFON.com.ar  Tue Apr  1 14:51:02 2003
From: FBatista at uniFON.com.ar (Batista, Facundo)
Date: Tue, 1 Apr 2003 16:51:02 -0300 
Subject: PEP 313 - Roman numeral literals
Message-ID: <A128D751272CD411BC9200508BC2194D0188535F@escpl.tcp.com.ar>

#- 
#- In addition, any calculations involving the Roman calendar 
#- (cf http://www.clubs.psu.edu/aegsa/rome/romec.html) would 
#- clearly benefit
#- from such a module.
#- 
#- My kalend.py module, currently running at 87,543 lines of code, could
#- be shrunk substantially if it could do its calculations using the
#- new roman type.  *Thank you*, Mike, for starting the process 
#- of getting 
#- this valuable feature into Python.

I think Mike is proposing a built-in function. If this is the case, I'll
like these two:

	literal IntToRoman(int)
	int RomanToInt(literal)

But if we're thinking in a module, I'll like to have an object called Roman
(or so) where I could make something like this:

	myRoman1 = Roman()
	myRoman1.fromInt(15)

	myRoman2 = Roman()
	myRoman2.fromString("X")

	myRoman3 = Roman()
	myRoman3 = myRoman1 + myRoman2

	print myRoman3.toInt()
		25

	print myRoman3.toString()
		"XXV"

A nice function in the module could be too:

	Roman.verify(literal)
		returning True if the string is a "well written" roman
number (after, if we call Roman.fromString is allright, othwerwise it will
raise an error)


I salute you.

.elm.	




From win at pfaffenhofen.de  Mon Apr 28 03:41:31 2003
From: win at pfaffenhofen.de (Bob Wheldon)
Date: Mon, 28 Apr 2003 09:41:31 +0200
Subject: Mega-Job-Search_MachineMega Job Search Machine
Message-ID: <3EACDB2B.9010509@pfaffenhofen.de>

Is anyone interested in developing a Mega Job Search Machine with me for 
non-US positions?

-- 
Regards,

Bob Wheldon
Bob.Wheldon at gmx.de
Tel.      +49 8444/7309
Mobile +49 179 5061083







From alloydflanagan at attbi.com  Mon Apr 21 09:24:42 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 21 Apr 2003 06:24:42 -0700
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <a8b7f07a.0304161158.29c0ef6b@posting.google.com> <b7kk53$fe7$1@slb2.atl.mindspring.net>
Message-ID: <a8b7f07a.0304210524.3d26aff8@posting.google.com>

"Andrew Dalke" <adalke at mindspring.com> wrote in message news:<b7kk53$fe7$1 at slb2.atl.mindspring.net>...
> A. Lloyd Flanagan
> > This mathematical analysis of an algorithm's behavior (and best-case,
> > average, and worst-case performance) can be trivial.  For example, x +
> > 1 is O(1).
> 
> Or not so trivial.  If x can be an arbitrary size integer, represented
> in memory as a chunk of bits, the value 'n' for the number of bits,
> and a sign flag, then addition could be O(n) == O(log(x)).  Eg,
> suppose    x =  1111111111111111111111111111111111111111
> add 1 to get = 10000000000000000000000000000000000000000
> so 49 bits were affected.
> 
> But for fixed size integers, the above is true.
> 
>                     Andrew
>                     dalke at dalkescientific.com


I stand corrected.  See, this algorithm-analysis stuff can tricky!



From theller at python.net  Mon Apr  7 07:41:42 2003
From: theller at python.net (Thomas Heller)
Date: 07 Apr 2003 13:41:42 +0200
Subject: Calling C libraries from a Python script
References: <e81e0f3d.0304050639.37b59655@posting.google.com> <r88gc37v.fsf@python.net> <ddc19db7.0304070250.7ecaf65e@posting.google.com>
Message-ID: <3ckuttzd.fsf@python.net>

a-steinhoff at web.de (Armin Steinhoff) writes:

> Thomas Heller <theller at python.net> wrote in message news:<r88gc37v.fsf at python.net>...
> > joshua at goodish.org (Joshua Goodlett) writes:
> > 
> > > I'm trying to find examples of calling C library functions from within
> > > a Python script.  Any advice or pointers on where to look for examples
> > > would be appreciated.
> > 
> > If the C library is a dll, you could use ctypes.
> 
> Such DLLs must also be conform to the  _cdecl call convention ... isn't it ?

No. ctypes supports __stdcall as well, and at least one user
has wrapped a Delphi dll (__stdcall, but with reverse argument order,
IIRC).

For __stdcall, use the windll object to access the dll, for __cdecl
use cdll.

Thomas



From aleax at aleax.it  Sat Apr 26 06:22:42 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 10:22:42 GMT
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com>
Message-ID: <S5tqa.9420$K35.308092@news2.tin.it>

Anna wrote:

> My daughter just asked me about this "python stuph". So I pulled up
> livewires.org.uk and showed her some basics.
> 
> She just coded her first while loop and for loop... she thought it was
> kewl! In fact, we had to kick her off the computer and send her to bed cuz
> she's got so much to do tomorrow...
> 
> FWIW, she's 12...

Wow, cool!  I think livewires was just the right choice, too.

I consider it a key root of Python's amazing strength, that a bright
12-years-old CAN start grasping it so fast... while at the same time
Python has ample power for experienced programmers to use it for such
major projects as, e.g., mailman, Zope, Chandler, and the like!!!  [Of
course, it DOES take a somewhat-special 12-y-o, but I'm anything but
surprised that your daughter should be!-)]

Now, knowing that your daughter _is_ about to turn into a teenager (and
we all know how teenagers tend to rebel against parental authority), the
best approach to stimulate her interest might well be to forbid her to
dabble in it further, it being dangerous, addictive, adults-only stuff;-)


Alex




From missive at frontiernet.net  Tue Apr  8 11:25:42 2003
From: missive at frontiernet.net (Lee Harr)
Date: Tue, 08 Apr 2003 15:25:42 GMT
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com> <b6soav$7p2$1@panix1.panix.com> <Qdpka.2889$Zu2.2627@news01.roc.ny.frontiernet.net> <a8b7f07a.0304080603.473a8d90@posting.google.com>
Message-ID: <WRBka.134$jS3.16@news02.roc.ny.frontiernet.net>

> If all your classes only have one parent, it doesn't make a
> difference.  If your classes a) use multiple inheritance, b) several
> of them define a method with the same name, and c) all of them call
> super() to execute that method, then python does some really neat
> stuff to ensure that all the methods of that name get called, and that
> it occurs in the "correct" order.
> 
> So super() is preferable if a) you're redefining a method but want to
> call the base class method(s), and b) your class is part of a multiple
> inheritance graph OR may be used in a multiple inheritance graph.
> 
> Disclaimer:  I'm new to this stuff myself, so if somebody's got a
> better explanation, please speak up!


I guess I do not get it.


>>> class A(object):
...   def f(self):
...     print 'Af'
... 
>>> class B(object):
...   def f(self):
...     print 'Bf'
... 
>>> class C(A, B):
...   def f(self):
...     super(C, self).f()
... 
>>> c=C()
>>> c.f()
Af


Should this be showing
Af
Bf
?

How would I even get that output?  I know I can do:

class C(A, B):
  def f(self):
    A.f(self)
    B.f(self)


but I can't seem to get that out of super()




From tim.one at comcast.net  Mon Apr 21 23:17:22 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 21 Apr 2003 23:17:22 -0400
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <g8n5avgooim13l90bv5t9gu6ilg1oar101@4ax.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCCEOFEDAB.tim.one@comcast.net>

[Courageous]
> ...
> In environments that I work on where continuations aren't available,
> but the same features must exist, the hand-rolled equivalent is a set
> of "parked" threads whereby some scheduler is written to selectively
> activate them. That's how they work, conceptually. A continuations is
> a marker for a "place one can return to whenever one might want". No
> big deal, really. Doesn't have to be complicated.

Except that a collection of threads isn't as powerful as continuations.  As
soon as "the program counter" changes in one of those threads, the
continuation it's standing in for has been changed.  A true continuation can
be invoked any number of times, starting at exactly the same state each
time.  You'd have to augment those threads with some platform-specific
grunge to clone the thread control state top-to-bottom to repair that
weakness.

> 'Course the implementation details may be a bit harder to explain,
> depending. :-)

That's the rub.  Continuations are simple from a very abstract view, in
function spaces.  Implementing them efficiently on real hardware with finite
resources is a challenge.  Alas, people tend to overlook that implementing C
call semantics efficiently on real hardware with finite resources is a
challenge too -- and C was designed to be easy to implement, while
continuations were designed to scratch a theoretical itch.





From johnroth at ameritech.net  Sun Apr  6 08:58:32 2003
From: johnroth at ameritech.net (John Roth)
Date: Sun, 6 Apr 2003 08:58:32 -0400
Subject: properties and get/set methods
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org>
Message-ID: <v908s7echa191b@news.supernews.com>

"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
news:mailman.1049604615.22077.python-list at python.org...
> Dan Bishop wrote:
>
> >Timothy Grant <tjg at craigelachie.org> wrote in message
news:<mailman.1049581810.25706.python-list at python.org>...
> >
> >
> ...
>
> >>Is it possible to determine which attribute was used to as the
trigger for the
> >>call to setvar() or getvar()? so that set_appropriate_attribute() or
> >>appropriate_attribute() can set the correct attribute or get the
correct
> >>attribute?
> >>
> >>
> >
> >Why not just use two different getvar and setvar functions?
> >
>
> When you're doing generic "framework" level programming, it's just not
> acceptable to have to define 2 (really 3) new functions for every
> property.  Compare:
>
>     def setMyProperty( self, value ):
>        do_the_generic_thing( self, 'myProperty', value)
>        self.__dict__["myProperty"] = value
>     def getMyProperty( self ):
>        do_the_generic_thing( self, 'myProperty', value)
>        return self.__dict__["myProperty"]
>     def delMyProperty( self ):
>        del self.__dict__[ "myProperty" ]
>     myProperty = property( getMyProperty, setMyProperty,
delMyProperty,
> """Some documentation""")
>
> versus:
>
>     myProperty = StringProperty(
>        "myProperty", """Some documentation""",
>        defaultValue = "this value",
>     )
>
> In these cases, you're using properties to simplify the modelling of
> domain objects, or to provide some *generic* functionality over and
> above a regular value get/set.  You don't want to have to specify that
> functionality for every property, you just want to say "this is one of
> _those_ properties".  The code bloat (not to mention mindless
> duplication of code) when you get up to 15 or 20 properties for an
> object would be ridiculous.

Well, I suppose since this is Sunday, it's appropriate to bring
religion into it.

John Roth
>
>
> Enjoy,
> Mike






From mcherm at mcherm.com  Thu Apr 10 16:22:44 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Thu, 10 Apr 2003 13:22:44 -0700
Subject: (no subject)
Message-ID: <1050006164.3e95d294cd18f@mcherm.com>

Tim Peters writes:
> Alas, list.insert() existed before sequence indices were generalized to give
> a "count from the right end" meaning to negative index values.  When the
> generalization happened, it appears that list.insert() was just overlooked.
> 
> I'd like to change this.  If I did, how loudly would people scream?

I wouldn't mind, but I'm not your test case... the amount of Python
code I have lying around to maintain is small enough that I easily
checked to see that this wouldn't break anything.

Would the correct way to do this be to enable warnings for one release
for any code that passed a negative first argument, then make the fix
in the next release? That's annoyingly slow for those who like the
feature (but they lived without it this long, they can continue to do
so), and it's quite polite to those who live in fear of code breakage.

-- Michael Chermside





From andy at wild-flower.co.uk  Fri Apr 18 07:30:08 2003
From: andy at wild-flower.co.uk (Andy Jewell)
Date: Fri, 18 Apr 2003 12:30:08 +0100
Subject: A new and very robust method for doing file locking over NFS?
In-Reply-To: <lcu1cw525a.fsf@gaffa.mit.edu>
References: <lcu1cw525a.fsf@gaffa.mit.edu>
Message-ID: <200304181230.08478.andy@wild-flower.co.uk>

On Thursday 17 Apr 2003 8:56 pm, Douglas Alan wrote:
> I'd like to do file locking over NFS without using lockd.  The reason
> I want to avoid using lockd is because many lockd implementations are
> too buggy.
>
> It is fairly easy to avoid using lockd -- just avoid using lockf() to
> lock a file.  Instead of using lockf(), lock a file by creating a lock
> file that you open with the O_CREAT | O_EXCL flags.  To unlock the
> file, you merely unlink the lock file.  This method is fairly
> reliable, except that there is a small chance with every file lock or
> unlock that something will go wrong.
>
> The two failure symptoms, as I understand things, are as follows:
>
> (1) The lock file might be created without the client realizing that
> it has been created if the file creation acknowledgement is lost due
> to severe network problems.  The file being locked would then remain
> locked forever (until someone manually deletes the lock) because no
> process would take responsibility for having locked the file.  This
> failure symptom is relatively benign for my purposes and if needed it
> can be fixed via the approach described in the Red Hat man page for
> the open() system call.
>
> (2) When a process goes to remove its file lock, the acknowledgement
> for the unlink() could be lost.  If this happens, then the NFS driver
> on the client could accidentally unlink a lock file created by another
> process when it retries the unlink() request.  This failure symptom is
> pretty bad for my purposes, since it could cause a structured file to
> become corrupt.
>
> I have an idea for a slightly different way of doing file locking that
> I think solves problem #2 (and also solves problem #1).  What if,
> instead of using a lock file to lock a file, we rename the file to
> something like "filename.locked.hostname.pid"?  If the rename()
> acknowledgement gets lost, the client will see the rename() system
> call as having failed due to the file not existing.  But in this case
> it can then check for the existence of "filename.locked.hostname.pid".
> If this file exists, then the process knows that the rename() system
> call didn't actually fail--the acknowledgement just got lost.  Later,
> when the process goes to unlock the file, it will rename the file back
> to "filename".  Again, if the rename system call appears to fail, the
> process can check for the existance of "filename.locked.hostname.pid".
> If the file no longer exists, then it knows the rename call really did
> succeed, and again the acknowledgement just got lost.
>
> How does this sound?  Is this close to foolproof, or am missing
> something?
>
> I'm not much of an NFS expert, so I am a bit worried that there are
> details of NFS client-side caching that I don't understand that would
> prevent this scheme from working without some modification.
>
> |>oug

It depends on how other processes are playing with your files; what if another 
program is just scanning directories and opening files at random?  If this 
program doesn't know your locking scheme, it'll just walk all over your 
process...

However, I have used a similar method to this (on windows) when I was playing 
with a distributed file migration tool I wrote.  I had a control directory 
which contained a queue of command-files waiting to be processed, and each 
participating process had a directory named <hostname>.<pid>, and a log 
directory.  When a process 'claimed' a command-file, it would try to move it 
to it's own subdirectory; if that failed, it would just assume that another 
process got there first and then pick another one.  Once each process had 
completed the specified work, it would delete the command-file and write a 
log file, named <hostname>.<pid>.<command-file-name> to the log directory 
specifying how everything had gone. This worked fine with about 6 'helpers', 
each running two or three processes.  The 'engine' ran fine - with no race 
conditions evident and no file corruptions or collisions.  It was a bit of a 
bar-steward to *STOP* (just as any multi-process program), as busy 'remote 
threads' would have to complete their work before they would see the 'all 
stop' flag.  The most difficult thing, however, was recovering from a fatal 
error, (such as when I hit Python's recursion limit), where the supervisor 
process died, but the workers continued to try to process, or when I'd made 
some command error or such and the workers all 'sulked'.  It was fun doing it 
though, in the end, the performance boost wasn't enough to warrant all the 
hassle, so I reverted back to the linear version :-)

Anyway, the point I'm making is that, if you have complete control over the 
location of the files you are trying to lock, and know that no rogue 
processes will interfere, you should be ok.   Remember that other programs 
will not honour your locks unless you modify them to do so (with windows that 
may be difficult...).  I would suggest moving the file to another directory, 
though, rather than simply renaming it, as it's easier to see what's going 
on.  It's also a good idea to disambiguate your pid's by adding the hostname 
to them - pid's are only guaranteed to be unique within a host, so two remote 
processes on different machines could possibly (read: will eventually) end up 
with the same pid .

Windows file locking is just as unpredictable as NFS, particularly when you 
are mixing different versions on the same network.

Using this renaming scheme does give you one particular edge - portability, as 
this should work equally well on windoze, mac and [u|li]n[i|u]ix.

Good luck,

-andyj




From oktaysafak at ixir.com  Tue Apr  8 02:59:48 2003
From: oktaysafak at ixir.com (Oktay Safak)
Date: Tue, 8 Apr 2003 09:59:48 +0300
Subject: unicode bug in turkish characters?
References: <mailman.1049401867.14209.python-list@python.org> <b6l685$14j$05$1@news.t-online.com>
Message-ID: <000d01c2fd9c$75aea320$0200a8c0@b>

Martin, thank you very much for all your help.

Regards,

Oktay


> Ah, I see. Windows. My example was from Linux, where tr_TR is a
> valid locale name.
>
> Python delegates the toupper call to the operating system. It does not,
> in itself, include a database of lower/upper case conversions for all
> locales.
>
> So there is nothing we can do about that; ask Microsoft.
>
> > Locale defaults as determined by getdefaultlocale():
> > ------------------------------------------------------------------------
> > Language:  tr_TR
> > Encoding:  cp1254
>
> Please understand that getdefaultlocale does *not* determine
> the locale used by the program. Instead, it determines (sometimes)
> the locale that Python would use if you had invoked setlocale
> with "", i.e. asking for the user's preference (sometimes,
> getdefaultlocale just doesn't work).
>
> Regards,
> Martin





From peter at engcorp.com  Sun Apr 20 21:47:26 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 20 Apr 2003 21:47:26 -0400
Subject: Mail and text extraction
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com>
Message-ID: <3EA34DAE.EC67C1E3@engcorp.com>

Dino Levy wrote:
> 
> And all you want to do is to connect to your POP3 server, and download
> the section called Raises ( for example, two or three lines after the
> title ), so all you want to do is to download those few lines containing your
> text... ( ONLY those lines, not the whole e-mail )

Why not download the whole message, then scan it after downloading?
You provide no good reason, so far, for why you wouldn't just do
this much simpler step.  Keep the design simple, modular, clean,
and you can do most of what you want with simple, modular, existing
solutions in standard modules.  Why wouldn't you do that?

-Peter



From bdesth at removethis.free.fr  Sat Apr  5 14:18:50 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Sat, 05 Apr 2003 21:18:50 +0200
Subject: Why use a GUI framework?
References: <1g5u8vc1k8t97jg8vsm2gi81sfnu7cm1hf@4ax.com>
Message-ID: <3E8F2C1A.5080008@removethis.free.fr>

Jane Doe wrote:
> Hi,
> 
> 	I know a bit more about the multiple options available to
> build a GUI-based application (Windows, in our case), but I haven't
> found information on the point of using another tool since the
> win32all extension is available to use the native Win32 widgets. 
> 
> Do wxWindows et al. just make it easier, or is it a must-use for
> people who wish to write OS-independant Python applications?

Both, I think...
The fact is that win32 API based GUI won't run on anything else.

Laotseu




From skywalkerpackage at hotmail.com  Wed Apr  9 05:55:00 2003
From: skywalkerpackage at hotmail.com (Johannes Eble)
Date: Wed, 09 Apr 2003 09:55:00 GMT
Subject: Visio like library for Python?
Message-ID: <3e93e50f.676072541@news.online.de>

Hello community,

is there a library available for Python that behaves similar to MS
Visio? For example, I would like to create vector graphic objects and
connect the with a line. If I move, flip or rotate a connected object,
the line should follow the connection.

I would like to draw UML diagrams.

There is an open-source application called dia
(http://www.lysator.liu.se/~alla/dia/) which comes near to what I
want.

But the program is in c and it is tightly coupled with GTK+. It seems
to be an impossible mission to control dia from another program, at
least in Windows.

If something like this isn't available for Python, a simple vector
graphic library could be useful as well. The file format should be XML
based (like the dia format).


Any pointers are appreciated.


Regards


Johannes








From donn at drizzle.com  Sun Apr  6 14:40:36 2003
From: donn at drizzle.com (Donn Cave)
Date: Sun, 06 Apr 2003 18:40:36 -0000
Subject: curses: xterm vs. linux terminal
References: <3E9065FE.1090108@sympatico.ca>
Message-ID: <1049654435.583332@yasure>

Quoth Brian Alexander <brian094 at sympatico.ca>:
| Why is it that the curses special (alt) characters do not display on xterms?

Because they aren't in the font.  Look for a font that has the special
line drawing characters.

	Donn Cave, donn at drizzle.com



From aussie2010 at yahoo.com  Wed Apr  2 13:52:06 2003
From: aussie2010 at yahoo.com (Chris)
Date: 2 Apr 2003 10:52:06 -0800
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <3E8974E5.88B861D7@engcorp.com> <435d40e0.0304011457.56e32bd0@posting.google.com> <sKSdnfwAUb0mrxejXTWcog@comcast.com>
Message-ID: <435d40e0.0304021052.736293b3@posting.google.com>

"Terry Reedy" <tjreedy at udel.edu> wrote in message news:<sKSdnfwAUb0mrxejXTWcog at comcast.com>...
> > What are the reasons for not having multi-line comments?
> 
> Because we more or less have them already -- the compiler can dump '''
> ''' strings that are not assigned to anything or otherwise used.
> 
> > I just remembered that IDLE will color-code """ as a string and not
>  a
> > comment - even when it is intended only as a comment.
> 
> Unlike the compiler, IDLE does not know the differrence.

But if there were multi-line comments everything would display with
the correct color code, in IDLE, or Emacs, etc.

> 
> Once could propose that ###, for instance, delimit multiline comments,
> but I suspect Guido would say, "Why bother?".

I believe I read that the forced indentation of Python was developed
after work done with beginning and/or casual programmers. In that
vein, commmenting out code would best be implemented by the use of
multi-line comments, and not "if 0:" or """ ... """. It can't be that
big of a change - the logic would seem to be almost identical to the
""" logic, except that you always discard the string, rather than
sometimes assigning it to a variable.

I would go with /* */ since it is already the multi-line comment in at
least 4 languages - Java, C, C++, Rexx (interpreted).

> 
> TJR



From claird at lairds.com  Mon Apr 21 11:07:20 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 21 Apr 2003 15:07:20 -0000
Subject: Testing methods
References: <9d8ac130.0304190938.755d4b4c@posting.google.com> <b7s2r9$11ka$1@news.peterlink.ru> <Zohoa.37253$LB6.847098@news1.tin.it>
Message-ID: <va8298t6r8hb13@corp.supernews.com>

In article <Zohoa.37253$LB6.847098 at news1.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>Anton Muhin wrote:
>   ...
>> try:
>>      foo.absent_method()
>> except (AttributeError, TypeError), e:
>>      print "Failed. Message @%s@" % e, "Type: %s" % e.__class__.__name__
>> else:
>>      print "OK"
>
>Careful!!!  There's a LITTLE trap here.  If foo DOES have absent_method
>BUT that method's body propagates one of these exceptions, this will
>catch the propagated exception AS IF it had been generated in the attempt
>to access / call it.  You can split the access-and-call in two to help:
>
>try:
>    bound = foo_absent_method
			.
		[important stuff
		on related topics]
			.
			.
  bound = foo.absent_method
is what Alex intended to write, I believe.

Typographical errors rarely merit follow-up--and Alex certainly
commits very, very few of them.  As the original poster described
himself as a beginner, I suspect there might be value in this case
in swift correction.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From p-abel at t-online.de  Wed Apr  9 07:12:24 2003
From: p-abel at t-online.de (Peter Abel)
Date: 9 Apr 2003 04:12:24 -0700
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com> <b70g6p$ecq$1@netlx020.civ.utwente.nl>
Message-ID: <13a533e8.0304090312.362b1e6f@posting.google.com>

Fred Gansevles <gansevle at ewi.utwente.nl> wrote in message news:<b70g6p$ecq$1 at netlx020.civ.utwente.nl>...
> Jim Shady wrote:
> 
> > Hello,
> > 
> > I am a beginner at Python and I just joined this group.
> > 
> > I have a string:
> > 
> > abcd/df/a/iiwk/abcdefghijkl/b/c
> > 
> > I need to get the longest string between the /s of the string. For
> > example, longest_str() for the above line should return
> > 'abcdefghijkl'. How do I go about doing this?
> > 
> > Thanks in advance.
> > 
> > Regards,
> > Jim
> 
> try this one:
> 
>         def ls(s, c='/'):
>             parts = s.split(c)
>             d = {}
>             for p in parts:
>                 if not d.has_key(len(p)):
>                     d[len(p)] = []
>                 d[len(p)].append(p)
>             k = d.keys()
>             k.sort()
>             return d[k[-1]]
> 
>         print ls('abcd/df/a/iiwk/abcdefghijkl/b/c')
> 
> it returns a list with the longest strings

Maybe one of the shortest possibilities - but not the most comprehensible - is:
>>> s='abcd/df/a/iiwk/abcdefghijkl/b/c'
>>> print max(map(lambda item:(len(item),item),s.split('/')))
(12, 'abcdefghijkl')
>>> 
Regards Peter



From tanzer at swing.co.at  Fri Apr 11 12:52:51 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Fri, 11 Apr 2003 18:52:51 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "Fri, 11 Apr 2003 14:25:54 +0200."
             <3e96b435$0$66259$edfadb0f@dread11.news.tele.dk>
Message-ID: <E1941lj-00079O-00@tswings.swing.cluster>

"Anders J. Munch" <andersjm at dancontrol.dk> wrote:

> "Christian Tanzer" <tanzer at swing.co.at> wrote:
> > Carl Banks <imbosol-1049990205 at aerojockey.com> wrote:
> > > I think dynamic scpoing is a wholly bad idea, and simplifying a very
> > > misguided programming practice doesn't make it any less so.
> >
> > You are of course right. Surprisingly though, it works very well for
> > Emacs. Passing arguments would *not* work in the context of Emacs.
> >
>
> What makes you say that?

The part you didn't quote <wink>. Here it is:

=> What you're missing is that Globals are Evil(tm), and furthermore,
=> that if your program depends on changing a global in a dynamic scope,
=> it badly needs to be redesigned.  (To wit, that functions that refer
=> to the global you want to change need to be changed to accept it as an
=> argument.)

`You are of course right` referred only to this paragraph.

The `Passing arguments would *not* work in the context of Emacs`
referred ony to the `To wit, ...` snippet.

> Why wouldn't lexical scoping supplemented by
> a temporary-global-binding construct work just as well or better?

Dynmaic scoping is used for lots of things in Emacs. With a different
design many of those uses could certainly be replaced by a simpler
design. I doubt that one could easily replace all of them, though.

As someone else pointed out, dynamic scoping is used to change values
of functions called indirectly -- thus converting them into arguments
passed around doesn't really work. One could use a
temporary-global-binding approach (or better temporary-object-binding
[as in temporarily overriding a buffer attribute]), but without special
forms like `with-` style macros it would look rather clumsily.

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From martin at v.loewis.de  Wed Apr 30 15:55:17 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 30 Apr 2003 21:55:17 +0200
Subject: Some oddities in 2.3b1, esp. with Tkinter
References: <b8mpp1$25hm$1@nntp6.u.washington.edu> <3EAEF1D6.30900@v.loewis.de> <b8p336$27pg$1@nntp6.u.washington.edu>
Message-ID: <m3smrzg3q2.fsf@mira.informatik.hu-berlin.de>

"Russell E. Owen" <no at spam.invalid> writes:

> >> 1) A filter to remove fonts with non-ascii names (used to filter out 
> >> stuff that I could not display in a pop-up menu) broke.
> >>             try:
> >>                 unicode(fontName, 'ascii', 'strict')
> >>                 return True
> >>             except UnicodeError:
> >>                 return False
[...]
> This is not true on MacOS X. I found it out the hard way. The font names 
> with unicode appear to be asian fonts, at least for the most part.

I see. Please find out what fontName is, and submit a bug report to SF.

> >Are you using Debian by any chance? Otherwise, you should not see
> >a threaded Tcl.
> 
> Hmm...I installed Aqua Tcl/Tk 8.4.2 from a MacOS X binary installer I 
> found on the net. I'm not sure what options it was built with (though 
> I'd have guessed the default ones). I can certainly install it from 
> source if that would help.

Ah, so the Aqua Tcl is multi-threaded. Bad choice. There is not much
we can do: Tcl threading and Python threading are incompatible. A Tcl
interpreter can be used only from the thread in which it is created.

> >If your Tcl installation is threaded, you should use 
> >tkapp.createfilehandler instead.
> 
> Hmmm...so I can first try my usual way of creating a file handler, and 
> if that fails, then try this? 

I recommend to *always* use
tkapp.createfilehandler. _tkinter.createfilehandler is only supported
for backwards compatibility, and cannot be supported in threaded Tcl.

> I'd be happy to. Could you please tell me where the Tkinter project 
> lives?

sf.net/projects/python

Regards,
Martin



From alanmk at hotmail.com  Tue Apr 29 14:50:36 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Tue, 29 Apr 2003 19:50:36 +0100
Subject: HTTP application streams a response to the client gradually - how 
 deal with it?
References: <Xns936CB5697A6D4cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <3EAEC97C.1EADEBF7@hotmail.com>

Harald Massa wrote:

> I want to connect to an HTTP application, that streams its response to 
> the client gradually.

Hmmm, this is actually done by HTTP, at a "software timescale", i.e.
when programs are talking to programs. For example, servers send
responses in "chunks", when they don't know how long the final entity is
going to be. The client software assembles the chunks, and presents the
finished whole to the user.

But there is a general assumption in HTTP that the process generating
the output will not delay unnecessarily, and the transaction will be
completed without unnecessary delay. 

> (it is a "web chat")

And if you tried to make the same thing happen at "human timescales", I
think you'd find yourself swimming against the current, i.e. find it
very difficult to do.

> From reading documentes and google, I learned that urllib will not fit.
> Also httplib --- I only see methods that provide me with whole HTTP-
> responses.

That's because HTTP has always been, and probably always will be, a
request/response protocol. One HTTP request (which may transfer several
resources) results in exactly one HTTP response (which may contain
multiple resources).

I think that HTTP is fundamentally mismatched to a chat application,
unless every "sentence" of the chat is POSTed to a central server, and
the clients regularly refresh the content from the central server. It is
not possible for the server to push content to the clients with HTTP.
Clients must initiate the conversation, and must do so regularly if they
want regular updates (this can be automated with meta-refresh tags in
HTML).

However, HTTP is just a protocol layered over sockets. 

I'm sure they're are several other protocols that more closely, or
probably precisely, match your requirements. For example, twisted seems
to have plenty of functionality in this regard:-

http://www.twistedmatrix.com/products/words
http://www.twistedmatrix.com/documents/examples/

If you disagree that HTTP is not suitable for a chat application, please
feel free to say why.

> Is there any way to deal with "streamed http" without going to socket-
> programming?

If you're going to do some socket programming, then the first thing
you'll probably end up doing is giving up on the idea of using HTTP,
which would make your life a *whole* lot simpler.

If your issue is with the use of port 80 (or whatever HTTP port is in
use), often being the only port that goes through firewalls, then that's
a different issue from trying to use HTTP for what it wasn't designed
for.

Hopefully someone will follow up this message with a list of protocols
that are suitable for chat apps. Anyone?

regards,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From bokr at oz.net  Tue Apr  1 05:49:43 2003
From: bokr at oz.net (Bengt Richter)
Date: 1 Apr 2003 10:49:43 GMT
Subject: Breaking out of a while loop with a key press?
References: <3e885ad5@mail.hmgcc.gov.uk> <3e8938e1$1@mail.hmgcc.gov.uk>
Message-ID: <b6bqs7$qlk$0@216.39.172.122>

On Tue, 1 Apr 2003 07:59:44 +0100, "Richard" <richardd at hmgcc.gov.uk> wrote:

>Hi guys,
>
>Thank you all for your suggestions. I'm sorry my question was a bit
>ambiguous with regard to the 'console window' reference. I was in fact
>referring to a Win2K Command Prompt (cmd.exe).
>
>I will try catching the KeyboardInterrupt exception and check that I have no
>bare 'except's anywhere that could be catching this for me.
>
Note that catching the exception means allowing the exception to interrupt
wherever in your loop execution may be, so you can not expect such a loop exit
to give you a final complete loop execution. See suggestions (mine and
others) for using signal handlers to accomplish orderly loop termination
based on modifying a while loop condition from a signal handler (which
returns to continue execution of the loop where interrupted).

Regards,
Bengt Richter



From gregory_r_warnes at groton.pfizer.com  Tue Apr  1 12:40:21 2003
From: gregory_r_warnes at groton.pfizer.com (Warnes, Gregory R)
Date: Tue, 1 Apr 2003 12:40:21 -0500 
Subject: Precidence of unary plus/minus relative to exponentiation
Message-ID: <D7A3CFD7825BD6119B880002A58F06C202F2C760@groexmb02.pfizer.com>

The current operator precedence places unary plus and minus higher than
exponentiation.  This leads the counterintuitive result:

	>>> -2.0**2.0
	-4.0
	
Python's operator predicence is causing this to be interpreted as:

	>>> -(2.0**2.0)

rather than the normal mathematical precidence, which would give

	>>> (-2.0)**2.0

Is there a good reason for the Python order of precedence?

-Greg


LEGAL NOTICE
Unless expressly stated otherwise, this message is confidential and may be privileged. It is intended for the addressee(s) only. Access to this E-mail by anyone else is unauthorized. If you are not an addressee, any disclosure or copying of the contents of this E-mail or any action taken (or not taken) in reliance on it is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately.




From hellprout at yahoo.com  Thu Apr 17 05:38:47 2003
From: hellprout at yahoo.com (hellprout)
Date: 17 Apr 2003 02:38:47 -0700
Subject: python and macro excel
Message-ID: <5b8834c2.0304170138.12f1a393@posting.google.com>

hi , i'm french 
i want to integrate a macro in excel with python and execute it 

what can i do ? or where can i find a idea ? 
i found some news on some win32.client site , so not my information 

thanks 

georges



From lost at soul.net  Fri Apr 25 19:29:48 2003
From: lost at soul.net (lost)
Date: Sat, 26 Apr 2003 00:29:48 +0100
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it>
Message-ID: <pan.2003.04.25.23.29.48.204375@soul.net>

On Fri, 25 Apr 2003 22:51:12 +0000, Alex Martelli wrote:

> lost wrote:
<snip>
> It's trivially easy to make each of these assumptions in some
> OTHER direction, of course, but, unless you give precise specs,
> people who try to help you will have to just take their best
> guesses at what you might have meant.
> 
Hi Alex,
sorry i wasn't a bit more specific in what i needed, but what you posted
was exactly what i was looking for!  Thanks a lot.






From jjl at pobox.com  Fri Apr 25 09:56:41 2003
From: jjl at pobox.com (John J. Lee)
Date: 25 Apr 2003 14:56:41 +0100
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com> <mailman.1050978337.8048.python-list@python.org> <87wuhm2rsu.fsf@pobox.com> <Bkipa.1154$Hm6.104194064@newssvr14.news.prodigy.com> <87adefkbs9.fsf@pobox.com> <m7Ypa.156$Jr1.14453940@newssvr15.news.prodigy.com>
Message-ID: <87znme8yue.fsf@pobox.com>

"Dennis Reinhardt" <DennisR at dair.com> writes:
[John Lee wrote:]
> > Yes, I read your post.  You said you did some kind of manual hack to
> > package your app, but clearly you have some system for your web page
> > packager (neat, BTW :-), which you didn't specify.
> 
> Errr..., that's right, I didn't specify how I did it.  I have abandoned this

I didn't intend any implied criticism, Dennis -- I was just explaining
why I naturally assumed you were using a pre-existing solution rather
than your own thing to make your Python-packaging web page.


> particular approach myself and describing what is going on is tough   There
> is lots of non-portable stuff in there.
> 
> Instead, I have posted code for the approach I am actually using above.

Cool.


John



From ianb at colorstudy.com  Tue Apr 22 14:46:35 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 22 Apr 2003 13:46:35 -0500
Subject: Overloading and?  was <RE: Should I prefer an external
	database>
In-Reply-To: <b8420j$629$1@slb3.atl.mindspring.net>
References: <mailman.1051023275.29338.python-list@python.org>
	 <b8420j$629$1@slb3.atl.mindspring.net>
Message-ID: <1051037195.702.185.camel@lothlorien>

On Tue, 2003-04-22 at 13:38, Andrew Dalke wrote:
> BTW, what I've done for other projects which build up a parse
> tree then evalute it is to hijack the arithmetic operators.  I've
> used '*' for and, '+', for or, and '-' for 'but not'.  In that case, the
> above expression is
> 
> > for paper in Paper.select(Paper.q.title == 'foo' +
> >                                  Paper.q.author.startswith('Bob')):

This won't work, though, because of the order of precedence.  It parses
like (Paper.q.title == ('foo' + Paper.q.author.startswith('Bob'))).

Better operators are &, |, and ~ (the binary logic operators).  You can
use these with SQLObject, but I don't like to, because they have the
same problems of precedence and they are more trouble than they are
worth.  AND() isn't pretty, but it's very straight-forward to use and
unlikely to cause surprise.

  Ian






From mis6 at pitt.edu  Fri Apr 25 14:30:21 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 25 Apr 2003 11:30:21 -0700
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk>
Message-ID: <2259b0e2.0304251030.30fd1f94@posting.google.com>

Michael Hudson <mwh at python.net> wrote in message news:<7h37k9ilm5v.fsf at pc150.maths.bris.ac.uk>...
> Thomas Heller <theller at python.net> writes:
> 
> > Michael Hudson <mwh at python.net> writes:
> > 
> > > mis6 at pitt.edu (Michele Simionato) writes:
> > > 
> > > > The meta-type conflict is rather annoying, indeed. According to the
> > > > metaclass book, the language should generate the correct metaclass
> > > > that solves the conflict automatically, without burden for the
> > > > programmer. Python is not that magic (yet) therefore you have to
> > > > solve the conflict yourself.
> > > 
> > > I guess a potential problem is that if you do
> > > 
> > > class C(object):
> > >     __metaclass__ = A
> > > 
> > > class D(object):
> > >     __metaclass__ = B
> > > 
> > > class E(C, D):
> > >     pass
> > > 
> > > class F(C, D):
> > >     pass
> > > 
> > > you'd like both E and F to have the *same* (rather than merely
> > > equivalent) metaclasses.  This would seem to be tricky to arrange.
> > 
> > Wouldn't a caching class factory be able to do this?
> 
> Yes.  I only said "tricky", not impossible.
> 
> class MetaGit(type):
>     def __eq__(self, other):
>         # aha! this makes class objects unhashable, cackle!
>         return self is other
> 
> I don't know whether you want to cope with things like that.
> 
> Cheers,
> M.

Two questions:

1) I don't see how this solve the problem of avoiding duplications of
metaclasses;

2) why you say this makes classes unhashable ?

It seems I can still use A and B as keys for a dictionary: 

class A: __metaclass__ = MetaGit
class B: __metaclass__ = MetaGit

dic={A:1, B:2}

print dic[A]  #=>1 okay
print dic[B]  #=>2 okay

?? Maybe I am missing something obvious ...


                       Michele



From peter at engcorp.com  Tue Apr 15 22:42:37 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 15 Apr 2003 22:42:37 -0400
Subject: find non-empty value in a dictionary
References: <b7ib8b$3rk$1@lust.ihug.co.nz>
Message-ID: <3E9CC31D.E5705394@engcorp.com>

matthew wrote:
> 
> say you have a dictionay d =
> {'X': 0.0, 'Y'': 240.2, 'Z': 0.0, 'P': 13.3445}
> 
> where p divides pretty much evenly into Y; in this case approx 18
> 
> P will always be present and only one of X,Y,Z will have a non-zero value.
> 
> How can I easily find which X,Y, or Z is non-zero; esp. given that
> comparing floats is 'inaccurate' as in:
> 
> if X != 0:
>         <stuff>

I don't really understand half of what you're asking, but in this
particular case you might not need to be so concerned.  

While comparing floats is "inaccurate" in one sense, comparing one
float zero against another will always return True, provided you 
haven't arrived at these values by *numerical calculation*.  In
other words, if these values are actually read in from some data
file or something, and they originally started out as, say, 
strings containing "0.0", then you can definitely compare them
with 0.0 without problem.

If they are only *nearly* zero, as a result of some calculation,
then you're quite right in avoiding the direct comparison.

-Peter



From exarkun at intarweb.us  Mon Apr 21 00:49:42 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 21 Apr 2003 00:49:42 -0400
Subject: python dynamic scoping question
In-Reply-To: <b7vg9t01mdo@enews2.newsguy.com>
References: <79ad5955.0304201513.6ffd4f83@posting.google.com> <b7vg9t01mdo@enews2.newsguy.com>
Message-ID: <20030421044942.GA11679@meson.dyndns.org>

On Sun, Apr 20, 2003 at 07:49:55PM -0500, Greg Krohn wrote:
> [snip]
> 
> From the docs:
> 
> globals() - Return a dictionary representing the current global symbol
> table. This is always the dictionary of the current module (inside a
> function or method, this is the module where it is defined, not the module
> from which it is called).
> 
> >>> def f():
> ...   def g():
> ...     x = 5
> ...     globals()['x'] = 5
> ...   g()
> ...   print x
> ...
> >>> f()
> 5
> 
> 
> Cool?
> 

  No.  The right way to do this is:

    def f():
        def g():
            global x
            x = 5
        g()
        print x

  But isn't the answer to the original poster's question, afaict.

  Jp

-- 
        "I quite agree with you," said the Duchess; "and the moral of
that is -- Be what you would seem to be' -- or, if you'd like it put
more simply -- Never imagine yourself not to be otherwise than what it
might appear to others that what you were or might have been was not 
otherwise than what you had been would have appeared to them to be
otherwise.'"       -- Lewis Carrol, "Alice in Wonderland"
-- 
 up 32 days, 1:02, 3 users, load average: 0.21, 0.21, 0.25




From qas20030408 at qa-systems.com  Tue Apr  8 12:09:26 2003
From: qas20030408 at qa-systems.com (QA Systems)
Date: Tue, 8 Apr 2003 18:09:26 +0200
Subject: Java Software Quality Problems?
Message-ID: <200304081609.h38G9QU18857@qarouter.qasystems.com>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030408/c216952f/attachment.html>

From jepler at unpythonic.net  Mon Apr 28 20:48:11 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Mon, 28 Apr 2003 19:48:11 -0500
Subject: In which versions is list.sort stable?
In-Reply-To: <b8kg8g$1d10$1@ns.felk.cvut.cz>
References: <mailman.1051479463.462.python-list@python.org> <mailman.1051562029.16746.python-list@python.org> <b8kg8g$1d10$1@ns.felk.cvut.cz>
Message-ID: <20030429004810.GA22245@unpythonic.net>

On Tue, Apr 29, 2003 at 02:15:12AM +0200, Frantisek Fuka wrote:
> Can you please explain to a newbie what this thread means?
> 
> Should I not use SORT in Python 2.2.2? Docs do not mention anything 
> about "unstability". Are there any other "bacis" functions (and SORT 
> seems pretty basic to me) that I should avoid???

It's not that list.sort is buggy in some versions of Python.

The "stability" of a sort refers to the result of a sort on a list where
there is not a total ordering.  For instance, consider the following list:
	[4, 3a, 2, 3b]
where 3a and 3b are equal to each other, but distinct.

One possible sorted version of that list is
	[2, 3b, 3a, 4]
but if your sort routine gives this result then it is not "stable", because
3a was before 3b in the input, but not in the output.  The "stable" output
is
	[2, 3a, 3b, 4]

In Python 2.3, the sort happens to have the property that it will always
give the *second* result, while in other versions either result may be
given.  Unless the sort is "stable", either result is acceptable.

A non-stable sort can be made stable by sorting on a tuple of (l[i], i)
because now there are no equal but distinct elements.  This is one use of
the decorate-sort-undecorate idiom.  A "sort stably" function was posted
elsewhere in this thread that (I think) achieves the result in this way.
As Tim (author of the new sort) also posted, this may be faster anyway(!).

Jeff




From tjreedy at udel.edu  Tue Apr 22 13:08:49 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 22 Apr 2003 13:08:49 -0400
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <f6a31cb8.0304220810.547f5acd@posting.google.com>
Message-ID: <oRKdnYDxiYyb5TijXTWcpA@comcast.com>

"Ted Drain" <teddrain at earthlink.net> wrote in message
news:f6a31cb8.0304220810.547f5acd at posting.google.com...
> My $.02:
>
> We use Python as a scripting interface to an underlying C++ system.
> The users write Python code to do what they want.  We ended up
writing
> __repr__ and __str__ in our classes to be the same thing because our
> users were very annoyed when they typed:
>
> >>> s
> >>> print s
>
> and got different results.

If you think of repr(o) as the codewriter view of an object o and
str(o) as the (passive) user view of o, then printing repr(o) in
interactive (codewriter) mode makes perfect sense.

There can, of course, be conflicts when users write code while still
thinking like users ;-)

Terry J. Reedy





From a.schmolck at gmx.net  Tue Apr 15 21:21:25 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 16 Apr 2003 02:21:25 +0100
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <mailman.1050453385.17091.python-list@python.org>
Message-ID: <yfsel43b5ka.fsf@black132.ex.ac.uk>

Jp Calderone <exarkun at intarweb.us> writes:

> On Tue, Apr 15, 2003 at 11:17:26PM -0000, Dave Benjamin wrote:
> > In article <yfs7k9vcu2p.fsf at black132.ex.ac.uk>, Alexander Schmolck wrote:
> > [snip]
> > 
> > >> keyword-argument trick wasn't really that bad, once you got used to it. In
> > >> any case, with dynamic scoping, this problem has completely disappeared.
> > >                  ^^^^^^^^^^^^^^^
> > > not quite.
> > 
> > Am I using the wrong terminology here? I'm specifically talking about the
> > named-argument hack to get around Python's previous lack of nested scopes.
> > This specific problem, as far as I know, is gone now. Am I missing something?
> > 
> 
>   I think you meant "partial lexical scoping" ;)  Dynamic scoping is what
> Python used to do.  Since it isn't quite lexical scoping, it's best to just
> stick with the term "nested scopes".

Aargh. No! You also have a look at:

http://www.supelec.fr/docs/cltl/clm/node43.html

'as



From martin at v.loewis.de  Sat Apr 12 06:48:38 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 12 Apr 2003 12:48:38 +0200
Subject: pickle and security
References: <pan.2003.04.10.18.52.25.740636@physics.purdue.edu> <m3wui2gjuw.fsf@mira.informatik.hu-berlin.de> <slrnb9d4ad.1m0.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <m3znmw3sah.fsf@mira.informatik.hu-berlin.de>

Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> writes:

> > Depending on your version of cPickle, it also contains a call to eval,
> > to unpickle a string. This is believed to be safe (as only safe
> > strings are passed to eval), but you may want to review that specific
> > fragment of code.
> 
> What about the attack suggested in
> <mailman.1049909846.24938.python-list at python.org> (which is to use
> "".__class__.__class__.__subclasses__("".__class__.__base__) to access
> supposedly inaccessible classes)?

It is unrelated to the context of my posting, as you cannot exploit
this in the eval calls that pickle/cPickle do.

I believe you can also not exploit this in pickle at all, since you
cannot have pickle perform arbitrary getattrs on unpickling.

Regards,
Martin




From BPettersen at NAREX.com  Thu Apr 24 11:44:51 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Thu, 24 Apr 2003 09:44:51 -0600
Subject: Question about accessing class-attributes.
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE0EA@admin56.narex.com>

> From: Duncan Booth [mailto:duncan at NOSPAMrcp.co.uk] 
> 
> Alex Martelli <aleax at aleax.it> wrote in 
> news:EARpa.120$3M4.2728 at news1.tin.it:
> 
> > ..isn't this smoother and handier?  Or does it just look that way
> > to me, due to overexposure to custom metaclasses?-)
> > 
> Yes, I agree that it is better, but I have one niggling worry 
> about this:
> 
> Your metaclass works fine for the one requirement given here, 
> and other metaclasses you have proposed work fine in their 
> place, but > what happens if you want to combine these and 
> get e.g. a Bunch which counts its instances?

Well, for some reason Python requires that if C1.__metaclass__ == M1,
C2.__metaclass__ == M2, and C2 is a subclass of C1, then M2 must be a
subclass of all metaclasses of of C2.__bases__, in particular M1. I
haven't seen a rationale for this (and it intuitively seems like an
arbitrary restriction... E.g. given two objects a and b, if b is a
subtype of a, i.e. if fn(b) and fn(a) both work [trying not to mention
contravariance], why would you expect a.__class__ and b.__class__ to
have any relation? [translate obj->class
obj.__class__->class.__metaclass__]).

-- bjorn




From just at xs4all.nl  Tue Apr 29 07:16:13 2003
From: just at xs4all.nl (Just)
Date: Tue, 29 Apr 2003 13:16:13 +0200
Subject: ttList-module problem
References: <mailman.1051609184.11588.python-list@python.org>
Message-ID: <just-95EA24.13161329042003@news1.news.xs4all.nl>

In article <mailman.1051609184.11588.python-list at python.org>,
 "Mehta, Anish" <Anish.Mehta at enst-bretagne.fr> wrote:

> Hello !
> 
> I am having some problem in understanding this code. Firstly i am not 
> finding TTFont which is being imported with the command
> 
> from fontTools.ttLib import TTFont
> 
> I have searched in the ttLib directory but i am unable to find TTFont so 
> that i can read about it. As far as i am getting is that this module is 
> taking input and then some information about      tags( ? ) and then 
> printing out the format in the form of checksum, length and offset. 
> Please elaborate this a bit more.
> 
> def ttList(input, output, options):
>         ttf = TTFont(input)
>         reader = ttf.reader
>         tags = reader.keys()
>         tags.sort()
>         print 'Listing table info for "%s":' % input
>         format = "    %4s  %10s  %7s  %7s"
>         print format % ("tag ", "  checksum", " length", " offset")
>         print format % ("----", "----------", "-------", "-------")
>         for tag in tags:
>                 entry = reader.tables[tag]
>                 checksum = "0x" + hex(entry.checkSum)[2:].zfill(8)
>                 print format % (tag, checksum, entry.length, entry.offset)
>         print
>         ttf.close()
> 
> 
> I'll be waiting for your response.

TTFont is in ttLib/__init__.py

Just



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Wed Apr 23 15:25:48 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Wed, 23 Apr 2003 21:25:48 +0200
Subject: ANN: Pyro 3.2 released
Message-ID: <3ea6e8bc$0$49105$e4fe514c@news.xs4all.nl>

I'm glad to announce the latest version of Python's own powerful
remote method invocation technology -- Pyro 3.2!

You can get it via http://pyro.sourceforge.net, then go
to the SF project homepage download area.

Most important new stuff since 3.1:
* 2-way mobile code
* Customizable authentication mechanism
* Windows NT Services, Unix Daemons
* The regular buch of bug fixes
* Documentation updates and clarifications
* As usual the complete change history can be found in the manual.


---> What is Pyro?
Pyro is an acronym for PYthon Remote Objects. Pyro is an advanced and
powerful Distributed Object Technology system written entirely in Python,
that is designed to be very easy to use.

It is extremely easy to implement a distributed system with Pyro, because all
network communication code is abstracted and hidden from your application.
You just get a remote Python object and invoke methods on the object on
the other machine.

Pyro offers you a Name Server, an Event Service, mobile objects, remote
exceptions, dynamic proxies, remote attribute access, automatic
reconnection, a very good and detailed manual, and many examples to get
you started right away.

Have fun!
--Irmen de Jong.




From rmunn at pobox.com  Tue Apr 15 17:30:25 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 15 Apr 2003 21:30:25 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it>
Message-ID: <slrnb9o9ob.493.rmunn@localhost.localdomain>

Alex Martelli <aleax at aleax.it> wrote:
> Robin Munn wrote:
>    ...
>> How is try ... finally "broken and fragile code"? How is this:
>> 
>>     try:
>>         old_stdout = sys.stdout
>>         new_stdout = cStringIO.StringIO()
>>         sys.stdout = new_stdout
>>         function_whose_output_I_want_to_capture()
>>         do_something_with(new_stdout)
>>     finally:
>>         sys.stdout = old_stdout
>> 
>> in any way inferior to this (making up a semi-Pythonic syntax):
>> 
>>     with sys.stdout = cStringIO.StringIO():
>>         function_whose_output_I_want_to_capture()
>>         do_something_with(sys.stdout)
>>
[snip...]
>> 
>> What I really want to know is why you would consider sample #1 "broken
>> and fragile code". I can't think of any circumstances where sample #1
>> would fail to achieve its desired effect. Can you?
>
[snip...]
> 
> Another difference between the two snippets, which I don't
> understand, comes when function_blah_blah does its own rebinding
> of sys.stdout.  In that case, the first snippet "ignores" the
> rebinding (calls do_something_with with the value you had just
> bound to sys.stdout BEFORE calling function_blahblah), the
> second snippet "accepts" the rebinding (calls do_etc with the
> value CURRENTLY bound to sys.stdout). Since I don't know what's
> SUPPOSED to be happening under such circumstances, I can't tell
> which of the two versions (if either) is wrong, but I do find
> the difference puzzling.  Assuming the second hypothetical snippet
> has the correct semantics, then, maybe the first one should be:
> 
>     old_stdout = sys.stdout
>     sys.stdout = cStringIO.StringIO()
>     try:
>         function_whose_output_I_want_to_capture()
>         do_something_with(sys.stdout)
>     finally:
>         sys.stdout = old_stdout
> 
> to achieve the same semantics -- specifically, what we want
> to put in the try clause is just what would go in the with
> clause's body if it existed, while the "saving the previous
> state" (implicit in the hypothetical 'with') and the actual
> initialization (the assignment that follows the with keyword
> itself in the hypothetical case) need to be BEFORE the try
> clause.

You're right; using the new_stdout variable name was unnecessary and
would have rejected any sys.stdout rebinding that function_etc() might
have done.

I suppose I just illustrated how a try ... finally block might be
considered "broken and fragile". I tried to use it, and missed some
subtleties. I'm still not convinced that a "with"-like statement is
necessary in Python, but I think I just went from -1 to -0 on it.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From max at alcyone.com  Thu Apr 24 19:10:18 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 24 Apr 2003 16:10:18 -0700
Subject: Exception and finally question
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com>
Message-ID: <3EA86EDA.EFD023C0@alcyone.com>

Tung Wai Yip wrote:

> How to I access the exception object in a handler?
>
> try:
>     raise NameError, 'HiThere'
> except NameError:
>     #how to print the message 'Hi There' here?

	try:
	    raise NameError, "Hi, there."
	except NameError, e:
	    print e

Note that e is actually the exception _object_, not just the string (or
whatever other argument) it was constructed with:

>>> try:
...  raise NameError, "What's up, dog?"
... except NameError, e:
...  print repr(e)
...  print str(e)
... 
<exceptions.NameError instance at 0x815b3b4>
What's up, dog?

> Why except and finally can't be used together? I like to log a error
> message if an exception is throw. And I a file to be guaranteed to be
> close. What can't I do them together?

By design.  Originally, as I understand it, Python supported a unified
try: ... except: ... finally: ... construct, but evidently it caused a
lot of confusion with newbies (personally, I don't see where the
confusion would be, but I guess that's because I'm familiar with the
construct from Java).  So in modern Python the try can either have an
except (perhaps several) or a finally, but not both.

You can get the effect you're probably looking for by just wrapping a
try: ... except: ... inside a try: ... finally: ...

	try:
	    try:
	        somethingThatMayRaise
	    except ...:
	        handleErrorCaseOrMaybeReraise
	finally:
	    doThisRegardless	

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Insight -- the titillating knack for hurting!
\__/ Colette
    Blackgirl International / http://www.blackgirl.org/
 The Internet resource for black women.



From hmm at iki.fi  Tue Apr  1 06:23:24 2003
From: hmm at iki.fi (Markku =?iso-8859-1?q?H=E4nninen?=)
Date: 01 Apr 2003 14:23:24 +0300
Subject: use for jython?
References: <pan.2003.03.29.12.59.36.983303@gamebox.net>
Message-ID: <87el4mea1f.fsf@skootteli.hmm.iki.fi>

"A Nilsson" <anders at gamebox.net> writes:
> I am just curious: is there anyone using jython for a "real"
> project or is it just useful for quick-and-dirty java tests?

We have embedded Jython to our Java/J2EE-product as a scripting language.

I have also delivered some servlets to a customer using jython. The original
plan was to prototype with jython and write the actual code in java but
the schedule was so tight that we ended up shipping the jython code. The 
customer has actually already done some own servlets with jython after
that (without any knowledge of it before).

I personally think that Jython will be a "killer" in that sense that it
combines the vast available Java-libraries with the ease of Python and would
like to see that Jython would follow the Python releases more closely. 

-- 
   Markku H?nninen     /
     hmm at iki.fi       /   "Actually it works, it just doesn't look that way." 



From mwm at mired.org  Tue Apr  1 12:52:43 2003
From: mwm at mired.org (Mike Meyer)
Date: Tue, 01 Apr 2003 17:52:43 GMT
Subject: PEP 313 - Roman numeral literals
Message-ID: <x7y92uhzpw.fsf@guru.mired.org>

PEP: 313
Title: Adding Roman Numeral Literals to Python
Version: $Revision: 1.2 $
Last-Modified: $Date: 2003/04/01 17:41:34 $
Author: Mike Meyer <mwm at mired.org>
Status: Draft
Type: Standards Track
Content-Type: text/plain
Created: 01-Apr-2003
Python-Version: 2.4
Post-History:


Abstract

    This PEP proposes adding Roman numerals as a literal type.  It
    also proposes the new built-in function "roman", which converts an
    object to an integer, then converts the integer to a string that
    is the Roman numeral literal equivalent to the integer.


Rationale

    Roman numerals are used in a number of areas, and adding them to
    Python as literals would make computations in those areas easier.
    For instance, Superbowls are counted with Roman numerals, and many
    older movies have copyright dates in Roman numerals.  Further,
    LISP provides a Roman numerals literal package, so adding Roman
    numerals to Python will help ease the LISP-envy sometimes seen in
    comp.lang.python.  Besides, the author thinks this is the easiest
    way to get his name on a PEP.


Syntax for Roman literals

    Roman numeral literals will consist of the characters M, D, C, L,
    X, V and I, and only those characters.  They must be in upper
    case, and represent an integer with the following rules:

    1.  Except as noted below, they must appear in the order M, D, C,
    L, X, V then I.  Each occurence of each character adds 1000, 500,
    100, 50, 10, 5 and 1 to the value of the literal, respectively.

    2.  Only one D, V or L may appear in any given literal.

    3.  At most three Is, Xs and Cs may appear in any given literal.

    4.  A single I may appear immediately to the left of the single V,
    followed by no Is, and adds 4 to the value of the literal.

    5.  A single I may likewise appear before the last X, followed by
    no Is or Vs, and adds 9 to the value.

    6.  X is to L and C as I is to V and X, except the values are 40
    and 90, respectively.

    7.  C is to D and M as I is to V and X, except the values are 400
    and 900, respectively.

    Any literal composed entirely of M, D, C, L, X, V and I characters
    that does not follow this format will raise a syntax error,
    because explicit is better than implicit.


Builtin "roman" Function

    The new builtin function "roman" will aide the translation from
    integers to Roman numeral literals.  It will accept a single
    object as an argument, and return a string containing the literal
    of the same value.  If the argument is not an integer or a
    rational (see PEP 239 [1]) it will passed through the existing
    builtin "int" to obtain the value.  This may cause a loss of
    information if the object was a float.  If the object is a
    rational, then the result will be formatted as a rational literal
    (see PEP 240 [2]) with the integers in the string being Roman
    numeral literals.


Compatability Issues

    No new keywords are introduced by this proposal.  Programs that
    use variable names that are all upper case and contain only the
    characters M, D, C, L, X, V and I will be affected by the new
    literals.  These programs will now have syntax errors when those
    variables are assigned, and either syntax errors or subtle bugs
    when those variables are referenced in expressions.  Since such
    variable names violate PEP 8 [3], the code is already broken, it
    just wasn't generating exceptions. This proposal corrects that
    oversight in the language.


References

    [1] PEP 239, Adding a Rational Type to Python
        http://www.python.org/peps/pep-0239.html

    [2] PEP 240, Adding a Rational Literal to Python
        http://www.python.org/peps/pep-0240.html

    [3] PEP 8, Style Guide for Python Code
        http://www.python.org/peps/pep-0008.html


Copyright

    This document has been placed in the public domain.


^L
Local Variables:
mode: indented-text
indent-tabs-mode: nil
sentence-end-double-space: t
fill-column: 70
End:

-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From exarkun at intarweb.us  Wed Apr  2 16:36:48 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 16:36:48 -0500
Subject: Comments
In-Reply-To: <b6ff2r$g1a$1@news.schlund.de>
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <435d40e0.0304011457.56e32bd0@posting.google.com> <sKSdnfwAUb0mrxejXTWcog@comcast.com> <435d40e0.0304021052.736293b3@posting.google.com> <b6fc6e$1fe$1@panix2.panix.com> <b6ff2r$g1a$1@news.schlund.de>
Message-ID: <20030402213648.GA15307@meson.dyndns.org>

On Wed, Apr 02, 2003 at 09:52:59PM +0200, Carsten Gaebler wrote:
> Roy Smith wrote:
> >Multi-line comments are evil
> >
> >The first time you spend a half-hour trying to understand a block of
> >ugly code and finally get to the end and find a "*/", you'll
> >understand why they're evil.
> 
> ACK.
> 
> But multi-line strings are evil, too. At least when they occur in an 
> indented block and leading whitespace is not allowed in the string 
> your're building (e.g. in an HTTP request or something). That looks 
> quite awful. :-)

  Agreed, which is what literal concatenation is for

    c = ("Here is the first line\n"
         "Here is another line\n"
         "Last line, here I am.\n")

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 13 days, 18:00, 7 users, load average: 0.10, 0.09, 0.08




From gerritvd at decillion.net  Fri Apr  4 01:00:59 2003
From: gerritvd at decillion.net (Gerrit van Dyk)
Date: Fri, 04 Apr 2003 08:00:59 +0200
Subject: Distributed environment
Message-ID: <3e8d1edc$0$235@hades.is.co.za>

Hi all,

I am in the process of designing a system that will need to use a 
distributed environment. 

The system will consist of multiple (thin) clients += 100-200 that will need 
to connect to a central server. The client applications is not 
browser-based as they need to do a lot of data manipulation, draw charts, 
run scenarios etc. The idea is to do most of the processing on servers 
except for the GUI stuff. 

Several different services will be registered with the central server to 
supply different types of services (rate feeds, risk analysis, forecasting 
etc). 

The central server(HUB) needs to control all the services and clients. When 
a request comes in from a client, the HUB needs to distribute this request 
to the correct service. Multiple services could be registered for the same 
type of service. This will enable multiple clients to use the same type of 
service concurrently. The services could also make use of relational 
databases or other types of data stores. The client applications will have 
no concept of data storage and will have no direct access to the databases.

In some cases, some of the services will be needing the services of other 
services, in essence also becoming clients for the other services.

The central server (system) needs to keep track of all the clients and their 
requests. This is to ensure that if a machine or service fails the system 
should be able to recover without affecting the clients at all. The system 
needs to be up 24/7/52.

These services will also be distributed across multiple machines for load 
balancing.

My questions are.

Is this feasable? 
Any pointers or pitfalls I might step in?
Which environments would you recommend I should look at, for doing this type 
of thing?

Regards
Gerrit



From ianb at colorstudy.com  Tue Apr 29 06:06:27 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 29 Apr 2003 05:06:27 -0500
Subject: Web programming and a different "type" problem (was: Slice
	confusion : a[n:p] is a list exclude the last  element p)
In-Reply-To: <vaqvrb3ipqblf2@corp.supernews.com>
References: <mailman.1051532982.25352.python-list@python.org>
	 <UXara.22428$K35.745468@news2.tin.it>  <vaqvrb3ipqblf2@corp.supernews.com>
Message-ID: <1051610786.684.2678.camel@lothlorien>

On Mon, 2003-04-28 at 14:22, Cameron Laird wrote:
> 'Doesn't work, though.  Here's where the special case
> enters:  if only one datum with the indicated name is
> present as input, Zope and the others return it as a
> simple string, rather than a list with that string as 
> its sole member. 

I'm not a fan of using field names (like field:list, or field*, or
field[]), though I suppose it isn't that bad.  Without that you have the
option of a custom getter, like maybe:

valueList = request.multiple('field')

Which promotes a single string to a list, and if the value is missing
returns an empty list (which the field names don't do).

Otherwise you got to use a schema.  I made a simple schema system
recently, as I've been playing around with forms.  It's the
DictCall.(txt,py):

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/webware-sandbox/Sandbox/ianbicking/FormEncode/

It has several ways of defining simple schemas which are used to
translate a dictionary of input into a function call.  So you can do
something like:

# goto delete?id=1&id=2&id=3
def delete(id_set_int): 
    # id_set_int is a list of integers...

You could also define the schema through a function attribute or the
docstring.

Really just an experiment, but I'm still playing with the ideas.

  Ian






From awheat at imaging.robarts.ca  Mon Apr  7 14:23:13 2003
From: awheat at imaging.robarts.ca (Andrew Wheatley)
Date: 7 Apr 2003 11:23:13 -0700
Subject: Tkinter - tracking toplevel windows
Message-ID: <d44e8ee7.0304071023.1d571c2b@posting.google.com>

I have a Tkinter Gui with a button which generates other toplevel windows.
Whenever a toplevel is generated, I append its id to a global list.  What I'd like
to do with wm_protocol/WM_DELETE_WINDOW method is remove from the list
the window id currently being killed and then destroy the window.

My question is:

How can I tell which window is being destroy when I kill it with the X in the 
top right hand corner.



From no at mail.com  Tue Apr 29 09:29:41 2003
From: no at mail.com (Jeroen Wolff)
Date: 29 Apr 2003 13:29:41 GMT
Subject: os.spawn
Message-ID: <Xns936C9D9E91B1Bthewolffman@194.109.133.29>

Hi,

I want to run some shell commands from within python. 
Wat i normal do at the shell prompt is:
'flow-cat <dirname> | flow-stat <param> > outputfile.txt'

I like to do this from within python and wait for it to finish and the 
process the outputfile.txt
Is this possible? The problem is with '|' and '>" piping....

Can somebody please help?

Thanks,

Jeroen Wolff



From a.schmolck at gmx.net  Thu Apr 10 09:41:54 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 10 Apr 2003 14:41:54 +0100
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org>
Message-ID: <yfsbrzeh3kt.fsf@black132.ex.ac.uk>

Jeff Epler <jepler at unpythonic.net> writes:

> Ignoring the fact that "globals" are module-level, you can obtain
> "stacking" of global variables with only a single level of name lookup,
> including automatic removal on return, in Python.  All that you need is a
> convenent way to write complex things in the 'body' argument.  Untested
> code, of course
> 
> _let_undef = object()
> def let(body, **kw):
> 	restore = {}
> 	g = globals()
> 	for k, v in kw.items():
> 		restore[k] = g.get(k, _let_undef)
> 		g[k] = v
> 	try:
> 		return body()
> 	finally:
> 		for k, v in restore.items():
> 			if v is _let_undef:
> 				del g[k]
> 			else:
> 				g[k] = v
>

Sure, but this is a but a clumsy, almost unusable imitation of scheme's
(non-standard, but common) fluid-let macro. So what are you doing for
multithreaded code (which scheme's faked dynamic extent via fluid-let doesn't
cover easily but CL's special variables do)?

Maybe I'm missing something, but non-dynamically scoped globals just seem like
a bad idea to me.


'as




From martin at v.loewis.de  Sun Apr 27 15:52:11 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Sun, 27 Apr 2003 21:52:11 +0200
Subject: 2.3b1 and Tk8.4.2
In-Reply-To: <3eac2d62$0$594$ba620e4c@reader1.news.skynet.be>
References: <3eac2d62$0$594$ba620e4c@reader1.news.skynet.be>
Message-ID: <b8hcd5$f5d$07$1@news.t-online.com>

Helmut Jarausch wrote:
> Is there a fix other than keeping Tk8.3 just for Python?

I recommend to analyse the problem first.

Why is there a reference for TkSetClassProcs in _tkinter.so?
_tkinter.c does not call that function.

Regards,
Martin




From bokr at oz.net  Sun Apr 27 14:39:06 2003
From: bokr at oz.net (Bengt Richter)
Date: 27 Apr 2003 18:39:06 GMT
Subject: Reading a non-standard floating point format
References: <mailman.1051400681.24630.python-list@python.org> <b8fd9r$6fo$0@216.39.172.122> <c76ff6fc.0304262323.7350200c@posting.google.com>
Message-ID: <b8h84a$e5n$0@216.39.172.122>

On 27 Apr 2003 00:23:05 -0700, sjmachin at lexicon.net (John Machin) wrote:

>bokr at oz.net (Bengt Richter) wrote in message news:<b8fd9r$6fo$0 at 216.39.172.122>...
>> On Sat, 26 Apr 2003 19:41:43 -0400, "Tim Peters" <tim_one at email.msn.com> wrote:
>> [...]
>> >There's no way to guess without examples, though.  If you can, show us the
>> >48-bit byte strings for these numbers:
>> >
>> >    0.0
>> >    1.0
>> >    1.25
>> >    2.9e-39
>> >    1.7e38
>> >
>> >That should be enough to fill in the missing pieces.
>> >
>> Maybe include
>>     -1.25
>> also?
>> And can minus zero be generated?
>> Are numbers sign/magnitude or?
>> 
>
>I would ask for 1.125 as well as 1.25 (to be sure of bit order within
>a byte[*]) and also 1.03125 (1+1/32, 8 bits onwards from 1.25) to be
>sure of byte order within the 48-bit stucture.
>
That might be easier to interpret than Tim's numbers, but his do include
a bunch of bits in the values (for double):

 >>> prb(2.9e-39)
 '.0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
 000000000000000000000000000000000011111100101000000010100100111100010000100111011001011'
 >>> prb(1.7e+38)
 '11111111110010011001111000111100011001101111110101101000000000000000000000000000000000000000000
 00000000000000000000000000000000'
 >>> prb(1.25)
 '1.01'

Or just the significant ones
 >>> prb(1.7e+38).rstrip('0')
 '11111111110010011001111000111100011001101111110101101'
 >>> prb(2.9e-39)[1:].lstrip('0')
 '11111100101000000010100100111100010000100111011001011'

>The OP might also consider posting on alt.sys.alpha-micro -- surely
>somebody there will be able to provide better docs than the quoted
>excerpt from the AlphaBasic manual. You could ask about access to the
>AMs that are still running -- it could be very handy to mount a
>"chosen-plaintext" attack on the problem i.e. write an AlphaBasic
>program which wrote various chosen f.p. numbers to a "random access"
>(i.e. binary file) then ogle it with the aid of a hex dumper.
>
>* ... bit order within a byte: don't laugh, we're in 1980s
>microcomputer territory here; no f. p. h/ware, all done in ASM by
>heavy bit-bashing. I recall different variants of "standard" 32-bit
>CRCs -- the variants arising from whether the bits were shifted in
>from the left/right/high-end/low-end/east/west/...
>
ISTR a software variant that would postpone normalization also, to benefit
from things happening to be lined up already in intermediate results
of a computation ... back when we were happy if we could make all the ops
but floating point divide and square root under a millisecond.

So how do you feel about nixie tubes and biquinary ;-)

Regards,
Bengt Richter



From mcherm at mcherm.com  Wed Apr 16 11:48:58 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Wed, 16 Apr 2003 08:48:58 -0700
Subject: How many of you are Extreme Programmers?
Message-ID: <1050508138.3e9d7b6a1d520@mcherm.com>

Christopher writes:
> So how many of you guys use XP processes?

Well, I'm not sure if I qualify or not. I really LIKE most of the points
that XP makes, but I find many of its proponants overly exuberant. I
think exhaustive, automated tests are a wonderful idea. I believe in
refactoring. I think pair programming works very well most of the time.
I think super-short release cycles (with code that grows feature-by-
feature) is a great idea. I think having a customer on-site is extremely
helpful.

I DON'T think that it's wise to just start coding without a bit of design
up front (although one must keep YAGNI in mind). I don't think that using
the code itself as the only documentation (no design docs) works well.
Most of all, I don't like mandating any particular methodology... the
needs vary from project to project and team to team.

But none of that really matters, because I don't get to set the policy
where I work. We don't do unit tests (except for the ones that _I_ write)
because "they would take too long". We deliver functionality in a single
big release. We don't do pair programming except for very brief periods.
Perhaps someday I can change those things, but for now, I don't get to
practice much of this, except in my at-home-on-my-own-time projects
(basically non-existant since my son was born 4 months ago).

-- Michael Chermside





From sholden at holdenweb.com  Wed Apr  9 05:40:52 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 09 Apr 2003 09:40:52 GMT
Subject: atexit functionality
References: <mailman.1049821814.20582.python-list@python.org>
Message-ID: <EURka.30982$It5.2640@news2.central.cox.net>

[posted and mailed]

----- Original Message -----
From: <viszneki at telerama.com>
Newsgroups: comp.lang.python
Sent: Tuesday, April 08, 2003 1:30 PM
Subject: atexit functionality


> I am not a member of the Python mailing list, I dont' have the time to
read any
> mailing lists at the moment. So if someone could please respond to me
> personally, that would be great.
>
> Generally speaking I have a lot of objects that clean themselves up before
> python actually exits, so I was wondering if a deregister function could
be
> provided.
>
> Rather than a deregister function, I thought about making the following
> modification to atexit.py after _exithandlers.append((func,targs,kargs))
>
> return _exithandlers[len(_exithandlers)]
>
> My intention is to be able to store the returned value as a member
variable in
> a class that registers its own cleanup functions. But if the class cleans
> itself up, the class could "clean up" its cleanup registration with atexit
by
> saying something like del self.cleanup
>
> I hope my ideas are well received! I think they will be because this
> modification is fully backward compatible, and doesn't introduce any
negative
> side effects or anything like that.
>
> Thanks for your time,
> Donny Viszneki.
>
> ps, could someone please personally respond to my email address
> viszneki at telerama.com about any decisions made about my idea? thanks a
lot.
>
Donny:

Effectively, what you just did is to say "I'm far too busy to engage fully
in this news group, so please waste your time responding privately to my
suggestion in such a way that the group is unlikely to effectively benefit
from anything positive that might happen as a result of this post."

Not terribly good netiquette, that. I hope you get the replies you wanted as
well as this rather grumpy message.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From j4nff5m02 at sneakemail.com  Fri Apr 18 03:41:33 2003
From: j4nff5m02 at sneakemail.com (Roel Schroeven)
Date: Fri, 18 Apr 2003 07:41:33 GMT
Subject: Random web page?
In-Reply-To: <b7neqo$knk$1@news.hccnet.nl>
References: <b7m4ql$ril$1@news.hccnet.nl> <3E9EAA36.BFB1018D@engcorp.com> <IQmdnQg6xagDdQOjXTWcpA@august.net> <b7neqo$knk$1@news.hccnet.nl>
Message-ID: <N_Nna.50877$t_2.4682@afrodite.telenet-ops.be>

Anton Vredegoor wrote:

> But why do the DNS-lookup if it's possible to load a page using an ip?

Because not all websites can be reached that way, due to virtual 
hosting. Take my website for example (don't look at it, it's really not 
interesting :) ). http://roelschroeven.net gives a normal webpage, while 
http://64.191.37.195 (which is the ip-address for that domain) gives a 
generic page (in this case something that shouldn't really be there IMHO).

-- 
"Codito ergo sum"
Roel Schroeven




From alanpong at hkstar.com  Fri Apr 18 04:31:56 2003
From: alanpong at hkstar.com (Alan Pong)
Date: 18 Apr 2003 01:31:56 -0700
Subject: Problem in Accessing C++ Class Instance Global Variable (near success, help please!!)
Message-ID: <ea0e2298.0304180031.49129224@posting.google.com>

python 2.3, vc++6, win32 exe project

I has one global c++ class instance variable.  I want to call its
member
function through wrapper function which is accessed by python.
However, as seen in wrapper.cpp, it give me another new variable copy
so that phython actually accessing another instance.

Does anybody has idea to fix it?
thanks.
rgds.
alan

-- test.cpp (will be compiled into test.exe)
CTest gtest; //one and only one global variable
main()
{
	gtest.SetString("abcd");
	Py_Initialize();
	PyImport_AddModule( "myext1" );
	PyRun_SimpleFile( fp, "test.py" );
	prinf("%s",gtest.GetString());	//still "abcd", unchange
}


-- wrapper.cpp (will be compiled into test.exe)
extern CTest gtest;

extern "C" CTest * getInstance()
{
	return &gtest;  //***Question is here: a new variable!!! why?***
}

extern "C" void SetString(char* str)
{
	gtest.SetString(str); 
	messagebox(null,str,null,0); //shows "2345",i.e. the script works!
}

-- myext1.c (will be compiled into myext1.dll)
extern char *getInstance();
extern void SetString(char *str);

static PyObject *
import_getInstance( PyObject *self, PyObject *args )
{
  char *obj = getInstance();
  return Py_BuildValue("i", obj);
}

static PyObject *import_SetString(PyObject *self, PyObject *args )
{
	char *msg;

	if (!PyArg_ParseTuple(args, "s", &msg))
	    return NULL;
	SetString(msg);
	Py_INCREF(Py_None); 
	return Py_None;
}

-- test.py
from myext1 import *
SetString("2345")

-- END



From glc at well.com  Tue Apr 15 10:02:21 2003
From: glc at well.com (Greg Chapman)
Date: Tue, 15 Apr 2003 14:02:21 GMT
Subject: Implementing Metatypes in C
References: <mailman.1050275480.8590.python-list@python.org> <lk2o9v8bamjhgdmas6ce2g2qvuthril237@4ax.com>
Message-ID: <3l3o9v0ks3b69fn67pv0i15hov8jlj53ki@4ax.com>

On Tue, 15 Apr 2003 13:34:47 GMT, Greg Chapman <glc at well.com> wrote:

>I believe your solution may break in 2.3a2 if subtypes declare tp_members, since
>the PyMemberDefs for these are kept in the variable part of the type structure
>which follows the things declared in etype

Actually, that's bogus.  It's not tp_members which get stored in the etype's
members, but the PyMemberDefs for __slots__.  Nevertheless, without the patch,
you can run into problems trying to extend the etype structure in a metatype.

---
Greg Chapman




From ckea25d02 at sneakemail.com  Tue Apr  8 00:55:04 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 16:55:04 +1200
Subject: exceptions with spaces don't work??
In-Reply-To: <Pine.LNX.4.44.0304080156480.1201-100000@alice>
References: <b6t48i$mvd$1@news.doit.wisc.edu> <Pine.LNX.4.44.0304080156480.1201-100000@alice>
Message-ID: <b6tk5f$8rudj$1@ID-169208.news.dfncis.de>

John J Lee wrote:
> Maybe it happens to work sometimes because Python is caching
> short strings, so the two occurrences of "someerror" happen to be the same
> object in memory.

Essentially, yes. The caching is known as "interning" (a term
which goes back to the early days of Lisp). The bytecode
compiler interns any string literals which happen to resemble
Python identifiers (because this speeds lookup of names at
runtime).

Putting spaces in the string is one way to make it
not look like an identifier, so it doesn't get interned.
There are many other ways, of course.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From tanzer at swing.co.at  Mon Apr 14 04:00:46 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Mon, 14 Apr 2003 10:00:46 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "Sat, 12 Apr 2003 06:40:42 GMT."
             <KxOla.1317$ah6.258@nwrdny03.gnilink.net>
Message-ID: <E194ytS-0002aI-00@tswings.swing.cluster>

Carl Banks <imbosol-1050127854 at aerojockey.com> wrote:

> Christian Tanzer wrote:
> > Dynmaic scoping is used for lots of things in Emacs. With a different
> > design many of those uses could certainly be replaced by a simpler
> > design. I doubt that one could easily replace all of them, though.
> >
> > As someone else pointed out, dynamic scoping is used to change values
> > of functions called indirectly -- thus converting them into arguments
> > passed around doesn't really work.
>
> No, but lexical closures can do this just fine:
>
>     def a():
>         value_to_be_passed_indirectly = some_calculation()
>         def b():
>             do_something_with(value_to_be_passed_indirecty)
>         call_function_that_calls_callback(b)
>
>
> Not to mention you can do similar stuff with a class:
>
>     def a(self):
>         closure = B(some_calculation())
>         call_function_that_calls_callback(closure.b)
>
>     class B:
>         def __init__(self,value):
>             self.value_to_be_passed_indirectly = value
>         def b(self):
>             do_something_with(self.value_to_be_passed_indirecty)
>
>
> The fact is, arguments have been passed indirectly in Python for a
> long time, happily, and without dynamic scoping.  It will not be hard
> to replace this use of dynamic scoping.

You didn't understand what I tried to express. One last try:

Emacs has lots (thousands?) of variables influencing the behavior of
the editor. Some are global, some are buffer-local, some are
frame-local, some are bound to text properties, ...

To make this more concrete, lets look at an example. Emacs provides
text properties which allow customization of cursor movements in part
of a buffer. Using these influences the behavior of all Emacs
builtins, e.g., goto-char. If you want to implement a function which
needs unencumbered movement for whatever reason, it simply `let`s
`inhibit-point-motion-hooks` to a true value before calling other
functions. What gets called inside the `let` can be an arbitrary
function and that function calls other arbitrary functions the
programmer might know nothing about. But somewhere down the call tree,
some builtin will be called which behaves differently depending on the
value of `inhibit-point-motion-hooks` and the text properties in the
buffer.

All these functions (except the one you're writing) might have been
implemented years before the text properties and
`inhibit-point-motion-hooks` were introduced.

> > One could use a
> > temporary-global-binding approach (or better temporary-object-binding
> > [as in temporarily overriding a buffer attribute]), but without special
> > forms like `with-` style macros it would look rather clumsily.
>
> Even if you did want to do this, it would hardly be clumsy to use a
> try ... finally.

Just compare:

    old_a = buffer.a
    old_b = buffer.b
    old_c = buffer.c
    try :
       buffer.a = fct_a(...)
       buffer.b = fct_b(...)
       buffer.c = fct_c(...)
       some_other_function(...)
    finally :
       buffer.a = old_a
       buffer.b = old_b
       buffer.c = old_c

with :

    (let ((a (fct-a ...))
          (b (fct-b ...))
          (c (fct-c ...)))
      (some-other-function ...))

Note, that you have to guarantee that all temporary changes are popped
before you leave the scope and that you might get exceptions from some
functions called to compute the temporary values.

Of course, you could provide `buffer.push_context/buffer.pop_context`
functions but note that the variables in question might not all be
buffer-local. In Elisp, you don't have to know where they come from
(which is just as well, because that changes over time).

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From cnetzer at mail.arc.nasa.gov  Sun Apr  6 17:19:24 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 06 Apr 2003 14:19:24 -0700
Subject: Is there swap function in Python?
In-Reply-To: <b6q1so$1cm2$1@mail.cn99.com>
References: <b6q0n6$1c55$1@mail.cn99.com>
	 <3r119vosrdukh6til7bdmgredlhhehqg64@4ax.com>  <b6q1so$1cm2$1@mail.cn99.com>
Message-ID: <1049663963.1698.42.camel@sayge.arc.nasa.gov>

On Sun, 2003-04-06 at 13:16, Xin Liu wrote:
> >
> > What do you mean by the swap function? What should it do
> > precisely?
> >
> > -- 
> > Ren? Pijlman
> It is a function like swap(a, b) . Use to swap the value of a and b.

Python doesn't work this way.  Python references refer to an object. 
You can swap references, but you can't (in general) swap the memory
contents of objects.  Nor do you need to or want to.

In C++ it makes sense to talk about swapping the "value of a and b".  In
Python, the value of a and b would not change, but the labels in the
current namespace (ie. a and b) can be swapped.  The objects themselves
would remain the same.

So rather than calling a functions swap(a, b), you just do a,b = b,a

In general, python functions do not mess with their caller's namespace,
so it could not operate like the C++ swap() with references.

So, this makes sense in Python:

a = 1
b = 2

def swap(t1, t2):
    return t2, t1

a,b = swap(a, b)   # After this point, a == 2 and b == 1


But there is not way (other than abusing globals or the module
namespace) to do it like this:

a = 1
b = 2

def swap(t1, t2):
    t2, t1 = t1, t2
    return

swap(a, b)
# After this point, a == 1 and b == 2.  The calling namespace is
# not changed.


Using classes, you can create namespaces that can be manipulated by
callable objects (the class instance's methods), and in that case, it
*may* be meaningful to have a swap() method.  But still, it is only
manipulating a namespace within which the objects are referred to, not
the swapped objects themselves (or their contents in machine memory). 
The below is illustrative only, not meant to be used for a practical
use:


class Pair:
    def __init__(self, t1, t2):
        self.t1 = t1
        self.t2 = t2

    def get(self):
        return self.t1, self.t2

    def swap(self):
        self.t1, self.t2 = self.t2, self.t1


a = 1
b = 2
pair = Pair(a, b)
pair.get()   # returns 1,2
pair.swap()  # changes the namespace of the pair object
pair.get()   # returns 2,1

a == 1
b == 2       # The a and b labels did NOT change


-- 
Bay Area Python Interest Group - http://www.baypiggies.net/

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From mertz at gnosis.cx  Tue Apr  8 16:53:45 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 16:53:45 -0400
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
 <+6xk+kKkXQeJ092yn@gnosis.cx> <mailman.1049829736.3719.python-list@python.org>
 <45yk+kKkXAlY092yn@gnosis.cx>
Message-ID: <Zbzk+kKkXULM092yn@gnosis.cx>         

"Mike C. Fletcher" <mcfletch at rogers.com> wrote previously:
|>    '"%s"' % repr(aString)[1:-1]
|Except when there's a " character in the string ;) .

Ahh... what I MEANT to write was, of course:

    '"%s"' % repr(aString)[1:-1].replace('"',r'\"')

Emphasizing the "not pretty" part further.  (I still think it is
prettier than Mike's though :-)).

Btw. The OP wrote "string", so obviously I am not interested in
repr(u"x").

Yours, Lulu...





From exarkun at intarweb.us  Fri Apr 11 14:36:31 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 11 Apr 2003 14:36:31 -0400
Subject: Python and Schools
In-Reply-To: <v9e21mhemm8c2@corp.supernews.com>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com>
Message-ID: <20030411183631.GA483@meson.dyndns.org>

On Fri, Apr 11, 2003 at 01:24:02PM -0500, Paul Watson wrote:
> Were you required to learn how to refine crude oil into petrol before you
> were taught how to drive a car?
> 
> Surely you must have studied the basic physics of fluid mechanics before you
> could be successful in the WC.
> 
> We need to teach students correct design priciples to build something
> greater than what already exists.  We will never get very far if we require
> everyone to start with a quark or atom.  Yes, of course we need some people
> who design silicon and create microcode.  They will learn the low-level
> details what they need to know as they need it.  Knowing great design and
> organization principles will enable them to make the most of it.
> 

  +1.

  Programming and computer science are related, but not the same.  Teaching
someone computer science and expecting this to make them a good programmer
is a mistake.

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 22 days, 15:01, 3 users, load average: 1.05, 1.11, 1.16




From aleax at aleax.it  Wed Apr 30 19:16:11 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 23:16:11 GMT
Subject: Raising objects
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it> <mailman.1051678851.21661.python-list@python.org> <OFLra.36816$K35.1143840@news2.tin.it> <mailman.1051719837.14426.python-list@python.org>
Message-ID: <%OYra.27334$3M4.759346@news1.tin.it>

Steven Taschuk wrote:

> Quoth Alex Martelli:
>> Steven Taschuk wrote:
>   [inability to raise new-style objects]
>> > I don't have a real use case either, but it does seem rather
>> > arbitrary.  Is there an implementation reason for this restriction?
>> 
>> In a sense, yes: there is no way to check if something "is a new-style
>> class" as opposed to "any type whatsoever" -- so one could end up
>> raising lists, tuples, strings (shades of the past...), whatever.
> 
> How about this?
> 
>     >>> def newstyle(obj):
>     ...     """Whether obj is an instance of a new-style class."""
>     ...     return hasattr(obj, '__class__') \
>     ...            and issubclass(obj.__class__, object)

This so-called "newstyle" accepts EVERYTHING except old-style classes.  E.g:

>>> hasattr(23, '__class__')
True
>>> issubclass((23).__class__, object)
True
>>>

So, it IS accepting "any type whatsoever" -- just not classic classes.
As I said, this includes lists, tuples, strings -- and also integers,
and so forth.


> It would produce counterintuitive results in such use if it were
> possible for a new-style class object to be an instance of an
> old-style class...  is that possible?

I guess you can try using a classic-class as a metaclass but it
would be a HUGE lot of work.  That's because of the semantic level
confusion in the classic object model that was removed in the
newstyle object model.  E.g.:

>>> class icmeta:
...   def __getattr__(self, name):
...     return getattr(type, name)
...   def __call__(*args, **kwds):
...     print 'BUH!'
...     return type.__call__(*args, **kwds)
...
>>> class weirdo(icmeta): pass
...
>>> ww = weirdo()
>>> ww
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: descriptor '__repr__' of 'type' object needs an argument
>>> ww()
BUH!
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "<stdin>", line 5, in __call__
TypeError: descriptor '__call__' requires a 'type' object but received a 
'instance'
>>>

See?  Calling the INSTANCE ww suprisingly ends up using the
METACLASS's __call__ -- eeek.

I very much doubt anybody would WANT to use this kind of thing.


> This function does report that strings and lists are new-style.
> Aren't they?

If by "new-style" you mean "anything EXCEPT classic classes" then
yes, of course -- strings, lists, numbers, etc, aren't classic
classes.  But then it's simpler to check for types.ClassType, no?


>> What I hope to see happen eventually is a spec that raise can use:
>> 
>>   -- any subclass of Exception (by that time a new-style class), OR
>>   -- for backwards compatibility any old-style class, OR
>>   -- for VERY backwards compatibility, any string
>> 
>> [and also INSTANCES as well as classes for the first two cases].
> 
> That seems reasonable.
> 
> But is there a problem with being able to raise anything?  Perhaps
> some class of easy-to-produce bugs that I'm not seeing?

When you raised strings (still supported for backwards compatibility)
the except statement worked by IDENTITY -- it didn't catch a string
that was equal but not identical.  This avoided deuced bugs such as:

in module primo.py:
error = 'error'

in module secondo.py, independently developed:
error = 'error'


in module terzo.py, using both the others...:

import primo, secondo

...

try:
    primo.blah()
    secondo.bluh()
except primo.error:
    print "error in blah"


where secondo.bluh might raise secondo,error -- if except had
worked by equality, it would have made an unwanted catch here.


So -- would all objects be caught by identity?  Consider the
horrible, perennial constraints this would impose on obvious
possible future optimizations -- why couldn't a very smart
compiler be able to "fold" equal immutable objects into a
single one?  Answer -- it couldn't do so without breaking
code using catch-by-identity as above.


Now we catch-by-class instead (when class instances are raised)
and the problem evaporates (except for "legacy" code as above).
Moreover, thanks to inheritance, we can easily catch large
"groups" of exceptions either jointly or severally.

Moreover, if what you raise is always derived by Exception, the
Exception base class can ensure all sort of nice error-reporting
etc -- I consider it a minor design bug that in the switch to
class-based exceptions the constraint that classes raises must
subclass Exception (directly or indirectly) wasn't introduced.


> What I'm imagining at the moment is this: suppose you can raise
> anything, and except clauses identify the exceptions to be caught
> by their class.  The only problem I can see with this is that it
> clashes with the present (backwards-compatible) treatment of
> raised strings:

Exactly.

>     try:
>         raise 'foo'
>     except 'foo':
>         print 'wah'
> wouldn't print 'wah', since 'foo' is not an instance of 'foo'.

It might or might not do so today, depending on the id() of the
two 'foo' literals (which Python doesn't guarantee either way).

> You'd have to do something like
>     try:
>         raise 'foo'
>     except str, e:
>         if e == 'foo':
>             print 'wah'
>         else:
>             raise
> which is ugly.

It would also have different semantics (== instead of is).


> So, suppose that strings were special-cased for backwards
> compatibility:
> 
>     def exceptmatch(exception, classtomatch):
>         """Whether an except clause specifying classtomatch should
>         catch exception.
>         """
>         if isinstance(classtomatch, str):
>             return exception == classtomatch
>         else:
>             return isinstance(exception, classtomatch)
> 
> (I ignore for simplicity except clauses which match more than one
> class.)
> 
> Is there a problem with this scheme?

The horrid complexity and specialcasing is problematic enough
for me.  Allowing such generality in what is raised buys you
nothing pragmatically useful AND you pay a price in complexity.
No thanks!!!

Allowing only [a] subclasses of (say) NewException, a hypothetical
new newstyle class; or [b] classic classes, and [c] strings, to
be raised, with b and c there only for backwards compatibility,
is FAR simpler, neater, easier to explain, and more pragmatical.


>> Exactly: even in my hoped-for new spec, you'd have to wrap this
>> in a subclass of Exception.  Facilitating the [ab]use of the
>> exception mechanism for generalized control flow is, I suspect,
>> NOT on the list of priorities in the BDFL's mind;-).  Feel free
>> to open a PEP to settle that beyond any doubt, though...;-)
> 
> Heh.  I imagine you're right.  I'd argue in favour of a scheme
> such as above chiefly on the basis of regularity (rather than for
> making my silly example work).

I think it's LESS simple and thus regular than my a/b/c scheme --
where a is the mainstream and b/c sad legacies of the past.


> I just don't see a good reason not to be able to raise new-style
> objects, so to me it's an arbitrary restriction and as such makes
> the language slightly uglier.  (It is, of course, a very minor
> issue; I hadn't even noticed that Exception is old-style until
> this thread.)

I think it's a restriction that is FAR from arbitrary: on the
contrary, in my opinion it would be quite useful and neat.


Alex




From aleax at aleax.it  Mon Apr 14 12:46:40 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 16:46:40 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain>
Message-ID: <QBBma.17307$T34.545916@news2.tin.it>

Robin Munn wrote:
   ...
> How is try ... finally "broken and fragile code"? How is this:
> 
>     try:
>         old_stdout = sys.stdout
>         new_stdout = cStringIO.StringIO()
>         sys.stdout = new_stdout
>         function_whose_output_I_want_to_capture()
>         do_something_with(new_stdout)
>     finally:
>         sys.stdout = old_stdout
> 
> in any way inferior to this (making up a semi-Pythonic syntax):
> 
>     with sys.stdout = cStringIO.StringIO():
>         function_whose_output_I_want_to_capture()
>         do_something_with(sys.stdout)
> 
> Yes, sample #2 is shorter. But it would require adding a new keyword to
> the language, which is a pretty heavy price to pay. And sample #2
> doesn't need a temporary variable -- but sample #1's temporary variable
> is a function-local variable, so speed and namespace-clutter costs are
> minimal.
> 
> What I really want to know is why you would consider sample #1 "broken
> and fragile code". I can't think of any circumstances where sample #1
> would fail to achieve its desired effect. Can you?

Just to play devil's advocate, I think sample #1 would behave very
strangely if its _first_ statement failed -- before propagating the
exception, it would (weirdly...) bind sys.sdout to some old value
for old_stdout (if any) -- and if there was no previous value for
old_stdout it would propagate the WRONG exception, namely the
UnboundLocalError met in the finally clause, rather than the
AttributeError (say) met in the first statement of the try clause.

In other words, I think your try clause encompasses too much -- a
common error, which only bites when some of the statements erroneously
placed into the try-clause happens to raise an exception.  Specifically
I believe that in this case you should optimally place the first THREE
statements, that are currently in your try-clause, OUTSIDE of it --
that's not critical, but, conceptually, until you're done rebinding
sys.stdout, there's no reason to ensure sys.stdout later gets bound
back to its previous (saved) value.

try/finally is quite error-prone that way, because somehow it SEEMS
we'd want to put into the try clause all the "initialization"
statements that, in fact, what we really want to have in it is what
comes AFTER "initialization".

Another difference between the two snippets, which I don't
understand, comes when function_blah_blah does its own rebinding
of sys.stdout.  In that case, the first snippet "ignores" the
rebinding (calls do_something_with with the value you had just
bound to sys.stdout BEFORE calling function_blahblah), the
second snippet "accepts" the rebinding (calls do_etc with the
value CURRENTLY bound to sys.stdout). Since I don't know what's
SUPPOSED to be happening under such circumstances, I can't tell
which of the two versions (if either) is wrong, but I do find
the difference puzzling.  Assuming the second hypothetical snippet
has the correct semantics, then, maybe the first one should be:

    old_stdout = sys.stdout
    sys.stdout = cStringIO.StringIO()
    try:
        function_whose_output_I_want_to_capture()
        do_something_with(sys.stdout)
    finally:
        sys.stdout = old_stdout

to achieve the same semantics -- specifically, what we want
to put in the try clause is just what would go in the with
clause's body if it existed, while the "saving the previous
state" (implicit in the hypothetical 'with') and the actual
initialization (the assignment that follows the with keyword
itself in the hypothetical case) need to be BEFORE the try
clause.  (Possibly there is some subtlety about what is
supposed to happen if cStringIO.StringIO rebinds sys.stdout
and then raises an exception?  In that case, then maybe
the second assignment I'm putting here before the try: needs
to go inside the try clause).  I think this kind of issues
shows try/finally isn't quite as simple to use as one might 
hope, so, perhaps, thinking of potential alternatives to
try/finally isn't necessarily a bad idea.


Alex




From member7758 at dbforums.com  Wed Apr 30 13:45:35 2003
From: member7758 at dbforums.com (andyabel)
Date: Wed, 30 Apr 2003 17:45:35 +0000
Subject: Using a C-header file in Python
Message-ID: <2825864.1051724735@dbforums.com>

I have a binary data file that I'd like to read from Python.
The file is complex and is described by a 3rd-party C-header file with
definitions like the following:

#define  TRANS_LEN                      64
struct trs_general_
{
    unsigned char opcode;
    unsigned char sub_opcode;
    unsigned char data[TRANS_LEN - 2 - sizeof(struct trans_tail_)];
    struct   trans_tail_  tail;
}
(And trans_tail_ is also defined as a struct in the file etc.)

Can I "include" this C-header file in my Python program (without
altering the header file) so I can read the file using standard Python
filehandling functions.

Thanks! (Please keep the answer simple, I'm a newbie to Python :-)

--
Posted via http://dbforums.com



From gh at ghaering.de  Wed Apr 16 05:16:30 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 16 Apr 2003 11:16:30 +0200
Subject: Any good donothing?
In-Reply-To: <slrnb9q42j.10n.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix>
References: <slrnb9p2vf.t3.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix> <mailman.1050453505.22283.python-list@python.org> <slrnb9q42j.10n.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix>
Message-ID: <3E9D1F6E.4060002@ghaering.de>

Dialtone wrote:
> * Gerhard H?ring <gh at ghaering.de> [160403 00:37]:
>>From what I remember other such solutions just use pipes to pipe the
>>output of cdda2wav into the mp3 encoder. No synchronization problems,
>>then.
> 
> Uhmmm... I tried to use pipes but I didn't managed it since it didn't
> take the right input file to encode.

There sure must be a way to use pipes, i. e. have cdda2wav write to 
stdout and the mp3 encoder read from stdin.

If one of them is really so broken as to not allow this, you can create 
a "fake" file for this. It's called a fifo. You can use mkfifo(1) or 
os.mkfifo from Python.

-- Gerhard





From sholden at holdenweb.com  Mon Apr  7 19:19:19 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Mon, 07 Apr 2003 23:19:19 GMT
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <v938h26mdl9s8f@corp.supernews.com> <h469m-4et.ln1@news.lairds.org>
Message-ID: <XHnka.22636$It5.16795@news2.central.cox.net>

"Kyler Laird" <Kyler at news.Lairds.org> wrote in message
news:h469m-4et.ln1 at news.lairds.org...
> claird at lairds.com (Cameron Laird) writes:
>
> >I exaggerate a bit.  There are indeed legitimate roles
> >for obfuscating technologies.  Managers think they're
> >far more common than is true, though, and programmers
> >far, FAR overestimate the objective value of source
> >code.
>
> Hey, I program and I think source code is *extremely*
> valuable!  It's rare that I can be convinced to use
> software for which I can not get the source code.  I
> certainly wouldn't ever want to *depend* on software
> that I can't control.
>
> What kind of business *pays* to become dependent on
> software they don't control?!
>

Can we say "Microsoft customers"? "Oracle customers"? ... the list is
endless!

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From skip at pobox.com  Tue Apr  8 12:34:52 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 8 Apr 2003 11:34:52 -0500
Subject: exceptions == errors?
In-Reply-To: <1049817240.3158.883.camel@utonium.pdi.com>
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
        <16018.5597.933725.584314@montanaro.dyndns.org>
        <1049817240.3158.883.camel@utonium.pdi.com>
Message-ID: <16018.64044.25718.688696@montanaro.dyndns.org>

    Jim> It's always been my impression that exceptions impose considerable
    Jim> performance overhead and therefore should be used to handle only
    Jim> the truly exceptional cases; for any predictable case, one should
    Jim> perform validation and/or error-handling outside the exception
    Jim> mechanism.

Well, yeah, there's a performance hit.  There is, after all, no free
lunch. ;-) The hit is minimal when no exception is raised, and substantially
higher when an exception is raised.  You have to balance the cost against
the clarity of the code.

Here's a trivial example.  If you expect most of the time that an object
will have a particular attribute, it's generally cheaper to just access it
within a try/except statement like so:

    try:
        x += obj.attr
    except AttributeError:
        pass

If, on the other hand, you expect that most of the time the attribute won't
exist, it's generally cheaper to call hasattr():

    if hasattr(obj, "attr"):
        x += obj.attr

Also, this whole thing changes across versions as different features are
added to the Python runtime and different optimizations are made to it.
What's true in 1.5.2 or 1.6 may well not be true in 2.3.  You have to try it
and see.

Here's some timeit output against Python CVS (just cvs up'd today since
Jeremy's and Tim's recent foray into the garbage collector).  The first set
of trials tries both ways to test for the nonexistent attribute
str.__nonzero__:

    >>> t = timeit.Timer(stmt="try:\n  str.__nonzero__\nexcept AttributeError:\n  pass")
    >>> t.timeit()
    17.419214010238647
    >>> u = timeit.Timer(stmt="if hasattr(str, '__nonzero__'):\n  pass")
    >>> u.timeit()
    4.5997350215911865

The next shows access to int.__nonzero__, which does exist:

    >>> t = timeit.Timer(stmt="try:\n  int.__nonzero__\nexcept AttributeError:\n  pass")
    >>> t.timeit()
    1.469681978225708
    >>> u = timeit.Timer(stmt="if hasattr(int, '__nonzero__'):\n  pass")
    >>> u.timeit()
    2.2288650274276733

Skip




From roy at panix.com  Tue Apr 15 07:39:31 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 15 Apr 2003 07:39:31 -0400
Subject: Built-ins vs. modules
References: <mailman.1050378626.14581.python-list@python.org>
Message-ID: <roy-EC4C93.07393015042003@reader1.panix.com>

Paul Berkowitz <berkowit at silcom.com> wrote:
> Why import the string module if you can do the same things anyway? The page
> of built in String Methods bears an uncanny resemblance to the page on the
> string module. Is there a historical reason? Are the built in type methods
> quite recent?

The string module came first.  Later (I forget exactly which release, 
but I'm guessing either 2.0 or 2.1) the string type was overhauled to 
make most of the functionality of the string module into built-in string 
methods.

The string module is basicly obsolete at this point.  All new code you 
write really should use the string methods.



From jmiller at physics.purdue.edu  Thu Apr 10 14:52:26 2003
From: jmiller at physics.purdue.edu (Jason Miller)
Date: Thu, 10 Apr 2003 14:52:26 -0400
Subject: pickle and security
Message-ID: <pan.2003.04.10.18.52.25.740636@physics.purdue.edu>

As I understand it, the major security issue with unpickling untrusted
sources is that it may cause python to instantiate objects, calling
constructors that could do just about anything.  If I only want to
unpickle objects that are not class instances, and (using cPickle) I set
find_global to None, are there any security concerns that remain? It would
be very nice if doing this were sufficient to make it safe to unpickle
from, say a socket, but I would like to know if that is just a bad idea.


     -Jason



From exarkun at intarweb.us  Thu Apr 10 02:34:13 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 10 Apr 2003 02:34:13 -0400
Subject: Python/C API
Message-ID: <20030410063413.GA13598@meson.dyndns.org>

  How does one write generators in C, or does one not?

  Jp

-- 
1.79 x 10^24 furlongs per picofortnight - It's not just a good idea, 
			It's the law!
-- 
 up 21 days, 3:01, 5 users, load average: 2.15, 2.20, 1.79




From mwh at python.net  Wed Apr  9 06:14:41 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 9 Apr 2003 10:14:41 GMT
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org> <kauu6b.f2d1.ln@boundary.tundraware.com> <7h3wui4528k.fsf@pc150.maths.bris.ac.uk> <mailman.1049874794.11476.python-list@python.org>
Message-ID: <7h3n0j03rla.fsf@pc150.maths.bris.ac.uk>

James Gregory <james at anchor.net.au> writes:

> On Tue, 2003-04-08 at 18:27, Michael Hudson wrote:
> > Tim Daneliuk <tundra at tundraware.com> writes:
> > 
> > > I am unclear, though, on whether a mathematician understands 1.0 to
> > > mean 1.0000000... (a single number) or the _neighborhood_ around 1
> > > with one digit precision.
> > 
> > I'm a mathematician, albeit not a numerical analyst, and I don't think
> > considering a float as the interval of real numbers to which it is the
> > closest floating point approximation is a particularly helpful point
> > of view.
> 
> To my mind, 1.0 is a member of R, 1 is a member of Z. But I do crazy
> things like think of "1/2" as the number that yields 1 when multiplied
> by 2. Not 0.5. Nor the quantity of apple you get when sharing an apple
> with one other person.
> 
> But it's context dependent, and that's one of the properties that having
> all "numbers" as objects affords you. It means that my software can tell
> if a particular number is a member of GF(5) (The Galois Field of 5
> elements), and then it can infer that "1/2" actually means 3 (3 * 2 = 6,
> 6 % 5 = 1, as per the above definition).
> 
> What does 1/2 mean to you in Z?

Nothing.

> Does it mean the same thing in R?

Well, no :-)

Cheers,
M.

-- 
  The above comment may be extremely inflamatory. For your
  protection, it has been rot13'd twice.
                           -- the signature of "JWhitlock" on slashdot



From blunck at gst.com  Wed Apr 30 22:22:05 2003
From: blunck at gst.com (Christopher Blunck)
Date: 30 Apr 2003 19:22:05 -0700
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com> <mailman.1051741964.7925.python-list@python.org>
Message-ID: <1147e466.0304301822.54858fcd@posting.google.com>

"Nik H." <nNOSPAMhaldimann at gmx.ch> wrote in message news:<mailman.1051741964.7925.python-list at python.org>...
> Christopher Blunck wrote:
> > What advice can you offer on spreading the testing bug around?  
> 
> Teach it at universities.

Nik - I see you post from a .ch domain.  Are you referring to Swiss
universities?


> This is my first year as a comp sci major. I've been doing unit testing 
> before, and I was surprised that we've actually been taught how to write 
> unit tests right from the start in programming classes, along with other 
> agile concepts like iterative development and pair programming (alas 
> using Java not Python). It's interesting to see what effect this has on 
> people that have done little or none real software development before - 
> unit testing comes absolutely natural to them.

Such a subtle yet potent point.  If I knew how to unit test while in
college, I'd kick the living crap out of every computer science
project I was ever handed.  Part of my difficulty in learning to
program (and this related to a significant amount of stress) was *not
knowing* why something did or did not work.  It was infuriated to see
a block of code I thought worked fail under a particular scenario. 
Was it the block of code or was it something that called that block of
code?  If only I had known about unit testing back then I could've
gotten straight As.  Well maybe not, but I certainly would've done
better than I actually did.  But that's a Monday morning quaterback
type statement to make, so I'll retract it.

I am happy you brought this point out tho, as I think it is quite
applicable...


> I read some accounts on universities still teaching the waterfall model 
> and stuff. So this post is meant as a heads up. There might be a new 
> generation of programmers with mad unit testing skills on the rise. <wink>

Can't wait for you all to get on the market.  ;-)



From staschuk at telusplanet.net  Tue Apr  8 16:36:00 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 8 Apr 2003 14:36:00 -0600
Subject: Best way to make a weighted sum
In-Reply-To: <867ka46awj.fsf@ieee.org>; from godoy@metalab.unc.edu on Tue, Apr 08, 2003 at 04:34:36PM -0300
References: <867ka46awj.fsf@ieee.org>
Message-ID: <20030408143600.A1578@tibia.amotlpaa.bogus>

Some comments on your code, and an alternative approach.

Quoth Jorge Godoy:
  [...]
> class IEModulo11:

I don't know why you're using a class for this.  A simple function
which calculates (and/or verifies) the checksum seems more natural.

Consider that, to calculate the checksum for a given value of ie
with your class, I have to do this:
	calculator = IEModulo11(ie)
	calculator.EstipulaPesos('RJ')
	calculator.CalculaSoma()
	# now sum is in calculator.soma
What a pain.  With a simple function for this purpose, it's just:
	sum = checksum(ie, 'RJ')

And, despite the class name, you don't actually do a mod 11.  I
assume this is a bug.

>     def __init__(self, ie=None, uf=None):
>         
>         self.estados = ['CE', 'AP', 'RJ'] 
>         self.ie = ie
>         self.soma = 0

Why isn't the uf argument used?

>     def EstipulaPesos(self, uf=None):
>       """
>       First digit indicates the number position, the second indicates
>       it's weight.
>       """
> 
>       if uf in ['CE', 'AP']:
>             self.pesos = {8:9, 7:8, 6:7, 5:6, 4:5, 3:4, 2:3, 1:2}
>       else:
>             if uf in ['RJ']:
>                 self.pesos = {7:2, 6:7, 5:6, 4:5, 3:4, 2:3, 1:2}

If uf is not in ['CE', 'AP', 'RJ'], then self.pesos simply does
not get set; this will cause CalculaSoma to fail later, in a
mysterious way.  Better to have an informative exception raised at
this point, surely.

It's also more natural imho to have the weights in a dict keyed by
uf.  That dict can replace your self.estados, which I note is not
otherwise used.

Also, the weights are conceptually a sequence, not a map; why not
make them a list?

>     def CalculaSoma(self):
> 
>         ie = self.ie
>         iter = len(ie)
> 
>         for i in ie:
>             produto = i * self.pesos[iter]
>             self.soma = self.soma + produto
>             iter = iter - 1

Why not return the sum?  See above.

It's also not good form to use 'iter' as a variable name, since
'iter' is the name of a built-in function.

> What I want is to sum the products of a certain 'ie' number where each
> of its digits is multiplied by a certain weigth, depending on its
> position inside the number. (This is for some documents verification
> and 'ie' can vary in the amount of digits as well as its weigth can
> vary from one place to another --- I know this is a mess, but it's the
> way some Brazilian states did it...)

An alternative:

	import operator

	weights = {}
	weights['CE'] = weights['AP'] = [9, 8, 7, 6, 5, 4, 3, 2]
	weights['RJ'] = [2, 7, 6, 5, 4, 3, 2] # is this right?

	def checksum(ie, state):
		w = weights[state][-len(ie):]
		return reduce(operator.add, [x*y for x, y in zip(ie, w)]) % 11

Imho this is much simpler, more straightforward, and easier to use.

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From nanthan79 at lycos.com  Thu Apr 10 03:47:50 2003
From: nanthan79 at lycos.com (Pathmenanthan Ramakrishna)
Date: Thu, 10 Apr 2003 07:47:50  0000
Subject: Python Interrupts & Exceptions! Codes
Message-ID: <NEOJIEJEFHCKGEAA@mailcity.com>

hi,

im new to python.currently im doing few testing of programs to analyze the pattern of system calls of different level of execution like ftp,ssh,telnet.Where can i get simple python codes that uses linux interrupts or exceptions e.g: service request?

i hope i get some resourses and guidelines...


thanks

Nanthan.R


____________________________________________________________
Get 25MB, POP3, Spam Filtering with LYCOS MAIL PLUS for $19.95/year.
http://login.mail.lycos.com/brandPage.shtml?pageId=plus&ref=lmtplus




From aleax at aleax.it  Fri Apr 25 19:54:20 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 23:54:20 GMT
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.37.08.286778@soul.net>
Message-ID: <MUjqa.8444$K35.255841@news2.tin.it>

<posted & mailed>

lost wrote:

> On Fri, 25 Apr 2003 22:51:12 +0000, Alex Martelli wrote:
> 
>> for each_file in os.dir('/var/path/to/files/'):
> 
> actually, not quite what i was looking for :)
> At this line, I get:
> AttributeError: 'module' object has no attribute 'dir'

os.listdir -- sorry, if I wrote os.dir it must have been a typo.


Alex




From aahz at pythoncraft.com  Fri Apr  4 10:32:06 2003
From: aahz at pythoncraft.com (Aahz)
Date: 4 Apr 2003 10:32:06 -0500
Subject: snakefarm?
References: <3cky9wjv.fsf@python.net> <b6k5ph$5rk$1@news.schlund.de>
Message-ID: <b6k8hm$isu$1@panix2.panix.com>

In article <b6k5ph$5rk$1 at news.schlund.de>,
Carsten Gaebler  <news at snakefarm.org> wrote:
>Thomas Heller wrote:
>>
>> Then I remembered the snakefarm. Is it possible to get access to the
>> snakefarm to test Python software on different platforms?
>> 
>> Whom do I have to ask?
>
>In case you mean snakefarm.org, this is my private homepage with some 
>(very few) Python thingies. And I'm not willing to give anyone access to 
>it. :-)

No, I think this is what Thomas is referring to:

http://lists.lysator.liu.se/mailman/listinfo/snake-farm
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From aussie2010 at yahoo.com  Sun Apr 13 06:08:43 2003
From: aussie2010 at yahoo.com (Chris)
Date: 13 Apr 2003 03:08:43 -0700
Subject: split encloser
References: <mailman.1050107298.22320.python-list@python.org>
Message-ID: <435d40e0.0304130208.1e51b4db@posting.google.com>

"Tiller, Jason" <jtiller at sjm.com> wrote in message news:<mailman.1050107298.22320.python-list at python.org>...
> Hi, "Aussie", :)
> 
> Now, to be totally honest, I kicked myself after Robin pointed out that you
> can get a CSV add-on module for Python.  DOH!  ::sigh::  Of course something
> like this is available - this is hardly the first-ever application of CSV
> (or TSV or whatever)...
> 
> I apologize if the commented regex wasn't detailed enough to completely
> explain its function.  If you have any  questions, don't hesitate to ask.  I
> can tell that the base regex I described in my previous message doesn't meet
> all the possible situations that a more sophisticated CSV module might.
> 
> If you're just interested in understanding regex'es for potential
> application in your own scripts, the regex chapter in "Programming Perl" is
> /very/ good, IMHO.  There's always Jeffrey Freidl's "Mastering Regular
> Expressions" - all of the solid concepts are there, regardless of language.
> I apologize that I can't really offer insight into Python-based regex
> processing - I just haven't studied it well, and the Programming Python
> (1.3) book that I have devotes only five or so pages to the entire topic
> ("Programming Perl"'s regex chapter is 70 pages, and "Mastering..." is 308
> pages).  I would hope that the more modern "Programming Python" does a
> better job...
> 
> However, for solving your particular problem, I encourage you to take a look
> at the CSV module that Robin mentioned in his post.
> 
> Good luck!
> 
> ---Jason
> 
>
Jason, 
Sorry not to be clear. I didn't have a problem with your explanation,
it's just that my brain freezes trying to understand regular
expressions (beyond the basics). I have the book you mention and maybe
I can dust it off and give it a go!



From martin at v.loewis.de  Sat Apr 12 04:51:25 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Sat, 12 Apr 2003 10:51:25 +0200
Subject: Is curses thread-safe?
In-Reply-To: <3E97204A.5030106@sympatico.ca>
References: <3E97204A.5030106@sympatico.ca>
Message-ID: <b78k26$8o6$00$2@news.t-online.com>

Brian Alexander wrote:

> Is the curses module thread-safe? Can I do this:
> 
> class windowThread(threadClass):
>  def __init__(self):
>   threadClass.__init__()
>   self.thisWindow = curses.newwin()
>   ...
> 
> ... and expect every thread created to behave well during calls the the 
> curses I/O routines?

Yes, curses is protected by the global interpreter lock: No two calls
into curses are allowed simultaneously.

Of course, if you have interleaving calls from two threads, you may 
notice strange effects (e.g. if both change the cursor position, the
second call takes precedence)

Regards,
Martin




From mcfletch at rogers.com  Thu Apr  3 11:18:55 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Thu, 03 Apr 2003 11:18:55 -0500
Subject: [wxPython] wxMDIChildFrame
In-Reply-To: <b6hl4f$hfj$1@demo2.univ-lyon1.fr>
References: <b6hl4f$hfj$1@demo2.univ-lyon1.fr>
Message-ID: <3E8C5EEF.8050207@rogers.com>

Normally the best place to ask questions about wxPython is on the 
wxPython-users list.

Regarding your question (this is going to sound pedantic, 'cause I'm 
trying to show how you would solve the problem, more than the particular 
solution):

    The problem you're describing can be thought of this way:

        What is the current focus of the application/user when they are
        issuing this command via the menu?  That is, which particular
        view/window should be considered (part of) the "context" of the
        action?

    Current GUI libraries, including wxPython, embody a concept called
    "focus", which is essentially a pointer to the currently-active
    control/window (there are all sorts of problems with the "single
    focus" model when you get into multi-modal interfaces, but you
    probably don't care about that.  Single focus is the norm in WIMP
    interfaces, and wxPython is definitely a WIMP interface engine).

        * wxWindow::FindFocus()

    Note also that windows/controls are able to retrieve references to
    their parents:

        * wxWindow::GetParent()

    Then realise that you can test whether a particular window is an
    instance of your image-holding-frame just by doing a standard Python
    isinstance(...) test. 

    With that, you can walk up the sequence of parents of the
    currently-focused control looking for a window of your
    wxMDIChildFrame class.  That's then usable as (part of) the context
    of your menu action.  You can see this pattern in a number of
    places, it's dirt-common.  I've built a few of these mechanisms over
    the years, works quite nicely, however...

    Consider that this is a very common operation which is critical for
    the proper operation of MDI interfaces and wonder why people
    wouldn't have a built-in method for doing this stuff ;) :

        Then, poke around in the wxMDIParentFrame documentation for
        "GetActiveChild"

Smile so the depressed half of the world considers you a threat :) ,
Mike


Roro wrote:

>Hello everybody,
>
>I want to create a window with others windows inside : My goal is to open
>several Images and apply filters on it.
>
>In concrete terms, my main window is with a bar menu : I click on  'Open
>File' Item and select images witch I wanna see.
>And I open  Images in a window by image, and I use wxMDIParentFrame and
>wxMDIChildFrame to create my windows.
>
>
>To create Windows, I use these lines :
>
> def OnNewWindow(self, evt, ):
>        / Open Dialog window witch ask witch files to open
>        /
>        self.winCount = self.winCount + 1
>        self.win = wxMDIChildFrame(self, Self.winCount+100, "Child Window:
>%d" % self.winCount)
>        / ...
>        /
>        self.win.Show(true)
>
>The problem comes now, I want to select a window and apply a filter on it
>(filters are in my menu bar). So, I want to select image's window, click in
>the menu bar the filter and see the result in the  selected window.
>I can only apply my filter on the last opened window. the others windows
>become ghost-windows (I can see them but i can't kill them). I think, it's
>because 'win' points on the last window and run over the previous created
>window.
>I want to manage my windows, but I didn't find any solution.
>
>Is it possible to create an indirection for 'win', by passing an argument
>with the new name for the window  ?
>Can I use 'Self.winCount' to refer to my ghost-windows ? Using their ID
>number ?
>
>
>Thanks in advance for any help.
>
>
>  
>

-- 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From danny at iamlearning.com  Wed Apr  2 16:39:37 2003
From: danny at iamlearning.com (danny at iamlearning.com)
Date: Wed, 2 Apr 2003 15:39:37 -0600 (CST)
Subject: How to compile a python script to *.pyo
In-Reply-To: <3E8B53BA.DBB0C84C@engcorp.com>
References: <mailman.1049309105.22947.python-list@python.org>
        <b6fj57$626$1@news-reader11.wanadoo.fr>
        <3E8B53BA.DBB0C84C@engcorp.com>
Message-ID: <3045.143.166.82.240.1049319577.squirrel@iamlearning.com>

> Philippe Lafoucri?re wrote:
>>
>> danny at iamlearning.com wrote:
>>
>> > I'm new on this, please bear with we.
>> >
>> > Let's say I have this script:
>> >
>> > --------------------------
>> > #!/usr/bin/env python
>> > x = 'Hello Python'
>> >
>> > print x,
>> >
>> > --------------------------
>> >
>> > How can I compile this script to *.pyo?
>> >
>> > Thanks in advance,
>> >
>> > Danny
>>
>> python -O yourscript.py
>
> Generally this doesn't work.  You have to import
> a module (or use other techniques) for it to be compiled.
> Python does not store a .pyc or .pyo file for the main
> module, perhaps surprisingly.
>
> -Peter
> --
> http://mail.python.org/mailman/listinfo/python-list


What are other techniques I can use?
I want my users run *.pyo or *.pyc files instead of *.py.

-Danny







From johnroth at ameritech.net  Wed Apr  2 07:22:57 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 2 Apr 2003 07:22:57 -0500
Subject: ternary operator
References: <33b491f.0304011848.d87dcc6@posting.google.com>
Message-ID: <v8ll9ngl81o682@news.supernews.com>

"daniel w. moore" <dan65536 at hotmail.com> wrote in message
news:33b491f.0304011848.d87dcc6 at posting.google.com...
> anyone heard about the results ofPEP 308 voting?  What do people think
> of this, particularly Pythonic Internals and Syntax Parsing experts (i
am not one!).

I don't belive Guido has said what (if anything) he's going to do.
Right now, he's busy with release 2.3. When he makes up his mind,
you'll see it in the PEP 308 writeup. Either he'll reject it, or he'll
put
in the way he intends to implement it. (At least, I hope he will.)

John Roth
>
> -dan





From amk at amk.ca  Tue Apr  1 16:21:02 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Tue, 01 Apr 2003 15:21:02 -0600
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de>
Message-ID: <CO6cnfDQxO4jnxejXTWc3Q@speakeasy.net>

On Tue, 01 Apr 2003 13:58:42 +0200, 
	Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote:
> I think about moving from Zope to
> ZODB and (Quixote or Twisted).

You're mixing levels here.  ZODB is an object database; Zope is for content
manager and application development, built on top of that object database.
Twisted is a framework for writing network services; Quixote is a framework
for writing HTTP/web applications, and can be hosted on top of Medusa or
Twisted.  (See the March archive of the quixote-users list for some example
code to run Quixote applications on top of Twisted.)

So the choices are more like:
   ZODB or SQL
   Apache or AOLserver or Twisted or Medusa 
   Quixote or Zope or Webware or ...
   
Zope provides content management features that Quixote and Webware don't.
You could write a content manager in Quixote or Webware, but no one has done
so yet, as far as I know.

--amk



From blunck at gst.com  Wed Apr 30 22:15:03 2003
From: blunck at gst.com (Christopher Blunck)
Date: 30 Apr 2003 19:15:03 -0700
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com> <3ea624c6.0304301049.52c4027c@posting.google.com>
Message-ID: <1147e466.0304301815.75ddb614@posting.google.com>

rcs at russellsalsbury.com (Russ Salsbury) wrote in message news:<3ea624c6.0304301049.52c4027c at posting.google.com>...
> blunck at gst.com (Christopher Blunck) wrote in message news:<1147e466.0304290638.287687f6 at posting.google.com>...
>  
> > What advice can you offer on spreading the testing bug around?  
> 
> It's very much a cultural issue.  First and foremost you have to have
> to use it yourself.  If you are an indifferent user, you can't
> convince others to use it.  Be consistent, even for trivial cases.

Yes, but how did you combat the allegation of being "difficult" or
"not a team player".  There are lots of labels that unsympathetic
coworkers can apply to you to subvert your methodology proposals. 
I've found that reputation wins out when it gets down to this level
(My code has unit tests and only has 2 bugs reported so far.  What
about you?  Oh, you don't have unit tests?  Wow.).  What has been your
experience?


> If you are a manager, find an evangilist in your group.  Get some
> successes before you impose the full rigor on everyone.  You don't
> need a revolt at the very beginning.  That said, requiring tests for
> checkin should be done from the beginning.

My manager and I now use the phrase "I believe I hit a vein" when
describing the infectious nature of unit testing.  When you sit down
with a reasonable person (even if they are 20+ years older than you)
and explain the benefits of unit testing, they (assuming they are
rational) usually jump on board (warning:  this is a generalization). 
Having the manager as the catalyst (the person that says 'Thou must
write unit tests') is a great motivator, but rational and level headed
developers take to it quite easily.  What have you done in situations
where you have NOT had rational and level headed developers AND you
lacked the support of your manager and/or team lead?  (This to me
seems like an impossible situation and a change of position seems
required.  But I'm interested in hearing your opinion.)


> Another possible ally is the QA manager, who could enforce the
> requirement for tests before checkin.  S/he should also include unit
> tests in the regression/acceptance tests.

Our QA manager has been influential in coming on board with the unit
testing initiative I've proposed.  But this isn't always the case (QA
engineers sometimes do not recognize the value of unit tests).  I'm
interested in what you've done to overcome that situation.


> > What practices did you follow?  
>
> I recruited an eager and talented QA tester who wanted to move into
> development to set up the JUnit framework and become the first
> user/evangelist.

What questions did you ask to determine if the candidate fit your
needs?  I have my own opinions of what questions to ask, but I'm
interested in what you asked since you seem to indicate that you had
some success.


> > Did you beat people over the head with it until they wised up?  
> 
> A developer may be able to do that, but a manager usually fails at it.
>  There are too many ways for the disgruntled to sabotage it, esp. in
> larger organizations.

A good point to make.  I'll keep it in mind.


> > Were you softer in your approach?  
> 
> Apparently not soft enough, because I got canned a few months later.
> ;-)
> Actually, it worked in my group, but I was chartered to improve the
> development practices of the entire company and when the President was
> replaced and the VP of Engineering and the Chief Architect left, the
> disgruntled had their way.  Pythonistas do not always triumph. ;-)

People of moderate intelligence always triumph over the mediocre 9 -
5'ers.  Our relative importance (like the market) fluctuates, but in
the long haul, our practices are sound.  I've never heard anyone say
that unit testing is a Bad Thing (tm.).  But, I've heard lots of
people explain away *not unit testing* because it was *too difficult*,
*too time consuming*, *not rewarding*, or .. (again) *too difficult*. 
To those I'd say let them sink with their own ship.  While we
pythonistas may go down with the ship, we were the smart souls who
came on board with a life preserver and were able to float to safety
when everyone else drowned and put *yet another* failed project on
their resume.  Don't be deterred - we are in the right, even if the
rest of the jabroni's don't think so.  :)


> > What were the long term implications?
> 
> The company went bankrupt a few months later. :-)

Wow - you are telling me that a company that did not make an
investment in quality software but instead focused on reckless feature
building went under?!  Wasn't that the same thing that happened to
thousands of companies between 1996 and 2001?!  What a surprise. 
Might be worth pointing out.  ;-)


> The long term implications:  no matter what happens you are a better
> programmer if you do rigorous automated(!) unit testing.  Your company
> produces a better product and has a better chance of long term success
> if everyone does.

And you learn a great deal during unit testing.  Someone on my team
just remarked to me (after underestimating the significance of unit
testing) that the junit tree was as complex and required as much
design as the main code base.  This was flattering to me as it
displayed his recognition that unit testing is as important as regular
development and should be paid the same level of respect.  As you've
pointed out, those that make an investment in regressable unit tests
are the hero's that will be recognized long after they leave the
project.  It's only a shame that they are not there to see the fruits
of their own labor...

-c



From mertz at gnosis.cx  Mon Apr  7 22:07:57 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Mon, 07 Apr 2003 22:07:57 -0400
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com>
Message-ID: <97ik+kKkXUxY092yn@gnosis.cx>         

Tim Daneliuk <tundra at tundraware.com> wrote previously:
|As I understand it, integers and floats are distinct mathematical
|entities. A colleague of mine claims, that insfar as we use them in
|computing, ints are merely a proper subset of floats.

Floats are peculiar creatures.  They really are not much like Real or
Rational numbers, even though it is often convenient to pretend they
are.  The trick about floating point numbers is that although they are
extremely useful for representing real-life (fractional) quantities,
operations on them do not obey the arithmetic rules we learned in middle
school:  associativity, transitivity, commutativity; moreover, many very
ordinary-seeming numbers can be represented only approximately with
floating point numbers.  For example:

     >>> 1./3
     0.33333333333333331
     >>> .3
     0.29999999999999999
     >>> 7 == 7./25 * 25
     0
     >>> 7 == 7./24 * 24
     1

Tim knows this, of course, but it is worth emphasizing.  Moreover, a
quote the Timbot found for this list a couple years ago is worth
reiterating:

  Many serious mathematicians have attempted to analyze a
  sequence of floating point operations rigorously, but found
  the task so formidable that they have tried to be content
  with plausibility arguments instead.
  -- Donald Knuth (_The Art of Computer Programming_, Third Edition,
     Addison-Wesley, 1997; ISBN:  0201896842, vol.  2, p.  229):

Yours, Lulu...

--
Keeping medicines from the bloodstreams of the sick; food from the bellies
of the hungry; books from the hands of the uneducated; technology from the
underdeveloped; and putting advocates of freedom in prisons.  Intellectual
property is to the 21st century what the slave trade was to the 16th.




From whisper at oz.net  Tue Apr 15 14:25:44 2003
From: whisper at oz.net (David LeBlanc)
Date: Tue, 15 Apr 2003 11:25:44 -0700
Subject: Visio like library for Python?
In-Reply-To: <3e9bdd3d.100422968@news.online.de>
Message-ID: <GCEDKONBLEFPPADDJCOEKEIPJIAA.whisper@oz.net>

<snip>
> >> I would like to draw UML diagrams.
> >
> >Maybe not quite what you asked but you might want to look at
> >ObjectDomain which is a UML diagram editor that is fully scriptable in
> >Python (or rather Jython).
> 
> I would prefer an Open Source product.
> 
<snip>
> 
> Regards
> 
> 
> Johannes
> 

PyUT - 100% python.

http://pyut.sourceforge.net/

Dave LeBlanc
Seattle, WA USA




From vlindberg at verio.net  Wed Apr 30 18:29:55 2003
From: vlindberg at verio.net (VanL)
Date: Wed, 30 Apr 2003 22:29:55 GMT
Subject: fcntl in windows
In-Reply-To: <gsh0bvkcan4tsr4kc3m0s4tcreh6tuq780@4ax.com>
References: <k2rvav06nhjgtr578fa5ostftfl69rethp@4ax.com> <mailman.1051717847.12855.python-list@python.org> <HSSra.42028$K35.1214886@news2.tin.it> <7h33cjz51y0.fsf@pc150.maths.bris.ac.uk> <gsh0bvkcan4tsr4kc3m0s4tcreh6tuq780@4ax.com>
Message-ID: <D7Yra.11797$8Z5.69632@sea-read.news.verio.net>

See portalocker... that does some of it for both windows and Unix.

VanL




From b.maryniuk at forbis.lt  Fri Apr 25 14:19:23 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Fri, 25 Apr 2003 20:19:23 +0200
Subject: What means exactly "Memory error"?
In-Reply-To: <20030425133405.GJ15695@frobozz.local>
References: <200304241406.20833.b.maryniuk@forbis.lt> <200304241917.47911.b.maryniuk@forbis.lt> <20030425133405.GJ15695@frobozz.local>
Message-ID: <200304252019.23702.b.maryniuk@forbis.lt>

On Friday 25 April 2003 15:34, Andrew Bennetts wrote:
> Why can't you use SAX if all you are doing is splitting by <HR>?  Sounds
> like a perfect job for SAX to me -- you just write the XML stream out to a
> file as you read it in, and switch to a new file everytime you hit <HR>.
> The memory requirement for this technique would be minimal.

Because I only SPLIT by <HR>! ;-) To split I need only doc.split(). But all 
the rest I do with DOM: too much changes in the HTML between different code 
chunks.

-- 
Regards, Bogdan

Never attribute to malloc that which can be adequately explained by
stupidity.
		-- From the .sig of joerg at raleigh.ibm.com (Joerg Pommnitz)





From pobrien at orbtech.com  Wed Apr 16 11:52:55 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 16 Apr 2003 10:52:55 -0500
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it> <slrnb9o9ob.493.rmunn@localhost.localdomain> <eX%ma.18570$G <7h31y02srj6.fsf@pc150.maths.bris.ac.uk>
Message-ID: <m2adeqh220.fsf@orbtech.com>

Michael Hudson <mwh at python.net> writes:

> Carl Banks <imbosol-1050441179 at aerojockey.com> writes:
> 
> > First, you're basing it on a bad example.  Rebinding sys.stdout is bad
> > programming, period.
> 
> Whoah there!  Says who?  Would you like to try using IDLE or any other
> IDE without binding sys.stdout?

Or how about one that includes a Python shell (like PyCrust) where the
user can rebind any of stdin, stdout, stderr, by entering the
appropriate code at the interactive prompt, but the shell itself needs
to keep on working?  If rebinding sys.stdout is bad, I'm happy to say
that PyCrust is bad, really bad.  ;-)

    def runsource(self, source):
        """Compile and run source code in the interpreter."""
        stdin, stdout, stderr = sys.stdin, sys.stdout, sys.stderr
        sys.stdin, sys.stdout, sys.stderr = \
                   self.stdin, self.stdout, self.stderr
        more = InteractiveInterpreter.runsource(self, source)
        # If sys.std* is still what we set it to, then restore it.
        # But, if the executed source changed sys.std*, assume it was
        # meant to be changed and leave it. Power to the people.
        if sys.stdin == self.stdin:
            sys.stdin = stdin
        if sys.stdout == self.stdout:
            sys.stdout = stdout
        if sys.stderr == self.stderr:
            sys.stderr = stderr
        return more

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From antoni3 at otenet.gr  Mon Apr  7 11:36:10 2003
From: antoni3 at otenet.gr (Antonis Kaklis)
Date: Mon, 7 Apr 2003 18:36:10 +0300
Subject: Parsing data
Message-ID: <b6s5lt$5c1$1@usenet.otenet.gr>

hello to everybody...

A have a file with values ordered by columns:
Say (e.g force values, in time steps of 0.0200sec)

   .2161473E-05   .2183854E-05   .2177804E-05   .2174099E-05   .2254829E-05
   .2145163E-05   .2322683E-05   .2112049E-05   .2092585E-05   .2069618E-05
   .2052460E-05   .2572402E-05   .2625725E-05   .2701999E-05   .2771539E-05
   .1884220E-05   .1864868E-05   .1835972E-05   .3150803E-05   .3245766E-05
   .3343058E-05   .3437211E-05   .3543562E-05   .1512718E-05   .1483842E-05
   .1440723E-05   .1408216E-05   .1363429E-05   .1340211E-05   .1291136E-05
   .1260124E-05   .4968635E-05   .5089609E-05   .5228032E-05   .5364493E-05
   .7569464E-06   .7437904E-06   .5947033E-05   .5593801E-06   .6339938E-05
   .6471278E-05   .2300983E-06   .1903562E-06   .1526042E-06   .7414982E-05
     ...            ...              ...          ...              ...

how can i parse the values (in order to proccess or plot or whatever else i
want)
of a structure like that?

Please help
Thanks in advance





From aleax at aleax.it  Fri Apr 25 19:52:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 23:52:46 GMT
Subject: Getting values from python interpreter from Java
References: <mailman.1051288050.12209.python-list@python.org>
Message-ID: <iTjqa.8440$K35.255695@news2.tin.it>

<posted & mailed>

Lichtenstein, Mark (Contractor) wrote:

> I would like to see how to get the value of a local variable from Java.
> In Java, I am saving the values of the local dictionary in a PyObject Java
> object.
> Using a PythonInterpreter object called "interp", I use setLocals and
> getLocals methods to
> get and restore the local dictionary.
> 
> interp.setLocals(localsObject);
> interp.getLocals(localsObject);
> 
> My question is from Java how does someone go into the PyObject contatining
> the locals dictionary to get the value of a local dictionary variable?

As I already answered on python-help, you can get the PyObject value
of a variable by calling localsObject.__getitem__("varname") from Java.

Please don't post questions to python-help AND the general Python
list too, choose just one or the other, thanks.


Alex




From exarkun at intarweb.us  Wed Apr  9 19:05:56 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 9 Apr 2003 19:05:56 -0400
Subject: Python application deployment?
In-Reply-To: <14bd1399.0304091426.1b08e005@posting.google.com>
References: <14bd1399.0304091426.1b08e005@posting.google.com>
Message-ID: <20030409230556.GC10833@meson.dyndns.org>

On Wed, Apr 09, 2003 at 03:26:09PM -0700, kkennedy wrote:
> Is there a way to deploy the Python "runtime" with your Python
> scripts?  I am only concerned with Windows (since most Linux distros
> would already have it installed).  I would like it to be smaller than
> the 7 MB download available from python.org, and be able to be
> installed transparent to the user.  In other words, it would be nice
> to be able to simply unzip the files needed to run my scripts without
> the user having to download and install the full Python windows setup,
> and then download and install my Python scripts seperately.
> 
> Are there any tools to make this doable by the common programmer?

  This question is asked at least once a week.  The FAQ and the newsgroup
archives will answer it for you.

  HTH,

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 20 days, 19:01, 1 user, load average: 0.21, 0.15, 0.05




From martin at v.loewis.de  Wed Apr 23 01:38:55 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 23 Apr 2003 07:38:55 +0200
Subject: SGMLParser questions
References: <igmbavg5ughls6ksiden2rgnael1s0ub8q@4ax.com>
Message-ID: <m3n0ihdb80.fsf@mira.informatik.hu-berlin.de>

Tung Wai Yip <tungwaiyip at yahoo.com> writes:

> 2. Does it supports namespace?

No.

> 3. handle_starttag() turn tags into lower case. This work perfectly
> for HTML. But can it works case sensitively?

No.

To get these features, you could subclass SGMLParser. However, you
probably have to copy a lot of code into your subclass.

Regards,
Martin




From cben at techunix.technion.ac.il  Wed Apr  9 12:19:10 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Wed, 9 Apr 2003 19:19:10 +0300 (IDT)
Subject: Visio like library for Python?
In-Reply-To: <3e93e50f.676072541@news.online.de>
References: <3e93e50f.676072541@news.online.de>
Message-ID: <Pine.GSO.4.51.0304091917090.1222@techunix.technion.ac.il>

Johannes Eble wrote on 2003-04-09:

> Hello community,
>
> is there a library available for Python that behaves similar to MS
> Visio? For example, I would like to create vector graphic objects and
> connect the with a line. If I move, flip or rotate a connected object,
> the line should follow the connection.
>
WxWindows features something called Object Graphics Library (OGL) that
ca nd othese things.  I've read OGL's demo source and it seems
relatively easy to program but I neer got to actually use it.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From jack at performancedrivers.com  Sat Apr  5 13:09:26 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Sat, 5 Apr 2003 13:09:26 -0500
Subject: Q: The roles of Python
In-Reply-To: <5LEja.3680$945.10664@tor-nn1.netcom.ca>; from waldbie@attglobal.net on Sat, Apr 05, 2003 at 01:08:24PM -0800
References: <5LEja.3680$945.10664@tor-nn1.netcom.ca>
Message-ID: <20030405130926.A22750@localhost.localdomain>

On Sat, Apr 05, 2003 at 01:08:24PM -0800, Carl Waldbieser wrote:
> I think Python offers a lot of advantages over the other two big choices in
> the Windows scripting world (JScript and VBScript)

Python is a full-blown programming language, JScript and VBScript are scripting
hacks for browsers.  Or to be kinder, single purpose mini-languages.

It would be more appropriate to compare python with perl.

-jack




From ianb at colorstudy.com  Sat Apr  5 14:30:27 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 05 Apr 2003 13:30:27 -0600
Subject: Pythonic way of web-programming
In-Reply-To: <a3b5b608.0304051049.7048d76@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
Message-ID: <1049571027.6023.59.camel@lothlorien>

On Sat, 2003-04-05 at 12:49, Giorgi wrote:
> The question is that the only way to establish Python as a language
> for corporate level web-programming is to have the standard approach
> if not introduced, then at least supported by the Python consortium...

I don't think that's entirely true.  It's difficult for *Python* to be a
language for certain areas of web programming (perhaps corporate), but
certain Python environments could still achieve that.  The corporation
would choose an environment, not a language, and if a good environment
is written in Python then indirectly they'll have chosen Python.

But I'm not saying it wouldn't be nice to have more cohesion... that's
just not the way it is now.  It's harder to make it happen, because
unlike Java or .NET, the Python community is not a top-down community. 
Guido has some authority, but I think he's almost as much a decision
process as a decider -- he decides on things where there's some level of
consensus in the community, and doesn't make dictates otherwise.  The
standard library has some authority, but its fairly neutral on these
things.  Besides that there's no authority at the moment.

The next step I can envision would be for Twisted to become the
canonical infrastructure, where higher-level web frameworks are all
built ontop of it.  In that way we can start to build consensus one
piece at a time.

  Ian






From Andreas.Schumann at schupag.ch  Thu Apr 10 06:06:08 2003
From: Andreas.Schumann at schupag.ch (Andreas Schumann)
Date: Thu, 10 Apr 2003 12:06:08 +0200
Subject: Problem passing LPVOID
Message-ID: <DNEJLDBGFEPPNGGFNGBECEJJCHAA.Andreas.Schumann@schupag.ch>

Hi

I'm writing my first extension to Python to use a Windows dll. Now I'm
having a problem to pass a parameter of typ LPVOID from Python to the dll.

Example:

The dll function is defined in the header file as follows:
DWORD dll_function (LPVOID  data);

data can be a pointer to anything (int, string, etc.). I.e. in Python I
should be able to use this function passing any type of data, e.g.:
res = dll_function(12)
res = dll_function('aaa')

How do I have to use PyArg_ParseTuple to get data to be passed to the dll
function? Do I have to use "O&" and write my own converter function?

Thanks
Andy Schumann

 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: winmail.dat
Type: application/ms-tnef
Size: 6024 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030410/ea0d8daa/attachment.bin>

From andy at reportlab.com  Mon Apr 14 21:30:49 2003
From: andy at reportlab.com (Andy Robinson)
Date: Tue, 15 Apr 2003 02:30:49 +0100
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>	 <mailman.1049530929.9225.python-list@python.org>	 <a3b5b608.0304051049.7048d76@posting.google.com>	 <mailman.1049571010.25625.python-list@python.org>	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>	 <mailman.1049653811.11491.python-list@python.org>	 <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com>
Message-ID: <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>

>...it might even be possible for such "super-frameworks" to paper over
>the fundamental but desirable differences between the underlying
>frameworks. For example, some people might want a threaded model,
>whilst others might prefer a process-forking model together with
>similarly convenient data sharing semantics.

Hear, hear.

I think it's sad that Python has this slogan "there's only one way
to do it", but that nowhere is this less true than web development.
At Python-UK I opined that this might be holding us back in comparison
to other languages and tools.  Anyone coming into Java or .NET
development gets a very clear feeling of "one way to do it".

We have a bunch of competing frameworks, but if we analyze them
into layers then they compete less, and a more compatible approach
must be possible.

First, let's note the half-a-dozen different templating schemes
available.  DTML, Woven, Quixote, YAPTU, preppy.

Second, it has always seemed to me that there is a very well
understood model of a web app consisting of a controller, Request,
Response and a few other things, handled in very similar ways by Java
servlets, ASP and WebKit.  This is the basic 'dispatch loop'. It would
be wonderful if this could be systematized in something like the 'db
api', with some interfaces defined in the standard Python library.  

Finally, there's a transport mechanism.  Here one could use
CGI, Twisted, Zope, BaseHTTPServer or whatever else.

Wouldn't it be nice if one day we could have some standardized
interfaces, so that one could switch between these without undue pain
and reduce the confusion people must face?

Best Regards,

Andy Robinson
ReportLab Inc.



From riemer at ppprs1.phy.tu-dresden.de  Tue Apr  8 07:47:55 2003
From: riemer at ppprs1.phy.tu-dresden.de (Tilo Riemer)
Date: Tue, 8 Apr 2003 13:47:55 +0200
Subject: Announce: apm.py 0.1.0
In-Reply-To: <komam-r2d.ln1@nb2.stroeder.com>
References: <mailman.1049750436.1395.python-list@python.org> <komam-r2d.ln1@nb2.stroeder.com>
Message-ID: <20030408114755.GA14666@ppprs1.phy.tu-dresden.de>

Hello Michael,

> What about ACPI? That became standard on some Linux distros (e.g. SuSE 
> 8.1+).

acpi.py is under way ;-)
I think I can publish the first version this week (only for Linux yet).

 
> Maybe it would be better to call the module powerman or similar to subsume 
> APM, ACPI and any other standard popping up.

ACPI and APM do the same things more or less but have nothing to do with
each other... perhaps it's useful to mix it if they provide the same
interface. but ACPI is very more powerful. In the best case APM is a subset
of ACPI...

Any suggestions?


Best regards, Tilo




From Vector180W at netscape.net  Tue Apr  8 03:01:42 2003
From: Vector180W at netscape.net (vector)
Date: Tue, 08 Apr 2003 17:01:42 +1000
Subject: pygtk install
Message-ID: <3E9273D6.7030500@netscape.net>

Im having problems installing pygtk on my windows XP machine.
I installed
pygtk-1.99.16.win32-py2.2.exe
from
http://www.pcpm.ucl.ac.be/~gustin/win32_ports/
and my test prog
import pygtk
pygtk.require('2.0')
import gtk
gives me this error
Traceback (most recent call last):
   File 
"G:\Development\Python\Lib\site-packages\Pythonwin\pywin\framework\scriptutils.py", 
line 301, in RunScript
     exec codeObject in __main__.__dict__
   File "E:\Md_prj\md_422\Code\Python\testgtk1.py", line 3, in ?
     import gtk
   File 
"G:\Development\Python\Lib\site-packages\gtk-2.0\gtk\__init__.py", line 
39, in ?
     import gobject as _gobject
ImportError: DLL load failed: The specified module could not be found.

I then tried
GTK+-Runtime-Environment-2.2.1.1.exe
and reloaded python etc all the same




From edo4 at emich.edu  Tue Apr  8 16:15:57 2003
From: edo4 at emich.edu (Ngoc)
Date: 8 Apr 2003 13:15:57 -0700
Subject: How to build Numeric in Debug mode
Message-ID: <99457f2.0304081215.214a5a9c@posting.google.com>

I want to use Numeric.array from Numeric module.  I can run python
with import Numeric in release but not debug mode. What is the command
to build Numeric module in debug version ?
>>python_d setup.py install doesn't work??
Thanks in advance.
Eve



From NineOfSix at gmx.de  Sat Apr 26 06:40:13 2003
From: NineOfSix at gmx.de (Axel Grune)
Date: Sat, 26 Apr 2003 12:40:13 +0200
Subject: MD5 to long?
References: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de>
Message-ID: <b8dnlo$8sco1$1@ID-126101.news.dfncis.de>

Shoot me! eval works well;-)

 >>> longnumber = eval( '0x' + 'fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8' )

Axel




From gregory_r_warnes at groton.pfizer.com  Sat Apr 12 11:46:08 2003
From: gregory_r_warnes at groton.pfizer.com (Warnes, Gregory R)
Date: Sat, 12 Apr 2003 11:46:08 -0400
Subject: PEP 754: IEEE 754 Floating Point Special Values
Message-ID: <D7A3CFD7825BD6119B880002A58F06C202F2C7F9@groexmb02.pfizer.com>

I've just submitted PEP 754, "IEEE 754 Floating Point Special Values",
included below, and would like feedback.

-Greg
================================

PEP: 754
Title: IEEE 754 Floating Point Special Values
Version: $Revision: 1.1 $
Last-Modified: $Date: 2003/04/12 13:39:34 $
Author: Gregory R. Warnes <warnes at users.sourceforge.net>
Status: Draft
Type: Standard Track
Content-Type: text/x-rst
Created: 28-Mar-2003
Python-Version: 2.3
Post-History:


Abstract
========

This PEP proposes an API and a provides a reference module that
generates and tests for IEEE 754 double-precision special values:
positive infinity (Inf), negative infinity (-Inf), and not-a-number
(NaN).


Rationale
=========

The IEEE 754 standard defines a set of binary representations and
algorithmic rules for floating point arithmetic.  Included in the
standard is a set of constants for representing special values,
including positive infinity (Inf), negative infinity (-Inf), and
indeterminate or non-numeric results (NaN).  Most modern CPUs
implement the IEEE 754 standard, including the (Ultra)SPARC, PowerPC,
and x86 processor series.

Currently, the handling of IEEE 754 special values in Python depends
on the underlying C library.  Unfortunately, there is little
consistency between C libraries in how or whether these values are
handled.  For instance, on some systems "float('Inf')" will properly
return the IEEE 754 constant for positive infinity.  On many systems,
however, this expression will instead generate an error message.

The output string representation for an IEEE 754 special value also
varies by platform.  For example, the expression "float(1e3000)",
which is large enough to generate an overflow, should return a string
representation corresponding to IEEE 754 positive infinity.  Python
2.1.3 on x86 Debian Linux returns "inf".  On Sparc Solaris 8 with
Python 2.2.1, this same expression returns "Infinity", and on
MS-Windows 2000 with Active Python 2.2.1, it returns "1.#INF".

Adding to the confusion, some platforms generate one string on
conversion from floating point and accept a different string for
conversion to floating point.  On these systems ::

    float(str(x))

will generate an error when "x" is an IEEE special value.

In the past, some have recommended that programmers use expressions
like::

    Inf = 1e300**2
    NaN = Inf/Inf

to obtain positive infinity and not-a-number constants.  However, the
first expression generates an error on current Python interpreters.  A
possible alternative is to use::

    Inf = 1e300000
    NaN = Inf/Inf

While this does not generate an error with current Python
interpreters, it is still an ugly and potentially non-portable hack.
In addition, defining NaN in this way does solve the problem of
detecting such values.  First, the IEEE 754 standard provides for an
entire set of constant values for Not-a-Number.  Second, the standard
requires that ::

    NaN != X

for all possible values of X, including NaN.  As a consequence ::

    NaN == NaN

should always evaluate to false.  However, this behavior also is not
consistently implemented.  [e.g. Cygwin Python 2.2.2]

Due to the many platform and library inconsistencies in handling IEEE
special values, it is impossible to consistently set or detect IEEE
754 floating point values in normal Python code without resorting to
directly manipulating bit-patterns.

This PEP proposes a standard Python API and provides a reference
module implementation which allows for consistent handling of IEEE 754
special values on all supported platforms.


API Definition
==============

Constants
---------

NaN
    IEEE 754 "Not a Number" value

Inf, PosInf
    IEEE 754 Positive Infinity value

NegInf
    IEEE 754 Negative Infinity value


Functions
---------

is_NaN(value)
    Determine if the argument is a IEEE 754 NaN (Not a Number) value.

is_Inf(value), is_PosInf(value)
    Determine if the argument is a IEEE 754 positive infinity value

is_NegInf(value)
    Determine if the argument is a IEEE 754 negative infinity value

is_Finite(value)
    Determine if the argument is an finite IEEE 754 value (i.e., is
    not NaN, positive or negative infinity)

is_Infinite(value)
    Determine if the argument is an infinite IEEE 754 value (positive
    or negative infinity)


Example
-------

(Run under Python 2.2.1 on Solaris 8.)

>>> import fpconst
>>> val = 1e30000 # should be cause overflow and result in "Inf"
>>> val
Infinity
>>> fpconst.is_Inf(val)
1
>>> fpconst.Inf
Infinity
>>> nval = val/val # should result in NaN
>>> nval
NaN
>>> fpconst.is_NaN(nval)
1
>>> fpconst.is_NaN(val)
0


Implementation
--------------

The reference implementation is provided in the module "fpconst" [1]_,
which is written in pure Python by taking advantage of the "struct"
standard module to directly set or test for the bit patterns that
define IEEE 754 special values.  Care has been taken to generate
proper results on both big-endian and little-endian machines.

The current implementation is pure Python, but some efficiency could
be gained by translating the core routines into C.


References
==========

See http://babbage.cs.qc.edu/courses/cs341/IEEE-754references.html for
reference material on the IEEE 754 floating point standard.

.. [1] Further information on the reference package is available at
   http://software.biostat.washington.edu/statsoft/snake/fpconst.


Copyright
=========

This document has been placed in the public domain.

 
..
   Local Variables:
   mode: indented-text
   indent-tabs-mode: nil
   sentence-end-double-space: t
   fill-column: 70
   End:


LEGAL NOTICE
Unless expressly stated otherwise, this message is confidential and may be privileged. It is intended for the addressee(s) only. Access to this E-mail by anyone else is unauthorized. If you are not an addressee, any disclosure or copying of the contents of this E-mail or any action taken (or not taken) in reliance on it is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately.




From jeremy at zope.com  Tue Apr 22 11:16:37 2003
From: jeremy at zope.com (Jeremy Hylton)
Date: 22 Apr 2003 11:16:37 -0400
Subject: Should I prefer an external database
In-Reply-To: <u3ckbymri.fsf@mail.afm.dtu.dk>
References: <u3ckbymri.fsf@mail.afm.dtu.dk>
Message-ID: <1051024596.12838.17.camel@slothrop.zope.com>

On Tue, 2003-04-22 at 04:11, Brian Elmegaard wrote:
> I am working on a script for organizing a program for a conference. I
> started out using a connection to mySQL, but then I thought that it
> would be easier to keep it within python. So I implemented a class:
> 
> class Paper:
>         def __init__(title,author,session):
>                 self.title=title
>                 self.author=author
>                 self.session=session
> I then have a list to which each new paper instance is appended. I can
> sort by any of the paper instances variables by using the
> decorate-sort-undecorate method.
> 
> I now have it working, but I probably could do better, so:
> 
> Is this a more pythonic way to implement this?

You've chosen a simple and obvious way to implement it, which sounds
pythonic to me.  (Note that you forgot to include self in the argument
list to __init__.)

You may find ZODB helpful for this project, because it allows you to
store your objects in a database without warping your application's data
model to fit in an SQL database.  There's a short guide to help get you
started at:  http://www.zope.org/Wikis/ZODB/guide/index.html

> To do a select I think I need to use for loops specifying the names of
> the variables, i.e:
> 
> for paper in paperlist:
>         if paper.session=='foo':
>                 print paper.title
> 
> Probably this might be done more elegantly, but how do I know which
> variables (database columns) the instance has?

Who put the instances in the list?  That is, I assume that you have
control over what kind of objects get put in the list.  So you should
know what attributes they have.

> Would I be better of with a list of list than with a list of class
> instances? 
> What about a class of classes? Could such be made and inherit from the
> list class to support index and sort?

A list of class instances seems like the clearest way to write it.  In
ZODB, you could create an index of objects keyed by the session variable
if you expected to do this kind of query a lot.

Jeremy







From Patrick.Bussi at space.alcatel.fr  Mon Apr  7 05:03:32 2003
From: Patrick.Bussi at space.alcatel.fr (Patrick.Bussi at space.alcatel.fr)
Date: Mon, 7 Apr 2003 11:03:32 +0200
Subject: Headers issue with urllib2 and ClientCookie (was : ClientCookie
	.read() failing on some servers )
Message-ID: <C1256D01.0031EBA1.00@vzmta01.netfr.alcatel.fr>

Hello,

Really sorry for a previous incomplete mail sent by error to the list.

I am concerned about a piece of soft which has different behaviour depending on
the web site whether I send it.

Here is :

[1] short test program, which simply creates an instance of urllib2 object, adds
headers to it, sends the request to the server and get the response. (uncomment
the one or the other line to use ClientCookie or urllib2)

[2] extract of the dump of results for www.python.org (correct)

[3] dump of results for urllib2 opening www.lycos.fr, showing an
urllib2.HTTPError: HTTP Error 302

[4] dump of results for ClientCookie opening www.lycos.fr

[5] tcpdump trace on the PPP link, showing that the headers seem not taken into
account, which could explain the failure.

Questions :

A/ does the wrong headers justify the failure in server response ?
B/ what is wrong in my test program headers ?


[1]----------snip----------
#! /usr/bin/env python
'''usage:
         $ python test.py www.python.org
'''

def test(h):
    '''h is the host name. Caution : no protection against mistakes
    '''
    import ClientCookie, urllib2, urllib
    from urllib2 import Request
    req = urllib2.Request('http://'+h)
    req.add_header('Host',h)
    req.add_header('User-agent',
'Mozilla/5.5.(X11;.U;.Linux.2.4;.en-US;.0.8).Gecko/20010409')
    req.add_header('Accept','*/*')
    req.add_header('Accept-Language','en')
    req.add_header('Accept-Encoding','gzip,deflate,compress,identity')
    req.add_header('Keep-Alive','300')
    req.add_header('Connection','keep-alive')
    print '\n'.join(['%s' %k for k in (req.get_full_url(), req.headers)])

    response=ClientCookie.urlopen(req)
#    response=urllib2.urlopen(req)
    print '\n', response.info()
    if response: print response.read()

if __name__ == '__main__':
    import os, sys
    print 'Linux', ''.join(['%s'*3 %os.uname()[2:]])
    print 'Python', sys.version
    try: h=sys.argv[1]
    except: sys.exit(__doc__)
    test(h)
----------snip----------

[2]----------snip----------
Linux 2.4.19-pb11#3 SMP Sun Mar 30 06:32:53 CEST 2003i686
Python 2.2.2 (#3, Apr  7 2003, 05:07:52)
[GCC 2.96 20000731 (Mandrake Linux 8.2 2.96-0.76mdk)]
http://www.python.org
{'Host': 'www.python.org', 'Accept-Language': 'en', 'Accept-Encoding':
'gzip,deflate,compress,identity', 'Connection': 'keep-alive', 'Keep-Alive':
'300', 'Accept': '*/*', 'User-agent':
'Mozilla/5.5.(X11;.U;.Linux.2.4;.en-US;.0.8).Gecko/20010409'}
Date: Sun, 06 Apr 2003 16:00:56 GMT
Server: Apache/1.3.26 (Unix)
Last-Modified: Fri, 04 Apr 2003 15:59:44 GMT
ETag: "5a7585-3876-3e8dabf0"
Accept-Ranges: bytes
Content-Length: 14454
Keep-Alive: timeout=15, max=100
Connection: Keep-Alive
Content-Type: text/html
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"><?xml-stylesheet
href="./css/ht2html.css" type="text/css"?>
[...cut...]
----------snip----------


[3]----------snip----------
Linux 2.4.19-pb11#3 SMP Sun Mar 30 06:32:53 CEST 2003i686
Python 2.2.2 (#3, Apr  7 2003, 05:07:52)
[GCC 2.96 20000731 (Mandrake Linux 8.2 2.96-0.76mdk)]
http://www.lycos.fr
{'Host': 'www.lycos.fr', 'Accept-Language': 'en', 'Accept-Encoding':
'gzip,deflate,compress,identity', 'Connection': 'keep-alive', 'Keep-Alive':
'300', 'Accept': '*/*', 'User-agent':
'Mozilla/5.5.(X11;.U;.Linux.2.4;.en-US;.0.8).Gecko/20010409'}
Traceback (most recent call last):
  File "test111-03.py", line 33, in ?
    test(h)
  File "test111-03.py", line 22, in test
    response=urllib2.urlopen(req)
  File "/usr/lib/python2.2/urllib2.py", line 138, in urlopen
    return _opener.open(url, data)
[...cut...]
  File "/usr/lib/python2.2/urllib2.py", line 425, in http_error_302
    self.inf_msg + msg, headers, fp)
urllib2.HTTPError: HTTP Error 302: The HTTP server returned a redirect error
that wouldlead to an infinite loop.
The last 302 error message was:
Found
----------snip----------


[4]----------snip----------
[...cut...]
   File "/usr/lib/python2.2/site-packages/ClientCookie/_ClientCookie.py", line
2098, in http_error_302    raise HTTPError(req.get_full_url(), code,
----------snip----------

[5]----------snip----------
0x0000 4500 014b bc2f 4000 4006 528e 3e93 a00c     E..K./@. at .R.>...
0x0010 c26d 89e2 813c 0050 bceb f313 2dbb 0d7b     .m...<.P....-..{
0x0020 5018 16d0 03cf 0000 4745 5420 2f20 4854     P.......GET./.HT
0x0030 5450 2f31 2e30 0d0a 486f 7374 3a20 7777     TP/1.0..Host:.ww
0x0040 772e 7079 7468 6f6e 2e6f 7267 0d0a 5573     w.python.org..Us
0x0050 6572 2d61 6765 6e74 3a20 5079 7468 6f6e     er-agent:.Python    <--------
???
----------snip----------


Thanks for any help


---
Patrick Bussi
patrick.bussi at space.alcatel.fr


Any opinions expressed are my own and not necessarily those of my Company.






From donn at drizzle.com  Fri Apr 25 00:02:52 2003
From: donn at drizzle.com (Donn Cave)
Date: Fri, 25 Apr 2003 04:02:52 -0000
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth>
Message-ID: <1051243370.926621@yasure>

Quoth Jim Richardson <warlock at eskimo.com>:

| I have come across a weirdness. I have a number of floats, (lat long
| pairs) which I stuff into a dict, but when I take them out, my nice %2f4
| number, (say, 47.5894) is something long and *almost* right, like
| 47.589333333333333339. What gives? I am not doing anything special as
| far as I can tell, just stuffing the numbers into a dict, and recalling
| it later. This is with Python 2.2 on Linux, and no, it's not an original
| Pentium :)

They are floats, they were born weird.

You're looking at the repr() output, which for reasons we won't
go into here renders the value to its ultimate precision.  If
you run the same numbers through str(), they'll look different.

Maybe the big question here is whether your notion of "right"
has any computational significance, that is, does it matter to
your application which of these renditions more accurately
describes the value.  Because in that respect, repr is right
and str is almost right.

	Donn Cave, donn at drizzle.com



From gh at ghaering.de  Tue Apr 15 08:22:57 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: 15 Apr 2003 12:22:57 GMT
Subject: Inserting HTML files as strings in database
References: <mailman.1050407951.26474.python-list@python.org>
Message-ID: <slrnb9nu28.170.gh@haering.opus-gmbh.net>

vivek kumar <gupt_vive at hotmail.com> wrote:
> hi all,

> I am trying to make a search engine in python. I want to save the html
> files retrieved by the crawler into a Mysql database as strings
> (MEDIUMTEXT). But the characters like ' " etc are giving problems while
> inserting it as a string. [...]

It's generally a bad idea to insert the values into the SQL string
yourself. This, and quoting the values appropriately, is the job of the
DB-API module:

#v+
a_really_ugly_string = "".join([chr(x) for x in range(1,255)])

# cx is the connection object
cu = cx.cursor()
cu.execute("""
    INSERT INTO MYTABLE(FOO) VALUES (%s)
    """, (a_really_ugly_string,))
cu.commit()
#v-

For inserting binary data (i. e. data containing the null byte) you'll need
a binary column type.

-- Gerhard



From mc at mclaveau.com  Tue Apr 29 02:29:02 2003
From: mc at mclaveau.com (Michel Claveau)
Date: Tue, 29 Apr 2003 08:29:02 +0200
Subject: Calling dotNet components from Python.
References: <924a9f9c.0304282138.15aa7635@posting.google.com>
Message-ID: <b8l63s$amc$1@news-reader13.wanadoo.fr>

Hi !

Nice surname, Paradox !
You are welcome on the NG news://news.zoo-logique.org/programmation.Paradox

@-salutations
--
Michel Claveau





From tweedgeezer at hotmail.com  Mon Apr 14 02:24:44 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 13 Apr 2003 23:24:44 -0700
Subject: Parsing strings?
Message-ID: <698f09f8.0304132224.6a28c76f@posting.google.com>

Is there any way to turn a string (in Python format, escapes and all)
into an actual Python string without using eval()?

I.e., I'd like to turn the four-character string "\\x00" into the
one-character Python string "\x00" without using eval().  Is there any
way to do that?

Thanks,
Jeremy



From robin at jessikat.fsnet.co.uk  Sat Apr 19 16:11:25 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Sat, 19 Apr 2003 21:11:25 +0100
Subject: Parent references: hot or not?
References: <mailman.1050762692.5175.python-list@python.org>
 <fFhoa.6843$_N.1979@twister.nyroc.rr.com>
Message-ID: <n+OI4LAt1ao+EwHb@jessikat.fsnet.co.uk>

In article <fFhoa.6843$_N.1979 at twister.nyroc.rr.com>, darrell
<dgallion1 at yahoo.com> writes
>Check the weakref module.
>The child can keep a weakref to the parent. This won't set up a cycle.
>
>--Darrell
>
>
except when the parent container is a tuple or list.
-- 
Robin Becker



From jepler at unpythonic.net  Fri Apr 11 12:27:10 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 11 Apr 2003 11:27:10 -0500
Subject: a module who knows its caller
In-Reply-To: <2259b0e2.0304110812.4eea37de@posting.google.com>
References: <2259b0e2.0304110812.4eea37de@posting.google.com>
Message-ID: <20030411162708.GC924@unpythonic.net>

On Fri, Apr 11, 2003 at 09:12:35AM -0700, Michele Simionato wrote:
> I have a module (say m.py) that can be called by different scripts (say a.py 
> and b.py). I want m.py to be able to tell if it has been called by a.py or
> by b.py.

this is the sort of thing you abuse sys._getframe() for.

# a.py and b.py
import m

# m.py
import sys
try:
	f = sys._getframe(1)
except ValueError:
	print "not imported by anyone"
else:
	print "imported from file", f.f_code.co_filename
	# could use f_globals['__name__'] but this doesn't exist
	# when the importer is __main__
	print "imported from module", f.f_globals['__name__']


now, several ways to cause m to be imported:
[jepler at parrot cb]$ python -c 'import a'
imported from file a.py
imported from module a
[jepler at parrot cb]$ python a.py
imported from file a.py
imported from module __main__
[jepler at parrot cb]$ python -c 'import m'
imported from file <string>
imported from module __main__
[jepler at parrot cb]$ python m.py
not imported by anyone
[jepler at parrot cb]$ python -c 'import a; import b'
imported from file a.py
imported from module a

Notice that when m is imported twice (once from a and once from b), only
one message is printed.  That's a sign that this kind of perversity will
ultimately do you wrong.

Jeff




From fperez528 at yahoo.com  Wed Apr  9 17:17:10 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 09 Apr 2003 15:17:10 -0600
Subject: Wraparound problems with time.clock() on Linux
References: <b722d0$bau$1@peabody.colorado.edu>
Message-ID: <b722km$bau$2@peabody.colorado.edu>

Fernando Perez wrote:


> The problem I am having is that in certain long runs, I get negative
> values
> for the total() function.  I was wondering if there are wraparound issues
> I
> should be aware of.  Here is an example of the output from two different
> runs of my code:
[snip]
> This one, on the other hand, generated negative times.  I know (from
> timestamps on files made by the code as it ran), that the code took about
> 50 minutes of wall time to actually execute.  Since time.clock() is
> supposed to return floats, I wasn't expecting wraparound problems.

Sorry, scratch that.  I went looking further in the docs, and it seems that
time.clock() calls the system's clock(). From man clock:

       Note  that  the  time  can  wrap  around.   On  a  32bit  system 
where
       CLOCKS_PER_SEC  equals 1000000 this function will return the same
value
       approximately every 72 minutes.



Now, my code ran in less than 72 minutes, but it's possible that redhat
messed up the CLOCKS_PER_SEC  value so that the wraparound is hit before 72
minutes (half that, maybe?).  At any rate, now I know there _are_
wraparound issues.

Anyone know of a reliable way to time long-running codes?

Cheers,

f



From cajones at datafast.net.au  Fri Apr  4 18:20:33 2003
From: cajones at datafast.net.au (Casey Jones)
Date: Sat, 05 Apr 2003 09:20:33 +1000
Subject: I'm new to programming
Message-ID: <oprm4ogjb9acxezr@news.datafast.net.au>


-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

I've been parusing the 'Learn Python in an Afternoon' tutorial and my head 
feels comfortable with it's
syntax etc. I've looked at VB, C, C++, blah it gives me a friggin 
headache!..

I suppose the problem is a goal "What the hell do I want to program and I 
just don't know".
I feel that it would give me a better understanding if I had a goal.

I know html I can look at Javascript and know what to look for when editing 
for roll overs etc but get
me to program something from scratch huh no can do I'm at a complete loss.

It's like I guess. "I know the moves in chess but ask me to checkmate a 
chessmaster in three moves forget it".

Inspiration would be appreciated. 



From aleax at aleax.it  Mon Apr 14 06:53:34 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 10:53:34 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi>
Message-ID: <Oqwma.11524$LB6.319489@news1.tin.it>

Marcus Alanen wrote:

> On Sat, 12 Apr 2003 20:14:51 -0400, Roy Smith <roy at panix.com> wrote:
>>One of the very few things I like about C++ over Python is that the STL
>>is very careful to document the complexity of the various operations.
>>AFAICT, no such documentation exists for Python.
> 
> Indeed, that was the first thing I wondered about when I started
> learning Python.

Interesting!  What other languages (besides C++ for a subset of its
standard libraries) give O() specifications for their operations?
In other words, what other languages did you learn, before Python,
that didn't leave you wondering about such performance issues?


>>As another example, I can guess that Python dictionaries work like STL
>>maps on the inside, but that would be just a guess.  Is inserting an
>>element logN?  Is keys() linear?  Is del on an element logN?  Is
>>has_key() logN?  Those would be my guesses.  But they're just guesses.
> 
> Others answered this very thoroughly, I'd just like to emphasize the
> linearity of keys() --- most often you can manage with .iterkeys() or
> .itervalues() just fine, and it's a lot faster.

Both keys and iterkeys are O(N) [as, of course, is itervalues].

As for the performance issues, I wouldn't be so sure.  E.g.:

[alex at lancelot python2.3]$ python2.3 -O timeit.py -s'''
> d=dict.fromkeys(range(200))''' 'for x in d.keys(): pass'
10000 loops, best of 3: 24.3 usec per loop
[alex at lancelot python2.3]$ python2.3 -O timeit.py -s'''
d=dict.fromkeys(range(200))''' 'for x in d.iterkeys(): pass'
10000 loops, best of 3: 28.8 usec per loop
[alex at lancelot python2.3]$ python2.3 -O timeit.py -s'''
d=dict.fromkeys(range(200))''' 'for x in d: pass'
10000 loops, best of 3: 28.1 usec per loop

[alex at lancelot python2.3]$ python2.3 -O timeit.py -s'''
d=dict.fromkeys(range(200))''' 'for x in d.keys(): pass'
10000 loops, best of 3: 24.7 usec per loop
[alex at lancelot python2.3]$ python2.3 -O timeit.py -s'''
d=dict.fromkeys(range(200))''' 'for x in d: pass'
10000 loops, best of 3: 28.9 usec per loop
[alex at lancelot python2.3]$ python2.3 -O timeit.py -s'''
d=dict.fromkeys(range(200))''' 'for x in d.iterkeys(): pass'
10000 loops, best of 3: 28.2 usec per loop
[alex at lancelot python2.3]$

In this case, and of course specifically on my machine (1.2 GHz
Athlon running Mandrake 9.0 and Python 2.3a2), we see that the
performance advantage is rather for looping on d.keys(), which
ends up being around 15% faster than looping on d.iterkeys()
[or equivalently looping directly on d itself].  I observe
similar results for Python 2.2.2 on the same machine, btw
(but performance advantage for keys reduced to just 3-5% as both
loops are substantially slower -- 2.3 is quite a bit faster
than 2.2 just about across the whole spectrum of operations).

That's most definitely NOT a significant performance difference, 
of course, but calling iteration on iterkeys "a lot faster" than 
iteration on keys would appear to be substantially misleading, at 
least without a lot of qualifications and hedging!-)


Alex




From mlh at furu.idi.ntnu.no  Wed Apr  9 20:03:40 2003
From: mlh at furu.idi.ntnu.no (Magnus Lie Hetland)
Date: Thu, 10 Apr 2003 00:03:40 +0000 (UTC)
Subject: PEP305 csv package: from csv import csv?
References: <mailman.1049814434.25284.python-list@python.org>
Message-ID: <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>

In article <mailman.1049814434.25284.python-list at python.org>, Hamish
Lawson wrote:
>According to the documentation in progress at
>
[snip]

Hey -- we finally get a csv package! Hooray!

>Is there some reason why the cvs

I assume you mean csv?

> package's __init__.py doesn't import the 
>required names from cvs.py,

csv.py, right? (I consistently make this mistake myself...)

> so allowing the shorter form below?

I agree that this should be done (assuming that that's what you think
too... :)

As far as I can see, csv.csv.__all__ contains the relevant names, so
simply adding

  from csv import *

to csv.__init__.py should do the trick nicely.

As for reasons -- the checkin message for the __init__.py (made by
Cliff Wells) is:

  "Make cvs [sic] into a package so we can have a util package"

So it seems that prior to this decision, csv was simply a module, and
the names were directly available. I'd think this to be preferrable,
and more consistent with existing library modules. We may have modules
containing classes with the same name, but this convention of placing
a module inside a package with the same name does not occur at all in
the libs. (Yes, I checked ;)

Of my currently installed packages, I see that numarray uses the
convention, for example, in its RandomArray2 module/package. However,
it also uses the above suggested convention of placing

  from RandomArray2 import *

in RandomArray/__init__.py.

So ... Are any of the PEP authors reading this? (Just checking -- as
this probably belongs on the discussion list... Sorry about that ;)

-- 
Magnus Lie Hetland               "Nothing shocks me. I'm a scientist." 
http://hetland.org                                   -- Indiana Jones



From ray at rays-web.com  Wed Apr 23 19:00:47 2003
From: ray at rays-web.com (Ray Smith)
Date: 23 Apr 2003 16:00:47 -0700
Subject: Off topic: What usenet news reader to use?
References: <mailman.1051036295.31460.python-list@python.org>
Message-ID: <5654fff9.0304231500.6673cf1f@posting.google.com>

Hi,

I know it's not answering your question ... but as an alternative I just use
google.  
I don't respond to many messages, speed is OK'ish for what I use it for and 
posting isn't really a big hassle anyway.  Worth looking at:

http://groups.google.com/groups?hl=en&group=comp.lang.python

Ray Smith


"Wai-Yip Tung \(wtung\)" <wtung at cisco.com> wrote in message news:<mailman.1051036295.31460.python-list at python.org>...
> I haven't used Usenet news for a long time. Now that I have subscribed
> to python-list I find that email does not scale well to the high volumne
> of discussion. Instead of deleting emails all day I think I should go
> back to Usenet news. Is there any good Windows news reader you would
> recommend. I'm happily using Opera as the browser. But it's
> revolutionary news and email reader eludes me. What are you using?
> 
> By the way I'm also looking for a free or low cost email client. My
> company is using Outlook but I want to find one for personal use. Thank
> you for you suggestions.
> 
> Wai Yip Tung



From the_3_project at yahoo.com  Tue Apr 29 05:39:07 2003
From: the_3_project at yahoo.com (Joe Green)
Date: 29 Apr 2003 02:39:07 -0700
Subject: _winreg newbie question
References: <a5b86892.0304282156.208b81b1@posting.google.com>
Message-ID: <a5b86892.0304290139.6870fb2b@posting.google.com>

The previous examples were incomplete (and contained errors). Here is
a more complete example:


# pyreged.py

from myregedit import *
def main():
  "Main() function"

  key='LOCAL_MACHINE'
  subkey='SOFTWARE'
  myr = myregedit()
  keylist = myr.IterKeys(key=key, subkey=subkey)
  if (keylist):
    print keylist
  else:
    print "Unable to iterate keys"


if __name__ == "__main__":
  main()

# myregedit.py
import sys, string
from _winreg import *

class myregedit:

  def __init__(self, compname=None, debug=None):
    "Class initilization 'constructor'"

    try:
      self.LOCAL_MACHINE = ConnectRegistry(compname,
HKEY_LOCAL_MACHINE)
      self.CURRENT_CONFIG = ConnectRegistry(compname,
HKEY_CURRENT_CONFIG)
      self.CURRENT_USER = ConnectRegistry(compname, HKEY_CURRENT_USER)
      self.DYN_DATA = ConnectRegistry(compname, HKEY_DYN_DATA)
      self.PERFORMANCE = ConnectRegistry(compname,
HKEY_PERFORMANCE_DATA)
      self.USERS = ConnectRegistry(compname, HKEY_USERS)
    except:
      print "Unable to connect to the registry."
      sys.exit(1)
    self.reg = [self.LOCAL_MACHINE, self.CURRENT_CONFIG,
self.CURRENT_USER,
                self.DYN_DATA, self.PERFORMANCE, self.USERS]
    
  def IterKeys(self, key=None, subkey=None):
    "Returns type and data of a given value"
    mykey=-1  

    if (key is "LOCAL_MACHINE"):
      print "HEYU!"
      mykey = 0
    elif (key is "CURRENT_CONFIG"):
      mykey = 1
    elif (key is "CURRENT_USER"):
      mykey = 2
    elif (key is "DYN_DATA"):
      mykey = 3
    elif (key is "PERFORMANCE_DATA"):
      mykey = 4
    elif (key is "USERS"):
      mykey = 5
    else:
      print "unableto assign mykey any value other than -1"
      print "[%s]\t[%s]\t[%d]" % (key,subkey,mykey)

    if (mykey is -1):
      return None

    subkeylist = []
    try:
      i = 0
      while 1:
        subkeylist.append(EnumKey(OpenKey(self.reg[mykey], subkey),
i))
        i += 1
    except EnvironmentError:
      pass
    return subkeylist


-------------
Doesn't make sense to me, I've been banging my head against the wall
for several hours now! I can't figure it out.. and I know it's
something very simple that I'm overlooking. I just don't see staring
at it any longer to be productive :)

Thanks



From fperez528 at yahoo.com  Fri Apr 18 15:25:40 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Fri, 18 Apr 2003 13:25:40 -0600
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <b7pjfl$bc6$1@peabody.colorado.edu>

Bill Martin wrote:

> I'm wondering about the value of allowing a class definition like this:
> 
>              class C:
>                pass
> 
> Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or
> some such. If I try to print a.x2 or b.x1, I get an exception message
> basically saying those member variables don't exist. It seems to me this
> defeats one of the basic ideas behind OOP, that is to assign members to
> the class definition, thus ensuring that all instantiations of a class
> have the same member variables and methods. I know there must be a
> reason that python allows this kind of thing that I described, but I
> can't figure out what that reason could be. To me, it looks like a very
> dangerous thing to do. Can someone help me understand why we'd like to
> do it? Thanks,

You are confusing the class and its instances.  a.x1 assigns to a's private
dict, so b can't see the x1 attribute, as it shouldn't.

If you want all instances of a class to see something, you must assign it to
the class:

In [1]: class C:
   ...:     pass
   ...:

In [2]: a=C()

In [3]: b=C()

In [4]: a.x1='in a'

In [5]: b.x2='in b'

In [6]: C.y='in all'

In [7]: a.x1
Out[7]: 'in a'

In [8]: b.x2
Out[8]: 'in b'

In [9]: a.x2
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)

?

AttributeError: C instance has no attribute 'x2'

In [10]: b.x1
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)

?

AttributeError: C instance has no attribute 'x1'

In [11]: a.y
Out[11]: 'in all'

In [12]: b.y
Out[12]: 'in all'


No surprises here.  Classes and instances are two different things.

hth,

f.



From ckea25d02 at sneakemail.com  Tue Apr  8 01:18:55 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 17:18:55 +1200
Subject: Calling python from a C function that is embedded in python?
In-Reply-To: <mailman.1049722812.17109.python-list@python.org>
References: <mailman.1049722812.17109.python-list@python.org>
Message-ID: <3E925BBF.2090308@sneakemail.com>

Greg Ewing (using news.cis.dfn.de) wrote:
 > VanL wrote:
 >> I want python to be able to call wrapped functions like
 >> they are python functions, and a wrapped C library to call wrapped
 >> python functions like they were in C.
 >
 > I have a demo showing how to do it with Pyrex, but I don't have it
 > here right now.

Well, I found the demo. It shows how you can pass a
Python function to a C library as a callback.

cheesefinder.h:
-------------------------------------------------------
typedef void (*user_func)(char *name, void *user_data);
void find_cheeses(user_func f, void *user_data);
-------------------------------------------------------

cheesefinder.c:
-------------------------------------------------------
#include "cheesefinder.h"

static char *cheese_list[] = {
   "cheddar",
   "camembert",
   "that runny one",
};

void find_cheeses(user_func f, void *user_data) {
   char **p = cheese_list;
   while (*p) {
     f(*p, user_data);
     ++p;
   }
}
-------------------------------------------------------

cheese.pyx:
-------------------------------------------------------
cdef extern from "cheesefinder.h":
   ctypedef void (*user_func)(char *name, void *user_data)
   void find_cheeses(user_func f, void *user_data)

def find(f):
   find_cheeses(callback, <void*>f)

cdef void callback(char *name, void *f):
   (<object>f)(name)
-------------------------------------------------------

run_cheese.py:
-------------------------------------------------------
import cheese

def report_cheese(name):
   print "Found cheese:", name

cheese.find(report_cheese)
-------------------------------------------------------

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From skip at pobox.com  Fri Apr 18 18:08:13 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 18 Apr 2003 17:08:13 -0500
Subject: Job Opportunity
In-Reply-To: <NHBBIABMOLGOBKFNCIFJKEHCCAAA.mike@edgelinkrecruiters.com>
References: <NHBBIABMOLGOBKFNCIFJKEHCCAAA.mike@edgelinkrecruiters.com>
Message-ID: <16032.30541.71198.936403@montanaro.dyndns.org>

    Mike> I am a Portland-based recruiter looking for an embedded Linux
    Mike> software engineer with Python, c/c++ programming skills.  Music
    Mike> interest/background is a must.  The position is full time and
    Mike> located in the Beaverton/Hillsboro area. Do you know anyone I
    Mike> could speak to?  Thanks.

Not off the top of my head, but if you visit

    http://www.python.org/Jobs.html

and follow the directions there, someone at jobs at python.org will add an
announcement for you in the next day or two.

Skip




From martin at v.loewis.de  Mon Apr 14 01:12:13 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 14 Apr 2003 07:12:13 +0200
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk> <3e970abc$0$49101$e4fe514c@news.xs4all.nl> <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk> <1e740$3e99a2f6$516049d2$15395@nf1.news-service.com> <2+aM4KAFkam+EwV+@jessikat.fsnet.co.uk> <m37k9y2l08.fsf@mira.informatik.hu-berlin.de> <m5b4piAB+dm+Ewn+@jessikat.fsnet.co.uk>
Message-ID: <m3smsld5n6.fsf@mira.informatik.hu-berlin.de>

Robin Becker <robin at jessikat.fsnet.co.uk> writes:

> so how do master slave implementations handle thousands of slaves? If
> the master is really a 'master' and not a SETI type server it would seem
> that these resource limits might play a part. How do GRID systems work
> this?

In a compute-intensive application, it is sensible to close the TCP
connection after you have communicated the job. Keeping the connection
alive is important to avoid the TCP connection setup, however, that is
neglible if you spend several minutes or more in computation.

I don't actually know how grid computing protocols work, but I believe
they are not concerned about networking performance, as they use HTTP
and SOAP to communicate jobs.

Regards,
Martin




From xlp223 at sina.com.cn  Sun Apr 27 00:18:51 2003
From: xlp223 at sina.com.cn (xlp223)
Date: 26 Apr 2003 21:18:51 -0700
Subject: a problem about running php on twisted 1.0.4 under window?
Message-ID: <4dbec4b4.0304262018.3ed4a8f3@posting.google.com>

my system is 
window me
python 2.2.2
twisted 1.0.4
php 4.3.1
mysql
phpbb 2.0.4
i want to run phpbb on twisted 1.0.4 under window,
when i installed above software on window xp home edition,it was successful.
but when i installed above software on window me,it failed.

Step of my doing is 
1)create dir c:\pydoc and copy c:\python22\doc to c:\pydoc 
  create c:\pydoc\phpbb2 and copy phpbb2 to it

2)edit hello.rpy in c:\python22\scripts and save it in this dir
from twisted.web import static, twcgi

class PHPScript(twcgi.FilteredScript):
    filter = 'C:\\PHP\\php'

resource = static.File("c:\\pydoc")
resource.processors[".php"] = PHPScript
resource.indexNames = ['index.html','index.php']

3)run the command:
  python mktap.py web -p 80 --resource-script=hello.rpy -i index.html,index.php

4)run the command:
  python twistd.py -f web.tap -r win32

After the fourth step ran,i entered into the ie browser and input
http://localhost/   it is ok to show the html.
but while i input http://localhost/phpbb2 ,
it show 500 error. 

why? who can help me to solve it?



From rmunn at pobox.com  Wed Apr  2 10:35:30 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 15:35:30 GMT
Subject: converting from shell script to python
References: <kfn0jaoig1.fsf@shaoz.activesky.com.au> <b6bccj$3fsej$1@ID-15245.news.dfncis.de> <b6db2p$48k88$1@ID-169208.news.dfncis.de> <b6e0h8$4agg8$2@ID-15245.news.dfncis.de>
Message-ID: <slrnb8lbqa.ccp.rmunn@localhost.localdomain>

Mathias Waack <mathias-usenet at valpo.de> wrote:
> "Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com>
> wrote:  
>> Alternatively, you can substitute things from a
>> dictionary instead of a tuple:
>> 
>>    vars = {'ROOTDIR': '/usr/lib'}
>>    CLASSPATH = \
>>      "%{ROOTDIR}s/a/a.jar:%{ROOTDIR}s/b/b.jar:%{ROOTDIR}s/c/c.jar
>>      " % vars 
> 
> Didn't know that before. Its nice - thanks for the hint. 

All the gory details on using the % operator can be found here:

    http://www.python.org/doc/current/lib/typesseq-strings.html

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From ckea25d02 at sneakemail.com  Tue Apr  1 21:26:02 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 14:26:02 +1200
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <mailman.1049226964.14755.python-list@python.org>
References: <x7y92uhzpw.fsf@guru.mired.org> <mailman.1049226964.14755.python-list@python.org>
Message-ID: <b6dh8c$43bo6$1@ID-169208.news.dfncis.de>

Dave Brueck wrote:
> Further, I
> propose that Aahz be the official vote counter. Does anybody have any
> preferences as to the voting method we should use?

The proponent of the PEP shall be placed in a gladiatorial
arena together with half a dozen hungry lions, and permitted
to debate the merits of the proposal with them. The BDFL
shall preside over the process, and at its conclusion, vote by
raising or lowering his thumb.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From danny at iamlearning.com  Thu Apr  3 21:43:34 2003
From: danny at iamlearning.com (danny at iamlearning.com)
Date: Thu, 3 Apr 2003 20:43:34 -0600 (CST)
Subject: How to compile a python script to *.pyo
In-Reply-To: <3E8C8A57.9080603@gatwick.westerngeco.slb.com>
References: <2977.143.166.82.240.1049308819.squirrel@iamlearning.com>
        <3E8C8A57.9080603@gatwick.westerngeco.slb.com>
Message-ID: <3407.192.168.2.15.1049424214.squirrel@iamlearning.com>

>
> I'll ignore the why's and wherefore's and give you what I consider a
> straight answer.....
>
> pydoc compileall
>
>   Python Library Documentation: module compileall
>
> NAME
>      compileall - Module/script to "compile" all .py files to .pyc (or
> .pyo) file.
>
> FILE
>      /usr/local/lib/python2.2/compileall.py
>
> DESCRIPTION
>      When called as a script with arguments, this compiles the
> directories given as arguments recursively; the -l option prevents
> it from recursing into directories.
>
>      Without arguments, if compiles all modules on sys.path, without
> recursing into subdirectories.  (Even though it should do so for
> packages -- for now, you'll have to deal with packages separately.)
>
>      See module py_compile for details of the actual byte-compilation.
>
> FUNCTIONS
>      compile_dir(dir, maxlevels=10, ddir=None, force=0, rx=None)
>          Byte-compile all modules in the given directory tree.
>
>
> .....  snip .....
>
>
>
> I use this for yes you've guessed it compiling my applications, on the
> fly as they are installed.
>
>
>
> Regards
> Martin

For Martin,
Thanks for your advice. That's what I need.
If this mailing list got more people like you, there are less junk
messages to read.










From bokr at oz.net  Thu Apr  3 19:09:57 2003
From: bokr at oz.net (Bengt Richter)
Date: 4 Apr 2003 00:09:57 GMT
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <3E8CBC89.733D01D1@alcyone.com>
Message-ID: <b6iigl$mk3$0@216.39.172.122>

On Thu, 03 Apr 2003 14:58:17 -0800, Erik Max Francis <max at alcyone.com> wrote:

>Michael Hudson wrote:
>
>> I certainly did.
>
>Me too.  I don't have any huge opinion on the proposition one way or
>another, but I must say I do find it reassuring in Python to know that
>(at present, without the proposal) whenever I see an assignment, I
>always know that there's going to be a rebinding/attribute assignment
>taking place.  Assignment seems so fundamental an operation that I'm not
>sure whether it would be net good or net bad to have it overridable.
>
So how do you feel about properties? Seems like the OP's proposal would
open up another way to implement at least the assignment part. Perhaps
there should be an analogous __getvalue__ method for the right hand side.
(Already ducking ;-)

Regards,
Bengt Richter



From gk_2345 at yahoo.com  Sat Apr 26 10:46:46 2003
From: gk_2345 at yahoo.com (George Kinney)
Date: Sat, 26 Apr 2003 10:46:46 -0400
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.29.48.204375@soul.net> <RZlqa.4182$lf3.712485@cletus.bright.net> <pan.2003.04.26.14.19.35.262492@soul.net>
Message-ID: <Y%wqa.4238$lf3.720331@cletus.bright.net>

"lost" <lost at soul.net> wrote in message
news:pan.2003.04.26.14.19.35.262492 at soul.net...
> thanks for the info, however that doesn't seem to work either? here's the
> whole program that I'm trying:
> #!/usr/bin/python
> from glob import glob
> file_list = glob.glob('/var/www/html/') for each_file in file_list:
>         print each_file
> print "done"
>
> Running that gives me:
> Traceback (most recent call last):
>   File "files.py", line 6, in ?
>     file_list = glob.glob('/var/www/html/')
> AttributeError: 'function' object has no attribute 'glob'
>
> Any ideas where I'm going wrong?

Sorry, I should have pointed out that glob does wild-card expansion (not
just a dir list).
So if you tried :
file_list = glob.glob('/var/www/html/*')   # (notice the *, could also be
*.html or whatever)
 for each_file in file_list :
     print each_file
print "done"

It'll behave like you expect.







From BPettersen at NAREX.com  Tue Apr 29 13:21:58 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Tue, 29 Apr 2003 11:21:58 -0600
Subject: Question about accessing class-attributes.
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE15E@admin56.narex.com>

> From: Alex Martelli [mailto:aleax at aleax.it] 
[...]
> > however I haven't been able to find a good description of attribute
> > lookup for new-style classes... (references?) E.g. why is 
> > the metaclass' __getattr__ never called?
> 
> Uh, what do you mean with this last question...?
> 
> >>> class meta(type):
> ...   def __getattr__(self, name):
> ...     if len(name)==1: return name
> ...     raise AttributeError, name
> ...
> >>> __metaclass__=meta
> >>> class X: pass
> ...
> >>> X.h
> 'h'
> >>>

Correct (although I must admit I didn't test regular attributes :-)
Here's the case I'm confused about:

>>> class meta(type):
...   def __getattr__(cls, name):
...     if name == '__len__':
...        print "meta.__getattr__('__len__')"
...        return lambda: 42
...     else:
...        print 'meta.__getattr__', name
...        return name
...
>>> class S(object):
...   __metaclass__ = meta
...
>>> S.__len__()
meta.__getattr__('__len__')
42
>>> len(S)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: len() of unsized object
>>>

I was told that special method "foo(x, arg)" was implemented as
"type(x).__foo__(x, arg)", which doesn't seem to be the case always...
Compare:

>>> class meta(type):
...   def __len__(cls):
...     return 42
...
>>> class S(object):
...   __metaclass__ = meta
...
>>> S.__len__()
42
>>> len(S)
42
>>>

So, it looks like it's looking up __len__ in the metaclass, but not
falling back on __getattr__ when it isn't there? I've looked at the C
code and it seems like special methods each have their own way of
finding the function they're needing, e.g. for len, it looks like it
uses:

  m = o->ob_type->tp_as_sequence;
  if (m && m->sq_length)
    return m->sq_length(o);

  return PyMapping_Size(o);

which would never call the metaclass' __getattr__ (since tp_as_sequence
would presumably be NULL?).

The relevance this has to super, is that you couldn't do a (contrived)
example like:

  for item in super(MyClass, self):
      ...

(in general, you can't pass super(MyClass, self) to a function expecting
a MyClass instance, even if MyClass defines no new behavior...)

Perhaps a more realistic exmaple?:

  class Secure(object):
    def __init__(self, name):
      self.name = name
  
    def __getattr__(self, attr):
      return getSecureMethod(self.name, attr, getUserPrivileges())
        
  class LogFoo(Secure):
    def __init__(self):
      Secure.__init__(self, 'SalaryDB')
      
    def foo(self):
      print 'calling foo'
      return super(LogFoo, self).foo()
      
  lf = LogFoo()
  print lf.foo()
 
Any insight?

-- bjorn




From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr  3 14:21:28 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 3 Apr 2003 19:21:28 GMT
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <qp5o8voto8li6p854klt7vo6dhon69hk2o@4ax.com>
Message-ID: <slrnb8p2dn.174.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <qp5o8voto8li6p854klt7vo6dhon69hk2o at 4ax.com>, Richie Hindle wrote:
<snip> 
> So the line `page.message = "New message"` changes the content of the tag
> with id 'message'.  In this context, it makes sense that assigning to an
> object *changes* that object, rather than *replacing* it with another.

That's assigning to an attribute, not to an object.

> What you can't do with PyMeld is this:
> 
>>>> message = page.message
>>>> message = "Another new message"
>>>> print page
><html><body>
><textarea id="message">Another new message.</textarea>
></body></html>
> 
> There are workarounds, but over loadable assignment would make this much
> neater.
<snip>

No, it would make it very weird, and massively incompatible with existing
code.  Binding and mutating are fundamentally different and shouldn't be
combined in one operator.  (Look at the confusion people sometimes have with
initialisation and assignment in C++.)  Python's augmented assignment
operators already have this horrible behaviour, and I stay clear of them
because of this.



From cnetzer at mail.arc.nasa.gov  Tue Apr 15 17:23:10 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 15 Apr 2003 14:23:10 -0700
Subject: Big-O notation (was: Re: Python and Schools)
In-Reply-To: <3E9C6C62.8010808@rogers.com>
References: <qGqla.61$Tj7.91655@news.uswest.net>
	 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
	 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
	 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
	 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
	 <3E9C6C62.8010808@rogers.com>
Message-ID: <1050441790.11992.18.camel@sayge.arc.nasa.gov>

On Tue, 2003-04-15 at 13:32, Mike C. Fletcher wrote:
> Greg Brondo wrote:

> Here's a crack at explaining it:

Good job.  I have a few corrections and minor additions.


>     * O(1) -- time is linear

You meant "time is constant".  The algorithm requires the same number of
operations regardless of input size.

>     * O(N log N) -- see O(N) and O(log N), basically, you do an amount
>       of work tied to the size of the whole set for every stage in an O(
>       log N ) operation (this is comparatively rarely discussed in my
>       experience).

Really?  Some of the most important classes of algorithms are O( N log N
), specifically, many common sort algorithms.  Also, there are many of
these algorithms that have degenerate cases that become O( N**2 )
algorithms, with naive implementations or just by the nature of the data
set, and it is important to be wary of this phenomenon.

Also, one further class that is import to know is O( 2**N ), such as the
brute-force solution to the "travelling salesman problem".  These are
much, MUCH worse than O( N**2 ) algorithms, and even today can rarely be
computed for problems sets much larger than 100 (whereas O( N**2 )
problem sets can grow to thousands or perhaps even millions and still be
practically solved).  Rather than computing "solutions" with these types
of algorithms, it is required to instead come up with "close-enough" or
"best guess" type algorithms that CAN be computed practically on large
data sets.

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From johnroth at ameritech.net  Sun Apr  6 13:24:40 2003
From: johnroth at ameritech.net (John Roth)
Date: Sun, 6 Apr 2003 13:24:40 -0400
Subject: properties and get/set methods
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org> <v908s7echa191b@news.supernews.com> <mailman.1049643850.17962.python-list@python.org>
Message-ID: <v90of6avvh1leb@news.supernews.com>

"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
news:mailman.1049643850.17962.python-list at python.org...
> John Roth wrote:
>
> >"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
> >news:mailman.1049604615.22077.python-list at python.org...
> >
> >
> ...
>
> >>In these cases, you're using properties to simplify the modelling of
> >>domain objects, or to provide some *generic* functionality over and
> >>above a regular value get/set.  You don't want to have to specify
that
> >>functionality for every property, you just want to say "this is one
of
> >>_those_ properties".  The code bloat (not to mention mindless
> >>duplication of code) when you get up to 15 or 20 properties for an
> >>object would be ridiculous.
> >>
[snip]
> >>
> Seriously though, yes, this is a "religious" argument.  The morality
> going along these lines:
>
>     * Needless duplication of code creates bugs and increases
>       maintenance costs
>     * Explicit is better than implicit
>     * Practicality beats purity, but value purity when you find it
>     * Beliefs change
>
> Properties as they were originally designed have a "canon" of use
which
> says they're for wrapping getter/setter methods up into objects that
> mediate access to an attribute via x.y notation.  That's fine.  It's
> useful sometimes, though not spectacularly often in my experience.
>
> When people start using properties heavily, the definition migrates
> closer to "describing a property/attribute slot of an object" with the
> access mediation of getter/setters being a side-effect of that
> description.  So after a while, the people using them start seeing
them
> in a different light than those for whom they are still just for
> wrapping getter/setter methods.
>
> The generic characteristics and systemic functions of these new
> properties get captured in the new systems (e.g. as sub-classes of
> property objects) in such a way as to dramatically simplify some types
> of programming.  For some, the new properties *become* the interface
to
> the canon as they are the only property type used in day-to-day
> programming.  In response, the canon of "properties" either expands to
> embrace the new usage, or fractures/is redefined.
>
> Just an example...
>
> class ElevationGrid( nodetypes.Geometry ):
>     PROTO = 'ElevationGrid'
>     zDimension = field.newField( 'zDimension', 'SFInt32', 0, 0)
>     xSpacing = field.newField( 'xSpacing', 'SFFloat', 0, 0.0)
>     creaseAngle = field.newField( 'creaseAngle', 'SFFloat', 0, 0.0)
>     normal = field.newField( 'normal', 'SFNode', 1, node.NULL)
>     solid = field.newField( 'solid', 'SFBool', 0, 1)
>     ccw = field.newField( 'ccw', 'SFBool', 0, 1)
>     texCoord = field.newField( 'texCoord', 'SFNode', 1, node.NULL)
>     height = field.newField( 'height', 'MFFloat', 0, list)
>     color = field.newField( 'color', 'SFNode', 1, node.NULL)
>     colorPerVertex = field.newField( 'colorPerVertex', 'SFBool', 0, 1)
>     xDimension = field.newField( 'xDimension', 'SFInt32', 0, 0)
>     zSpacing = field.newField( 'zSpacing', 'SFFloat', 0, 0.0)
>     normalPerVertex = field.newField( 'normalPerVertex', 'SFBool', 0,
1)
>
> that's a node which has default-values for each field, as well as
> per-field watchability, introspection, type-checking and coercion.
> (Incidentally, each property shows up as a single entry in
documentation
> systems (not as 39 accessor methods + 13 properties)).  Sure, you can
do
> that with getter/setter methods, but... well... why?  Using property
> sub-classes is the solution that seems right to me.
>
> Mike

I agree, that's a better solution although there is one big difficulty.
The approach works quite well in Java, where you've got the ability
to declare a variable as "final." You don't have that ability in Python,
so you run the risk of someone accidentally rebinding the field rather
than
using the extra level of indirection needed to access it correctly.

Following on to this idea, would you say that the property syntax
introduced in Python 2.2 was a mistake? It should have been more
general?

John Roth







From aahz at pythoncraft.com  Thu Apr  3 23:45:16 2003
From: aahz at pythoncraft.com (Aahz)
Date: 3 Apr 2003 23:45:16 -0500
Subject: Roman assignments (Re: Overloadable Assignment PEP)
References: <9383481e.0304012011.78f05548@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <a8b7f07a.0304031432.3c585657@posting.google.com> <b6in8i$600fi$1@ID-169208.news.dfncis.de>
Message-ID: <b6j2ks$8gg$1@panix2.panix.com>

In article <b6in8i$600fi$1 at ID-169208.news.dfncis.de>,
Greg Ewing (using news.cis.dfn.de) <ckea25d02 at sneakemail.com> wrote:
>A. Lloyd Flanagan wrote:
>>
>> Oh, good, if this had turned out to be a joke, I'd be afraid I won't
>> get my Roman numeral literals after all.
>
>The Roman numerals PEP seems unlikely to get passed
>in any case, but perhaps we can rescue something from
>it to help with this problem: we could apply
>augmented assignment semantics when and only when
>the name being assigned to resembles a Roman numeral.
>
>This very neatly solves the "voltage" problem,
>because then
>
>   V = 3
>
>would be an augmented assignment, whereas
>
>   Voltaire = 3
>
>would just be an ordinary assignment.

...and if we add a

    from __past__ import FORTRAN

it only applies when the names are I, J, K, L, M, or N.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From salvatore at wanadoo.fr  Thu Apr  3 12:15:17 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Thu, 03 Apr 2003 19:15:17 +0200
Subject: Creating instance
In-Reply-To: <JUoia.305$8h6.197@news02.roc.ny.frontiernet.net>
References: <b6d191$j7s$1@news-reader12.wanadoo.fr> <JUoia.305$8h6.197@news02.roc.ny.frontiernet.net>
Message-ID: <b6hq4h$ck3$1@news-reader12.wanadoo.fr>

Lee Harr wrote:

> In article <b6d191$j7s$1 at news-reader12.wanadoo.fr>, Salvatore wrote:
> 
>>Hello,
>>
>>I would like to create an instance of class
>>from the string name of the class
>>
>>class Foo:
>>	pass
>>
>>classname = 'Foo'
>>instance = create_instance('Foo')
>>
>>
> 
> 
> 
> I think it depends on which namespace the class lives in,
> and where you want to get at it from.
> 
> In the interactive interpreter, you might do:
> 
> class Foo:
>     pass
> instance = locals()['Foo']()
> 
> 
> 
> If the class lived in a module, say FooMod.py you could do:
> 
> # FooMod.py
> import FooMod
> 
> class Foo:
>     pass
> instance = getattr(FooMod, 'Foo')()
> 
Great too ;-)




From gupt_vive at hotmail.com  Tue Apr 22 01:06:22 2003
From: gupt_vive at hotmail.com (vivek kumar)
Date: Tue, 22 Apr 2003 10:36:22 +0530
Subject: widget like listview
Message-ID: <Law15-F76RArVc4Iu1B00004b25@hotmail.com>

hello david,

  Thanks for this valuable information. sorry for replying late. but for 
previous two days i was trying to use the bwidget and mclistbox modules. 
More over as i am a newbie in the Python world, so i didn't know how to use 
tcl directly in your python code. So i was trying to explore the subject. 
but was unable to find any help/tutorial on web. So i read the code in the 
files provided with the python distribution. However i was unable to use 
these modules. After installing the bwidget module when i try to call it 
like:

root.tk.call("package","require","BWidget")

then it says that no package name BWidget. The same problem occurrs with the 
Python Wrapper provided with the distro for BWidgets. I have tried each and 
everything (changing paths, installing ActiveTcl etc..) for using it but i 
was unable to use it. I have also tried the NovaGrid.py but it also requires 
Bwidgets.

Is there some help/tutorial available as how to embed/use tcl code in 
python.

TIA,
VIvek Kumar


>From: "David LeBlanc" <whisper at oz.net>
>To: "vivek kumar" <gupt_vive at hotmail.com>,<python-list at python.org>
>Subject: RE: widget like listview
>Date: Mon, 21 Apr 2003 02:16:55 -0700
>
> > 1. Once you have fixed the widths of the columns the user can not
> > chnage the
> > widths of the columns as per his/her preference.
>
>You might find the bwidgets package more interesting than Pmw. I think it's
>mc listbox is user changabe. See http://www.effbot.org/downloads/
>
>If not, the TkGrid control is. One python wrapping of it of it can be found
>in "NovaGrid" (as can a different implementation of bwidgets). See
>http://sourceforge.net/projects/novagrid
>
> > 2. What about linux. As this module uses the win32all extensions.
> > the code
> > no more remains portable. Currently whatever i have coded i have coded
> > keeping in mind that i have to run the application on both windows and
> > linux.
>
>Pmw doesn't use win32all. It's pure python, as is bwidgets (and tkgrid can
>be built on/for linux).
>
><snip>
>
> > THX,
> > Vivek Kumar
> >
>
>hth,
>
>Dave LeBlanc
>Seattle, WA USA
>


_________________________________________________________________
MSN Instyle. Keeps you in sync. http://server1.msn.co.in/instyle/index.asp 
Always!





From andy at wild-flower.co.uk  Mon Apr 21 18:22:26 2003
From: andy at wild-flower.co.uk (Andy Jewell)
Date: Mon, 21 Apr 2003 23:22:26 +0100
Subject: Multiple assignment?
In-Reply-To: <3EA3C010.5090002@antares.enst-bretagne.fr>
References: <3EA3C010.5090002@antares.enst-bretagne.fr>
Message-ID: <200304212317.10987.andy@wild-flower.co.uk>

On Monday 21 Apr 2003 10:55 am, Mehta, Anish wrote:
> Can someone tell me the what is being done here in this line of code.
>
> dir, file = os.path.split(input)
>
> Waiting for reply
>
> Thanks
>
> Anish MEHTA


1) Check the docs at python.org.

2) os.path.split cuts a file path into two parts, the directory bit, and the 
file name bit, and returns it as a tuple.  

3) Example of tuples in use:

>>> def swap(a,b):   # This is NOT the best way to do this - example only
	return b,a

>>> swap(1,2)
(2, 1)
>>> x=1
>>> y=2
>>> x,y=swap(x,y)
>>> x
2
>>> y
1

4) The final answer is that the statement:

    dir, file = os.path.split(input)

'unpacks' the tuple returned by os.path.split() into the variables dir and 
file.

5) 'file' is a bad choice for a variable name, as it is also a builtin class 
for opening files (depending on your version of python).






From woodsplitter at rocketmail.com  Mon Apr 14 08:55:10 2003
From: woodsplitter at rocketmail.com (David Rushby)
Date: Mon, 14 Apr 2003 08:55:10 -0400
Subject: PIL binary for windows?
References: <fe1cb5be.0304120417.5b756718@posting.google.com>
Message-ID: <o4yma.19580$c42.5968@fe01.atl2.webusenet.com>

"M?rten Ask" <majbritt37 at hotmail.com> wrote in message
news:fe1cb5be.0304120417.5b756718 at posting.google.com...
> Does anyone know where I can find a PIL binary for windows and Python
> 2.3? Or maybe some help in building it myself?

http://thor.prohosting.com/~drushby/files/PIL_113_for_py23_custom_compiled.z
ip






From carroll at tjc.com  Fri Apr  4 18:00:28 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Fri, 04 Apr 2003 15:00:28 -0800
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com>

On 3 Apr 2003 16:51:25 GMT, Ben Hutchings
<do-not-spam-ben.hutchings at businesswebsoftware.com> wrote:

>Or you can edit nntplib to add the option of storing XOVER output to a
>file.  It doesn't look like this would be too hard.  The 'file' could
>be an object that processes each line as it comes, rather than a real
>file.  Then send your change to nntplib back to the author and hope it's
>accepted into the official version.

That's not a bad idea.  David, you should consider taking this
approach.

Looking at nntplib.py, most of the logic to do this is already there:
xover calls longcmd to actually issue the command, and longcmd can
take an optional File parameter to which getlongresp (which actually
retrieves the response) will write its output, if present.

It wouldn't take much to add an optional file parameter to xover, and
modifying xover to pass it through to longcmd.  Right now, this is
only used in the body method, but I don't see why it wouldn't work for
xover or, for that matter, most of the other methods.





From staschuk at telusplanet.net  Tue Apr  1 00:54:34 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 31 Mar 2003 22:54:34 -0700
Subject: Coordinating access to a non-thread-safe db
In-Reply-To: <3E88D3DB.2010602@verio.net>; from vlindberg@verio.net on Mon, Mar 31, 2003 at 04:48:43PM -0700
References: <3E88D3DB.2010602@verio.net>
Message-ID: <20030331225434.A8765@tibia.amotlpaa.bogus>

Quoth VanL:
> I am writing some code that needs to respond to two circumstances:
> 1. User input via an event-driven interface
> 2. The passing of a certain amount of time (a la cron)
> 
> I am planning to use the sched and threading modules to do this.  
> However, I need to access a database that is only threadsafe at level 1 
> (sqlite) and so I am unsure how to coordinate database access such that 
> nothing gets corrupted.  [...]

The simplest strategy is to open a new connection in each thread
which needs one.  Of course, this is undesirable if opening a
connection takes a long time or is resource intensive.

If you want to open a connection just once, and run all database
interaction through that one connection, then you need a lock (see
threading.RLock) to enforce the mutual exclusion.  Code which
needs to use the connection first locks the lock, then uses the
connection, then releases the lock.  If another thread has locked
the lock, the first step will block until the lock is released.

If you trust all your code to do this dance right, it's easy to
implement.  For safety, though, some extra work is required; a
thread could release the lock but continue to make use of the
connection object, causing chaos.  To guard against this, the
function which locks the lock and provides the connection object
to the caller can instead return a wrapper with a release method
which not only releases the lock but sets a flag in the wrapper
which makes all its methods raise an exception.  Note that this
wrapper also needs to wrap objects returned by the backing
connection, such as cursors; also it should have a __del__ method
which calls release(), in case a caller forgets to release it
explicitly.

(If some database operations take longer than other threads can
wait, you probably need the middle ground: a connection pool.)

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From exarkun at intarweb.us  Sun Apr 13 14:48:30 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 13 Apr 2003 14:48:30 -0400
Subject: Terminal Emulation
In-Reply-To: <pan.2003.04.13.07.40.04.733053@web.de>
References: <pan.2003.04.13.07.40.04.733053@web.de>
Message-ID: <20030413184830.GA12415@meson.dyndns.org>

On Sun, Apr 13, 2003 at 09:40:29AM +0200, Hans-Joachim Widmaier wrote:
> All of my embedded software offers a means of debugging via a serial
> interface. You just hook a terminal onto it and start introspection. These
> days, of course, the "terminal" is usually a computer running a terminal
> emulator, and the embedded software allows you to do more than just view
> memory contents, e. g., download a new software. Of course, I handle this
> download with a python program. Now you have 2 programs: the terminal
> emulator and the download utility. Not too much hassle under unix, as the
> downloader ignores that the serial port is already used by the terminal
> emulator, it just does its job and leaves the port in the same state it
> had found it, the emulator doesn't even notice.
> 
> This doesn't work under Windows, though. So I thought about creating a
> terminal emulator with some added functionality. Some thinking revealed
> that I didn't know a widget that was handy for the job. Some more
> thinking gave me an idea for a unix solution: xterm in slave mode.
> Obviously, this only works where I don't really need it. A google search
> showed that some others also had a use for such a terminal emulation
> widget, but none seems to exist.
> 
> Not knowing wxWindows, pyQt and what-have-you, the Tkinter Text widget is
> about the only thing left. I'd have to grab all keyboard events away from
> it (yuck) and do the escape sequence interpretation by myself (not too
> bad, as I only need a subset of the whole vt100/vt320 stuff). The Text
> widget just seems way too "heavy" for this.
> Handling serial input/output as well as GUI events gives me problems
> enough already. ;-)
> 
> Has anyone found a nicer solution in the meantime? No problems if it meant
> using wxPython or pyGTK. (No pyQt, though, as it's not free under Windows.)
> 

  Paul Swartz has done some work in this area, based off the Tk Canvas
widget, for a "GUI" ssh client he wrote.  The relevant bits are distributed
with Twisted, in the conch/ui package.

  It is very slow, but I think it supports a lot of vt100.  It might be
worth basing your work off of.

  ViewCVS link for the package is 

     http://twistedmatrix.com/users/jh.twistd/viewcvs/cgi/viewcvs.cgi/twisted/conch/ui/?cvsroot=Twisted

  HTH,

  Jp

-- 
Somewhere, something incredible is waiting to be known.
                -- Carl Sagan
-- 
 up 24 days, 15:01, 1 user, load average: 0.99, 0.90, 0.85




From jack at performancedrivers.com  Wed Apr  9 16:05:51 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Wed, 9 Apr 2003 16:05:51 -0400
Subject: string() - Is this some kind of sick joke?
In-Reply-To: <36dbf824.0304091142.527af834@posting.google.com>; from chlump@yahoo.com on Wed, Apr 09, 2003 at 12:42:44PM -0700
References: <36dbf824.0304091142.527af834@posting.google.com>
Message-ID: <20030409160551.B1039@localhost.localdomain>

On Wed, Apr 09, 2003 at 12:42:44PM -0700, Daniel Gowans wrote:
> There is some sick joke in the python 2.2.2 documentation I have. 
> This is located in the ActivePython documentation:
> 
> strip(s[, chars]) 
>   Return a copy of the string with leading and trailing characters
> removed. If chars is omitted or None, whitespace characters are
> removed. If given and not None, chars must be a string; the characters
> in the string will be stripped from the both ends of the string this
> method is called on.
> 
> The same goes (in the docs) for rstrip and lstrip.
> 
> But does it actually work that way?  No!

I just checked my 2.2.1 and CVS HEAD versions of stringobject.c
for 2.2.1, it takes no args (and the docstring says it takes no args)
for 2.3 (CVS HEAD) it takes an optional argument

You might be looking at the docs for one version of python, and using another.
Or it could be a documentation incongruity in 2.2.2

try this in the interpreter

Python 2.2.1 (#4, Sep  1 2002, 20:16:35) 
[GCC 2.95.4 20011002 (Debian prerelease)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>a ="hi Mom!"
>>help(a.strip)
>>> help(a.strip)
Help on built-in function strip:

strip(...)
    S.strip() -> string
    
    Return a copy of the string S with leading and trailing
    whitespace removed.

>>> a.strip("!")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: strip() takes no arguments (1 given)




From staschuk at telusplanet.net  Tue Apr 29 19:24:35 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 17:24:35 -0600
Subject: Another itertool function?
In-Reply-To: <TXCra.34117$K35.1091440@news2.tin.it>; from aleax@aleax.it on Tue, Apr 29, 2003 at 10:23:47PM +0000
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <45pra.20060$xw4.19534@nwrdny01.gnilink.net> <C0ura.19455$3M4.563854@news1.tin.it> <mailman.1051650464.17483.python-list@python.org> <TXCra.34117$K35.1091440@news2.tin.it>
Message-ID: <20030429172435.B2224@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
> Steven Taschuk wrote:
  [...]
> > Alexander Schmolck proposed such a function back in February [1];
> > his version included a "step" argument doing what you suggest.
> > 
> > In the discussion of whether such a function was useful enough to
> > be included in itertools, I provided several use cases [2], but
> > was unable to find a convincing one which needed a step other than
> > one.  Do you have one in mind?
> 
> Simplest example: I have a flat sequence k1 v1 k2 v3 k3 v3 ... and I
> want a mapping k1->v1, k2->v2 etc.  dict(windowed(x, 2, 2)) would do it.

Ah, yes.  Sorry, I should have mentioned the case step = window
size, which I agree is important.  In Schmolck's original
proposal, that case even has its own function, xgroup.

> Second example: if I have a sequence of daily earnings and I want a "moving 
> daily average" -- for each day, the average of the seven days centered on
> it -- imap(sum, windowed(earnings, 7)) would suffice [ok, sum is not the
> same as avg, but dividing each number by 7 ain't hard if needed].  But say
> I want a "moving weekly average" instead -- for each WEEK, the average
> (or sum) of the three weeks centered on it.  Why not:
> imap(sum, windowed(earnings, 21, 7)) -- it seems simplest. Or maybe faster:
> imap(sum, windowed(imap(sum, windowed(earnings, 7, 7)), 3)).

With xgroup at hand, I feel xwindow(xgroup(earnings, 7), 3) is
clearer in this case than xwindow(earnings, 21, 7).  In full:

    def avg(x):
        return sum(x)/len(x)
    weeklyearnings = imap(sum, xgroup(earnings, 7))
    movavg = imap(avg, xwindow(weeklyearnings, 3))

(Imho: if you want a 3-week moving average, write one; don't write
a 21-day moving average that skips six out of seven days.)

> In general, 1 and N are going to be the most popular 'steps' for a
> window of length N, but other values may help for moving averages &c.

Agreed that 1 and n would be most common.  If there's no use case
for other values, I feel having a step argument in xwindow is a
premature generalization.  (Note that it can be added easily
later, if use cases surface.)

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From news at snakefarm.org  Wed Apr  2 14:52:59 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: Wed, 02 Apr 2003 21:52:59 +0200
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <435d40e0.0304011457.56e32bd0@posting.google.com> <sKSdnfwAUb0mrxejXTWcog@comcast.com> <435d40e0.0304021052.736293b3@posting.google.com> <b6fc6e$1fe$1@panix2.panix.com>
Message-ID: <b6ff2r$g1a$1@news.schlund.de>

Roy Smith wrote:
> Multi-line comments are evil
> 
> The first time you spend a half-hour trying to understand a block of
> ugly code and finally get to the end and find a "*/", you'll
> understand why they're evil.

ACK.

But multi-line strings are evil, too. At least when they occur in an 
indented block and leading whitespace is not allowed in the string 
your're building (e.g. in an HTTP request or something). That looks 
quite awful. :-)

cg.




From joze_storme at yahoo.com  Wed Apr 16 00:53:24 2003
From: joze_storme at yahoo.com (Joze)
Date: 15 Apr 2003 21:53:24 -0700
Subject: Sparse matrix data structures?
Message-ID: <20cb355e.0304152053.5d5caaec@posting.google.com>

Hi,

Are there any Python packages implementing some of the most usual data
structures for sparse matrices (coordinates, compressed rows,
compressed columns, etc.)? I found SparsePy on the web, but it doesn't
seem to be maintained any more, and the link to the source is broken.
How well do these integrate with numarray and/or SciPy?

Thanks for the help!
Joe.



From wesc at deirdre.org  Fri Apr  4 02:24:04 2003
From: wesc at deirdre.org (wesc at deirdre.org)
Date: Thu, 3 Apr 2003 23:24:04 -0800 (PST)
Subject: ANN: Python II course (Silicon Valley, CA)
Message-ID: <200304040724.h347O4S02758@alpha.ece.ucsb.edu>

this is a friendly reminder to ask you to please sign up
for Python II ASAP if you're interested.  the class starts
on Monday, April 14th... more details here:

http://artemis.ucsc-extension.edu/~wesc/024d50cd.htm

Register online, or contact Sherry at 408-861-3765 or
smirkarimi at ucsc-extension.edu to enroll.

On a related note, this will be the last course i will be
instructing for UC Santa Cruz Extension for a long time.
Work, family, and the Python community are making it hard
to devote the time necessary.  if you're interested in
training new Silicon Valley engineers about the wonders of
Python, contact me privately, and i will help u get started.

-wesley

ps. 'twas great seeing familiar & new faces @ PyCon last wk!

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

"Core Python Programming", Prentice Hall PTR, (c)2001
    http://starship.python.net/crew/wesc/cpp/

Silicon Valley-San Francisco Bay Area Python Users Group (BayPIGgies)
    http://www.baypiggies.net

wesley.j.chun :: wesc at deirdre.org or wesc at fuzzyorange.com
cyberweb.consulting : henderson, nv : cyberweb at rocketmail.com
http://roadkill.com/~wesc/cyberweb/

E-Mail using vi(m)/emacs with M$ outlook:
    http://groups.google.com/groups?threadm=fcd9edc1.0208281104.752da4bd%40posting.google.com




From neal at metaslash.com  Mon Apr  7 17:27:28 2003
From: neal at metaslash.com (Neal Norwitz)
Date: Mon, 07 Apr 2003 17:27:28 -0400
Subject: segfault with imp.load_module()
In-Reply-To: <1049561326.22972.14.camel@musicbox>
References: <mailman.1049534708.5967.python-list@python.org>
 <pan.2003.04.05.13.23.43.694594.12686@metaslash.com>
 <1049561326.22972.14.camel@musicbox>
Message-ID: <20030407212727.GC9536@epoch.metaslash.com>

On Sun, Apr 06, 2003 at 04:48:33AM +1200, Rob Brown-Bayliss wrote:
> On Sun, 2003-04-06 at 01:23, Neal Norwitz wrote:
> 
> > What version of python, what platform, what compiler, what options?
> > Assuming item = 'oda' this code works for me.
> 
> Redhat 8
> [rob at musicbox rob]$ python
> Python 2.2.1 (#1, Aug 30 2002, 12:15:30)
> [GCC 3.2 20020822 (Red Hat Linux Rawhide 3.2-4)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> 
> 
> If item = "abc" then I get (as expected) ImportError: No module named
> abc
> 
> So any way, I did some more playing and my module had a line:
> 
> from bhh import verbose
> 
> But the bhh module is not in the path, it's in the parent of the
> ./items/ directory.  If I remove that line then there is no segmentation
> fault.

As Michael mentioned, if you can provide the source to a complete
example, we can probably fix the bug.  Please open a SF bug report.

Neal




From theller at python.net  Wed Apr 16 13:39:57 2003
From: theller at python.net (Thomas Heller)
Date: 16 Apr 2003 19:39:57 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com>
Message-ID: <d6jmqr2q.fsf@python.net>

"Terry Reedy" <tjreedy at udel.edu> writes:

> "Thomas Heller" <theller at python.net> wrote in message
> news:n0iqy1zg.fsf at python.net...
> ======================================================================
> > FAIL: test_failUnlessEqual (__main__.FailingTests)
> > --------------------------------------------------------------------
> --
> > TestFailed: 0 != 1
> >   File "xunit.py", line 12, in test_failUnlessEqual
> >     self.failUnlessEqual(self.a, self.b)
> 
> I like having the actual values too.  It should make debugging an
> unexpected failure easier.

Actually, this is nothing that is new (as I also only discovered today).
You just have to use self.failUnlessEqual(x, y) instead of
self.failUnless(x == y).

>   But what if each value is more that a few
> bytes?  Example (from my future plans): two lists of all permutations
> of range(8), one produced by an obvious recursion and the other by a
> much less obvious iteration.

You are right, this is a pain. In unittest, in doctest, whereever
long lists are printed. It gets somewhat easier for the eye if
pprint.pprint() is used instead.


> Hmmm.  What I would really like is a structure comparison function
> that spit out the first difference found  so I did not have to print
> each and compare by eye.  That is not directly in the scope of your
> problem, but it would be the most useful succinct display.

As you say, it's a different problem.
The patches simply rearrange the output.

Thomas



From pinard at iro.umontreal.ca  Sun Apr 13 08:46:31 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 13 Apr 2003 08:46:31 -0400
Subject: Pyrex - undefined struct/union
In-Reply-To: <b7b7ht$jga$2@as201.hinet.hr>
References: <b7b7ht$jga$2@as201.hinet.hr>
Message-ID: <oqfzomlg48.fsf@titan.progiciels-bpi.ca>

[Haris Bogdanovic]

> I have a problem with this piece of code :

> cdef extern from "portaudio.h":
>     struct PaDeviceInfo:
>         int structVersion

[...]

> [...] when compiling C file (generated by pyrexc) I get errors that left
> from all this "info" struct members is an undefined struct/union.  What
> does that mean and how can I fix it ?

When you create declarations within a `cdef extern from "portaudio.h"', you
tell Pyrex about these so it can understand your code, but at the same time,
you ask it to _not_ bother to generate declarations in the C generated code,
on the premise that `#include "portaudio.h" will have done it already.

So, in a word, if C `struct PaDeviceInfo' is declared in "portaudio.h", it
is OK to include the Pyrex `struct Pa PaDeviceInfo' within the `cdef
extern'.  If C `struct PaDeviceInfo' is _not_ declared in "portaudio.h", you
should declare it to Pyrex outside any `cdef extern', and then Pyrex will
forward a C declaration in the C code it generates.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From andrew at acooke.org  Thu Apr 24 10:43:59 2003
From: andrew at acooke.org (andrew cooke)
Date: Thu, 24 Apr 2003 10:43:59 -0400 (CLT)
Subject: Data-driven testing
In-Reply-To: <X8Mpa.30197$y3.2398527@news010.worldonline.dk>
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> 
     <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> 
     <X8Mpa.30197$y3.2398527@news010.worldonline.dk>
Message-ID: <2092.200.27.214.82.1051195439.squirrel@127.0.0.1>

i'm not a great testing enthusiast (it's just another tool, not a magic
bullet), but sometimes you can write tests during exploratory coding.  in
general, for example, you can have a spearation between "invoking" the
code and the implementation - then i'd put code for running the program
and printing the output (or whatevere) in a "test" (even if it doesn't
actually test anything).  then i play with the implementation, running the
"test" to see what teh results are.  once i like the way things are going,
the test can be extended to include some assertions.

this isn't always possible - it works for the classes where you can add a
main method to play around from the command line.  the test simple
replaces that main method...

andrew

Max M said:
> While I do understand the principles of tdd, testing first and all that,
> I have a hard seing why it should be an advantage to do it in the early
> stages of a project.
[...]

-- 
http://www.acooke.org/andrew




From johnroth at ameritech.net  Wed Apr 23 20:23:13 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 23 Apr 2003 20:23:13 -0400
Subject: Creating dynamically named lists/dictionaries
References: <ydFpa.15$IO5.10324@newsfep2-gui.server.ntli.net>
Message-ID: <vaebjej8ov4427@news.supernews.com>

"Rogue 9" <rogue9 at ntlworld.com> wrote in message
news:ydFpa.15$IO5.10324 at newsfep2-gui.server.ntli.net...
> Hi all,
> I'm new to both programming and Python in particular and have run in
to a
> snag while trying to write my own program.I have looked through books
and
> the web but haven't as yet found an answer so I though I might try a
post
> here to see if anyone could point me in the right direction.
> Okay my program is for analysing a Lottery Draw results text file I
have
> pre-formatted so that each line contains 7 comma separated values
(Draw
> number and six values representing the numbers drawn in numerically
> ascending order.
> I've sussed out how to open the file read in the lines and I have
split
> the strings first by the '\n' newline symbol and then I have split it
> again by ','.I've also turned the values back into integers as they
were
> strings.Now what I've ended up with is a list containing one draw
result
> for each week (768 at the last count)as individual items in that list
e.g
> list =[[767,11,22,33,44,45,46]......[1,20,33,41,45,46,47]]
> What I really want is a list or dictionary for each individual weeks
> result but I can't seem to work out how to dynamically create a list
or
> dict with a unique name for each week.

You've already got a list for each week. Since I don't know where you're
going with this (that is, I don't know what you mean by "analyze") I
can't give you more specific answers.

So I'm going to assume that what you want is to be able to find the
results by week number. The following (untested) code should give
you an idea.

biglist.sort() # turn the list around so week 1 is at the front.

aWeek = biglist[week-1][1:] # gives you a list of the six result numbers
for the week.

> The results file will grow and I
> don't fancy creating 768+ lines of code just to instantiate a list or
> dict.At first I tried looping a variable equal to the length of the
file
> as in:
> x=len(filelength)
> 'draw'+str(x)=[drawresult]
> I thought this would dynamically make a whole bunch of lists such as
> draw1,draw2 etc. but all I got was errors.

Unfortunately, you can't do that. There is a way to add names to
a namespace in Python, but it's not what you want. If you want to create
a name for each list, you need a dict. The following (untested) code
will
build a dict with a name for each week of the lottery, and with the six
numbers as a list.

weekDict = {} # create a dictionary
for weeksNumbers in biglist: # biglist is the list you built on your
input
    weekDict["draw" + str(weeksNumbers[0])] = weeksNumbers[1:]  # notice
the way it slices the week!

The keys in weekDict will now be "draw1", "draw2" and so forth,
assuming your input list is the one you mentioned before.

> I hope I have explained my 'problem' sufficiently for someone to help
or
> to at least point me in the right direction to work towards a
solution.
> Yours hopefully,
> Rogue9

HTH
John Roth





From joe at strout.net  Wed Apr  9 13:29:40 2003
From: joe at strout.net (Joe Strout)
Date: Wed, 09 Apr 2003 17:29:40 GMT
Subject: no Python mail archiving utility?
Message-ID: <joe-17FB5F.10302009042003@orngca-news02.socal.rr.com>

Yesterday morning I posted asking where to find a web mail 
archive/search program written in Python.  I got nothing but the sound 
of chirping crickets in response.

Does this mean that no such utility exists?  Are all the dozens of 
mailing lists about Python-related topics serving up their archives via 
programs written in C and Perl?

Thanks,
- Joe

,------------------------------------------------------------------.
|    Joseph J. Strout         Check out the Mac Web Directory:     |
|    joe at strout.net           http://www.macwebdir.com             |
`------------------------------------------------------------------'



From sjmachin at lexicon.net  Sat Apr 19 21:53:50 2003
From: sjmachin at lexicon.net (John Machin)
Date: Sun, 20 Apr 2003 01:53:50 GMT
Subject: When not to use an RE -- an example
References: <c76ff6fc.0304191459.2adf73a@posting.google.com> <7f9e1817.0304191715.4a9cd49c@posting.google.com>
Message-ID: <3ea1f9fd.23623538@news.lexicon.net>

On 19 Apr 2003 18:15:56 -0700, imcmeans at telus.net (Ian McMeans) wrote:

>How about using a dictionary? I wish we had dictionary comprehensions
>:(
>
>>>> user_input = " ZZZZZZZ\n\n"
>>>> d = dict( [(x, None) for x in list(user_input.strip())] )
>>>> len(d.keys())
>1
>

I don't understand what your dictionary solution is doing. My need is
to detect a large list of possible junk strings -- say 95 printable
ASCII characters by say 30 characters width of the text field; loading
all of those into a dictionary is not exactly an elegant solution.

>I don't really understand your regular expression...
>"^(?:(.)(?=\1))+\1\Z"
>Why the lookahead assertion following the first capture group,
>followed by the capture group outside the parantheses? It seems crazy
>to me.

Crazily overcomplicated, as already pointed out by Alexander, yes. But
it worked. Whether the groups are capture groups or not is irrelevant
to the meaning.

>
>".{2,}" is repeated characters, ".{2,}\s*$" is repeated characters
>followed by optional whitespace, with nothing following it.

".{2,}" does *not* match repeated characters. It matches *any* string
of length two or more.

All of the above  strengthens my original implicit proposition: don't
*try* to use an RE when you don't need to.

Of course there will always be the folks who think that Friedl's big
e-mail address matcher is a shining example of software engineering
good practice :-)







From gustav at morpheus.demon.co.uk  Wed Apr 16 17:30:13 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Wed, 16 Apr 2003 22:30:13 +0100
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <3e9af967$0$49112$e4fe514c@news.xs4all.nl>
 <ptnosvmm.fsf@morpheus.demon.co.uk>
 <slrnb9qb9m.1bc.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <of36qgey.fsf@morpheus.demon.co.uk>

Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> writes:

> In article <ptnosvmm.fsf at morpheus.demon.co.uk>, Paul Moore wrote:
> <snip> 
>> PS Is there any way of forcing the Windows 2000/XP console to display
>>    Unicode in any form (UTF-8, maybe) rather than having to deal with
>>    code pages???
>
> UTF-8 is code page 65001.

Thanks. I didn't know that.

> Strangely, though, I get 'permission denied' when I run "chcp 65001" and
> then try to print a UTF-8-encoded Euro sign.  I don't know what could be
> going wrong there.

Someone followed up to this on python-dev. There seems to be some
level of extreme oddness with the Windows console running in UTF-8
mode.

C:\Data
>chcp 65001
Active code page: 65001

C:\Data
>python -c "print u'h\xf2l\xe1'.encode('utf-8')"
h?l?
Traceback (most recent call last):
  File "<string>", line 1, in ?
IOError: [Errno 2] No such file or directory

Using TYPE in CMD.EXE to type a file containing UTF-8 using codepage
65001 seems to work. But in 4NT.EXE it doesn't...

Sigh....

Paul.
-- 
This signature intentionally left blank



From aleax at aleax.it  Tue Apr 29 10:00:43 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 14:00:43 GMT
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com> <zAara.22169$K35.739509@news2.tin.it> <roy-621090.18204128042003@reader1.panix.com>
Message-ID: <fAvra.30824$K35.974064@news2.tin.it>

Roy Smith wrote:

> Alex Martelli <aleax at aleax.it> wrote:
>> No statistically significant / measurable difference.  The tiny
>> difference in conciseness is hardly compelling one way or another,
>> either.
> 
> I disagree (respectfully).
> 
> I'm all about writing code that's easier to read.  I'll admit that the
> += form is only marginally simplier to read, but marginal is better than
> nothing.

When what you're incrementing is a long name or an expression of
any substantial complication, then "left += right" may indeed offer
_significant_ advantages.

When what you're incrementing is "i" (as in the original question),
this is quite moot -- the alleged "increase in simplicity" is so
small that experimental error may be < 0...;-).

And I don't think that "one obvious way to do it" applies here --
just as, when I'm adding two numbers a and b, I'm pretty happy to 
get to choose between a+b and b+a depending on what's more readable
in the given context, quite similarly the choice between "i=i+1"
and "i+=1" is not necessarily preordained

> I tend to pick-and-choose from the XP menu, but one of the items I
> really like is "refactor mercilessly".  Why write a variable name twice
> when you can write it once?  Easier to read, and less error prone.  Have

When the variable name is short and sharp, writing it twice is no
problem -- and sometimes it can be more readable e.g. by showing
some parallelism:

    if override_with_new:
        current = incoming_new + increment
    else:
        current = current + increment

The choice between this, the "refactored" forms such as:

    if override_with_new:
        current = incoming_new
    current += increment

etc, etc, is indeed strictly one of readability, and thus of style.
I'm glad to have += as a CHOICE, but wouldn't like it as a MUST:-).


Alex




From tungwaiyip at yahoo.com  Thu Apr 24 18:58:17 2003
From: tungwaiyip at yahoo.com (Tung Wai Yip)
Date: Thu, 24 Apr 2003 15:58:17 -0700
Subject: Exception and finally question
Message-ID: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com>

Hello,

How to I access the exception object in a handler?

try:
    raise NameError, 'HiThere'
except NameError:
    #how to print the message 'Hi There' here?

Why except and finally can't be used together? I like to log a error
message if an exception is throw. And I a file to be guaranteed to be
close. What can't I do them together?

Wai Yip Tung



From tj_scarlet at yahoo.com  Wed Apr 23 21:03:40 2003
From: tj_scarlet at yahoo.com (Tj)
Date: 23 Apr 2003 18:03:40 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <1azna.115402$It5.22896@news2.central.cox.net> <tyfhe8xqhmy.fsf@pcepsft001.cern.ch> <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com> <b86lc3$tnd$1@plover.com> <ccc7084.0304231430.6b81effb@posting.google.com>
Message-ID: <ccc7084.0304231703.6c55570b@posting.google.com>

Hmm, must be sleepy because I could have sworn that you wrote that you
asked him this question, not Tom Christiansen.

Tj

tj_scarlet at yahoo.com (Tj) wrote in message news:<ccc7084.0304231430.6b81effb at posting.google.com>...
> mjd at plover.com (Mark Jason Dominus) wrote in message news:<b86lc3$tnd$1 at plover.com>...
> > I don't believe that's correct.  At an early Open Source Conference,
> > perhaps around 1998 or so, I was present at a conversation between
> > Guido and Tom Christiansen.  Tom asked Guido why Python didn't have
> > closures, and Guido's reply (and I think this is an exact quote) was
> > "Closures aren't important."
> > 
> > I remember it because I was so astonished by the obtuseness of the answer.
> 
> Well, it seems the burden of proof is on you.  Please explain the
> usability advantages of closures vs. Python's current scope system.



From aleax at aleax.it  Tue Apr 22 16:12:04 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 20:12:04 GMT
Subject: instance creatiion as an "event"
References: <mailman.1051021054.11597.python-list@python.org>
Message-ID: <omhpa.76858$T34.1982274@news2.tin.it>

Arthur wrote:

> Alex:
> 
>>I don't understand this. When you call:
> 
>>p = TestPoint(args)
>>p.redraw()
> 
>>it's EXACTLY AS IF TestPoint.__init__'s very last statement was
> 
>>self.redraw()
> 
> Standalone, yes.  But, unless I'm missing something, the effects in an
> inheritance chain is different. If I inherit from TestPoint, and need to
> call TestPoint.__init__ in the Point_Child.__init__ to pick up
> initilization routines common to the chain, I will now unavoidably be
> calling TestPoint's
> redraw method before I call Point_Child's.

Not with the once-habitual Python idiom for that purpose.  Try it!

class TestPoint:
    def __init__(self):
        self.tp = 23
        self.redraw()
    def redraw(self):
        print 'TestPoint.redraw', self.tp

class PointChild(TestPoint):
    def __init__(self):
        self.pc = 42
        TestPoint.__init__(self)
    def redraw(self):
        print 'PointChild.redraw', self.tp, self.pc

p = PointChild()


and not with super either, in the new object model:


class TestPoint(object):
    def __init__(self):
        self.tp = 23
        self.redraw()
    def redraw(self):
        print 'TestPoint.redraw', self.tp

class PointChild(TestPoint):
    def __init__(self):
        self.pc = 42
        super(PointChild, self).__init__()
    def redraw(self):
        print 'PointChild.redraw', self.tp, self.pc

p = PointChild()


Of course, this relies on the ability of being able to call the superclass's 
__init__ at the APPROPRIATE point in the subclass's __init__ ...!  If you
cannot rely on that, e.g. because implementation issues are just too
intricate, use the wrapper factory functions I suggested in the previous 
post (or similar but more sophisticated approaches based e.g. on a
custom metaclass!).


  When the main reason I am
> inheriting from TestPoint is to override its redraw method, and in fact to
> avoid calling it.  If TestPoint has attributes that Point_Child does not
> that are referenced in its redraw method, I will have an attribute error
> on the calling of TestPoint's redraw form Point_Child.

I don't know what you're saying.  TestPoint is the baseclass.  How can
it have attribute that its subclass PointChild lacks?!?!?!

I'm not just thinking of the theoretical horror of that.  In practice
too, instance attributes of an instance of TestPoint will be set in
TestPoint.__init__ -- if you call that from PointChild.__init__, in
whatever way (as you normally should), then the instance WILL have
those attributes set, period.

> suggestion as to working with the condition
> 
> if hasattr(sys, 'ps1')

That's for when you want to do different things in interactive
_sessions_ than in programs (doing different things at an interactive
_prompt_ is subtler -- and perhaps best achieved by wrappers, as
per my previous post).


Alex




From claird at lairds.com  Thu Apr 24 19:55:29 2003
From: claird at lairds.com (Cameron Laird)
Date: Thu, 24 Apr 2003 23:55:29 -0000
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <va956b925cj1df@corp.supernews.com> <25a672ee.0304231444.1ce0dd1b@posting.google.com> <r_Lpa.89137$T34.2223764@news2.tin.it>
Message-ID: <vagubhpul80f76@corp.supernews.com>

In article <r_Lpa.89137$T34.2223764 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>Denis Agar wrote:
>   ...
>> Ok, try to ignore the thickness of my skull, but you're saying that it
>> is possible to create progams using mfc42.dll, a GUI that most people
>> have already? I have the win32 extensions and am in the process of
>
>Yes, with Python and a simple add-on you can choose to reuse any DLL
>you may have lying around.  win32all (PythonWin in particular) does
>that for MFC specifically.  Other add-ons such as the oldie (but
>goldie) calldll, and the newer (and more powerful) ctypes, let you
>do that for ANY dll (with more Python-side work to adapt it).  Good
>luck -- it IS, of course, lower-level programming than typical Python
>work, since you want to use DLLs designed for lower-level languages.
			.
			.
			.
>The anygui project, anygui.sf.net, aims to design a uniform wrapper
>that can use most any GUI toolkit as a back-end -- this would include
>(at least potentially) MFC, or the bare Win32 API's, etc.  It's not
			.
			.
			.
All true, of course.

I want to emphasize again that, for the specific question
of construction of Python-scripted GUI applications, it is
not necessary to comprehend the full generality of these
possibilities.  Mark Hammond and his beneficiaries have 
successfully used win32all (or PythonWin) to write such
programs.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From max at alcyone.com  Fri Apr 18 14:01:08 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 18 Apr 2003 11:01:08 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <mailman.1050634170.24420.python-list@python.org>
Message-ID: <3EA03D64.B5C5C33B@alcyone.com>

Steven Taschuk wrote:

> Reading this thread has made me wonder whether current practice is
> uniform.  It would not be desirable imho for the language
> reference to say, "people do different things with __repr__,
> including <list of three things>", and not endow one of them with
> some normative weight.
> 
> ... How about this?  (Draft.)

That sounds far superior to the current documentation, since it is more
verbose about what __str__ and __repr__ are used for in the field.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If you think you're free, there's no escape possible.
\__/ Baba Ram Dass
    Esperanto reference / http://www.alcyone.com/max/lang/esperanto/
 An Esperanto reference for English speakers.



From mariobot at yahoo.com  Tue Apr  1 18:25:56 2003
From: mariobot at yahoo.com (Mario A. Botto)
Date: 1 Apr 2003 15:25:56 -0800
Subject: adjacent differences with a list comprehension
References: <69413f9.0303241331.62f0295f@posting.google.com>
Message-ID: <5f93bf16.0304011525.41e6a3a@posting.google.com>

pschmidt at omnimn.com (Phil Schmidt) wrote in message news:<69413f9.0303241331.62f0295f at posting.google.com>...
> Given a list of numbers, such as:
> 
> L = [2, 5, 8, 3, 9, 1]
> 
> I want to generate a list containing the differences between adjacent
> elements, i.e.,
> 
> Ld = [3, 3, -5, 6, -8]
> 
> I don't see how I can do this (easily/elegantly) with list
> comprehensions.
> 
> Any suggestions? Thanks!

Hi.
How about this? 
If, say, L1 = [a1, a2, a3], L2 = [b1, b2, b3, b4, b5], then 
zip(L1, L2) = [(a1,b1), (a2,b2), (a3, b3)]. Hence, 
[a - b for (a,b) in zip(L1, L2)] = [a1-b1, a2-b2, a3-b3]

This solves your problem if you set L1 = L, L2 = L[1:]. That is,

Ld = [a - b for (a,b) in zip(L[1:], L)]
print 'Ld: ', Ld
>> Ld:  [3, 3, -5, 6, -8]
 
Cheers.



From patrick at fraley.de  Thu Apr 17 08:34:16 2003
From: patrick at fraley.de (Patrick W. Fraley)
Date: Thu, 17 Apr 2003 14:34:16 +0200
Subject: Zope Sequences and Python Scripting
Message-ID: <pan.2003.04.17.12.34.15.777976@fraley.de>

Hi,

first off:  I am pretty new to python and have just started with Zope, so
please excuse and fopars on my side :)

I have a python script which returns a [] containing {}.  Now I can go
thru the returned array (I know this is perl lingo, sorry that is my
background) by using the <dtml-in> method, but I can not
access the values of the dictionary contained within the sequence item.

And here is what I am trying to do with the returned array:

<dtml-in filter_row>
<dtml-var filter_id><br>
<dtml-var bez_ref><br>
<dtml-var col_span><br>
<hr>
</dtml-in>

Here is the python script (filter_row):

# Import a standard function, and get the HTML request and response objects.
from Products.PythonScripts.standard import html_quote

request = container.REQUEST
RESPONSE =  request.RESPONSE

colspan = 1
curFilterId = -1
curFilter = {}

seqFilters = []

for spalte in context.get_spalten_liste(tabellen_id=request.tabellen_id):
  if curFilterId != spalte.filter_id:
    if len(curFilter)>0:
      if curFilterId>0:
        curFilterData = context.get_filter(filter_id=curFilterId)
        curFilter['bez_ref'] = curFilterData[0].bez_ref
        curFilter['filter_id'] = curFilterId
        seqFilters.append(curFilter)
      else:
        curFilter['filter_id'] = curFilterId
        curFilter['bez_ref'] = "&nbsp;"
        seqFilters.append(curFilter)

    curFilterId = spalte.filter_id
    curFilter = {}
    curFilter['col_span'] = 1
  else:
    curFilter['col_span'] = curFilter['col_span'] + 1

if len(curFilter)>0:
  if curFilterId>0:
    curFilterData = context.get_filter(filter_id=curFilterId)
    curFilter['bez_ref'] = curFilterData[0].bez_ref
    curFilter['filter_id'] = curFilterId
    seqFilters.append(curFilter)
  else:
    curFilter['bez_ref'] = "&nbsp;"
    seqFilters.append(curFilter)

return seqFilters





From donnal at donnal.net  Thu Apr 17 05:39:07 2003
From: donnal at donnal.net (Donnal Walter)
Date: Thu, 17 Apr 2003 04:39:07 -0500
Subject: wxPython: Intercepting keyboard events from multiple widgets?
In-Reply-To: <b7le9j$266le$1@ID-169208.news.dfncis.de>
References: <b7le9j$266le$1@ID-169208.news.dfncis.de>
Message-ID: <b7lskp$l5l$1@main.gmane.org>

Greg Ewing (using news.cis.dfn.de) wrote:
> I have a wxPython problem:
> 
> I have a wxNotebook, each wxPanel of which contains a
> number of controls.
> 
> I want to catch certain keystrokes and use them for
> things pertaining to the whole collection of controls,
> such as switching from one tab of the notebook to another.
> 
> Is there any way I can do this without having to attach
> event handlers individually to every control?
> 
> In the wxWindows docs under "Pluggable Event Handlers"
> it hints at it being possible to do things like
> intercepting all the events directed at a dialog box,
> but it doesn't go into any details, and it doesn't
> seem to be obvious how you would do this. Attaching
> an event handler to the dialog itself doesn't seem
> like it would work, because keyboard events seem to
> originate in whatever control has the keyboard focus
> and propagate upwards.
> 
> Can anyone point me to an example which shows how
> to do this, if it's really possible at all?


Would an accelerator table do the job?

http://www.wxwindows.org/manuals/2.4.0/wx28.htm#wxacceleratortable

You might also take a look at wxPython.lib.evtmgr.py

Regards,
Donnal Walter
Arkansas Children's Hospital






From carsten at gehling.dk  Tue Apr 29 07:43:32 2003
From: carsten at gehling.dk (Carsten Gehling)
Date: Tue, 29 Apr 2003 13:43:32 +0200
Subject: SV: Function to convert fetchone() result from a list to a dictionary
In-Reply-To: <b8lfrr$21c$1@nntp0.reith.bbc.co.uk>
Message-ID: <BHEFKIOIMEDGILGDACKDKEBEDFAA.carsten@gehling.dk>

> Fra: python-list-admin at python.org
> [mailto:python-list-admin at python.org]Pa vegne af Neil Padgen
> Sendt: 29. april 2003 11:16

> Take a look at the cursorclass argument:

<snip>

> >>> c = db.cursor(cursorclass=MySQLdb.cursors.DictCursor)

<snip>

Hmm... Nice. However, is this MySQLdb.cursors.DictCursor available in other
DB modules - say, if I someday changed to (shudder) Postgres or Oracle?

- Carsten







From tjreedy at udel.edu  Thu Apr 10 17:39:22 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Thu, 10 Apr 2003 17:39:22 -0400
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com> <b73vov$d92$00$1@news.t-online.com> <HzGdnQNo1Z2-AAijXTWcog@comcast.com> <b74ckl$itb$1@newsg4.svr.pol.co.uk>
Message-ID: <8c2dnXyBdvemeQijXTWcqw@comcast.com>

"Duncan Smith" <buzzard at urubu.freeserve.co.uk> wrote in message
news:b74ckl$itb$1 at newsg4.svr.pol.co.uk...
> Yes, you do have the link function as well as the linear predictor,
and the
> cut-off would be 0.5 rather than 1.  But in terms of absolute
efficiency
> there is (according to the literature) little to choose between
linear
> discriminant analysis and logistic regression.  Changing the cut-off
would
> be a simple way of attempting to minimise Alex's 'error-cost'.

For at least some types of data distributions, I can believe that.  My
comment about stepwise construction of a feature set versus
over-fitting with everything you can think of applies equally well to
logistic regression.  As far as I know, logistic regression cannot be
used for more than two mutually-exclusive groups.  I don't know
whether Alex might ever need to.

Terry J. Reedy





From andrews at pcug.org.au  Tue Apr 29 05:42:14 2003
From: andrews at pcug.org.au (Avery Andrews)
Date: Tue, 29 Apr 2003 09:42:14 GMT
Subject: Tkinter exit problem in Win98
Message-ID: <Pine.GSO.4.21.0304291934400.25385-100000@supreme.pcug.org.au>

Has anyone encountered a problem with Tkinter apps not shutting down
properly in Wi98?  I find that they (forex the hello example on pg 309 of
Learning Python, and some actually useful ones I have written) tend not to
terminate properly, but leave the Python window up and hard to close
(sometimes 3f salute/end task will shut it down, sometimes not).

This happens with a clean install of Win98, and both P22 and Py23b1.
Doesn't matter whether the script is run from a command window or by
clicking on the file name in the explorer.

 thanks
   Avery Andrews






From jepler at unpythonic.net  Fri Apr 25 11:32:21 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 25 Apr 2003 10:32:21 -0500
Subject: pythong referencing/aliasing
In-Reply-To: <ylcqa.9116$h%2.963225@read1.cgocable.net>
References: <ylcqa.9116$h%2.963225@read1.cgocable.net>
Message-ID: <20030425153219.GB30690@unpythonic.net>

A simple assignment such as
    v = expr
only re-binds v to the result of expr.  There is thus no way to directly
translate C code such as
    void set_to_one(int &x) { 
        x = 1;
    }
into Python.

A complex assignment, such as
    x.y = expr
    x[y] = expr
does not re-bind the name "x".  Instead, it mutates (changes) the thing
named by x.  If something else names the same object, it will continue
to refer to the same object (with a new value).

Of course, only some objects are mutable in Python.  Others (strings,
tuples, and numbers) are immutable.

This is a fairly common FAQ.  You may find that reading the FAQ will
give you a clearer explanation.  I can only add one other piece of
information: I rarely am hurt by the fact that Python doesn't have
references of the type you're talking about.

Jeff




From andrew at acooke.org  Wed Apr 23 19:15:11 2003
From: andrew at acooke.org (andrew cooke)
Date: Wed, 23 Apr 2003 19:15:11 -0400 (CLT)
Subject: Why functional Python matters
In-Reply-To: <ccc7084.0304231430.6b81effb@posting.google.com>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> 
     <1azna.115402$It5.22896@news2.central.cox.net> 
     <tyfhe8xqhmy.fsf@pcepsft001.cern.ch> 
     <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com> 
     <b86lc3$tnd$1@plover.com> 
     <ccc7084.0304231430.6b81effb@posting.google.com>
Message-ID: <58789.127.0.0.1.1051139711.squirrel@127.0.0.1>

a related question - why does queue have to be passed to helper() in the
code below?  i tried to define it inside walktree (but outside helper)
since it's global to the function, but python objects saying that the
local variable "queue" in helper is accessed before it's defined.  this
seems to be an example of where closures would work perfectly...

#!/usr/bin/python2.2

from __future__ import generators # needed for Python 2.2
import os

def walktree(basepath=".", postorder=True, depthfirst=True,
             ignorelinks=True):

    """Noah Spurrier's code, modified to allow depth/breadth-first
    traversal.  The recursion is there *only* to allow postorder
    processing as the stack rolls back - the rest of the algorithm is
    imperative and queue would be declared outside helper if I knew
    how."""

    # why can't queue be defined here as [basepath] (and then omitted
    # from calls to helper).

    def helper(queue):

        while queue:

            if depthfirst: dir = queue.pop(-1)
            else: dir = queue.pop(0)

            children = os.listdir(dir)

            dirs, nondirs = [], []

            for name in children:
                fullpath = os.path.join(dir, name)
                if os.path.isdir(fullpath) and not \
                       (ignorelinks and os.path.islink(fullpath)):
                    dirs.append(name)
                    queue.append(fullpath)
                else:
                    nondirs.append(name)

            if not postorder:
                yield dir, dirs, nondirs

            if postorder:
                for rest in helper(queue):
                    yield rest
                yield dir, dirs, nondirs

    return helper([basepath])

def test():
     for basepath, dirs, nondirs in \
             walktree(postorder=False, depthfirst=False):
         for name in dirs:
             print os.path.join(basepath, name)
         for name in nondirs:
             print os.path.join(basepath, name)

if __name__ == '__main__':
         test()



tj_scarlet at yahoo.com said:

> mjd at plover.com (Mark Jason Dominus) wrote in message
> news:<b86lc3$tnd$1 at plover.com>...
>> I don't believe that's correct.  At an early Open Source Conference,
>> perhaps around 1998 or so, I was present at a conversation between
>> Guido and Tom Christiansen.  Tom asked Guido why Python didn't have
>> closures, and Guido's reply (and I think this is an exact quote) was
>> "Closures aren't important."
>>
>> I remember it because I was so astonished by the obtuseness of the
>> answer.
>
> Well, it seems the burden of proof is on you.  Please explain the
> usability advantages of closures vs. Python's current scope system.
>
> Tj
> --
> http://mail.python.org/mailman/listinfo/python-list
>
>


-- 
http://www.acooke.org/andrew




From gh at ghaering.de  Sun Apr 20 19:47:21 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: Mon, 21 Apr 2003 01:47:21 +0200
Subject: Keyboard Commands
In-Reply-To: <Pine.SOL.4.51.0304201825001.6411@herald.cc.purdue.edu>
References: <Pine.SOL.4.51.0304201825001.6411@herald.cc.purdue.edu>
Message-ID: <20030420234721.GB1084@mephisto.ghaering.test>

* richard.a.charts.1 <charts at purdue.edu> [2003-04-20 18:27 -0500]:
> Hey,
> I'm working on a program and I want to use keyboard shortcuts for some
> of my menu items.  I can't seem to find the way to implement this.
> Can anyone point me in the right direction?  ex.  Ctrl-x, ctrl-F1,
> etc.  [...]

Which toolkit are you using? wxPython, PyQt, Tkinter, curses, ..., ?

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/




From jolderdissen at ixiacom.com  Wed Apr 30 17:40:20 2003
From: jolderdissen at ixiacom.com (Jan Olderdissen)
Date: Wed, 30 Apr 2003 14:40:20 -0700
Subject: network programming problem??
References: <14bd1399.0304301236.7fe19be5@posting.google.com>
Message-ID: <p8g0bvs1hd2slcscdv0smoc34pm712cccp@4ax.com>

I strongly suspect there is something wrong with the DNS setup on that
Linux box. If, say, the first DNS server IP configured was bogus,
you'd expect symptoms like this. You might want to try the litaral IP
address instead to see if this is the case:

urlopen("http://194.109.137.226").read()

I ran the script locally on RH 7.3 and I got values at around 1s per
request.

Jan

>Here are the results of running the script:
>
>-------------------- LINUX --------------------------
>Run 0 . 16.69
>Run 1 . 16.54
>Run 2 . 17.04
>Run 3 . 16.55
>Run 4 . 16.55
>Run 5 . 16.54
>Run 6 . 16.55
>
>-------------------- WIN XP -------------------------
>Run 0 . 2.69
>Run 1 . 0.59
>Run 2 . 0.61
>Run 3 . 0.66
>Run 4 . 0.59
>Run 5 . 0.61
>Run 6 . 0.59




From fnord at u.washington.edu  Fri Apr 11 14:07:33 2003
From: fnord at u.washington.edu (Lonnie Princehouse)
Date: 11 Apr 2003 11:07:33 -0700
Subject: Loss of precision / data corruption with Tkinter.DoubleVar ?!
Message-ID: <f70e3538.0304111007.b0a72c3@posting.google.com>

Greetings, 

I stumbled across this curiosity yesterday:

>>> x = 0.33333333333333098
>>> x
0.33333333333333098     
>>> import Tkinter
>>> y = Tkinter.DoubleVar(Tkinter.Tk())
>>> y.set(x)
>>> y.get()
0.33333333333300003  

What the....?

Sure, it's only on the order of 10e-13 delta, 
but why is there any difference at all??


I hit this because at some point I was trying to do:

y.set(x)
.
.
.
if y.get() == x:
    ...

and of course it wasn't working properly!

I have tried this on Python 2.2 for Win32 and Linux/x86




-Lonnie Princehouse



From guy at obstruction.com  Tue Apr 22 14:45:52 2003
From: guy at obstruction.com (Guy Middleton)
Date: 22 Apr 2003 14:45:52 -0400
Subject: another python style question -- copying a data file
References: <zTepa.127177$Vzu.30273@news02.bloor.is.net.cable.rogers.com> <mailman.1051033654.17248.python-list@python.org>
Message-ID: <b842l0$elb$1@chaos.obstruction.com>

In article <mailman.1051033654.17248.python-list at python.org>,
=?ISO-8859-1?Q?Gerhard_H=E4ring?=  <gh at ghaering.de> wrote:
> Guy Middleton wrote:
> > Ok, here's another style question.
> > 
> > I want to copy a data file, but if the file doesn't already exist, I don't
> > care.  I do care about any other errors, such as wrong permissions.
> > 
> > I have the following code, is this a typical Python way to do this?
> > 
> >     try:
> >         os.stat(DATAFILE)
> >         os.system("cp %s %s.new" % (DATAFILE, DATAFILE))
> >     except OSError, arg:
> >         # ENOENT is ok, we didn't do the copy
> >         # anything else is a problem
> >         if arg.errno != errno.ENOENT:
> >             raise
> 
> Here's what I'd do:
> 
> if os.path.exists(DATAFILE):
>      shutil.copyfile(DATAFILE, "%s.new" % DATAFILE)

Aha, I didn't know about shutil.  This is obviously better.  :-)



From david.bear at asu.edu  Wed Apr 30 12:44:20 2003
From: david.bear at asu.edu (David Bear)
Date: 30 Apr 2003 09:44:20 -0700
Subject: win32 adding a tcpip alias to an interface
Message-ID: <m3el3k2avv.fsf@moroni.pp.asu.edu>

I've never used the win32 extensions to python.  Most of my python
tinkering has been around unix.  However, I am contemplating a script
that would add a second ip address to an interface.  Here's the
general idea:
1) when a windows box starts, it gets a dhcp lease on a non-routable
IP address, web browsing is handled through NAT
2) if the users needs a fully routable ip address, he authenticates to
something -- this could be Active Directory, Samba, Kerberos --
something else.
3) after successful authentication the user is given a fully routable
ip address 

So, the python script would need to handle 2 and 3.  There are
probably classes someone has written to do authentication (haven't
looked yet)  But, I'm worried about how one might add an ip address to
a windows tcpip stack.  Anyone have any ideas how to do this?

tia



From al2000 at udev.org  Sun Apr 27 00:30:55 2003
From: al2000 at udev.org (Antony Lesuisse)
Date: Sun, 27 Apr 2003 06:30:55 +0200
Subject: trouble with xml.sax and unknow entities
Message-ID: <20030427043055.GA13452@udev.org>

I'm not on the list, please cc: me the answers.

I'm having trouble to parse the folowing xml with the default python xml.sax
api. I'm using python2.2 on debian unstable powerpc (python2.2-xmlbase).

'<?xml version="1.0"?><html><body>hello &nbsp; </body></html>'

See the code at the end.

xml.sax._exceptions.SAXParseException: <unknown>:1:39: undefined entity

The parser halt on &nsbsp; because it doesn't know about this entity. The
problem is cannot find a way to tell him what this entity is.

(1)
Is there a way to have a callback the parser arrive on &nbsp; ? None of the
folowing handler functions (resolveEntity,notationDecl,unparsedEntityDecl) are
called.

I thought resolveEntity had to be called in that situation but i probably
misunderstand the sax api.

(2)
Is there a way to register entities before the parsing begin ?
Something like:
    parser.registerEntity('&nbsp;','blahblah')

(3)
Or is there a way to register an external DTD where those entities can be
defined ?  Something like:
    parser.registerExternalDTD('xhtml.dtd')

Thank you for your help.

-----------------------------------------------------------
#!/usr/bin/python
import StringIO,sys,xml.sax,xml.sax.handler

class CHandler(xml.sax.handler.ContentHandler):
    def startElement(self, name, attrs):
        print name
    def characters(self, ch):
        print ch.encode('Latin-1')

class EResolver(xml.sax.handler.EntityResolver):
    def resolveEntity(self,publicId,systemId):
        print " resolveEntity  ",publicId,systemId
        sys.exit()
class DHandler(xml.sax.handler.DTDHandler):
    def notationDecl(name, publicId, systemId):
        print " notationDecl ",publicId,systemId
        sys.exit()
    def unparsedEntityDecl(name, publicId, systemId, ndata):
        print " unparsedEntityDecl ",publicId,systemId,ndata
        sys.exit()

xmlstr = '<?xml version="1.0"?><html><body>hello &nbsp; </body></html>'
parser = xml.sax.make_parser()
parser.setContentHandler(CHandler())
parser.setEntityResolver(EResolver())
parser.setDTDHandler(DHandler())
parser.parse(StringIO.StringIO(xmlstr))


-- 
Antony Lesuisse
GPG EA2CCD66: 4B7F 6061 3DF5 F07A ACFF  F127 6487 54F7 EA2C CD66




From stehule at kix.fsv.cvut.cz  Fri Apr 11 04:25:23 2003
From: stehule at kix.fsv.cvut.cz (Pavel Stehule)
Date: 11 Apr 2003 08:25:23 GMT
Subject: pyorbit1.99.4  and Python2.3 problem
Message-ID: <slrnb9cuvj.49s.stehule@kix.fsv.cvut.cz>

Hello

I tried pyorbit with new version  of Python and when I try run test I get
error: 
[root at stehule pyorbit-1.99.4]# cd tests/
[root at stehule tests]# python everything_server.py
TypeError: MRO conflict among bases Object, B1, BaseServer, B2
TypeError: MRO conflict among bases Object, B1, BaseServer, B2

** (process:13354): WARNING **: could not generate stub for base
IDL:orbit/test/C1:1.0

Is working orbit with 2.3 Python?
Thank you
Pavel



From wilk-spamout at flibuste.net  Tue Apr 15 05:47:00 2003
From: wilk-spamout at flibuste.net (Wilk)
Date: 15 Apr 2003 11:47:00 +0200
Subject: DB API for ADO/DAO?
References: <2402252b.0304142353.2b4711ae@posting.google.com>
Message-ID: <873ckk14a3.fsf@flibuste.net>

john.burton at jbmail.com (John Burton) writes:

> I want to be use an .MDB database from a python application on
> windows.
> 
> The problem with odbc is that you need to manually set up data sources
> using the control panel which isn't something I want the users to have
> to do.
> 
> So I'm using DAO via com to access the database and it's working fine
> except that I have to use the DAO com api instead of the standard
> python api. So I was wondering if there is a suitable DB API wrapper
> for DAO or ADO I can use? Or a better solution to my problem?

http://adodbapi.sourceforge.net/

-- 
William Dode - http://flibuste.net



From mertz at gnosis.cx  Tue Apr  8 15:10:54 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 15:10:54 -0400
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
Message-ID: <+6xk+kKkXQeJ092yn@gnosis.cx>         

tweedgeezer at hotmail.com (Jeremy Fincher) wrote previously:
|If it's not, then is it possible to achieve the same thing (everything
|escaped properly, etc.) through other means?

It's not beautiful, but this does what you want:

    '"%s"' % repr(aString)[1:-1]

--
mertz@  | The specter of free information is haunting the `Net!  All the
gnosis  | powers of IP- and crypto-tyranny have entered into an unholy
.cx     | alliance...ideas have nothing to lose but their chains.  Unite
        | against "intellectual property" and anti-privacy regimes!
-------------------------------------------------------------------------





From gustav at morpheus.demon.co.uk  Mon Apr 14 15:54:09 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Mon, 14 Apr 2003 20:54:09 +0100
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <3e9af967$0$49112$e4fe514c@news.xs4all.nl>
Message-ID: <ptnosvmm.fsf@morpheus.demon.co.uk>

Irmen de Jong <irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> writes:

> Paul Moore wrote:
[...]
>>>>>unicodedata.name(u'\u20a0')
>> 'EURO-CURRENCY SIGN'
>>
>
> Hmmm
>
>
>  >>> import unicodedata
>  >>> unicodedata.name(u'\u20ac')
> 'EURO SIGN'
>  >>> unicodedata.name(u'\u20a0')
> 'EURO-CURRENCY SIGN'
>  >>>

Ah. I got 0x20a0 from an example in a document I read on the web. It
never occurred to me that it might be inaccurate :-(


> I've always used 0x20ac as Euro Sign, ?.
> This works nicely in all circumstances, also printing it out on the
> windows console using CP1252 encoding.

>>> print u"\u20ac".encode("cp1252")
?

Brilliant. It looks like my understanding was fine, it was just the
data I was using which was wrong. Wrong Unicode character, and
originally, wrong encoding as well. I knew today was going to be a bad
day :-)

> I didn't know what 0x20a0 is, so I looked it up [1],[2]:

Those are really useful pointers. Thanks a lot. (I know, unicode.org,
obvious place to look. Sorry, I'll do my homework better next time...)

Paul.

PS Is there any way of forcing the Windows 2000/XP console to display
   Unicode in any form (UTF-8, maybe) rather than having to deal with
   code pages???
-- 
This signature intentionally left blank



From roger_hancock at hotmail.com  Thu Apr 17 23:00:56 2003
From: roger_hancock at hotmail.com (Roger Hancock)
Date: 17 Apr 2003 20:00:56 -0700
Subject: Memory leak while looping
References: <1020cbf8.0304162100.6fc49e6f@posting.google.com>
Message-ID: <1020cbf8.0304171900.110a1675@posting.google.com>

Thanks for the help.  The xrange call solved my problem.  However, I'm
still a little puzzled by the behavior I'm seeing with the range call.
 With my now better understanding of the range call, I would expect to
see my function take a big chunk of memory until the function exited. 
Instead, the function continues to grab more and more memory.  First,
it takes most of my physical memory (256 MB) and then proceeds to eat
up my swap memory.  Eventually my machine reboots.  Pretty impressive,
huh?  Just imagine the damage I could do with 4 lines of code.

Thanks again,

Roger



From skip at pobox.com  Mon Apr 14 17:04:47 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 16:04:47 -0500
Subject: Seperating HTML content from logic ( was: Re: Inserting Python
        commands into HTML documents)
In-Reply-To: <200304141348.18694.scott_list@mischko.com>
References: <zpqdne0Kztcb0AejXTWcow@comcast.com>
        <xrtma.10655$LB6.301836@news1.tin.it>
        <200304141348.18694.scott_list@mischko.com>
Message-ID: <16027.8815.289798.405298@montanaro.dyndns.org>

    Scott> I'd also like to know which Python toolkits/frameworks that are
    Scott> aimed at implementing this model.

Zope implements this model if you restrict yourself suitably (use Zope Page
Templates and Python scripts instead of DTML).

Skip





From dave at boost-consulting.com  Sun Apr 20 05:21:41 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Sun, 20 Apr 2003 05:21:41 -0400
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com>
 <a3b5b608.0304080302.c26ad68@posting.google.com>
 <af6b5bdd.0304181001.558da6f@posting.google.com>
Message-ID: <84of317cd6.fsf@boost-consulting.com>

bitbucket at safe-mail.net (poiboy) writes:

>> > Of course you can use template systems. But this is not always a good
>> > economic decision.
>>
>> OK, but can't one use XML/XLST as a sole solution for the view?
>
> I use XSLT [partially] for that reason; it is a general templating solution. 
> The main reason I use XSLT however is because templating is the best (most 
> relevant) excuse I have to keep another X specification in my repetoire.
>
> The primary drawback to XSLT is having to fudge an XML document out of your
> template variables before being able to transform them. If you're retrieving
> data from a native XML source, great, otherwise you need to build an imaginary
> tree and make sure it matches up with the paths in the template.  
>  
> Next, variable representation in the XSL template is likely to take up a whole
> lot more room than a typical {{template_variable}}. Instead of standing out
> nicely against a backdrop of HTML tags, the template variables in XSL have to
> be searched for in small pools of xsl:for-each|value-of|copy-of|... tags and
> select="/path/to/var" attributes (and if you've ever had trouble matching up
> variable names in templates, you'll have a lot of fun with variable paths).
>
> In the amount of time spent figuring out how to ensure unwanted text would
> not appear without running my variable tree through a schema
> (<xsl:template match="text()"/>), I could have implemented five other methods
> to render my page (just guessing).
>
> Neglecting the inherent joy of an education, I believe the cost of using XSLT
> exceeds what benefits it may have (portability, and..) and is likely more
> expensive than successive rewrites with other template systems.

I've always said that humans shouldn't be forced to read/write XML and
that goes for programmers, too.  Writing a programming language in XML
format seems like the perversion of a sadist to me.

I don't care much about XML programming (haven't had a use for it),
but at PyConDC recently I got together with a couple of guys from the
Twisted project who do care, and designed a nice little meta-language
in Python for doing this sort of template processing.  Here's a little
example of it in action:

    template = body[
        table(id="outer", width="100%", height="100%", border="0")[
            tr(valign="bottom")[
                td(id="output", width="75%", valign="top", model="latestOutput")[
                    div(pattern="listItem", view="html")[
                        "Foo"
                    ]
                ],
                td(id="room-contents", valign="bottom")[
                    strong[
                        "Stuff you have"
                    ],
                    div(model="playerInventory", view="List")[

                        # Note programmatic elements
                        if_(not arg1)
                        [    
                            div(_class="item")["Nothing"]
                        ].else_
                        [
                            for_each(arg1)
                            [
                                div(
                                    style=["color: red", "color: blue", None]
                                  , view="item"
                                  , controller="look")[arg1]
                                )
                            ]

                        ]
                    ]
                ]
            ]
        ] 
    ]

    # invoke the template with some data
    template(['foo', 'bar', 'baz', 'another item'])

which generates:

    '''<body>
        <table id="outer" width="100%" height="100%" border="0">
            <tr valign="bottom">
                <td id="output" width="75%" valign="top" model="latestOutput">
                    <div pattern="listItem" view="html">
                        Foo
                    </div>
                </td>
                <td>
                    <strong>Stuff you have</strong>
                    <div model="playerInventory" view="List">
                        <div style="color: red" view="item" controller="look">foo</div>
                        <div style="color: blue" view="item" controller="look">bar</div>
                        <div view="item" controller="look">baz</div>
                        <div style="color: red" view="item" controller="look">another item</div>
                    </div>
                </td>
            </tr>
        </table>
    </body>'''
    

I really loved this because it was the result of healthy
cross-pollination between the cultures and techniques of two very
different programming communities: Python and C++.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From sismex01 at hebmex.com  Tue Apr  8 09:47:53 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Tue, 8 Apr 2003 08:47:53 -0500 
Subject: Datagrid-Widget for Python?
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F062330445915B@mail.hebmex.com>

> 
> On my lengthy search for Python stuff I had already found some
> information on tkTable. May be I misunderstood something --I dropped
> that, because I thought, tkTable is written in TCL, and there is no
> Python-interface.
>

Yes, tkTable is written for Tk, but you can easily integrate it
into Tkinter, because you can make direct tk calls from tkinter,
so you can load it and call it's functions from a Python class.
I did a bit with it a while ago, it's a nice piece of code.

-gus




From staschuk at telusplanet.net  Fri Apr 25 14:50:38 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 12:50:38 -0600
Subject: Divided by a common language (was: OO + Python Developers Wanted)
In-Reply-To: <yyyvfx2zdmo.fsf@adams.patriot.net>; from nav@adams.patriot.net on Fri, Apr 25, 2003 at 01:32:47PM -0400
References: <mailman.1051183297.29110.python-list@python.org> <yyyznmezekm.fsf@adams.patriot.net> <vais01o44vtc0d@corp.supernews.com> <yyyvfx2zdmo.fsf@adams.patriot.net>
Message-ID: <20030425125038.B1034@tibia.amotlpaa.bogus>

Quoth Nick Vargish:
> claird at lairds.com (Cameron Laird) writes:
> 
> > >Which "City" would that be?
> > <URL: http://www.cityoflondon.gov.uk/ >
> 
> Oh, then he should have specified, "The City", and I would have known
> he meant Mother London. Here in Amerika, we have a number of urban
> locales that consider themselves "the City"...

Typical American arrogance!  Sheesh!

"The City" is obviously Toronto.  (Unless you're in the west.)

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From twegener at fastmail.fm  Wed Apr 30 01:56:08 2003
From: twegener at fastmail.fm (Tim Wegener)
Date: Wed, 30 Apr 2003 15:26:08 +0930
Subject: Removing last line while appending to file
Message-ID: <20030430152608.3c51bc77.twegener@fastmail.fm>

Hi, 

Is there a simple way to remove the last line (or arbitrary trailing
portion) of a file, when appending to it? 

I want to write a log file in xml, which requires root elements around
all the log data. This means that when I want to append to the log I
must first remove the trailing </log> element. 

I have tried using file.seek() and file.tell() to no avail. I would
prefer not to parse the entire file each time I want to simply append an
entry. 

Thanks, 
Tim





From a.schmolck at gmx.net  Sun Apr  6 22:54:17 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 07 Apr 2003 03:54:17 +0100
Subject: Handling slices
References: <mailman.1049676910.30971.python-list@python.org>
Message-ID: <yfsistrkofa.fsf@black132.ex.ac.uk>

Francois Pinard <pinard at iro.umontreal.ca> writes:

> Hi, Python friends.
> 
> I'm currently using Python 2.2.1.  How does one support slices when given
> to __getitem__?  Suppose I get a slice object.  How can I use it to get
> a slice of a local array?  In other words, how one does apply a slice?

array[slice]

won't work for lists

> 
> The documentation says that __getslice__ is deprecated since 2.0, but then,
> I need an alternative way for using it.  What is it?

__getitem__

Here is a simple example:

class SliceMaker:
    """Convinience class to make slices."""
    def __getitem__(self, a): return a


>>> a = Numeric.array([1,2,3,4])
>>> a[SliceMaker()[3:1:-1]
array([4, 3])


alex



From dgallion1 at yahoo.com  Thu Apr 17 20:35:33 2003
From: dgallion1 at yahoo.com (darrell)
Date: Fri, 18 Apr 2003 00:35:33 GMT
Subject: Downloading TV listings with urllib
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4> <mailman.1050602375.2763.python-list@python.org> <E196EHo-0003Gb-00@swan.mail.pas.earthlink.net> <mailman.1050608946.14831.python-list@python.org>
Message-ID: <pLHna.782$_N.471@twister.nyroc.rr.com>

This worked.

        print gzip.GzipFile(fileobj=cStringIO.StringIO(tvg().read()) ).read()

--Darrell



Gerhard H?ring wrote:

> Josh wrote:
>> On 17 Apr 2003, you wrote in comp.lang.python:
>>>Josh wrote: [...]
>>>>        h.putheader('Accept-Encoding:',
>>>>        'gzip,deflate,compress;q=0.9') [...]
>>>>when I use err.read() to read the object returned I get everything in
>>>>Hex.
>>>
>>>My guess is that you get what you're asking for - gzip-compressed
>>>data.
>> 
>>  Well I don't want gzipped data. I want HTML so I can parse it and
>>  capture
>> the listings table into a list or something. Any ideas?
> 
> Don't tell the webserver that you'll accept gzip-compressed data?
> 
> I. e. remove the line I quoted above.
> 
> -- Gerhard




From camelo at esss.com.br  Mon Apr 14 16:08:51 2003
From: camelo at esss.com.br (Marcelo A. Camelo)
Date: Mon, 14 Apr 2003 17:08:51 -0300
Subject: Variable inside module 
Message-ID: <000a01c302c1$aaa464f0$0d00000a@esss.com.br>

Hi!

Can someone please explain-me what is 
happening here:

Given the followin bar.py file (placed 
inside a foo/ directory):

	var = 'foo'

	def SetVar():
	    global var
	    var = 'bar'

	def PrintVar():
	    global var
	    print 'var:', var
      
In the same foo/ directory, I've placed
the following __init__.py:

	from bar import *

Then, in the interactive python interpreter:

	>>> import foo
	>>> print foo.var
	foo
	>>> foo.PrintVar()
	var: foo
	>>> foo.SetVar()
	>>> foo.PrintVar()
	var: bar
	>>> print foo.var
	foo
	>>>
      
What I don't understand is that foo.var seems 
to be two different variables: one accessed 
inside the bar.py file and one accessed from 
within the python interpreter. What is going on
here?

Thank you,
 

Marcelo A. Camelo





From martin at v.loewis.de  Sun Apr 27 13:39:05 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 27 Apr 2003 19:39:05 +0200
Subject: Why doesn't the uu module give you the filename?
References: <mailman.1051458642.22944.python-list@python.org>
Message-ID: <m3bryrlu12.fsf@mira.informatik.hu-berlin.de>

Daniel Berlin <dberlin at dberlin.org> writes:

> While it's simple enough to get the uu module to uudecode a string
> (using StringIO), it's impossible to get it to handle you the filename
> the uuencoded thing specifies.

Because nobody thought of this before. Please do submit a patch,
including changes to documentation. Please make the patch backwards
compatible if possible.

Regards,
Martin




From mhammond at skippinet.com.au  Wed Apr 30 18:01:16 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Wed, 30 Apr 2003 22:01:16 GMT
Subject: ActiveScripting: import email => no module named _codecs
In-Reply-To: <db81c492.0304300631.28ec1f36@posting.google.com>
References: <db81c492.0304300631.28ec1f36@posting.google.com>
Message-ID: <MIXra.8244$lD4.90282@news-server.bigpond.net.au>

My guess is that you are running this from inside IE.  IE executes 
Python code in a restricted environment.  My guess is that this 
restricted environment is preventing the Unicode codecs from being imported.

You could try to fix this by opening 
win32comext\axscript\client\framework.py and locating the line:

	ok_builtin_modules = rexec.RExec.ok_builtin_modules + ('win32trace',)


Change this to:

	ok_builtin_modules = rexec.RExec.ok_builtin_modules + 
('win32trace','_codecs')

And see if that helps.  Let me know either way.

Mark.


vadim wrote:

> I am doing rather simple thing after installing win32com 
> extensions (IE):
> 
> <script language="Python" >
> import email.Parser
> </script>
> 
> and getting following error:
> ...
> File "<ScriptBlock> ", line 2, in ?
>   import email.Parser
> SystemError: Failed to load the builtin codecs: No module named _codecs
> 
> Does anyone could help what it means?




From cnetzer at mail.arc.nasa.gov  Mon Apr  7 18:55:43 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 07 Apr 2003 15:55:43 -0700
Subject: [Slightly OT]: More on ints and floats
In-Reply-To: <fjss6b.1021.ln@boundary.tundraware.com>
References: <fjss6b.1021.ln@boundary.tundraware.com>
Message-ID: <1049756142.686.37.camel@sayge.arc.nasa.gov>

On Mon, 2003-04-07 at 15:03, Tim Daneliuk wrote:
> He argues that
> any place one can use an int, these days (with cheap FP hardware), one
> could use a float 0 extended to the precision of the machine and get
> equivalent computational results. Say the hardware supported 4 digits of
> precision. He is arguing that:
> 
> 
>     3/4.0000  is equivalent to 3.0000/4.0000

If you have a 32 bit int, and a 32 bit float, the ints are NOT a proper
subset of the float operations.  You need at least 64 bit doubles, to
have the same granularity as a 32 bit int.  If you use 64 bit ints, then
even a "long double" may not have 64 bits of integer granularity.  If
you try to use a float as a file pointer, or a double as a 64 bit long
file pointer, you WILL run into a whole lot of problems on large data
sets.  See below:

$ python
Python 2.2.2 (#1, Mar 21 2003, 23:01:54)
>>> import sys
>>> import Numeric as Num
>>>
>>> a = Num.array( sys.maxint, Num.Float32 )
>>> b = Num.array( 1, Num.Float32 )
>>>
>>> sys.maxint
2147483647
>>> a
2147483648.0
>>> a == sys.maxint
0
>>> b
1.0
>>> a-b == sys.maxint
0
>>> a-b
2147483648.0
>>> a-b-b
2147483648.0

So, subtracting 1 from a large valued float does not change it's value
(and in fact subtracting 64 from this same float will not change it).


Because you need (typically) a 64 bit double to have all the granularity
of a 32 bit int.  On most modern CPUs, although the CPUs have made much
progress on doing fast float/double operations, ints are still typically
faster.  Often there are more integer execution units then float units
in the chip itself (ie. more integer operations can be retired than
floats or doubles, in the same amount of time).  Which "vector" floating
point processors, the difference in speed is not as cut and dried, but
the precision porblem still exists.


So, I'd say your friend is basically wrong.





From alanmk at hotmail.com  Sun Apr 27 11:57:19 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Sun, 27 Apr 2003 16:57:19 +0100
Subject: trouble with xml.sax and unknow entities
References: <mailman.1051453782.26417.python-list@python.org>
Message-ID: <3EABFDDF.7E7C50A1@hotmail.com>

Antony Lesuisse wrote:
> 
> I'm not on the list, please cc: me the answers.
> 
> I'm having trouble to parse the folowing xml with the default python xml.sax
> api. I'm using python2.2 on debian unstable powerpc (python2.2-xmlbase).
> 
> '<?xml version="1.0"?><html><body>hello &nbsp; </body></html>'

Your base problem here is that this is not a "well-formed" XML document,
and as such, parsers are *obliged* to report it as an error. The
relevant
section of the XML spec is

http://www.w3.org/TR/REC-xml.html#sec-references

and says:

"In a document without any DTD, a document with only an internal DTD
subset which contains no parameter entity references, or a document
with "standalone='yes'", for an entity reference that does not occur
within the external subset or a parameter entity, the Name given in
the entity reference must match that in an entity declaration that
does not occur within the external subset or a parameter entity,
except that well-formed documents need not declare any of the 
following entities: amp, lt, gt, apos, quot."

The way to solve your problem is to introduce an "internal DTD
subset", like so

"""<?xml version="1.0"?>
<!DOCTYPE[
<!ENTITY nbsp "&#160;">
]>
<html><body>hello &nbsp; </body></html>
"""

> The parser halt on &nsbsp; because it doesn't know about this entity. The
> problem is cannot find a way to tell him what this entity is.
> 
> (1)
> Is there a way to have a callback the parser arrive on &nbsp; ? None of the
> folowing handler functions (resolveEntity,notationDecl,unparsedEntityDecl) are
> called.
> 
> I thought resolveEntity had to be called in that situation but i probably
> misunderstand the sax api.

The resolveEntity function is intended only for use with "external
entities", i.e. entities which reside entirely in a separate container.
The resolveEntity function is basically there so that you can control
the resolution of the addresses of XML documents and document fragments,
in relation to the address of the document you're dealing with.

> (2)
> Is there a way to register entities before the parsing begin ?
> Something like:
>     parser.registerEntity('&nbsp;','blahblah')

Here is the essential difficulty. If anyone else were to parse your
XML document, they would also have to configure their parser to 
"register" the same entity. The correct way to register the entity 
is in the document itself.

> (3)
> Or is there a way to register an external DTD where those entities can be
> defined ?  Something like:
>     parser.registerExternalDTD('xhtml.dtd')

Again, such a document whose dtd was *only* registered in this way,
i.e. it did not contain a doctype declaration, would not be a "valid"
XML document, i.e. its structure could not be checked against the
structure rules declared in the actual DTD: it could only be
"well-formed".

Hmmm. Or maybe not. I just tried to find the relevant section of the
XML spec that states the above, but I couldn't find it. All I could
find was

http://www.w3.org/TR/REC-xml.html#sec-prolog-dtd

"Definition: An XML document is valid if it has an associated document
type declaration and if the document complies with the constraints
expressed in it."

Which, to me, doesn't explicitly state that the doctype has to be
declared inside the document text itself.

Here's my version of your code.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
import StringIO,sys,xml.sax,xml.sax.handler

class CHandler(xml.sax.handler.ContentHandler):
    def startElement(self, name, attrs):
        print name
    def characters(self, ch):
        print ch.encode('Latin-1')

class EResolver(xml.sax.handler.EntityResolver):
    def resolveEntity(self,publicId,systemId):
        print " resolveEntity  ",publicId,systemId
        sys.exit()
class DHandler(xml.sax.handler.DTDHandler):
    def notationDecl(name, publicId, systemId):
        print " notationDecl ",publicId,systemId
        sys.exit()
    def unparsedEntityDecl(name, publicId, systemId, ndata):
        print " unparsedEntityDecl ",publicId,systemId,ndata
        sys.exit()

xmlstr = """<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html [
<!ENTITY nbsp "&#160;">
]>
<html><body>hello &nbsp; </body></html>
"""
parser = xml.sax.make_parser()
parser.setContentHandler(CHandler())
parser.setEntityResolver(EResolver())
parser.setDTDHandler(DHandler())
parser.parse(StringIO.StringIO(xmlstr))
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

regards,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From rjroy at takingcontrol.com  Wed Apr 23 22:19:31 2003
From: rjroy at takingcontrol.com (Robert Roy)
Date: Thu, 24 Apr 2003 02:19:31 GMT
Subject: pmw listbox selection loses focus
References: <d44e8ee7.0304230934.37035ed@posting.google.com>
Message-ID: <3ea74933.56070296@nntp.flfrd.phub.net.cable.rogers.com>

On 23 Apr 2003 10:34:49 -0700, awheat at imaging.robarts.ca (Andrew
Wheatley) wrote:

>I've created a Pmw.ScrolledListBox on Redhat Linux 8.0.  If I mouse
>click on a Terminal window or Emacs window for example (but not its
>decoration) after selecting a selection, the selection is turned off
>graphically.  However, the selectioncommand is not called.  How can I
>prevent the selection(s) from being deselected when I click inside
>other windows?
>
>Andrew

try setting  exportselection=0 on the listbox



From mwh at python.net  Wed Apr 30 13:29:46 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 30 Apr 2003 17:29:46 GMT
Subject: fcntl in windows
References: <k2rvav06nhjgtr578fa5ostftfl69rethp@4ax.com> <mailman.1051717847.12855.python-list@python.org> <HSSra.42028$K35.1214886@news2.tin.it>
Message-ID: <7h33cjz51y0.fsf@pc150.maths.bris.ac.uk>

Alex Martelli <aleax at aleax.it> writes:

> Jp Calderone wrote:
> 
> > On Wed, Apr 30, 2003 at 03:26:29PM +0000, Afanasiy wrote:
> >> Is there an fcntl in Windows? A stub, a wrapper, etc?
> > 
> >   http://python.org/doc/lib/module-fcntl.html
> 
> The page you indicate says "Availability: Unix", and indeed
> trying:
> 
> import fcntl
> 
> on Windows gives an ImportError: No module named fcntl
> 
> 
> As far as I know, nobody's written a fcntl-emulator for
> Windows -- the functionality is clearly available (a bit
> in msvcrt, all of it in win32all, or maybe also through
> ctypes) but the interfaces are, alas, quite different.

Indeed, it would seem to me that implementing a version of fcntl on
Windows would be flat out insane -- surely the sensible thing to do
would be provide two implementations of whatever it is you need, the
unix one of which calls fcntl.

The interface "design" of the fcntl syscall is not something that begs
to be spread around...

Cheers,
M.

-- 
6. The code definitely is not portable - it will produce incorrect 
   results if run from the surface of Mars.
               -- James Bonfield, http://www.ioccc.org/2000/rince.hint



From a.schmolck at gmx.net  Sun Apr 13 09:06:54 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 13 Apr 2003 14:06:54 +0100
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk> <3E977928.B3A6DC7E@engcorp.com>
Message-ID: <yfsznmueec1.fsf@black132.ex.ac.uk>

Peter Hansen <peter at engcorp.com> writes:
> What is this "symbol" that you keep talking about?  If you

I was refering to interned strings as symbols, because the only reason I can
think of why you'd care about (address) identity of strings is because you'd
like to use them as symbols.


> are simply using that as an alias for "interned string", then
> I believe you are making up reasons for those which they are
> *not* intended to have.  You said "the idea of symbols is their
> uniqueness".  Well, I don't know what a "symbol" is to you,

I was using (primitive) symbols loosely as in:

stuff you use to name (designate) other stuff. You can use any stuff as
symbols of which can produce sufficient amounts of unique instances and that
allows you to establish (some form of) identity relationship. But obviously
it's nice if

a) you have some convinient notation for writing and displaying symbols
   (strings do have that property)
b) creation and establishing identity are efficient (string equality would
   also do but is of course much slower than pointer comparision, i.e.
   identity via "is")

> but I believe that "interned strings" are *not* intended to 
> guarantee some "uniqueness" property which you may count on 
> in your code.  Any such property is a side effect of the

Well, unless python has a radically different idea from anyone else on what
interning a string (and "returning *the* interned string" string) means, I'd
really think that you can rely on:

>>> type(a) == type(b) == str and b == a and inter(b) is intern(a)
1

(but maybe one of the cognoscenti would like to affirm that?)

> implementation or something.  I believe the sole purpose of

Hmm, what implementation alternative do you have in mind? Unless intern is not
guaranteed to have any effect (i.e. it is a mere suggestion of some vague
optimization intent to the interpreter) I can't really think of what else it
could reasonably do.

> an interned string ("symbol"?) is to *optimize* access to those
> strings, i.e. for performance reasons.  

Sure, but the reason you want SYMBOL to be identical and not merely equal
(only) to itself is largely performance reason (apart from maybe that using
``X == SYMBOL`` might not be safe (depending on what X can be), so you'd have
to write your own predicate).

> 
> See http://www.python.org/doc/current/lib/built-in-funcs.html#built-in-funcs
> under "intern()" for what I believe is the only official purpose
> of interning strings: "to gain a little performance on dictionary lookup".

I agree the wording there is not particularly clear (and python doesn't much
encourage "symbolic computation" -- so the intent is not to have people write
intern("foo") all over the place where a lisper would write e.g. :foo), but
I'd be really suprised if the intern semantics were intended to be unspecified
w.r.t. the identity property above.

> 
> -Peter

'as



From scott_list at mischko.com  Tue Apr 22 19:38:37 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Tue, 22 Apr 2003 16:38:37 -0700
Subject: Can someone send me the latest Gadfly?
In-Reply-To: <0Ljpa.2547$%_3.2056417@newssrv26.news.prodigy.com>
References: <200304220934.38013.scott_list@mischko.com> <mailman.1051030174.3152.python-list@python.org> <0Ljpa.2547$%_3.2056417@newssrv26.news.prodigy.com>
Message-ID: <200304221638.37852.scott_list@mischko.com>

Google for "gadfly download"  You'll see where I went wrong.
First two links point to the old site!

On Tuesday 22 April 2003 15:54, drs wrote:
> "Gerhard H?ring" <gh at ghaering.de> wrote in message
> news:mailman.1051030174.3152.python-list at python.org...
>
> > Scott Chapman wrote:
> > > www.chordate.com appears to be seriously down.  Their DNS is resolving
>
> to
>
> > > ns1.rackshack.net.  I called Rackshack and they don't have a zone file
>
> for
>
> > > chordate.com.
> > >
> > > Chordate hosts the Gadfly database and I would like a copy of it on my
>
> RH8
>
> > > system but can't download it.  Google isn't helpful in finding another
> > > location to get it from.
> > >
> > > Can someone help out here?
> >
> > http://sourceforge.net/projects/gadfly
> > http://sourceforge.net/project/showfiles.php?group_id=662
>
> I am very confused ... You obviously did a reasonable amount of research to
> find this, and you are also clearly a sophisticated internet user ... a
> google search on "gadfly python" turned up the db as the second result, and
> an informational page linking to the db as the first.  what google searches
> did you try?
>
> -d





From dan at grassi.org  Wed Apr  2 11:56:15 2003
From: dan at grassi.org (Dan Grassi)
Date: Wed, 2 Apr 2003 16:56:15 +0000 (UTC)
Subject: How to obtain python file path
References: <20030401163212631-0500@news.covad.net> 
	<slrnb8jusc.5u1.ramen@lackingtalent.com> 
	<20030401165546102-0500@news.covad.net> 
	<slrnb8lr3q.15s.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <20030402115351679-0500@news.covad.net>

This works only if the python script is imported, not if it is executed 
directly.

Gee, I'm surprised that this is such a difficult thing to do yet 
conceptually simple.

Dan

In <slrnb8lr3q.15s.do-not-spam-ben.hutchings at tin.bwsint.com> Ben 
Hutchings  wrote:
> In article <20030401165546102-0500 at news.covad.net>, Dan Grassi wrote:
>> Not quite, that only provides the path to the initial .py file.
>> 
>> With a directory structure:
>> fa
>>     a.py
>>     fb
>>         b.py
>> 
>> cd .../fa
>> execute a.py
>> where a adds fb to sys.path
>> a imports b.py
>> 
>> I need b.py to report it's path which would be .../fa/fb
> 
> Assuming the current directory hasn't changed since the file was
> loaded, or the file was loaded with an absolute filename:
> 
>     def mydir():
>         import os.path
>         return os.path.abspath(os.path.dirname(__file__))
> 



From ckea25d02 at sneakemail.com  Tue Apr 15 21:00:30 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 13:00:30 +1200
Subject: Python and Schools
In-Reply-To: <mailman.1050436647.12826.python-list@python.org>
References: <mailman.1050436647.12826.python-list@python.org>
Message-ID: <b7i9al$17ik5$1@ID-169208.news.dfncis.de>

Jordan McCoy wrote:
> The order of an algorithm is O(f n), if a function f and a constant k
> exists such that k * f(n) is not more then the execution time of the
> algorithm

Just a few corrections:

I think you mean "the execution time of the algorithm is not
more than k * f(n)".

> The goal of algorithm development is to achieve ... O(n)
> ... as this represents constant execution time;

No, constant time is O(1) -- i.e. completely independent
of n. You're talking about linear time.

> 1. The common but inefficient bubble sort has an order of O(n^2). ...
 > each addition to the problem set will result in an exponential time
 > increase.

That's called polynomial time, not exponential. Exponential time
is O(a^n) for some a, which is even worse.

> 2. The more efficient quick sort has an order of O(n * log2 n). Worse
> then constant time, but much better then exponential time.

Again, I think you mean polynomial. Although it's better than
exponential time, too!

By the way, you can just write that as O(n * log n). The base
of the log is irrelevant, since it only changes things by a
constant factor.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From staschuk at telusplanet.net  Tue Apr  8 21:47:00 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 8 Apr 2003 19:47:00 -0600
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <BAB8B22D.2022%dogen@mindspring.com>; from dogen@mindspring.com on Tue, Apr 08, 2003 at 05:06:53PM -0400
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org> <BAB8B22D.2022%dogen@mindspring.com>
Message-ID: <20030408194700.A1973@tibia.amotlpaa.bogus>

Quoth Jon Brown:
> Actually the program does exactly what you describe, except from the inside
> out because it starts in the middle, and it cuts out of the loop if the
> characters don't match.

I see that I haven't expressed clearly what I'm suggesting.

The combinations() function recursively produces *all* k-lists of
words from the word list, whether or not it can be determined
partway through the assembly that no possible palindrome starts
with the list assembled thus far.  For example, if you ask for all
4-word-long palindromes given a word list of 6 words, it makes 360
= (6 choose 4)*4! calls to rPal().

My suggestion is that, if the word list is, say,
	abandon
	major
	wombat
	articulate
	orange
	royal
then the program could start with 'abandon', notice that no word
in the word list ends with 'a', and immediately, er, abandon
trying to find palindromes starting with this word, and move on to
finding palindromes starting with 'major'.  The code as it stands
does not make this inference, and so unnecessarily generates the
(5 choose 3)*3! = 60 sequences of 4 words which start with
'abandon', and tests them individually for palindromicity.

> To look for pals it spins off big, big numbers of combinations, true enough.
> That's what behind the need for speed. Since Python is inherently on the
> pokey side, I expected the slowness but wonder if I'm missing a cool way to
> pick up the pace. But I suspect the solution will involve C++, and I've
> never linked C/C++ with Python.

Going to C might yield some improvement, but it will do nothing to
deal with the combinatoric explosion which is in my opinion the
real performance sink in this problem.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"What I find most baffling about that song is that it was not a hit."
                                          -- Tony Dylan Davis (CKUA)




From mlh at furu.idi.ntnu.no  Mon Apr  7 18:19:42 2003
From: mlh at furu.idi.ntnu.no (Magnus Lie Hetland)
Date: Mon, 7 Apr 2003 22:19:42 +0000 (UTC)
Subject: Lazy string with re.match?
Message-ID: <slrnb93ubu.esr.mlh@furu.idi.ntnu.no>

I'm using regular expressions in a semi-recursive-descent parser, and
I'd like to read the data in a lazy manner. However, since I don't
know how much of the data a given regexp will need to find out whether
it matches or not, I thought about writing a lazy string class, and to
pass that to the match method of the various regexps. I first tried
simply implementing __getitem__, but was told in no uncertain terms
that a string or buffer was expected by the re object. So... I tried
to subclass str, but that didn't help -- it seemed to simply ignore my
__getitem__ method (not surprisingly, perhaps).

Does anyone have any info on how (if at all) I might trick a regexp to
use my __getitem__ method, or something similar?

The goal is, I suppose, to either let the regexp pump the lazy string
on its own, or to check whether it exhausted the given string material
(which may be a completely ordinary string) and, in that case, give it
some more until either it matches or does _not_ exhaust the string...

Any ideas?

(Since I'm not working with huge files at the moment, simply reading
in the entire thing as a single string is probably the easiest and
best solution, though...)

-- 
Magnus Lie Hetland               "Nothing shocks me. I'm a scientist." 
http://hetland.org                                   -- Indiana Jones



From mwh at python.net  Thu Apr 10 12:15:33 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 10 Apr 2003 16:15:33 GMT
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <7xwui295xq.fsf@ruckus.brouhaha.com> <yfsel4ah3oc.fsf@black132.ex.ac.uk> <mailman.1049987720.5681.python-list@python.org>
Message-ID: <7h3of3ez5uj.fsf@pc150.maths.bris.ac.uk>

Jim Meyer <jmeyer at pdi.com> writes:

> On Thu, 2003-04-10 at 06:39, Alexander Schmolck wrote:
> > I'd recommend against using __slots__ (unless really needed for optimization),
> > because it cripples the usefulness and generality of your classes to your and
> > also *to other people's* code (weakrefs, pickling, reflection -- all
> > compromised). The first two you could fix (but are unlikely to, unless you
> > *yourself* need to use weakrefs or pickling of instances of those classes),
> > the last one you can't.
> 
> I'm feeling very slow this morning; could you give a few explicit
> examples of how __slots__ interferes with weakrefs, pickling, and
> reflection?

This sort of thing:

/>> class C(object):
|..  pass
\__ 
->> import weakref
->> weakref.ref(C())
<weakref at 8288a9c; dead>
/>> class C(object):
|..  __slots__ = []
\__ 
->> weakref.ref(C())
Traceback (most recent call last):
  File "<input>", line 1, in ?
TypeError: cannot create weak reference to 'C' object
/>> class C(object):
|..  __slots__ = ['__weakrefs__']
\__ 
->> weakref.ref(C())
Traceback (most recent call last):
  File "<input>", line 1, in ?
TypeError: cannot create weak reference to 'C' object
/>> class C(object):
|..  __slots__ = ['__weakref__']
\__ 
->> weakref.ref(C())
<weakref at 8276374; dead>
->> 

Cheers,
M.

-- 
    FORD:  Just pust the fish in your ear, come on, it's only a
           little one.
  ARTHUR:  Uuuuuuuuggh!
                    -- The Hitch-Hikers Guide to the Galaxy, Episode 1



From stuffduff at cox.net  Fri Apr 25 17:13:47 2003
From: stuffduff at cox.net (stuffduff)
Date: Fri, 25 Apr 2003 21:13:47 GMT
Subject: Zope-Dev: Curious Problem += Anxiety
References: <45jiavkpsr5kl41u0ajs2jud1o0qd7rm0d@4ax.com>
Message-ID: <ov8javg3n16tsc874ulgh1gefsreu922bt@4ax.com>

I spent the day rebuilding my laptop.

python 2.1.3 from source and then zope 2.51 built & running on the new
python.

I don't think python's the problem, I think zope is.

I even tried moving the class definition inside the function
definition, (how could zope miss it there?); but still no joy.

Something in zope must be intercepting something is my guess,

Curiosity += Anxiety, (or it's going to be a very long weekend)

Sean

home: stuffduff at cox.net
work: swduffy at unmc.edu

On Fri, 25 Apr 2003 14:58:27 GMT, "<stuffduff>"
<just.say.no.to.spam at cox.net> wrote:

>Hi,
>
>I have a python external method which reads from an external database,
>unpickles a class object, and concatenates some of the object's
>properties into a string which is returned by the external method.
>The class definition is included in the module.
>
>When I run it from the linux prompt in the Extensions directory it
>behaves exactly as it should. The class is successfully unpickled, the
>properties are read and the string is returned.
>
>However, when I set it up as an external method and call it I get the
>infamous:
>
>Error Type: SystemError
>Error Value: Failed to import class MyObject from module __main__
>
>From my perspective Zope should be 'blissfully unaware' of the
>existence of the class or the instantiation of the object.  Since the
>object is only accessed in the external method, the module contains
>the class definition, and only a string is returned, why is this
>happening?
>
>What am I missing here?
>
>Platform: Red Hat 7.3, Python 2.1.3, Zope 2.51
>
>Curious,
>
>Sean
>
>P.S. Also posted to zope-dev but I wanted the 'python' perspective as
>well




From staschuk at telusplanet.net  Mon Apr 28 01:32:07 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 27 Apr 2003 23:32:07 -0600
Subject: overloading print behavior
In-Reply-To: <vapb1f5mcqf0cd@corp.supernews.com>; from cere@u.washington.edu on Sun, Apr 27, 2003 at 09:10:30PM -0700
References: <vapb1f5mcqf0cd@corp.supernews.com>
Message-ID: <20030427233207.B23591@tibia.amotlpaa.bogus>

Quoth Cere Davis:
  [...]
> When I:
> 
> print st
> 
> I get:
> 
> ['abc', <__main__.C instance at 0x40288fac>]
> 
> Which is what I expect,  but I would like it if print or perhaps the 
> str() coercion that is happening behind the scenes would walk through my 
> C instance
> and resolve the name/value paris within the C class and print the 
> values.  [...]

Actually it's repr() and not str() in the above example, as I will
describe below.

To specify str() conversion (which also applies for print
statements) for your class, you need to write a __str__ method:

    class Foo(object):
        def __init__(self, n):
            self.n = n
        def __str__(self):
            return 'Foo with n = %s' % self.n

To specify repr() conversion, you need to write a __repr__ method:

        def __repr__(self):
            return 'Foo(%r)' % self.n

The difference between str and repr is described in the docs:
    <http://www.python.org/doc/2.2.2/ref/customization.html>
(This part of the documentation is presently the subject of some
scrutiny here; some feel it is unclear, misleading, wrong.  The
discussion is in this thread:
    <http://groups.google.com/groups?th=24b817d49ec3a59b>
    (or <news:uw1i4nqpl33h.1l3wpteil4jb0.dlg at 40tude.net> et seq.)
If the docs leave you in doubt about whether you should be writing
a __str__ method or a __repr__, have a look through those posts.
You might want to skip the digression about orbital mechanics.)

The other thing to know is that, when you print a list (or dict
or tuple) containing your objects, those objects will be printed
by repr conversion, not by str conversion.  (Lists etc. actually
have no __str__, so Python falls back to the list's __repr__,
which in turn uses the __repr__ of the list's elements.)  There
are those whom this behaviour annoys; see the aforementioned
thread for some discussion (and an explanation of why it's the
way it is).

If you do want to include all the attributes of your objects in
the repr of those objects (this would not really be appropriate
for str, imho), then you could do it this way:

    class Foo(object):
        def __init__(self, a, b):
            self.a = a
            self.b = b
        def __repr__(self):
            pairs = ['%s = %r' % (name, value)
                for name, value in self.__dict__.items()]
            return '<a Foo with ' + ', '.join(pairs) + '>'

Then, for example:

    >>> Foo(3, 4)
    <a Foo with a = 3, b = 4>

-- 
Steven Taschuk                            staschuk at telusplanet.net
"Our analysis begins with two outrageous benchmarks."
  -- Clinger et al., "Implementation strategies for continuations"




From matthew at newsgroups.com  Wed Apr  9 19:30:02 2003
From: matthew at newsgroups.com (matthew)
Date: Thu, 10 Apr 2003 11:30:02 +1200
Subject: convert dictionary to string
Message-ID: <b72a7l$ifj$1@lust.ihug.co.nz>

hi,

can anyone tell what is the fastest way to convert a dictionary to a string?

eg: d={'A':45.10, 'F':56.50,...}
to 'A45.10F56.50...'

Thanks very much. matthew




From jkraska at san.rr.com  Sat Apr 12 13:19:19 2003
From: jkraska at san.rr.com (Courageous)
Date: Sat, 12 Apr 2003 17:19:19 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net> <Xns935B779CE4612cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <6cig9v02o69ocad9ss94utneqghk993hg5@4ax.com>

>And then there is the self. It is not bad. It does not hurt your OO 
>programming. It's like the nipples of a beautifull girl turning into a 
>different color during intercourse. You think it's wired for the first 
>time. You wonder, why is this necessary.

Possibly it is necessary for the same reason that many coding standards
(such as in C++, using Microsoft's approach), one must put m_ before all
class member attributes. What that coding standard amounts to is an
example of the standards writer actually _fighting_ the existence of
implicit this in C++. In other words, they are using attribute-naming
to disambiguate between class attributes and local variables of the same
name. In Python, this problem does not exist, because "m_" is spelled
"self".

C//




From max at alcyone.com  Wed Apr 16 06:52:44 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 03:52:44 -0700
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net>
Message-ID: <3E9D35FC.FCAD17C9@alcyone.com>

user at domain.invalid wrote:

> What do the back ticks in method "__str__" mean?

The backquotes are a way of saying "evaluate the repr of this
expression."  That is, `x` is the same as repr(x).  Most consider the
use of backquotes for this purpose gauche.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Maybe soul mates exist / After all
\__/ Des'ree
    Bosskey.net: Aliens vs. Predator 2 / http://www.bosskey.net/avp2/
 A personal guide to Aliens vs. Predator 2.



From llothar at web.de  Mon Apr  7 09:26:38 2003
From: llothar at web.de (Lothar Scholz)
Date: 7 Apr 2003 06:26:38 -0700
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <yfsn0j3kot2.fsf@black132.ex.ac.uk>
Message-ID: <6ee58e07.0304070526.41488e8@posting.google.com>

Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfsn0j3kot2.fsf at black132.ex.ac.uk>...
> llothar at web.de (Lothar Scholz) writes:

> > To make clear, i'm using XEMACS 8 hours a day for years now. But it's
> > absolutely losing ground compared to modern editors. And i don't think
> > that
> > extensibility in the emacs sense is really necessary today or a plus,
> > because
> > people want out of the box solutions that are specialized to their
> > domain.
> 
> The plausibility of your suggestion that the majority of emacs's long time
> user base is formed by mindless drones depresses me.

I never said something like this. Its just the observation that nearly
all of the hard core Emacs users i meet have never done more then
adding simple key bindings to there .emacs file.

I have learned elisp 8 years ago and tried to add some features to the
Eiffel mode. And it was pure hell, even with experience in Common
Lisp.



From sholden at holdenweb.com  Mon Apr  7 21:03:25 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 08 Apr 2003 01:03:25 GMT
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com>
Message-ID: <xdpka.23011$It5.19472@news2.central.cox.net>

"Tim Daneliuk" <tundra at tundraware.com> wrote in message
news:fjss6b.1021.ln at boundary.tundraware.com...
> OK, I don't want to resurrect another interminable discourse in ints,
> floats, and how they are/ought/can be handled in the language. But, I
> have a sort of theoretical mathematical question which the whole
> business brought to mind.
>
> As I understand it, integers and floats are distinct mathematical
> entities. A colleague of mine claims, that insfar as we use them in
> computing, ints are merely a proper subset of floats. He furthermore
> asserts that (again as regards to computing) the distinction between
> them was made as a purely as a practical matter because floating point
> arithmetic was historically computationally expensive. He argues that
> any place one can use an int, these days (with cheap FP hardware), one
> could use a float 0 extended to the precision of the machine and get
> equivalent computational results. Say the hardware supported 4 digits of
> precision. He is arguing that:
>
>
>     3/4.0000  is equivalent to 3.0000/4.0000
>
> (Never mind the old Python modulo vs. divison debate.)
>
> In effect he is saying that, unless there is a practical
> performance/cost issue at hand, there is no real reason to
> differentiate between ints and floats in practical programming
> problems.
>
Well, that depends on how each is represented. Suppose you had a platform
with 64-bit integers and 64-bit floats. It would be impossible to represent
every integer as a floating-point number because the mantissa in the floats
will be shorter than the integer precision.

Of course, now we are moving towards infinite-precision integers (with
hidden widening to long when required) this problem will always exist.

> As a matter of 'pure' mathematics, I argued that ints and floats are
> very different critters. My argument (which is no doubt formally very
> weak) is that the integer 3 and the float 3.0000 are different because
> of the precision problem.  For instance, the integer 3 exists at a single
> invariant point on the number line, but 3.0000 represents all numbers
> from 2.99995 through 3.00004.
>
As a matter of pure mathematics the integers are a subset of the reals,
which are a subset of the complex numbers.

Unfortunately pure mathematics and floating-point arithmetic only tend to
intersect in the brains of people like Tim Peters.

> Could one of the genius mathematicians here bring some clarity to this
> discussion, please?
>
So I guiess it's the timbot you want to hear from then, and not me :-)

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From pinard at iro.umontreal.ca  Sat Apr  5 17:08:09 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 05 Apr 2003 17:08:09 -0500
Subject: Input from a file as a command line argument
In-Reply-To: <pan.2003.04.05.16.31.02.87447.878@thomas-guettler.de>
References: <3E8EDBDA.7020701@pbezone.net>
	<pan.2003.04.05.16.31.02.87447.878@thomas-guettler.de>
Message-ID: <oqfzow7g3a.fsf@titan.progiciels-bpi.ca>

[Thomas Guettler]

> import sys
> for line in sys.stdin.xreadlines():
>     print line

In Python 2.2, this should be better written:

   import sys
   for line in sys.stdin:
       print line

However, I noticed this morning that whenever sys.stdin is connected to an
interactive terminal, input lines are buffered instead of passed one by one
to the application after each `Enter'.  So, for this application, I had to
revert to the older paradigm:

   import sys
   line = sys.stdin.readline():
   while line:
       print line
       line = sys.stdin.readline()

I wonder if this might be considered as a Python bug, or not.  What is the
general opinion? :-)

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From pinard at iro.umontreal.ca  Wed Apr 23 17:37:53 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 23 Apr 2003 17:37:53 -0400
Subject: Off topic: What usenet news reader to use?
In-Reply-To: <6h0eavc1dgmgd4rphnaodgvh9o6qigs62g@4ax.com>
References: <mailman.1051036295.31460.python-list@python.org>
	<d97bavkeagbpfkvm3649725tptqhkskds9@4ax.com>
	<6h0eavc1dgmgd4rphnaodgvh9o6qigs62g@4ax.com>
Message-ID: <oqsms8x5ce.fsf@alcyon.progiciels-bpi.ca>

[Tung Wai Yip]

> On Tue, 22 Apr 2003 20:55:13 +0100, Dale Strickland-Clark
> <dale at riverhall.NOTHANKS.co.uk> wrote:
> >"Wai-Yip Tung \(wtung\)" <wtung at cisco.com> wrote:

> >>Is there any good Windows news reader you would recommend.

> >In no particular order, have you tried: Outlook Express, Free Agent,
> >Mozilla, XNews, [...] Opera 7.10?

> I get back Free Agent[...]  Actually Opera's new reader works OK.
> Its M2 email client eludes me.  [...]

I'm not a MS Windows user.  Yet I use Gnus (over Emacs) for reading and
replying to both email and news, and am rather satisfied with it.  However,
it surely has hundreds of commands and knobs, which might overall frighten
a newcomer, so I do not know if it reasonable to recommend it or not. :-)

Gnus allows handling of high volumes with ease, in my opinion.  I heard
that people successfully run Gnus in a MS Windows environment, so it
might be a solution to your seek, who knows!

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From cswiger at mac.com  Sun Apr  6 15:58:39 2003
From: cswiger at mac.com (Chuck Swiger)
Date: Sun, 06 Apr 2003 15:58:39 -0400
Subject: Pythonic way of web-programming
In-Reply-To: <jQ_ja.12641$7w2.9369@nwrddc01.gnilink.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <e9570f37.0304060944.79564f88@posting.google.com> <jQ_ja.12641$7w2.9369@nwrddc01.gnilink.net>
Message-ID: <3E9086EF.9000802@mac.com>

Neil Schemenauer wrote:
[ ... ]
> For example, we have a web application that interacts with
> FedEx's servers in order to create new shipping labels, track
> packages, etc.  With an event driven system, waiting for a
> response for FedEx would freeze the entire server.  Obviously
> that can be fixed the hooking into the event loop.  It's just a
> matter of programming, right?

Apparently almost anything can be regarded as a matter of programming, 
as "Strauser family release #3" demonstrates (congratulations!)...  :-)

Do your queries against FedEx's system need to be atomic, and in what 
ways?  If you want is to make multiple outstanding transactions against 
a server asyncronously using an event-driven system without blocking, 
can't you have a multithreaded component that generates queries, listens 
until it gets the response, and then adds each completed transaction as 
an event to a single event queue?

The event-loop & the logic of your application doesn't need to block 
simply because there are some outstanding queries which haven't all been 
answered; you'd only block if there aren't any responses available to 
work on.

> Another example: we have an application that does some CPU
> intensive operations.  We have multiple CPUs in our server.  An
> event driven system would not (normally) take full use of the
> processors.  Also, clients who have IO bound requests would be
> blocked by one CPU bound request.

Hmm.  Well, to continue the paradigm I was using above for this example, 
you could pull off as many responses off the event queue as you have 
processors available; each time a processor finishes with one event, it 
then grabs the next event from the queue.

Sure, if the events might be related to each other-- where the system 
might need to update data which is in use (being processed, has some 
dependency, etc), then the system gets more complex.  You might need to 
have mutexes around the event queue or other data structures, use a 
2-phase transaction with explicit commit strategy, scan for and coalese 
related events in the queue to catch updates, dispatch notifications of 
changed data, interrupt & restart existing calculations, whatever.

Anyway, all of those are solutions to complexities resulting from the 
application's needs; if what you're doing doesn't need them, then the 
"one line of customers feeding multiple bank attendants" type of event 
dispatch model should work fine.

> I have nothing against the event driven model.  However, it is
> not the solution to every network server problem.

I certainly agree with your general notion: programming paradigms may be 
well-suited for some tasks and not well-suited for others.  Perhaps it's 
not clear to me what else you had in mind: non-blocking I/O like 
select() or poll()?  What kind of SMP dispatch system are you using now?

Network server architectures tend to be event driven in design because 
most flavors of network IPC are quantized into discrete network packets 
by the lower levels of the system; local IPC via SysV shared memory is a 
notable exception, though.  Event-driven models tend to be used a lot 
for various aspects of system UIs, and some of those systems might also 
display an unfortunate tendency to have a UI that becomes modal or 
single-threaded while processing events like mouse-drags, but the two 
are not the same thing.

-Chuck






From gradha at titanium.sabren.com  Sun Apr 13 15:23:42 2003
From: gradha at titanium.sabren.com (Grzegorz Adam Hankiewicz)
Date: Sun, 13 Apr 2003 21:23:42 +0200
Subject: an ugly file-reading pattern
In-Reply-To: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
Message-ID: <20030413192342.GA10327@pedos.es>

On 2003-04-13, Istvan Albert <ialbert at mailblocks.com> wrote:
> if I choose to process the file line by line the "Learning Python"
> book advises me to use the following atrocity:
> 
> while 1:
> 	line = file.readline()
> 	if not line: break
> 
> Maybe I'm picky here but having to start an infinite loop then
> needing a conditional to break out of it is anything but an
> elegant pattern.

An idiom I prefer over that which works with every python version:

        line = file.readline()
        while line:
                do_stuff()
                line = file.readline()

But it also makes for an infinite loop if you happen to forget the
last line of the loop. Life is tough, you have to accept that or
move to newer python versions, as suggested.

-- 
 Please don't send me private copies of your public answers. Thanks.




From theller at python.net  Wed Apr 16 10:02:27 2003
From: theller at python.net (Thomas Heller)
Date: 16 Apr 2003 16:02:27 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com>
Message-ID: <n0iqy1zg.fsf@python.net>

"Terry Reedy" <tjreedy at udel.edu> writes:

> > What do people think? Is this a useful change, and should I submit a
> > patch?
> 
> I plan to start using inittest Real Soon Now.  For normal use,  the
> irrelevant display of inittest internals constitutes noise and I
> believe I would prefer its suppression.  And yet, it is a bit
> disconcerting: the last displayed line does *not* raise the displayed
> exception.  So there is a visual 'mismatch'.  Two suggestions for
> further revision.
> 
> 1. The phrase 'most recent call last' is not longer true.  Substitute
> 'inittest internals suppressed' or somesuch.
> 
> 2. Add some visual indicator of the omission, such as an ellipsis line

> 
> This patch effectively assumes that inittest works perfectly now and
> after all future patches.  While I would make the suppression the
> default, I suspect that normal tracebacks might sometimes be a needed
> option.

These are valid points, IMO. Thanks for these suggestions.
Here is the next round. The append script prints this output,
is this better? Note that the first testcase (test_exception)
is an error, and the other three are simply failed tests.

------------------------------------------------
EFFF
======================================================================
ERROR: test_exception (__main__.FailingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "xunit.py", line 9, in test_exception
    a = 1/0
ZeroDivisionError: integer division or modulo by zero

======================================================================
FAIL: test_failIfEqual (__main__.FailingTests)
----------------------------------------------------------------------
TestFailed: 0 == 0
  File "xunit.py", line 15, in test_failIfEqual
    self.failIfEqual(self.a, self.a)

======================================================================
FAIL: test_failUnlessEqual (__main__.FailingTests)
----------------------------------------------------------------------
TestFailed: 0 != 1
  File "xunit.py", line 12, in test_failUnlessEqual
    self.failUnlessEqual(self.a, self.b)

======================================================================
FAIL: test_failUnlessRaises (__main__.FailingTests)
----------------------------------------------------------------------
TestFailed: wrong exception: 'ValueError: 10'
  File "xunit.py", line 18, in test_failUnlessRaises
    self.failUnlessRaises(TypeError, self._raise, ValueError, 10)

----------------------------------------------------------------------
Ran 4 tests in 0.010s

FAILED (failures=3, errors=1)
------------------------------------------------

#------------------
import unittest

class FailingTests(unittest.TestCase):
    a = 0
    b = 1
    def test_exception(self):
        a = 1/0

    def test_failUnlessEqual(self):
        self.failUnlessEqual(self.a, self.b)

    def test_failIfEqual(self):
        self.failIfEqual(self.a, self.a)

    def test_failUnlessRaises(self):
        self.failUnlessRaises(TypeError, self._raise, ValueError, 10)

    # helper method
    def _raise(self, exc_class, value):
        "Helper method to raise an exception"
        raise exc_class, value

if __name__ == '__main__':
    unittest.main()
#------------------



From berkowit at silcom.com  Fri Apr 18 12:50:07 2003
From: berkowit at silcom.com (Paul Berkowitz)
Date: Fri, 18 Apr 2003 09:50:07 -0700
Subject: Which documentation for 2.2?
Message-ID: <BAC57ACF.35B11%berkowit@silcom.com>

Can someone tell me why there are two separate sets of Documentation at
python.org for Python 2.2? One was released 21 December 2001, the other on
29 March 2002. When you go to the latter it says it's Python 2.2p1. Is this
an updated (corrected?) documentation for the same Python 2.2 release as the
other (in which case why is the earlier set still published), or is it
documentation for a 2.2p1 update of Python itself (in which case why doesn't
it say so on the main page?)

Thanks for any clarification which can be provided. I'm trying to discover
which set applies to the Python 2.2 installed by Mac OS 10.2. I will ask at
PythonMac SIG list as well.

-- 
Paul Berkowitz






From maxm at mxm.dk  Wed Apr  9 03:54:00 2003
From: maxm at mxm.dk (Max M)
Date: Wed, 09 Apr 2003 09:54:00 +0200
Subject: sorting strings written to files
In-Reply-To: <b70igq$9n0ff$2@ID-99293.news.dfncis.de>
References: <2745895.1049870828@dbforums.com> <b70igq$9n0ff$2@ID-99293.news.dfncis.de>
Message-ID: <3E93D198.1030801@mxm.dk>

William Park wrote:
> onefatman <member27857 at dbforums.com> wrote:
> 
>>Can someone help me?
>>
>>I have a small phone directory .dat file that has the entries
>>written to it as strings with three fields in each, separated by \t
>>and ended by \n.
>>
>>e.g.
>>phbook.write("Joe\tNobody\t62907243\n")
>>phbook.write("Daniel\tStevens\t62962328\n")
>>phbook.write("Ellen\tWay\t62910758\n")
>>phbook.write("Jason\tGreig\t62314587\n")
>>
>>it has a function that allows you to write new entries, but i need to
>>know how to sort them by first name, and then by last name when the
>>first names are the same.
> 
> 
> If it's already written to file, then 
>     man sort
> If it's still in Python, then
>     list.sort()


This could give problems:

l = [
     'a\ta',
     'a\tb',
     'a\tc',
     'a\tad',
     'a\tbe',
     'a\tcf',
     'aa\ta',
     'ab\tb',
     'ac\tc',
     'ad\tad',
     'ae\tbe',
     'af\tcf',
]

l.sort()
print l
 >>> ['a\ta', 'a\tad', 'a\tb', 'a\tbe', 'a\tc', 'a\tcf', 'aa\ta',
  'ab\tb', 'ac\tc', 'ad\tad', 'ae\tbe', 'af\tcf']

You would need to split them on the '\t' first.

l = [
     'a\ta',
     'a\tb',
     'a\tc',
     'a\tad',
     'a\tbe',
     'a\tcf',
     'aa\ta',
     'ab\tb',
     'ac\tc',
     'ad\tad',
     'ae\tbe',
     'af\tcf',
]

# decorate sort undecorate
l2 = [i.split('\t')[:2] + [i] for i in l]
l2.sort()
sorted = [i[-1] for i in l2]
print sorted

 >>> ['a\ta', 'a\tad', 'a\tb', 'a\tbe', 'a\tc', 'a\tcf', 'aa\ta',
'ab\tb', 'ac\tc', 'ad\tad', 'ae\tbe', 'af\tcf']

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From andymac at bullseye.apana.org.au  Sat Apr  5 23:05:47 2003
From: andymac at bullseye.apana.org.au (Andrew MacIntyre)
Date: Sun, 6 Apr 2003 15:05:47 +1100 (edt)
Subject: Why deleting large variables doesn't free up any memory?
In-Reply-To: <xizja.6994$pK2.7213@news.indigo.ie>
Message-ID: <Pine.OS2.4.44.0304061439560.45265-100000@tenring.andymac.org>

On Sat, 5 Apr 2003, Nick Hilliard wrote:

> sdieselil wrote:
>
> > I'm using Python 2.2.2 under FreeBSD. When I create a large variable
> > (using for example "a = range(1000000)") 'top' command indicates that
> > Python ate 12MB of swap space. But when I execute "del a" nothing
> > happens, it doesn't release these 12MB! Using "gc.collect()" also
> > doesn't help. What's the problem? What happened to Python's garbage
> > collector?
>
> Regardless of the other answers in this thread, I suspect that this problem is
> probably less important than you think.  If FreeBSD thinks that the memory isn't
> being used, it will be swapped out if there's a memory shortfall at any stage.
> The important factor is amount of memory which is being actively used at a
> particular time.

There's an additional subtlety with the OP's example: a only references
the list.

Each individual element in the list is an integer object in its own right.
As a performance optimisation, the memory for all integer objects created
is kept in a single list, and never free()ed.  Integer object memory
released by garbage collection of unreferenced integers (except -5..100,
which are interned) is reused for new integer references.

So when a is derefenced, only about 4MB (the size of the list vector) is
actually free()ed, which may or may not be returned to the OS by the
malloc implementation.

xrange() may help the OP, depending on the actual usage of a.

--
Andrew I MacIntyre                     "These thoughts are mine alone..."
E-mail: andymac at bullseye.apana.org.au  | Snail: PO Box 370
        andymac at pcug.org.au            |        Belconnen  ACT  2616
Web:    http://www.andymac.org/        |        Australia





From dogen at mindspring.com  Tue Apr  8 17:06:53 2003
From: dogen at mindspring.com (Jon Brown)
Date: Tue, 08 Apr 2003 17:06:53 -0400
Subject: Palindrome finder: help find ways to speed up?
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org>
Message-ID: <BAB8B22D.2022%dogen@mindspring.com>

Steven,

Actually the program does exactly what you describe, except from the inside
out because it starts in the middle, and it cuts out of the loop if the
characters don't match.

To look for pals it spins off big, big numbers of combinations, true enough.
That's what behind the need for speed. Since Python is inherently on the
pokey side, I expected the slowness but wonder if I'm missing a cool way to
pick up the pace. But I suspect the solution will involve C++, and I've
never linked C/C++ with Python.

Jonedary 

> A smarter search would do better.  One approach would be to build
> the palindrome from both ends instead of from just one end.  For
> example, suppose your code understood that a palindrome starting
> with 'orange' must end with a word 'o', 'ro', 'aro', 'naro',
> 'gnaro', 'egnaro', or a word ending with 'egnaro'.  If there are
> no such words, it can skip 1/n of the search space, namely all the
> combinations which start with 'orange'.




From jdhunter at ace.bsd.uchicago.edu  Thu Apr 10 14:34:43 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Thu, 10 Apr 2003 13:34:43 -0500
Subject: pygtk install
In-Reply-To: <3E9273D6.7030500@netscape.net> (vector's message of "Tue, 08
 Apr 2003 17:01:42 +1000")
References: <3E9273D6.7030500@netscape.net>
Message-ID: <m2u1d68am4.fsf@mother.paradise.lost>

>>>>> "vector" == vector  <Vector180W at netscape.net> writes:
    vector> load failed: The specified module could not be found.

    vector> I then tried GTK+-Runtime-Environment-2.2.1.1.exe and
    vector> reloaded python etc all the same

Did you set your PATH variable to include the GTK lib and bin subdirs?

Here are my notes for installing the GTK 2.2.1.1 libs

   Extract these to C:\GTK. Obtained from
   http://sourceforge.net/projects/gtk-win/. This will install a bunch
   of files to C:\GTK. You must add this dir to your PATH environment
   variable.  Go to your My Computer icon, right-click it, and select
   the Properties tab. Under the Advanced tab, choose Environment
   Variables. Double click on the system variable path and add a
   semicolon followed by the dir C:\GTK\lib;c:\GTK\bin. Click OK,
   closing the window. Or edit autoexec.bat on windows 98.

If your path is set correctly, pygtk *should* work.  If it fails, fire
up a copy of Dependency Walker and try to load the pyd module into
Dependency Walker and see what modules are missing.

John Hunter




From mis6 at pitt.edu  Fri Apr 18 09:55:02 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 18 Apr 2003 06:55:02 -0700
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de> <mailman.1050586468.29594.python-list@python.org> <rozna.115488$It5.70451@news2.central.cox.net> <b7mn1i$2icfr$2@ID-148859.news.dfncis.de> <LBBna.43583$T34.1210957@news2.tin.it> <b7nttn$2p743$3@ID-148859.news.dfncis.de>
Message-ID: <2259b0e2.0304180555.6c063ec3@posting.google.com>

Haran Shivanan <hs2l2ATyahoodotcom at yahoo.com> wrote in message news:<b7nttn$2p743$3 at ID-148859.news.dfncis.de>...
> module "Alex Martelli" core dumped with the following output:
> [snipped informative stuff]
> > 
> > This prints 11, as apparently required by your specs.  I hope
> > this can help you make the custom metaclass you need to
> > implement your actual, exact specifications, of course.
> Thanks *a lot* for your help.
> Do you know of any on-line resources where I can get info
> on metaclasses and using them?
> I was only dimly aware of their presence until your post.
> Thanks.

Alex didn't want to mention it, but his "Python in a Nushell" contains a
discussion of metaclasses and the metametaBunch example (plus many other
interesting things not necessarely related to metaclasses ...)

Good luck,

                    Michele



From guruprasad_p at lycos.com  Tue Apr  8 05:56:47 2003
From: guruprasad_p at lycos.com (Guruprasad Prahallad)
Date: 8 Apr 2003 02:56:47 -0700
Subject: Need help with Python porting
Message-ID: <78ada8a6.0304080156.6a780fe2@posting.google.com>

Hi all,

I am currently porting Python to Netware. I would really appreciate if
someone could tell me where I could find a reference implementation
for Netware. I downloaded the 2.2.2 source from python.org for Linux.
Could someone give me a pointer as to how to go about modifying this
code so that it works for NetWare.

I will be extremely grateful if any help is provided in this regard.

Thanks in advance,
Guru



From claird at lairds.com  Tue Apr 15 09:52:11 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 15 Apr 2003 13:52:11 -0000
Subject: book on Python for Fortran programmers
References: <3064b51d.0304150458.595afc30@posting.google.com> <b7h09k$q6u1$1@ID-134488.news.dfncis.de>
Message-ID: <v9o3kb2476qt3e@corp.supernews.com>

In article <b7h09k$q6u1$1 at ID-134488.news.dfncis.de>,
Jos? Rui Faustino de Sousa  <jrfsousa.removeFF2A45B2this at esoterica.pt> wrote:
>beliavsky at aol.com wrote:
>> started in Python?
>>
>
>There is a welt of information on Python's website:
>
>http://www.python.org/
>
>More directly you can try:
>
>http://www.python.org/doc/Newbies.html
>
>http://www.python.org/doc/Intros.html
>
>http://www.python.org/cgi-bin/moinmoin/PythonBooks
>
>And there is always:
>
>news://comp.lang.python
>
>It is pretty high traffic newsgroup (maybe too much...)
>
>And do not forget to have fun learning Python ;-)
			.
			.
			.
Mr. Beliavsky, <URL: http://python.org/topics/scicomp/ > and
<URL: http://biopython.org/ > are pages that you as a Fortran
adept might find particularly inviting.  Also, <URL: http://
www.ddj.com/topics/pythonurl/ > addresses the availability of
digests for those who wish to "keep up" with Python news with
a minimum of invested time.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From donn at u.washington.edu  Fri Apr  4 15:13:56 2003
From: donn at u.washington.edu (Donn Cave)
Date: 4 Apr 2003 20:13:56 GMT
Subject: Why deleting large variables doesn't free up any memory?
References: <169a7283.0304040936.69e2b9a1@posting.google.com> <b6khbt$ftf$1@panix2.panix.com> <slrnb8rmbn.rs.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <b6kp24$1l46$1@nntp6.u.washington.edu>

Quoth Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com>:
| In article <b6khbt$ftf$1 at panix2.panix.com>, Aahz wrote:
|> In article <169a7283.0304040936.69e2b9a1 at posting.google.com>,
|> sdieselil <sdieselil at yahoo.com> wrote:
|>>
|>> I'm using Python 2.2.2 under FreeBSD. When I create a large variable
|>> (using for example "a = range(1000000)") 'top' command indicates that
|>> Python ate 12MB of swap space. But when I execute "del a" nothing
|>> happens, it doesn't release these 12MB! Using "gc.collect()" also
|>> doesn't help. What's the problem? What happened to Python's garbage
|>> collector?
|> 
|> That's got nothing to do with Python; generally speaking, once the OS
|> allocates memory to a process, the process keeps it until it exits.
|> However, if you allocate and de-allocate big chunks of memory, you'll
|> notice that Python's size stays static -- the memory gets re-used.
|
| Actually it's an issue for the heap manager, not the OS.  Under Windows
| and Unix at least, it is possible for a process to return memory to
| the OS.  In practice allocation is done in chunks of at least a page
| and a page can't be freed until every object in it has gone.  Further-
| more the memory will probably be needed again, and if it isn't it will
| be paged out to disk.  So most heap managers don't bother to do this,
| except for very large allocations.

Now that's interesting.  Could you provide a reference, say a man page
for a system call, for how a user process would return memory, on a
UNIX platform?  (Memory that doesn't just happen to be at the top of
the data segment, of course.)

Thanks,
	Donn Cave, donn at u.washington.edu



From maxm at mxm.dk  Fri Apr 25 06:33:10 2003
From: maxm at mxm.dk (Max M)
Date: Fri, 25 Apr 2003 12:33:10 +0200
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
In-Reply-To: <oafgavg8e0pf512712g4aenldqktvmb0dr@4ax.com>
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <oafgavg8e0pf512712g4aenldqktvmb0dr@4ax.com>
Message-ID: <I78qa.30798$y3.2515733@news010.worldonline.dk>

Stephen Horne wrote:

> Experts are only expert in particular fields. Just because Einstein
> was a genius in the field of physics, that did not qualify him to
> state that only 10% of the brain is unused. Assuming he claimed it at
> all - I never checked, but it does get attributed to him now and again
> - it was probably never intended as a statement of strict biological
> fact. Benjamins view of Dijkstra's point can probably be seen in a
> similar light.


If only 10% of the brain is used, why don't we hear:

"Well he was shot in the head, but the bullet only went through the 90% 
of the brain that he doesn't use. So no sweat mate."

;-)

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From exarkun at intarweb.us  Tue Apr  1 02:16:11 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 1 Apr 2003 02:16:11 -0500
Subject: When Python outruns C++ ...
In-Reply-To: <slrnb8hrv9.mk0.chasm@galileo.rift>
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift>
Message-ID: <20030401071611.GA31357@meson.dyndns.org>

On Tue, Apr 01, 2003 at 01:48:26AM +0000, Julian Tibble wrote:
> [snip]
> 
> 
> Julian
> 
> p.s. maybe my python version is not very good, feel free to make
> it go faster:
> 
> #
> # Count the number of occurences of each word in a file
> # -- Julian Tibble
> #
> 
> import re
> 
> # "Word -> Number of occurences" mapping
> wordcount = {}
> 
> # Regular expression to match a word 
> re_word = re.compile('[a-z]+')
> 
> # Read in whole file and split into words
> try:
> 	while 1:
> 		line = raw_input().lower()
> 		line = re_word.findall(line)
> 		for word in line:
> 			if wordcount.has_key(word):
> 				wordcount[word] += 1
> 			else:
> 				wordcount[word] = 1
> except EOFError:
> 	pass
> 
> 
> # Order by descending number of occurences
> sortable = zip(wordcount.values(), wordcount.keys())
> sortable.sort()
> sortable.reverse()
> 
> for n,w in sortable:
> 	print w,n
> 

  It looks like this could easily be improved with the application of a
slightly smarter searching algorithm, such as Rabin-Karp search or
Boyer-Moore find.

  Both of these use a technique of doing a little extra work up front to
make each pass through the loop much less expensive.  Google for details.

  Jp

-- 
"There is no reason for any individual to have a computer in their
home."
                -- Ken Olson, President of DEC, World Future Society
                   Convention, 1977
-- 
 up 12 days, 2:00, 6 users, load average: 0.00, 0.00, 0.00




From hupp at cs.wisc.edu  Tue Apr 29 09:55:54 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Tue, 29 Apr 2003 08:55:54 -0500
Subject: os.spawn
In-Reply-To: <Xns936C9D9E91B1Bthewolffman@194.109.133.29>
References: <Xns936C9D9E91B1Bthewolffman@194.109.133.29>
Message-ID: <20030429135554.GA27957@upl.cs.wisc.edu>

On Tue, Apr 29, 2003 at 01:29:41PM +0000, Jeroen Wolff wrote:
> Hi,
> 
> I want to run some shell commands from within python. 
> Wat i normal do at the shell prompt is:
> 'flow-cat <dirname> | flow-stat <param> > outputfile.txt'
> 
> I like to do this from within python and wait for it to finish and the 
> process the outputfile.txt
> Is this possible? The problem is with '|' and '>" piping....

You want to use os.system or os.popen.  With popen you can read the
results directly without using the intermediate file. 

os.system('flow-cat <dirname> | flow-stat <param> > outputfile.txt')

or

flow = os.popen('flow-cat <dirname> | flow-stat <param>')

for i in flow.readlines():
	...


-Adam




From syver-en+usenet at online.no  Sun Apr  6 08:26:29 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 06 Apr 2003 14:26:29 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <u65pr7qx6.fsf@online.no>

pobrien at orbtech.com (Patrick K. O'Brien) writes:

> Sound like fun?

Sounds like Pymacs ;-)
http://www.iro.umontreal.ca/~pinard/pymacs/
-- 

Vennlig hilsen 

Syver Enstad



From max at cNOvSisiPonAtecMh.com  Tue Apr 29 14:13:23 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Tue, 29 Apr 2003 18:13:23 GMT
Subject: stupid q-n (wxPy)
References: <Ukxra.33100$7M5.3263531@twister.nyc.rr.com> <3EAEB850.50D9DEB9@engcorp.com>
Message-ID: <7hzra.27540$su3.3180311@twister.nyc.rr.com>

The only detail I have is that the script was copy-pasted out straight of
wxPython demo app. There were no specific instructions in the readme for
running the code so I just gave it a shot. Since wxPython is so popular I
fugured someone would just tell me "didn't you bla-bla, (idiot)?".
anyway, thanks.

--
========================================
Max Khesin, software developer -
max at cNvOiSsPiAoMntech.com
[check out our image compression software at www.cvisiontech.com, JBIG2-PDF
compression @
www.cvisiontech.com/cvistapdf.html]


"Peter Hansen" <peter at engcorp.com> wrote in message
news:3EAEB850.50D9DEB9 at engcorp.com...
> xam wrote:
> >
> > I am trying to run some wxPy samples and I am getting an error here:
> > if __name__ == '__main__':
> >     import sys,os
> >     import run
> >     run.main(['', os.path.basename(sys.argv[0])])
> >
> > - "import run" fails. I know this is something really stupid, but I
could
> > not find the answer on the net.
> > thanks.
>
> What is "run"?  Where is it?  Are you sure the script you are running
> can find "run.py"?  If there's no run.py in the directory of the script
> you show above, insert a "print sys.path" after the first import statement
> and make sure the list it prints contains the directory where run.py is
> actually installed.
>
> A little more detail/context would be helpful next time.
>
> -Peter





From aleax at aleax.it  Fri Apr 18 17:40:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 21:40:52 GMT
Subject: __getattr__ for globals?
References: <20030418100631860-0400@braeburn.themorgue.org>
Message-ID: <Eh_na.52810$T34.1444328@news2.tin.it>

Kevin Smith wrote:

> Is there a way to get the same type of effect of the __getattr__ method
> on class instances with global variables?  I want to be able to look up
> the value of global variables from an external source if they don't
> exist in the current set of globals.

Module objects don't offer this functionality -- when you access a
global X in module Y (whether by syntax Y.X from elsewhere, or by
just X from code inside Y), you get a direct access to Y.__dict__['X'],
period.  And you cannot rebind Y.__dict__ , either.

So why don't you use a class instance to hold your "globals"
instead?

class Globals:
    "whatever you wish"
glob = Globals

and use glob.X, glob.Y and so on rather than the corresponding
barenames.  This way you get to do whatever you wish AND your
code arguably gets more readable too...


Alex




From afd00 at aub.edu.lb  Sun Apr  6 06:16:21 2003
From: afd00 at aub.edu.lb (Ali Dada)
Date: Sun, 6 Apr 2003 10:16:21 GMT
Subject: PYTHONPATH question
Message-ID: <E1927CG-0007k5-00@fast.aub.edu.lb>

hi all:

i am working on RedHat 8.0
shouldn't i be able to read PYTHONPATH directly like i read any other env 
variable? because i am saying
echo $PYTHONPATH
and nothing is being displayed. the thing is i want to add an entry to 
PYTHONPATH and i don't want to use sys.path because i want the new entry to 
remain permanently.
thanks for any advice..










From daniele.trovato at obtree.com  Wed Apr 16 05:50:18 2003
From: daniele.trovato at obtree.com (Daniele Trovato)
Date: Wed, 16 Apr 2003 11:50:18 +0200
Subject: PyQT and Eric
References: <b7is7g$i1d$1@rex.ip-plus.net> <mailman.1050479366.18545.python-list@python.org>
Message-ID: <b7j8uk$jit$1@rex.ip-plus.net>

Hi Phil

I also tried to run PyQT or Python examples with idle and this tool gave me
also an error, but i can't remember exactly. I just tried to open a python
file and then i had a lot of red messages in the idle tool and i also have
thoose red error messages when want do something with the tool e.g "open a
new file". But I can tell this evening more about the errors, because i
don't have install SuSE on my business workstation.

Thanks Daniele



"Phil Thompson" <phil at river-bank.demon.co.uk> wrote in message
news:mailman.1050479366.18545.python-list at python.org...
> On Wednesday 16 April 2003 7:13 am, Daniele Trovato wrote:
> > Hi
> >
> > I tried to install Eric, but it won't work and I don't no why.
> > I'm using SuSE 8.2 and I installed Python, PyQT and PyQT Develop,
> > QScintilla, SIP ..... packages from the SuSE distribution.
> > My problem is, when I'm trying to install Eric with the following
command
> > "python install.py", the shell returns me an error: "Sorry, Please
install
> > PyQT"  but I have PyQT already installed.
> > Does anybody now where the problem is or how i can solve the problem ?
>
> Can you run any of the PyQt examples?
>
> What's the output of running the following (there shouldn't be any)?
>
> python -c "from qt import qVersion, PYQT_VERSION"
>
> The above is what the eric installer is running to see if PyQt is
installed.
>
> What version of PyQt?
>
> Phil
>





From martin at v.loewis.de  Mon Apr 21 16:26:35 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Mon, 21 Apr 2003 22:26:35 +0200
Subject: Locale settings
In-Reply-To: <3ea437a1_1@dns.sd54.bc.ca>
References: <3ea437a1_1@dns.sd54.bc.ca>
Message-ID: <b81k5m$822$06$1@news.t-online.com>

Bob van der Poel wrote:

> Now, for the most part this works. However, one user of my program has 
> just upgraded to Linux Mandrake 9.1 and (I think) he has set his 
> language to "de". In any event, the getdefaultlocal() seems to return 
> 'de' which isn't a valid python encoding (??) and the above code then 
> converts to his 'ascii'.

Using getdefaultlocale() is broken, and cannot be fixed. On Unix, use

locale.nl_langinfo(locale.CODESET)

instead; if supported, this will tell you the user's preferred encoding.
Be aware that
a) neither nl_langinfo nor CODESET may be supported by the system,
b) they may return a codeset for which a Python codec is not available,
c) you may have to invoke setlocale(locale.LC_CTYPE, "") before
    nl_langinfo returns a reasonable response (e.g. on Solaris)

Python 2.3 has a function locale.getpreferredencoding which does all
that. I recommend you copy its source code into your application, and
use that instead.

Regards,
Martin




From aleax at aleax.it  Mon Apr 14 15:31:04 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 19:31:04 GMT
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il>
Message-ID: <Y%Dma.13754$LB6.370780@news1.tin.it>

Avner Ben wrote:

> Question:
> 
>     Does standard Python have case-insensitive string comparison and
>     search?

No.  You normally accomplish them either with regular expressions (which
CAN be compiled into case-insensitive mode) or by coercing strings to
one fixed case (e.g. with the .lower method of string objects).


Alex




From duncan at NOSPAMrcp.co.uk  Wed Apr  9 06:20:01 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 9 Apr 2003 10:20:01 +0000 (UTC)
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de>
Message-ID: <Xns9358723026D8Cduncanrcpcouk@127.0.0.1>

skywalkerpackage at hotmail.com (Johannes Eble) wrote in
news:3e93e50f.676072541 at news.online.de: 

> is there a library available for Python that behaves similar to MS
> Visio? For example, I would like to create vector graphic objects and
> connect the with a line. If I move, flip or rotate a connected object,
> the line should follow the connection.
> 
> I would like to draw UML diagrams.

Maybe not quite what you asked but you might want to look at
ObjectDomain which is a UML diagram editor that is fully scriptable in
Python (or rather Jython). 

> But the program is in c and it is tightly coupled with GTK+.
> It seems to be an impossible mission to control dia from
> another program, at least in Windows.

A quick look at the Dia website seems to indicate that DiaPython exists.

-- 
Duncan Booth                                            
duncan at rcp.co.uk int month(char
*p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3" 
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure? 



From mwh at python.net  Thu Apr 10 05:38:33 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 10 Apr 2003 09:38:33 GMT
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org> <XSXja.346441$3D1.186570@sccrnsc01> <mailman.1049741353.6320.python-list@python.org> <b71pmn$ibu$03$1@news.t-online.com>
Message-ID: <7h33ckq1ylj.fsf@pc150.maths.bris.ac.uk>

Dominic <oblivious at web.de> writes:

> > But it's especially nice to wrap C with Pyrex because it VERY easy to adapt
> > interface to pythonic way.
> 
> Pyrex is a very nice tool :) I like it very much.
> 
> I have used it to glue xmms-control with a
> python script. The char* strings returned
> by xmms need to be freed which can be
> easily done with Pyrex and a
> two line utilitiy function.
> My previous swig wrapped functions
> simply leaked the memory :(

Heh, I just rewrote libxmms in Python...

Cheers,
M.

-- 
  No.  In fact, my eyeballs fell out just from reading this question,
  so it's a good thing I can touch-type.
                                    -- John Baez, sci.physics.research



From aleax at aleax.it  Tue Apr 15 08:21:45 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 12:21:45 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <mailman.1050173000.12088.python-list@python.org> <slrnb9nm7j.1bc.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <tPSma.16430$LB6.428266@news1.tin.it>

Ben Hutchings wrote:

> In article <mailman.1050173000.12088.python-list at python.org>,
> Alan McIntyre wrote:
>> Just off the cuff here, it seems to me that the time required to do
>> "string += string" is going to be proportional to the size of the
>> original string whether you do it in Python or with the STL's
>> std::string or with C strings (so long as you don't know beforehand how
>> long the final string will be).
> <snip>
> 
> Mutable string implementations may allocate more memory than their
> current contents require, in order to avoid future reallocation and
> copying.  A string implementation that grows its buffer exponentially
> (e.g. doubling it every time it needs to grow) will have amortised
> linear time complexity for additions, just like the STL vector or
> Python list.

I don't know if a std::string implementation would be *ALLOWED* to
"waste" arbitrary amounts of memory to achieve amortized O(1)
performance [I think I recall it is -- there being no O() constraints
on std::string memory OR runtime performances, I believe], but it's
certainly not *REQUIRED* to do so (sometimes an std::vector<char>
gives you much better performance than an std::string does...).

For an interestingly-off-mainstream implementation strategy I
suggest reading:
http://www.sgi.com/tech/stl/Rope.html
http://www.sgi.com/tech/stl/ropeimpl.html


Alex




From igorr at ifi.uio.no  Sat Apr 12 10:22:50 2003
From: igorr at ifi.uio.no (Igor V. Rafienko)
Date: 12 Apr 2003 16:22:50 +0200
Subject: hex and unsigned longs
Message-ID: <xjvof3bu75x.fsf@margvisi.ifi.uio.no>

Hi,

The documentation for the built-in function hex() says that hex()
always returns an unsigned literal. However, on my 2.2.1 python, I get
the following results:

>>> hex( 1 )
'0x1'
>>> hex( 1L )
'0x1L'
>>> hex( -1 )
'0xffffffff'
>>> hex( -1L )
'-0x1L'
>>> 

The literal in the last case is obviously not unsigned. What am I
missing? (I.e. why are longs treated differently than ints by hex?)

TIA,





ivr
-- 
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?



From frank at niessink.com  Tue Apr 29 14:37:43 2003
From: frank at niessink.com (Frank Niessink)
Date: Tue, 29 Apr 2003 20:37:43 +0200
Subject: Sorting without transitivity
In-Reply-To: <20030419213405.GA588@obstler>
References: <20030419213405.GA588@obstler>
Message-ID: <20030429183743.GA612@obstler>

On Sat Apr 19 23:34:05 2003, Frank Niessink wrote:
> Hi all,
> 
> [It's been while since I've dealt with transitivity and stuff, so
> I hope I got the terminology right]
> 
> I have objects that are not transitivily ordered, e.g.:
> 
> A > B
> A == C
> B == C
> 
> Now I have a list [A, B, C] and I want to sort it.
> The resulting list should have B before A, I don't care where
> C is. However, the Python list.sort() method (Python 2.2.2 on cygwin)
> doesn't put A and B in the right order.
> 
> How can I get the desired behavior? Should I implement my own
> sort algorithm? Any pointers?

Hi all,

First, thanks for all the help and sorry for not responding sooner.
I really appreciate the cheerful and supporting nature of this group.

I've looked into the existing topological sorting implementations in
Python, but all of them seemed to assume that the user of the method
already knows what items are smaller than which other items. 

In my case I don't have that information readily available, so I 
wanted the algorithm to take care of that. Also, giving an exception 
when there are cycles was not acceptable. 

So, I've been working on another implementation. Below is a tlist 
class that sorts its contents topologically. List items should 
provide a __lt__ method and an __eq__ method.

Any comments? Suggestions for improvements?

(Of course I have tests, mail me if you'd like to see the tests as well.)

---------- tlist.py --------------------

class tlist(list):
    ''' Topological list class. This class uses topological sort to
        sort its contents. List items should provide __lt__ and 
        __eq__ methods. '''
        
    def __init__(self, initial=None):
        list.__init__(self, initial or [])
        self.sort()

    def sort(self):
        result = []
        while self:
            result.append(self.popmin())
        self[:] = result

    def popmin(self):
        min = self.min()
        self.remove(min)
        return min

    def min(self):
        smallest, remainder = self[0], self[1:]
        found = False
        while not found:
            for item in remainder:
                if item < smallest:
                    remainder.remove(item)
                    smallest = item
                    break
            else:
                found = True
        return smallest

----------------------------------------

Thanks again, Frank

-- 
The Nutri-Matic was designed and manufactured by the Sirius Cybernetics 
Corporation whose complaints department now covers all the major land masses
of the first three planets in the Sirius Tau Star system.
	-- Douglas Adams, 'The Hitch Hiker's Guide to the Galaxy'




From claird at lairds.com  Mon Apr 28 16:02:04 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 28 Apr 2003 20:02:04 -0000
Subject: please comment on technologies
References: <slrnbalplq.254.luc@trillian.dont-panic.info>
Message-ID: <var25si68t3425@corp.supernews.com>

In article <slrnbalplq.254.luc at trillian.dont-panic.info>,
luc wastiaux  <luc at nospam.com> wrote:
>Hello, I am writing the specifications for a school project (the subject 
>is free for us to choose), I was thinking about doing a news-server 
>archive database just like google groups, but on a smaller scale (it would 
>only archive a limited number of newsgroup, not all of USENET). I got this 
>idea because my school uses a news server and a lot of valuable 
>information is lost when old messages are erased from the server's spool.
>
>The project would be developped by a group of 4 to 6 CS students 
>(including me), over a three months timespan (but not full time, we have 
>to go to classes and other stuff)
			.
			.
			.
>I'm (almost) sure we will make use of python as the main language, and use
>either Mysql or postgresql for the database (I'm familiar with mysql but
>maybe it's too limited for what I want to do?). I just found out about
			.
			.
			.
I'll say a few things I've read from no one else.

If a commercial client told me that he depended on Usenet, and
it was working well, but "a lot of valuable information is lost
...", my first recommendation would be to tune his Usenet 
service.  That's a way to re-use a lot of existing infrastruc-
ture.  You can set up a special-purpose NetNews server, with 
its own local datastore of just the groups you want, and
configure expiration times to "never".  That's the most con-
servative approach that comes to my mind, and, in
return-on-investment calculations, I'm a conservative.

Also, while it's fine with me that you practice your Mysql or
PostGresql skills, I don't get the point.  You're just storing
messages, right?  It's a write-once-read-many-delete-never
model?  You can do worse than just dumping stuff into the file
system.  Again, look at the existing NetNews server implementa-
tions.

These simplifications free attention to be focused on the
query model, which is likely to be a more interesting problem
than you realize.

For historical interest, I'll note that <URL: http://
phaseit.net/claird/news.lists/newsgroup_archives.html > indexes
a few thousand (well, it used to; maybe a couple hundred are
still live) special-purpose Usenet archives.  Perhaps it'll
amuse you to see how others have worked in this area before
you.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From jack at performancedrivers.com  Tue Apr 15 22:23:22 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Tue, 15 Apr 2003 22:23:22 -0400
Subject: Why functional Python matters
In-Reply-To: <20030416004537.GB7550@meson.dyndns.org>; from exarkun@intarweb.us on Tue, Apr 15, 2003 at 08:45:37PM -0400
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <slrnb9p3vo.vg8.ramen@lackingtalent.com> <20030416004537.GB7550@meson.dyndns.org>
Message-ID: <20030415222322.K1039@localhost.localdomain>

On Tue, Apr 15, 2003 at 08:45:37PM -0400, Jp Calderone wrote:
>     reduce(int.__add__, [1, 2, 3])
>
>   Unfortunately, this breaks when ints get promoted to longs, because you
> can't pass longs to int.__add__!

Or when the passed in list is empty.
Why is the 3rd argument to reduce optional again?

> 
>     map(ClassObj.method, [x, y, z])
> 
>   works, so long as x, y, and z are instances of ClassObj.  But if they are
> instances of different types, all of which merely happen to implement
> method, you're out of luck!  This is resolved with list comprehensions:
> 
>     [i.method(foo) for i in [x, y, z]]
> 
>   But what I'd like to see is a way to do:
> 
>     map(message("method", foo), [x, y, z])
> 
>   Of course, you could easily implement this as:
> 
>     def message(name, *args, **kwarg):
>         return lambda o: getattr(o, name)(*args, **kwargs)
>
>   But this is stupendously inefficient.  It'd be really great if there were
> some built-in MessageType that could be used this way and also managed to
> carry out the work in an efficient manner.

this adds 3 extra function calls, 1 for message(), 1 for lambda, and 1 for
the getattr.

doing plane-jane
  map(lambda x:x.method(foo), [x, y, z])
just has one extra level of function call overhead and does what you want

I was a little disapointed that the implementaion in bltinmodule.c doesn't
special case lambdas but instead does a regular function call.  I don't think
it is possible to change the lambda since they are anonymous, so they could be
speical cased like list comps (which generate the bytecode for a for loop 
during compile)

If lambdas were special-cased in this way the message("method", foo) could
only be _as efficient_ as lambdas even if it were made as core a part of the
language as list comps.  And the message() syntax is harder to read to boot.

-jack




From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr  4 14:13:59 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 4 Apr 2003 19:13:59 GMT
Subject: Why deleting large variables doesn't free up any memory?
References: <169a7283.0304040936.69e2b9a1@posting.google.com> <b6khbt$ftf$1@panix2.panix.com>
Message-ID: <slrnb8rmbn.rs.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <b6khbt$ftf$1 at panix2.panix.com>, Aahz wrote:
> In article <169a7283.0304040936.69e2b9a1 at posting.google.com>,
> sdieselil <sdieselil at yahoo.com> wrote:
>>
>>I'm using Python 2.2.2 under FreeBSD. When I create a large variable
>>(using for example "a = range(1000000)") 'top' command indicates that
>>Python ate 12MB of swap space. But when I execute "del a" nothing
>>happens, it doesn't release these 12MB! Using "gc.collect()" also
>>doesn't help. What's the problem? What happened to Python's garbage
>>collector?
> 
> That's got nothing to do with Python; generally speaking, once the OS
> allocates memory to a process, the process keeps it until it exits.
> However, if you allocate and de-allocate big chunks of memory, you'll
> notice that Python's size stays static -- the memory gets re-used.

Actually it's an issue for the heap manager, not the OS.  Under Windows
and Unix at least, it is possible for a process to return memory to
the OS.  In practice allocation is done in chunks of at least a page
and a page can't be freed until every object in it has gone.  Further-
more the memory will probably be needed again, and if it isn't it will
be paged out to disk.  So most heap managers don't bother to do this,
except for very large allocations.



From oblivious at web.de  Wed Apr  9 14:58:04 2003
From: oblivious at web.de (Dominic)
Date: Wed, 09 Apr 2003 20:58:04 +0200
Subject: Macros in Python?
Message-ID: <b71qde$q71$07$1@news.t-online.com>


(defmacro with-root-privs (() &body body)
   (let ((oldidsym (gensym)))
     `(let ((,oldidsym (geteuid)))
        (seteuid 0)
        (unwind-protect
	   (progn , at body)
	 (seteuid ,oldidsym)))))

Is there an easy way to produce similar code
in Python which gurantees that certain things
happen around some code? (Maybe by using
generators?)

The normal approach would be to enclose everything
with try: finally: but this would scatter the
code all over.

Is something like this feasible?

def server(a,b):
   if cond:
       ...
   else:
     With_Root_Privs:
        if x in user:
           active.append(x)
          (...)

Ciao,
  Dominic




From graham__fawcett at hotmail.com  Mon Apr  7 13:23:03 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 7 Apr 2003 10:23:03 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com>
Message-ID: <e9570f37.0304070923.70612d28@posting.google.com>

Lothar Scholz wrote:
> lekishvili at python.qartu.com (Giorgi) wrote in message news:<a3b5b608.0304040904.146dfff5 at posting.google.com>...
> 
>>- Is Python a real web-programming language at all? I mean, has
>>somebody ever used it for real, commercial tasks? Please, do not
>>consider my question too trivial. There is no language equally
>>suitable for all tasks, however, Python seems to me a language
>>naturally tailored for web-programming; this or that way, I have more
> 
> 
> No its not a naturally tailored web-programming language, because it
> is
> very hard to integrate in Server Pages. You have to change the
> indentation syntax/semantic of python to embedd it in a real useable
> way.
> This is done by Webware but it looks very strange to normal python
> programmers.
> 
> Of course you can use template systems. But this is not always a good
> economic decision.

Lothar, "not integrating well in Server Pages" is an incomplete argument 
against Python as a Web programming language, IMHO.

First, "Server Pages" are a subset of a much larger range of 
"templating" solutions. Server Page schemes tend to have a lot of code 
embedded between pseudo-HTML tags, interspersed with HTML snippets. They 
also tend to be unmaintainable, and often get polluted with application 
logic, when they should be focusing on presentation alone.

If the script in your Server Page is doing much more than iterating a 
list of elements, calling code from a Python module (that you wrote, or 
that's in the standard library), or doing a bit of string manipulation, 
then it's time to stop writing code for a while, Google "separate logic 
from presentation" and think about your approach until you're really 
sure you've got your concerns separated.

I would agree with you that Python isn't overly well suited to the 
Server Page approach, though I've used it in ASP with very few problems 
caused by indentation. Let's not count the language out quite yet...

Python has some outstanding templating options. I think, e.g., that 
ZopePageTemplates (doesn't require Zope, btw) is a work of art, and 
avoids many of the inherent design flaws that mar most templating 
approaches. If you need it to, ZPT will coexist very peacefully with Web 
designer tools (Dreamweaver, etc.), *far* better than most other 
templating systems will, regardless of what language those systems 
favour. And, best, ZPT makes it very *hard* to put application logic 
into the page, thus forcing you to consider presentation alone, a very 
helpful design constraint indeed.

Or, take a look at Quixote's approach to templating. It turns the whole 
concept inside out, and gives a programmer-friendly way of building Web 
output without inventing an pseudo-HTML minilanguage to struggle with. A 
great approach that builds on Python's flexibility and strengths.

But the big point is that there is far more to Web application 
programming than templating. Templating is the easy part! There's domain 
modelling, business logic, scripting, persistence management, database 
access, code maintainability, RAD, developer effectiveness, etc. Python, 
most lithe of modern languages, seems an excellent fit to these more 
serious needs.

Best,
-- Graham



From tjreedy at udel.edu  Tue Apr  1 22:16:24 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 1 Apr 2003 22:16:24 -0500
Subject: ternary operator
References: <33b491f.0304011848.d87dcc6@posting.google.com>
Message-ID: <9-OdnQcW6NMyyRejXTWc3w@comcast.com>

"daniel w. moore" <dan65536 at hotmail.com> wrote in message
news:33b491f.0304011848.d87dcc6 at posting.google.com...
> anyone heard about the results ofPEP 308 voting?

The results were posted.  I believe status quo more or less won,
depending on interpretation.

TJR





From graham__fawcett at hotmail.com  Thu Apr  3 12:08:26 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 3 Apr 2003 09:08:26 -0800
Subject: web server in python
References: <b6gnau$d7q$1@news.freedom2surf.net>
Message-ID: <e9570f37.0304030908.2f9c6afd@posting.google.com>

"Tom Saunders" <toms at f2s.com> wrote in message news:<b6gnau$d7q$1 at news.freedom2surf.net>...
> i am curerntly trying to make simple dynamic site in python which i intend
> to expand on. i can make the html using python but i have a few questions.
> using the BaseHTTPServer class, is it then possible to write all the code
> for a site inside one python program, which then serves based on requests
> from a website?
> ive been having trouble finding information about this and i would be
> greatfull if someone could give me an example of how to start.
> Hope you can help
> 
> Tom

It's possible to write a BaseHTTPRequestHandler subclass that will let
you "write all the code for a site in one python program". You can
write a do_GET() method that can look at the request URI and, based on
what was requested, dispatch the request to a specific handler. So if
you ask for the page 'http://localhost/hello', your do_GET method
might call your do_hello() method to fulfill the request. Your do_POST
method could dispatch your form-submission requests in a similar
fashion.

But you will end up writing a lot of support code that others have
already provided for you (for free).

There's an excellent list of Python Web Programming tools at:

http://www.python.org/cgi-bin/moinmoin/WebProgramming

Check out the Application Frameworks section in particular. Not all of
these tools will meet your "one file only" criterion, but many will.

-- Graham



From JReed at NaviSite.com  Thu Apr 24 11:57:20 2003
From: JReed at NaviSite.com (Reed, Judith)
Date: Thu, 24 Apr 2003 11:57:20 -0400
Subject: Problem getting zlib.o built on solaris 8 system
Message-ID: <DA8EA9A30F42A4428ACD87B38A32940E3FCF63@knight.empire.clrbl.com>

Greetings, Python folks. I am not a subscriber to this list, but am
hoping you can help me with a problem installing

python 2.2.2 on Solaris 8. I can configure, but when I do "make", I get
the following errors on the zlib extension build:

 

building 'zlib' extension

gcc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -fPIC -I. 

      -I/usr/local/src/Python-2.2.2/./Include -I/usr/local/include 

      -I/usr/local/src/Python-2.2.2/Include
-I/usr/local/src/Python-2.2.2 

      -c /usr/local/src/Python-2.2.2/Modules/zlibmodule.c 

      -o build/temp.solaris-2.8-sun4u-2.2/zlibmodule.o

gcc -shared build/temp.solaris-2.8-sun4u-2.2/zlibmodule.o 

      -L/usr/local/lib -lz -o build/lib.solaris-2.8-sun4u-2.2/zlib.so

Text relocation remains                         referenced

    against symbol                  offset      in file

<unknown>                           0x734
/usr/local/lib/libz.a(inflate.o)

<unknown>                           0x738
/usr/local/lib/libz.a(inflate.o)

                  :

<unknown>                           0xe18
/usr/local/lib/libz.a(infblock.o)

<unknown>                           0xe1c
/usr/local/lib/libz.a(infblock.o)

<unknown>                           0xe20
/usr/local/lib/libz.a(infblock.o)

                                    :

                                    etc...

 

I have zlib 1.3.1 installed, so I am baffled as to what is going on
here. Can anyone make any suggestions?
Please reply via email, as I'm not subscribed, I'll summarize if there
is interest.

 

Huge TIA!!!

 

Judith Reed

jreed at navisite.com

"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety." 

Benjamin Franklin

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030424/b0328da4/attachment.html>

From staschuk at telusplanet.net  Tue Apr  8 13:53:56 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 8 Apr 2003 11:53:56 -0600
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <080420031230508991%camel@oasis.com>; from camel@oasis.com on Tue, Apr 08, 2003 at 12:30:50PM -0400
References: <080420031230508991%camel@oasis.com>
Message-ID: <20030408115356.A407@tibia.amotlpaa.bogus>

Quoth dromedary:
> I've written a palidrome finder two ways (see below), both of which
> work, but verrrrrry slooooooooooowly. Would anybody out there care to
> look over the code and offer an idea or two about how to speed it up?
  [...]

Your approach to the problem is inherently slow, I think.  You go
through all combinations of k words from your word list, and check
each for palindromicity.  That's (n choose k) * k! combinations
checked.

A smarter search would do better.  One approach would be to build
the palindrome from both ends instead of from just one end.  For
example, suppose your code understood that a palindrome starting
with 'orange' must end with a word 'o', 'ro', 'aro', 'naro',
'gnaro', 'egnaro', or a word ending with 'egnaro'.  If there are
no such words, it can skip 1/n of the search space, namely all the
combinations which start with 'orange'.

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From donnal at donnal.net  Mon Apr 21 14:40:47 2003
From: donnal at donnal.net (Donnal Walter)
Date: Mon, 21 Apr 2003 13:40:47 -0500
Subject: wxBoxSizers in window and two embedded panels.
In-Reply-To: <14641910474.20030421211947@tyumbit.ru>
References: <14641910474.20030421211947@tyumbit.ru>
Message-ID: <b81ds6$a9f$1@main.gmane.org>

Igor Prischepoff wrote:
> Very strange. Panel is like window, so what the difference ?
> It's still a mistery for me.
> Any way - workaround solution is still the solution.

 From the docs:

"""A panel is a window on which controls are placed. It is usually 
placed within a frame. It contains minimal extra functionality over and 
above its parent class wxWindow; its main purpose is to be similar in 
appearance and functionality to a dialog, but with the flexibility of 
having any window as a parent."""

You can add wxControls to either a wxWindow or a wxPanel, but you can't 
(or at least you shouldn't) add a wxPanel to a wxWindow. Plus, when you 
add wxControls to a wxPanel, you get a tab order, but if you add them to 
a wxWindow there is no tab order.

> Thank for your help!

You are welcome.

Donnal Walter
Arkansas Children's Hospital






From adalke at mindspring.com  Wed Apr 23 12:58:18 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 23 Apr 2003 10:58:18 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
References: <mailman.1051023275.29338.python-list@python.org> <b8420j$629$1@slb3.atl.mindspring.net> <mailman.1051043856.10118.python-list@python.org>
Message-ID: <b86gim$60r$1@slb5.atl.mindspring.net>

Steven Taschuk
> __and__ could be passed other as a callable which computes and
> returns the second operand.  Then the standard __and__ could be
>     def __and__(self, other):
>         if bool(self):
>             return self
>         return other()

So a new idiom and new code (to generate a function from a set of
unevaluated code), just to allow overriding booleans?  Seems
rather excessive to me.

Gets pretty complicated with code like

print value or sys._getframe().f_lineno

since the function introduces a new frame.

OTOH, this could lead to an easy 'if/else' expression function.
And generically speaking, adding code blocks like this leads
towards Smalltalk and Ruby.

                    Andrew
                    dalke at dalkescientific.com





From adalke at mindspring.com  Tue Apr 15 00:15:34 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Mon, 14 Apr 2003 22:15:34 -0600
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl> <3E9AF8BD.BC918E5E@engcorp.com> <3E9B520F.7D6E0360@alcyone.com> <b7fltn$b7v$1@slb6.atl.mindspring.net> <3E9B6821.AACF1ADD@alcyone.com>
Message-ID: <b7g0ps$hi7$1@slb3.atl.mindspring.net>

Erik Max Francis
> There are even in some circumstances more fuel-minimizing transfers than
> Hohmann, like bielliptic transfers (which is sort of like a
> double-Hohmann transfer to a point way out in space), but they take vast
> amounts of time (e.g., a Hohmann-like transfer out to Neptune and then
> back simply to go from Earth to Mars).  Under some circumstances this
> can take less fuel than a simple Hohmann transfer (although it won't be
> energy minimizing).

I hadn't heard of a bielliptic transfer before, but then again, nearly
everything I know about this topic comes from reading Heinlein.  ;)
(Okay, and reading magazines like Nature and various articles
about space flight.)

Still, a gravitational assist can get things out faster and with less fuel
than either of those two.  But they are non-trivial to compute, and
Heinlein sure hadn't heard of them, given especially the constraint
of having to work it all out on paper and sliderule.

'Course, once Kilgore figures out his energy transformation equations[1]
and we work out the (*ahem*) "astrogation and control difficulties"[2],
we'll just get there at 1g the whole way.  <wink>

                    Andrew

[1]  Heinlein's "Farmer in the Sky" used a matter-energy conversion
power plant to boost at 1g for 4 hours, then coasted for two months.
The older atomic powered ships took "two years and nine months"
using, I assume, a Hohmann orbit.  "Your grandchildren will make
the trip in a week, blasting at one g all the way"

[2] After all, in "Have Spacesuit, Will Travel," a slipstick is
still "the best thing since girls."





From adalke at mindspring.com  Mon Apr 14 21:09:54 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Mon, 14 Apr 2003 19:09:54 -0600
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl> <3E9AF8BD.BC918E5E@engcorp.com> <3E9B520F.7D6E0360@alcyone.com>
Message-ID: <b7fltn$b7v$1@slb6.atl.mindspring.net>

Erik Max Francis:
> >>> info(stationaryEarthOrbit, highTritonOrbit)
> Source: stationary orbit around `Earth'
> Destination: medium orbit around `Triton'
 ...
>  Transfer: Hohmann from `Earth' to `Neptune'
>   Duration: 965260524.876 s
>   Burns: 11651.5022539, 4053.91436026 m/s

You wouldn't really do a Hohmann orbit all the way
out to Neptune, would you?  Aren't there any
gravitational assists that could help out?  After all,
it took Voyager 2 only 12 years or so to get there,
not 30, though the planets were well aligned at
the time.

                    Andrew
                    dalke at dalkescientific.com






From tim.one at comcast.net  Wed Apr 16 16:22:08 2003
From: tim.one at comcast.net (Tim Peters)
Date: Wed, 16 Apr 2003 16:22:08 -0400
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <b7jv2v$1vd4$1@nntp6.u.washington.edu>
Message-ID: <BIEJKCLHCIOIHAGOKOLHIEAKFGAA.tim.one@comcast.net>

[Donn Cave]
> But the idea that repr should be the function that conveys that
> "whole truth" is suspiciously consistent with the marshalling idea.

Indeed, the marshal module uses repr() to encode floats in code objects.
The idea behind __repr__ seems prettly clear in the docs:

    http://www.python.org/doc/current/ref/customization.html

    __repr__(self)
    Called by the repr() built-in function and by string conversions
    (reverse quotes) to compute the ``official'' string representation of
    an object.  If at all possible, this should look like a valid Python
    expression that could be used to recreate an object with the same
    value (given an appropriate environment).  If this is not possible, a
    string of the form "<...some useful description...>" should be returned.
    The return value must be a string object.   This is typically used for
    debugging, so it is important that the representation is information-
    rich and unambiguous.





From aahz at pythoncraft.com  Sun Apr  6 11:16:38 2003
From: aahz at pythoncraft.com (Aahz)
Date: 6 Apr 2003 11:16:38 -0400
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain>
Message-ID: <b6pgcm$1t3$1@panix3.panix.com>

In article <m2k7e7addl.fsf at localhost.localdomain>,
Patrick K. O'Brien <pobrien at orbtech.com> wrote:
>
>Sure.  The new editor, PyAlaMode, is in the PyCrust CVS repository on
>SourceForge:
>
>http://sourceforge.net/projects/pycrust
>
>While it works, I plan to revise it quite a bit.  Right now it just
>uses the standard Scintilla keybindings, and there is no way to change
>them.  In fact, there is no way to customize it at all.  But before I
>add all those kind of features I want to design the architecture and
>make it based upon and semi-compatible with Emacs.  If we that and
>make it easy to integrate with Bicycle Repair Man that should make it
>closer to a Smalltalk type environment.  But first I need to get the
>foundation in good shape before I pile on features.  I'll take all the
>help I can get.

Using "semi-compatible" confuses the issue, I think.  What you want is
"comparable in power".  Another project you might want to look at is the
Python binding for vim.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From aleax at aleax.it  Sat Apr 12 17:11:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 21:11:39 GMT
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com> <ugham-t2f.ln1@grendel.myth>
Message-ID: <fi%la.7870$T34.222571@news2.tin.it>

Jim Richardson wrote:

> On Sat, 05 Apr 2003 20:13:45 +0100,
>  Simon Faulkner <news at titanic.co.uk> wrote:
>> I have a list of about 5000 numbers in a text file - up to 14 digits
>> each.
>> 
>> I need to check for duplicates.
>> 
>> What would people suggest as a good method?
>> 
>> Simon
> 
> I'd use sort|uniq, but I don't know if that's available for MS type osen.

You can get implementations of sort and uniq for MS, but a short
Python script is better IMHO.


> In python, just stuff them all in a dictionary, any repeats, will be
> eliminated. But this is rather crude and probably slow. But it would
> work.

Anything but slow!  Python dictionaries are quite fast.  But removing
duplicates is not the same as 'checking for duplicates' -- Simon
might rather want (e.g.) a list of all numbers that WERE in fact
duplicate.  A script that plays with a Python dict is still no doubt
the right solution, but it's hard to write one without more precise
specifications regarding what is desired.


Alex




From mathias-usenet at valpo.de  Tue Apr  1 01:42:29 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 1 Apr 2003 06:42:29 GMT
Subject: converting from shell script to python
References: <kfn0jaoig1.fsf@shaoz.activesky.com.au>
Message-ID: <b6bccj$3fsej$1@ID-15245.news.dfncis.de>

Shao Zhang <shao at cia.com.au> wrote:

> #!/bin/sh
> ROOTDIR=/usr/lib
> CLASSPATH="$ROOTDIR/a/a.jar:$ROOTDIR/b/b.jar:$ROOTDIR/c/c.jar"
> java -classpath $CLASSPATH foo

How about: 

import os

ROOTDIR='/usr/lib'
CLASSPATH="%s/a/a.jar:%s/b/b.jar:%s/c/c.jar" % \
          (ROOTDIR,ROOTDIR,ROOTDIR)
os.system('java -classpath %s foo' % (CLASSPATH))

Sorry that it comes without lists/array/join etc;)

Mathias



From g.szolderits at gmx.at  Wed Apr 16 15:17:10 2003
From: g.szolderits at gmx.at (=?ISO-8859-15?Q?G=FCnter?= Szolderits)
Date: Wed, 16 Apr 2003 21:17:10 +0200
Subject: PyQT and Eric
References: <b7is7g$i1d$1@rex.ip-plus.net> <mailman.1050479366.18545.python-list@python.org> <b7k998$2c9e$1@news.imp.ch>
Message-ID: <3e9dabdf$0$53334$91cee783@newsreader02.highway.telekom.at>

Daniele Trovato wrote:

> @G?nter
>>Install PyQt sources and recompile. Fixed the problem for me
> I will try it, but I'm not an expert in recompile sources an installing
> it.

It's quite easy:

1. install the source package with YaST
2. go to /usr/src/packages/SOURCES and unpack the PyQt package
   (tar xjf PyQt-x11-gpl-3.5.tar.bz2)
3. cd into the PyQt-x11-gpl-3.5 directory
4. run "python build.py"
5. su (must be root for make)
6. run "make"
7. run "make install"

Maybe you have to install other packages as well because of dependencies
(qt-devel?).

Good luck,
G?nter



From peter at engcorp.com  Wed Apr 30 07:45:24 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 30 Apr 2003 07:45:24 -0400
Subject: Removing last line while appending to file
References: <20030430152608.3c51bc77.twegener@fastmail.fm> <20030430164323.256ee71a.twegener@fastmail.fm>
Message-ID: <3EAFB754.41CE2258@engcorp.com>

Tim Wegener wrote:
> 
> Someone emailed me directly with the solution: use open(filename, "r+")
> This solved my problem.
> (I was using the wrong mode for open())

Maybe mode "rb+" would be a little better for you, if these needs
to be used on Windows or to be portable to Windows.

-Peter



From anton at vredegoor.doge.nl  Wed Apr 23 08:02:41 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Wed, 23 Apr 2003 14:02:41 +0200
Subject: Regarding Dictionaries in python
References: <mailman.1050957134.26691.python-list@python.org>
Message-ID: <b85vgf$6q1$1@news.hccnet.nl>

"Mehta, Anish" <Anish.Mehta at enst-bretagne.fr> wrote:

> >>>D= {'a' : 1, 'b' : 2, 'c' : 3, 'd' : 4}
> >>>D.items()
>[ ('b', 2), ('c', 3), ('d', 4), ('a', 1)]
>
>
>why it is not showing the items in order in which i have inserted. 
>Please tell me why it is happing like this.

Others have already explained this phenomenon. I would like to add
that the language designers do not guarantee that dictionaries will
always return their items in the same order across implementations.

So, for example on another Python interpreter:

Python 2.2.1 (#1, Jun 25 2002, 10:55:46) 
[GCC 2.95.3-5 (cygwin special)] on cygwin
Type "help", "copyright", "credits" or "license" for more information.
>>> D= {'a' : 1, 'b' : 2, 'c' : 3, 'd' : 4}
>>> D.items()
[('a', 1), ('c', 3), ('b', 2), ('d', 4)]
>>> 

This gives the language designers enough flexibility to always provide
us with "state of the art" implementations of the dictionary type.

At the same time this makes the dictionary type the place where Python
hides most of its ambiguity.

Anton








From wilfred_s at webtv.net  Wed Apr 16 13:22:06 2003
From: wilfred_s at webtv.net (W S)
Date: Wed, 16 Apr 2003 13:22:06 -0400 (EDT)
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <15946-3E9D913E-63@storefull-2373.public.lawson.webtv.net>

I asked:

>Are the nutshell and cookbook O'Reilly Python
>books easier to read?

Your comments were helpful and I thank you all. I'll read the on-line
tutorials, get the books and keep at it.

Wilfred




From adalke at mindspring.com  Tue Apr 22 14:17:04 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Tue, 22 Apr 2003 12:17:04 -0600
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net> <nCUoa.41676$LB6.941312@news1.tin.it> <pan.2003.04.22.03.08.58.526973@mn.rr.com> <3japa.44301$LB6.997057@news1.tin.it> <roy-6ECAE1.08483422042003@reader1.panix.com>
Message-ID: <b840q9$ggs$1@slb5.atl.mindspring.net>

Roy Smith:
> I'm with Guido j^2 on this one.

Don't you mean -j**2?

  :)
                    Andrew
                    dalke at dalkescientific.com





From dklaussen at hotmail.com  Thu Apr 24 09:32:42 2003
From: dklaussen at hotmail.com (Daniel K)
Date: 24 Apr 2003 06:32:42 -0700
Subject: =?ISO-8859-1?Q?Vim_Newbie_=96_Python_Edit/Compile/Run_c?=
 =?ISO-8859-1?Q?ycle_and_integrating_py_help_questions?=
Message-ID: <a791ab19.0304240532.5a025b81@posting.google.com>

I can't seem to find how to ask Vim to "compile" the current buffer
the way I could in emacs (using c-c, c-c.)  What I'm hoping for is to
run the current buffer, show the output in a split window
(horizontal?) and move the cursor in the buffer to the first error. 
Is there a Vim way of doing things?

Is there any way to integrate some python language reference to be
accessed from within Vim?  Something like hot-key keyword lookup to
the docs.

Is asking for keyword completion in Vim going against the grain of
what Vim stands for?  (Simplicity)

Any suggestions or pointers to where I can find these answers would be
greatly appreciated.  I'm really enjoying the power of Vim!!



From SBrunning at trisystems.co.uk  Mon Apr  7 07:02:10 2003
From: SBrunning at trisystems.co.uk (Simon Brunning)
Date: Mon, 7 Apr 2003 12:02:10 +0100 
Subject: negative values in profile output
Message-ID: <31575A892FF6D1118F5800600846864D012007A1@intrepid>

> From:	Matthew M. Boedicker [SMTP:mboedick at mboedick.org]
> 
> When using the profile module to profile a long-running program,
> I am getting some negative values:
> 
> 6753628 -4036.027   -0.001 -4036.027   -0.001 gobject_sql.py:36(<lambda>)
 
Probably all down to Tim Peters. I know he'd had some success speeding
list.sort() up, but I didn't know he'd got *that* far...

Cheers,
Simon Brunning




-----------------------------------------------------------------------
The information in this email is confidential and may be legally privileged.
It is intended solely for the addressee. Access to this email by anyone else
is unauthorised. If you are not the intended recipient, any disclosure,
copying, distribution, or any action taken or omitted to be taken in
reliance on it, is prohibited and may be unlawful. TriSystems Ltd. cannot
accept liability for statements made which are clearly the senders own.




From staschuk at telusplanet.net  Wed Apr 16 17:53:16 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 15:53:16 -0600
Subject: Big-O notation
In-Reply-To: <Xns935F94F66A99Eduncanrcpcouk@127.0.0.1>; from duncan@NOSPAMrcp.co.uk on Wed, Apr 16, 2003 at 01:39:41PM +0000
References: <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <roy-92A049.08504216042003@reader1.panix.com> <Xns935F94F66A99Eduncanrcpcouk@127.0.0.1>
Message-ID: <20030416155316.F4266@tibia.amotlpaa.bogus>

Quoth Duncan Booth:
  [...]
> I believe you are making some unreasonable assumptions here. Remember that 
> if I have an algorithm that is O(N^2) that is really just a shorthand for 
> saying that it will have a running time a*N^2 + b*N * c where a, b, and c 
> are constants but for sufficiently large N only the N^2 term matters.

Not to dispute your general point, but a nit: O(n^2) doesn't imply
a polynomial expansion such as you have above; the actual runtime
could be, for example, 3n^2 + log n + 18/n.  That is, the
additional terms might be anything which (asymptotically) grows
slower than n^2.

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From skip at pobox.com  Wed Apr  2 11:00:57 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 2 Apr 2003 10:00:57 -0600
Subject: Socket timeouts for HTTP connections (httplib)?
In-Reply-To: <Pine.LNX.4.51.0304020916520.17427@ns0.tbye.com>
References: <9df3d134.0304020550.7e32bf6b@posting.google.com>
        <Pine.LNX.4.51.0304020916520.17427@ns0.tbye.com>
Message-ID: <16011.2361.330921.405719@montanaro.dyndns.org>

    Dave> In newer versions of Python sockets can have timeouts, but someone
    Dave> recently posted that they don't work for the high-level APIs.

Fixed in CVS. ;-)

Skip




From max at alcyone.com  Fri Apr 25 18:27:25 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 25 Apr 2003 15:27:25 -0700
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <3EA9B64D.3208D730@alcyone.com>

Frantisek Fuka wrote:

> You can say to me "Don't forget to always include the parenteses" but
> I'm still curious if this cannot be somehow configured, so that I get
> error when I try to just access the method pointer (".isGreen")
> instead
> of calling the method (".isGreen()"). I understand that there has to
> be
> the possibility of accessing the method pointers but I usually don't
> need it and for a beginner like me this makes the applications very
> hard
> to debug.

There might be a PyChecker option, but I wouldn't count on it, since due
to Python's dynamic nature there's no real way for the interpreter (or
PyChecker) to know that you meant to call the attribute rather than test
its value as a Boolean.  (After all, at the time that code gets run,
object.isGreen could be anything at all.)

One thing that PyChecker, at least, is more likely to catch is "code has
no effect" statements like

	sys.exit

Without the parentheses, this clearly can't do anything useful, just
like

	2 + 2

I can't recall offhand whether or not PyChecker detects this type of
logical error; I would suspect it probably does or will soon.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ To perceive is to suffer.
\__/ Aristotle
    Python chess module / http://www.alcyone.com/pyos/chess/
 A chess game adjudicator in Python.



From rmunn at pobox.com  Fri Apr 11 08:49:13 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 11 Apr 2003 12:49:13 GMT
Subject: BEGINNER: converting integer to short month name?
References: <OF47A94545.30289BFA-ON85256D03.006E4C80-85256D03.006FA823@seminole.fl.us> <mailman.1049921355.15625.python-list@python.org> <b73066$evv$1@slb9.atl.mindspring.net> <3E950823.DE77C48D@alcyone.com> <1d22f12e.0304110250.278ebb3c@posting.google.com>
Message-ID: <slrnb9cpno.dd8.rmunn@localhost.localdomain>

Graham Breed <usenet at microtonal.co.uk> wrote:
> I get
> 
> Python 2.2 (#1, Dec 31 2001, 15:21:18)
> [GCC 2.95.3-5 (cygwin special)] on cygwin
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import calendar
>>>> calendar.month_abbr[0]
> Segmentation fault (core dumped)

WHen I do that, I get the empty string ''. Note that month 0 is
undefined; if you want 'Jan', you need to do calendar.month_abbr[1].

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From theller at python.net  Thu Apr 17 10:33:12 2003
From: theller at python.net (Thomas Heller)
Date: 17 Apr 2003 16:33:12 +0200
Subject: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net> <b7mbnb$pp2$06$1@news.t-online.com>
Message-ID: <8yu9gpnb.fsf@python.net>

"Harald Schneider" <h_schneider at marketmix.com> writes:

> No permission probs, logged in as Admin. Platform is Win2K, not updated from
> NT.
> There are no exceptions. Reading a registry key just claims it would not
> exist. When I run the same code outside the service wrapper, I can read the
> key.
> 

Strange.

Do you mean *you* are logged in as Admin, or does the *service* run under
the Admin account?

Per default a service runs under the system account. Maybe the registry
keys are really not there then? HKEY_CURRENT_USER, probably?

What happens when you run the service as not-py2exe'd service (via
PythonService.exe)?

Thomas

PS: As far as i can tell, there are no differences between the 2.2 and 2.1
release of py2exe. Not yet ;-)



From dale at riverhall.NOTHANKS.co.uk  Wed Apr 16 06:46:07 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 16 Apr 2003 11:46:07 +0100
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il>
Message-ID: <d7aq9v0a5ecoue90gccfos9jtkkqcr73di@4ax.com>

"Avner Ben" <avner at skilldesign.com> wrote:

>Question:
>
>    Does standard Python have case-insensitive string comparison and search?
>
>        Avner.
>

This might help. 

This is a class you use in place of a string and it emulates a string
in all practical ways except that comparisons and lookups are case
insensitive. All uses of the string for assignments, however, yield
the original case.

We've found this very handy when using strings as dictionary indices
or when searching for Windows filenames in lists.

Some examples:
>>> poem = iStr("Mary had a little LAMB")
>>> poem == "mary had a little lamb"
1
>>> poem.find("LITTLE")
11
>>> poem.title()
'Mary Had A Little Lamb'
>>> poem
iStr('Mary had a little LAMB')
>>> 

Note that the hash value is the lower case value. If you use an iStr
as a dictionary index, you will need to make sure that lookups are
lower case too:

>>> filelist = {iStr('WINNT'): 0, iStr('Temp'): 0}
>>> 'winnt' in filelist
1
>>> 'WINNT' in filelist
0
>>> 'temp' in filelist
1
>>> 

The replace() method of the str class hasn't been tackled. If you use
replace on an iStr, it will operate as on a normal string.

I've submitted this to the Python Cookbook.


class iStr(str):
    """Case insensitive strings class.
    Performs like str except comparisons are case insensitive."""

    def __init__(self, strMe):
        str.__init__(self, strMe)
        self.__lowerCaseMe = strMe.lower()

    def __repr__(self):
        return "iStr(%s)" % str.__repr__(self)

    def __eq__(self, other):
        return self.__lowerCaseMe == other.lower()

    def __lt__(self, other):
        return self.__lowerCaseMe < other.lower()

    def __le__(self, other):
        return self.__lowerCaseMe <= other.lower()

    def __gt__(self, other):
        return self.__lowerCaseMe > other.lower()

    def __ne__(self, other):
        return self.__lowerCaseMe != other.lower()

    def __ge__(self, other):
        return self.__lowerCaseMe >= other.lower()

    def __cmp__(self, other):
        return cmp(self.__lowerCaseMe, other.lower())

    def __hash__(self):
        return hash(self.__lowerCaseMe)

    def __contains__(self, other):
        return other.lower() in self.__lowerCaseMe

    def count(self, other, *args):
        return str.count(self.__lowerCaseMe, other.lower(), *args)

    def endswith(self, other, *args):
        return str.endswith(self.__lowerCaseMe, other.lower(), *args)

    def find(self, other, *args):
        return str.find(self.__lowerCaseMe, other.lower(), *args)
    
    def index(self, other, *args):
        return str.index(self.__lowerCaseMe, other.lower(), *args)

    def rfind(self, other, *args):
        return str.rfind(self.__lowerCaseMe, other.lower(), *args)

    def rindex(self, other, *args):
        return str.rindex(self.__lowerCaseMe, other.lower(), *args)

    def startswith(self, other, *args):
        return str.startswith(self.__lowerCaseMe, other.lower(),
*args)

    
--
Dale Strickland-Clark
Riverhall Systems Ltd



From a.schmolck at gmx.net  Thu Apr 10 09:39:47 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 10 Apr 2003 14:39:47 +0100
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <7xwui295xq.fsf@ruckus.brouhaha.com>
Message-ID: <yfsel4ah3oc.fsf@black132.ex.ac.uk>

Paul Rubin <http://phr.cx at NOSPAM.invalid> writes:

> Alex Martelli <aleaxit at yahoo.com> writes:
> > __slots__ is, essentially, an optimization (in terms of memory
> > consumption) intended for classes that are likely to be created 
> > in very large numbers: you give up the normal flexibility of
> > being able to add arbitrary attributes to an instance, and what
> > you gain in return is that you save one dictionary per instance
> > (dozens and dozens of bytes per instance -- that will, of course,
> > not be all that relevant unless the number of instances that are
> > around is at the very least in the tens of thousands).
> 
> I generally use __slots__ as a way to "declare" instance variables,
> so I get an exception if I accidentally assign to an undeclared one.
> I'd hope that was one of __slots__'s intended uses.

I'd recommend against using __slots__ (unless really needed for optimization),
because it cripples the usefulness and generality of your classes to your and
also *to other people's* code (weakrefs, pickling, reflection -- all
compromised). The first two you could fix (but are unlikely to, unless you
*yourself* need to use weakrefs or pickling of instances of those classes),
the last one you can't.

The fact that someone's code using __slots__ makes other peoples code less
useful (because things that work fine for instances without slots suddenly
screw up) makes me think introducing __slots__ was a rather bad idea to start
with.


'as



From mail at manuelmgarcia.com  Sat Apr 19 19:12:42 2003
From: mail at manuelmgarcia.com (Manuel M Garcia)
Date: Sat, 19 Apr 2003 23:12:42 GMT
Subject: Are reference cycles bad form?
References: <3o11avop6vkh3ake8dgbtst93t1or771ct@4ax.com> <mailman.1050772931.15672.python-list@python.org>
Message-ID: <5sl3avsia0v41d361c6se63at9hd59po2b@4ax.com>

On Sat, 19 Apr 2003 11:16:10 -0600, Steven Taschuk
<staschuk at telusplanet.net> wrote:

>Imho this is an engineering question, not a style question.  It's
>a trade-off between the benefits of the algorithms you can write
>when the data structure has reference cycles and the slightly
>inferior garbage collection properties of such structures.  The
>right trade-off to make depends on the circumstances.

In my current case, it feels more like a style question, instead of an
engineering question.  My typical Python program at work is swamped by
accessing a database over the network, so even if my algorithms ran
infinitely fast, it would only take off a small percentage of the
whole runtime.  And since I am working in Windows, my memory usage is
a small fraction of what Windows requires to display my desktop
wallpaper, 32-bit color icons, anti-aliased fonts, etc.

If I am coding in C, reference cycles don't bother me, because I have
to clean up everything anyway.  But when I am using a language that
does memory management for me, I am prejudiced against creating any
reference cycles.  Probably irrationally prejudiced, undoubtedly
because of the years I coded in Visual Basic and older versions of
Python, which could not reclaim memory in a reference cycle.

Happily, Python code is easy to refactor, so I am more likely in
Python to experiment with different data structures and algorithms,
and pick the one that feels 'cleanest'.

Manuel



From ruediger.maehl at web.de  Tue Apr  1 02:40:55 2003
From: ruediger.maehl at web.de (Rüdiger Mähl)
Date: 1 Apr 2003 07:40:55 GMT
Subject: Breaking out of a while loop with a key press?
References: <3e885ad5@mail.hmgcc.gov.uk>
Message-ID: <Xns9350627F44471ruedigermaehlwebde@193.101.67.2>

"Richard" <richardd at hmgcc.gov.uk> wrote:

> Hi,
> 
> Can anyone suggest a simple way of detecting a key press and existing
> the program? I have a program which at the moment uses a while True:
> statement. However this is not ideal as I have to close down the
> console window to exist the program. Can anyone tell me the proper way
> of doing this? So that a CTRL-C, for example, can quit the program
> correctly? 

Hi Richard, 

perhaps, you are looking for a simple approach like 
the following sample.

HTH, Ruediger

# ========================================
import sys
import time
import atexit
import signal

CONTROL_C = False

def program_exit():
    # You may do some clean-up here, but you don't have to.
    print "Exiting application"


def ctrlCHandler(*whatever):
    # Just sets the value of CONTROL_C
    global CONTROL_C
    CONTROL_C = True
    print
    print "Interrupt caught! Please wait..."
    print
    # You must check CONTROL_C in your program

# call this procedure, if control-c is pressed.
signal.signal(signal.SIGINT, ctrlCHandler)

# program_exit is called, when sys.exit is executed.
atexit.register(program_exit)

while 1:
    print "Waiting for Ctrl-C to be pressed..."
    time.sleep(2)
    if CONTROL_C: sys.exit(0)
# ========================================
    



From max at alcyone.com  Tue Apr 15 22:07:58 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 15 Apr 2003 19:07:58 -0700
Subject: Big-O notation (was: Re: Python and Schools)
References: <qGqla.61$Tj7.91655@news.uswest.net>
		 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
		 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
		 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
		 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
		 <3E9C6C62.8010808@rogers.com> <mailman.1050441947.6447.python-list@python.org>
Message-ID: <3E9CBAFE.8016BFB2@alcyone.com>

Chad Netzer wrote:

> Also, one further class that is import to know is O( 2**N ), such as
> the
> brute-force solution to the "travelling salesman problem".  These are
> much, MUCH worse than O( N**2 ) algorithms, and even today can rarely
> be
> computed for problems sets much larger than 100 (whereas O( N**2 )
> problem sets can grow to thousands or perhaps even millions and still
> be
> practically solved).  Rather than computing "solutions" with these
> types
> of algorithms, it is required to instead come up with "close-enough"
> or
> "best guess" type algorithms that CAN be computed practically on large
> data sets.

Note also (I'm sure you know this, Chad, I'm just elaborating for the
general benefit of everyone reading), since big-O notation represents
the behavior of algorithm complexity as the number of elements tends
toward infinity, it means that only the "biggest" function counts. 
e.g., something that might scale as n (n + 1) = n^2 + n -- for instance,
having a collection of n elements and selecting all pairs i, j where i
!= j -- only is written as O(n^2), since as n -> oo, n^2 grows much
faster than n.

It also means that one doesn't bother distinction between logarithms
base 10, e, 2, etc.; O(ln n) ~ O(lg n) ~ O(lb n).  Similarly, O(2^n) ~
O(10^n) ~ O(e^n) too.

There's also amortized notation, where on average you'd expect a certain
type of behavior but occasionally it will be another.  For instance,
adding an element to the end of a dynamically resizing vector (like a
list in Python) is amortized O(1), since typically the operation will
take constant time, but occasionally (like when you're at the end of the
vector's capacity and it needs to be resized) it will take O(n). 
Amortized is when you expected it to take O(f(n)), but only about every
n operationss, so one calls it "amortized O(f(n)/n)"; in this case, the
worst case is O(n), which should only happen every n operations, so you
speak of "amortized O(n)" (n/n = 1).

As Roy Smith mentioned, this is not merely of academic interest,
although formally defining it takes knowledge of calculus.  It is
vitally important that you know the complexity of the algorithms you're
using and how they'll interrelate to avoid scalability problems.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ I want a martini that could be declared a disaster area.
\__/ Capt. Benjamin "Hawkeye" Pierce
    Crank Dot Net / http://www.crank.net/
 Cranks, crackpots, kooks, & loons on the Net.



From pinard at iro.umontreal.ca  Mon Apr 28 13:44:12 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 28 Apr 2003 13:44:12 -0400
Subject: Pyrex Hard Case
In-Reply-To: <20030428171412.ZMGU12530.mta2@mta2.real.mail.entelchile.net>
References: <20030428171412.ZMGU12530.mta2@mta2.real.mail.entelchile.net>
Message-ID: <oqfzo2y0sz.fsf@alcyon.progiciels-bpi.ca>

> > I'm not fully sure I'm replying to your question, but this might work:
> > 
> > ---------------------------------------------------------------------->
> > cdef extern from *:
> >     MACRO_NAME(EXPECTED_ARGUMENTS)
> > ----------------------------------------------------------------------<

[rdrb at 123.cl]

> I had already done that.
> extern MACRO_NAME(EXPECTED_ARGUMENTS)

This is not the same thing.  Your macro declaration, in Pyrex, really
ought to be within the `cdef extern from *:' and indented within it.
You do not need to repeat `extern', yet I presume it does not hurt either.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From oren-py-l at hishome.net  Wed Apr 16 04:15:29 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Wed, 16 Apr 2003 04:15:29 -0400
Subject: Post-OOP: not just an "unfortunate acronym"?
In-Reply-To: <pan.2003.04.15.21.27.33.796814@mn.rr.com>
References: <pan.2003.04.15.21.27.33.796814@mn.rr.com>
Message-ID: <20030416081529.GA27879@hishome.net>

On Tue, Apr 15, 2003 at 09:26:16PM +0000, Anna wrote:
> The 1960s and '70s saw a crusade to root out "spaghetti code"
> and replace it with "structured programming." Since the 1980s software
> development has been dominated by a methodology known as object-oriented
> programming, or OOP. 

The biggest innovation in OOP is that the kind of people that used to
write spaghetti code now design spaghetti class hierarchies.

> Now there are signs that OOP may be running out of
> oomph, and discontented programmers are once again casting about for the
> next big idea. 

And the reason these programmers feel their discontent has anything to do 
with the methodology they are using is... ?  

    Oren




From LogiplexSoftware at earthlink.net  Tue Apr  8 11:37:45 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 08 Apr 2003 08:37:45 -0700
Subject: wxPython update problem
In-Reply-To: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no>
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no>
Message-ID: <1049816265.3724.51.camel@software1.logiplex.internal>

On Mon, 2003-04-07 at 03:05, Nils Grimsmo wrote:
> hi!
> 
> i have a problem getting wxPython to update the gui while the thread is
> running user procedures. i have tried calling wxWindow.Refresh() and
> wxWindow.Update() in various orders both on the wxFrame and the wxButton, 
> but the gui doesn't update.

There are a couple of examples of updating the GUI from a thread in the
wxPython demo.  Did you look at them?   There are two approaches: 1) use
a thread and send update events to the GUI (see throbber in the demo) or
2) use a wxTimer (see the demo of the same name).

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From roy at panix.com  Fri Apr 18 15:04:07 2003
From: roy at panix.com (Roy Smith)
Date: Fri, 18 Apr 2003 15:04:07 -0400
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net> <roy-B8E873.12412018042003@reader1.panix.com> <mailman.1050686436.32569.python-list@python.org>
Message-ID: <roy-4474E4.15040718042003@reader1.panix.com>

Steven Taschuk <staschuk at telusplanet.net> wrote:
> class c:
>     def __setattr__(self, attribute, value):
>         setattr(self.__class__, attribute, value)
> 
> Did you have something cleverer in mind?

That was the general idea, yes.



From gk_2345 at yahoo.com  Sat Apr 26 10:43:52 2003
From: gk_2345 at yahoo.com (George Kinney)
Date: Sat, 26 Apr 2003 10:43:52 -0400
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.29.48.204375@soul.net> <RZlqa.4182$lf3.712485@cletus.bright.net> <b9tqa.9443$K35.308314@news2.tin.it>
Message-ID: <eZwqa.4237$lf3.720329@cletus.bright.net>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:b9tqa.9443$K35.308314 at news2.tin.it...
> Yes, but - I suggest you AVOID using names of builtin types as names
> for your own variables; it's a habit that will eventually "bite" you,
 Oops. :)





From Padraig at Linux.ie  Wed Apr  9 06:54:27 2003
From: Padraig at Linux.ie (Padraig at Linux.ie)
Date: Wed, 09 Apr 2003 11:54:27 +0100
Subject: Visio like library for Python?
In-Reply-To: <3e93e50f.676072541@news.online.de>
References: <3e93e50f.676072541@news.online.de>
Message-ID: <3E93FBE3.2050301@Linux.ie>

Johannes Eble wrote:
> Hello community,
> 
> is there a library available for Python that behaves similar to MS
> Visio? For example, I would like to create vector graphic objects and
> connect the with a line. If I move, flip or rotate a connected object,
> the line should follow the connection.
> 
> I would like to draw UML diagrams.

The method used by http://gaphor.sourceforge.net/

Padraig.




From neal at metaslash.com  Sat Apr  5 08:23:50 2003
From: neal at metaslash.com (Neal Norwitz)
Date: Sat, 05 Apr 2003 08:23:50 -0500
Subject: segfault with imp.load_module()
References: <mailman.1049534708.5967.python-list@python.org>
Message-ID: <pan.2003.04.05.13.23.43.694594.12686@metaslash.com>

On Sat, 05 Apr 2003 04:22:51 -0500, Rob Brown-Bayliss wrote:


> Hi, I can do this in interactive python:
> 
>>>> f,n,d = imp.find_module("oda",["./items/"]) print f,n,d
> <open file './items/oda.py', mode 'r' at 0x82c6db8> ./items/oda.py
> ('.py', 'r', 1)
>>>> it = imp.load_module("oda",f,n,d)
>>>> print it
> <module 'oda' from './items/oda.pyc'>
> 
> but this code:
> file, filename, description = imp.find_module(item, ["./items/"]) 
> print file, filename, description
> item_module = imp.load_module(item,file, filename, description)
> file.close()
> 
> gets this:
> 
> [rob at musicbox BHH-V]$ python utils.py 
> <open file './items/oda.py', mode> 'r' at 0x81c2af8> ./items/oda.py 
> ('.py', 'r', 1)
> Segmentation fault

What version of python, what platform, what compiler, what options?
Assuming item = 'oda' this code works for me.

Neal



From no at spam.invalid  Wed Apr 30 13:46:22 2003
From: no at spam.invalid (Russell E. Owen)
Date: Wed, 30 Apr 2003 10:46:22 -0700
Subject: Python 2.3b1: RuntimeError using rexec
References: <b8mh4l$2oo0$1@nntp6.u.washington.edu> <m31xzl6pac.fsf@mira.informatik.hu-berlin.de> <45tr+kKkXstQ092yn@gnosis.cx> <mailman.1051648246.19116.python-list@python.org>
Message-ID: <b8p25f$20ou$1@nntp6.u.washington.edu>

Clever idea!

So far I have been using the standard repr output, and in any case it is 
nice to have everything on one line. But then trying to sort separators 
from strings containing separators becomes an issue. The new csv module 
would probably make this easy. However, I also want to keep the code 
compatible with 2.2.2...

By the way, I ran into a problem with the proposed safe eval: it turns 
True and False started getting into my prefs files (as a side effect of 
the recent change to Tkinter of returning objects instead of strings), 
and the safe eval-with-no-builtins can't recognize those values. I'm 
sure one could pass a minimal builtins dict that just included True, 
False and perhaps None, and I could certainly convert True and False to 
1 and 0 before writing the file, but...

Fortunately plain eval will work just fine for my application. The 
security holes are irrelevant -- a malicious user could find much easier 
ways to cause damage. It has always been intended that users can execute 
arbitrary Python code (e.g. write their own scripts).

It is a interesting how seemingly small changes can cause problems. 
However, I really like most of the changes, so I'm happy to accommodate.

-- Russell

In article <mailman.1051648246.19116.python-list at python.org>,
 Lulu of the Lotus-Eaters <mertz at gnosis.cx> wrote:

>|"Russell E. Owen" <no at spam.invalid> writes:
>|> I stumbled across this when testing some existing code. This code uses
>|> rexec to convert a string representation of a dictionary to a dictionary
>|> as part of reading in font preferences.
>|> Any suggestion on how to do that without rexec would also be
>|> appreciated. (I was quite disappointed that "dict" didn't do it!)
>
>I'm not sure exactly what string representation of a dictionary you use.
>But it's not too hard to use dict().  E.g.:
>
>  >>> d = '''
>  ... a:b
>  ... c:d
>  ... e:f
>  ... '''
>  >>> dict([x.split(':') for x in d.split()])
>  {'a': 'b', 'c': 'd', 'e': 'f'}
>...



From duncan at NOSPAMrcp.co.uk  Tue Apr 15 07:47:18 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 15 Apr 2003 11:47:18 +0000 (UTC)
Subject: hex -> 16bit signed int (newbie)
References: <TTRma.1502$KH1.321574@news20.bellglobal.com>
Message-ID: <Xns935E81FC55CACduncanrcpcouk@127.0.0.1>

"zif" <zf_acs at DELETETHIShotmail.com> wrote in 
news:TTRma.1502$KH1.321574 at news20.bellglobal.com:

> How can I get 16bit signed int from a hex string?
> Eval('0x...') gives me 16bit unsigned,
> struct.pack('i', '\x..\x..') works fine, but
> Python doesn't allow me to concatenate strings
> with '\' (str(92) gives me '\\', which doesn't
> work as well).
> 
How about this?

>>> def signedFromHex16(s):
	v = int(s, 16)
	if not 0 <= v < 65536:
		raise ValueError, "hex number outside 16 bit range"
	if v >= 32768:
		v = v - 65536
	return v

>>> signedFromHex16("ffff")
-1
>>> signedFromHex16("8000")
-32768
>>> signedFromHex16("7fff")
32767
>>> 

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From elabuschagne at gmsonline.co.za  Tue Apr  8 09:16:45 2003
From: elabuschagne at gmsonline.co.za (Etienne Labuschagne)
Date: Tue, 08 Apr 2003 15:16:45 +0200
Subject: SystemError in weakrefobject.c
Message-ID: <5.2.0.9.0.20030408150555.030de210@esrimail>

Hi all,

I have a block of code within a try except that does message parsing 
work.  This block of code is never supposed to fail, it must log the error, 
reject the message which caused the parsing to fail and continue.  So, I 
catch all exceptions, log the error and continue normally with the rest of 
the messages.

I am however getting SystemErrors which seems to go past the except and 
then crashes the thread.  In fact, it not only crashes the thread, but the 
whole program.

Looking at the exception, it seems something more serious is happening and 
I can't figure out what is wrong.  Here are two of the exceptions which I 
have logged so far:

Traceback (most recent call last):
File "C:\!Projects\Python\GlobalTrack\MessageDecoders.py", line 743, in decode
'\n'.join(['%s = %s' %(k,dict[k]) for k in dict.keys()])))
SystemError: 
D:\trentm\as\Apps\ActivePython-2_2\src\Core\Objects\weakrefobject.c:690: 
bad argument to internal function

and

Traceback (most recent call last):
File "C:\!Projects\Python\GlobalTrack\MessageDecoders.py", line 80, in __call__
return self.decode()
File "C:\!Projects\Python\GlobalTrack\MessageDecoders.py", line 344, in decode
cannedId = int(self._destinationAddress)
SystemError: 
D:\trentm\as\Apps\ActivePython-2_2\src\Core\Objects\weakrefobject.c:690: 
bad argument to internal function

Any help would be greatly appreciated.

Thanks
Etienne





From jmeyer at pdi.com  Thu Apr 10 11:14:18 2003
From: jmeyer at pdi.com (Jim Meyer)
Date: 10 Apr 2003 08:14:18 -0700
Subject: __slots_ and inheritance
In-Reply-To: <yfsel4ah3oc.fsf@black132.ex.ac.uk>
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com>
	 <7xwui295xq.fsf@ruckus.brouhaha.com>  <yfsel4ah3oc.fsf@black132.ex.ac.uk>
Message-ID: <1049987657.19661.15.camel@utonium.pdi.com>

On Thu, 2003-04-10 at 06:39, Alexander Schmolck wrote:
> I'd recommend against using __slots__ (unless really needed for optimization),
> because it cripples the usefulness and generality of your classes to your and
> also *to other people's* code (weakrefs, pickling, reflection -- all
> compromised). The first two you could fix (but are unlikely to, unless you
> *yourself* need to use weakrefs or pickling of instances of those classes),
> the last one you can't.

I'm feeling very slow this morning; could you give a few explicit
examples of how __slots__ interferes with weakrefs, pickling, and
reflection?

Thanks in advance,

--j, always looking to learn.
-- 
Jim Meyer, Geek at Large                                  jmeyer at pdi.com





From christopher.saunter at durham.ac.uk  Sat Apr 26 12:35:40 2003
From: christopher.saunter at durham.ac.uk (Christopher Saunter)
Date: Sat, 26 Apr 2003 16:35:40 +0000 (UTC)
Subject: GUI testing with python  (SendKeys)
References: <b8c6c9$p2m$0@198.102.102.163>
Message-ID: <b8ecgs$47i$1@sirius.dur.ac.uk>

Hi,
  You might want to check out Girder - http://www.girder.nl
It offers more robust and flexible ways of remote controlling Win32 GUI 
stuff than sendkeys.  There is a very noddy Python client for talking to 
Girder at http://www.dur.ac.uk/christopher.saunter/soft.html

HTH
cds

pj (pjdurai at hotmail.com) wrote:
: Greetings

: Is there a python module similar to Perl's Win32::GuiTest?
: Basically something that would implement SendKeys and friends on Win32 ?

: Before I went ahead and rolled my own, I just wanted to check.
: Web searches didnt turn up anything useful.

: cheers
: pj






From peter at engcorp.com  Thu Apr 24 11:39:32 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 24 Apr 2003 11:39:32 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it>
Message-ID: <3EA80534.4B3FFA43@engcorp.com>

Alex Martelli wrote:
> 
> Peter Hansen wrote:
> 
> > Aahz wrote:
> >>
> >> Which brings up the question: do the unit-testers here write tests for
> >> throwaway scripts?
>     ...
> > Basically, with a literally throwaway script, you do lose much of
> > the business value of the tests, so as far as I can see you write
> > them only if they truly help you write the code properly, which
> > should once you're proficient at TDD and then only sometimes, since
> > you're a very experienced programmer.  Otherwise, and certainly in some
> > cases no matter who you are, they really would be a waste of time.
> > IMHO.  Except maybe to help you develop the habits.   _Four_ reasons...
> 
> Oh come on -- we all know that some "throw-away" scripts happen
> to NOT get thrown away, but rather lie around and get reused and
> tweaked and grown and...  It's hard to quantify HOW MANY such
> "throw-away" scripts are in fact thrown away, and how many end
> up getting kept (and reused and grown) instead -- guess it depends
> on how good one is at planning / forecasting -- but as for me,
> I would paraphrase Brooks' MMM Chapter 11's famous title:
> 
>         Plan NOT to throw one away -- you won't, anyway.

I could have said that, but I know Aahz knows when something is really
throwaway, and there are such things, so I thought I'd try answering
the question directly for a change.  ;-)  And I did emphasize that this 
was for _literally_ throwaway scripts, which do exist.  If nothing else,
you could call it a spike solution which is expected to be used, once
only, then tossed.  If it turns out it's got to be used again, then it 
should be re-written anyway so it's not a serious problem that it lacks 
tests.

For me, if the script has any identifiable subroutines, let's say as
an example a routine which produces a filtered list of files with a
particular extension and which have been written in the last day, then
I would drive that with tests, even for a throwaway.  As Timothy Grant
pointed out, you generally want to know that even your throwaway scripts
work properly.  I'm just less anal about testing the ones I think are
throwaway than the ones I "know" are not.  If I'm wrong... no big deal:
I rewrite as a real script with proper tests.  Since the script was 
only twenty or thirty lines of code (any more and it could not possibly
be throwaway, right?) it isn't really a big deal.

-Peter



From alloydflanagan at attbi.com  Thu Apr  3 09:30:54 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 3 Apr 2003 06:30:54 -0800
Subject: PEP 313 - Roman numeral literals
References: <mailman.1049227024.16768.python-list@python.org>
Message-ID: <a8b7f07a.0304030630.37733ff0@posting.google.com>

"Batista, Facundo" <FBatista at uniFON.com.ar> wrote in message news:<mailman.1049227024.16768.python-list at python.org>...
> #- 
...
> But if we're thinking in a module, I'll like to have an object called Roman
> (or so) where I could make something like this:
> 
> 	myRoman1 = Roman()
> 	myRoman1.fromInt(15)
> 
> 	myRoman2 = Roman()
> 	myRoman2.fromString("X")
> 

FWIW, I would instead suggest Roman.from(), where the function could
handle either, such as:
def from(self, some_roman):
     try:
          self.__intval = int(some_roman)
     except ValueError:
          self.__intval = self.__stringToRoman(some_roman)
     return

So who's doing the prototype implementation? :)



From ramen at lackingtalent.com  Thu Apr 24 04:16:16 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Thu, 24 Apr 2003 08:16:16 -0000
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com>
Message-ID: <slrnbaf7bi.gb9.ramen@lackingtalent.com>

In article <3E9FFCB5.79AEA22E at engcorp.com>, Peter Hansen wrote:
> Aahz wrote:
>> 
>> In article <mailman.1050587068.2733.python-list at python.org>,
>> Jp Calderone  <exarkun at intarweb.us> wrote:
>> >
>> >--
>> >It is practically impossible to teach good programming style to
>> >students that have had prior exposure to BASIC: as potential
>> >programmers they are mentally mutilated beyond hope of
>> >regeneration.        -- Dijkstra
>> 
>> I like to believe that I'm a counter-example to this generalization.
> 
> I too learned BASIC first (from a book which I found in the public 
> library, a book with lots of pictures and big fonts and stuff... 
> I forget what it was called though :-( ).  I also hope my code is more 
> an example of good style than of bad.
> 
> Nevertheless, we might give Dijkstra the benefit of the doubt and
> accept that a *student* who had exposure to BASIC might be 
> unteachable as far as improving her style goes.  Perhaps only once
> she advances beyond the student stage, and has some years of actual
> _practice_, will she discard the early mental mutilation of BASIC
> and find her way to Good Style.
> 
> (Though I doubt Dijkstra intended to allow this loophole in his
> potential "law". :-)

Well, I think Dijkstra's basic point was that you can't really "unlearn"
anything. You can ignore what you've learned, but it's still there. Once
you've learned some BASIC, it's a part of your brain, and that imprint will
remain there FOR THE REST OF YOUR LIFE. But it's okay, it hasn't stopped me
either. Maybe someday we'll start a therapy group.

Dave



From gerald at vamphq.com  Wed Apr  9 14:05:44 2003
From: gerald at vamphq.com (Gerald Bauer)
Date: 9 Apr 2003 11:05:44 -0700
Subject: GUI Toolkits(Python and XUL)
References: <mailman.1047143834.15642.python-list@python.org> <3E6A40CB.6010301@rsballard.com> <mailman.1047151994.28453.python-list@python.org> <abf9d5f7.0303110826.6e73b917@posting.google.com> <mailman.1047401240.31100.python-list@python.org>
Message-ID: <c4a49f85.0304091005.533ae82@posting.google.com>

I know it's a little late for posting, but here I go anyway:

You might wonna check out Luxor - a free GPL'ed open-source project -
that tries to bring XUL and Python to the masses without using the
Mozilla Gecko machinery. In a way it's a free-standing, clean-room
implementation of XUL built using Java. It currently is available in a
Swing UI toolkit flavor and an Eclipse UI toolkit flavor (aka Standard
Widget Toolkit/SWT) is in the works.

Find out more @ http://luxor-xul.sourceforge.net

To get started using XUL and Python grab the luxor-python-examples
package in the download section.

  - Gerald



From Scott.Daniels at Acm.Org  Thu Apr  3 23:22:23 2003
From: Scott.Daniels at Acm.Org (Scott David Daniels)
Date: Thu, 03 Apr 2003 20:22:23 -0800
Subject: Finding field widths of floats for neat printing
In-Reply-To: <2830c89c.0304030645.51ad149c@posting.google.com>
References: <2830c89c.0304030645.51ad149c@posting.google.com>
Message-ID: <3e8d0881$1@nntp0.pdx.net>

Andrew Gregory wrote:

 > I want to write lists of floating point numbers to text files in neat
 > columns with decimal points aligned, right-hand edges straight (rather
 > than ragged), and without unnecessary trailing zeros, e.g.
 > 1.1    -34.67     0.006
 > 2.0     -1.05     0.010
 > etc.
 >
 > In order to do this I need to loop through the lists of numbers for
 > each column to find the field sizes m and n need to achieve this. I
 > can then produce a format specifier string (fss) for each column,
 > e.g. fss = "%" + "%i.%if" % (m,n)
 >
 > Given a number, -34.67 for example is there a SMART Python way of
 > determining the field sizes? Obviously I need to work at less than the
 > precision of the machine, and ignore the fact that -34.67 might be
 > represented as
 > -34.6699999999999
 >

How about something involving:
     import os.path

     def reverse(str):
         l = list(str)
         l.reverse()
         return ''.join(l)

     numbers = [1.1, 11.77, 23, 107.333]

     final = 12 - len(os.path.commonprefix(['0'*12]+
              [reverse('%.12f' % x)
               for x in numbers])

-Scott David Daniels
Scott.Daniels at Acm.Org




From mwh at python.net  Mon Apr  7 06:45:58 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 7 Apr 2003 10:45:58 GMT
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <7h3llymio0q.fsf@pc150.maths.bris.ac.uk>

Ben Finney <bignose-hates-spam at and-zip-does-too.com.au> writes:

> On Sun, 06 Apr 2003, Erik Max Francis wrote:
> >  Ben Finney wrote:
> > > In ork.py I want:
> > >     import eek
> >  
> > If the toplevel package is put in a place where Python can find it,
> >  	import monty.foo.eek
> 
> It's not.  Even if I make it so on my computer, it's not going to be so
> on someone else's computer when they download my package tree.

It is a Fact Of Life that Python packages (and modules therein) need
to know where they are in the global module namespace.

Cheers,
M.

-- 
  Programming languages should be designed not by piling feature on
  top of feature, but by removing the weaknesses and restrictions
  that make the additional features appear necessary.
               -- Revised(5) Report on the Algorithmic Language Scheme



From antonmuhin at sendmail.ru  Mon Apr 14 12:16:35 2003
From: antonmuhin at sendmail.ru (Anton Muhin)
Date: Mon, 14 Apr 2003 20:16:35 +0400
Subject: Displaying Unicode on the console (Windows)
In-Reply-To: <182bcf76.0304140744.1b3d43ee@posting.google.com>
References: <182bcf76.0304140744.1b3d43ee@posting.google.com>
Message-ID: <b7emt5$69h$1@news.peterlink.ru>

Paul Moore wrote:
>>>>print u'\u20a0'.encode("cp1258")
> [BTW, in case it's relevant - I know it probably is - the output from
> "chcp" at the console prompt is "Active code page: 1252"]
Did you try:
	print u'\u20a0'.encode("cp1252") # Note 1252, not 1258
?

hth, anton.




From nirina at mail.blueline.mg  Tue Apr  8 17:21:59 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Wed, 9 Apr 2003 01:21:59 +0400
Subject: packaging Python files
References: <pan.2003.04.07.14.44.41.50125.1629@presidency.com>
Message-ID: <019501c2fe2c$c96a53c0$f7a2383e@raseliar>

"Rajarshi Guha" <rajarshi at presidency.com> wrote:
>I was wondering whether there was any framework to create a
> single Python file that would contain all the required files of the
> project so that a user would simply run the file. Is this a good idea?
> Has it been implemented?

Distutils , i think. Describe your program in a file 'setup.py', run

python setup.py sdist

to create the package. Users have just to unpack it and run

python setup.py install

to ... install

Is that suit your needs?

nirinA







From m.faassen at vet.uu.nl  Tue Apr  1 18:03:31 2003
From: m.faassen at vet.uu.nl (Martijn Faassen)
Date: 1 Apr 2003 23:03:31 GMT
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de>
Message-ID: <b6d5s3$lvo$2@newshost.accu.uu.nl>

Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote:
> I looked at zope3 but it seems very bloated.

It isn't done yet, obviously. How does it seem bloated? Too much
framework for your tastes?

Regards,

Martijn
-- 
History of the 20th Century: WW1, WW2, WW3?
No, WWW -- Could we be going in the right direction?



From luc at nospam.com  Sat Apr 12 16:43:18 2003
From: luc at nospam.com (luc wastiaux)
Date: Sat, 12 Apr 2003 20:43:18 +0000 (UTC)
Subject: readline clear buffer
References: <slrnb9edf0.tds.luc@grizzly.dont-panic.info> <FWRla.4654$T34.116245@news2.tin.it>
Message-ID: <slrnb9guj7.uvs.luc@grizzly.dont-panic.info>

In article <FWRla.4654$T34.116245 at news2.tin.it>, Alex Martelli wrote:
> luc wastiaux wrote:
> 
>> Hello,
>> when using readline, I can retrieve the contents of the line buffer using
>> get_line_buffer()
>> 
>> but is there a way to clear this buffer ? none of the provided functions
>> seem to allow that: http://www.python.org/doc/lib/module-readline.html
> 
> Right: it seems that Python's readline.c does not expose interfaces to
> rl_delete_text, rl_kill_text, or other GNU Readline functions that would
> allow such behavior.  I suspect the simplest solution would be to patch
> readline.c to add such interfaces (also patch the docs accordingly), and
> submit the patch to sourceforge, where I think it would be welcome (and
> still in time to get into Python 2.3, for a modest addition, I think).

what would be my "deadline" for submitting such a patch and still make it 
into python 2.3 ?

-- 
luc wastiaux
 $ finger luc at info.4002.org



From nospam at nospam.net  Fri Apr 11 22:07:46 2003
From: nospam at nospam.net (c42)
Date: Sat, 12 Apr 2003 02:07:46 GMT
Subject: Newbie: Example of GUI front-end for SQL
Message-ID: <pan.2003.04.12.02.09.40.159546@nospam.net>

I'm researching a new project that will include a GUI front-end
(several screens) to an SQL server (probably Postgre or MySQL).

Can someone recommend an open source program that I can review that will
give me an idea if this is possible with python?

It is also important that the program be able to run on Windows and Linux.



From joshway_without_spam at myway.com  Tue Apr 29 10:48:51 2003
From: joshway_without_spam at myway.com (Josh)
Date: Tue, 29 Apr 2003 14:48:51 +0000 (UTC)
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com>
Message-ID: <b8m3cj$8vu$1@ginger.mathworks.com>

Christopher Blunck <blunck at gst.com> wrote:
...

> What advice can you offer on spreading the testing bug around?  

If you have an established codebase, it might be difficult to convince
people to go back and add unit tests for all functionality.  One good
way to ease people into it is to get them to write a unit test for
every bug-fix they make.



From squan at web.de  Sat Apr  5 01:44:20 2003
From: squan at web.de (Stefan Quandt)
Date: 4 Apr 2003 22:44:20 -0800
Subject: pickling new-style extension types: examples?
References: <slrnb8scn8.d6f.mac@mac.dgp.toronto.edu>
Message-ID: <ae5c48d7.0304042244.58c8e0fe@posting.google.com>

Maciej Kalisiak <mac at die.spammer.die.dgp.toronto.edu> wrote in message news:<slrnb8scn8.d6f.mac at mac.dgp.toronto.edu>...
> I've implemented a new extension type in C, and am now attempting to make it
> picklable, with little success.  Can someone point out any online examples of
> doing this, or even post some code snippets?  Ideally I would want to implement
> the pickling functionality in C, but will settle for a Python class wrapper
> around the extension type, if necessary.  Either the __reduce__ protocol, or
> using the "copy_reg" module will do.

See
http://www.ibm.com/developerworks/library/l-pypers.html?n-l-1172

To pickle an extension type you need 
 - a mechanism to "serialize" it (convert to a string) and
 - one to reconstruct it from a string.
Then you write a Python wrapper class where you implement the special
methods
__getstate__() and __setstate()__ which are called by the pickle
modules dump()/load():

class Wrapper( object ):
  def __init__( self ): self.c_object = CExtensionType( whatever )
  def __getstate__( self ):
    # Replace extension type instance by string converted version
    self.c_object_str = serialize( self.c_object )
    del self.c_object
    return self.__dict__
  def __setstate__( self, state ):	# state is a dict
    state[ c_object ] = CExtensionType( state.c_cobject_str ) #
Reconstruct from string
    del state[ c_object ] # delete string representation
    self.__dict__.update( state )



From tweedgeezer at hotmail.com  Tue Apr  1 15:01:04 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 1 Apr 2003 12:01:04 -0800
Subject: else: w/o except: - why not?
References: <mailman.1049156943.12744.python-list@python.org> <v8j9pb3v2cnra6@corp.supernews.com>
Message-ID: <698f09f8.0304011201.1a4939a1@posting.google.com>

claird at lairds.com (Cameron Laird) wrote in message news:<v8j9pb3v2cnra6 at corp.supernews.com>...
> And, for you, does
>   try:
>       f1()
>   except:
>       f2()
>       raise
> become *not* bad practice?

Wouldn't that be better written like thiS?

try:
  f1()
finally:
  f2()


Jeremy



From tungwaiyip at yahoo.com  Fri Apr 25 13:56:10 2003
From: tungwaiyip at yahoo.com (Tung Wai Yip)
Date: Fri, 25 Apr 2003 10:56:10 -0700
Subject: variable scoping problem
References: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com> <yfs65p3wf5e.fsf@black132.ex.ac.uk>
Message-ID: <ilmiavspd10m814g2gcpfe14dpg4rnh1v4@4ax.com>

On 25 Apr 2003 02:15:57 +0100, Alexander Schmolck <a.schmolck at gmx.net>
wrote:

>Tung Wai Yip <tungwaiyip at yahoo.com> writes:
>
>> Consider this
>> 
>> >>> x=1
>> >>> def foo():
>> ...   print x
>> ...
>> >>> def bar():
>> ...   #print x   <- referenced before assignment error
>> ...   x=2
>> ...   print x
>> ...
>> >>> foo()
>> 1
>> >>> bar()
>> 2
>> >>> print x
>> 1
>> >>>
>> 
>> What rules in Python explain the scoping of x? Is there a name for the
>> scope of the first x? 'module' scope? Why is bar() not able to access
>> x but foo() can?
>
>Q1: global scope
>
>Q2: Because when python sees that x is assigned to in a function it is treated
>as a local variable *anywhere* in that function (even *before* the
>assignment); *unless* you've declared it as global in that function. As long
>as you don't assign to it, you're fine.
>
>This makes sense -- otherwise you could easily introduce devious bugs by
>changing global variables without any intent of doing so. OTOH, you don't want
>to have to declare a module, function or a constant global just so that you
>can use it from within the function.
>
>Consider (all untested):
>
>def aTest():
>    print "HI"
>
>def bTest():
>    def aTest(): print "Just a local definition"
>    aTest()
>
>then consider:
>
>>>> aTest
>HI
>>>> bTest()
>Just a local definition
>>>> aTest()
>HI
>
>def aTest():
>    print "HI"
>
>def bTest():
>    aTest() 
>    def aTest(): print "NEVER"
>    return aTest(3)
>>>> bTest()
>
>
>and finally:
>
>def aTest():
>    print "HI"
>def bTest():
>    global aTest
>    aTest()
>    def aTest(): print "REDEFINED GLOBALLY"
>    aTest()
>>>> aTest()
>Hi
>>>> bTest()
>HI
>REDEFINED GLOBALLY
>>>> aTest()
>REDEFINED GLOBALLY
>
>
>'as

Bingo. The 'global' keyword solved my problem. All I'm trying to do is
to have a global variable to shared value so that one function can
read the value saved by a previous function.

For people with C background Python's way is rather different. It
threats a name differently when it is used for read only versus it is
written to. I understand your example. But I'm not sure which problem
is worst, to threat a name differently depends on usage or to
unintentional overwritten a global variable. C's way has its problem
but at least I'm very used to it. The most unambigous way is to force
people to declare variable first (but I guess this would be unpythonic
:).




From fperez528 at yahoo.com  Wed Apr  9 20:18:24 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 09 Apr 2003 18:18:24 -0600
Subject: Wraparound problems with time.clock() on Linux
References: <b722d0$bau$1@peabody.colorado.edu> <b722km$bau$2@peabody.colorado.edu> <slrnb9983v.vss.news@cipw08.physik.uni-bonn.de>
Message-ID: <b72d8h$ffe$1@peabody.colorado.edu>

Carsten Gaebler wrote:

> In article <b722km$bau$2 at peabody.colorado.edu>, Fernando Perez wrote:
>> Sorry, scratch that.  I went looking further in the docs, and it seems
>> that time.clock() calls the system's clock(). From man clock:
>> 
>>        Note  that  the  time  can  wrap  around.   On  a  32bit  system
>> where
>>        CLOCKS_PER_SEC  equals 1000000 this function will return the same
>> value
>>        approximately every 72 minutes.
>> 
>> 
> 
>> Anyone know of a reliable way to time long-running codes?
> 
> Without having tested it for more than 72 minutes, I'd try
> resource.getrusage().

Thanks for the tip.  I'll give it a shot.  It still has 32-bit issues with
time_t shoved in there, but hopefully the CLOCKS_PER_SEC  factor won't kill
me quite so early.

OTOH, getrusage() isn't nearly as user-friendly as a simple clock() call.
Bummer :(

Cheers,

f.



From maxm at mxm.dk  Mon Apr  7 15:28:31 2003
From: maxm at mxm.dk (Max M)
Date: Mon, 07 Apr 2003 21:28:31 +0200
Subject: Python & Music: Performance question
In-Reply-To: <3E91C3D9.E274C521@easystreet.com>
References: <1759581e.0304060218.164d4cd4@posting.google.com> <1759581e.0304070549.3b3a4caa@posting.google.com> <3E91C3D9.E274C521@easystreet.com>
Message-ID: <ehkka.6614$y3.832923@news010.worldonline.dk>

achrist at easystreet.com wrote:
> Is there anything in Python like the Midi support in the java
> sdk?  (under MS Windows) It's pretty easy in java to send midi data 
> straight from the program to the midi sequencer or synthesizer
> (no need to write a file or deal with some external player).   
> Anything like that for Python?

Not really no. Especially not something that is cross-platform.

It really is a pitty, as Python would be exceptionally good for 
algorithmic composition.

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From lost at soul.net  Fri Apr 25 19:37:09 2003
From: lost at soul.net (lost)
Date: Sat, 26 Apr 2003 00:37:09 +0100
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it>
Message-ID: <pan.2003.04.25.23.37.08.286778@soul.net>

On Fri, 25 Apr 2003 22:51:12 +0000, Alex Martelli wrote:

> for each_file in os.dir('/var/path/to/files/'):

actually, not quite what i was looking for :)
At this line, I get:
AttributeError: 'module' object has no attribute 'dir'

btw, I'm using a valid file path, not the one in the example - if that
makes any difference.
My python version is 2.2.1, on Red Hat Linux 8.

Thanks for the help!



From dgallion1 at yahoo.com  Sun Apr 13 11:30:28 2003
From: dgallion1 at yahoo.com (darrell)
Date: Sun, 13 Apr 2003 15:30:28 GMT
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk>
Message-ID: <oofma.65265$e8.23168@twister.nyroc.rr.com>

Looks like the only reference to your tuple is a weak one.
So it would die immediately.
Something must have more than a weakref to your tuple.

--Darrell


Robin Becker wrote:

> What's the correct incantation to get a weak reference to a tuple? The
> rather obvious wrapper approach fails as the wrapper dies immediately
> 
> ie weakref.ref(Wrapper((....)))
> 
> or we make a non anonymous wrapper which is what we want to avoid.




From adalke at mindspring.com  Tue Apr 22 14:38:30 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Tue, 22 Apr 2003 12:38:30 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
References: <mailman.1051023275.29338.python-list@python.org>
Message-ID: <b8420j$629$1@slb3.atl.mindspring.net>

Bjorn Pettersen:
> From: Ian Bicking [mailto:ianb at colorstudy.com]
[... + SQLObject example]

> for paper in Paper.select(AND(Paper.q.title == 'foo',
>                               Paper.q.author.startswith('Bob'))):

> Does anyone know the reason for not allowing an overload of the and
> operator?

The 'and' and 'or' operators work on the boolean-ness of the objects,
and apply short-circuit behaviour.  This is different than other binary
operators, which don't short-circuit.

For example, consider

counter = 0
class Funky:
  def __nonzero__(self):
    global counter
    print "Hello!"
    assert counter == 0
    counter = 1
    return 1

>>> counter = 0
>>> a = Funky() or Funky()
Hello!
>>> a
1
>>>

Note that __nonzero__ was only called once, not twice.

There's no way to implement that behaviour in Python, so
it cannot be overridden.  Eg, suppose there is an "__and__".
You might think it could look like

  def __and__(self, other):
      if bool(self): return self
      if bool(other): return other
      return False

However, this doesn't work because of the short-circuiting

  print 1 and 1/0

prints '1', and never evalutes 1/0.


BTW, what I've done for other projects which build up a parse
tree then evalute it is to hijack the arithmetic operators.  I've
used '*' for and, '+', for or, and '-' for 'but not'.  In that case, the
above expression is

> for paper in Paper.select(Paper.q.title == 'foo' +
>                                  Paper.q.author.startswith('Bob')):

                    Andrew
                    dalke at dalkescientific.com





From no at spam.invalid  Tue Apr 29 14:43:31 2003
From: no at spam.invalid (Russell E. Owen)
Date: Tue, 29 Apr 2003 11:43:31 -0700
Subject: Python 2.3b1: RuntimeError using rexec
Message-ID: <b8mh4l$2oo0$1@nntp6.u.washington.edu>

I'm trying out Python 2.3b1 and am having trouble with rexec:

[rowen:~] rowen% python
Python 2.3b1 (#1, Apr 29 2003, 10:41:11) 
[GCC 3.1 20020420 (prerelease)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
running pythonstartup.py
>>> import rexec
>>> r = rexec.RExec()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File 
"/Library/Frameworks/Python.framework/Versions/2.3/lib/python2.3/rexec.py
", line 184, in __init__
    raise RuntimeError, "This code is not secure in Python 2.2 and 2.3"
RuntimeError: This code is not secure in Python 2.2 and 2.3

Any hints?

I've looked at the what's new and the standard documentation for rexec 
and it seems normal -- no hint that the module has been deprecated or 
that one needs to jump through some "yes I know it's dangerous" hoop to 
use it.

Similarly a google search for python rexec runtimeerror turned up 
nothing interesting.

I stumbled across this when testing some existing code. This code uses 
rexec to convert a string representation of a dictionary to a dictionary 
as part of reading in font preferences.

Any suggestion on how to do that without rexec would also be 
appreciated. (I was quite disappointed that "dict" didn't do it!)

Here are some sample entries from the prefs file. I can change the file 
format if necessary but it'll be a headache for existing users. The 
prefs file handles lots of things, but dictionaries are only used for 
font descriptions and are my only use for rexec:
Data Font = {'weight': 'normal', 'family': 'Helvetica', 'overstrike': 
'0', 'underline': '0', 'slant': 'roman', 'size': '12'}
Background Color = '#fcfcfc'

-- Russell



From gupt_vive at hotmail.com  Tue Apr 15 07:57:17 2003
From: gupt_vive at hotmail.com (vivek kumar)
Date: Tue, 15 Apr 2003 17:27:17 +0530
Subject: Inserting HTML files as strings in database
Message-ID: <Law15-F42Tziwi7Qf4v0000175b@hotmail.com>

hi all,
  I am trying to make a search engine in python. I want to save the html 
files retrieved by the crawler into a Mysql database as strings 
(MEDIUMTEXT). But the characters like ' " etc are giving problems while 
inserting it as a string. Is there any way to escape these characters ( like 
\' for ') without reading the whole file and checking for these special 
charaters???
  And also as it is the first time i am trying to implement a search engine 
type of thing. I have implemented a basic crawler using urllib2 ( coz of 
proxy ). Can u please help me by giving some pointers??

I am using Python 2.2.2, MySQL 3.23.55, MySQLdb 0.9.2 on windows 2000.

thanks
Vivek Kumar

_________________________________________________________________
Vrroooom?? Fasten your seatbelts. 
http://server1.msn.co.in/msnSpecials/formula2003/index.asp Get set for F1 
2003





From hemanexp at yahoo.com  Tue Apr 29 20:15:33 2003
From: hemanexp at yahoo.com (perl lover)
Date: Tue, 29 Apr 2003 17:15:33 -0700 (PDT)
Subject: Problem in scalling............
Message-ID: <20030430001533.3975.qmail@web41708.mail.yahoo.com>

hi,
   i have a problem in scalling a canvas widget in
tkinter. I have created a canvas of size 1000 * 1000.
When i move the mouse over canvas the x,y values of
mouse pointer are showned in status bar. It is working
fine with normal canvas(ie, before scalling). I got
762*477 as bottom right pixel value.  When i scallled
the canvas stil i got the same value (762*477) if the
mouse is placed at bottom right of the canvas. Since i
used 2 as scalling factor for both x,y, I expect
bottom right value is as half as normal canvas (ie,
381,248). How can i get it? My program is given below.


from Tkinter import *

class App:
	def __init__(self,parent):
		self.myparent = parent
		self.canvas =
Canvas(parent,width="1000m",height="1000m")
	
self.canvas.create_rectangle(100,100,200,200,fill='blue')
self.canvas.bind('<B1-ButtonRelease>',self.ZoomIn)
		self.canvas.bind('<Motion>',self.mm)
		self.canvas.pack()

	def ZoomIn(self,event):
		self.canvas.scale(ALL,2,2,1.5,1.5)

	def mm(self,event):
		 print
self.canvas.canvasx(event.x),self.canvas.canvasy(event.y)



root = Tk()
myapp =  App(root)
root.mainloop()

Thanx

__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com




From chris.gonnerman at newcenturycomputers.net  Thu Apr 17 02:34:26 2003
From: chris.gonnerman at newcenturycomputers.net (Chris Gonnerman)
Date: Thu, 17 Apr 2003 01:34:26 -0500
Subject: Hi
References: <FA13E89D0804CB4DB5FB420B70E2C0CC0ADD952E@kecmsg07.ad.infosys.com> <3E9D5D58.7090607@ghaering.de>
Message-ID: <005c01c304ab$649c7600$ba01010a@local>

----- Original Message -----
From: "Gerhard Haering" <gh at ghaering.de>


> 201005- M.S.Shivakumaraswamy wrote:
> > Hi
> >
> > I am working on windows
> >
> > But there is no curses module for the windows
>
> There are alternatives to curses on Windows, like /F's console module [1].

Or my WConio module:

    http://newcenturycomputers.net/projects/wconio.html

Chris Gonnerman -- chris.gonnerman at newcenturycomputers.net
http://newcenturycomputers.net






From Kyler at news.Lairds.org  Fri Apr 11 15:23:15 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Fri, 11 Apr 2003 19:23:15 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <boejm-khk.ln1@news.lairds.org>

Harald Massa <cpl.19.ghum at spamgourmet.com> writes:

>- if you learn programming with Python and later have to use Java or 
>Visual Basic, it will be the most frustrating experience for the young 
>fellows.

After reading this thread, I am not sure that we're all making
what I consider a vital distinction.  When we talk about
"teaching Python" are we talking about "teaching students how
to program with Python as the primary language" or are we
"teaching students how to solve problems with Python as the
primary tool"?

If we're just teaching the use of Python as a way to solve
problems (my particular interest), it doesn't matter that the
student will be frustrated when going to another language -
heck, it's a mark of how appropriate Python is that it's more
difficult to achieve the same results with another language.

OTOH, if we're teaching programming (I have a Scheme course
going now.) it's a whole different ballgame.  There I feel it
is vital to understand the basics and be able to build them
from scratch.

(I recently used Python matrix operations to solve some GIS
problems.  I know terrifyingly little about either but I was
able to stumble through and get the job done with surprising
ease.  Do I really understand what I did?  No.  Could I code
it in C?  Probably not.  Was it adequate to solve my needs?
Yes!  What a great tool!)

I recall sub teaching "Computer Science" at a middle school a
few years ago.  The task for the day was to follow step-by-
step instructions on running MS Word to edit and print a file
on a floppy disk.  We don't all define "Computer Science" the
same. 

So...I think you probably all understand the difference
between Python as a tool and Python as a base language for
teaching programming.  I'd just like to see a more deliberate
distinction so that we don't confuse the goal of "Python and
Schools".

Thank you.

--kyler



From charts at purdue.edu  Sun Apr 20 19:27:21 2003
From: charts at purdue.edu (richard.a.charts.1)
Date: Sun, 20 Apr 2003 18:27:21 -0500
Subject: Keyboard Commands
Message-ID: <Pine.SOL.4.51.0304201825001.6411@herald.cc.purdue.edu>

Hey,
I'm working on a program and I want to use keyboard shortcuts for some of
my menu items.  I can't seem to find the way to implement this.  Can
anyone point me in the right direction?  ex.  Ctrl-x, ctrl-F1, etc.

Thanks for any help.
And please feel free to email ur answers to me directly.

Charts



From linux_pt2003 at netcabo.pt  Wed Apr  9 05:35:01 2003
From: linux_pt2003 at netcabo.pt (Carlos Sousa)
Date: Wed, 09 Apr 2003 10:35:01 +0100
Subject: PyQT
References: <3e92e21e$0$1986$a729d347@news.telepac.pt> <8DDka.4761$Du.16503@newsc.telia.net>
Message-ID: <pan.2003.04.09.09.34.52.420758@netcabo.pt>

Em Tue, 08 Apr 2003 17:26:28 +0000, Tim Gahnstr?m /Bladerman escreveu:

> "Carlos Sousa" <linux_pt2003 at netcabo.pt> >
> 
>> Can someone tell me who to install eric3 with PyQT ?
>> I try from the sources but something goes wrong
> 
> you must tell a lot more.
> 
> What platform are you using?
> What have you tried already?
> What goes wrong?
> What kind of errormessages do you get?
> Have you searched google groups for your particular error?
> 
> Tim

ok,I wiil be more specific:

I have mandrake 9.1.I build and install (with root account)the sources in following
order: QT(3.1), QScintilla, Sip, PyQT and then eric3.

After I execute "make" I got a lot of "errors", like that:

qextscintilla.cxx:2055: `SendScintilla' cannot be used as a function
qextscintilla.cxx:2056: `SendScintilla' cannot be used as a function
qextscintilla.cxx: At global scope:
qextscintilla.cxx:2064: syntax error before `::' token
qextscintilla.cxx:2069: 'QString' is used as a type, but is not defined as a type.
qextscintilla.cxx:2071: syntax error before `.' token
{standard input}: Assembler messages:
{standard input}:161: Error: symbol `bstart_off' is already defined
{standard input}:167: Error: symbol `words' is already defined
{standard input}:173: Error: symbol `bend_off' is already defined
{standard input}:191: Error: symbol `SendScintilla' is already defined
{standard input}:197: Error: symbol `buf' is already defined
{standard input}:203: Error: symbol `SendScintilla' is already defined
{standard input}:209: Error: symbol `buf' is already defined
{standard input}:215: Error: symbol `SendScintilla' is already defined
{standard input}:227: Error: symbol `SendScintilla' is already defined
{standard input}:313: Error: symbol `_ZZ7connectzE7connect' is already defined
{standard input}:319: Error: symbol `_ZZ7connectzE13SendScintilla' is already defined
make: ** [qextscintilla.o] Erro 1

After that I tried with rpm's as all seems to be OK, but I can't execute
"eric3.py".

The error is:
[root at a200-10-189-84 qt]# eric3
Traceback (most recent call last):
  File "/usr/lib/python2.2/site-packages/eric3/eric3.py", line 6, in ?
    from qt import QApplication, QTranslator, QTextCodec, QPixmap, QPoint, \
  File "/usr/lib/python2.2/site-packages/qt.py", line 39, in ?
    import libqtc
ImportError: /usr/lib/python2.2/site-packages/libqtcmodule.so: undefined symbol: _ZNK9QSGIStyle9classNameEv


What I ask is "who to install eric3 ?" (clean install) from someone tried
before me.

I search in some sources but I'm confuse about that, if you can help...I
will apreciate.








From exarkun at intarweb.us  Wed Apr 23 10:09:49 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 23 Apr 2003 10:09:49 -0400
Subject: Testing if string contains a substring
In-Reply-To: <b865dq$28gi$1@ns.felk.cvut.cz>
References: <b865dq$28gi$1@ns.felk.cvut.cz>
Message-ID: <20030423140949.GA9664@meson.dyndns.org>

On Wed, Apr 23, 2003 at 03:44:57PM +0200, Frantisek Fuka wrote:
> I want a simple test to see if str contains any occurences of substr. I 
> am currently using this:
> 
> import string
> if string.find(str,substr) + 1:
>     ...do something...
> 
> But it somehow looks "ugly" to me. Is there more "standard" way to test 
> this?

  The preferred way is:

    if s.find(substr) != -1:
        ...

  In 2.3, this new way will work:

    if substr in s:
        ...

  Jp

-- 
"One World, one Web, one Program." - Microsoft(R) promotional ad
"Ein Volk, ein Reich, ein Fuhrer." - Adolf Hitler
-- 
 up 34 days, 11:03, 4 users, load average: 1.96, 1.78, 1.70




From exarkun at intarweb.us  Thu Apr 17 09:52:15 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 17 Apr 2003 09:52:15 -0400
Subject: Python Thread Question
In-Reply-To: <bd993a2f.0304170534.245aa2d5@posting.google.com>
References: <bd993a2f.0304170534.245aa2d5@posting.google.com>
Message-ID: <20030417135215.GD25373@meson.dyndns.org>

On Thu, Apr 17, 2003 at 06:34:11AM -0700, Anand B Pillai wrote:
> Hi Pythonistas
> 
>  I have written this application which is a kind of intranet
>  web-spider. It crawls a given url and retrives the files in
>  the url and saves it to the disk. 
> 
>  Now when I do this using multiple threads(python threads), 
>  assigning each url to a thread I find that the download gets
>  completed faster than if it were in a single thread. I assume
>  that the reason for this must be simple, that when you use 
>  a single thread idiom, the app has to wait till a file is 
>  downloaded. Whereas if you use a thread for each download, 
>  the app can spawn other threads for other downloads, so no 
>  wait is needed. I am firing off a group of threads (limited
>  by a maxthread count) and pooling them in a threadgroup. 
>  Once the threads are fired for download, the app does not
>  try to control them until they finish/killed or a network 
>  time-out occurs.
> 
>  Ideally speaking, multithreading need not improve the speed
>  of an application but in examples like this which involve
>  bottlenecks like network traffic, it does. My questions about
>  this are:

  Unless you use non-blocking sockets.  But that is another story.

> 
>  1. Does python threads work only if the native platform supports
>     threading ?  i.e, is python firing 'C' threads which in turn
>     fire the platform API threads (Win32 for windows/ pthreads for
>     linux etc)?

  Yes, native threads.

>  2. Can a software API (Win32/pthreads) do multithreading even if
>     the CPU does not support multithreading ? (might seem like a
>     superfluous question when almost all cpus does in this age, but
>     the question is still valid). Or is multithreading ultimately
>     related to how the CPU handles threads ?

  I am not aware of a modern CPU on which threading cannot be implemented. 
Perhaps a more useful piece of information to give you is this: threading
APIs can be implemented either in a kernel or in a userspace library; they
don't require hardware support.

>  3. Is the apparent increase in speed in my program using multiple
>     threads attributable to the CPU or the platform API or python ?

  It is attributable to none of the above, as far as I can tell.  When you
perform blocking reads, your app ends up spending most of its time idle. 
When you launch 50 threads and do 50 blocking reads, your app spends
slightly less of its time idle, but is still mostly just waiting for those
reads to return something.  The difference is, data is probably coming in
for all 50 of those reads at the same time, maybe from different hosts - so
you get higher network throughput, and so your program seems faster.

> 
>  4. Can I safely say that multithreading will improve my application
>     performance if it has similar work to do on many resources at the
>     same time ? (egs: a web parser/ spider/ a disk-to-disk file copier/
>     directory synchronizer) Or does it depend upon the nature of the
>     task at hand ?

  Nope.  Threading costs you context switches.  A multi-threaded app using
blocking IO will appear "faster" than a single-thread app doing the same,
but both take longer to run than a single written app using non-blocking IO.

  Context switches are usually relatively cheap compared to what your app is
actually doing, though, so the difference between doing blocking IO in a
multi-thread app and non-blocking IO in a single-thread app isn't always
obvious.  (Other things associated with multi-threaded apps, such as
deadlocks and race conditions are, though ;)

  Hope this helps,

  Jp

-- 
It is practically impossible to teach good programming style to
students that have had prior exposure to BASIC: as potential
programmers they are mentally mutilated beyond hope of
regeneration.        -- Dijkstra
-- 
 up 28 days, 9:02, 5 users, load average: 0.04, 0.03, 0.00




From anton at vredegoor.doge.nl  Sat Apr 19 09:03:38 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Sat, 19 Apr 2003 15:03:38 +0200
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <Gu_na.34381$LB6.779929@news1.tin.it>
Message-ID: <b7rhgk$4of$1@news.hccnet.nl>

Alex Martelli <aleax at aleax.it> wrote:

>[alex at lancelot python2.3]$ python -O timeit.py -s"lst=[('a',1), ('b',2), 
>('c',3)]*100" -s'import operator' 'x=0' 'for y in lst: x+=y[1]'
>10000 loops, best of 3: 103 usec per loop
>
>Using lists 10 or 100 times as long seems to show O(N) behavior for
>each case, as could be expected.  So, the simplest (and I would agree
>thereby most Pythonic) approach is over twice as fast as the fancy ones
>(measurements on 2.2.2 seem to show similar results, all approaches
>being about 20% slower than in 2.3, uniformly).

Unpacking the tuple early seems to be 9% faster, if Knuth allows it.

Anton

Anton Vredegoor at hyperspace ~
$ python -O timeit.py -s"lst=[('a',1), ('b',2), 
('c',3)]*100" -s'import operator' 'x=0' 'for y in lst: x+=y[1]'
10000 loops, best of 3: 125 usec per loop

Anton Vredegoor at hyperspace ~
$ python -O timeit.py -s"lst=[('a',1), ('b',2), 
('c',3)]*100" -s'import operator' 'x=0' 'for i,j in lst: x+=j'
10000 loops, best of 3: 115 usec per loop





From aleax at aleax.it  Sun Apr 13 17:47:32 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 21:47:32 GMT
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <b7aj86$q9i$1@slb6.atl.mindspring.net> <j%9ma.9030$T34.282951@news2.tin.it> <b7cl9v$63t$1@slb2.atl.mindspring.net>
Message-ID: <UVkma.9969$LB6.266285@news1.tin.it>

Matt Gerrans wrote:

> "Alex Martelli" wrote:
>> > Learning Python is an excellent book, but unfortunately quite out of
> date.
>>
>> Yes to both.  Fortunately, a new version IS in the works, and indeed
>> at quite an advanced stage...
> 
> That's great news!   I can't wait.    But don't worry, it won't stop me
> from buying a copy of your recently released Python in a Nutshell -- which
> has been garnering rave reviews -- in the meantime.  ;-)

Not worrying in the least, thanks -- I think that the books complement
each other quite well;-).


Alex




From rmunn at pobox.com  Thu Apr 10 13:43:00 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 10 Apr 2003 17:43:00 GMT
Subject: BEGINNER: converting integer to short month name?
References: <mailman.1049919678.686.python-list@python.org>
Message-ID: <slrnb9amil.bom.rmunn@localhost.localdomain>

MCollins at co.seminole.fl.us <MCollins at co.seminole.fl.us> wrote:
> that works fine for the numerical month names.  now i'm trying to convert
> the numerical month into it's short month name equivalent to grab those
> files.
> 
> i could always use a long if , elif, elif to do the job, but i was
> wondering if there was built-in function already capable of that.  i was
> just looking for something a little more elegant.

Try the time.strftime function:

    http://www.python.org/doc/2.2/lib/module-time.html

You might have to construct a tuple:

    time_tuple = (2000, the_month_num, 1, 0, 0, 0, 0, 0, 0)
    month_abbr = time.strftime('%b', time_tuple)

This will get you 'Jan', 'Feb', 'Mar', etc. depending on the value of
the_month_num.

Interesting, looks like you don't have to even specify a valid year or
day:

    the_month_num = 2
    time_tuple = (0, the_month_num, 0, 0, 0, 0, 0, 0, 0)
    month_abbr = time.strftime('%b', time_tuple)
    print month_abbr

prints out "Feb".

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From tjreedy at udel.edu  Fri Apr 25 11:46:34 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Fri, 25 Apr 2003 11:46:34 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com> <058qa.30797$y3.2515229@news010.worldonline.dk>
Message-ID: <nJGdnZlE070sxTSjXTWcog@comcast.com>

"Max M" <maxm at mxm.dk> wrote in message
news:058qa.30797$y3.2515229 at news010.worldonline.dk...
> Actually I write a LOT of "throwaway" scripts, that I know I will
not
> need again, or that if I will use them again, I will just rewrite
them.
>
> It is far more important for me that these scripts are up and
running
> fast than that they are correct.
>
> I use them for massaging large amount of text. If ie. I need to make
a
> website for a customer, and I have a large amount of similar texts.
It
> could be something like a list of employees that is keept in a flat
> file, text document, excell sheet or similar.

This sort of thing is what I first learned Python for also.
In my case, it was been 'flat', rectangular data table files.

> Usually these files are pretty similar, but I know they are full of
> errors, misspellings etc. So I have to edit them manually anyway.

Example throwaway scripts for error discovery:

for line in file('path.ext', 'r').readlines:
    if len(line) != N: print line # N = expected line length
-or-
    if line[M] not in ' mf': print line # M = 'sex' data column

This is followed by maual editing to try to fix offending lines.
Such a script *is* a test script.  Should I write a test for the
tests?*

(Having said that, if I were to pull such snippets together to write a
datatest.py module, having checked that none exists, I should and
probably now would write test data with errors and corresponding  unit
tests.)

> I write a small script that can save me a lot of copy-pasting,
adding
> files etc. by automating most of the process. These scripts only
have to
> be "good enough" as it is faster to manually edit minor errors in
the
> resulting output than to rewrite the script.

Or the fix may require human judgement that is hard to program, like
finding the extra/missing space to delete/add or it may require
physical action like referring back to original data sheet or even
making a phone call to correct a mis-entered value.


*To paraphrase Socrates/Plato: "Who tests the testers?"

My Python2.2/Lib/test directory seems *not* to have a
'test-unitest.py' file!  -- though perhaps it should -- and could (see
Kent Beck's book).  In any case, there are no automated tests of all
the other test files.  Test bugs and omissions are
found, when they are, the old-fashioned way -- re-inspection, perhaps
by another pair of eyes,  or investigation of anomalous experience.

My point: test code is useful when it is easier to verify than the
payoff code it is verifying.  It is less useful or even a nuisance
when it it not.  Expected revision frequency of payoff code versus
test code factors into 'easier'.

Terry J. Reedy





From peter at engcorp.com  Tue Apr  1 21:44:00 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 21:44:00 -0500
Subject: stop exeuction and give my a prompt
References: <b6dara$5mv$1@news.asu.edu> <u3cl1d98y.fsf@online.no>
Message-ID: <3E8A4E70.AF4A830@engcorp.com>

Syver Enstad wrote:
> 
> David Bear <david.bear at asu.edu> writes:
> 
> > I would like to have the ability to insert a statement in my code that
> > would stop program execution at that point and give me a python prompt. 

> Use the python debugger. In the most primitive way is to use it from
> the interactive prompt.
> 
> Do:
> import pdb # at the top of your sourcefile
> 
> Do:
> pdb.set_trace() # at the point you want to debug.

And note that when you do that, the context of the prompt
is *inside* set_trace(), or somewhere like that, so to see
the local context from which it was called, you have to execute
a "return" or "r" command.  

Also be careful of trying to work with one-character variable
names, as the debugger tends to use many of these as shortcuts
for various commands.  

And there are probably other gotchas, but it's pretty damn
convenient if you ask me!

-Peter



From abpillai at lycos.com  Mon Apr  7 06:49:04 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 7 Apr 2003 03:49:04 -0700
Subject: PYTHON EDITOR!!
References: <mailman.1049690233.24009.python-list@python.org>
Message-ID: <bd993a2f.0304070249.5365e166@posting.google.com>

If you are on Linux, you have some of the best choices
for python-editing by default. Emacs & vim have python-mode
plugins written by programmers. 

For python-mode in emacs go here ->
http://www.python.org/emacs/python-mode/

I am not a vim user so you can post to comp.editors.vi for 
help on python-mode on vim. I am sure there is one :-)

 If you havent used emacs/vim, then this is also probably
a good way to get started. For emacs try http://gnu.org/software/emacs
and vim of course is in http://vim.org.

 Python comes with its own editor on installation, 'IDLE'.
It is some nice useful features like code completion, syntax
highlighting, class browsing. But I find that it is somewhat
unstable on Windoze 98. 

  I dont have very high opinion on the commercially / shareware
editors available out there. They come with obscure syntax files
which none can edit for customization. I dont know anything about
cool-edit. Perhaps you can try out Ultraedit or Editplus. Google
them to find out the links.

For a complete online collection, point & click ->
http://www.python.org/cgi-bin/moinmoin/PythonEditors

Enjoy pythoning!

Anand Pillai

Visit Python Chamber - http://members.fortunecity.com/anandpillai

nanthan at siswa.utm.my (Pathmenanthan RamaKrishna) wrote in message news:<mailman.1049690233.24009.python-list at python.org>...
> hi,
> 
> im Nanthan, a postgraduate student from Malaysia.im very new to 
> python language but im already getting use of it.Currently,i 
> have been working in windows, but my development actually is on 
> Linux, so can u suggest me the best editor(i mean use friendly 
> and graphical,i dont want command line editor)for python?
> 
> is cooledit the right choice.
> 
> i hope u'll reply soon
> 
> thanks
> 
> Nanthan.R



From max at alcyone.com  Mon Apr 14 03:16:15 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 00:16:15 -0700
Subject: more informative stack traces
References: <e0bf36b5.0304132100.4324c3b8@posting.google.com>
Message-ID: <3E9A603F.77BC3767@alcyone.com>

"L.C." wrote:

> I am proposing a different style of stack trace which I found more
> useful in debugging problems with live code. The main change involves
> showing the actual parameter values in function calls, gdb-style.
> It sounds simple because it really is, but it saves lots of guesswork
> when navigating through the stack to see what caused a problem.

Looks pretty neat.  I find myself wondering what the arguments were in
the call stack myself.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The hour which gives us life begins to take it away.
\__/ Seneca
    Church / http://www.alcyone.com/pyos/church/
 A lambda calculus explorer in Python.



From none at none.com  Mon Apr 28 05:58:24 2003
From: none at none.com (Iwan van der Kleyn)
Date: Mon, 28 Apr 2003 09:58:24 +0000
Subject: Gvim unstable on Mdk 9.1?
Message-ID: <3eacdf1b$0$145$e4fe514c@dreader9.news.xs4all.nl>

I use Vim and Gvim as my default editor(s). On Mandrake 9.1 I've lots of 
problems with Gvim, though. It keeps on 'hanging' with a guaranteed 
crash when I start the program with a path as parameter. So its' 
impossible to open a file in Gvim from Konqueror, for example.

Does anyone know a fix? Or do I have to install Vim from source?

Regards,

Iwan




From carroll at tjc.com  Fri Apr 11 02:57:21 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Thu, 10 Apr 2003 23:57:21 -0700
Subject: string() - Is this some kind of sick joke?
References: <36dbf824.0304091142.527af834@posting.google.com>
Message-ID: <1dpc9v4beon9upq2sd3clfb2plh7nlqm7p@4ax.com>

On 9 Apr 2003 12:42:44 -0700, chlump at yahoo.com (Daniel Gowans) wrote:

>>>> y="'hello'"
>>>> import string
>>>> string.strip(y,"'")
>Traceback (most recent call last):
>  File "<interactive input>", line 1, in ?
>TypeError: strip() takes exactly 1 argument (2 given)
>>>> 
>
>What gives?  Did I get the bogus version?  Is this wishful thinking
>documentation?  I also tried  y.strip("'")  with the same results.

I get that, too, on string.strip(y,"'"), but y.strip("'") works fine:

>>> import string
>>> y="'hello'"
>>> y
"'hello'"
>>> string.strip(y,"'")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: strip() takes exactly 1 argument (2 given)
>>> y.strip("'")
'hello'
>>>

Are you sure y.strip fails for you?  

The strip definition in string.py does not match the doc; it only
takes one argument:

  # Strip leading and trailing tabs and spaces
  def strip(s):
     """strip(s) -> string

     Return a copy of the string s with leading and trailing
     whitespace removed.

     """
     return s.strip()





From aahz at pythoncraft.com  Tue Apr 15 08:17:31 2003
From: aahz at pythoncraft.com (Aahz)
Date: 15 Apr 2003 08:17:31 -0400
Subject: instance creatiion as an "event"
References: <mailman.1050095779.7490.python-list@python.org> <pan.2003.04.11.23.20.04.65410@webone.com.au> <mailman.1050117319.5102.python-list@python.org>
Message-ID: <b7gt8r$em0$1@panix1.panix.com>

In article <mailman.1050117319.5102.python-list at python.org>,
Steven Taschuk  <staschuk at telusplanet.net> wrote:
>Quoth Simon Burton:
>> On Fri, 11 Apr 2003 17:15:32 -0400, Arthur wrote:
>>>
>>> The only thing I seem to be missing is some kind of "hook" to trigger a 
>>> general update cycle of all existing class instances upon the creation 
>>> of a new class instance at the interactive prompt
>> 
>> class Member:
>> 	members = []
>> 	def __init__(self):
>> 		for m in Member.members:
>> 			m.notify(self)
>> 		Member.members.append(self)
>> 	def notify(self,other):
>> 		print self,"a new member:",other
>
>Though you will probably want to use weak references in
>Member.members, unless you intend that all instances of the class
>live as long as the class does.

Blech.  It's essentially impossible to delete a class as long as
instances of it exist.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From fperez528 at yahoo.com  Wed Apr 16 18:28:13 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 16 Apr 2003 16:28:13 -0600
Subject: LinearAlgebra vs LinAlg.py (access to full LAPACK routines)
References: <1ada6ef5.0304161421.2c81eeea@posting.google.com>
Message-ID: <b7kldt$rqs$1@peabody.colorado.edu>

Jason Taylor wrote:

> In the documentation for the LinearAlgebra package it states:
> 
>     # This module is a lite version of LinAlg.py module which contains
>     # high-level Python interface to the LAPACK library.  The lite
> versioho
>     # only accesses the following LAPACK functions: dgesv, zgesv,
> dgeev,
>     # zgeev, dgesdd, zgesdd, dgelsd, zgelsd, dsyevd, zheevd, dgetrf,
> dpotrf.
> 
> Does LinAlg.py exist? If so, where can it be found?
> 
> The reason I want it, is so that I can have access to all LAPACK
> routines (If it provides that).

>From Numeric's setup.py:

# delete all but the first one in this list if using your own LAPACK/BLAS
sourcelist = [os.path.join('Src', 'lapack_litemodule.c'),
              os.path.join('Src', 'blas_lite.c'),
              os.path.join('Src', 'f2c_lite.c'),
              os.path.join('Src', 'zlapack_lite.c'),
              os.path.join('Src', 'dlapack_lite.c')
             ]
# set these to use your own BLAS
library_dirs_list = []
libraries_list = []


You can build NumPy with your own (full) lapack/blas, and even with an
atlas-optimized version.

If you use scipy (http://scipy.org), it is designed to detect a full
lapack/blas build (including an atlas-based one) and give you the whole
thing in one package.  Note that scipy is a superset of Numpy, so you get
all of Numeric plus much more.

Cheers,

f



From hat at se-46.wpa.wtb.tue.nl  Mon Apr  7 07:36:33 2003
From: hat at se-46.wpa.wtb.tue.nl (Albert Hofkamp)
Date: Mon, 7 Apr 2003 11:36:33 +0000 (UTC)
Subject: building a .py files with distutils
References: <mailman.1049559387.7474.python-list@python.org>
Message-ID: <slrnb92om2.fnk.hat@se-46.wpa.wtb.tue.nl>

On Sat, 5 Apr 2003 10:15:19 -0600, Jeff Epler <jepler at unpythonic.net> wrote:
> I'm writing a program which uses a generated .py file (it's generated based
> on a yapps grammar file).  I know how to generate the .py file either as a 
> shell command (python yapps.py gram.g) or as a python expression
> (yapps.generate("gram.g"))

I had a similar problem with 'building' documentation files as preparation
of creating a distribution.

I wanted to do that in setup.py (i.e. I want to automate the building
process, I have a running Python interpreter, I have setup.py, so what's
the problem?)
After a few attempts to control the building process from Python (i.e.

import ftlib,distutils,sys
build_documentation()
distutils.build_my_package(srcdir='.',tgt=['src','rpm'],version=sys.argv[1])
ftplib.ftp_result_to_website()

), I switched to having a shell-script outside distutils that did the
preparation, then it calls distutils to build the source and binary
distributions.

There are still a few problems (for example, I'd like to have distutils
take the version number as a parameter instead of a hard-coded value in
setup.py), but it works.



Albert
-- 
Unlike popular belief, the .doc format is not an open publically available format.



From zephyr01 at alltel.net  Fri Apr 25 13:06:06 2003
From: zephyr01 at alltel.net (Hans Nowak)
Date: Fri, 25 Apr 2003 13:06:06 -0400
Subject: Vim Newbie ? Python Edit/Compile/Run  cycle and integrating py
 help questions
In-Reply-To: <cb035744.0304250832.71aef2d9@posting.google.com>
References: <a791ab19.0304240532.5a025b81@posting.google.com> <3EA7EBD3.3010605@ghaering.de> <mailman.1051243975.12570.python-list@python.org> <cb035744.0304250832.71aef2d9@posting.google.com>
Message-ID: <3EA96AFE.4090905@alltel.net>

Will Stuyvesant wrote:
> [Hans Nowak]
> 
>>I have been trying to edit Python source code with vim on Mac OS X.  Most of 
>>the settings in my .vimrc file (copied from Windows) work just fine, except for 
>>one.  I use Tab to indent and Shift-Tab to dedent (a line, not a block).  Tab 
>>behaves as usual, but Shift-Tab doesn't seem to work; it just behaves like I 
>>pressed Tab.
> 
> 
> In gvim on windows and linux (untested on Mac, since I have no access
> to a Mac anywhere):
> 
> Start block selection with the v key, or with <SHIFT>v.  And then of
> course the j key for line down and the l key for going right etc.
> 
> To indent a line, or a selected block, use <SHIFT>>>.  To dedent it do
> <SHIFT><<.

Thanks for your reply; however, I know how to do indents and dedents in vim, 
the problem is more how to map Shift-Tab to do the dedent.  The reason is 
simply that I have been using it for years and don't want to unlearn it and get 
used to a different key combination. :(

I'll see if I can do something with the key combinations you mentioned.

Cheers,






From cliechti at gmx.net  Thu Apr  3 13:37:17 2003
From: cliechti at gmx.net (Chris Liechti)
Date: 3 Apr 2003 20:37:17 +0200
Subject: Python commands
References: <3e8bb9c5$1_1@news.iprimus.com.au>
Message-ID: <Xns9352D1FE6C1E0cliechtigmxnet@62.2.16.82>

"Paul Dexter" <dextech at terrigal.net.au> wrote in
news:3e8bb9c5$1_1 at news.iprimus.com.au: 
>                 I have just discovered Python and I was wondering if
>                 anyone 
> could help me find an equivalent command in Python for the "OUT" and
> "INP" commands in Qbasic. I wish to convert a Qbasic program to
> Python. 
>     The following snippet is an example of the task I wish to perform;
> 
> base0 = &H378:
> 
> interface:
> OUT base0, &HA0
> OUT base0, &H80
> OUT base0 + 2, &H8
> 
> V = (INP(base0 + 1) AND &H10) * &H8

appart from the solutions others already posted, i have also a pyparallel 
module, found on http://pyserial.sf.net

chris

-- 
Chris <cliechti at gmx.net>




From jon+usenet at unequivocal.co.uk  Tue Apr 15 19:59:45 2003
From: jon+usenet at unequivocal.co.uk (Jon Ribbens)
Date: 15 Apr 2003 23:59:45 GMT
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <mailman.1049653811.11491.python-list@python.org> <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
Message-ID: <slrnb9p77h.hth.jon+usenet@snowy.squish.net>

In article <d5mm9v8em9j41kus2tvj2raipd7dagv18k at 4ax.com>, Andy Robinson wrote:
> Wouldn't it be nice if one day we could have some standardized
> interfaces, so that one could switch between these without undue pain
> and reduce the confusion people must face?

I could suggest the interface defined in jonpy
(http://jonpy.sf.net/)... The low-level interface is very similar to
Java servlets and hence may be popular. There are implementations for
standard CGI, mod_python and FastCGI, and other implementations should
be entirely possible.

Is there a mailing list for this sort of thing?



From mwh at python.net  Mon Apr 28 07:10:15 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 28 Apr 2003 11:10:15 GMT
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com> <3ea947e1$0$198$a1866201@newsreader.visi.com> <mailman.1051301620.25438.python-list@python.org> <3ea9ac4c$0$179$a1866201@newsreader.visi.com> <b8camh$g65$05$1@news.t-online.com>
Message-ID: <7h3ptn6khds.fsf@pc150.maths.bris.ac.uk>

"Martin v. L?wis" <martin at v.loewis.de> writes:

> Grant Edwards wrote:
> > But those C runtime library routines aren't redundant. they're
> > providing functionality that Python doesn't implement.
> 
> They are *also* providing functionality that Python
> doesn't implement. There is quite some redundancy, too. E.g.
> a Posix C library supports regular expressions, which Python
> also does, and Python doesn't use the C implementation.

And this trend is increasing, due to rampant bugs and or weirdnesses
in various platforms C libraries (yes, I mean *you*, strptime).

Cheers,
M.

-- 
  Of course, it obviously is beta hardware so such things are to be
  expected, but that doesn't mean that you can't point your fingers
  and generate a nelson style HAHA at a multi billion dollar
  corporation's expense.                   -- CmdrTaco on slashdot.org



From claird at lairds.com  Tue Apr  8 07:44:16 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 08 Apr 2003 11:44:16 -0000
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049722991.18682.python-list@python.org> <v938h26mdl9s8f@corp.supernews.com> <3E923E72.AA57E192@engcorp.com>
Message-ID: <v95dggqr41de5a@corp.supernews.com>

In article <3E923E72.AA57E192 at engcorp.com>,
Peter Hansen  <peter at engcorp.com> wrote:
		.
		.
		.
>Source code with decent unit and acceptance tests is an asset.
>Source code without decent tests is a liability.
>
>-Peter Hansen

And the generation procedures.  The three things the
Spanish Inquisition always ... no, wait; the generation
procedures, and the installer; the *four* ...

Fondly,

-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From max at alcyone.com  Thu Apr 10 01:58:59 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 09 Apr 2003 22:58:59 -0700
Subject: BEGINNER: converting integer to short month name?
References: <OF47A94545.30289BFA-ON85256D03.006E4C80-85256D03.006FA823@seminole.fl.us> <mailman.1049921355.15625.python-list@python.org> <b73066$evv$1@slb9.atl.mindspring.net>
Message-ID: <3E950823.DE77C48D@alcyone.com>

Matt Gerrans wrote:

> Looks like it would be
>    calendar.month_abbr[ the_month + 1 ]
> 
> Interestingly, calendar.month_abbr[ 0 ] is 'Saturday'.   Hmm...

Python 2.2.2 (#2, Oct 14 2002, 17:32:20) 
[GCC egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import calendar
>>> calendar.month_abbr[0]
''

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If love is the answer, could you rephrase the question?
\__/ Lily Tomlin
    Official Omega page / http://www.alcyone.com/max/projects/omega/
 The official distribution page for the popular Roguelike, Omega.



From jhibschman at yahoo.com  Thu Apr 10 14:52:44 2003
From: jhibschman at yahoo.com (Johann Hibschman)
Date: Thu, 10 Apr 2003 18:52:44 GMT
Subject: Macros in Python?
References: <b71qde$q71$07$1@news.t-online.com> <slrnb9b01m.108.do-not-spam-ben.hutchings@tin.bwsint.com> <b74bem$oi3$02$1@news.t-online.com>
Message-ID: <m2znmymbaj.fsf@yahoo.com>

Dominic <oblivious at web.de> writes:

> > Alternatively, you can curry the above function:
> >     def with_root_privs(func):
> >         def call(func, *args, **kwargs):
> >             save_euid = os.geteuid()
> >             os.seteuid(0)
> >             try:
> >                 func(*args, **kwargs)
> >             finally:
> >                 os.seteuid(save_euid)
> >         return lambda *args, **kwargs: call(func, *args, **kwargs)

> Thanks. This one is also very good. :-)
> Interesting idea to use a closure to build
> a function wrapper.

Hm.  Why not just do:

def with_root_privs(func):
    def call(*args, **kwargs):
        save_euid = os.geteuid()
        os.seteuid(0)
        try:
            func(*args, **kwargs)
        finally:
            os.seteuid(save_euid)
    return call

I don't see what that extra layer of indirection is gaining you.

-Johann



From whisper at oz.net  Sun Apr  6 19:30:09 2003
From: whisper at oz.net (David LeBlanc)
Date: Sun, 6 Apr 2003 16:30:09 -0700
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2y92n5ih8.fsf@localhost.localdomain>
Message-ID: <GCEDKONBLEFPPADDJCOEIEJAJFAA.whisper@oz.net>

> PyCrust and PyAlaMode are both based on Scintilla as well.  In
> wxPython the Scintilla wrapper is called wxStyledTextCtrl.  The
> difference between PyAlaMode and Scite is that the autocompletion and
> calltips in PyAlaMode are dynamically generated based on the namespace
> of the module being edited whenever the user decides to update the
> namespace.
>
> In short, I'm already doing what you're suggesting.  ;-)
>

Umm... If you're using Scintilla (or it's wxPython analog), you're not quite
doing what I suggested. Unlike Scintilla, Scite has buffer, file management
and other bells and whistles added that Scintilla does not have. A review of
the Scite classes' public methods might be instructive :)

Regards,

Dave LeBlanc
Seattle, WA USA





From dirk at gerrits.homeip.net  Thu Apr 10 02:18:32 2003
From: dirk at gerrits.homeip.net (Dirk Gerrits)
Date: Thu, 10 Apr 2003 08:18:32 +0200
Subject: Odd behaviour of dir(object) with metaclasses.
In-Reply-To: <BIEJKCLHCIOIHAGOKOLHGELAFEAA.tim.one@comcast.net>
References: <b71toc$k7m$1@main.gmane.org> <BIEJKCLHCIOIHAGOKOLHGELAFEAA.tim.one@comcast.net>
Message-ID: <b732ap$74q$1@main.gmane.org>

Tim Peters wrote:
> [Dirk Gerrits]
> 
>>I am a bit puzzled by the behaviour of the dir builtin.
[snip]
>>
>>Why is there an unnatural (IMHO) distinction between type/class objects
>>and other non-module objects? I assume that a class's class is not
>>considered become dir comes from the pre-metaclass era?
> 
> 
> No, dir() was rewritten from scratch for 2.2.  The comments in the C code
> say
> 
> 	/* Elif some form of type or class, grab its dict and its bases.
> 	   We deliberately don't suck up its __class__, as methods belonging
> 	   to the metaclass would probably be more confusing than helpful. */
> 
> That's what Guido believed at the time, and I agreed at the time, so that's
> what the code does.  Across at least seven pre-releases of 2.2, nobody
> disagreed <wink>.

Well I think its understandable, but in my mind, what dir does to 
classes should be called something like instancedir. Anyway, I guess you 
have a point. People who don't use metaclasses won't care anyway. And 
those who do might not care either. ;)

> 
>>Shouldn't the docs really read
>>...
> 
> 
> No, becuase the suggested rewording doesn't describe the actual behavior --
> as you noted above, the current wording does.

Yeah I meant to say 'Shouldn't the docs really read ... with the 
behaviour changed accordingly?'. But I assume you'll say no to that as 
well. ;)

>>...
>>I hope you'll be able to clarify these things for me a bit.
> 
> 
> dir() was meant to be a convenience in interactive mode, and that's all.
> Because of that, no two people will agree on what it "should do" in all
> respects in all cases.  Changing it always makes people scream too, so I
> doubt it will be changed again.  If you don't like what it does, write a dir
> that does what you like -- doing so in Python is easy enough.

Yes, one of Python's many great strengths. :)

Thanks for clearing this one up.

Dirk Gerrits






From alloydflanagan at attbi.com  Thu Apr  3 17:50:42 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 3 Apr 2003 14:50:42 -0800
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n ...)
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com> <mailman.1049298845.7655.python-list@python.org> <3E8B4094.F61515B3@engcorp.com>
Message-ID: <a8b7f07a.0304031450.76fa610c@posting.google.com>

Peter Hansen <peter at engcorp.com> wrote in message news:<3E8B4094.F61515B3 at engcorp.com>...
> Jp Calderone wrote:
> > 
> forth my perspective on this.  I'm coming to *count* on this as one of
> the most effective solutions for certain kinds of testing.  (Specifically,
> by creating a mock file system by intercepting open() calls in my own
> method, to allow generating of various file system errors and situations
> and to increase performance of unit tests by avoiding the real file
> system.)
> 
> -Peter

Hey, that's a great idea, thanks!



From aleax at aleax.it  Sat Apr 26 15:27:56 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 19:27:56 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <ZFyqa.7824$3M4.220335@news1.tin.it> <oprn827qp6mu3s0q@news.cis.dfn.de>
Message-ID: <05Bqa.8473$3M4.233690@news1.tin.it>

Klaus Meyer wrote:

> On Sat, 26 Apr 2003 16:42:33 GMT, Alex Martelli <aleax at aleax.it> wrote:
> 
>> as you proceed to learn that "to call callables, you need to CALL them"
> 
> You misunderstand me (and the OP, i think). It's not a problem of
> understanding, it's a simple type error or a "false remember" (or a
> "Fl?chtigkeitsfehler" in German, i don't know a translation).
> Programmers sometimes make simple mistakes ;-)

I think the misunderstanding is probably the other way around.  I know
perfectly well that we all make simple mistakes: but if the right way
to do something is constantly reinforced, then gradually and eventually
we LEARN, and the mistake becomes extremely rare.

Take addition and multiplication, for example.

How often do you make the mistake of omitting parentheses in THAT case?
I.e. coding "a+b*c" when you actually MEAN "(a+b)*c"?  Pretty rarely, I
suspect -- because the relative priorities of + and * have been drilled
into you for decades, you know DEEPLY that you need parentheses here.

So: the problem is a transient one -- reinforcing the right way to do
it until the neural pathways associated with it dominate the ones that
are associated with the wrong way, giving error messages for the error
as soon as feasible to minimize the cost of the learning.  By LEARNING
here I mean DEEP learning -- the kind that makes errors extremely rare --
not just the superficial kind, limited to the conscious level, where you
have to think and pay deliberate attention to avoid the mistake.


> from the OP:
> 
> if object.isGreen:
>         do something...
> 
> PyChecker could (optional?) warn in such cases if isGreen is a function.

I suspect there may be a very deep misunderstanding here about the
nature of PyChecker -- which works by STATICAL examination of your
module's bytecode -- or else of the degree of difficulty involved in
determining what type "object.isGreen" could possibly BE, in general,
by means of such a statical examination (as opposed to EXECUTING the
code).  To prove that "object.isGreen" is indeed a function, without
actually executing the code but rather by statical examination,
PyChecker would have to prove it can examine ALL pathways the code
could possible have been taken from the execution of the class statement
from which 'object' (what a horrid name choice, duplicating the name
of a built-in type!) is instantiated, or the point of instantiation,
up to this point of execution -- and that in NONE of those paths is it
EVER possible for 'object.isGreen' to have be rebound to any instance
of any different type.  The problem in the general case is obviously
insoluble (it's Turing's so-called "halting problem", or isomorphic to
it).  I very much doubt that the special cases that might perhaps be
resolved with a bounded effort of static examination are worth the
huge amount of development effort.

A vastly more feasible pathway is the one which I proposed, and which
you appear to keep arguing against by stubbornly going back to the
desire of PyChecker (i.e., statical examination) taking care of it --
INSTRUMENT the code, so that when it actually EXECUTES, an exception
is raised at runtime IF the bound-method is evaluated in a boolean
context such as this one.

However, if the very people I'm trying to help here -- those who are
prone to forgetting parentheses when they need to call something
without arguments -- keep stubbornly arguing that the approach I
propose, and which I believe is the only viable one (although very
partial indeed in its support for them), is not what they want -- well,
tough luck... I know when I'm defeated.  If the only (partial) solution
you can possibly HAVE is so irredeemably distant from anything you
would ACCEPT, then there is absolutely nothing either I, or anybody
else, can do for you.  Please accept my apologies for misguidedly
trying to help you guys.


Alex




From ianb at colorstudy.com  Sun Apr  6 14:33:01 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 06 Apr 2003 13:33:01 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <e9570f37.0304060944.79564f88@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
	 <mailman.1049571010.25625.python-list@python.org>
	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
	 <e9570f37.0304060944.79564f88@posting.google.com>
Message-ID: <1049653981.650.49.camel@lothlorien>

On Sun, 2003-04-06 at 12:44, Graham Fawcett wrote:
> Neil Schemenauer <nas-usenet at arctrix.com> wrote in message news:<%5Oja.10776$7w2.9784 at nwrddc01.gnilink.net>...
> > Ian Bicking <ianb at colorstudy.com> wrote:
> > > The next step I can envision would be for Twisted to become the
> > > canonical infrastructure, where higher-level web frameworks are all
> > > built ontop of it.
> > 
> > That's not feasible.  Some applications would require extensive
> > modification to work on top of an event driven framework.
> > 
> >   Neil
> 
> Neil, I may just be suffering from
> daylight-savings-adjustment-syndrome, but what kinds of applications?
> We are talking only about Web application frameworks, which are
> request handlers and are therefore event-driven by definition.

In the Twisted/async model everything that's not really fast has to be
done with events.  So instead of doing:

f = open("somefile")
contents = f.read()
unsafeHTML.sub('', contents)
...

You would do that with events and callbacks.  I can't give a good
example of the translation because I'm not well-versed in that sort of
programming, but you'd compose it as a number of functions, with the
file read calling those functions once it's been completed.

It's pretty weird, and yes it isn't feasable, but luckily you don't have
to use that style to use Twisted (just if you want to program Twisted
protocols).

  Ian






From mcherm at mcherm.com  Thu Apr  3 12:18:46 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Thu,  3 Apr 2003 09:18:46 -0800
Subject: else: w/o except: - why not?
Message-ID: <1049390326.3e8c6cf67d3d2@mcherm.com>

Michael Chermside wrote:
> I think that a "bare except" is useful in only two ways. The first
> is like a "finally" clause that's skipped if everything works.
[...snip...]
> The other situation is running "untrusted plug-in" code of some 
> sort.

David Eppstein replies:
> I have a bare except wrapped around the main event loop of some python 
> gui code.  It provides an opportunity to log the problem and give the 
> user a chance to clean up instead of just immediately crashing.
> Doesn't seem to fit into either of your two uses, but I think it's a 
> legit use.

David Bolen replies:
> In our case [...], a large amount of our Python code
> deals with distributed systems management and communications, and
> there are quite a few cases where we have bare excepts in order to
> catch any exceptions to be packaged and either reported (as part of a
> plugin or dynamic module system as you show above), propagated to some
> client, or just logged and execution resumed.

Good point guys. I think I should add a third senario... that of 
capturing exceptions for logging or changing how they're reported
to the user.

-- Michael Chermside





From roy at panix.com  Sat Apr 12 08:59:43 2003
From: roy at panix.com (Roy Smith)
Date: Sat, 12 Apr 2003 08:59:43 -0400
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>
Message-ID: <roy-6956D3.08594312042003@reader1.panix.com>

Mads Orbesen Troest <mads at troest.NEVERMORE.dk> wrote:
> If I create a class which I want to be able to print like this ...
> 	tst = myclass()
> 	exp = r"test: %(intepol)s";
> 	print exp % { 'intepol':tst }
> 
> ... it seems I have two choices, either of which appear to work. I can 
> overload the __str__ or the __repr__ method.

Most likely, what you want to do is define a __str__ method for your 
class.

The basic concept is that __str__ produces a string the is easy for 
humans to read, and __repr__ produces a string which the python parser 
can read and regenerate the original object.  For complex data 
structures, __repr__ is usually of limited value.



From mhammond at skippinet.com.au  Tue Apr 29 18:59:46 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Tue, 29 Apr 2003 22:59:46 GMT
Subject: simplest question on win32com.client
In-Reply-To: <a5b86892.0304290534.31528b9e@posting.google.com>
References: <a5b86892.0304290534.31528b9e@posting.google.com>
Message-ID: <CtDra.7016$lD4.74951@news-server.bigpond.net.au>

Joe Green wrote:

> I can't seem to get this code to run.
> 
> 
> --
> from win32com.client import Dispatch
> 
> xl = Dispatch("Excel.Application")
> --
> 
> It comes back with:
> 
> The instruction at "0x1e039e79" referenced memory at "0x1e3840fc". The
> memory could not be "read".
> 
> Click OK to terminate the program
> Click on CANCEL to debug the program

I've seen this happen when an old state excel process is hanging around. 
  Check using the Task Manager that there is no Excel currently running.

Mark.




From seppi at seppi.de  Mon Apr 14 10:16:52 2003
From: seppi at seppi.de (Sebastian Wilhelmi)
Date: 14 Apr 2003 07:16:52 -0700
Subject: accessing a functions var from a subfunction.
Message-ID: <a3cae241.0304140616.1833e5c4@posting.google.com>

Hi,

I would like to do the following:

-------8<-------8<-------8<-------8<-------
def test ():
    count = 0
    def inc_count ():
        count += 1
    inc_count ()
    inc_count ()
    print count

test ()
-------8<-------8<-------8<-------8<-------

This doesn't work (and I even understand, why ;-)

One solution is the following, which I however do not see as very
clean or nice.

-------8<-------8<-------8<-------8<-------
def test ():
    count = [0]
    def inc_count ():
        count[0] += 1
    inc_count ()
    inc_count ()
    print count[0]

test ()
-------8<-------8<-------8<-------8<-------

Now my question: Is there some way to achieve this with a nicer
syntax?
Using 'global' would not count, as that would make a variable
unnecessarily known to the outside.

Thanks,
Sebastian



From tuco at interfree.it  Mon Apr 14 17:18:19 2003
From: tuco at interfree.it (Nicola Mingotti)
Date: Mon, 14 Apr 2003 21:18:19 +0000 (UTC)
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com>
Message-ID: <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org>

"sdhyok" <sdhyok at yahoo.com> wrote in message
news:420ced91.0304141131.73a6206e at posting.google.com

> I am looking for a way to express a date in a real number
> in Python.
> 
> For instance, in MatLab,
> n = datenum(2001,12,19,18,0,0) returns n = 731204.75.
> 
> This kind of expression turned out to be very helpful
> when handling or storing dates.
> 
> Is there any funtion for the purpose in Python (e.g. in datetime module)?
> 
> Daehyok Shin


import time
t = time.time()
# t is the number you asked for 

If you want to know more about this look at 
the "time" module .


bye.





-- 
Posted via Mailgate.ORG Server - http://www.Mailgate.ORG



From oren-py-l at hishome.net  Thu Apr 10 03:23:59 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Thu, 10 Apr 2003 03:23:59 -0400
Subject: Python/C API
In-Reply-To: <20030410063413.GA13598@meson.dyndns.org>
References: <20030410063413.GA13598@meson.dyndns.org>
Message-ID: <20030410072359.GA64924@hishome.net>

On Thu, Apr 10, 2003 at 02:34:13AM -0400, Jp Calderone wrote:
>   How does one write generators in C, or does one not?
> 

You can write something similar to Python generators using Duff's Device.

#define GEN_BEGIN(switch(self->state) { case 0:
#define GEN_END case -1: } return NULL;
#define GEN_YIELD(value) self.state = __LINE__; return (value); case __LINE__:
#define GEN_RETURN(value) self.state = -1; return (value);

PyObject* myobject_next(myobject *self)
{
    GEN_BEGIN
    while(...) {
         ...
         GEN_YIELD(...)
         ...
         if(...) {
             ...
             GEN_YIELD(...)
             ...
             if(...) {
                 GEN_RETURN(...)
             } 
         }
         ...
    }
    GEN_END
}

Local variables will not be preserved after yielding, so all state data
must be stored in the instance.

    Oren





From news at 92tr.freeserve.co.uk  Sun Apr  6 12:06:00 2003
From: news at 92tr.freeserve.co.uk (ddoc)
Date: Sun, 6 Apr 2003 16:06:00 +0000
Subject: "Automation" of OpenOffice apps via Python?
References: <mailman.1049616555.27075.python-list@python.org> <8351bb33.0304060549.51f7e8a5@posting.google.com>
Message-ID: <b6pj9s$qec$1@newsg1.svr.pol.co.uk>

THis or something near it might also be expcted to be useful.  Maybe.
http://udk.openoffice.org/common/man/tutorial/office_automation.html 
How to do OLE automation on Star Office and Open Office

> I have been meaninig  to try this out but haven't got to it yet. I'm
> sure other people here would be interested to learn of your
> experiences with python-bridge

I shall be very interested to hear of your experiences with Python-bridge 
or anything else you use for this purpose.

-- 
A



From jack at performancedrivers.com  Tue Apr 29 18:38:47 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Tue, 29 Apr 2003 18:38:47 -0400
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 29)
In-Reply-To: <20030429143703.C1336@tibia.amotlpaa.bogus>; from staschuk@telusplanet.net on Tue, Apr 29, 2003 at 02:37:04PM -0600
References: <vatflcjn7dum95@corp.supernews.com> <20030429143703.C1336@tibia.amotlpaa.bogus>
Message-ID: <20030429183847.A1050@localhost.localdomain>

On Tue, Apr 29, 2003 at 02:37:04PM -0600, Steven Taschuk wrote:
> Quoth I:
>   [...]
> >     Jack Diederich exhibits two real-life uses for metaclasses:
> >     registering classes with a factory, and named prototypes.
> >         <http://groups.google.com/groups?th=296a0181247593f3>
> 
> It looks like Google has lost this thread since I assembled the
> list of links.  (In fact, I can't find on Google any post to
> c.l.py from 27 April...)  The other Google groups links in the
> previous list are still good, as of the time of this post.
> 
> Jack Diederich's post can be found in the python.org archives:
>     <http://mail.python.org/pipermail/python-list/2003-April/160318.html>

1) I put an augmented and reformatted version on my web site, includes feedback
   from c.l.py

   http://jackdied.com/python/

2) What happened on the 27th to cause a blind spot in google?
   Best conspiracy theory wins a cookie.

-jackdied




From cnetzer at mail.arc.nasa.gov  Wed Apr  9 16:42:06 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 09 Apr 2003 13:42:06 -0700
Subject: atexit functionality
In-Reply-To: <1049915746.3e947162ca34b@webmail.telerama.com>
References: <mailman.1049821814.20582.python-list@python.org>
	 <10074955979500@aluminium.rcp.co.uk>
	 <1049915746.3e947162ca34b@webmail.telerama.com>
Message-ID: <1049920925.601.34.camel@sayge.arc.nasa.gov>

On Wed, 2003-04-09 at 12:15, viszneki at telerama.com wrote:
>  Gee I'm such a jerk.

Relax.  Let it slide.

> 
> If anyone has anything positive to say about my idea, please say so.

I think Duncan's suggestion about maintaining your own list, was a very
good one.

>  In my last 
> email to this mailing list, I posted some hypothetical code to do the job:
> 
> return _exithandlers[len(_exithandlers)]

This will fail by definition, since indexing begins at 0.  You probably
meant:

return _exithandlers[len(_exithandlers) - 1]

or

return _exithandlers[-1]


> Shortly after having sent it, I realized it wouldn't work because of the way 
> Python does things

Hence, Duncan's suggestion, and use of the remove() method of lists.

However, my view is that if you have objects that can "cleaned up"
before their reference count goes to zero (ie. other things can still
refer to it, and potentially call its methods), then you have a design
problem and need to reconsider what you want.  Or you want to make use
of existing facilities that Python already provides (see below).

> I tried this myself, by adding that line of code to the end of 
> atexit.register, just to make sure, because I am admittedly no master at 
> pointer dynamics in Python (something I wish I knew more about.)

Well, first, don't think in terms of "pointers".  Objects all have
lifetimes based on references, which are just a name to refer to an
object.  You can decide to eliminate a reference to an object, but you
can't (generally) decide to destroy an object at any specific time
unless you know you control all the references to an object (ie, such as
lists, dictionaries, etc.)

With regards to your cleanup idea, if you want to allow people to call
cleanup() functions, and thus free up and invalidate object resources,
before the lifetime of the object is ended (ie. being able to close
files or remove references to other objects), then you should generally
protect your member functions from using those freed up resources after
the cleanup has occured.  This could be done by having a "cleanedup" or
"invalid" flag on the object, which is set True when an early cleanup
has occurred.  this could be a lot of extra work, for little gain,
however.

Perhaps you should investigate the "weakref" standard module, which
gives you a way to refer to another object without increasing it's
reference count (and thus potentially preventing it from "cleaning up"
when it otherwise could.)  So, you could hand out references to an
object, and still allow it to be cleaned up only after all non-weak
references have been dropped (ie. avoiding early cleanup).  Users of
weakrefs will generally be forced to always check the reference before
using it anyway, which formalizes the usage I outlined above.


> Thank you everybody for your time,
> Donny Viszneki.

-- 
Bay Area Python Interest Group - http://www.baypiggies.net/

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From lost at soul.net  Fri Apr 25 18:39:14 2003
From: lost at soul.net (lost)
Date: Fri, 25 Apr 2003 23:39:14 +0100
Subject: Dealing with files...
Message-ID: <pan.2003.04.25.22.39.14.432753@soul.net>

Hi guys!
I'm starting to deal with files, and need to know how to deal directly
with the filesystem. what I'm looking for, is similar to this pseudo code:

for each_file in "/var/path/to/files/":
	print each_file #show the NAME of the file 
	write the NAME of the file	into file_list.txt
print "done"

This would print the name of each file in /var/path/to/files (for example,
a list of html files), and then add the name of the file into
file_list.txt.

Any help on this would be very much appreciated! Thanks.



From dialtone at despammed.com  Mon Apr 21 15:50:13 2003
From: dialtone at despammed.com (Dialtone)
Date: 21 Apr 2003 21:50:13 +0200
Subject: coverting list back to string
References: <3cf72d52.0304211132.799d8afb@posting.google.com>
Message-ID: <878yu33a16.fsf@vercingetorix.vercingetorix>

regnivon at netscape.net (scn) writes:

> my quest:  i know that output.write(remove) is not correct.  you
> cannot use the 'write' method on a list.  is there a better way to
> convert the list to a string after the first element of the list has
> been removed and then output the strings to a file?  i'm a newbie
> programmer and need guru help. thank you.

Sure:

input = file('c:/python22/programs/linetest.txt', 'r')
output = file('c:/python22/programs/output.txt', 'w')

for line in input:
    remove = line.split()
    remove = remove[1:]
    output.write(''.join(remove))

input.close()
output.close()


-- 
try: troll.uses(Brain)
except TypeError, data: 
   troll.plonk()
Co-Responsabile Hardware Information & Technology http://hit.edengames.net



From exarkun at intarweb.us  Wed Apr  2 01:36:49 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 01:36:49 -0500
Subject: nntplib, huge xover object
In-Reply-To: <ZCidndFDJdNGwBejXTWc3g@giganews.com>
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com>
Message-ID: <20030402063649.GB4169@meson.dyndns.org>

On Tue, Apr 01, 2003 at 11:52:52PM -0500, David Sfiligoi wrote:
> Hi,
> 
> I built a small script that use the xover function in the nntplib module.   
> The problem that I came across is that what xover can return a huge tuple 
> when there is 1000s of article in a newsgroup(which is frequent)

  There are a couple options.

  One is to use a library that doesn't store the entire result of the
protocol action in memory, or at least gives you the option not to.  This is
a good solution, except it probably requires that you not use the standard
nntplib, so it is a bit of extra work.

  Another is to limit the number of articles returned by xover.  This is
done very easily, by simply specifying sane values for start and end, and
making repeated calls until the entire dataset has been retrieved,
processing or dumping to secondary storage whatever intermediate results you
receive.  This is all accomplished simply by selecting appropriate values
for "start" and "end" in your code example.

> 
> testxover_resp,testxover_subs = s.xover(start,end)
> 

  Jp

-- 
"The problem is, of course, that not only is economics bankrupt but it has
 always been nothing more than politics in disguise ... economics is a form
 of brain damage."  -- Hazel Henderson
-- 
 up 13 days, 2:00, 3 users, load average: 1.00, 1.00, 1.03




From b.maryniuk at forbis.lt  Thu Apr 24 12:43:04 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Thu, 24 Apr 2003 18:43:04 +0200
Subject: What means exactly "Memory error"?
In-Reply-To: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net>
References: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net>
Message-ID: <200304241843.04939.b.maryniuk@forbis.lt>

On Thursday 24 April 2003 17:26, Tim Peters wrote:
> [Bo M. Maryniuck. gets a traceback ending with
>     File "/usr/lib/python2.2/site-packages/_xmlplus/dom/Event.py",
>         line 93, in initEvent
>     self.bubbles = canBubbleArg
>     MemoryError
> ]
> I can't help with your app, but I can tell you what MemoryError means. 

[skipp]

> The most likely cause is that you've
> simply run out of memory (#1 above) -- but that's more a symptom than an
> explanation.  Don't expect it to be easy to resolve.

Yes, you're right: I'm just out of memory, though server has 4G RAM... :/ Have 
you at least any ideas how to decrease/pack collected data to avoid fat 
growing?

P.S. Or at least which library I should use to parse HTML with DOM using 
something else, instead HtmlLib...?

-- 
Regards, Bogdan

99 little bugs in the code, 99 bugs in the code,
fix one bug, compile it again...
101 little bugs in the code....





From sholden at holdenweb.com  Fri Apr 18 12:30:57 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Fri, 18 Apr 2003 16:30:57 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com> <b7kvuh$20aem$1@ID-169208.news.dfncis.de> <3EA03AFF.1030309@removethis.free.fr>
Message-ID: <5LVna.126512$It5.11343@news2.central.cox.net>

"laotseu" <bdesth at removethis.free.fr> wrote ...
> Greg Ewing (using news.cis.dfn.de) wrote:
> > Tj wrote:
> >
> >> Lambda could also stand renaming;
> >> Pythonistas probably think it's an arrogant name.  (Or exotic?  I
> >> can't tell.)
> >
> >
> > Actually, we think it's kind of cute.
> >
> >   Mary had a little lambda,
> >   Its syntax white as snow,
> >   And every program Mary wrote,
> >   She wrote in Lisp, you know.
>
> lol !-)
>
This *finally* gives me the chance to ask: when a lambda grows up, does it
become a sheepda?

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From giles_brown at hotmail.com  Wed Apr  9 08:33:31 2003
From: giles_brown at hotmail.com (Giles Brown)
Date: 9 Apr 2003 05:33:31 -0700
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <57de9986.0304090433.64753952@posting.google.com>

iamshady at rediffmail.com (Jim Shady) wrote in message news:<9afd3f36.0304082245.33670341 at posting.google.com>...
> Hello,
> 
> I am a beginner at Python and I just joined this group.
> 
> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string. For
> example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?

You can split the string into a list of (length, substring) tuples.
The 'max' of these tuples will be the longest substring.

Python 2.2.1 (#34, Apr  9 2002, 19:34:33) [MSC 32 bit (Intel)] on win32
Type "copyright", "credits" or "license" for more information.
IDLE 0.8 -- press F1 for help
>>> def longest(s, sep='/'):
	return max([(len(sub), sub) for sub in s.split(sep)])[1]

>>> longest('abcd/df/a/iiwk/abcdefghijkl/b/c')
'abcdefghijkl'
>>> 

Giles



From peter at engcorp.com  Tue Apr 15 20:50:11 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 15 Apr 2003 20:50:11 -0400
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org>
Message-ID: <3E9CA8C3.71C2819E@engcorp.com>

Dan Jacobson wrote:
> 
> I need to use one liners every day, especially in Makefiles.  I thus
> need to know how to make any kind of python program into a big long one
> liner. However, right out of the starting gates I run into
> $ python -c 'print 2;for i in (1,4):print i'
>   File "<string>", line 1
>     print 2;for i in (1,4):print i
>               ^
> SyntaxError: invalid syntax
> 
> Help. BTW tell me what to do in all the other cases when ";" can't be
> a replacement for newline.  I.e. does python's nifty indentation rules
> spell the death of one-liners?
> 
> BTW, how do I import a module from the command line when using -c? Is
> there an option switch for importing modules?
> 
> This makes no output and returns no error value to the shell:
> $ python -c 'import math' -c 'for i in range(200,500,50): print
>   i,360/2/math.pi*math.asin(i/1238.4)'

I think you're looking for this (all on one line, with a space instead
of the line break):

C:\>python -c "exec '''import math\nfor i in range(200,500,50):\n print 
i,360/2/math.pi*math.asin(i/1238.4)\n'''"

200 9.29389730029
250 11.6465309088
300 14.0192608728
350 16.4168005237
400 18.8442931045
450 21.3074328157

-Peter



From karam_chand02 at yahoo.com  Wed Apr 16 04:05:30 2003
From: karam_chand02 at yahoo.com (Karam Chand)
Date: 16 Apr 2003 01:05:30 -0700
Subject: Embedding Python Engine into a windows executable
Message-ID: <433b8cf0.0304160005.651339b2@posting.google.com>

Hello,

Is is possible to embed Python engine completely into a windows
executable? Basically, I want to give some kind of python scripting to
allow users extend my applicaion and I dont want to ship python*.dll
separately.

Regards,

Karam



From luc at nospam.com  Sat Apr 26 17:39:45 2003
From: luc at nospam.com (luc wastiaux)
Date: Sat, 26 Apr 2003 23:39:45 +0200
Subject: please comment on technologies
References: <slrnbalplq.254.luc@trillian.dont-panic.info> <HqCqa.8931$3M4.241359@news1.tin.it>
Message-ID: <slrnbalv50.254.luc@trillian.dont-panic.info>

In article <HqCqa.8931$3M4.241359 at news1.tin.it>, Alex Martelli wrote:
>> The project would be developped by a group of 4 to 6 CS students
>> (including me), over a three months timespan (but not full time, we have
>> to go to classes and other stuff)
> 
> The most valuable thing you'll learn will be how to structure your
> work.  I heartily recommend incremental development (most crucial
> features first), test-driven design, pair programming, and a healthy
> diffidence towards "big design up front" - you may get the temptation
> of designing in some big chunks of infrastructure preparing for future
> enhancements... that may never come.
> 

In short, you recommend following extreme programming guidelines, did I 
understand your advice correctly ?


>> Spyce today and it looks like I could use it for the web interface.
> 
> Sure, that's one possibility.  Personally I'd choose Webware, which
> offers an amount of flexibility I appreciate, but Spyce's OK too.

I will have a look a webware. I was looking forward to using spyce because 
it features <?php style tags to embed code into HTML templates, does 
webware have that ?
I just know webware from this url: 
http://colorstudy.com/docs/shootout.html

> 
>> Then I have this idea about using XSL for templating, the idea is not only
>> to separate the data from the representation, but also allow programmers
>> to "leech off" the search engine (it's easier to do this if the output is
>> XML).
> 
> The ability for other programmers to send "queries" to your search
> engine and get XML output is good.  However, personally, I would avoid
> XSL (unless you guys are already well familiar with it -- or NEED to
> get very familiar with it anyway to pursue your studies).  It's simpler
> to have separate URLs, or a form field requesting XML or HTML output,
> and just use different templates with the same contents -- doing your
> templating in Cheetah (or preppy/Spyce/whatever).

Ok, I thought XSL was the definitive answer to all templating needs, but 
you don't seem very excited about it ?

thanks a lot for all your comments and advice!

-- 
luc wastiaux
 $ finger luc at info.4002.org



From rstephens at vectron.com  Thu Apr 24 09:16:45 2003
From: rstephens at vectron.com (Ron Stephens)
Date: 24 Apr 2003 06:16:45 -0700
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org> <QxAoa.540832$3D1.296696@sccrnsc01> <a5415304.0304210544.1342b436@posting.google.com> <3EA490F5.30606@removethis.free.fr> <slrnbaf8vd.gb9.ramen@lackingtalent.com>
Message-ID: <a5415304.0304240516.1b305131@posting.google.com>

My impression is that PythonCard has some similarities to the old
Macintosh HyperCard development tool, in that it uses a Resource File
Editor to set up the windows and widgets, and in that some of the key
PythonCard developers are familiar with HyperCard and like some of
HyperCard's features.

However, PythonCard is entirely new technology and is being built from
the ground up, and the differences from HyperCard are so great that
there has been much debate as some wished to change the name so as to
avoid confusion. Still, I love the name PythonCard and evidently it is
here to stay.

Really, PythonCard is a work of love by Kevin Altis and several other
developers who are putting their best original thinking into its
architecture. It is innovative, and truly a collaborative project. If
it uses some good ideas that were also used in HyperCard, that's all
to the better ;-))))

If you subscribe to the PythonCard mailing list, you will be able to
follow some interesting collaborative development.



From menno at netbox.biz  Thu Apr  3 01:35:50 2003
From: menno at netbox.biz (Menno Smits)
Date: Thu, 3 Apr 2003 16:35:50 +1000
Subject: quopri module with large files
Message-ID: <20030403163550.4bb97cc7.menno@netbox.biz>

decode() in the standard quopri module loads the entire file being
decoded into memory and sends the result to binascii.a2b_qp().  This
can potentially lead to memory exhaustion problems if the file being
decoded is large. 

Can anyone see a problem with replacing the offending code in decode()
which looks like this:

    if a2b_qp is not None:
        data = input.read()
        odata = a2b_qp(data, header = header)
        output.write(odata)
        return

with this? :

    if a2b_qp is not None:
        while 1:
            line = input.readline()
            if not line:
                break
            output.write(a2b_qp(line, header=header))

This would be more efficient memory-wise when it comes to large email
attachments. The base64 module takes a similar approach.

I've compared the two implementations using a few test cases and they
both appear to give identical output. However, maybe someone can see a
problem with the line-by-line approach that I'm not seeing. A brief
check of RFC 2045 indicates there shouldn't be a problem.

I'm looking at the Python 2.2.2 implementation BTW.

-- 
Menno Smits
Senior Development Engineer
NetBox | Oxcoda
E-mail: menno at netbox.biz
Web: netbox.biz




From gtalvola at nameconnector.com  Tue Apr  8 11:31:35 2003
From: gtalvola at nameconnector.com (Geoffrey Talvola)
Date: Tue, 8 Apr 2003 11:31:35 -0400 
Subject: [ANN?] Boston PIGs
Message-ID: <61957B071FF421419E567A28A45C7FE59AF39D@mailbox.nameconnector.com>

Jack Diederich [mailto:jack at performancedrivers.com] wrote:
> Would anyone be interested in attending a regular Boston 
> Python meeting?
 
I would be interested.

- Geoff




From aahz at pythoncraft.com  Sun Apr 27 11:35:30 2003
From: aahz at pythoncraft.com (Aahz)
Date: 27 Apr 2003 11:35:30 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <b8fsvk$kkg$1@panix2.panix.com> <1ELqa.9765$3M4.264114@news1.tin.it>
Message-ID: <b8gtc2$j9e$1@panix2.panix.com>

In article <1ELqa.9765$3M4.264114 at news1.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>Aahz wrote:
>>
>> Anyway, all objects that do not implement a __nonzero__ method are true:
>
>...unless they implement __len__, which is used as a fallback if it's
>present but __nonzero__ is absent (I'm sure Aahz know this, I'm just
>dotting t's and crossing i's for other readers who might not).

Right.  Thank you, I was in a hurry.  (I'm amused to see that someone
else uses the "cross i's, dot t's" expression.)
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From duncan at NOSPAMrcp.co.uk  Wed Apr 23 04:45:06 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 23 Apr 2003 08:45:06 +0000 (UTC)
Subject: capturing output from os.system() in Windows?
References: <f405d9e7.0304230003.63baab05@posting.google.com>
Message-ID: <Xns9366626B53C4Eduncanrcpcouk@127.0.0.1>

gabnash at hotmail.com (David) wrote in 
news:f405d9e7.0304230003.63baab05 at posting.google.com:

> Being a complete novice at python, I thought I'd try a simple script
> to run a NET VIEW (which runs in a DOS box) command under Windows and
> then manipulate the resulting data.
> 
> The first problem I have encountered is how to capture the output from
> the 'NET VIEW' command.
> 
> os.system('NET VIEW') returns 0 under windows, although the DOS box
> pops up and the NET VIEW executes correctly.
> 
> Is there some other command I should be using instead? I guess an
> alternative might be to run the 'NET VIEW' under dos and redirect the
> output into a python script using a BAT file?

Us os.popen, os.popen2, os.popen3 or os.popen4 according to your exact 
needs.
e.g. lines = os.popen('net view').readlines()

N.B. Be careful if you try to separate stdout and stderr from the command 
you are running as you can get deadlocks if a pipe buffer fills up while 
you are trying to read from the other pipe.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From cce at clarkevans.com  Thu Apr 10 03:01:23 2003
From: cce at clarkevans.com (Clark C. Evans)
Date: Thu, 10 Apr 2003 07:01:23 +0000
Subject: Python/C API
In-Reply-To: <20030410063413.GA13598@meson.dyndns.org>
References: <20030410063413.GA13598@meson.dyndns.org>
Message-ID: <20030410070123.GA6652@doublegemini.com>

On Thu, Apr 10, 2003 at 02:34:13AM -0400, Jp Calderone wrote:
| How does one write generators in C, or does one not?

You basically implement your own secondary call stack.  Instead of 
calling a function directly, you call a procedure which puts your
function on the secondary call stack and then executes it.  Then,
when the function returns you don't pop the secondary call stack.
For each generator, you get another call stack, etc.   

Clark

P.S. I'm (in the background cycles) working on a yaml parser
in C, which uses 'generator' technique for parsing.  I only
have a crude call stack impl, if someone has a good one...




From news at yebu.de  Tue Apr 15 03:24:18 2003
From: news at yebu.de (Karl Scalet)
Date: Tue, 15 Apr 2003 09:24:18 +0200
Subject: vim configuration help
In-Reply-To: <P4Oma.78974$rd4.2975295@twister.austin.rr.com>
References: <P4Oma.78974$rd4.2975295@twister.austin.rr.com>
Message-ID: <b7gc3g$lb54$1@ID-141451.news.dfncis.de>

Hi Randall,

Randall Smith wrote:
> I'm learning Python and I've been looking for a good IDE to use with it. 
>    I haven't found an IDE that has text editing capabilities close to 
> Vim/Emacs.
> 
> Much of my previous programming is with Emacs/PHP and I love the 
> auto-indenting, tab-indenting, well thought-out key strokes, and the 
> ability to ssh into a server and run Emacs.
> 
> I'd like to try out Vim while writing Python code (I'm using Webware). 
> Right now, the default settings using color coding, have an 8 space tab 
> and no autoindenting.  I would like to have a tab space of 2 or 4 spaces 
> and have auto-indenting.
> 
> I have no idea how to configure Vim.  I'm a total newbie.  Can someone 
> please help me get started.  I'm looking to configure Vim for Python. 
> Autoindenting and 4 space tabs are high priority now.

looking in my vimrc file:

set autoindent
set tabstop=4
set shiftwidth=4
set expandtab

could give you a start

Karl




From danb_83 at yahoo.com  Sat Apr  5 20:38:59 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 5 Apr 2003 17:38:59 -0800
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
Message-ID: <ad052e5c.0304051738.387077bb@posting.google.com>

Simon Faulkner <news at titanic.co.uk> wrote in message news:<1lau8vcctfn4g2pi9k5df3km7qgg5r72sv at 4ax.com>...
> I have a list of about 5000 numbers in a text file - up to 14 digits
> each.
> 
> I need to check for duplicates.
> 
> What would people suggest as a good method?

Sort the list, then check whether any 2 consecutive numbers are equal.



From antonmuhin at sendmail.ru  Thu Apr 17 04:52:18 2003
From: antonmuhin at sendmail.ru (Anton Muhin)
Date: Thu, 17 Apr 2003 12:52:18 +0400
Subject: do python's nifty indentation rules spell the death of one-liners?
In-Reply-To: <mailman.1050548909.29232.python-list@python.org>
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org> <1050544015.99640.1@iris.uk.clara.net> <mailman.1050548909.29232.python-list@python.org>
Message-ID: <b7lq06$26ot$1@news.peterlink.ru>

Evan Simpson wrote:
> Garry Knight wrote:
> 
>> Maybe I didn't completely understand what the OP meant, but this isn't a
>> one-liner. It's a five-liner.
> 
> 
> python -c """exec('''print 2$$for i in (1,4):$$  print 
> i$$'''.replace('$$','\n'))"""
> 
> Yech,
> 
> evan @ 4-am
> 
> 
> 


Sorry for posting it as a reply to Evan's message...

Back to OP's problem: I'd personally prefer to write a small .py file 
and reduce all the one-liners to calls of functions from this module.

hth,
Anton.




From sholden at holdenweb.com  Thu Apr 17 10:49:33 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 14:49:33 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <87k7dt5m3j.fsf@interim.henrik-motakef.de>
Message-ID: <1azna.115402$It5.22896@news2.central.cox.net>

"Henrik Motakef" <henrik.motakef at web.de> wrote ...
> laotseu <bdesth at removethis.free.fr> writes:
>
> > Lispers friends, let's have you're opinion on that point : is Common
> > Lisp a functionnal language ?
>
> Common Lisp is whatever you want it to be: You can write
> functional-style programs in it if you want, but is certainly not a
> "pure" functional language. It supports several other paradigms
> equally well, and even more important, it will let you integrate
> yet-unkown paradigms when you think that they are useful.
>
> There are really ideas to remove the functional stuff from Python? I
> wonder why people frequently think that making things impossible will
> improve programming languages.
>
> just-because-it's-flexible-doesn't-mean-it's-perl-ly y'rs


I can see that, but wouldn't you agree that removing redundant features also
removes cruft from the source, and eases the maintenance burden, in turn
reducing the likelihood of implementation bugs? While Guido is known to
regret the inclusion of lambda in Python, I don't think he would want to
take it out on purely ideological grounds. Indeed, the suggestion was not so
much that it would be taken out of the existing language, rather that it
would not be implemented in some "futurePython", normally called Python 3 or
Python 3000.

It would be nice to have One True Python, but that's about as likely as One
True Lisp. The Python world is pragmatic enough to be able to live with "one
*most obvious* way to do it", but if there are two similar features in the
language they have to be maintained in lockstep as the language develops.

just-becuase-it's-there-doesn't-mean-it's-helpful-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From michael at stroeder.com  Thu Apr  3 19:30:38 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Fri, 04 Apr 2003 02:30:38 +0200
Subject: ANN: python-ldap-2.0.0pre07
Message-ID: <mailman.1049583241.7523.clpa-moderators@python.org>

HI!

Find a new pre-release of python-ldap:

   http://python-ldap.sourceforge.net/

python-ldap provides an object-oriented API to access LDAP directory
servers from Python programs. It mainly wraps the OpenLDAP 2.x libs for
that purpose. Additionally it contains modules for other LDAP-related
stuff (e.g. processing LDIF, LDAPURLs and LDAPv3 schema).

Download link:
http://prdownloads.sourceforge.net/python-ldap/python-ldap-2.0.0pre07.tar.gz?download

Changes since 2.0.0pre06:

LDAPObject.c:
   * Wrapped OpenLDAP's ldap_search_ext()
   * Removed empty __doc__ strings
   * Removed fileno
   * Removed all stuff related to caching in OpenLDAP libs

ldap.ldapobject:
   * Fixed SASL rebind in ldap.ldapobject.ReconnectLDAPObject
   * use search_ext() instead ldap_search()
   * new class attribute timeout for setting a global time-out
     value for all synchronous operations

ldap.schema:
* Fixed two typos in ldap.schema.models
* Some attempts to improve performance of parser/tokenizer
* Completely reworked to have separate OID dictionaries for
   the different schema element classes
* Fixed the Demo/schema*.py to reflect changes to ldap.schema

Documentation updates and various __doc__ string modifications.

ldapurl:
   * Removed all Unicode stuff from module ldapurl
   * Consistent URL encoding in module ldapurl

ldif:
   * Removed ldif.FileWriter
   * Proper handling of FILL (see RFC 2849)





From danra at walla.co.il  Wed Apr 16 09:24:15 2003
From: danra at walla.co.il (Danra)
Date: 16 Apr 2003 06:24:15 -0700
Subject: Typing arguments in python
Message-ID: <fd1c664e.0304160524.295bbdc1@posting.google.com>

Here's an idea I've been playing with for a syntax for type checking
in python. Instead of trying to explain it, I'll just start off with
an example:

Let's say I want to pass a file object to function f. I could to one
of the following definitions:

def f(fileobj):
    .
    .
    fileobj.write('...')
    .
    .

Or:

def f(fileobj):
    if not hasattr(fileobj,'write'):
        raise TypeError, "Given arg must have the 'write' attribute."
    .
    .
    fileobj.write('...')
    .
    .

The advantages of the second definition are obvious - The fact that
the argument passed doesn't have the write method would be detected
immediately, rather than when trying to call the it.

Why is that so important? Well, for starters in many cases it isn't.
In a simple script or application it really is sufficient to get the
exception only when calling write.

But - what if:

1) The stuff you do before calling write takes a lot of time, which
could be saved if the error is detected when the function is called.
2) The object gets passed around from function to function, and maybe
changes its name on the way. In this case it would be more difficult
to trace the problem from the errorneous call to the original function
call.

Now, once the advantage of the second approach have been verified,
what if in my application, I have a LOT of functions which pass around
file objects? Should I copy the write-attribute checking routine to
all of them? If I have to check a lot of  attributes for a lot of
objects in each function, it becomes a bit of a nuisance, and less
readable.

So the obvious solution is to write a function:

def verify_writable(x):
    if not hasattr(x,'write'):
        raise TypeError,...

and then we could use:

def f(fileobj):
    verify_writable(fileobj)
    .
    .

But it would be more readable if we could use the syntax:

def writable(x):
    if not hasattr(x,'write'):
        raise TypeError,...

def f(writable fileobj):
    .
    .

and also

def f( (readable,writable) fileobj ):
    .
    .

or for value checking:

def f( positive x ):
    .
    .

Of course using such 'typing' for an argument would be permissable,
not mandatory. Basically cause it would mostly be used in big
applications, not just for your everyday script.

To summarize - why do I think such syntax is good?
Because in addition to the advantages of type and value checking
before actually starting to do the work in the function, which could
be done using available syntax - by just calling such checking
functions in the beginning of your function - you get more readable
and obvious code, mostly because such value and type checking *belong*
in the argument list, not inside the function body, IMHO.

P.S. I am not a expert on python or anything of the sort. It could be
that all of the above is just balderdash in your opinions, in which
case feel free to say so :)



From a.schmolck at gmx.net  Tue Apr  8 12:53:45 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 08 Apr 2003 17:53:45 +0100
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E9267F6.9020605@kyborg.dk>
Message-ID: <yfsu1d9hqw6.fsf@black132.ex.ac.uk>

Kim Petersen <kp at kyborg.dk> writes:

> Alexander Schmolck wrote:
> > Kim Petersen <kp at kyborg.dk> writes:
>  >> (snip)
> >>function in elisp - do a shallow copy of local variable space (dictionary) and
> >>send it as the local vars - you can safely throw away the dict after function
> >>returns).
> > Ahem, there are several thousand variables kicking about in global in emacs,
> > you want to copy them all for each function call?
> 
> look above - *Local* variable space ... globals stay globals.

Maybe we have a misunderstanding. 

I thought in your scheme every call to a lisp function foo (from python *or*
another lisp function) is to be implemented by making a copy of the locals and
passing it to foo and this copy is discarded after return. If so, this won't
work (because foo can mess with its callers *locals*; also globals *don't*
stay globals -- a let binding of a global name means it can no longer be
globally modified further up the call tree).

A better example to clarify:

(defun foo()
  (message "when foo is called folding is: %s" case-fold-search)
  (setq case-fold-search :evil))
(defun bar()
  (message "when bar is called folding is: %s" case-fold-search)
  (let ((case-fold-search t))
    (foo)
    (message "what did foo do to folding? %s!" case-fold-search))
  (message "bar's last folding is: %s again" case-fold-search))
(defun quux()
  (message "before bar:%s" case-fold-search)
  (bar)
  (message "after bar:%s" case-fold-search))
(quux)


before bar:nil
when bar is called folding is: nil
when foo is called folding is: t
what did foo do to folding? :evil! ;; !!!
bar's last folding is: t again 
after bar:nil

Tell me if you think this would work, too and I'll just reread your message.


'as



From mgerrans at mindspring.com  Thu Apr  3 01:08:48 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Wed, 2 Apr 2003 22:08:48 -0800
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org>        <b6fj57$626$1@news-reader11.wanadoo.fr>        <3E8B53BA.DBB0C84C@engcorp.com> <mailman.1049319846.7316.python-list@python.org> <3E8B67AD.7AD74C0C@engcorp.com>
Message-ID: <b6gj7b$huv$1@slb6.atl.mindspring.net>

I wanted to release some Python scripts that would be on a very large number
of systems and had to be reliable.   I wanted to have have at least a little
protection from tinkering, so used this little compiler:

# pycompile.py
if __name__=='__main__':
   import os
   exec( 'import %s' %
os.path.splitext(os.path.basename(os.sys.argv[1]))[0])

It was run as part of a release script which put together the package as
well as updating version stamps and whatnot.

Later I added a little MD5 self-check of the whole package (many Python and
other files), so the Python scripts can be included, but the first thing I
can ask someone when they report that my stuff "doesn't work" is to look at
a log file and see if it says anything about failing a self-integrity check.
That lets me know whether I'm the guilty party or they are, which is a good
starting point.  ;-)

By the way, this is the kind of script you only use on an internal project
that you use yourself, because it uses exec() on a command line parameter
(probably a bad idea in most cases) and has no error handling.   It could
also be improved to use glob(), for instance, if you want to compile many
scripts in one swell foop.





From peter at engcorp.com  Sat Apr 19 20:20:06 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sat, 19 Apr 2003 20:20:06 -0400
Subject: QuickTax uses Python 2.1
Message-ID: <3EA1E7B6.17959CD3@engcorp.com>

I was frustrated for a while by an obscure problem when I tried
running my copy of QuickTax.  It was working a couple of weeks
ago, but today it popped up a dialog indicating a problem with a
resource "c:\a\quicktax2002\QT2002\UserDict.pyc".

Huh?  What's it doing with a .pyc file?  I look in the install
directory and see a handful of files that look suspicious 
including _sre.pyd, python21.dll, and some others.  But no
UserDict.pyc.  Well, why should there be?  I didn't mean to
install Python 2.1 there.  Grumble grumble Windows grumble.

Okay... re-install QuickTax.  Check directory.  UserDict.pyc!
Ah... the light dawns (finally).  Along with interesting files 
like gherkin.pyc (obviously a pickle variant :-), qtcalcengine.pyc,
tebase.pyc, tedata.pyc, tecmd.pyc....  and text inside the
executable like "Error creating Python TaxCalcEngine".

(The missing UserDict.pyc was caused by a UserError several
days ago, when I unthinkingly blew away all .pyc files on
my hard drive in an attempt to free up some space.  I never
though any software except mine would be using them.)

Judging by the content of the .pyc files, the bulk of the 
core of the software is at least scripted with Python, if
not written in Python.  There is also a hint that there's a 
PDF generator there, or at the very least a call to a PDF 
generator.  _socket.pyd is present as well.  My guess is that
the final tax file generation and e-filing code is done in
Python.

Just thought I'd mention another appearance of Python in a
commercial product.  Suddenly I feel somewhat more comfortable
with the calculations.  Now how come I don't get a refund?!
(I'll submit a patch to CVS as soon as I find the bug. :-)

-Peter

P.S.: Judging by the whining at 
http://www.lehopictures.com/the_joint/archives/cat_digital_world_software.php
they are using Python for its portability.  I can't speak 
personally for their tech support, but I've been _more than 
happy_ with the product itself for the last three years.
Unfortunately, it doesn't help anyone but a Canadian...



From andrew-pythonlist at puzzling.org  Fri Apr 25 22:10:57 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sat, 26 Apr 2003 12:10:57 +1000
Subject: Date validation
In-Reply-To: <b8bnf3$i8m$1@main.gmane.org>
References: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au> <b8ap7f$4go$1@main.gmane.org> <20030425080658.GG15695@frobozz.local> <b8bnf3$i8m$1@main.gmane.org>
Message-ID: <20030426021057.GK15695@frobozz.local>

On Fri, Apr 25, 2003 at 06:24:51PM +0200, Fredrik Lundh wrote:
> Andrew Bennetts wrote:
> 
> > > > I mean I *could* write this myself, but I had to ask the computer how
> > > > many days in April - someone with a better grasp of the date must have
> > > > written this stuff before :)
> > >
> > > don't you have any knuckles on your left hand?
> >
> > And which knuckles do you use to remember which years are leap years?  ;)
> 
> in what part of the world does the number of days in April
> vary depending on the year?

Heh.  I was pointing out that there's more than just the number of days in a
month involved.  Was 2000 a leap year?  What about 1900?  Dates are just
complex enough that my gut instinct is to avoid writing this sort of code
myself, if I can.  I've no desire to reinvent the mistakes of thousands of
other programmers.

> > > if you want your program to check how many days there are in
> > > a given month, use the calendar module:
> >
> > He does.  In his words: "how can I check whether or not such strings
> > represent a valid date?"
> >
> > His problem is verifying that the entire date is correct.  By the time you
> > factor in the varying days of the months and leap years, I can understand
> > wanting to use a prewritten and tested module rather than an ad-hoc solution
> > for a problem that many people must have encountered before.
> 
> "calendar" is a prewritten and tested module, and has been part
> of Python's standard library since around 1066 AD.

:)

> by using calender, he can add 3-4 lines of obviously correct code
> to his program, and move on to the next problem...

For days in the month, yes, so long as you're willing to ignore leap years.
That's presuming that leap years are the only "gotcha" with date
validation... I might be forgetting something.

Or I might be turning a molehill into a mountain.  But I've far too often
written 3-4 lines of obviously correct code that turned out to be not as
correct as I thought it was :)

If the calendar module had an "isValidDate" function, then I'd be
wholeheartedly agreeing with you!

> in contrast, you're suggesting that he
> 
> 1) update to a newer Python version, and end up with a program
> that may or may not work if run on an older version, or

I didn't suggest that at all; I was merely pointing out that eventually this
annoying problem will be taken care of by the new batteries included in
Python 2.3.  I find people generally like knowing that their problem will be
solved by a newer Python, even if they can't use it just yet.

> 2) backport changes to a newer Python version to his own Python
> install, and end up with a nonstandard Python environment and a
> program that may or may not work if run on an older version, or
> 
> 3) add 550 lines of recently written, rather complex, and yet to be
> widely used code to his application, and end up with a program that
> may or may not work...

At a glance, I thought 2 and 3 were the same option, because the code looked
very similar and was written by the same person.  Thus, I concluded that the
550 lines of code from the recipe were tested -- for better or worse, I
trust the Python standard library to be well-tested and dependable :)

It turns out that the code in 2.3 is a little different; 2.3's _strptime.py
uses the new datetime module in 2.3 to do some of the work -- and in
particular, raise the exception on invalid dates.  So that kills that idea,
because the version in the recipe doesn't raise an exception.

In private email we continued the thread a little, and it looks like he'll
probably use the mxDateTime package, seeing as he was already had it
installed, and it has the behaviour he wants.

> but I can of course not stop him from doing things in the hardest
> possible way; I'm just pointing out that there's another way to solve
> his problem using tools that's already installed on his machine... if
> you have a problem with that, that's your problem.

My problem is that I consider writing code to be a problem :)

If someone else has already written the code, I'd rather use that than my
own effort, if possible.  Sometimes it's too much work to reuse someone
else's code, sometimes it's somehow incompatible with my existing code.
Sometimes it's buggy.  Sometimes I think I could write a better version.

Mostly I just want to get the job done -- in which case I'll quite happily
try someone else's code first before attempting to design, write and debug
my own effort.

> (it could be worse, of course: I'm still waiting for the let's-use-a-
> metaclass crowd to discover this thread...)

No, I'm sure what we *really* need is to use are descriptors that return
generators... <wink>.

-Andrew.





From philh at cabalamat.org  Sat Apr  5 10:27:43 2003
From: philh at cabalamat.org (phil hunt)
Date: Sat, 5 Apr 2003 16:27:43 +0100
Subject: how to test when correct output is "random"?
References: <mailman.1049493247.4823.python-list@python.org>
Message-ID: <slrnb8ttff.tc2.philh@cabalamat.uklinux.net>

On Fri, 4 Apr 2003 15:53:13 -0600, Jeff Epler <jepler at unpythonic.net> wrote:
>
>But other parts are based on randomness.  "Alternation rules" are the
>foundation on which this is built.  An alternation rule might read
>like this:
>    S: "a" | "b" | "c";
>Then, str(S) would yield one of the alternatives.  Over many runs, it
>should yield each choice about an equal number of times.  So how do
>I test this?  Run a large number of trials and do a statistical test?
>This doesn't seem to fit in with the philosophy of software testing,
>which stresses exact repeatability.

If your code exhibits randomness, it obviously won't output the same 
value every time. All you can do is test the values are reasonable: 
statistical test are good. Also, check values are within legal 
ranges, so for example a dice throw would have to be an integer from 
1 to 6.

-- 
Philip Hunt <philh at cabalamat.org>

Interested in adventure holidays in Spain? 
Look at: <http://www.cabalamat.org/advcon/>



From aussie2010 at yahoo.com  Thu Apr 10 20:53:53 2003
From: aussie2010 at yahoo.com (Chris)
Date: 10 Apr 2003 17:53:53 -0700
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org>
Message-ID: <435d40e0.0304101653.1eb11a6c@posting.google.com>

Jason Tiller <jtiller at sjm.com> wrote in message news:<mailman.1049417184.1536.python-list at python.org>...

> 
> Good luck!  I hope this pattern at least gives you a starting point
> for implementing the split() in Python.
> 
> ---Jason
> Sonos Handbell Ensemble

Jason, thanks a lot for the example and the explanation. I fired up
Perl and it worked like a charm. Now to try and understand it!  I
looked at stuff about Python regular expressions and there is
something called finditer and findall that might be useful for parsing
a string, but can't seem to find any documentation on them.



From andersjm at dancontrol.dk  Tue Apr  8 07:40:55 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Tue, 8 Apr 2003 13:40:55 +0200
Subject: Making python module using Borland C++ 5
References: <b6u7fp$d2c$1@news.cc.tut.fi>
Message-ID: <3e92b527$0$32016$edfadb0f@dread12.news.tele.dk>

"Rami Saarinen" <rs55862 at NOSPAMyahoo.com> wrote:
> I am trying to build a simple module for python (2.1.1) using the 
> Borland C++ Builder 5.

Piece of cake with distutils, assuming you're just writing an
ordinary extension module, and not trying to embed Python in your
application.

Just write a setup.py and run it with --compiler=bcpp.

If there's something about your setup that makes this straightforward
approach not work, please elaborate.

- Anders





From graham__fawcett at hotmail.com  Tue Apr  1 14:21:47 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 1 Apr 2003 11:21:47 -0800
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de>
Message-ID: <e9570f37.0304011121.1873dac2@posting.google.com>

Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote in message news:<pan.2003.04.01.13.58.42.127815.759 at thomas-guettler.de>...
> Hi,
> 
> I think about moving from Zope to
> ZODB and (Quixote or Twisted).
> 
> I looked at zope3 but it seems very bloated.
> 
> Up to now Zope seems the best http server for ZODB.
> 
> Has anyone experience with serving an application 
> which uses ZODB as database backend but does not use Zope?
> 
>  thomas

I'm working on such an application right now -- Quixote and ZODB. It's
working out very well, and is far more manageable (and enjoyable!)
than any of the Zope projects I've worked on.

Do you have a specific concern that you're seeking feedback about?

-- Graham



From rmunn at pobox.com  Sat Apr  5 13:11:50 2003
From: rmunn at pobox.com (Robin Munn)
Date: Sat, 05 Apr 2003 18:11:50 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com>
Message-ID: <slrnb8ti36.ldu.rmunn@localhost.localdomain>

carroll at tjc.com <carroll at tjc.com> wrote:
> Modified method works as follows:
> 
>    xover(start, end, [file])
>        Return a pair (resp, list). list is a list of tuples, one for
>        each article in the range delimited by the start and end
>        article numbers. Each tuple is of the form (article number,
>        subject, poster, date, id, references, size, lines).  
> 
>        If the file parameter is specified, then the output is stored
>        to a file rather than to list.  If file is a string, then the
>        method will open a file object with that name, write to it, 
>        then close it.  If file is a file object, then it will start
>        calling write() on it to store the lines returned by XOVER.  If
>        file is supplied, then the returned list is an empty list.
>   
>        This is an optional NNTP extension, and may not be supported
>        by all servers.

A couple points.

First, from what you've written above, it sounds like you're doing
type-checking to determine whether the parameter is a string or a file.
You may not be doing type-checking, in which case this advice doesn't
apply to you, but type-checking is the wrong way to go about that. Do
interface checking instead, e.g.:

    # Is it a file-like object?
    try:
        output_file.write
        print "output_file is a file-like object"
    except AttributeError:
        # No: it doesn't have a write() method. Is it a string?
        try:
            output_file.lstrip
            print "output_file is a string-like object"
        except AttributeError:
            # Neither a file nor a string
            raise ValueError, "output_file should be a file or a string"

Second, do *not* use the name "file" as your parameter name; you will
shadow the file builtin.

Third, I'm not sure I like the functionality of passing a string
filename. It sounds convenient, but all it really does is *add*
ambiguity to your function which could be avoided. Just let that
parameter be a file-like object that's already been opened for writing,
and do *not* close it when you're done. If the user really wants the
functionality you provide (open filename, write, close it when done),
they can do it as easily as this:

    xover(start, end, open('foo.txt', 'w'))

See? File-like object gets created, passed to your function, and then
automatically garbage-collected (and thus closed) when your function
returns.

Other than that, nice job.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From bdesth at removethis.free.fr  Wed Apr  2 10:31:53 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 02 Apr 2003 17:31:53 +0200
Subject: Best practice for object attributes?
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk> <3E8AEC44.2090909@removethis.free.fr> <b6eusn$avn$1@nntp0.reith.bbc.co.uk> <mailman.1049298075.30675.python-list@python.org>
Message-ID: <3E8B0269.5080000@removethis.free.fr>

Jp Calderone wrote:
> On Wed, Apr 02, 2003 at 04:12:36PM +0100, Michael Sparks wrote:
> 
>>[snip]
>>
>>
>>As far as I can tell the behaviour seems to be that you automatically
>>gain instance attributes that clone the state of the class attributes
>>at object creation time, which are then independent and can be modified
>>via the usual self.<attr> syntax later on. (They cannot however be
>>modified via the self.__dict__[<attr>] syntax since they're not stored
>>there)
>>
> 
> 
>   Not quite.  What really happens is simpler.  When an attribute lookup on
> an instance fails, the lookup proceeds to the instance's class object (if it
> fails there, it proceeds to the class object's bases, and so on).  The
> attributes aren't cloned, and the time of creation matters not.  

Of course.... thought I had a brain, but it seems I can't find it today. 
Someone seen my brain 'round here ?

Laotseu




From peter at engcorp.com  Tue Apr  8 19:17:02 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 08 Apr 2003 19:17:02 -0400
Subject: Newbie: Python + C Question
References: <3e92ebe8$1@rmcnt03a.rmc.ca> <3E933C8D.5020700@removethis.free.fr>
Message-ID: <3E93586E.D4AB1531@engcorp.com>

laotseu wrote:
> 
> bourassa wrote:
> > When Python uses the C extension does it just process
> > it in an interpreted fashion
> 
> Python is not a C interpreter.
> 
> > or does the code written in C run like a
> > compiled C program?
> 
> newbie to programming, uh ?-)
> 
> How do you think this all works ?

Actually, to be fair to the OP, this is a valid and even interesting
question.  Where, after all, does it say that the C code that is
in the extension is actually pre-compiled into an executable and
is run as machine code when called by Python?  (Maybe it's there
in the docs, but maybe it's not or maybe it's unclear.  Have you
checked?)

I could picture someone believing that Python had the ability to
execute C code interpretively via some fancy C-interpreter
extension, which would allow the existing code-base of C to be 
used within an otherwise-pure-Python program.  Definitely not
very likely to happen as a general feature, but still an interesting
concept.

-Peter



From aleax at aleax.it  Fri Apr 18 11:23:26 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 17:23:26 +0200
Subject: Teaching : Python, Scheme, Java...
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFDE@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFDE@admin56.narex.com>
Message-ID: <200304181723.26065.aleax@aleax.it>

On Friday 18 April 2003 04:52 pm, Bjorn Pettersen wrote:
> > From: Alex Martelli [mailto:aleax at aleax.it]
> >
> > Donnal Walter wrote:
> > > What is the difference between "recursive calls" and "tail
> > > recursion"? Sometimes I have written methods that call the
> > > same method on a parent or child object, with the depth of
> > > recursion being limited by a method of the same name NOT
> > > calling itself. In the past I have thought of this
> > > as tail recursion. Is this incorrect?
> >
> > A function-call is "tail" if it's the last thing the caller does,
> > i.e., basically, if the caller executes "return somefunction(args)"
> > or a semantically equivalent sequence.
>
> [..recursive..]
>
> > A function-call is "tail recursive" if it's tail and recursive.
> > Such calls can be optimized by avoiding the allocation of a new
> > stack frame -- the same frame object that's already on the
> > stack to handle this specific call can be re-used (if you're
> > willing to lose traceback information -- Python currently does
> > not allow that).
>
> [To digress, elaborate, use my MS for something :-)] It's acutally more
> general. Any tail calls (recursive or not) can be optimized, trivially

Nor did I say otherwise, please note -- IF you can reuse the stack
frame rather than having to generate a new one, which depends
of course on what's going on behind the scenes (can the function
being called use the same frame that was used to call the one
that's now tail-calling it -- depends on stackframe format).

> (assuming the return address is stored in sp, etc.), instead of:
>
>    return f()
>
> being compiled to (A) -- excuse my shorthand:
>
>          push sp, loc1
>          jump f
>    loc1: pop sp
>          reg1 = pop fn_result_stack # 1
>          push reg1, fn_result_stack # 2
>          jump sp
>
> (1 and 2 would disappear with a peephole optimizer, assuming there is
> one) tail call optimization creates (B):
>
>          jump f
>
> with the last two lines of f looking like the last two lines of (A) the
> semantics are the same.

You're not dealing with the question of the stack frame.  How does
f know it's being called without arguments?  Surely it will find the
needed indication in its stack frame.  Who removes the stack frame
from the stack, caller or callee?  What happens if different calls need
different-sized stack frames -- can the currently-topmost stack frame
be "resized", and if so, can it still be correctly removed afterwards
(e.g. in a caller-removes-frame approach, does the caller needs the
frame size information to do the removal, and if so, can it find it out
about a frame that may have been resized?).

Calling conventions are established to try and optimize many issues,
and the ability to optimize all tail calls is only one of many.  The case
of a recursive tail call can sometimes be optimized (because all stack
frames received by function f are in some manner "uniform", but ones
received by other functions need not be) even in call conventions that
do not allow the general-case optimization.


> [[Now what if you came up with an algorithm that converted any function
> to a function with tail call? Your code for return would be really fast,
> it would never "return" to a previous state, you wouldn't need a stack
> anymore, implementing continuations would be trivial... They did it for

Of course, such a design would ensure uniformity of stack frames, having
chosen to focus on this optimization.  BTW, for those who may have
dabbled with Forth and friends in the past, this has some similarities
to "threaded interpreters" (even though there need not be an interpreter
involved, and it looks more like direct-threading than indirect) -- in Forth
of course the stack frame problem is swept away by having the programmer
handle the stack explicitly.

BTW, you may claim you don't need a stack, but that's in some sense
word-trickery;-).  If a calls b calls c calls d calls e, at this point the 
"nonstack" frames for 4 ongoing calls need to be "somewhere" -- this
somewhere may e.g be a singly linked list, but that's just because such
a list is a rather good implementation of... ahem, a last-in-first-out queue,
yes, that's the ticket!  [must not say the S-word...]

> SML, and the various steps, algorithms, etc. are nicely described in:
>
>    Compiling with Continuations, Andrew Appel
>    Cambridge University Press 1992, ISBN 0-521-41695-7
>
> which is one of the more interesting books I read in college both for
> language and compiler theory (couldn't find an online-ref).]]

Nolo contendere (I didn't study it deeply, barely browsed it, 10 years
ago or so -- my work was Fortran, C and some C++ then...:-).


Alex





From puccio_13 at yahoo.it  Fri Apr 25 07:35:05 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Fri, 25 Apr 2003 13:35:05 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com>
Message-ID: <Dl9qa.26067$DN.734905@tornado.fastwebnet.it>

"Martin v. L?wis" wrote:

> Alessio Pace wrote:
> 
>> Hi, I'd like to make a GUI for a program written in python, meant to run
>> on mac, win and linux: which is the preferable solution? I saw around
>> tkinter, wxpython and pyqt....
> 
> I recommend Tkinter. It comes with the standard Python installation, and
> it works quite well atleast on Windows and X11 (I've never used it on
> the Mac).
> 
> Regards,
> Martin

That's what I had in mind, but after reading some tutorials on Tkinter I saw
that wxpython and pyqt have more features still remaining portable...mmm...
Thanks anyway.

-- 
bye
Alessio Pace



From bbfoto2002 at netscape.net  Wed Apr  2 12:46:23 2003
From: bbfoto2002 at netscape.net (BBFoto)
Date: Wed, 02 Apr 2003 17:46:23 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org> <3e8afeb9.1125578@news-server> <3E8B0060.D9C449A3@engcorp.com>
Message-ID: <3e8b1b0f.8380222@news-server>

I apologize for top posting.  I clicked on "Followup Usenet Message".
My mistake was unintentional.  I will try to find out more about it as
was suggested.  And as to why I tried to execute a .pyc file:  If I
knew what I was doing, then I would do it right.  I do appreciate all
attempts to help.

I did a search on the PySol directory.  Nowhere is there a file called
"PySol.py".  There is a series of files in sequence named
"PySol_15.pyc"  "PySol_16.pyc"  "PySol_20..pyc"  "PySol_21.pyc" and
"PySol_22.pyc".


I tried to execute this line,
 
execfile('c:/program files/pysol/pysol-4.80/data/PySol_22.pyc') 

The path is correct.  After it failed I copied this error message from
the Dos screen:

  File "c:/program files/pysol/pysol-4.80/data/PySol_22.pyc", line 1
    -_
     ^
SyntaxError: invalid syntax


Thanks again,

BB




From paul at fxtech.com  Fri Apr 18 09:56:48 2003
From: paul at fxtech.com (Paul Miller)
Date: Fri, 18 Apr 2003 08:56:48 -0500
Subject: create a new instance in C of a Python subclass
References: <36ko9vknj9mm6m1ad18r3dm51q3t02pe03@4ax.com> <he8zegsn.fsf@python.net>
Message-ID: <4v00avocpki19ijlj8f8revh9tb032ht0f@4ax.com>

>You do the same as in Python: call the type.

How would I do that from C? 

Right now I'm just calling PyObject_NEW, but it doesn't seem to be calling
the subclass's (Python) __init__(). 





From cjw at sympatico.ca  Fri Apr 11 10:39:23 2003
From: cjw at sympatico.ca (Colin J. Williams)
Date: Fri, 11 Apr 2003 10:39:23 -0400
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <hfYka.2414$h%2.256728@read1.cgocable.net> <a3b5b608.0304100901.58ced140@posting.google.com>
Message-ID: <3E96D39B.4C60CA87@sympatico.ca>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030411/0caecde7/attachment.html>

From alanmk at hotmail.com  Wed Apr 30 05:37:56 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Wed, 30 Apr 2003 10:37:56 +0100
Subject: apache modules in python ?
References: <pan.2003.04.29.22.09.59.56666@sdf-eu.org>
Message-ID: <3EAF9974.6EF2DD7E@hotmail.com>

Geirgat Jonas wrote:
> 
> Is it possible to write modules for the apache http server in python ?

Just to be explicitly clear on what that means: You want to be able to
implement the Apache module API using python, and then dynamically load
the resulting module into Apache, and have it service requests?

Mod_python does indeed do exactly that: it implements the Apache module
API, in C, and permits the administrator to map URLs to python programs.
When it wants to execute a python program, mod_python has an embedded
python interpreter, through which it runs the programs, and serves the
resulting output back to the user.

If you were to write other Apache modules in python, say "mod_gzippy",
then you would need to also implement the Apache module API, in C. Code
for this is available in mod_python. 

Then you would want to call your python functionality which does
gzip'ping. Which means that your C code would need to call python code.
Due to the multithreaded nature of Apache, this can be a tricky
business. Python has a "Global Interpreter Lock", which allows only one
thread to be executing at a time. And it gets more complex when you are
using "foreign threads", i.e. Apache threads making function calls to
python threads. Again, mod_python implements all of this functionality. 

And does it correctly, having been thoroughly debugged over the years.

So, given the above considerations, is there a reason why you can't
simply use mod_python, which solves all of the hard problems for you?

regards,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From rstephens at vectron.com  Mon Apr 21 21:33:55 2003
From: rstephens at vectron.com (Ron Stephens)
Date: 21 Apr 2003 18:33:55 -0700
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org> <QxAoa.540832$3D1.296696@sccrnsc01> <a5415304.0304210544.1342b436@posting.google.com> <3EA490F5.30606@removethis.free.fr>
Message-ID: <a5415304.0304211733.364492e2@posting.google.com>

laotseu <bdesth at removethis.free.fr> wrote in message 
> Yep... There is another aspect of PythonCard that, imvho, makes it more 
> pythonic (than Boa), it's the fact that it's somehow a 'contained' 
> environnement. I mean that there is no clear distinction between the 
> environnement and the application, all the developpement tools being 
> parts of the final application - even if hidden. It's somehow like a 
> smalltalk environnement, where you modify an existing system until it 
> fits the application needs.


Hmm, I like that..I never thought of it that way before, but its very
true. Kind of a recursive tool-application-tool-application...

It really is a great 'framework'...



From justin at gnubia.net  Wed Apr  2 09:36:12 2003
From: justin at gnubia.net (Justin Ryan)
Date: 02 Apr 2003 08:36:12 -0600
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com>
References: <x7y92uhzpw.fsf@guru.mired.org>
	 <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <1049294172.31553.23.camel@qutang.gnubian.org>

howabout 'at most four, and no more than three in a row'

On Wed, 2003-04-02 at 08:03, Ben Hutchings wrote:
> In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer wrote:
> <snip>
> >     3.  At most three Is, Xs and Cs may appear in any given literal.
> <snip>
> 
> OK cleverclogs, how do we write 39?





From pobrien at orbtech.com  Sun Apr  6 15:15:25 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 14:15:25 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
Message-ID: <m2ptnz8mk2.fsf@localhost.localdomain>

Carl Banks <imbosol-1049649076 at aerojockey.com> writes:

> > Sound like fun?
> 
> Sounds like a lot of refactoring.
> 
> Here's how I would do it.  More experienced programmers, feel free to
> correct anything experience says won't work so well.

[great advice snipped]

Music to my ears.  That was the kind of feedback I was hoping to get.
Thank you very much, Carl.  The approach you described was one
approach I had in mind, but wasn't entirely sure was feasible, since I
haven't really delved into the internal of Emacs.  Of course, my
thoughts on the approach were very vague, while yours were very
specific.  You've given me some good things to think about.  Much
appreciated.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From andrew at acooke.org  Sun Apr 27 08:51:20 2003
From: andrew at acooke.org (andrew cooke)
Date: Sun, 27 Apr 2003 08:51:20 -0400 (CLT)
Subject: please comment on technologies
In-Reply-To: <xvMqa.9914$3M4.267925@news1.tin.it>
References: <slrnbalplq.254.luc@trillian.dont-panic.info> 
     <HqCqa.8931$3M4.241359@news1.tin.it> 
     <slrnbalv50.254.luc@trillian.dont-panic.info> 
     <xvMqa.9914$3M4.267925@news1.tin.it>
Message-ID: <53081.127.0.0.1.1051447880.squirrel@127.0.0.1>

Alex Martelli said:
>> Ok, I thought XSL was the definitive answer to all templating needs, but
>> you don't seem very excited about it ?
>
> I guess I shouldn't put my name to this -- it's not a "respectable"
> thing to say, and might inveigle me in flamewars I'd rather avoid --
[...]

not sure if this is an oblique reference to my comments about it being
difficult to sell XSL in another post, but i wasn't looking for a
flamewar, just relating direct experience.  i used XSL in my current
project and cannot get the client to use it.  instead they are writing
specific chunks of VB code (with hand rolled XML parsers!) to rearrange
data.  they were the people i was thinking of, not you!

if you've never used XSL then another reason to avoid it is that it can be
fiendishly tricky to solve some problems (even if it is turing complete). 
it's not a general purpose language by any means.  but it does rearrange
xml well (and i did manage to write something that verified the checksum
of chilean ID numbers once ;-)

andrew

-- 
http://www.acooke.org/andrew




From alloydflanagan at attbi.com  Mon Apr 14 10:46:52 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 14 Apr 2003 07:46:52 -0700
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it>
Message-ID: <a8b7f07a.0304140646.71675ca1@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<6EWla.4594$LB6.132415 at news1.tin.it>...
> A. Lloyd Flanagan wrote:
> 
> > Unless you absolutely have to have it, don't use it.
> 
> "Absolutely have to have it", just like "screws up big time", are
> silly overbids IMHO -- over-reactions that I don't see as justified
> at all!  Why can't we all just keep SOME perspective, please...?!
> 

Perspective?  Moi??

Seriously, I was going to clarify my statement, then realized that
you'd stated the trade-offs better than I could.  And thanks very much
for the benchmarks, that kind of data is really helpful.

Perhaps I should add a corollary to Knuth's law:  "Premature brevity
is the root of all misunderstandings"?

I think the root of the issue here is that the documentation could be
clearer about when to use __slots__ and when not to.  We could cut and
paste much of your last message into the docs, and it would help a
lot.  How do changes and updates make it into the official docs,
anyway?



From anton at vredegoor.doge.nl  Mon Apr 21 16:20:19 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 21 Apr 2003 22:20:19 +0200
Subject: __setitem__ puzzle
References: <b80q7c$cmd$1@news.hccnet.nl> <dOudnc1k_I7TrjmjXTWcqw@comcast.com>
Message-ID: <b81jrs$60c$1@news.hccnet.nl>

"Terry Reedy" <tjreedy at udel.edu> wrote:

>As Aahz correctly said, you should specify what l[tuple] is supposed
>to mean.

Probably, yes. But I didn't want to kill too many alternative
outcomes.

>Reading code and usage reveals the intention that
>l[(a,b,c)]==l[a][b][c], etc.

Yes. I couldn't believe that there wasn't a conversion function for
this in the standard distribution. Maybe there is, I still have a hard
time believing it.

>The solution is to directly do what you tried to do indirectly by
>making the copy.
>This also works with tlists as sublists.

This is appreciated very much! Please excuse my antinomic presentation
of the puzzle.

Anton





From sjmachin at lexicon.net  Fri Apr 25 22:37:08 2003
From: sjmachin at lexicon.net (John Machin)
Date: 25 Apr 2003 19:37:08 -0700
Subject: Date validation
References: <mailman.1051255479.31840.python-list@python.org>
Message-ID: <c76ff6fc.0304251837.5dc9fe6a@posting.google.com>

James Gregory <james at anchor.net.au> wrote in message news:<mailman.1051255479.31840.python-list at python.org>...
> Hi all,
> 
> I'm writing a killer web app in python and I want to validate all the
> input it gets. Amongst said input are dates represented as text strings
> of the format 'YYYY-MM-DD' - how can I check whether or not such strings
> represent a valid date? I had thought that

While the others are arguing about what to do with their knuckles, here's some code:

# dateok.py

days_in_month = [0, 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31]
#                    J   F   M   A   M   J   J   A   S   O   N   D

def leap(year):
   if year % 4: return 0
   if year % 100: return 1
   return year % 400 == 0
   
MIN_YEAR = 1800
MAX_YEAR = 2200
# suit yourself ...

def dateok(y, m, d):
   if not MIN_YEAR <= y <= MAX_YEAR: return 0
   if not 1 <= m <= 12: return 0
   if m == 2 and leap(y):
      lastday = 29
   else:
      lastday = days_in_month[m]
   return 1 <= d <= lastday

test_data_strg = """
2201 01 01 0
2200 12 31 1
1800 01 01 1
1799 12 31 0
2000 31 12 0
2000 12 31 1
2000 00 01 0
2000 02 29 1
2004 02 29 1
2001 02 29 0
2100 02 29 0
1900 02 29 0
2001 01 31 1
2001 02 31 0
2001 03 31 1
2001 04 31 0
2001 05 31 1
2001 06 31 0
2001 07 31 1
2001 08 31 1
2001 09 31 0
2001 10 31 1
2001 11 31 0
2001 12 31 1
2001 01 32 0
2001 02 32 0
2001 03 32 0
2001 04 32 0
2001 05 32 0
2001 06 32 0
2001 07 32 0
2001 08 32 0
2001 09 32 0
2001 10 32 0
2001 11 32 0
2001 12 32 0
2001 01 30 1
2001 02 30 0
2001 03 30 1
2001 04 30 1
2001 05 30 1
2001 06 30 1
2001 07 30 1
2001 08 30 1
2001 09 30 1
2001 10 30 1
2001 11 30 1
2001 12 30 1
2001 02 28 1
2001 01 00 0
2001 01 -1 0
2001 01 -2 1
"""
# The last entry is deliberate to show that the 
# testing mechanism actually does detect errors

if __name__ != "__main__":
   del test_data_strg
else:
   print
   npass = 0
   nfail = 0
   for line in test_data_strg.splitlines():
      if not line.strip(): continue
      y, m, d, expected = map(int, line.split())
      actual = dateok(y, m, d)
      if actual == expected:
         npass += 1
      else:
         nfail += 1
         print "dateok(%d, %d, %d) -> %d, expected %d" % (y, m, d, actual, expected)
   print
   print "Passed: %d, Failed: %d" % (npass, nfail)

 ... and not a class in sight, let alone a metaclass.

HTH,
John



From trentm at ActiveState.com  Tue Apr  1 13:43:16 2003
From: trentm at ActiveState.com (Trent Mick)
Date: Tue, 1 Apr 2003 10:43:16 -0800
Subject: Debug python for Windows?
In-Reply-To: <49CCE87D4A3BD511B1C500D0B7B69C0601334F01@exchange.youmeus.com>; from Sean.Slattery@aprsmartlogik.com on Mon, Mar 24, 2003 at 05:24:09PM -0000
References: <49CCE87D4A3BD511B1C500D0B7B69C0601334F01@exchange.youmeus.com>
Message-ID: <20030401104316.D9686@ActiveState.com>

[Sean Slattery wrote]
> 
> Neither ActiveState nor the .exe provided at www.python.org have debug
> versions of python.

Au contraire:
    http://downloads.activestate.com/ActivePython/etc/ActivePython-2.2.2-224-win32-ix86.debug.zip
The package is just relatively hidden. :)
(There is also an equivalent ftp://... path to this package.)

Unzip the package and run "python install.py".


Cheers,
Trent

-- 
Trent Mick
TrentM at ActiveState.com




From i at mindlace.net  Sat Apr 26 10:41:24 2003
From: i at mindlace.net (emf)
Date: Sat, 26 Apr 2003 10:41:24 -0400
Subject: Dealing with files...
In-Reply-To: <pan.2003.04.26.14.19.35.262492@soul.net>
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.29.48.204375@soul.net> <RZlqa.4182$lf3.712485@cletus.bright.net> <pan.2003.04.26.14.19.35.262492@soul.net>
Message-ID: <3EAA9A94.5040106@mindlace.net>

lost wrote:
> thanks for the info, however that doesn't seem to work either? here's the
> whole program that I'm trying:
> #!/usr/bin/python
> from glob import glob
> file_list = glob.glob('/var/www/html/') for each_file in file_list:
>         print each_file
 > print "done"
[...]
> Any ideas where I'm going wrong?
> Thanks again...

In this case you've imported just the function glob from module glob. So 
glob('/var/www/html/') would be sufficient; however, you've got other 
problems.

You can't put a logical block inside an assignment; in this case, you 
don't do anything with file_list aside from print each thing in it.

Here's one take, using glob:

#!/usr/bin/python
from glob import glob
for filename in glob('/var/www/html/*'):
     print filename
print "done"

Here's another using os.listdir:

#!/usr/bin/python
from os import listdir
dir = '/var/www/html/'
for filename in listdir(dir):
     print dir,filename
print "done"

The advantage of the latter is you get the invisible directories or 
files and, if you eventually want to expand this to something else, you 
have just the filenames instead of the full paths.

~mindlace






From michael at stroeder.com  Fri Apr 11 09:32:52 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Fri, 11 Apr 2003 15:32:52 +0200
Subject: ANN: python-ldap-2.0.0pre08
Message-ID: <mailman.1050073083.3825.clpa-moderators@python.org>

HI!

Find a new pre-release of python-ldap:

   http://python-ldap.sourceforge.net/

python-ldap provides an object-oriented API to access LDAP directory
servers from Python programs. It mainly wraps the OpenLDAP 2.x libs for
that purpose. Additionally it contains modules for other LDAP-related
stuff (e.g. processing LDIF, LDAPURLs and LDAPv3 schema).

Download link:
http://prdownloads.sourceforge.net/python-ldap/python-ldap-2.0.0pre08.tar.gz?download

Changes since 2.0.0pre07:

ldap.schema:
* For backward compability with Python versions prior to 2.2
   Lib/ldap/schema/tokenizer.py and Lib/ldap/schema/models.py use
   (()) instead of tuple() for creating empty tuples.





From fredrik at pythonware.com  Sun Apr  6 16:06:25 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sun, 6 Apr 2003 22:06:25 +0200
Subject: pb py32exe and PIL
References: <brzj4oqc.fsf@python.net> <20030406124725.26136.00000748@mb-fd.aol.com>
Message-ID: <b6q1be$uc8$1@main.gmane.org>

Ryomanga wrote:

> >You have to import the plugins you need in your script so that
> >py2exe knows they are used. You should be able to google for this
> >questions and find some answers.
> >
>
> I have imorted them

have you imported the PLUGIN modules you need?
(JpegImagePlugin, etc).

the following note discusses how to speed up initialization,
but the same approach can be used with pyexe:

http://www.pythonware.com/products/pil/articles/speeding-up-initialization.htm

</F>

<!-- (the eff-bot guide to) the python standard library:
http://effbot.org/zone/librarybook-index.htm
-->







From dreed at capital.edu  Mon Apr 14 19:52:57 2003
From: dreed at capital.edu (Dave Reed)
Date: Mon, 14 Apr 2003 19:52:57 -0400
Subject: splitting strings into variables
In-Reply-To: <pan.2003.04.14.23.33.00.456110@soul.net>
References: <pan.2003.04.14.23.33.00.456110@soul.net>
Message-ID: <200304141952.57775.dreed@capital.edu>

On Monday 14 April 2003 19:33, lost wrote:
> hi there,
> i've read about splitting strings, and can do
> strname.split(delimeter)
> but what i need is something like this:
> strname = hello:there
> strname.split(":")
> print strname
> hello
> 
> so :there is taken away, and i just have hello?  how can i do that 
with
> split?

strname = "hello:there" # your example was missing quotes here
strname = strname.split(":")[0]
print strname
hello

Dave





From skip at pobox.com  Fri Apr  4 16:59:48 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 4 Apr 2003 15:59:48 -0600
Subject: Invalid format character in string
In-Reply-To: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com>
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com>
Message-ID: <16014.84.301494.928376@montanaro.dyndns.org>

    Nikolai> ... since the text is actually a HTML file, it may contain a
    Nikolai> lot of characters with % (typically cell width definitions of
    Nikolai> tables).  If these are contained in the file, the string format
    Nikolai> statement complains about illegal format character.  

Just double any literal % characters, e.g.:

    <table width="%(width)s%%">

Skip




From janeaustine50 at hotmail.com  Sun Apr 13 07:39:42 2003
From: janeaustine50 at hotmail.com (Jane Austine)
Date: 13 Apr 2003 04:39:42 -0700
Subject: How to subclass datetime?
Message-ID: <ba1e306f.0304130339.3bece039@posting.google.com>

Running Python 2.3a2 on win XP.

I'm interested in the new datetime module and I want to subclass
date/datetime class to modify the behaviour. However, simply
inheritting the classes(like "class mydate(date):" ) didn't work.

Should I resort to delegation?

Thanks in advance.

Jane



From duncan at NOSPAMrcp.co.uk  Tue Apr  8 06:31:37 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 8 Apr 2003 10:31:37 +0000 (UTC)
Subject: passing no arguments to a callback function question
References: <b6tqig$93k$1@lust.ihug.co.nz> <3E9294D4.A9F95112@engcorp.com> <b6u60i$i6p$1@lust.ihug.co.nz>
Message-ID: <Xns935773F03FC90duncanrcpcouk@127.0.0.1>

matthew <matthew at newsgroups.com> wrote in 
news:b6u60i$i6p$1 at lust.ihug.co.nz:


> Thanks for the hint. However, functions like print_params are outside of 
> the class scope sitting in another module. The idea is that users can 
> write their own functions which they can add as listeners for certain 
> event types being scanned from a file. I 'pinched' the curry class from 
> the py.cookbook and I think I'm probably being tripped up by my only 
> vague understanding of argument passing. Here is the listener code 
> (which I'm sure must look pretty ugly to you guys):-. Thanks for further 
> assistance. matthew.


>              entry = (function, self.curry(function,args))

This passes an argument of None which will then be passed on to your 
function when it is called. You need to pass an empty sequence if you want 
your function to take no arguments. Also it looks as though you want to 
unpack a tuple 'args' into the *args argument to curry, which means you 
need to use the star on the call as well.

Try:

     if args is None: args = ()
     entry = (function, self.curry(function, *args))

or you could set the default argument to () instead of None for a similar 
effect.





-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From mwh at python.net  Wed Apr  2 10:48:05 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 2 Apr 2003 15:48:05 GMT
Subject: Slicing in Python 2.3a2
References: <mailman.1049200983.30254.python-list@python.org> <slrnb8jamj.421.ramen@lackingtalent.com>
Message-ID: <7h3pto4lx3f.fsf@pc150.maths.bris.ac.uk>

Dave Benjamin <ramen at lackingtalent.com> writes:

> I think it makes sense if you think in terms of "start", "stop", and "step".
> The step amount represents the value that is added to the implicit counter
> with each iteration. I think that the confusion actually comes from the
> defaults for start and stop if they are not supplied. It seems from your
> example (and the [::-1] form) that the defaults for start and stop are
> swapped if the step is negative.

Yep!  You wouldn't believe the amount of hair I tore out implementing
all that...

Cheers,
M.

-- 
  I'm not sure that the ability to create routing diagrams 
  similar to pretzels with mad cow disease is actually a 
  marketable skill.                                     -- Steve Levin
               -- http://home.xnet.com/~raven/Sysadmin/ASR.Quotes.html



From staschuk at telusplanet.net  Wed Apr 30 00:22:38 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 22:22:38 -0600
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 29)
In-Reply-To: <20030429183847.A1050@localhost.localdomain>; from jack@performancedrivers.com on Tue, Apr 29, 2003 at 06:38:47PM -0400
References: <vatflcjn7dum95@corp.supernews.com> <20030429143703.C1336@tibia.amotlpaa.bogus> <20030429183847.A1050@localhost.localdomain>
Message-ID: <20030429222238.A3128@tibia.amotlpaa.bogus>

Quoth Jack Diederich:
  [...]
> 2) What happened on the 27th to cause a blind spot in google?
>    Best conspiracy theory wins a cookie.

<paranoia> The Church of Scientology's lawyers sent a letter to
Google claiming copyright on our discussions that day and
threatening legal action, so Google removed them. </paranoia>

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From kjt at clmac-kjt.st-and.ac.uk  Tue Apr 15 07:25:13 2003
From: kjt at clmac-kjt.st-and.ac.uk (Kevin J Thomson)
Date: 15 Apr 2003 12:25:13 +0100
Subject: installing zope with python in a different place
Message-ID: <m2n0isknom.fsf@clmac-kjt.st-and.ac.uk>

Hello I am trying to install zope 2.6.1 on a solaris 8 box from scratch (i know there is a binary build)
and have installed python 2.1.3 in my own directory with threads /home/kjt/python

I have build zope with the right version of python and everthing builds fine (zope in /home/kjt/zope).

the problem occurs when i try to run the start script.

I get  

Traceback (most recent call last):
  File "/home/kjt/zope/z2.py", line 529, in ?
    import ZServer
  File "/home/kjt/zope/ZServer/__init__.py", line 86, in ?
    from HTTPServer import zhttp_server, zhttp_handler
  File "/home/kjt/zope/ZServer/HTTPServer.py", line 43, in ?
    from PubCore import handle
  File "/home/kjt/zope/ZServer/PubCore/__init__.py", line 14, in ?
    import ZRendezvous
  File "/home/kjt/zope/ZServer/PubCore/ZRendezvous.py", line 15, in ?
    from ZServerPublisher import ZServerPublisher
  File "/home/kjt/zope/ZServer/PubCore/ZServerPublisher.py", line 13, in ?
    from ZPublisher import publish_module
  File "/home/kjt/zope/lib/python/ZPublisher/__init__.py", line 23, in ?
    from Publish import publish_module, Retry
  File "/home/kjt/zope/lib/python/ZPublisher/Publish.py", line 19, in ?
    from Response import Response
  File "/home/kjt/zope/lib/python/ZPublisher/Response.py", line 13, in ?
    import HTTPResponse
  File "/home/kjt/zope/lib/python/ZPublisher/HTTPResponse.py", line 19, in ?
    import zlib, struct
ImportError: No module named zlib

does anybody have any suffestions to what I might have not set done wrong.

Thanks in advance.

Kevin.




From abpillai at lycos.com  Mon Apr  7 06:54:08 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 7 Apr 2003 03:54:08 -0700
Subject: Emacs python-mode problem
References: <m2of3ktq67.fsf@db.hit.edu.cn> <bd993a2f.0304062214.2027bed8@posting.google.com> <3E911848.D8AFE286@alcyone.com>
Message-ID: <bd993a2f.0304070254.18cece@posting.google.com>

Well. I am a prisoner of M$. :-O

Anand

Erik Max Francis <max at alcyone.com> wrote in message news:<3E911848.D8AFE286 at alcyone.com>...
> Anand B Pillai wrote:
> 
> > (setq auto-mode-alist
> >        (append '(("\\.C$"  . c-mode)
> >                 ("\\.cc$" . c++-mode)
> >                 ...
> 
> Note that .C (on case sensitive filesystems, anyway) is usually a _C++_
> file, not a C one.



From tundra at tundraware.com  Tue Apr  8 12:50:08 2003
From: tundra at tundraware.com (Tim Daneliuk)
Date: 08 Apr 2003 16:50:08 GMT
Subject: [Slightly OT]: More on ints and floats
In-Reply-To: <mailman.1049768113.16134.python-list@python.org>
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org>
Message-ID: <kauu6b.f2d1.ln@boundary.tundraware.com>

Lulu of the Lotus-Eaters wrote:
> Tim Daneliuk <tundra at tundraware.com> wrote previously:
> |As I understand it, integers and floats are distinct mathematical
> |entities. A colleague of mine claims, that insfar as we use them in
> |computing, ints are merely a proper subset of floats.
> 
> Floats are peculiar creatures.  They really are not much like Real or
> Rational numbers, even though it is often convenient to pretend they
> are.  The trick about floating point numbers is that although they are
> extremely useful for representing real-life (fractional) quantities,
> operations on them do not obey the arithmetic rules we learned in middle
> school:  associativity, transitivity, commutativity; moreover, many very
> ordinary-seeming numbers can be represented only approximately with
> floating point numbers.  For example:
> 
>      >>> 1./3
>      0.33333333333333331
>      >>> .3
>      0.29999999999999999
>      >>> 7 == 7./25 * 25
>      0
>      >>> 7 == 7./24 * 24
>      1
> 
> Tim knows this, of course, but it is worth emphasizing.  Moreover, a
> quote the Timbot found for this list a couple years ago is worth
> reiterating:
> 
>   Many serious mathematicians have attempted to analyze a
>   sequence of floating point operations rigorously, but found
>   the task so formidable that they have tried to be content
>   with plausibility arguments instead.
>   -- Donald Knuth (_The Art of Computer Programming_, Third Edition,
>      Addison-Wesley, 1997; ISBN:  0201896842, vol.  2, p.  229):
> 
> Yours, Lulu...

As a side note, this question was ultimately motivated by an entirely
different sort of "pure math" question: Does (integer) 1 = 1.0?
Engineers, physicists, and computer scientists all would
say "no" because of the precision problem with the latter number.
I am unclear, though, on whether a mathematician understands
1.0 to mean 1.0000000... (a single number) or the _neighborhood_
around 1 with one digit precision.


Thanks to all who took the time to respond...


-- 
----------------------------------------------------------------------------
Tim Daneliuk     tundra at tundraware.com
PGP Key:         http://www.tundraware.com/PGP/TundraWare.PGP.Keys.txt




From bdesth at removethis.free.fr  Wed Apr 16 18:58:44 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 16 Apr 2003 22:58:44 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz>
Message-ID: <3E9DE024.3020807@removethis.free.fr>

Paul Foley wrote:
> On Tue, 15 Apr 2003 23:49:37 +0000, laotseu  wrote:
> 
> 
>>Even for OO programmers, functionnal features in Python are IMHO a
>>great plus, and BTW functionnal and OO paradigm does not have to
>>conflict (that would be functionnal vs imperative and object vs
>>procedural). CLOS is one of the great system objects out here, and
>>it's been implemented on top of a functionnal language,
> 
> 
> Oh yes?  Which functional language would that be?
> 

CLOS means Common Lisp Object System.

Laotseu




From tim.one at comcast.net  Wed Apr 16 20:03:52 2003
From: tim.one at comcast.net (Tim Peters)
Date: Wed, 16 Apr 2003 20:03:52 -0400
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <b7km3s$238m$1@nntp6.u.washington.edu>
Message-ID: <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net>

[Tim, quotes the __repr__ docs
 http://www.python.org/doc/current/ref/customization.html
]

[Donn Cave]
> It's clear that the docs promote this idea, but they also hint at
> its half-baked quality - repr() evals to the original value, unless
> it doesn't.  I don't think the idea is clear at all.

I do, but the variety of objects is messy.  The spirit of __repr__ is "do
the most faithful job you can, within reason".  Some objects simply can't be
reproduced faithfully with a comprehensible Python expression, and the docs
admit to that reality up front.

> Taken together with the __str__ documentation, the original poster saw
> this stuff and did not find it clear.

I believe that, but am not concerned by it.  There's usually more than one
way to convert an object to string that may be useful, and for some objects
there may be many distinct ways, and Python has two convert-to-string
special methods to cover two frequent uses:  friendly (str) and anal (repr).
I think these work very well for the builtin scalar types, but not so well
for the builtin container types (because str(container) applies repr() to
the containees).  It's also sometimes unfortunate that the interactive
prompt is wedded to __repr__ by default.  So it goes.  What to do with each
method is in large part art, and the teachable seem to do OK learning by
example.  Here's an example new in 2.3:

>>> import datetime
>>> t = datetime.date.today()
>>> print repr(t)
datetime.date(2003, 4, 16)
>>> print str(t)
2003-04-16
>>>

Fred Drake decided to make repr() and str() do those specific things for
dates, and it was instantly obvious to me, and to Guido, that those are
appropriate things for each to do.  Is it actually hard for you to see that
both match the doc's intents?  It can't be hard to see that the repr() does,
so maybe you can't see that the str() result is "more convenient or
concise"?  Hard to believe <wink>.

>>> print eval(repr(t))
2003-04-16
>>> eval(repr(t)) == t
True
>>>

> He learned that repr is ``official'', vs. the ``informal'' str.  This is
not
> explained.

Not to the point of an algorithm everyone can apply to a class and come up
with identical results, no.  That's why I call it art.  Going back to the
date example, if str(t) produced '04-16-2003', that would have been a poor
str() result, but "explaining why" requires dragging in arguments from human
interface design.  You can't expect the __str__ docs to cover that:  the
goals are explained briefly, and the rest is up to the programmer.

> The requirements for repr conflict, and the primary one is conceded to
perhaps
> not be practicable.

That's because sometimes it's not.  So what?  Because it may not always be
possible, you want the language to cut off the possibility of trying?

> The requirements for str are on the other hand silent on its intended
> application.

Applications are the user's business, not the language spec's.  If you can't
think of an application for "more convenient or concise; doesn't need to try
to be a Python expression", stay away implementing __str__.

> The only kind of clarity this documentation possesses is when it lets you
read
> into it what you already thought.  Not the fault of the documentation,
though,
> but of the unsound ideas behind it.

It's evident that more than one way of converting to string may be useful,
and that some ways are more useful for developers and others for end users.
Those are "the unsound ideas behind it".  They're not precise ideas, nor do
I believe they can be made precise.  Please tell us how *you* would design
the to-string facilities in a language -- sticking strictly to "sound"
ideas, of course.






From dale at riverhall.NOTHANKS.co.uk  Thu Apr 17 17:22:55 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Thu, 17 Apr 2003 22:22:55 +0100
Subject: "pythonic"
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
Message-ID: <5l5u9v4s0oqc5q94ns5fqojb3hpdcpn75a@4ax.com>

Afanasiy <abelikov72 at hotmail.com> wrote:

>Too many times now I have seen the term "pythonic" used to vaguely
>describe something which is never explicitly defined, as if it cannot be.
>
>Now it seems I can just replace "pythonic" with "my idea of utopia".
>
>I do not know if this ignorance, laziness, or a meme, but I would like
>to know if there a real definition of "pythonic". Anyone?

It's mostly guff; frequently used by people struggling to attain the
moral high ground to help justify their position.

Once they've validated their argument as 'pythonic', it becomes
blasphemous to adopt a counter stance.

The first to play the 'pythonic' card tends to win the argument.

However, 'premature pythonic' can alert the wary to an unskilled
player who may then be subject to more scrutiny than they'd like. A
subsequent 'false pythonic' can damage your credibility and make it a
right bitch to get taken seriously again.

Hope that helps.

-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From cben at techunix.technion.ac.il  Fri Apr 11 07:40:09 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Fri, 11 Apr 2003 14:40:09 +0300 (IDT)
Subject: convert dictionary to string
In-Reply-To: <b72buh$9uo$1$8300dec7@news.demon.co.uk>
References: <b72a7l$ifj$1@lust.ihug.co.nz> <b72buh$9uo$1$8300dec7@news.demon.co.uk>
Message-ID: <Pine.GSO.4.51.0304111438170.15707@techunix.technion.ac.il>

Stephen D Evans wrote on 2003-04-10:

>
> d={'A':45.10, 'F':56.50,}
>
> # assuming that you would like 2 decimal places...
>
> s1 = ''.join(['%s%.2f' % item for item in d.iteritems()])
>
> # or without the list comprehension...
>
> s2 = ''
> for (key, value) in d.items():
>     s2 = '%s%s%.2f' % (s2, key, value)
>
Note that the later is using quadratic time.  Joining strings by
appending to them is almost always a bad idea (unless you have very
few of them); ''.join is the recommeded solution for most cases but I
guess [c]StringIO is very effecient too.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>

Never spend several weeks on a single homework - or you will miss a
nice Linux conference ;-(.




From mark at freelance-developer.com  Sat Apr 26 16:49:18 2003
From: mark at freelance-developer.com (Mark Nenadov)
Date: Sat, 26 Apr 2003 20:49:18 +0000
Subject: WML examples using mod_python?
References: <slrnbals94.2qp.grante@tuxtop.visi.com>
Message-ID: <uLFqa.9353$h%2.1023630@read1.cgocable.net>

Grant Edwards wrote:

> Can anybody point me towards any examples of using mod_python
> for WML?
> 
> Any other python-related WML examples (normal CGI, etc.) would
> also be welcome.
> 

Grant,

I haven't worked with WML in Python, though I have in other languages (PHP, 
etc). 

All you will really need to do is assemble your WML decks (which are just 
markup) and then output them along with the correct Content-type line. Once 
you have your WML generated, it is just a matter of outputing it. You 
shouldn't have to do anything special with Mod_Python

If you are looking for something to make WML-generation quicker, you might 
want to check out "pywml" (http://sourceforge.net/projects/pywml).


-- 
~Mark Nenadov



From Vector180W at netscape.net  Tue Apr  1 19:37:14 2003
From: Vector180W at netscape.net (vector)
Date: Wed, 02 Apr 2003 10:37:14 +1000
Subject: Entry widget bind problems
Message-ID: <3E8A30BA.7000805@netscape.net>

newbie
PythonWin 2.2.2

Im a little confused with an entry widget bind. I had it working b4 but 
In an attempt to tidy up the code I have confused the class/callback 
reference or something.
As you can see from the text below, there is a MyFrame instance which 
builds the TK Gui window, in which a frame contains an entry widget. 
after typing in text into the entry widget a <return> event(enter key) 
is supposed to run the doEntry, but all i get this error.
what am i doing wrong?

##error
    p=page.get()       #gets all text
AttributeError: Event instance has no attribute 'get'
##
def doEntry(page,*args):
     """Gets text from Entry widget on <enter> hit"""
     p=page.get()       #gets all text
     print p              #print it
     page.delete(0,END)        #clear entry widget

def entry(self,side,initTxt):
     """Make simple entry widget.Requires initial text"""
     self.entrVar = StringVar()      #assign variable
     w=Entry(self, relief=SUNKEN,textvariable=self.entrVar,
             font=(('MS','Sans','Serif'),'6'))
     #w.bind('<Return>',doEntry,self)	#tried to bind it here too
     w.pack(side=side, expand=YES,fill=BOTH)
     w.insert(0,initTxt)
     return w

class MyFrame(Frame):
     def __init__(self,master):
         Frame.__init__(self,master)
		..........snip
         self.outEntry=entry(Oframe,LEFT,'Enter_Command')
         self.outEntry.bind('<Return>',doEntry)
		......endsnip

ALSO	tried
         self.outEntry.bind('<Return>',doEntry,self.outEntry)




From mcherm at mcherm.com  Wed Apr 16 13:06:59 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Wed, 16 Apr 2003 10:06:59 -0700
Subject: How many of you are Extreme Programmers?
Message-ID: <1050512819.3e9d8db34e222@mcherm.com>

Christopher writes:
> So how many of you guys use XP processes?

Well, I'm not sure if I qualify or not. I really LIKE most of the points
that XP makes, but I find many of its proponants overly exuberant. I
think exhaustive, automated tests are a wonderful idea. I believe in
refactoring. I think pair programming works very well most of the time.
I think super-short release cycles (with code that grows feature-by-
feature) is a great idea. I think having a customer on-site is extremely
helpful.

I DON'T think that it's wise to just start coding without a bit of design
up front (although one must keep YAGNI in mind). I don't think that using
the code itself as the only documentation (no design docs) works well.
Most of all, I don't like mandating any particular methodology... the
needs vary from project to project and team to team.

But none of that really matters, because I don't get to set the policy
where I work. We don't do unit tests (except for the ones that _I_ write)
because "they would take too long". We deliver functionality in a single
big release. We don't do pair programming except for very brief periods.
Perhaps someday I can change those things, but for now, I don't get to
practice much of this, except in my at-home-on-my-own-time projects
(basically non-existant since my son was born 4 months ago).

-- Michael Chermside





From ckea25d02 at sneakemail.com  Tue Apr  1 19:31:31 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 12:31:31 +1200
Subject: Is Python overhyped (just like Java)?
In-Reply-To: <b6cobh$hck$1@news.island.liu.se>
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it>  <bd993a2f.0303310333.5a95ecb3@posting.google.com>  <roy-D262FA.09230131032003@reader2.panix.com>  <b6aam1$7o6$1@news.island.liu.se> <uk7ef3w3c.fsf@boost-consulting.com> <b6cobh$hck$1@news.island.liu.se>
Message-ID: <b6dahn$40kd1$1@ID-169208.news.dfncis.de>

Thomas Bellman wrote:
> It says that
> the motivation was for container classes, but IMHO that's
> essentially the same problem: you want to use types/classes with
> different inheritance in the same functions or data structures,
> as long as the interface is the same.

No, it's not the same problem, when you're dealing with a
static type system. Getting stuff *into* the collection is
only half the story. The real problem is that when you
get something *out* of the collection, you want the compiler
to know statically that it's of whatever type the collection
is supposed to contain.

Interfaces don't help you with that at all -- you need a
notion of parameterized types, which is what C++ templates
are.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From jolderdissen at ixiacom.com  Tue Apr 29 15:55:03 2003
From: jolderdissen at ixiacom.com (Jan Olderdissen)
Date: Tue, 29 Apr 2003 12:55:03 -0700
Subject: Raising objects
References: <453d10b2.0304290750.222fb453@posting.google.com>
Message-ID: <tpltav0bbeehseajdo3osampi7flsvhk6g@4ax.com>

Carlos,

since object is a type object
(http://www.python.org/doc/current/lib/bltin-type-objects.html) and
not a string, class or instance of a class you can't raise it or
classes derived from it.

That said, I question the point of deriving from object as classes
(and in fact all objects as far as I can tell) are derived from object
as it is. Since object isn't a class, deriving from it probably is
bogus to begin with. What exactly are you trying to do?

Jan

On 29 Apr 2003 08:50:42 -0700, cpitaar at yahoo.com.ar (Carlos P.) wrote:

>Hi all,
>
>I have this short question: can the operator 'raise' raise exceptions
>which are instances of object? I'm getting the following behaviour
>from python 2.2.2:
>
>>>> class F: pass
>
>>>> raise F()
>Traceback (most recent call last):
>  File "<stdin>", line 1, in ?
>__main__.F: <__main__.F instance at 0x8136bf4>
>
>>>> class E(object): pass
>
>>>> raise E()
>Traceback (most recent call last):
>  File "<stdin>", line 1, in ?
>TypeError: exceptions must be strings, classes, or instances, not E
>
>Thank you,
>Carlos




From xxdigitalhellxx at hotmail.com  Wed Apr 30 10:20:07 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Wed, 30 Apr 2003 08:20:07 -0600
Subject: How to call built in function 'open' from inside object that
 has 'open' method?
In-Reply-To: <db81c492.0304300526.7b8d3c86@posting.google.com>
References: <db81c492.0304300526.7b8d3c86@posting.google.com>
Message-ID: <b8om5t$984$1@heckle.cc.uregina.ca>

vadim wrote:
> I am new to Python and do not know the answer to very
> simple how to 'open' from inside another object that
> has 'open' method. I know that in C++ I should use
> following syntax ::open. 
> 
> More details on reasons. I have met this problem trying to 
> write a simple script use ActiveScripting  inside IE.
> I am trying "file = open(...) but because context window
> object has 'open' method that just opens window it is called
> first. Sorry for simple question. I was trying to find 
> the answer but either it is not on the surface or too simple.

Hmm, maybe I'm missing something but should open() call the builtin function 
and self.open() call your class function?

-- 
Ryan




From tvassila at siac.com  Mon Apr 14 17:42:02 2003
From: tvassila at siac.com (Thanos Vassilakis)
Date: Mon, 14 Apr 2003 17:42:02 -0400
Subject: Seperating HTML content from logic ( was: Re: Inserting Python commands
 into HTML documents)
Message-ID: <OF6352333F.BF2DEE10-ON85256D08.0076B46B@nsmtp.siac.com>

Whole point behind pso is to  separate HTML content from logic:

We use it at NYSE with success - but it is still very much beta.



http://sourceforge.net/projects/pso/


user guide: http://www.scriptfoundry.com/modules/pso/doc/pso-guide.html

and building a site:
http://www.scriptfoundry.com/modules/pso/doc/pso-example.html



-----------------------------------------
This message and its attachments may contain  privileged and confidential information.  If you are not the intended recipient(s), you are prohibited from printing, forwarding, saving or copying this email.  If you have received this e-mail in error, please immediately notify the sender and delete this e-mail and its attachments from your computer.





From djc at object-craft.com.au  Sun Apr 27 08:14:09 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 27 Apr 2003 22:14:09 +1000
Subject: Sybase module 0.36 released
Message-ID: <m3y91wjfxq.fsf@ferret.object-craft.com.au>

WHAT IS IT:

The Sybase module provides a Python interface to the Sybase relational
database system.  It supports all of the Python Database API, version
2.0 with extensions.

NOTES:

This is another incremental improvement to FreeTDS support.

You can build for FreeTDS like this:

     python setup.py build_ext -D HAVE_FREETDS -U WANT_BULKCOPY
     python setup.py install

If you are using FreeTDS 0.60 or 0.61 change the compile command to:

     python setup.py build_ext -D HAVE_FREETDS=60 -U WANT_BULKCOPY

or
     python setup.py build_ext -D HAVE_FREETDS=61 -U WANT_BULKCOPY


The module is available here:

    http://www.object-craft.com.au/projects/sybase/download/sybase-0.36.tar.gz

The module home page is here:

    http://www.object-craft.com.au/projects/sybase/

CHANGES SINCE 0.36pre6:

* Output parameters from stored procedures are supported via the
  OUTPUT() function.  The parameter to OUTPUT() determines the type
  and size of the buffer which will be allocated to receive the output
  value.

    c.callproc('test_proc', {'@type': 'business',
                             '@tot_sales': 5,
                             '@num_books': Sybase.OUTPUT(1)})

* Added Time() and TimeFromTicks() DB API functions.

* Fixed fmx.maxlength typo.

* Reorganised result fetching to use "fetcher" classes; _FetchNow,
  _FetchNowParams, and _FetchLazy.  This was in response to testing
  against the Python DB API 2.0 Anal Compliance Unit Test.

CHANGES SINCE 0.35:

* Fixed bug in ctx.c:ct_con_drop() which caused SEGFAULT with
  FreeTDS.

* Export version of FreeTDS used in sybasect.__have_freetds__

* Set Sybase._ctx as the global context soon as in Sybase.py to
  avoid problems with undefined global ctx.

* Look for SYBASE_OCS on all platforms.

* Fixed spelling error in ctx.c:global_ctx().

* Fixed cursor locking bug with .nextset() method.

* Added .debug_msg() method to context type which writes to the debug
  log when debugging is enabled for the context object.

* Merged AIX build changes from Dietmar Rothkamp.

* Now compiles with FreeTDS 0.61.

* Limit size of TEXT fields to 65536.

* Removed references to CS_PENDING and CS_BUSY in Sybase.py.

* Bind CS_VARBINARY_TYPE as CS_BINARY_TYPE.

* Do not attempt to ct_cancel() when handling an exception if have not
  connected to the server.  Bug fix in Sybase.py.

* Seems like FreeTDS reports the wrong maxlength in ct_describe() for
  CS_NUMERIC_TYPE and CS_DECIMAL_TYPE.  Ignore FreeTDS and assume a
  maxlength of sizeof(CS_NUMERIC).  Bugfix in databuf.c.

* Use correct T_STRING_INPLACE type in structure member descriptions.
  The code does not use Python API for these members.

* Debug output for DataFmt now includes scale and precision.

* More definitions added to freetds.h; CS_SRC_VALUE, CS_CLEAR.

-- 
http://www.object-craft.com.au





From lekishvili at python.qartu.com  Tue Apr  8 07:02:52 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 8 Apr 2003 04:02:52 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com>
Message-ID: <a3b5b608.0304080302.c26ad68@posting.google.com>

> No its not a naturally tailored web-programming language, because it
> is
> very hard to integrate in Server Pages. You have to change the
> indentation syntax/semantic of python to embedd it in a real useable
> way.
> This is done by Webware but it looks very strange to normal python
> programmers.
> 
> Of course you can use template systems. But this is not always a good
> economic decision.

OK, but can't one use XML/XLST as a sole solution for the view?



From donn at drizzle.com  Thu Apr 10 17:56:26 2003
From: donn at drizzle.com (Donn Cave)
Date: Thu, 10 Apr 2003 21:56:26 -0000
Subject: popen(), sendmail: Success?
References: None <3e95d63f.2635484@news.muenster.de>
Message-ID: <1050011784.966197@yasure>

Quoth mb at muenster.de (Martin Bless):
| To send mails from Python here's the example my provider provides:
| This http://faq.puretec.de/skripte/python/3.html
|
| The relevant lines are these:
|
| sendmail = os.popen("/usr/lib/sendmail -t", "w")
| sendmail.write(msg)
| sendmail.close()
|
| This works just fine. But is it enough? How do I know the operation
| succeeded?
|
| I've read the docs about os.popen[234] and the module popen2 but I'm
| not sure where to go. And since experimentation isn't easy in this
| case I'd appreciate some hints or a concrete example very much.

But since you use os.popen, you would want to read the docs for popen.

popen(command[, mode[, bufsize]])
   Open a pipe to or from command. The return value is an open file
   object connected to the pipe, which can be read or written depending
   on whether mode is 'r' (default) or 'w'. The bufsize argument has the
   same meaning as the corresponding argument to the built-in open()
   function. The exit status of the command (encoded in the format
   specified for wait()) is available as the return value of the close()
   method of the file object, except that when the exit status is zero
   (termination without errors), None is returned. 

>From this you can see that if the command completes without errors,
sendmail.close() returns None.

	Donn Cave, donn at drizzle.com



From ianb at colorstudy.com  Wed Apr 16 14:53:31 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 16 Apr 2003 13:53:31 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <7kCdnQb9MqHQ2gCjXTWcow@speakeasy.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
	 <mailman.1049571010.25625.python-list@python.org>
	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
	 <mailman.1049653811.11491.python-list@python.org>
	 <cQ_la.5597$LB6.158861@news1.tin.it>
	 <mailman.1050210449.24337.python-list@python.org>
	 <23891c90.0304140221.20e9e7fb@posting.google.com>
	 <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
	 <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net>
	 <mailman.1050470255.12198.python-list@python.org>
	 <7kCdnQb9MqHQ2gCjXTWcow@speakeasy.net>
Message-ID: <1050519211.30714.426.camel@lothlorien>

On Wed, 2003-04-16 at 07:06, A.M. Kuchling wrote:
> I'm willing to host a mailing list for web discussion on my DSL-connected
> machine.  That might be easier than jumping through the SIG hoops.  If
> you're interested, just suggest a name ('web'?) for the list and I'll create
> it.  I can't do CVS, but maybe we could just use a sandbox/ directory in the
> Python CVS tree.

I've put a fair amount of effort into Web-SIG, and I'm not quite ready
to give up on it yet (other things in life have taken me away from it
for the last few days is all).

  Ian






From do-not-spam-ben.hutchings at businesswebsoftware.com  Wed Apr 16 06:15:18 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 16 Apr 2003 10:15:18 GMT
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <3e9af967$0$49112$e4fe514c@news.xs4all.nl> <ptnosvmm.fsf@morpheus.demon.co.uk>
Message-ID: <slrnb9qb9m.1bc.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <ptnosvmm.fsf at morpheus.demon.co.uk>, Paul Moore wrote:
<snip> 
> PS Is there any way of forcing the Windows 2000/XP console to display
>    Unicode in any form (UTF-8, maybe) rather than having to deal with
>    code pages???

UTF-8 is code page 65001.

Strangely, though, I get 'permission denied' when I run "chcp 65001" and
then try to print a UTF-8-encoded Euro sign.  I don't know what could be
going wrong there.



From johnroth at ameritech.net  Wed Apr 16 14:07:13 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 16 Apr 2003 14:07:13 -0400
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <v9r6m7864jmic6@news.supernews.com>

"Christopher Blunck" <blunck at gst.com> wrote in message
news:1147e466.0304160630.5b510724 at posting.google.com...
> Was just reading a thread about how python implements protected and
> private methods.  I found JP's response quite interesting (this is
> something I've heard numerous friends of mine say of the language):
>
> [In Python] everything is permissible, but not
> everything is opportune.  And that is very much Python's philosophy:
> rather than focusing on trying to make some things impossible (and
> generally failing -- a simple cast in most implementations of C++ lets
> you blast away any "protected" that a silly library designer may have
> tried to impose on you;-), Python empowers and trusts the programmer.
>
>
> The natural response a non-Python programmer has to this statement is,
> "what?!  you __trust__ the programmer?!  that doesn't work in most
> environments."  Despite the ignorance this response demonstrates, I
> none-the-less thought about the statement and the answer I came up
> with was "we write lots of really thorough tests to demonstrate
> functionality."
>
> That got me to thinking:  continuous testing is a cornerstone of the
> XP methodology.  <generalization>And Python programmers typically
> write lots of tests</generalization> (at least more than their Java
> counterparts from what I've seen).  That being said, can that
> generalization be extended to "Python programmers subscribe to the XP
> methodology"?

Well, there's one other reason. Python delivers a unit testing tool
as part of the standard library. Java doesn't. I suspect that this
subliminal hint has its effects.

> So how many of you guys use XP processes?

I do. Writing tests before code is something that I've thought would
be a good idea for literally decades, but it took both XP and having
the test tool delivered with the language to get me to start.

John Roth





From ianb at colorstudy.com  Wed Apr 23 03:02:42 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 23 Apr 2003 02:02:42 -0500
Subject: True/False
In-Reply-To: <m3r87tdbo3.fsf@mira.informatik.hu-berlin.de>
References: <mailman.1051036237.30103.python-list@python.org>
	 <3EA5B173.9420324C@alcyone.com>
	 <mailman.1051056037.8077.python-list@python.org>
	 <3EA5DE1E.22704437@alcyone.com>
	 <m3r87tdbo3.fsf@mira.informatik.hu-berlin.de>
Message-ID: <1051081361.701.287.camel@lothlorien>

On Wed, 2003-04-23 at 00:29, Martin v. L?wis wrote:
> Erik Max Francis <max at alcyone.com> writes:
> 
> > but at this point it's looking awfully ugly.
> 
> That's why it will take a long time until Python warns about
> assignments to True/False. Essentially, when Python 2.3 is considered
> an old Python version (i.e. at Python 4.0), the warning may be
> enabled. If the warning hasn't been seen for some time (say, at Python
> 6.2), it becomes an error. 
> 
> So the typical process is that people will have to actively break
> backwards compatibility at some time by removing the assignment.
> There is still a point in writing
> 
> try:
>   True
> except NameError:
>   True, False = 1, 0
> 
> as this gives you the "proper" boolean objects in 2.3+.

Note that the idiom I've been using (and which I've seen elsewhere) is:

True, False = 1==1, 0==1

Which retains proper boolean objects.  Or maybe one might use:

True, False = not 0, not 1

I'm a fan of True and False as being distinct from integers, which is
why I use the names, and I certainly wouldn't want to clobber their
proper values.

> b) import True, False from a module that doesn't get installed
>    in Python 2.3+
> try:
>    from boolcompat import *
> except ImportError:
>    pass

You could also make a boolcompat like:

# boolcompat.py
try:
    True
except NameError:
    from _boolcompat import *

# _boolcompat.py
True, False = 1==1, 0==1


This protects against SyntaxWarning and doesn't require try:except: in
the calling module.  But, despite these advantages, I won't be using 
it :)

  Ian






From pan-newsreader at thomas-guettler.de  Mon Apr  7 16:13:39 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Mon, 07 Apr 2003 22:13:39 +0200
Subject: Updating CVS
References: <5Nhka.139$N5.25540@mencken.net.nih.gov> <pan.2003.04.07.19.52.22.351156.788@thomas-guettler.de> <l5jka.141$N5.25497@mencken.net.nih.gov>
Message-ID: <pan.2003.04.07.22.13.38.651912.788@thomas-guettler.de>

> The line "sys.stderr = sys.stdin" redirects any errors to the browser.
> My question is less cgi and more cvs. I guess I didn't state it clear
> enough. I just want to know why the previous script doesn't update my
> CVS directory. Is it because I trying to run in from a web script and
> not as myself being logged in? I'm using an Apache web server running on
> Solaris.

It depends how your apache is configured. Mine is running suexec: All CGI
scripts are running with my UID. (http://localhost/~guettli/...)

But it seems that yours is running as nobody or www-data

os.system("who loves mom") might help.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From aleax at aleax.it  Mon Apr 28 10:54:29 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 14:54:29 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net> <nCUoa.41676$LB6.941312@news1.tin.it> <pan.2003.04.22.03.08.58.526973@mn.rr.com> <3japa.44301$LB6.997057@news1.tin.it> <OM8ra.21447$WQ4.16996@news2.central.cox.net>
Message-ID: <Fgbra.22722$K35.750403@news2.tin.it>

Steve Holden wrote:

> "Alex Martelli" <aleax at aleax.it> wrote in message
> news:3japa.44301$LB6.997057 at news1.tin.it...
> 
> [why Tim and Guido are dead good ...]
   ...
>> & (                      # conditions for a weak-notrump opener
>>     s ( 1 g 4  3  3  3   # shape 4333 (any), or...
>>         2 g 4  4  3  2   #       4432 (any), or...
>>         3   3- 3- 3- 5   #       5332 w 5 clubs, or...
>>         4   3- 3- 5  3-  #       5332 w 5 diamonds;
>>       )
>>     < ( 11                    # moreover, a 12-14 range for
>>         \+ SHDC c( 4 3 2 1 0) # normal Milton-Work point-count
>>         15
>>       )
>>   )
>>
>> There -- aren't you glad I don't do language design any more?-)
> 
> On this evidence it's arguable that you never did! This language was
> *designed*?

Well yes.  Got anything against prefix-notation?  I just found it
marginally easier to parse FUNC ( ARGUMENTS ) than (FUNC ARGUMENTS),
otherwise it would have been a bit more Lisp-ish in terms of syntax.

\FUNC <suits> <expression> "scans" the expression over the suits
(Spades, Hearts, Diamonds, Clubs, in this case) and "reduces" the
tuple of results by passing them as ARGUMENTS to the given FUNC --
making this somewhat germane to the thread's Subject;-)... it IS
rather like reduce(func, [expr(suit) for suit in suits]), except,
more compact because that's such a common construct in evaluating
a bridge hand;-).


> but-anyway-let's-all-be-grateful-for-python-ly y'rs  - steve

Hmmm, yes, definitely, let's;-).


Alex




From martin at v.loewis.de  Mon Apr 28 16:34:26 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 28 Apr 2003 22:34:26 +0200
Subject: zipfile with umlaut in filename
References: <3ea9a4bf$1_1@news.vo.lu> <b8caue$g65$05$3@news.t-online.com> <3eaa52b1_1@news.vo.lu>
Message-ID: <m3fzo28ip9.fsf@mira.informatik.hu-berlin.de>

Patrick Useldinger <uselpa at myrealbox.com> writes:

> Well, doing the same thing with WinZip works ok and keeps the name as
> it should.
> Also, reading *that* archive back in Python yields, as a result of
> namelist():
> ['Der h\x94fische Charakter der Liebe.doc']
> for a file called
> 'Der h?fische Character der Liebe'.

And why would you think that \x94 is "o-umlaut"?

It appears that Python puts the non-ASCII characters into the zipfile
as the system reports them (i.e. in the ANSI code page on your
system); winzip apparently converts them to some other encoding first,
probably code page 437 or code page 850.

Neither approach is more correct than the other - they are just
incompatible, and the zip format specification does not specify a
right approach. Non-ASCII is simply not supported in zipfiles (WinZip
supports it, zipfile.py supports it as well, but with a different
approach).

To get the effect of Winzip in Python, when adding a file with name N,
do

  zfile.write(N, unicode(N, "cp1252").encode("cp437"))

and then try to open the zipfile with Winzip.

HTH,
Martin




From mwh at python.net  Wed Apr  9 06:13:41 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 9 Apr 2003 10:13:41 GMT
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org> <kauu6b.f2d1.ln@boundary.tundraware.com> <7h3wui4528k.fsf@pc150.maths.bris.ac.uk> <mailman.1049826042.680.python-list@python.org>
Message-ID: <7h3r88c3rmy.fsf@pc150.maths.bris.ac.uk>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> writes:

> |Floats are perfectly normal rational numbers; it's the *operations*
> |which are weird.
> 
> Aren't you being slightly disingenuous here?

Maybe.  But you've chopped the bit of my post where I was making my
real point: floating point numbers (in the compter sense) are not best
thought of as intervals of real numbers.

> Any numbers are more-or-less only meaningful in terms of the operations
> they enter into.  The Peano integers, or the Cauchy or Dedekind Reals,
> are not really interesting because of the beauty of their construction,

Well, in the case of the reals, the construction is of no interest
because there is only one model up to unique isomorphism...

And you're talking to the wrong mathematician: I do not relish in the
Bourbaki-style games of building all of mathematics up from axiomatic
set theory.  I'll grudgingly grant that it may be important that it
can be done, but all attempts to show me the constructions will be
studiously ignored.

> but rather because it is relatively easy to define operations on them
> that match our intuitions about how arithmetic is -supposed- to behave.
> The operations are the whole point of the entities.

Well, yeah, but there's a perfectly well defined, injective map from 

    {CS-floats} -> Q

but it doesn't commute with the arithmetic operations.  That's all I
meant.

> Likewise, a float isn't *really* a number at all.  It's really just a
> pattern of bits, such as '01000001010000010100000101000001'.  Maybe
> those bits are a way of representing the string "AAAA", or maybe they
> are a way of representing a -thing- that enters into operations that act
> (a little bit) like addition, multiplication, etc.  Just being a bit
> pattern doesn't make it a *number* at all.

Well, be like that then :-) It depends what you mean by "float".  I
always try very hard to know which category I'm working in.  Is {-1,
1} just a random subset of the set of integers?  No, it's the group of
units of the initial object of the category of commutative rings!

This is one thing Grothendieck got absolutely right: one should
always, always, always consider the relative situation (i.e. think
about morphisms rather than maps).  So a float is a bit pattern such
as '01000001010000010100000101000001', but to be a float it has to
come with an arrow that explains it's interpretation.

Cheers,
M.

-- 
  ZAPHOD:  Listen three eyes, don't try to outwierd me, I get stranger
           things than you free with my breakfast cereal.
                    -- The Hitch-Hikers Guide to the Galaxy, Episode 7



From opengeometry at yahoo.ca  Tue Apr 29 19:46:04 2003
From: opengeometry at yahoo.ca (William Park)
Date: 29 Apr 2003 23:46:04 GMT
Subject: Permanent Clock
References: <d613c14.0304251200.33f1b741@posting.google.com>
Message-ID: <b8n2rs$b9rlr$1@ID-99293.news.dfncis.de>

Mauricio <inforcati at yahoo.com.br> wrote:
> Is there some way of accessing a permanent clock, i.e., a clock that
> works even if the system is rebooted or poewered off?
> I would like to increment a variable each day, independent of the
> system date (once this one users can change!), but not leaving my
> script running all the time. Better explaining: every time tha script
> is run, it must now if the current day is the same of tha last time
> the script was started.If not, increment tha variable!
> 
>  thanks for helping,

Replace the CMOS battery on your motherboard.

-- 
William Park, Open Geometry Consulting, <opengeometry at yahoo.ca>
Linux solution for data management and processing. 



From bokr at oz.net  Tue Apr 29 17:07:15 2003
From: bokr at oz.net (Bengt Richter)
Date: 29 Apr 2003 21:07:15 GMT
Subject: Problem subclassing tuple
References: <mailman.1051631814.18940.python-list@python.org> <2MmdnZmgXowvMTOjXTWcpA@comcast.com>
Message-ID: <b8mpi3$vq7$0@216.39.172.122>

On Tue, 29 Apr 2003 12:48:48 -0400, "Terry Reedy" <tjreedy at udel.edu> wrote:

>
>"John Wilson" <tug at wilson.co.uk> wrote in message
>news:mailman.1051631814.18940.python-list at python.org...
>> Given this class:
>>
>> class Holder(tuple):
>>     def __init__(self, *values):
>>         tuple.__init__(self, values)
>>
>>
>> When I try to create an instance:
>>
>> Holder(1, 2)
>>
>> I get:
>>
>> TypeError: tuple() takes at most 1 argument (2 given)
>>
>> It looks like the constructor for the superclass is being called
>rather than
>> the __init__ method on Holder.
>>
>> I'm obviously missing something here. Do I have to supply a __new__
>method?
>> Any help would be appreciated.
>
>
>Because tuples are immutable, one cannot mutate them in an init
>method, even to set the initial values.  So, as you apparently
>suspected, they are initialized in the __new__ method from a sequence.
>This is the source of the exception (see below). tuple.__init__ should
>be a no op.
>
>>>> class Holder(tuple):
>...      def __init__(self, *values):
>...          print "Holder init entered"
>...
>>>> Holder(1,2)
>Traceback (most recent call last):
>  File "<stdin>", line 1, in ?
>TypeError: tuple() takes at most 1 argument (2 given)
># init not entered
>>>> Holder((1,2))
>Holder init entered
>(1, 2)
>
>While Holder.__init__() cannot modify the value of self, it *can* add
>attributes (which surprised me a bit) so it can have a use.
>
>>>> class Holder(tuple):
>...     def __init__(self, seq):
>...         self.y = 1
>...
>>>> a=Holder((1,2))
>>>> a.y
>1
>
>I don't know what more you could do in a __new__ method.
>
I suppose you could mess with the value, e.g.,

 >>> class H(tuple):
 ...     def __new__(cls, seq):
 ...         return tuple.__new__(cls,seq+seq)
 ...
 >>> H([1,2,3])
 (1, 2, 3, 1, 2, 3)
 >>> H('abc')
 ('a', 'b', 'c', 'a', 'b', 'c')
 >>> h= H('abc')
 >>> h
 ('a', 'b', 'c', 'a', 'b', 'c')
 >>> type(h)
 <class '__main__.H'>

Regards,
Bengt Richter



From sholden at holdenweb.com  Fri Apr 18 11:42:16 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Fri, 18 Apr 2003 15:42:16 GMT
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <s1Vna.126364$It5.107742@news2.central.cox.net>

"Bill Martin" <wcmartin at vnet.net> wrote ...
> I'm wondering about the value of allowing a class definition like this:
>
>              class C:
>                pass
>
> Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or
> some such. If I try to print a.x2 or b.x1, I get an exception message
> basically saying those member variables don't exist. It seems to me this
> defeats one of the basic ideas behind OOP, that is to assign members to
> the class definition, thus ensuring that all instantiations of a class
> have the same member variables and methods. I know there must be a
> reason that python allows this kind of thing that I described, but I
> can't figure out what that reason could be. To me, it looks like a very
> dangerous thing to do. Can someone help me understand why we'd like to
> do it? Thanks,
>

Well, if you think it's dangerous you probably *don't* want to do it, but
it's the easiest way to set up a data structure with named components - you
are just taking advantage of the fact that each instance has its own
namespace. It's a firly standard idiom, though it isn't required.

Now, other object-oriented languages can go to great pains to make sure that
an instance's attributes can only be accessed in very controlled ways, such
as using accessor methods (getthis(), setthat(), etc.). Python, on the other
hand, pays you the compliment of assuming you know what you are doing, and
"gives you enough rope to shoot yourself in the foot"!

If you choose to abuse that freedom, you end up with a limp. [Please, nobody
ask "a limp what?"].

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From tim at bladerman.com  Tue Apr  8 08:39:47 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Tue, 08 Apr 2003 12:39:47 GMT
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com> <urnka.4658$Du.16154@newsc.telia.net> <b68f71a7.0304080323.75b5ed1a@posting.google.com> <2me59vcigrdkoeapbtb267i46jf07kr9gr@4ax.com>
Message-ID: <nqzka.4693$Du.16451@newsc.telia.net>

----- Original Message -----
From: "Rainer Klueting" <rainer.klueting at epost.de>
> On 8 Apr 2003 04:23:24 -0700, alan_salmoni at yahoo.com (Alan James

> >is a spreadsheet written in it called Novagrid which you might like to
> >try:
> >
> >http://sourceforge.net/projects/novagrid

> There is a folder on my hard disc called "download", and in there is a
> file named novagrid-1.0.tgz.gz. ;-) I downloaded it about a week ago
> but didn't try it.
>
> On my lengthy search for Python stuff I had already found some
> information on tkTable. May be I misunderstood something --I dropped
> that, because I thought, tkTable is written in TCL, and there is no
> Python-interface.

It seems you have done your homework well and looked up alot of interesting
stuf about Python, now lets just hope you get hooked like the rest of us :-)

Anyway, tkinter is the python.interface to tcl and tk so tkTable and
Novagrid should not be a problem to use.

Tim





From aussie2010 at yahoo.com  Thu Apr  3 18:30:10 2003
From: aussie2010 at yahoo.com (Chris)
Date: 3 Apr 2003 15:30:10 -0800
Subject: split encloser
Message-ID: <435d40e0.0304031530.1c197742@posting.google.com>

string.split() takes a delimiter and works fine as long as the
delimiter isn't part of the data fields. But frequently they are.
e.g. 'John Doe,135 South Main St.,#122, Springfield, Iowa' or
      ' so long goodbye see ya'

Because the fields can contain the delimiter in some cases, an
encloser is usually used (typically "") to handle those fields.

The above strings would be written:
'John Doe,"135 South Main St., #122", Springfield, Iowa'
and 
'"so long" goodbye "see ya"'

I don't understand regular expressions but I was wondering if anyone
that did knew of a way to get re.split() to handle "enclosers" as used
above.



From cben at techunix.technion.ac.il  Wed Apr  9 13:53:10 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Wed, 9 Apr 2003 20:53:10 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <3E9444E8.4070805@kyborg.dk>
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr>
 <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com>
 <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com>
 <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk>
 <3E9267F6.9020605@kyborg.dk> <yfsu1d9hqw6.fsf@black132.ex.ac.uk>
 <3E9444E8.4070805@kyborg.dk>
Message-ID: <Pine.GSO.4.51.0304092040520.1222@techunix.technion.ac.il>

Kim Petersen wrote on 2003-04-09:

> No misunderstanding... and possibly not efficient - thats why i said
> that you could do push/pop of bindings instead... since then i
> remembered that elisp (or any other lisp for that matter *doesn't* have
> global variables - so it would be inefficient). But to make it efficient
> and usable you can do it with a class instead (that simulates a dictionary):
>
> [snip]

True.  Perhaps doing it indeed with classes (without simulating
anything :-) would be even better:

class Global:
    """This is the top namespace in this example"""
    case_fold_search = 'nil'

def foo(env):
    print "When foo is called folding is:", env.case_fold_search
    env.case_fold_search = 'evil'

def bar(env):
    print "When bar is called folding is:", env.case_fold_search
    class local(env):
        case_fold_search = 't'
    foo(local)
    print "What did foo to folding?", local.case_fold_search
    print "bar's last folding is:", env.case_fold_search

def quux(env):
    print "before bar:", env.case_fold_search
    bar(env)
    print "after bar:", env.case_fold_search

quux(Global)

>>> ## working on region in file /usr/tmp/python-10913_eN...
before bar: nil
When bar is called folding is: nil
When foo is called folding is: t
What did foo to folding? evil
bar's last folding is: nil
after bar: nil

Of course, one can use instances instead of classes, I'm not sure this
is needed.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From glenfant at NOSPAM.bigfoot.com  Mon Apr 14 15:03:40 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Mon, 14 Apr 2003 21:03:40 +0200
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <mailman.1050336802.11273.python-list@python.org> <vfxhrn7a.fsf@morpheus.demon.co.uk>
Message-ID: <b7evu6$1lup$1@biggoron.nerim.net>

"Paul Moore" <gustav at morpheus.demon.co.uk> a ?crit dans le message de news:
vfxhrn7a.fsf at morpheus.demon.co.uk...
> Skip Montanaro <skip at pobox.com> writes:
>
> >     ...
> >     Paul> [BTW, in case it's relevant - I know it probably is - the
output
> >     Paul> from "chcp" at the console prompt is "Active code page: 1252"]
> >
> > Seems that cp1252 is the most likely encoding you didn't try.  What
happens
> > if you execute
> >
> >     print u'\u20a0'.encode("cp1252")
>
> Doh. Dumb typo. But no joy with that:
>
> >chcp 1252
> Active code page: 1252
>
> >python
> Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
> Type "help", "copyright", "credits" or "license" for more information.
> >>> print u'\u20a0'.encode("cp1252")
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "C:\Apps\Python\lib\encodings\cp1252.py", line 18, in encode
>     return codecs.charmap_encode(input,errors,encoding_map)
> UnicodeError: charmap encoding error: character maps to <undefined>
>

Anyway, the console glyphs collection used when typing "chcp 1252" is buggy.
The Euro symbol has the code 128 in cp1252 encoding. When typing Alt+0128 in
a console after setting "chcp 1252" writes an upper "?".
When writing a chr(128) to a file and then viewing it through an editor
shows the Euro symbol...

--Gilles




From max at alcyone.com  Tue Apr 22 17:05:42 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 14:05:42 -0700
Subject: coverting list back to string
References: <3cf72d52.0304211132.799d8afb@posting.google.com> <3EA44A5F.6ADA09B8@alcyone.com> <3cf72d52.0304220537.c63e6f1@posting.google.com>
Message-ID: <3EA5AEA6.494DBC36@alcyone.com>

scn wrote:

> Erik.  This worked and thank you.  However, the lines of text are all
> one continuous string instead of line-by-line.

That's because with no arguments, S.split() delimits by any chunk of
whitespace, and a newline qualifies:

>>> "The quick brown fox did something really gross.\n".split()
['The', 'quick', 'brown', 'fox', 'did', 'something', 'really', 'gross.']

So your .split operation is removing the newline that exists at the end
of every line that's bein read.  If you want to put it back in, just do
it manually with something like:

	output.write(' '.join(remove) + '\n')

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Do not stand in a place of danger trusting in miracles.
\__/ (an Arab proverb)
    PyQStat / http://www.alcyone.com/pyos/qstat/
 A Python wrapper around QStat, the realtime game server status tool.



From tanzer at swing.co.at  Tue Apr 15 03:08:38 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Tue, 15 Apr 2003 09:08:38 +0200
Subject: Using property in classic class may not work
In-Reply-To: Your message of "Tue, 15 Apr 2003 00:00:08 GMT."
             <cYHma.5800$hF.46142@news-server.bigpond.net.au>
Message-ID: <E195KYY-0004Ry-00@tswings.swing.cluster>

Neil Hodgson <nhodgson at eb2.net.au> wrote:

>     The "property" feature in Python 2.2 appears very helpful to me and
> seems to work on both classic and new-style classes. However, pychecker
> gives this error message:
>
> DemonTerm.py:563: Using property (frame) in classic class DemonFrame may
> not work
>
>     The class in question is derived from a wxWindows class, so I can't
> make it derive just from object. The "What's New in Python 2.2" document
> only mentions property wrt new-style clases. Can I depend on property
> working for classic classes?

>>> class T :
...   def _get_x(self) :
...     return 42 * self.i
...   x = property(_get_x)
...   def __init__(self) :
...     self.i = 5
...
>>> t = T()
>>> t.x
210
>>> t.i+=1
>>> t.x
252
>>> t.x = "foo"
>>> t.x
'foo'

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From max at alcyone.com  Fri Apr  4 17:51:23 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 04 Apr 2003 14:51:23 -0800
Subject: Converting text to html
References: <3E8E0681.9005524D@easystreet.com>
Message-ID: <3E8E0C6B.66A2E5FD@alcyone.com>

achrist at easystreet.com wrote:

> Is there any recommended code available to convert plain text into
> text that can go into an html document without breaking it?  I
> think that I just have to do these three translations:
> 
>         < =>  &lt;      > =>  &gt;   &  =>  &amp;
> 
> But I'm not sure if there are any others.  I've got to do this
> quite a bit while the user is waiting, so it would be nice to
> have something that did it in a way that was not a lot slower
> than any of the alternatives.

Use the cgi.escape function.  It ha san optional "quote" parameter that
will also convert double quotes to &quot; for you as well.  Normally you
don't need that, but if you plan on putting something in a quoted
section you would.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If the sky should fall, hold up your hands.
\__/ (a Spanish proverb)
    The laws list / http://www.alcyone.com/max/physics/laws/
 Laws, rules, principles, effects, paradoxes, etc. in physics.



From pobrien at orbtech.com  Sun Apr  6 13:39:17 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 12:39:17 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <7xn0j3fsdu.fsf@ruckus.brouhaha.com>
Message-ID: <m21y0fa5kq.fsf@localhost.localdomain>

Paul Rubin <http://phr.cx at NOSPAM.invalid> writes:

> pobrien at orbtech.com (Patrick K. O'Brien) writes:
> > If you were crazy enough to think that you could create a program
> > along the lines of Emacs, but written in Python, how would you go
> > about it?  How would you design the domain model for files, buffers,
> > windows, and frames?  How would you allow the same level of
> > customizability?  How would you map functions (or methods or whatever)
> > to keys?  Any thoughts?
> 
> I think you're going to need some C extensions to represent buffers
> similar to how Emacs does it.  It's hard to move blocks of characters
> around in Python with the array module without constantly re-allocating 
> the arrays.  Of course it can't be done with strings because Python
> strings are immutable.

The Scintilla text control (written in C and wrapped in wxPython by a
control called wxStyledTextCtrl) handles the on-screen representation
and user-manipulation of the text.  So that part is actually easy.
Just read in a file, feed it to Scintilla, and go.  The next step is
wrapping all this in an Emacs-like framework.

> Aside from that, the general architecture can probably be similar to
> Emacs.  You might want to use a multi-threaded interpreter to make
> extensions more flexible (e.g. so they can respond to keystrokes while
> doing other stuff).  You generally won't need as much C code as Emacs
> used, because computers are much faster now than when Emacs was
> written.  Python and Emacs Lisp are fairly similar in speed, I think.

I shouldn't need any C, hopefully, since I'm using Scintilla.
Mimicking the Emacs architecture is where I want to focus next - how
to represent commands and keybindings and user-coded extensions and
customizations and so forth.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From bdesth at removethis.free.fr  Wed Apr  9 15:39:40 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 09 Apr 2003 19:39:40 +0000
Subject: Newbie: Python + C Question
References: <3e92ebe8$1@rmcnt03a.rmc.ca> <3E933C8D.5020700@removethis.free.fr> <3E93586E.D4AB1531@engcorp.com>
Message-ID: <3E9476FC.4050007@removethis.free.fr>

Peter Hansen wrote:
> laotseu wrote:
> 
>>bourassa wrote:
>>
>>>When Python uses the C extension does it just process
>>>it in an interpreted fashion
>>
>>Python is not a C interpreter.
>>
>>
>>>or does the code written in C run like a
>>>compiled C program?
>>
>>newbie to programming, uh ?-)
>>
>>How do you think this all works ?
> 
> 
> Actually, to be fair to the OP, this is a valid and even interesting
> question.  Where, after all, does it say that the C code that is
> in the extension is actually pre-compiled into an executable and
> is run as machine code when called by Python?  (Maybe it's there
> in the docs, but maybe it's not or maybe it's unclear.  Have you
> checked?)

Personnaly, no. BTW I didn't mean to be harsh (and I do apologize if it 
could be taken this way, the fact is that my english is somewhat limited 
so I can't always express my thoughts as clearly as I would like).

The fact is that most of Python docs seems to take for granted that 
everybody kons at least what are a compiled language, an interpreted 
language, byte-code, machine code etc... And the fact is that most 
programmers knows at least enough about those topics to get the picture. 
Hence my reflexion about the OP being newbie to programming in general.

> I could picture someone believing that Python had the ability to
> execute C code interpretively via some fancy C-interpreter
> extension, which would allow the existing code-base of C to be 
> used within an otherwise-pure-Python program.  

I must admit I never personnally imagined such a thing before reading 
the OP question ! BTW, most of the 'existing code-base of C' being 
mainly distributed as binaries, I think the misleading part of it is the 
'C extension' term itself.

Laotseu




From exarkun at intarweb.us  Sat Apr 19 13:47:48 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sat, 19 Apr 2003 13:47:48 -0400
Subject: Testing methods
In-Reply-To: <9d8ac130.0304190938.755d4b4c@posting.google.com>
References: <9d8ac130.0304190938.755d4b4c@posting.google.com>
Message-ID: <20030419174748.GA27805@meson.dyndns.org>

On Sat, Apr 19, 2003 at 10:38:40AM -0700, Ovid wrote:
> Hello,
> 
> Relatively new to Python and I thought I would teach myself the
> language by building a test harness.  After I'm done, I will probably
> be throwing it away in favor of Pyunit, but it seemed like a good way
> to learn some of the fiddly bits.
> 
> Fiddly bits such as "can this object do that?"  In my test code, I
> want to test whether or not a given class *or* instance of that class
> can execute a particular method (i.e., if it has or inherits the
> method).  For example, one way (in Perl), to test if object "$foo" can
> execute an arbitrary method would be:
> 
>   if ( $foo->can($some_method) )
> 
> For the life of me, I can't figure out the equivalent in Python.  I
> suppose I could do:
> 
>   if some_method in Foo.__dict__:
> 

  The Pythonic (Gee, I hope I'm still allowed to use that word) way to do
this is simply:

    try:
        foo.some_method()
    except:
        print "Foo can't do it"
    else:
        print "Foo can do it"

  Sometimes this is undesirable, but then the right answer depends on the
particulars of why it is undesirable.  There are other approaches, but none
that work as reliably.  For example:

    if hasattr(foo, 'some_method') and callable(foo.some_method):
        ...

  callable() may return false positives, though.  I don't think it will ever
return false negatives.

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 30 days, 13:02, 4 users, load average: 0.29, 0.22, 0.16




From aleax at aleax.it  Sun Apr 13 05:22:23 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 09:22:23 GMT
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <b7aj86$q9i$1@slb6.atl.mindspring.net>
Message-ID: <j%9ma.9030$T34.282951@news2.tin.it>

Matt Gerrans wrote:

> Learning Python is an excellent book, but unfortunately quite out of date.

Yes to both.  Fortunately, a new version IS in the works, and indeed
at quite an advanced stage...


Alex




From tebeka at cs.bgu.ac.il  Sun Apr 13 04:00:46 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 13 Apr 2003 01:00:46 -0700
Subject: wxPython - Can you explain why I crash
References: <33803989.0304090652.174ac9de@posting.google.com> <3E946842.5D897F5@engcorp.com>
Message-ID: <33803989.0304130000.3db41505@posting.google.com>

Hello Peter,
> > I've written this little client. It used to work fine but now no
> > matter which button I press the application crash.
> > I'm running Python2.2 and wxPython 2.4.0.7
> > Any ideas?
> 
> [snip 480 lines of code]
> 
> I have an idea: spend a little of your own time cutting down the
> program to a *much* smaller example which still crashes.  Then,
> and only then, post it to this newsgroup and ask for others' help.
Thanks Peter (and I'm not being ironic). I did just that and found the
problem.
I gave the 'Quit' button id of -1 and looks like it caught every
button click and terminated the program.

Next time I'll trying using my brain instead of the newsgroup :-(

Miki



From xxdigitalhellxx at hotmail.com  Thu Apr 24 10:10:05 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Thu, 24 Apr 2003 08:10:05 -0600
Subject: ftp problem on solaris machines with python
In-Reply-To: <88bc63c6.0304240407.674220cb@posting.google.com>
References: <88bc63c6.0304240407.674220cb@posting.google.com>
Message-ID: <b88rae$etu$1@heckle.cc.uregina.ca>

Doug Farrell wrote:
> Hi all,
> 
> We've written an installation tool in Python2.2 that tars up
> directories and ftp's the files over to the destination machines. This
> works great from linux machines to linux machines. However, going from
> solaris to solaris machines the process croaks if the tar files being
> transferred are to big. Has anyone seen problems like this using
> Python on solaris? And if so, any suggestions?

I assume that you're gzipping them, too?

What do you mean by 'croaks'?

-- 
Ryan




From gh at ghaering.de  Mon Apr 28 10:06:07 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: 28 Apr 2003 14:06:07 GMT
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com>
Message-ID: <slrnbaqdae.1gg.gh@haering.opus-gmbh.net>

Aahz <aahz at pythoncraft.com> wrote:
> In article <b8hc64$6n2$1 at slb9.atl.mindspring.net>,
> Andrew Dalke <adalke at mindspring.com> wrote:
>>
>>        i = i + 1
> 
> I'm a little curious why you didn't write this as 
> 
>     i += 1
> 
> Any reason (i.e. still need to support 1.5.2) or just habit?  If the
> latter, why haven't you changed your habit?

I use += and friends myself. But one could argue that the ability of the
inplace operators to return new objects is everything but clean.

-- Gerhard



From Dick.Zantow at lexisnexis.com  Thu Apr 17 09:09:57 2003
From: Dick.Zantow at lexisnexis.com (rzed)
Date: Thu, 17 Apr 2003 09:09:57 -0400
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com> <h3qna.143255$vs.16245277@news3.calgary.shaw.ca>
Message-ID: <b7m94s$ut$1@mailgate2.lexis-nexis.com>

Parzival Herzog wrote:
> "Aahz" <aahz at pythoncraft.com> wrote in message
> news:b7kqr8$jsq$1 at panix2.panix.com...
>> In article <Wrena.33096$T34.876856 at news2.tin.it>,
>> Alex Martelli  <aleax at aleax.it> wrote:
>>> Steve Holden wrote:
>>>>
>>>> You'd do better to call it "Python for VB programmers",
>>>> expecially since the change to VB6 is so radical that migrating
>>>> to a new and more sensible language might be an attractive
>>>> option to some!
>>>
>
> Maybe when you're done that, you could do "Python for RPG/400"
>
>      H*
>      H* RPG/400 VERSION OF THE BOTTLES PROGRAM *
>      H*
>      FSCREEN  O   F      80            WORKSTN
>      C                     MOVE 100       X       30
>      C           X         DOWGE0
>      C                     EXCPT
>      C                     SUB  1         X
>      C                     END
>      C                     SETON                     LR
>      OSCREEN  E
>      O                         X          3
>      O                                   26 'BOTTLES OF BEER ON THE'
>      O                                   31 'WALL,'
>      O                         X         36
>      O                                   53 'BOTTLES OF BEER'
>      O        E
>      O                                   22 'TAKE ONE DOWN AND PASS'
>      O                                   32 'IT AROUND'
>
> P.S. Can you believe it, at version 400 kicks at the can, they came
> up with this?
>

A little off the thread's topic, but ... I always liked RPG, though it
is far more pleasant when you stay reasonably close to its main
purpose -- to generate reports. I'm not sure what the "400 kicks at
the can" means, but I'm quite sure this is not version 400, but the
version of RPG used by the AS/400. The code looks essentially the same
as RPGIII, which was what I used first.

--
rzed





From oren-py-l at hishome.net  Wed Apr  2 02:02:32 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Wed, 2 Apr 2003 02:02:32 -0500
Subject: [Q]How to append paths in a platform-independent way?
In-Reply-To: <0d7ql-5t3.ln1@beastie.ix.netcom.com>
References: <420ced91.0304011946.56a22d2a@posting.google.com> <0d7ql-5t3.ln1@beastie.ix.netcom.com>
Message-ID: <20030402070232.GA54138@hishome.net>

On Tue, Apr 01, 2003 at 09:00:48PM -0800, Dennis Lee Bieber wrote:
> sdhyok fed this fish to the penguins on Tuesday 01 April 2003 07:46 pm:
> 
> > How to do the following jobs in a platform-independent way?
> > 
> > path = /home/sdhyok
> 
>         This is already platform dependent... it doesn't mean the same thing 
> in Windows.

Actually, it does. Windows accepts slash-separated paths just fine at the 
API level. Even mixed use works ok. The command processor and other 
command-line utilities require backslashes because they use forward slash 
for options. 

MacOS (not OS X) is probably the only major OS that does not accept 
slashes (any Mac owners out there can confirm that?)

    Oren




From ulope at gmx.de  Thu Apr  3 07:48:37 2003
From: ulope at gmx.de (Ulrich Petri)
Date: Thu, 3 Apr 2003 14:48:37 +0200
Subject: writing to windows memory
References: <3E8C01CE.6070601@netscape.net>
Message-ID: <b6haj8$5e8ne$1@ID-67890.news.dfncis.de>

"vector" <Vector180W at netscape.net> schrieb im Newsbeitrag
news:3E8C01CE.6070601 at netscape.net...
> newbie:
> pythonwin 2.2.2
> Id like to bind a button.widget that simply writes a one to windows
> address 440h. any ideas on where to start.(the button stuf is trivial) i
> need to know how to access the address space.

What do you exactly mean by  "windows address 440h"? Imho there is noch such
thing.
You seem to try some low level hacking, perhaps youre better off with C/C++
?

What is the goal of your program?

Ciao Ulrich





From jepler at unpythonic.net  Wed Apr  2 15:09:45 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Wed, 2 Apr 2003 14:09:45 -0600
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to Python. Good idea? Bad Idea? PEP?
In-Reply-To: <20030402155232.GB4884@meson.dyndns.org>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com> <20030402155232.GB4884@meson.dyndns.org>
Message-ID: <20030402200944.GB32009@unpythonic.net>

> On Wed, Apr 02, 2003 at 09:14:24AM -0500, WP wrote:
> > Of course, it all comes I suppose from integers being "first class objects" 
> > and thus, not actually integers, they are actually immutable objects which 
> > contain an integer, and so we get to the ugly condition where "i++" 
> > actually decref's one object, and either looks up or creates another 
> > object, and increfs it.  Why does it come from this?
> 
On Wed, Apr 02, 2003 at 10:52:32AM -0500, Jp Calderone wrote:
>   Furthermore, what language are you talking about, when you talk about
> "i++"?  Surely not Python.
> 
> > Hardly a fast operation to implement, but certainly one of the most common
> > idioms in all of programming.
> 
>   Ahh, worrying about efficiency again.
[...]

I implemented a scheme for object tag bits, so that the PyObject* could
either store ints in the range -2^29..^29-1 or a pointer to a full
Python object.

There was no performance increase, because many places in Python (anywhere
a PyXXX_Check macro was invoked, PyObject_CheckFull() must be called to
see if it's a full object or not) an extra branch was introduced to check
the tag bit.  The incref and decref macros had an additional branch too.
BINARY_ADD and other 'inlined in ceval.c for int OP int case' instructions
were complicated, and yet another set of overflow points were created.
(I don't recall whether I overflowed to Python int or long at that
point)

At the same time, I kept discovering bugs all along due to more places
where PyObject_CheckFull() was needed but not called (for instance, in
type()).  Those bugs would be a long time showing up.

In summary, I demonstrated to my own satisfaction that 'ints are full
objects, not native types' is probably not the performance problem in
any Python code I use, and taking a lisp-style route to fix the
'problem' is a step in the wrong direction.

Jeff




From dcengija_remove_ at inet.hr  Fri Apr 11 07:26:32 2003
From: dcengija_remove_ at inet.hr (Davor Cengija)
Date: Fri, 11 Apr 2003 13:26:32 +0200
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com> <b766fi$p3b$1@newstree.wise.edt.ericsson.se>
Message-ID: <16411664.REEpLcBOk5@lothar.cengija.com>

Frithiof Andreas Jensen wrote:

> 
> "kkennedy" <titanrebel at comcast.net> wrote in message
> news:14bd1399.0304091426.1b08e005 at posting.google.com...
>> I would like it to be smaller than
>> the 7 MB download available from python.org, and be able to be
>> installed transparent to the user.
> 
> On windows, Biig == Good!!
> 
> Users are so used to huge installs & downloads that they expect them and
> they will be suspicious of anything that takes less than 30 minutes to
> install - especially if it costs money. The "no-brain-install" is the
> right idea, thoug.

(really off-topic, but worth of mentioning)

I wrote a simple python program with gui for a friend of mine and his 
projects at local university. His co-workers didn't want to use it since 
'it's too small, obviously it doesn't do what we need'. It's about 45k of 
python source. Then I created self-standing exe with mcmillan's installer. 
It's now about 7mb in size. Now it looks 'respectable' and 'worth of trying 
but not for serious work since it doesn't have setup.exe with EULA and 
other things -- so it isn't a real program'. Now I'm looking for a free 
install-shield like software.

So, it isn't the quality of a software what only counts. Actually, I'm not 
sure that the quality is the most important thing these days.
-- 
Davor Cengija, dcengija_remove_ at inet.hr



From mis6 at pitt.edu  Fri Apr 18 17:08:11 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 18 Apr 2003 14:08:11 -0700
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <2259b0e2.0304181308.316737a4@posting.google.com>

Bill Martin <wcmartin at vnet.net> wrote in message news:<3EA01A9B.4030209 at vnet.net>...
> I'm wondering about the value of allowing a class definition like this:
> 
>              class C:
>                pass
> 
> Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or 
> some such. If I try to print a.x2 or b.x1, I get an exception message 
> basically saying those member variables don't exist. It seems to me this 
> defeats one of the basic ideas behind OOP, that is to assign members to 
> the class definition.

Some time ago I posted this recipe to make Python anal in the way you
want:

  def frozen(self,name,value):
      if hasattr(self,name):
          object.__setattr__(self,name,value) # standard __setattr__
      else:
          raise AttributeError("You cannot add attributes to %s" % self)

  class Frozen(object):
      """Subclasses of Frozen are frozen, i.e. it is impossibile to add
       new attributes to them and their instances"""
      __setattr__ = frozen
      class __metaclass__(type):
          __setattr__ = frozen


 Cheers,


                  Michele



From frithiof.jensen at removethis.ted.ericsson.dk  Wed Apr 30 11:13:53 2003
From: frithiof.jensen at removethis.ted.ericsson.dk (Frithiof Andreas Jensen)
Date: Wed, 30 Apr 2003 17:13:53 +0200
Subject: Self modifying systems question
References: <mailman.1051708787.23125.python-list@python.org>
Message-ID: <b8op7o$q90$1@newstree.wise.edt.ericsson.se>

"Sten Kvamme" <skvamme at as4-2-8.an.g.bonet.se> wrote in message
news:mailman.1051708787.23125.python-list at python.org...
> Hello,
> Thanks to all of you who answered my question.

There was an example on modifying objects at runtime on
http://aspn.activestate.com/ASPN/Python/Cookbook/ - search for "mixin" (or
"mix in") which is the OO term.

> Is it also possible to save the changes made to an instance in a
> persistant way?

The ZODB can do that - http://www.zope.org/Wikis/ZODB/StandaloneZODB - has a
link to an "Object Database Only" installation of Zope via. distutils.






From brian at sweetapp.com  Thu Apr  3 18:32:08 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Thu, 03 Apr 2003 15:32:08 -0800
Subject: distributed computing implementations
In-Reply-To: <m3of3n6vfu.fsf@mira.informatik.hu-berlin.de>
Message-ID: <005801c2fa39$3e9d20c0$21795418@dell1700>

> It is true that you have to write interface definitions [for CORBA]. 
> However, considering that you are doing distributed computing, I 
> can't really see this as "overhead". Overhead compared to what?

XML-RPC and SOAP don't require interface definitions. And how is it not
overhead? I'm sure that a lot of C++/Java programmers NEVER think of
defining interfaces as overhead.

> Code written by yourself? Compared to what? A CORBA client in Python
> is really short. 

This short?

from xmlrpclib import Server
s = Server(uri)
s.method(*args)

> A CORBA server is larger, but then, the servers
> for other distributed computing infrastructures are also larger.

True. XML-RPC servers are huge compared to XML-RPC clients:

from SimpleXMLRPCServer import SimpleXMLRPCServer
server = SimpleXMLRPCServer(("localhost", 8000))

server.register_function(func1)
server.register_function(func2)

server.server_forever()

> That depends on the implementation you use. For a client, it is
> certainly true that XML-RPC libraries are significantly (factor 10)
> smaller than the Fnorb libraries.

Are the CORBA Python libraries written in pure Python?

Cheers,
Brian





From pan-newsreader at thomas-guettler.de  Wed Apr  9 02:58:37 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 09 Apr 2003 08:58:37 +0200
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <pan.2003.04.09.08.58.37.390478.811@thomas-guettler.de>

On Wed, 09 Apr 2003 08:45:42 +0200, Jim Shady wrote:

> Hello,
> 
> I am a beginner at Python and I just joined this group.
> 
> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string. 

Hi,

this is one solution

s="x/aa/"
s=s.split("/")
max_item=None
max_len=-1
for item in s:
    if len(item)>max_len:
        max_item=item
        max_len=len(item)
print "Max:", max_item, max_len

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From randall at tnr.cc  Tue Apr 15 02:58:55 2003
From: randall at tnr.cc (Randall Smith)
Date: Tue, 15 Apr 2003 06:58:55 GMT
Subject: vim configuration help
Message-ID: <P4Oma.78974$rd4.2975295@twister.austin.rr.com>

I'm learning Python and I've been looking for a good IDE to use with it. 
    I haven't found an IDE that has text editing capabilities close to 
Vim/Emacs.

Much of my previous programming is with Emacs/PHP and I love the 
auto-indenting, tab-indenting, well thought-out key strokes, and the 
ability to ssh into a server and run Emacs.

I'd like to try out Vim while writing Python code (I'm using Webware). 
Right now, the default settings using color coding, have an 8 space tab 
and no autoindenting.  I would like to have a tab space of 2 or 4 spaces 
and have auto-indenting.

I have no idea how to configure Vim.  I'm a total newbie.  Can someone 
please help me get started.  I'm looking to configure Vim for Python. 
Autoindenting and 4 space tabs are high priority now.

Thanks,

Randall




From gabnash at hotmail.com  Wed Apr 23 04:03:33 2003
From: gabnash at hotmail.com (David)
Date: 23 Apr 2003 01:03:33 -0700
Subject: capturing output from os.system() in Windows?
Message-ID: <f405d9e7.0304230003.63baab05@posting.google.com>

Being a complete novice at python, I thought I'd try a simple script
to run a NET VIEW (which runs in a DOS box) command under Windows and
then manipulate the resulting data.

The first problem I have encountered is how to capture the output from
the 'NET VIEW' command.

os.system('NET VIEW') returns 0 under windows, although the DOS box
pops up and the NET VIEW executes correctly.

Is there some other command I should be using instead? I guess an
alternative might be to run the 'NET VIEW' under dos and redirect the
output into a python script using a BAT file?

Any ideas?



From aleax at aleax.it  Wed Apr 16 12:55:33 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 16:55:33 GMT
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <v9qot54nmk87cb@news.supernews.com> <u1cyqvfi.fsf@python.net>
Message-ID: <9Wfna.34058$T34.892831@news2.tin.it>

Thomas Heller wrote:
   ...
> Hm, the latest version of the patch which I posted typically prints this:
> 
> ======================================================================
> FAIL: test_failUnlessEqual (__main__.FailingTests)
> ----------------------------------------------------------------------
> TestFailed: 0 != 1
>   File "xunit.py", line 12, in test_failUnlessEqual
>     self.failUnlessEqual(self.a, self.b)
> 
> Is this poorly formatted? I won't argue about this.

It's sure WAY better than what it does out of the box!!!

> Anyway, this is IMO the information I want to have:
> filename, line number, source code line, and the actual#
> values which made the test fail.

Seconded.

> Concerning the 'more is better': If the test crashes (raises an
> unexpected exception), the full traceback is printed.
> If the test fails because a condition is not fulfilled, I'm *not*
> interested to see how unittest raises this error.

Absolutely!!!  I think unittest would be MUCH better with
your mods.


Alex




From axs at arbornet.org  Fri Apr  4 21:34:05 2003
From: axs at arbornet.org (axs)
Date: 4 Apr 2003 18:34:05 -0800
Subject: I'm new to programming
References: <oprm4ogjb9acxezr@news.datafast.net.au>
Message-ID: <6478eb43.0304041834.76c4574f@posting.google.com>

Casey Jones <cajones at datafast.net.au> wrote in message news:<oprm4ogjb9acxezr at news.datafast.net.au>...
> -- 
> Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
> 
> I've been parusing the 'Learn Python in an Afternoon' tutorial and my head 
> feels comfortable with it's
> syntax etc. I've looked at VB, C, C++, blah it gives me a friggin 
> headache!..
> 
> I suppose the problem is a goal "What the hell do I want to program and I 
> just don't know".
> I feel that it would give me a better understanding if I had a goal.
> 
> I know html I can look at Javascript and know what to look for when editing 
> for roll overs etc but get
> me to program something from scratch huh no can do I'm at a complete loss.
> 
> It's like I guess. "I know the moves in chess but ask me to checkmate a 
> chessmaster in three moves forget it".
> 
> Inspiration would be appreciated.

Look at sourceforge.net for some projects. Start simple. Incredibley
simple. Like parsing files or moving files or opening sockets and
readin sockets. Simple simple simple. Look at topcoder.com for some
puzzle problems.



From missive at frontiernet.net  Fri Apr 11 09:55:41 2003
From: missive at frontiernet.net (Lee Harr)
Date: Fri, 11 Apr 2003 13:55:41 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <3e96b088_2@127.0.0.1>
Message-ID: <xPzla.367$xm2.20@news02.roc.ny.frontiernet.net>

In article <3e96b088_2 at 127.0.0.1>, Jason Smith wrote:
> 
> I mean, if we're all studying sorting, and I turn in
> 
>   a = [ 'this', 'is', 'a', 'list' ]
>   a.sort()
> 
> well then clearly I just got an F.  But it's not Python's fault.
> 


Why? This looks good to me.

Well... ok. I see your point. But if you are studying sorting algorithms,
a.sort() should definitely be on the list :o)  And then I think almost
any student will see the advantage of a.sort() over a page of intricate code.

What I mean is that students should first learn how to use the tools before
they are required to know how to make them.

I think going the other way around just encourages the poor practice of
everyone creating their own incompatible and (very often) buggy private
libraries.




From mhammond at skippinet.com.au  Sun Apr  6 18:57:51 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Sun, 06 Apr 2003 22:57:51 GMT
Subject: Debugging embedded python
In-Reply-To: <u4r5bgyfx.fsf@online.no>
References: <u4r5bgyfx.fsf@online.no>
Message-ID: <Ph2ka.2428$Tb6.22760@news-server.bigpond.net.au>

Syver Enstad wrote:
> I have the python interpreter embedded in a daemon/service (this is
> win2k so it's a service). I'd really like to be able to debug the
> python code that is running. I was thinking of popping up a console
> window to be able to run pdb, how does one go about it to accomplish
> this? Are there any code doing similar things or articles.

You should be able to run the service in "debug" mode (ie, as a normal 
program" by using the "--debug" option.  Then you can debug however you 
like.

But assuming you do need to debug as a service, you really need an 
out-of-process debugger, like Komodo and most of the commercial IDEs. 
Pythonwin does *not* do out-of-process debugging, so is no good.

Mark.




From a.schmolck at gmx.net  Tue Apr 15 17:46:38 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 15 Apr 2003 22:46:38 +0100
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
Message-ID: <yfs7k9vcu2p.fsf@black132.ex.ac.uk>

Dave Benjamin <ramen at lackingtalent.com> writes:

> The next thing I read will probably not surprise you if you have ever done a
> Google search on functional programming. It was David Mertz's "Charming
> Python" series on IBM's DeveloperWorks. His papers were not easy to digest
> in one sitting, but I made a great effort to understand them and apply the
> techniques he described. At this point, I should mention that I was still
> programming only in PHP!

I think your life might be much happier if you just went to
http://www.drscheme.org/, dowloaded Dr Scheme and used *that* for
webdevelopment. Then you can use a real functional programming language and
read books by people who actually have a real clue about functional
programming (there is lots of *excellent* material available on-line and for
free, like SICP and HTDP, the latter uses Dr Scheme -- also have a look at
http://www.cs.brown.edu/courses/cs173/2001/Lectures/). Why settle for some
second-rate immitation? [1]

(Dr Scheme *does* have web libraries, I've never used them but I'd suspect
they shouldn't be too bad, because they seem to have hit on trick of using
webdevelopment as application domain to bring things like continuations to the
unwashed masses (viz. their sophomores) :).

>    - lambda
>    
>    Lambda has had the thorniest history of all of these, I think. Lambda has
> its roots in LISP and the Lambda calculus, and can be used to create
> "anonymous functions" at runtime that can be passed around as needed. This
> is a very powerful and useful feature for functional programming, as well as
> for event-driven programming. It baffles me that anyone would want to remove
> it.

Ain't gonna happen. Anyway you can always replace 

  foo(lambda x: x+3)
with: 
  def bar(x): return x + 3
  foo(bar)

 
> keyword-argument trick wasn't really that bad, once you got used to it. In
> any case, with dynamic scoping, this problem has completely disappeared.
                 ^^^^^^^^^^^^^^^
not quite.


> The second "problem" that everyone seems to keep bringing up is that lambda
> limits you to expressions, not statements. This makes me laugh every time,
> because this is one of the things that I *loved* about Python when I was
> first learning it! If you get used to writing functions that only return
> expressions, you are on the road to understanding the functional style! One

Only that python won't let you use rather helpful things like ``if`` as
an expression, but never mind.

> In fact, I believe that it is advantageous to design functions specifically
> for use with map and filter. If you design this way, you will benefit from
> cleaner syntax than list comprehensions can offer. For example:

I wonder how a function written for map would look different from one written
for a list comprehension...

> Rather than remove these important functional built-ins, I would like to see
> two last built-ins be added: unzip (to reverse the behavior of zip), and

There is already an unzip function and it's written ``zip(*...)``

> curry (see the example by Alex Martelli, et. al. in the Python Cookbook,

Why not just use lambda? Python's argument passing is fairly complex, so I'm
not sure whether one (or rather two) curry function(s) would be worth the pain
(at least not if you tell your emacs to display lambda as greek character).

'as

Footnotes: 
[1]  I'm talking about *functional* programming.



From a.schmolck at gmx.net  Tue Apr 15 20:03:11 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 16 Apr 2003 01:03:11 +0100
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net>
Message-ID: <yfsr883b96o.fsf@black132.ex.ac.uk>


Carl Banks <imbosol-1050288268 at aerojockey.com> writes:

Firstly:

> > Have you had traumatic childhood experiences with early perl or lisp
> > dialects?  :)
> 
> Yes

OK, let's try this. Just consider for a moment the following possibility:

  THERE IS NOT A *SINGLE* DRAWBACK TO DYNAMICALLY SCOPED *GLOBALS*.

Shocking, isn't it? And quite possibly wrong, but let's just forget about that
for the moment, OK? I really think that discussion (and replying to your
posts) is unduly complicated by the fact that you (maybe understandably) seem
to suffer some trauma induced kneejerk reflex.

As I tried to show (unsuccessfully it seems) in my last post, I see no reason
why dynamically scoped globals should in any way more error prone than the
current globals (or require "defensive naming schemes"). OK, let's pretend
there are no packages or classes and that:

  global foo 

declares a *dynamically scoped* variable. No variable in a function that
hasn't got such a declaration, repeat after me, *will ever be dynamically
scoped* -- you can *always* tell by just looking at a function itself, without
context, what the scope of its variables is. No difference for the worse
compared to python's current scheme, right? That's because it still is exactly
the same, up to now. OK, so how do we get the funky (let ((*foo* 3)) ...)
style shadowing? Or how do we introduce a non-global dynamically scoped
variable?. Simply with:

  global foo = 3 # shadows foo (or introduces a new dynamically scoped
                 # variable only visible to callees, and *ONLY* if those
                 # callees contain a ``global foo`` or ``global foo=...``
                 # statement)

so in summary:
  
  foo = 3       # lexical

  global foo    # declares dynamically scoped variable foo (if it doesn't
                # exist, it is created a global scope), pretty much like 'old'
                # global
 
  global foo=3  # shadows foo or declares and binds 3 to it

Maybe this is stupid and maybe it doesn't work and no I haven't thought deeply
about packages and classes so far (which as Beni already mentioned seem to
render a literal mapping of CL's approach unattractive) but this is what I
proposed as an illustration in my last posting and it would clearly seem to
contradict your claim that "you can't tell what it is" (you're right to say
that you can't tell "who set it", but neither of course can you with current
globals, so I have difficulties seeing your point there).

'as



From lbates at swamisoft.com  Mon Apr 28 15:00:48 2003
From: lbates at swamisoft.com (Larry Bates)
Date: Mon, 28 Apr 2003 14:00:48 -0500
Subject: python shell as standalone
References: <b8hh2b$9vc$1@news.tu-darmstadt.de>
Message-ID: <BQednc18TI_T5jCjXTWcoA@comcast.com>

Take a look at py2exe at http://py2exe.sourceforge.net/

You can create an "executable" with some supporting
libraries that doesn't require Python installation.

-Larry Bates


"Mirko Koenig" <koenig at v-i-t.de> wrote in message
news:b8hh2b$9vc$1 at news.tu-darmstadt.de...
> Hi
>
> I wrote a modul that gives me some extensions to the python shell.
> E.g. displaying complex number in different ways.
> Now i want to make the shell, with this modul loaded, a stand alone
> programm, so that i can run it unter win/dos without having python
> installed.
>
> Is that possible? If yes how?
> I can't find documentation on creating standalone code with python.
>
> Thanx
>
> Mirko Koenig





From lyle at users.sourceforge.net  Wed Apr  9 16:24:55 2003
From: lyle at users.sourceforge.net (Lyle Johnson)
Date: Wed, 09 Apr 2003 15:24:55 -0500
Subject: string() - Is this some kind of sick joke?
In-Reply-To: <36dbf824.0304091142.527af834@posting.google.com>
References: <36dbf824.0304091142.527af834@posting.google.com>
Message-ID: <3E948197.4090303@users.sourceforge.net>

Daniel Gowans wrote:

> What gives?  Did I get the bogus version?  Is this wishful thinking
> documentation?  I also tried  y.strip("'")  with the same results.
> 
> If this can't be done this way (oh, how I love chomp()) then is there
> a simple, elegant way to do it?

Starting with Python 2.2.2 (but *not* in Python 2.2.1), the strip(), 
lstrip() and rstrip() methods for string objects accept this optional 
argument. I would not have expected the 'string' module to pick up these 
extras, however, since it's deprecated.




From mike at bindkey.com  Tue Apr 29 14:58:33 2003
From: mike at bindkey.com (Mike Rovner)
Date: Tue, 29 Apr 2003 11:58:33 -0700
Subject: stupid q-n (wxPy)
References: <Ukxra.33100$7M5.3263531@twister.nyc.rr.com> <3EAEB850.50D9DEB9@engcorp.com> <7hzra.27540$su3.3180311@twister.nyc.rr.com>
Message-ID: <b8mhud$ppd$1@main.gmane.org>

"Max Khesin" <max at cNOvSisiPonAtecMh.com> wrote in message
news:7hzra.27540$su3.3180311 at twister.nyc.rr.com...
> The only detail I have is that the script was copy-pasted out straight of
> wxPython demo app. There were no specific instructions in the readme for
> running the code so I just gave it a shot. Since wxPython is so popular I
> fugured someone would just tell me "didn't you bla-bla, (idiot)?".
> anyway, thanks.

In wxPython/demo resides run.py which demo scripts import.
But *you* don't want to follow pretty complex demo framework in your
program.

Start with something very simple like that:

from wxPython.wx import *

class MainFrame(wxFrame):
    def __init__(self, parent, id, title):
        # create a frame for our demo
        wxFrame.__init__(self, parent, id, title)

        self.drawPanel = wxScrolledWindow(self, -1, style=wxSUNKEN_BORDER)
        self.drawPanel.SetBackgroundColour(wxWHITE)

        self.bmp=wxImage('test.jpg', wxBITMAP_TYPE_JPEG).ConvertToBitmap()
        EVT_PAINT(self.drawPanel, self.draw)

    def draw(self,evt):
        dc = wxPaintDC(self.drawPanel)
        self.drawPanel.PrepareDC(dc)
        dc.BeginDrawing()
        dc.DrawBitmap(self.bmp, 0, 0)
        dc.SetBrush(wxBLACK_BRUSH)
        dc.SetPen(wxGREY_PEN)
        dc.DrawLine(0,0,200,200)
        dc.EndDrawing()

class App(wxApp):
    def OnInit(self):
        wxInitAllImageHandlers()
        frame = MainFrame(NULL, -1, 'JPG Demo')
        frame.Show(true)
        self.SetTopWindow(frame)
        return true

app=App(0)
app.MainLoop()








From ianb at colorstudy.com  Mon Apr  7 13:56:09 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 07 Apr 2003 12:56:09 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m28yum76rt.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	 <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
	 <b6qs04$7eqi1$1@ID-169208.news.dfncis.de>
	 <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
	 <mailman.1049693590.23386.python-list@python.org>
	 <m28yum76rt.fsf@localhost.localdomain>
Message-ID: <1049738169.658.242.camel@lothlorien>

On Mon, 2003-04-07 at 08:53, Patrick K. O'Brien wrote:
> > Managing so many methods from diverse sources might be difficult.  I
> > want generic methods, I keep thinking of places where they'd be mighty
> > useful... like Mixins but better.
> 
> This I don't understand.  Could you elaborate?

All sorts of extensions will provide new commands that work on the
current buffer (or rather, work on a buffer, usually the current one). 
Adding methods via the mixin approach (actually assigning methods to the
class after the class's creation) will work, but in my experience it can
quickly become ugly... it's unclear where a method comes from, and
everything becomes way too dynamic.

Maybe generic methods aren't the best idea, so I mostly retract them,
though I think they are still interesting.  Events are a better
alternative to mixins, and more appropriate here.

The criticism that it's too hard to have a buffer argument to every
command is certainly false, though.

  Ian






From h_schneider at marketmix.com  Thu Apr 17 12:10:23 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Thu, 17 Apr 2003 18:10:23 +0200
Subject: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net> <b7mbnb$pp2$06$1@news.t-online.com> <8yu9gpnb.fsf@python.net> <b7mf9u$2jb$03$1@news.t-online.com>
Message-ID: <b7mjlc$6tm$03$1@news.t-online.com>

Running the service framework with

if __name__ == '__main__':
    win32serviceutil.HandleCommandLine(Service)

also works without problem. When I freeze the whole thing, it ceases to
function ... :-/

--Harald


"Harald Schneider" <h_schneider at marketmix.com> schrieb im Newsbeitrag
news:b7mf9u$2jb$03$1 at news.t-online.com...
> It's in HKEY_LOCAL_MACHINE.
>
> The registry calls are in another module, which is imported. The registry
> calls work, when I move them to the Service framework. So this seems to be
> an import problem which only happens on 2.2 ...
>
> > What happens when you run the service as not-py2exe'd service (via
> > PythonService.exe)?
>
> It just says
> Could not start the service - error 203
> Registering the Python Service Manager..
>
> Very strange ..
>
> --Harald
>
>
>
> "Thomas Heller" <theller at python.net> schrieb im Newsbeitrag
> news:8yu9gpnb.fsf at python.net...
> > "Harald Schneider" <h_schneider at marketmix.com> writes:
> >
> > > No permission probs, logged in as Admin. Platform is Win2K, not
updated
> from
> > > NT.
> > > There are no exceptions. Reading a registry key just claims it would
not
> > > exist. When I run the same code outside the service wrapper, I can
read
> the
> > > key.
> > >
> >
> > Strange.
> >
> > Do you mean *you* are logged in as Admin, or does the *service* run
under
> > the Admin account?
> >
> > Per default a service runs under the system account. Maybe the registry
> > keys are really not there then? HKEY_CURRENT_USER, probably?
> >
> > What happens when you run the service as not-py2exe'd service (via
> > PythonService.exe)?
> >
> > Thomas
> >
> > PS: As far as i can tell, there are no differences between the 2.2 and
2.1
> > release of py2exe. Not yet ;-)
>
>





From gh at ghaering.de  Mon Apr 28 12:42:12 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Mon, 28 Apr 2003 18:42:12 +0200
Subject: simple extension modules with GNU's C#
In-Reply-To: <80dac0f4.0304280830.756ed8c@posting.google.com>
References: <80dac0f4.0304280830.756ed8c@posting.google.com>
Message-ID: <3EAD59E4.7070201@ghaering.de>

Ionutz Borcoman wrote:
> Hi,
> 
> What's the simplest way to make an extension in C# ? [...]

AFAIK there is no simple way. You'd probably be better off learning C 
(or C++) and using this language for writing Python extensions.

Or do you need to interface a .NET library from Python? There might be a 
simpler way, then, like calling a webservice via SOAP.

-- Gerhard




From fredrik at pythonware.com  Fri Apr 25 03:48:56 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 25 Apr 2003 09:48:56 +0200
Subject: Date validation
References: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au>
Message-ID: <b8ap7f$4go$1@main.gmane.org>

James Gregory wrote:

> I mean I *could* write this myself, but I had to ask the computer how
> many days in April - someone with a better grasp of the date must have
> written this stuff before :)

don't you have any knuckles on your left hand?

if you want your program to check how many days there are in
a given month, use the calendar module:

>>> import calendar
>>> firstday, ndays = calendar.monthrange(2003, 4)
>>> print ndays
30
>>> print calendar.day_abbr[firstday]
'Tue'

</F>







From f.glasner at feldmann-mg.com  Tue Apr 15 15:15:44 2003
From: f.glasner at feldmann-mg.com (Franz Glasner)
Date: Tue, 15 Apr 2003 21:15:44 +0200
Subject: super-problem
In-Reply-To: <J_Xma.4475$Up3.2766033@twister.nyc.rr.com>
References: <J_Xma.4475$Up3.2766033@twister.nyc.rr.com>
Message-ID: <Pine.WNT.4.53.0304152114130.326@fagnt.zentrale.feldmann-mg.com>

super() works only for new style classes.

Derive your class B from object and super() works:

  class B(object):
     ...

Franz.


On Tue, 15 Apr 2003, Max Khesin wrote:

> I was playing with super() and here is my session - the suggested usage of
> super appears to produce an error. What am I missing? thanks -
> max.
>
>
> PythonWin 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32.
> >>> print super.__doc__
> super(type) -> unbound super object
> super(type, obj) -> bound super object; requires isinstance(obj, type)
> super(type, type2) -> bound super object; requires issubclass(type2, type)
> Typical use to call a cooperative superclass method:
> class C(B):
>     def meth(self, arg):
>         super(C, self).meth(arg)
> >>> class B:
> ...  def meth(self, arg):
> ...   print ('B::meth, arg = %s'%arg)
> ...
> >>> class C(B):
> ...  def meth(self, arg):
> ...   super(C, self).meth(arg)
> ...
> >>> c = C()
> >>> c.meth('Ho')
> Traceback (most recent call last):
>   File "<interactive input>", line 1, in ?
>   File "<interactive input>", line 3, in meth
> TypeError: super() argument 1 must be type, not class
>
> --
> ========================================
> Max Khesin, software developer -
> max at cNvOiSsPiAoMntech.com
> [check out our image compression software at www.cvisiontech.com, JBIG2-PDF
> compression @
> www.cvisiontech.com/cvistapdf.html]
>
>
>
>



From alet at librelogiciel.com  Thu Apr 17 18:37:12 2003
From: alet at librelogiciel.com (Jerome Alet)
Date: Fri, 18 Apr 2003 00:37:12 +0200
Subject: [ANNOUNCE] PyKota v1.03 is out !
Message-ID: <pan.2003.04.17.22.37.12.349093@librelogiciel.com>

Hi there,

I'm please to announce PyKota v1.03

PyKota is a python-written, internationalized, complete, centralized and
extensible print quota solution for CUPS (http://www.cups.org)

PyKota is distributed under the terms of the GNU GPL.

You can see the main features, view a screenshot of a print quota
report, give a look at the recent changes, download the CVS version,
or purchase an official package from :

  http://www.librelogiciel.com/software/PyKota/action_Presentation

You can see a sample report at :

  http://servmediup.unice.fr/cgi-bin/printquota.cgi

The list of changes from version 1.02 is impressive :

	- Group quotas work now

	- Possibility to charge per page and per job on a per printer basis

	- Possibility to limit printing by page quota or by account balance
          value on a per user or per group basis

	- Jobs history which will allow per-period reports in a next
          version

	- Support for AppleTalk printers

	- And much more ! Please read NEWS file for details.


Send comments, bug reports, or flames to : alet at librelogiciel.com

Thanks in advance

Jerome Alet




From aleax at aleax.it  Thu Apr 17 11:14:50 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 15:14:50 GMT
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de>
Message-ID: <Kxzna.27947$LB6.642977@news1.tin.it>

Haran Shivanan wrote:

> This seems like a fairly straight forward thing but I'm not sure about how
> to implement it.
> I have a base class from which I'm deriving several new classes:
> class A:
> mem_a=0
> mem_b=1
> class B(A):pass
> 
> I want B to inherit mem_a and mem_b from A without my having to explicit
> do something like:

That is exactly how Python's inheritance works!  Have you tried it?

>>> class A:
...   mema = 23
...   memb = 45
...
>>> class B(A): pass
...
>>> B.mema
23
>>> B.memb
45
>>>


We're talking about CLASS attributes here.  If you're confused between
class attributes and instance attributes (the latter being the kinds
that's typically assigned in __init__), then the right approach is to
ensure that A has a constructor (__init__) and each derived class uses
the base class's constructor (by not overriding it, or by explicitly
calling it if it does override it with its own __init__), which again
is the normal Pythonic way of doing things (doesn't necessarily happen
automatically, but when it doesn't you make it happen explicitly).


Alex




From adalke at mindspring.com  Wed Apr 16 22:31:08 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 16 Apr 2003 20:31:08 -0600
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com> <3E9DE5F0.623D4CD8@alcyone.com>
Message-ID: <b7l3e7$ff2$1@slb9.atl.mindspring.net>

Erik Max Francis:
> I certainly do extreme programming.  I always do my best programming
> while parachuting out of planes, bungee off of towers, and rappeling
> into ceiling light caves.

Lucky you - *I* get my best programming done cave diving, and
you know how hard it was to get my laptop waterproof?  At least you're
in air the whole time!

Whiteboards don't work very well either.  :(

                    Andrew
                    dalke at dalkescientific.com





From BPettersen at NAREX.com  Mon Apr 21 10:42:33 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Mon, 21 Apr 2003 08:42:33 -0600
Subject: python dynamic scoping question
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE01C@admin56.narex.com>

> From: Grzegorz Adam Hankiewicz [mailto:gradha at titanium.sabren.com] 
> 
> On 2003-04-20, Bjorn Pettersen <BPettersen at NAREX.com> wrote:
> > > I have a question regarding scoping in python.  I wish to have
> > > some mechanism which puts variables at a function level scope
> > > into the global scope, dynamically.  I cannot alter my design.
> > 
> > Good luck, given what you're saying, as opposed to only the
> > fragment of code, you would need perfect knowledge of all
> > assignments that have dynamically happened in a program, i.e. you
> > can't find these variables by static/lexical analysis. I'm almost
> > positive there doesn't exists a language where you can perform
> > this operation in this direction unless you consider making _all_
> > variables global (dynamic scoping does something similar, but in
> > the other direction :-)
> 
> Hmmm... maybe I don't understand the terminology well, but what
> you say seems to be done with TCL's upvar/uplevel keywords, where
> you can "overwrite" variables from the global scope, or even a
> specific caller level up the stack. I think it's quite funky, and
> can keep you from redesigning many things under tight deadlines.
> Abusing it of course is bad, as many other things in life.

You're correct! God, it's been a long time since I looked at it. What
you're describing above (and described just as mystically in the TCL
manuals) is simply call by reference (in the C++ sense of the term).
Bear with the TCL syntax for a moment ('cause nobody would think of
implementing it like that unless the syntax looked relatively clear..):

1  proc f {} {
2     set a = 1
3     g a
4     puts $a
5  }
6
7  proc g {v} {
8     uplevel set v 42
9  }

Semantics:

1  def f():  
2     a = 1
3     g('a')  # note, without the $, TCL passes the string]ified
variable name
4     print a

7  def g(v): 
      _tmp = sys._getframe(1) # get calling frame
8     exec 'string2variable(v) = 42' in _tmp.f_globals, _tmp.f_locals

The C equivalent would be:

  void f() {
     int* a = malloc(sizeof(int));
     *a = 1;
     g(a)
     printf("%d\n", *a);
 }

 void g(int* v) {
    *v = 42;
 }

Notice that in a calling context the variable object (the variable
itself, not what it's pointing to) is referenced/dereferenced very
similarly...

So, this isn't particularly interesting. If I say that "uplevel set x
42" will create the variable if it doesn't exists, and "uplevel 2 set x
42" does the same to the caller's caller things become troublesome
(theoretically at least). I can't think of a formal semantics that
wouldn't severely restrict the language in other areas.. (and with no
semantics, no correctness proof of at least compiler optimizations, and
probably programs themselves).

To the OP, it seems simple, use TCL and bind variables using (the #0
getting you to globals):

  uplevel #0 set varname value

Learned something new today!

-- bjorn




From adalke at mindspring.com  Thu Apr 17 16:58:50 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Thu, 17 Apr 2003 14:58:50 -0600
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk>
Message-ID: <b7n4b8$maq$1@slb0.atl.mindspring.net>

Robin Becker:
> interestingly genetic programming et al seem to perform worse when
> attempts are made to 'direct' the evolution.

Pardon?

I thought the point of the cost function was to 'direct' the
evolution.

                    Andrew
                    dalke at dalkescientific.com





From ramen at lackingtalent.com  Tue Apr  1 16:39:16 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 21:39:16 -0000
Subject: How to obtain python file path
References: <20030401163212631-0500@news.covad.net>
Message-ID: <slrnb8jusc.5u1.ramen@lackingtalent.com>

In article <20030401163212631-0500 at news.covad.net>, Dan Grassi wrote:
> I need to obtain the full path of the python.py file that the program is 
> executing from wether imported or run as main.

import os, sys
print os.path.abspath(sys.argv[0])

HTH,
Dave



From martin at v.loewis.de  Fri Apr 25 13:12:25 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 19:12:25 +0200
Subject: PyXML - setContentHandler??
In-Reply-To: <3EA93AD2.30208@ncsa.uiuc.edu>
References: <3EA91EA8.2050505@ncsa.uiuc.edu> <b8b87m$rs4$01$1@news.t-online.com> <3EA93AD2.30208@ncsa.uiuc.edu>
Message-ID: <b8bq9m$hu0$01$1@news.t-online.com>

Randy Heiland wrote:

>  >>> xml.sax.__file__
> '/usr/local/lib/python2.3/site-packages/_xmlplus/sax/__init__.pyc'
>  >>>
>  >>> parser.__class__.__module__
> 'xml.sax.drivers.drv_pyexpat'

Wow. That looks like a *very* old PyXML installation to me. These days,
you should always get stuff from xml.sax.drivers2.

This is a SAX1 class, so it has setDocumentHandler instead of 
setContentHandler.

Regards,
Martin




From altis at semi-retired.com  Fri Apr 25 18:19:02 2003
From: altis at semi-retired.com (Kevin Altis)
Date: Fri, 25 Apr 2003 15:19:02 -0700
Subject: GUI testing with python  (SendKeys)
References: <b8c6c9$p2m$0@198.102.102.163>
Message-ID: <hsiqa.1252$nl1.53881@news.uswest.net>

Did you try looking in the Cookbook? This might be what you're looking for.


http://aspn.activestate.com/ASPN/search?query=sendkeys&section=PYTHONCKBK&ty
pe=Subsection

You don't need to implement Sendkeys, just access it via Mark Hammond's
win32com module.

  http://starship.python.net/crew/mhammond/win32/

For the scripts in the cookbook, if you want to speed things up a bit,
change:

  shell = win32com.client.Dispatch("WScript.Shell")

to

  shell = win32com.client.gencache.EnsureDispatch("WScript.Shell")

ka

"pj" <pjdurai at hotmail.com> wrote in message
news:b8c6c9$p2m$0 at 198.102.102.163...
Greetings

Is there a python module similar to Perl's Win32::GuiTest?
Basically something that would implement SendKeys and friends on Win32 ?

Before I went ahead and rolled my own, I just wanted to check.
Web searches didnt turn up anything useful.

cheers
pj







From brian at sweetapp.com  Fri Apr  4 19:24:00 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Fri, 04 Apr 2003 16:24:00 -0800
Subject: cgi.py -- anybody using it?
In-Reply-To: <Pine.LNX.4.44.0304050051160.1193-100000@alice>
Message-ID: <002501c2fb09$a7dc1110$21795418@dell1700>

> I just tried testing some HTML form upload code using cgi.py, and
> discovered that it wants lower-case HTTP headers :-/
> 
> This is broken (unless I missed something in RFC 2616).  So, does
anybody
> actually use this module?

This does seem broken. File a bug report:
http://sourceforge.net/tracker/?group_id=5470&atid=105470

> I'm in two minds about the 'batteries included' principle, but I
suppose
> there is no obvious best trade-off point between convenience and the
> problems of maintainability, obsolescence, etc...

Probably this feature is rarely used and most user agents use lowercase
headers.

Cheers,
Brian





From nav at adams.patriot.net  Fri Apr 25 13:10:22 2003
From: nav at adams.patriot.net (Nick Vargish)
Date: 25 Apr 2003 13:10:22 -0400
Subject: Off topic: What usenet news reader to use?
References: <mailman.1051036295.31460.python-list@python.org>
Message-ID: <yyy4r4m1p1d.fsf@adams.patriot.net>

You could use XEmacs/Gnus as your newsreader, which should work well
under most operating systems, including Win32. The learning curve is a
little steep, especially if you aren't already an Emacs user, but it's
a wonderful newsreader, with great editing features when you want to
post or reply.

As far as mail clients go, I don't know what to recommend. When last I
looked, none of the Emacs based mail readers did well with IMAP
accounts (though this oculd well have changed). At work we are forced
to use Outlook (*sigh*). I use a webmail client (IMP) to read my
personal mail when at work, and Evolution when at home, though the
stability issues of Evolution are starting to annoy me. 

Nick

-- 
#include<stdio.h> /* SigMask 0.3 (sig.c) 19990429 PUBLIC DOMAIN "Compile Me" */
int main(c,v)char *v;{return !c?putchar(*v-1)&&main(0,v+ /* Tweaks welcomed. */
1):main(0,"Ojdl!Wbshjti!=obwAqbusjpu/ofu?\v\1");}  /* build: cc -o sig sig.c */



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Sat Apr 26 09:38:03 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Sat, 26 Apr 2003 15:38:03 +0200
Subject: RuntimeError at end of test.autotest in Python 2.3b on windows,
 tmpfile also fails
In-Reply-To: <3eaa75a0$0$49106$e4fe514c@news.xs4all.nl>
References: <3eaa75a0$0$49106$e4fe514c@news.xs4all.nl>
Message-ID: <3eaa8bb9$0$49109$e4fe514c@news.xs4all.nl>

Irmen de Jong wrote:

> Also, test_os fails:
> ERROR: test_tmpfile (__main__.TemporaryFileTests)
> -----------------------------------------------------------
> Traceback (most recent call last):
>   File "lib\test\test_os.py", line 49, in test_tmpfile
>     fp = os.tmpfile()
> OSError: [Errno 13] Permission denied

FWIW, there was another error message here:
ERROR: test_tmpnam (__main__.TemporaryFileTests)
---------------------------------------------------------
Traceback (most recent call last):
  File "lib\test\test_os.py", line 61, in test_tmpnam
    self.check_tempfile(os.tmpnam())
  File "lib\test\test_os.py", line 29, in check_tempfile
    open(name, "w")
IOError: [Errno 13] Permission denied: '\\sg4.'


Can't a regular user on windows XP use the tmpnam/tmpfile functions then?

--Irme de Jong




From canrocks at yahoo.ca  Mon Apr 21 19:04:09 2003
From: canrocks at yahoo.ca (Denis Agar)
Date: 21 Apr 2003 16:04:09 -0700
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com>
Message-ID: <25a672ee.0304211504.4f3cd067@posting.google.com>

I guess I understand the situation then. Is there any way to use a GUI
that most people already have such as the ones employed by Visual
Basic or Visual C++?



From news at datatailors.xs4all.nl  Wed Apr 16 10:40:42 2003
From: news at datatailors.xs4all.nl (Peter van Kampen)
Date: Wed, 16 Apr 2003 16:40:42 +0200
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <roy-92A049.08504216042003@reader1.panix.com>
Message-ID: <slrnb9qqr9.nvm.news@datatailors.com>

In article <roy-92A049.08504216042003 at reader1.panix.com>, Roy Smith wrote:

[snip excellent stuff]

Well, I'm amazed at, and grateful for, all the time and effort you
guys put in. I also received an excellent reply by email. So a big
thank you to all. 

PterK

P.S. I agree that this would be a valuable resource on a webpage as
Andrew suggested. It get's pointed out quite often that concatenating
strings in loop is an expensive operation in python and I have always
tried to keep that in mind but it is much nicer if you understand why...

-- 
Peter van Kampen
pterk -- at -- datatailors.com



From peter at engcorp.com  Wed Apr  2 17:43:57 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 17:43:57 -0500
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org>
	        <b6fj57$626$1@news-reader11.wanadoo.fr>
	        <3E8B53BA.DBB0C84C@engcorp.com> <mailman.1049319846.7316.python-list@python.org>
Message-ID: <3E8B67AD.7AD74C0C@engcorp.com>

danny at iamlearning.com wrote:
> 
> I want my users run *.pyo or *.pyc files instead of *.py.

To expand on Jp's response a little: please explain why you think
you want this, so we can better understand the need that's driving
it.  Maybe you are trying to do something you shouldn't be trying
to do, or maybe you're just going about it the wrong way.  What
makes you think doing what you are asking is a good thing?

-Peter



From do-not-spam-ben.hutchings at businesswebsoftware.com  Wed Apr  2 09:03:47 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 2 Apr 2003 14:03:47 GMT
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer wrote:
<snip>
>     3.  At most three Is, Xs and Cs may appear in any given literal.
<snip>

OK cleverclogs, how do we write 39?



From dkuhlman at rexx.com  Wed Apr  2 12:45:37 2003
From: dkuhlman at rexx.com (Dave Kuhlman)
Date: Wed, 02 Apr 2003 09:45:37 -0800
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com> <slrnb8lbo0.ccp.rmunn@localhost.localdomain> <3E8AFFB7.9516A49F@engcorp.com>
Message-ID: <b6f7ib$4rc09$1@ID-139865.news.dfncis.de>

Peter Hansen wrote:

> Robin Munn wrote:
>> 
>> sdieselil <sdieselil at yahoo.com> wrote:
>> > Hi
>> >
>> > How can I implement double-linked lists and other complex tree-like
>> > structures in Python?
>> 
>> Do you *need* doubly-linked lists? Is there something that Python's
>> builtin list type can't do for you? Consider whether it's worth taking
>> the time to code up a data structure, with all the attendant edge cases,
>> when you've already got such nice useful builtin types...
> 
> I can think of at least one example where one would *need* doubly-linked
> lists... a homework assignment.  ;-)

But suppose that I *do* implement double linked lists.  I'll have two 
objects that point at each other, in other words circular references.

In older versions of Python, circular references were not garbage collected.  
Is this still true in Python 2.2 and 2.3?  Do I need to use the gc module 
to *enable* garbage collection?  Or, is not *disabling* it good enough?  Do 
I still need to worry about this?  And, suppose you tell me that circular 
references are collected but I don't believe you.  How can I check?

After reading the documentation on the gc module ("3.2 gc --  Garbage 
Collector interface" in the Python Library Reference), I *think* I 
understand this, but am not sure.  Is there a document that explains this?  
I could not find one at www.python.org.  For those of us who want to 
promote Python, it would be nice to have a document that we could point to 
and say, "See. Problem solved. Nothing to worry about."

  - Dave

-- 
Dave Kuhlman
dkuhlman at rexx.com
http://www.rexx.com/~dkuhlman



From martin at v.loewis.de  Sun Apr 13 06:06:21 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 12:06:21 +0200
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <Wl9ma.6864$LB6.194823@news1.tin.it>
Message-ID: <m3adeuww2q.fsf@mira.informatik.hu-berlin.de>

Alex Martelli <aleax at aleax.it> writes:

> The crucial difference is that Python dicts are hashtables, while the
> std::map in the standard library of C++ can't be, because it's based
> on an order relationship, NOT on hashing.  So, std::map has an order,
> but its O(...) performance just can't be as good as a dict's.

Nit-pickers (like myself) will point out that the complexity of a
hash-table depends on the distribution of the hash function (in
addition to depending on the complexity of the hash function).  With a
carefully-chosen hash O(1) function, all dictionary operations will be
O(n). For std::map, O(log n) is the worst-case performance.

Regards,
Martin



From nomad*** at ***absamail.co.za  Sun Apr 13 17:08:32 2003
From: nomad*** at ***absamail.co.za (Nomad)
Date: Sun, 13 Apr 2003 23:08:32 +0200
Subject: PYTHONPATH question
References: <E1927CG-0007k5-00@fast.aub.edu.lb> <mailman.1049632149.9888.python-list@python.org> <5ai29vogfthpqvbmlic09pbllltun61d2l@4ax.com> <mailman.1049729471.8104.python-list@python.org>
Message-ID: <mb3j9vkgifh5ghc38mja4mhanc9urmuu76@4ax.com>

On Mon, 7 Apr 2003 10:30:19 -0500, Skip Montanaro <skip at pobox.com>
wrote:

>
>    nomad> For future safety, it might be an idea to make it:
>
>    nomad>     PYTHONPATH=$PYTHONPATH:<colon-separated list of directories>
>    nomad>     export PYTHONPATH
>
>What's the interpretation of an empty element in $PYTHONPATH?
>

My thinking was rather that, if there was a system-wide $PYTHONPATH
set in /etc/profile, it would be overwriten by any $PYTHONPATH
statement in ~/.profile that didn't explicitly include it, and that
might bugger up other 3rd party modules.

Laters

-- 
Nomad

Wondering of the vast emptyness of the 'net
in search of something cool.



From a.schmolck at gmx.net  Sun Apr  6 23:16:05 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 07 Apr 2003 04:16:05 +0100
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
Message-ID: <yfsd6jzkney.fsf@black132.ex.ac.uk>

Carl Banks <imbosol-1049649076 at aerojockey.com> writes:
> Next three functions are "save-current-buffer", "with-current-buffer",
> and "with-temp-buffer".  These are actually not functions, but special
> forms and macros.  Python has no analogue to these (barring some
> obscure metaclass tricks). 

Care to share the obscure metaclass trick to implement with-* style macros in
python?

> There is one potentially disasterous pitfall you should be aware of.
> Emacs does all kinds of strange things with namespaces.  It uses
> dynamic binding, unlike Python, which uses lexical binding.  Emacs has
> one monolithic global namespace, as opposed to Python, which has an
> organized infrastructure of "global" namespaces, that is, modules.
> Also, Emacs has weird namespaces like buffer local variables.  One
> thing you'll have to do is think about how to sort all these namespace
> issues out.

Yes, these are minor issues (BTW you forgot to mention that emacs also has
separate namespaces for functions/macros and variables (and, in a slightly
different way, lots of other things)).

But luckily enough reimplementing emacs in some "better" programming language
is a common enough computer scientist phantasy that a formal analysis of the
particular hurdles dynamic scope poses has already been done for you (buffer
local variables are also briefly mentioned):

http://citeseer.nj.nec.com/neubauer01down.html

 
> It seems to me that a lot of these issues won't matter much if you
> have a good and consistent strategy for organizing stuff and passing
> parameters around.

:)


alex



From alanmk at hotmail.com  Wed Apr 30 10:06:17 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Wed, 30 Apr 2003 15:06:17 +0100
Subject: How to call built in function 'open' from inside object that has 
 'open' method?
References: <db81c492.0304300526.7b8d3c86@posting.google.com>
Message-ID: <3EAFD859.75C9E3AE@hotmail.com>

vadim wrote:
> 
> I am new to Python and do not know the answer to very
> simple how to 'open' from inside another object that
> has 'open' method. I know that in C++ I should use
> following syntax ::open.
> 
> More details on reasons. I have met this problem trying to
> write a simple script use ActiveScripting  inside IE.
> I am trying "file = open(...) but because context window
> object has 'open' method that just opens window it is called
> first. Sorry for simple question. I was trying to find
> the answer but either it is not on the surface or too simple.

There should be no clash between the names if your object has
an "open" method.

When an object has an "open" method, it is either referred to
by

self.open()

or 

objname.open()

So you should be able to call (the builtin) open from within 
your object, and there should be no name clash, because you have not
explicitly specified the open method on your object (i.e. "self").

Here is an example:

class  Sketch:

  def  __init__(self, filename):
    # Note this calls __builtin__.open
    self.outchan = open(filename, "wt") 

  def  open(self):
    self.outchan.write("Opening window\n")

  def  doit(self, o):
    self.outchan.write("Goodbye, %s!\n" % o)

  def  close(self):
    self.outchan.write("Closing window\n")
    self.outchan.close()

  def  defenestrate(self, o):
    #  Note this calls self.open
    self.open()              
    self.doit(o)
    self.close()

>>> s = Sketch('sketch.log')
>>> s.defenestrate('tabby')

HTH,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From nraavi_news at yahoo.com  Tue Apr 29 19:57:28 2003
From: nraavi_news at yahoo.com (nraavi_news at yahoo.com)
Date: 29 Apr 2003 16:57:28 -0700
Subject: Building Python 2.2.2 - configure error
Message-ID: <65a7284e.0304291557.6b3e437f@posting.google.com>

Hi,

I am trying to configure Python2.2.2 on Alpha (running Tru64 UNIX
V4.0F, C compiler V6.4) with Tcl/Tk support and keep getting an error
about unresolved symbols in the Tk library when running configure.

A sample of the errors are:
22124:./conftest: /sbin/loader: Error: Unresolved symbol in
/opt/tcl/lib/libtk8.4.so: XCreatePixmapCursor
22124:./conftest: /sbin/loader: Error: Unresolved symbol in
/opt/tcl/lib/libtk8.4.so: XUngrabKeyboard
22124:./conftest: /sbin/loader: Error: Unresolved symbol in
/opt/tcl/lib/libtk8.4.so: XUngrabPointer
22124:./conftest: /sbin/loader: Error: Unresolved symbol in
/opt/tcl/lib/libtk8.4.so: XDrawString
...
...
...
22124:./conftest: /sbin/loader: Error: Unresolved symbol in
/opt/tcl/lib/libtk8.4.so: XFillPolygon
22124:./conftest: /sbin/loader: Error: Unresolved symbol in
/opt/tcl/lib/libtk8.4.so: XDrawString16
22124:./conftest: /sbin/loader: Error: Unresolved symbol in
/opt/tcl/lib/libtk8.4.so: XQueryPointer
resolve_symbols: loader error: dlopen: Unresolved symbols

ALL these symbols exist in the library. To help isolate the problem, I
create confdefs.h with the following line:
#define _POSIX_THREADS 1

conftest.c is:
-----------------------------
#line 4191 "configure"
#include "confdefs.h"
#include <pthread.h>
      void *foo(void *parm) {
        return NULL;
      }
      main() {
        pthread_attr_t attr;
        pthread_t id;
        if (pthread_attr_init(&attr)) exit(-1);
        if (pthread_attr_setscope(&attr, PTHREAD_SCOPE_SYSTEM))
exit(-1);
        if (pthread_create(&id, &attr, foo, NULL)) exit(-1);
        exit(0);
      }
----------------------------

Compiled the program using:
cc -o conftest conftest.c -I/opt/tcl/include -L/opt/tcl/lib -lm
-ltcl8.4 -ltk8.4 -lpthread

I get these unresolved symbols when I run conftest. My environment
variables are:

PATH=/opt/tcl/bin:.......
LD_LIBRARY_PATH=/opt/tcl/lib:/usr/shlib:

I've installed both Tcl 8.4.2 and Tk 8.4.2 in the same directory
/opt/tcltk. /opt/tcl and /opt/tk are links to /opt/tcltk. Tcl and Tk
both pass all the tests.

I would appreciate any help in debugging this matter. Please post your
response.

Thank you,

-Narendra



From cbuffer at removethis.rosecott.ukfsn.org  Tue Apr 15 18:39:29 2003
From: cbuffer at removethis.rosecott.ukfsn.org (Ken Parkes)
Date: Tue, 15 Apr 2003 23:39:29 +0100
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>	<slrnb9ooms.sho.ramen@lackingtalent.com>
Message-ID: <pan.2003.04.15.22.22.18.262135@removethis.rosecott.ukfsn.org>

On Tue, 15 Apr 2003 19:51:37 +0000, Dave Benjamin wrote:


> Also, don't forget, sometimes the best things in life are free:
> http://www.ibiblio.org/obp/thinkCSpy/


Have you tried printing this?  When I try to print it on an HP Deskjet
840C (CUPS on Mandrake 9) I lose the last letter on every line and the
last two lines on every page completely.  Have fiddled endlessly to no
avail....and it looks such a good intro.
Ken.



From tdelaney at avaya.com  Tue Apr  1 16:59:28 2003
From: tdelaney at avaya.com (Delaney, Timothy C (Timothy))
Date: Wed, 2 Apr 2003 07:59:28 +1000
Subject: I need help
Message-ID: <338366A6D2E2CA4C9DAEAE652E12A1DE3DD0A9@au3010avexu1.global.avaya.com>

> From: Jewlz408 [mailto:Jewlz408 at aol.com]
> 
> I need a lot of help
> What do you know about Arrays
> How many storage Locations will be allocated for each of the 
> given arrays?
> 1) Dim A(1 to 10)
> 2) Dim B(-3 to 5)
> 3) Dim C$( 12)
> 4) Dim D( 1 to 4, -2 to 2)
> 5) Dim E(4,6)

http://www.catb.org/~esr/faqs/smart-questions.html

In particular, please note the following section:

http://www.catb.org/~esr/faqs/smart-questions.html#forum

If this is an April Fools Joke, you have just done a very poor one. Only people who are well-known on the list have the right to post AFJs - the old-timers can laugh with them and at the newbies who believe it ;)

Tim Delaney




From aleax at aleax.it  Mon Apr 14 10:15:12 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 14:15:12 GMT
Subject: Complexity of standard Python data structures (Re: Speed of string += string)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi> <Oqwma.11524$LB6.319489@news1.tin.it> <slrnb9lbbq.v4f.marcus@infa.abo.fi>
Message-ID: <Qnzma.12335$LB6.335182@news1.tin.it>

Marcus Alanen wrote:

> On Mon, 14 Apr 2003 10:53:34 GMT, Alex Martelli <aleax at aleax.it> wrote:
>>Marcus Alanen wrote:
>>> Indeed, that was the first thing I wondered about when I started
>>> learning Python.
>>Interesting!  What other languages (besides C++ for a subset of its
>>standard libraries) give O() specifications for their operations?
>>In other words, what other languages did you learn, before Python,
>>that didn't leave you wondering about such performance issues?

Maybe I should clarify that I'm not _attacking_ your statement (nor
you for making it, Marcus) -- I'm trying to _understand_ exactly
what's involved, to help me teach Python better, and/or write better
introductory material about it.  So, I'm thankful for the opportunity
this thread may pose to refine my understanding of such learning
and teaching issues.


> It must've been a combination of issues:
> 
> 1) Coming from a C/C++ background, I want to use data structures
> efficiently.  This belongs to the "premature optimization" group. So
> sue me.

While performing optimizations may be premature, _understanding_
what's going on can hardly be (if the person trying to understand
has sufficient cultural background -- and somebody who programs in
C++ is quite likely to have it, anyway).

However, I still don't understand why this would be a problem
specific to Python rather than any other language -- including
C++ for those operations that are not part of that subset of C++'s
standard library for which O()-behavior is specified in the standard.

What's the O()-behavior of C+'s dynamic_cast<>, for example, and
on what parameters does that performance behavior depend?  I do
not know, and I don't think I *CAN* know -- every implementation is
free to do its best (or its worst...) to optimize dynamic_cast's
performance, time-vs-space tradeoffs, whatever.

So why did you perceive Python as being different in this respect?
C doesn't specify the O()-behavior of, e.g., qsort (indeed it's
often pretty horrible...!) -- C++ doesn't specify performance for
such things as RTTI, exception handling, or even for that matter
for the dispatch of virtual methods (you COULD have a conforming
C++ implementation that dispatches in time depending on the depth
of the inheritance graph or the number of virtual methods, though
in practice I know of no such implementation) -- so why did Python
leave you wondering, where C and C++ didn't, when the (lack of)
standard constraints on operations' performance seems so similar?


> 2) The project I was doing could with time lead to handling quite big
> sets of data, so algorithmic efficiency was a concern.

This helps explain why you care to understand such issues (though
I believe that a desire for such understanding is quite normal even
if one is involved in rather different things at a givem time!),
but still doesn't help me understand why Python was different for
you, in this respect, from other languages.


> 3) Python's "list" had on insert() routine, but also a __getitem__
> routine. Both probably couldn't be O(1), so I got a bit confused.
> I didn't want to choose a "slow" method on purpose.

Yes, I understand this point, I think.

> 4) And hey, I did want to learn the language properly.

And definitely this one.

> Number 3 was quite irritating, since C++ "list" is a double-linked
> list, while Python "list" is what is called a "vector" in C++.

Yes, the terminology clash is nasty.  Indeed I remember being peeved
back when C++ started using "list" to mean a *doubly-linked* list
(rather than a singly listed one)...


> Probably not very good excuses (except number 4), but I'm just telling
> what I felt at the time. On the other hand, I don't understand why the
> Python documentation simply can't say "speed O(N)", "speed amortized
> O(1)" for the various methods.

I suspect the motivations are akin to those the C++ Standard (and just
about every other language standardization document) has for not giving
constraints for the performance of the various operations it's
describing -- keeping future changes open.  The C++ Standard is quite
exceptional in giving O()-constraints for SOME small subset of the
operations it describes (I think it's to Stepanov's credit that there
is even that exception) -- so you end up with no guarantees about the
O() behavior of the most fundamental operations BUT with such guarantees
about conceptually higher-level stuff in the library (something that
always amused me quite a bit).

But -- DO Python beginners typically start learning the language with
a *REFERENCE* manual?  Probably not -- yet you explain that the time
when you were left wondering was just when you STARTED learning.  So,
perhaps, the document to retouch with a few small mentions of typical
performance behavior (perhaps in a footnote or two) could be the
tutorial, rather than either of the reference manuals (language and
library).  Clearly, the tutorial is being descriptive (of one specific
implementation it comes with) and not prescriptive (it does not give
any constraints to what future implementations might choose), which
may make it more acceptable to give such indications there.

So what do you think -- is it the tutorial that such changes should
be targeting...?


In "Python in a Nutshell", I took a more traditional approach --
*assuming* that the target reader would want to understand the
*functionality* of operations first, the *performance
characteristics* of opeations only quite a bit later.  So, I
have descriptions of fundamental functionality quite early, in
chapter 4, with almost no mention of performance issues (except
a very few ones, such as the fact that pow(a,b,c) returns the
same result as (a**b)%c "but faster" -- or the fact that
passing a comparison function to the sort method of a list
"slows down the sort").  Only much later (in Chapter 17,
"Testing, Debugging, and Optimizing") do I cover the big-O
behavior of operations on lists, strings and dicts -- including
the key facts, from which almost all others follow, that lists
are internally implemented with vectors, dicts with hash tables.

Do you think that this traditional approach (teach functionality
first, performance issues only later) is fundamentally flawed?

That's basically what I see the "first thing I wondered about"
comment as potentially implying, and what I'm trying to get a
deeper understanding of, particularly when you seem to single
out Python wrt other languages in this respect.  Of course, in
the Nutshell I was constrained by other considerations too (it's
primarily a *reference* work, only secondarily intended to be
read from the start onwards -- having decided to devote a part
of a chapter to optimization, it seemed to me that optimization
related information was best grouped there, rather than spread
around -- and _repeating_ information would go against the grain
of what's meant above all to be a *concise* reference...).  But,
that book may well not be the last time I have occasion to write
about Python (or anyway teach it...) -- so, I think I need to
understand these issues more deeply.


>>> Others answered this very thoroughly, I'd just like to emphasize the
>>> linearity of keys() --- most often you can manage with .iterkeys() or
>>> .itervalues() just fine, and it's a lot faster.
> 
>>As for the performance issues, I wouldn't be so sure.  E.g.:
> 
> Sorry, I really didn't explain that very well. In several cases I have
> had, I don't need to traverse the whole array of dict(), so iterkeys()
> does feel like saving some memory, and possibly speed. Naturally, both
> functions in general visit all elements, and are linear in speed.

Interesting -- since both keys and iterkeys (or direct iteration on
the dict) walk the keys in "random" order, I'm not too sure what a
typical use case of interrupting the loop prematurely would be --
"find any key k that satisfies predicate(k)" for example?  That does
happen some of the time, of course, but I don't think of such cases
as typical.  In any case, it's entirely and always true that keys()
takes O(N) _memory_ while iterkeys() most definitely doesn't, so, _if
memory constraints are involved_, iterkeys, or direct iteration on
the dict object, are going to be preferable.

> By the way, I tested a similar program with 10 million elements a >15%
> speed gain with iterkeys():

That may be a case where "memory constraints are involved" -- allocating
and filling and walking a list with 40 MB of data is going to bust (at
least) any cache your CPU might have, while, if you're lucky, iterkeys
might have better locality and _some_ of the data involved might remain
in the cache.  (Incidentally, cache behavior and other such issues of
memory hierarchies are by far the most complicated performance aspects
to understand, analyze or predict -- and yet far too often they can at
the same time be _dominating_ the performance behavior of your program...
and no language spec is going to be very helpful there;-).


Alex




From ekranawetter at bmw.co.at  Sat Apr 19 06:31:15 2003
From: ekranawetter at bmw.co.at (ekranawetter_steyr)
Date: 19 Apr 2003 03:31:15 -0700
Subject: wxPython 1.52 under Win95
Message-ID: <7d22d8d6.0304190231.4f6da5b1@posting.google.com>

Hi all,

I tried to run wxPython 1.52 under Win95 (because the application I
use is hard to port to version 2) and Python crashes on the command:

from wxPython.wx import	* 

This is my Python installation:

- py152.exe
- win32all-132.exe
- a lot of files in \Python\wxPython (wx22_2.dll from 21st Oct 00)
- win32com
- win32

How can I fix this problem? Please send an email also to
ekranawetter at utanet.at

Thanks, Erwin



From sophy.cheeran at nettasking.com  Sun Apr  6 23:45:35 2003
From: sophy.cheeran at nettasking.com (Sophy)
Date: 6 Apr 2003 20:45:35 -0700
Subject: Python ODBC interface crash on connection failure to MS SQL
Message-ID: <48e1ebc2.0304061945.3039c772@posting.google.com>

Hi,

I would appreciate ur advice on this problem that i am facing. 
I am running a python application and using the MSSQL database through
a python ODBC interface. I find that the python ODBC interface crashes
whenever there is a connection failure to the MSSQL. I get an "unable
to read memory" message coming up and a print message "disconnected"
which i later found in the ODBC.cpp source file in the method
odbcPrintError() which handles the dbi errors.
I use the same with MySQL and dont seem to have any problems with
connection failure handling.

Can somebody advice me as to how i can resolve this issue.

Thanks
Sophy



From forodejazz at yahoo.es  Mon Apr  7 09:49:41 2003
From: forodejazz at yahoo.es (=?ISO-8859-1?Q?David_Asorey_=C1lvarez?=)
Date: 7 Apr 2003 06:49:41 -0700
Subject: Python & Music: Performance question
References: <1759581e.0304060218.164d4cd4@posting.google.com>
Message-ID: <1759581e.0304070549.3b3a4caa@posting.google.com>

forodejazz at yahoo.es (David Asorey ?lvarez) wrote in message news:
<1759581e.0304060218.164d4cd4 at posting.google.com>...
> Hello, group.
> 
> I'm developing a small music program, and I have some doubts.
> 
>  [...]

  Thanks to all.
  I've tried some test with the first version and there is not any
problem, even with thousands of notes:

class Note:
   def __init__(self, pitch, duration):
      self.pitch = pitch
      self.duration = duration
   def change_pitch(factor): pass
   def change_duration(factor): pass


if __name__ ==  '__main__':
    import time
    notes = []
    t1 = time.clock()
    for x in range(10000):
        notes.append(Note(x, x))
    t2 = time.clock()
    print t2 - t1

Pentium 800, 128 Ram, W2K: 0.16 s
Pentium 133, 40 Ram, Linux (woody): 1.67 s

For my purposes, this are acceptable values.

Again, thanks for your advices.

David



From claird at lairds.com  Tue Apr  8 08:48:01 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 08 Apr 2003 12:48:01 -0000
Subject: The place of open-source (was: How to protect python code ?)
References: <mailman.1049758213.26256.python-list@python.org> <slrnb945tu.p4.bignose-hates-spam@rose.localdomain.fake> <5ppka.23067$It5.15751@news2.central.cox.net>
Message-ID: <v95h81ic01csee@corp.supernews.com>

In article <5ppka.23067$It5.15751 at news2.central.cox.net>,
Steve Holden <sholden at holdenweb.com> wrote:
			.
			.
			.
>Of course, the businesses that make a good living out of proprietary
>software encourage all sorts of FUD about open source. Consequently, large
>organizations implement rules forbidding the use of open source software
>products because "they aren't supported" when in fact the support for some
>open source products is actually superior to that of most proprietary
>products.
>
>The fact of the matter is that most companies would prefer to pay for a
>black-box proprietary product because ultimately there's assumed to be a
>responsible (and accountable) organization behind it. If they bothered to
>read the licenses, of course, they'd realize that the producer's liability
>is usually limited to the purchase price of the software.
			.
			.
			.
There certainly are policies in place that prevent
or make difficult use of some or all open-source
products.  The proximate reasons for those policies
are, as Steve knows, more varied than simply, "they
aren't supported", and they originate in more than
just FUD.

I make a point of this because I think it's important
to realize that rational use of open source will 
involve more than simply FUD battles.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From anthony.evershed at npl.co.uk  Mon Apr  7 10:13:55 2003
From: anthony.evershed at npl.co.uk (CTony)
Date: 7 Apr 2003 07:13:55 -0700
Subject: COM and MS Word
References: <2eb85cfb.0304070115.6aecb644@posting.google.com> <pan.2003.04.07.11.53.57.522248.788@thomas-guettler.de>
Message-ID: <2eb85cfb.0304070613.211f24ac@posting.google.com>

Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote in message news:<pan.2003.04.07.11.53.57.522248.788 at thomas-guettler.de>...
> On Mon, 07 Apr 2003 11:15:31 +0200, CTony wrote:
> 
> > It appears that some data files I need are in WordPerfect (.dat) format,
> > and I need to get them into Python for processing and later entry into a
> > database. I haven't been able to find a way of interfacing to
> > WordPerfect directly, but MS Word 2000 can read these files...
> 
> Maybe you can solve your problem by avoiding COM. You can create HTML
> from Word-Documents with wvWare. I think you can handle this easier than
> COM. Our save the word document as XML and parse this.
> 
> Sorry, this was no answer to your question, but maybe it helps, too.
> 
>  thomas

Thanks, however (you could hear it coming, couldn't you?) I'm trying
to avoid making explicit changes to the document in Word, and unless
wvWare is run from Python...?

I've just found out - there are literally a few hundred of these files
(ulp!), which makes any real manual work a bit of a pain! *g*

Thanks anyway,

Tony



From gh at ghaering.de  Mon Apr 14 09:58:32 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: 14 Apr 2003 13:58:32 GMT
Subject: Calling Python Code Snippets within Python Application
References: <236da4ed.0304140143.25abc577@posting.google.com>
Message-ID: <slrnb9lf9l.12k.gh@haering.opus-gmbh.net>

Marc Floessel <gildemere at hotmail.com> wrote:
> This is very likely a newbie question, but I'd still appreciate a
> pointer. I need to call Python Code snippets (saved in a file/database
> etc) within my main Python Application. I would like to do the
> following things: [...]
> In which direction should I look?

The "execfile" builtin function.

while you're at it, you might want to look at "exec" and "eval", too.
Unlike the functions "eval" and "execfile", "exec" is a statement, btw.

-- Gerhard



From aleax at aleax.it  Tue Apr 29 18:00:59 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 22:00:59 GMT
Subject: The behavior of List.insert
References: <b8m8dn$b03$1@pegasus.csx.cam.ac.uk> <TZOdnU3XYIdTMzOjXTWcpQ@comcast.com> <b8mc4b$e80$1@pegasus.csx.cam.ac.uk>
Message-ID: <vCCra.33976$K35.1085821@news2.tin.it>

York wrote:

>>> mylist = [1,2,3,4,5],
>>> mylist.insert(-1,99) will make mylist to be [99,1,2,3,4,5]
>>>
>>> just wonder why not let it to be [1,2,3,4,99,5]
>> 
>> Guido decided otherwise for whatever reason.  What you observed is
>> intended: from LibRef 2.2.6.4 Mutable Sequence Types
>> "(4)
>> When a negative index is passed as the first parameter to the insert()
>> method, the new element is prepended to the sequence. "
>> so this is not an implementation bug.
> 
> Thanks. I don't know what Guido's reason is, but I still believe that it
> is nice to allow list.insert distinguish different negative indice rather
> than treat all them as zero.

Agreed -- and indeed, in 2.3 beta 1...:

[alex at lancelot src]$ python
Python 2.3b1 (#1, Apr 26 2003, 10:44:25)
[GCC 3.2 (Mandrake Linux 9.0 3.2-1mdk)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> x=list('ciao')
>>> x.insert(-1, 'w')
>>> x
['c', 'i', 'a', 'w', 'o']
>>>

...the issue is finally fixed.


Alex




From unendliche at hanmail.net  Fri Apr 11 00:35:56 2003
From: unendliche at hanmail.net (Seo Sanghyeon)
Date: 10 Apr 2003 21:35:56 -0700
Subject: Can I define "static" (like in C++) methods in Python?
References: <169a7283.0304101430.2bec922b@posting.google.com>
Message-ID: <45e6545c.0304102035.f06e4fd@posting.google.com>

> Can I define methods which are attributes of class and not attributes of
> instance?

Yes. It is called staticmethod().

class Math:
  def add(x, y):
    return x + y
  add = staticmethod(add)

Math.add(2, 3)



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Fri Apr 25 13:09:35 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Fri, 25 Apr 2003 19:09:35 +0200
Subject: Python Package Index @ python.org down?
Message-ID: <3ea96bce$0$49106$e4fe514c@news.xs4all.nl>

I get a Python stacktrace when trying to access the Python Package Index,
http://www.python.org/pypi

"OperationalError: database is locked"

Tried three times. Anything wrong over there?

--Irmen




From aleax at aleax.it  Thu Apr 24 09:27:42 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 13:27:42 GMT
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com> <mailman.1051184857.22620.python-list@python.org>
Message-ID: <iDRpa.132$3M4.3351@news1.tin.it>

Andrew Bennetts wrote:

> On Thu, Apr 24, 2003 at 01:29:44PM +0200, Nikolai Kirsebom wrote:
>> Given the following class hierarchy:
>>       
>> class A(object):
>>       InstCount = 0
>>       def __init__(self):
>>               self.__class__.__dict__['InstCount'] += 1
> 
> A simpler way to do this line would be:
> 
>               self.__class__.InstCount += 1

However, this hides a bug with subclasses that fail to redefine
InstCount in their bodies and turns it into weird behavior instead.

I suggest a custom metaclass -- see my earlier post in this thread.


Alex




From escalation746 at yahoo.com  Thu Apr  3 14:49:05 2003
From: escalation746 at yahoo.com (robin)
Date: 3 Apr 2003 11:49:05 -0800
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <b6f66v$28k$1@slb3.atl.mindspring.net> <v8m9a4nk4dpjf7@corp.supernews.com>
Message-ID: <a626bbd7.0304031149.7abe84c0@posting.google.com>

claird at lairds.com (Cameron Laird) wrote:

> SOAP and such are just concessions to commercial misunder-
> standings about what business needs.

I would like to know more about what you mean by this.

Jp Calderone <exarkun at intarweb.us> wrote:

> Personally, I think SOAP is worth ignoring.  ;)  XML-RPC is what I would
> choose if I were going for that sort of solution, too.

I'd like more details here too. Why ignore SOAP?

-- robin



From carroll at tjc.com  Fri Apr 11 12:39:00 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Fri, 11 Apr 2003 09:39:00 -0700
Subject: string.split() documentation incorrect?
References: <mailman.1050076398.7023.python-list@python.org>
Message-ID: <qard9v4ndv0g89bffoseign9g3isvjt6og@4ax.com>

On Fri, 11 Apr 2003 09:52:32 -0600, Jerry Seutter
<jerry.seutter at pason.com> wrote:

>I'm trying to use string.strip to remove the comma off the end of 
>a string, eg:
>
>string.strip("id INT NOT NULL,", ",")
>
>The current documentation says I can do this.  Python 2.2.2 gives
>the following error:
>
>>>> strip('  hello  ', ',')
>Traceback (most recent call last):
>  File "<stdin>", line 1, in ?
>TypeError: strip() takes exactly 1 argument (2 given)
>
>So, I think either the documentation is incorrect or there is 
>a bug.  Thoughts?

Yeah, it's a bug, bug number 697220.  See
http://sourceforge.net/tracker/?group_id=5470&atid=105470&func=detail&aid=697220
there's a patch available.

However, use of the string module is deprecated in favor of methods of
the string class, which work as documented.  The string module only
calls the equivalent class method anyway, e.g., string.strip(s) turns
into s.strip().

Save yourself some trouble and use the class method directly, e.g.,
instead of these:

 >>> string.strip(s,",")
 >>> string.strip("id INT NOT NULL,",",")

use these:

 >>> s.strip(",")
 >>> "id INT NOT NULL,".strip(",")




From andrew.markebo at telia.com  Fri Apr 25 14:45:30 2003
From: andrew.markebo at telia.com (Andrew Markebo)
Date: Fri, 25 Apr 2003 18:45:30 GMT
Subject: Curses on Win32
References: <b86r8p$27n$1@gateway.northgrum.com>
Message-ID: <m3u1cms9f8.fsf@telia.com>

/ Dang Griffith <dmgriffith at tasc.com> wrote:
|    I'm running 2.3a2 and decided to spearmint with curses, only to find
| it's not available for import on Win32.  Is/was curses available in 2.2?
|  Is there an alternative text user interface library available?
|    Thanks,
|       --dang

Considered checking the cygwin-port of python to the win32 platform? 

           /Andy

-- 
 The parser of the compiler rests on the code!



From pobrien at orbtech.com  Mon Apr  7 10:12:42 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 07 Apr 2003 09:12:42 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3e91777f$0$49105$e4fe514c@news.xs4all.nl>
Message-ID: <m2k7e65rc5.fsf@localhost.localdomain>

Boudewijn Rempt <boudewijn at tryllian.com> writes:

> Patrick K. O'Brien wrote:
> 
> > If you were crazy enough to think that you could create a program
> > along the lines of Emacs, but written in Python, how would you go
> > about it?  How would you design the domain model for files, buffers,
> > windows, and frames?  How would you allow the same level of
> > customizability?  How would you map functions (or methods or whatever)
> > to keys?  Any thoughts?
> > 
> 
> About a year ago I began something like that (mostly to teach myself how
> an editor really worked). I extended Kalam -- the example application
> in my PyQt book which used the Qt edit control -- with a backend that
> represented buffers and a redisplay written in Python that painted the
> glyphs and cursor on a widget with QPainter. Was quite fun until I lost
> the source just as I had the redisplay working.
> 
> The best resource for someone implementing an editor from scratch is:
> 
> The Craft of Text Editing, or Emacs for the Modern World by Craig A. Finseth
> 
> http://www.finseth.com/~fin/craft/

Thanks for the link.  Someone else recommended it as well the other
day.  I read it, but most of it is low-level stuff already handled by
Scintilla.  What I need is the framework for modeling files,
documents, buffers, windows, commands, etc.  I guess I'll just invent
my own (which is what I've already started to do, but it sure feels
like reinventing the wheel).

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From claird at lairds.com  Fri Apr 25 12:33:32 2003
From: claird at lairds.com (Cameron Laird)
Date: Fri, 25 Apr 2003 16:33:32 -0000
Subject: pythong referencing/aliasing
References: <ylcqa.9116$h%2.963225@read1.cgocable.net> <mailman.1051284819.25791.python-list@python.org>
Message-ID: <vaioqs7e92v14d@corp.supernews.com>

In article <mailman.1051284819.25791.python-list at python.org>,
Jeff Epler  <jepler at unpythonic.net> wrote:
			.
		[other true stuff]
			.
			.
>give you a clearer explanation.  I can only add one other piece of
>information: I rarely am hurt by the fact that Python doesn't have
>references of the type you're talking about.
			.
			.
			.
I'll make a positive assertion of this:  in C and
C++, stylish programmers frequently use references
to good effect.  Python rarely, very rarely, needs
to do the same.  It might be instructive to choose
a typical C++ reference idiom, and illustrate how
its typical use case is better served in Python be
a derived return, or dictionary construction, or ...
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From sholden at holdenweb.com  Mon Apr 28 20:42:21 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 29 Apr 2003 00:42:21 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <20030417205405.A10672@tibia.amotlpaa.bogus> <mailman.1051310385.1948.python-list@python.org> <b8cgst$20ho$1@nntp6.u.washington.edu> <mailman.1051490038.11587.python-list@python.org>
Message-ID: <NTjra.23897$WQ4.21608@news2.central.cox.net>

"Steven Taschuk" <staschuk at telusplanet.net> wrote in message
news:mailman.1051490038.11587.python-list at python.org...
> Quoth Donn Cave:
> > Quoth Steven Taschuk <staschuk at telusplanet.net>:
> > ...
> > | I've submitted a documentation patch for the __str__ and __repr__
> > | docs, following ideas discussed in this thread:
> > |     <http://www.python.org/sf/727789>
>   [...]
> > I still think that the normal usage of __str__ can be directly
> > described as a data conversion to type string.  I know you don't
> > find that completely illuminating on its own, but I think it does
> > convey an essential idea that is otherwise likely to be missed.
> > See for example the __str__ and __repr__ definitions in Cookie.py.
>
> Hm... "conversion to type string".  So under this concept str()
> converts an object to a string, while repr() represents the object
> in a string, right?  Is this what you're saying when you
> distinguish "the data as a string" from "the object as a string"?
> If so, I quite like it.
>
</lurk>
Finally, the murk lifts and a dim light becomes perceptible. "str() is what
the users will recognize. repr() is what the programmers will thank you
for."

The print __str__()/__repr__() dichotomy is probably the best compromise for
the average Python user.

> (For some reason convert/represent sits better in my brain than
> data/object.  Don't ask me; I just think here.)
>
How about "use/program". It isn't valid to complain that programming an
object is using it, since that objection qualifies you as a programmer.

> > Since this apparently isn't as obvious to you as it is to me, you
> > may be in a better position to explain it if I can convince you
> > that there's something to it.  [...]
>
> *chuckle*  True.  (Not that I'm in any special position vis a vis
> changing the documentation.  Quite the contrary.)
>
Everybody is in the *same* position where changing the documentation is
concerned: if you can think of a change that retains existing clarity to
existing audiences and makes things plainer to others, go ahead and propose
the change. I think it's fabulous that I have been able to lurk this long as
this. Now the horrible truth can be exposed: you are *just as responsible as
I am* for maintaining the documentation *and no more*.

SourceForge is open source. Find out about bug reports: the developers
aren't [all] nazis ;-)

> > [...] If we can decide on some central
> > notion behind normal __str__ usage, the examples will make more
> > sense and it won't be necessary to describe __str__ as "not __repr__".
>
> I entirely agree.  I'll have to think about the notion of
> conversion to type string and see if I can devise a more
> satisfactory explanation of __str__ in those terms. (Further
> bulletins as events warrant.)
>
you guys-are-doing-pretty-well-by-youselves-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From anton at vredegoor.doge.nl  Mon Apr 21 13:57:42 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 21 Apr 2003 19:57:42 +0200
Subject: __setitem__ puzzle
References: <b80q7c$cmd$1@news.hccnet.nl> <EKUoa.67108$T34.1782963@news2.tin.it>
Message-ID: <b81bh0$jc9$1@news.hccnet.nl>

Alex Martelli <aleax at aleax.it> wrote:

>> I have subclassed the builtin list type to accept tuples for indexing.
>
>...apparently to deal with nested lists, if I read you right.  I don't
>like that; I've long thought that L[seq] should be [L[x] for x in seq],
>and L[seq]=seq1 should be like
>
>    for i, v in zip(seq, seq1): L]i] = v
>
>with the added constraint that len(seq) == len(seq1).

See the other post for some context for this class. I hope you'll like
it more after seeing that. However it's already too late to not start
writing it. (see Dijkstra's quote earlier in this thread).

Anton



From exarkun at intarweb.us  Thu Apr 17 11:26:21 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 17 Apr 2003 11:26:21 -0400
Subject: Python Thread Question
In-Reply-To: <b7meoo$mf6$1@panix2.panix.com>
References: <bd993a2f.0304170534.245aa2d5@posting.google.com> <mailman.1050587609.6119.python-list@python.org> <b7meoo$mf6$1@panix2.panix.com>
Message-ID: <20030417152621.GE25373@meson.dyndns.org>

On Thu, Apr 17, 2003 at 10:46:48AM -0400, Aahz wrote:
> [Mostly good answers from Jp, but one correction is needed]
> 
> > [snip]
> >
> >  Context switches are usually relatively cheap compared to what your
> >app is actually doing, though, so the difference between doing blocking
> >IO in a multi-thread app and non-blocking IO in a single-thread app
> >isn't always obvious.  (Other things associated with multi-threaded
> >apps, such as deadlocks and race conditions are, though ;)
> 
> This isn't quite true on a multi-CPU box.  Because I/O releases the GIL,
> the OS can schedule threads to run on all available CPUs.  So it's
> certainly possible for a threaded app to outrun non-blocking I/O.

  Quite right.  Thanks for the correction.

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 28 days, 11:02, 6 users, load average: 0.01, 0.03, 0.00




From jack at performancedrivers.com  Tue Apr  8 02:24:29 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Tue, 8 Apr 2003 02:24:29 -0400
Subject: [ANN?] Boston PIGs
Message-ID: <20030408022428.A1039@localhost.localdomain>

After six months of subscribing to c.l.py I haven't seen any activity that
specifically targets Boston.  As implied, that is where I am.   If there was a 
caucus I would attend a monthly Python Interest Group.  Heck, if arstechnica 
can pull fifty people a couple times a year it can't be all that hard.

Would anyone be interested in attending a regular Boston Python meeting?

Possible names abound,
  BAPIG (Boston Area Python Interest Group)
  BPIGs (Boston Python Interest Group)s
  BPUG (feh)
  BeePee (phonetical)

-jack




From andrew-pythonlist at puzzling.org  Tue Apr 22 23:58:25 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 23 Apr 2003 13:58:25 +1000
Subject: Twisted 1.0.4
In-Reply-To: <3EA60532.6ADC21C@alcyone.com>
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local> <341410000.1051007646@leeloo.intern.geht.de> <20030422121836.GA8567@frobozz.local> <1051031635.18957.24.camel@software1.logiplex.internal> <mailman.1051064256.4034.python-list@python.org> <3EA60532.6ADC21C@alcyone.com>
Message-ID: <20030423035824.GB10754@frobozz.local>

On Tue, Apr 22, 2003 at 08:14:58PM -0700, Erik Max Francis wrote:
> Andrew Bennetts wrote:
> 
> > Using gcc doesn't require you to link against anything, let alone LGPLed
> > libraries -- although of course that is the most common situation; most
> > systems using gcc will also be using glibc.  But the libraries linked
> > against are a seperate issue to the point I was making about gcc; using
> > gcc doesn't necessitate using LGPLed libraries, ...
> 
> While that's in some sense technically true, any use of gcc to build an
> application -- that is, actually create an application rather than
> simply compile to object code via the -c option -- _will_ silently link
> in those pesky LGPL libraries like libc or libg++.
> 
> In special forms of embedded systems where you in effect have your own
> operating system, one might use gcc in a way where no such linking takes
> place, but for any typical application it will, whether you realize it
> or not.  (I'm not sure whether or not you're aware of this, but I
> thought I'd point it out anyway.)

Yep, I'm aware of this (but I don't mind you pointing it out, because it
might not be obvious to some).

I've never tried this, but I believe you can use gcc to build a .dll for
windows -- e.g. a .pyd to accelerate some part of a python program (look,
ma, I'm still on-topic! ;).  This would be dynamically linked to Python and
perhaps some OS services when used on Windows, but there the standard C libs
aren't LGPL...

I believe gcc can also be used on proprietary unix systems where the
standard libc is *not* glibc, but the vendor's own implementation.  For that
matter, don't the various BSD flavours have their own, BSD-licenced, libc?

Again, my point is that gcc isn't relevant to the licence of C code compiled
with it; it's just a compiler, and compiled code isn't a derived work of the
compiler, or whatever the magic legalese in the GPL is.  The libraries
compiled into or linked with your code are a seperate matter, and can have
licences relevant to your code.

-Andrew.





From mwh at python.net  Mon Apr 28 12:28:02 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 28 Apr 2003 16:28:02 GMT
Subject: __base__ and __bases__[0]
References: <2259b0e2.0304280813.5a94c41f@posting.google.com>
Message-ID: <7h37k9ek2o6.fsf@pc150.maths.bris.ac.uk>

mis6 at pitt.edu (Michele Simionato) writes:

> I thought __base__ and __bases__[0] where just equivalent, but it
> seems it is not so:
> 
> >>> class C(object): pass
> ...
> >>> class M(C,type): pass
> ...
> >>> M.__bases__[0]
> <class '__main__.C'>
> >>> M.__base__
> <type 'type'>
> 
> Question: how __base__ is computed and what's the rationale about it ?

It's found by the function best_base in typeobject.c.  I suggest you
read that...

Cheers,
M.

-- 
  We've had a lot of problems going from glibc 2.0 to glibc 2.1.
  People claim binary compatibility.  Except for functions they
  don't like.                       -- Peter Van Eynde, comp.lang.lisp



From ianb at colorstudy.com  Tue Apr 29 23:32:42 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 29 Apr 2003 22:32:42 -0500
Subject: POD vs. reST for standalone writing?
In-Reply-To: <7gy91ssyrl.fsf@pikespeak.metacarta.com>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
	 <mailman.1051590229.26633.python-list@python.org>
	 <7gy91ssyrl.fsf@pikespeak.metacarta.com>
Message-ID: <1051673562.699.218.camel@lothlorien>

On Tue, 2003-04-29 at 17:51, eichin at metacarta.com wrote:
> and the description "The underline/overline must be at least as long
> as the title text."  My first problem with this is that if changing
> the content causes the markup to need changing, the markup changes
> will be a bit more error prone.  The next is that level for a given
> punctuation type is based on, if I'm not misinterpreting this, what
> punctuation usage has been seen *in this document*, which makes moving
> selections among a small set of documents a little harder, unless you
> have an explicit style guide (in which case, why not represent it
> outright in the code?)

Yes, I have the same problem with the non-portable nature of the header
hierarchy.  I have a vague convention of what I use in my own documents
(=, -, ~), but even they aren't consistent, and it would be worse
working with someone else.  And it's hard to tell just a localized
context what the document's conventions are -- you have to start at the
top and see what punctuation is used first to really be sure.

I like the underlining itself, and if there was a strict ordering of
punctuation I wouldn't be able to remember it, so that's out.  But it
would be nice if there was something easier...

Chapter 1
---------

*Section 1.1
------------

**Section 1.1.1
---------------

Or something in that style...

  Ian






From pobrien at orbtech.com  Sun Apr  6 11:55:19 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 10:55:19 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com>
Message-ID: <m27ka7aae0.fsf@localhost.localdomain>

aahz at pythoncraft.com (Aahz) writes:

> In article <m2k7e7addl.fsf at localhost.localdomain>,
> Patrick K. O'Brien <pobrien at orbtech.com> wrote:
> >
> >Sure.  The new editor, PyAlaMode, is in the PyCrust CVS repository on
> >SourceForge:
> >
> >http://sourceforge.net/projects/pycrust
> >
> >While it works, I plan to revise it quite a bit.  Right now it just
> >uses the standard Scintilla keybindings, and there is no way to change
> >them.  In fact, there is no way to customize it at all.  But before I
> >add all those kind of features I want to design the architecture and
> >make it based upon and semi-compatible with Emacs.  If we that and
> >make it easy to integrate with Bicycle Repair Man that should make it
> >closer to a Smalltalk type environment.  But first I need to get the
> >foundation in good shape before I pile on features.  I'll take all the
> >help I can get.
> 
> Using "semi-compatible" confuses the issue, I think.  What you want is
> "comparable in power".  Another project you might want to look at is the
> Python binding for vim.

No, I really want something very similar to Emacs, not just
"comparable in power."  The reason is that there is a ton of elisp
code out there, including most of Emacs itself, that could be
translated to work with PyAlaMode.  My goal isn't to be able to
convert in a completely automated fashion, but close enough to be
attractive to both an elisp programmer and a Python programmer.

That means all the Emacs primitives need to exist and be spelled the
same (other than using underscores instead of dashes in the names),
and the ability to register functions and bind keystrokes needs to be
there, including support for define-key, global-set-key,
local-set-key, and commands bound to multiple keystrokes using
ctl-x-map and esc-map.

Plus, much of Emacs works very well.  So I'd rather benefit from
whatever design lessons the Emacs folks learned than make all the same
mistakes myself.  (I like making new mistakes.)  And if we can remove
some Emacs warts, that would be even better.

And since Emacs has a Vi mode, maybe we'll even get you using
PyAlaMode.  ;-)

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From paul at prescod.net  Thu Apr 17 20:08:16 2003
From: paul at prescod.net (Paul Prescod)
Date: Thu, 17 Apr 2003 17:08:16 -0700
Subject: Why functional Python matters
In-Reply-To: <87k7dt5m3j.fsf@interim.henrik-motakef.de>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <87k7dt5m3j.fsf@interim.henrik-motakef.de>
Message-ID: <3E9F41F0.1080609@prescod.net>

Henrik Motakef wrote:
> ...
> 
> There are really ideas to remove the functional stuff from Python? I
> wonder why people frequently think that making things impossible will
> improve programming languages.
> 
> just-because-it's-flexible-doesn't-mean-it's-perl-ly y'rs

There is nobody suggesting to make anything impossible. The question is 
whether to have two or three syntactic variations for more or less the 
same thing. Lambda is syntactic sugar for def. You literally can't 
accomplish anything (useful) with it that you couldn't with def. map is 
very near to syntactic sugar for list comprehensions (or vice versa). 
apply is syntactic sugar for "*", "**", and so forth. If Python adds new 
stuff (like "*", and list comprehensions) and never removes anything 
then there will come a day when it is full of redundant features. I 
don't really feel strongly enough to argue about lambda et. al. but I 
want to make the point that there are two sides to this argument.

  Paul Prescod






From km-news1 at onlinehome.de  Sat Apr 26 19:20:22 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Sun, 27 Apr 2003 01:20:22 +0200
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <ZFyqa.7824$3M4.220335@news1.tin.it> <oprn827qp6mu3s0q@news.cis.dfn.de> <05Bqa.8473$3M4.233690@news1.tin.it>
Message-ID: <oprn9e38pfmu3s0q@news.cis.dfn.de>

> To prove that "object.isGreen" is indeed a function, without
> actually executing the code but rather by statical examination,
OK, PyChecker is out of the game.


> I know when I'm defeated.  If the only (partial) solution
Please, never give up! :-)

-- 
Gru? - regards Klaus :-)



From eppstein at ics.uci.edu  Sun Apr 20 13:37:27 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Sun, 20 Apr 2003 10:37:27 -0700
Subject: Sorting without transitivity
References: <20030419213405.GA588@obstler> <mailman.1050821403.13707.python-list@python.org> <7h3y925o596.fsf@pc150.maths.bris.ac.uk>
Message-ID: <eppstein-1BA930.10372720042003@news.service.uci.edu>

In article <7h3y925o596.fsf at pc150.maths.bris.ac.uk>,
 Michael Hudson <mwh at python.net> wrote:

> Steven Taschuk <staschuk at telusplanet.net> writes:
> 
> > Martelli's already given the magic words "topological sort"; this
> > post is just about the terminology.  (All from memory; corrections
> > welcomed.)
> 
> Do you (or anyone else) know why it's called a *topological* sort?
> Don't see no topology here, off hand.

Graph theory is sometimes considered to be a branch of topology.  The 
usage goes back to at least 1961, when Daniel Lasser published 
"Topological ordering of a list of randomly numbered elements of a 
network" in CACM [4(4):167-168].
http://portal.acm.org/citation.cfm?doid=355578.366314

I found in MathSciNet some earlier papers (Rothe in 1939 and Kyner in 
1956) using the phrase "topological order" to mean something else
(for them, "order" is a number associated with an individual point, 
rather than a way of placing points into a sequence) so I'm guessing the 
phrase comes not out of the mathematical literature but rather out of 
the Navy project to which Lasser refers.

The earlier mathematical phrase for the same concept seems to be "linear 
extension", e.g. Ladislas Fuchs' paper "On the extension of the partial 
order of groups" [Amer. J. Math.  72,  (1950). 191--194] (amusingly 
enough the MathSciNet review of this paper uses "order" in both senses.)

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From robin at jessikat.fsnet.co.uk  Sun Apr 13 14:14:29 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Sun, 13 Apr 2003 19:14:29 +0100
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>
 <3e970abc$0$49101$e4fe514c@news.xs4all.nl>
 <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>
 <1e740$3e99a2f6$516049d2$15395@nf1.news-service.com>
Message-ID: <2+aM4KAFkam+EwV+@jessikat.fsnet.co.uk>

In article <1e740$3e99a2f6$516049d2$15395 at nf1.news-service.com>, Duncan
Grisby <duncan-news at grisby.org> writes
>In article <71M$aHCTKxl+Ew8l at jessikat.demon.co.uk>,
> Robin Becker  <robin at jessikat.fsnet.co.uk> wrote:
>
>>sigh :(
>>for real distributed computing I wouldn't want to hold the sockets open.
>
>Why ever not!?  Most protocols people use for distributed computing do
>hold sockets open. As long as connections are seen as a cache, and can
>be closed if necessary, it's a good thing. TCP start-up overhead is
>huge, especially if the network has high latency.
>
>Cheers,
>
>Duncan.
>
Well after thinking about it I guess you're right. I guess I'm harking
back to times when only small numbers of file handles could be held and
thinking that sockets might be subject to the same kind of limits.
-- 
Robin Becker



From pprobert at facstaff.wisc.edu  Mon Apr  7 20:14:13 2003
From: pprobert at facstaff.wisc.edu (Paul Probert)
Date: Mon, 07 Apr 2003 19:14:13 -0500
Subject: exceptions with spaces don't work??
Message-ID: <b6t48i$mvd$1@news.doit.wisc.edu>

As a python newbie, I came across a problem:
If I raise an exception with a string, as in:
    raise 'somestring'
then my statement
    except 'somestring':
will catch it unless there are spaces. For instance
suppose I have a file ext.py:

def re():
   raise 'x y'
   return 3
	
try:
   print 'calling re'
   x=re()
except 'x y':
   print 'got it'

now execute it:
 >>> execfile('ext.py')
calling re
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "ext.py", line 7, in ?
     x=re()
   File "ext.py", line 2, in re
     raise 'x y'
x y

now replace the spaces with _:

def re():
   raise 'x_y'
   return 3
	
try:
   print 'calling re'
   x=re()
except 'x_y':
   print 'got it'

 >>> execfile('ext.py')
calling re
got it

Is this a bug??

-- 
Paul Probert
The University of Wisconsin-Madison




From dave at boost-consulting.com  Sun Apr 20 18:00:06 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Sun, 20 Apr 2003 18:00:06 -0400
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com>
 <a3b5b608.0304080302.c26ad68@posting.google.com>
 <af6b5bdd.0304181001.558da6f@posting.google.com>
 <84of317cd6.fsf@boost-consulting.com>
 <mailman.1050870230.2976.python-list@python.org>
Message-ID: <84n0ikommx.fsf@boost-consulting.com>

Ian Bicking <ianb at colorstudy.com> writes:

> On Sun, 2003-04-20 at 04:21, David Abrahams wrote:
>> I don't care much about XML programming (haven't had a use for it),
>> but at PyConDC recently I got together with a couple of guys from the
>> Twisted project who do care, and designed a nice little meta-language
>> in Python for doing this sort of template processing.  Here's a little
>> example of it in action:
>> 
>>     template = body[
>>         table(id="outer", width="100%", height="100%", border="0")[
>>             tr(valign="bottom")[
>>                 td(id="output", width="75%", valign="top", model="latestOutput")[
>>                     div(pattern="listItem", view="html")[
>>                         "Foo"
>>                     ]
>>                 ],
>>                 td(id="room-contents", valign="bottom")[
>>                     strong[
>>                         "Stuff you have"
>>                     ],
>>                     div(model="playerInventory", view="List")[
>> 
>>                         # Note programmatic elements
>>                         if_(not arg1)
>>                         [    
>>                             div(_class="item")["Nothing"]
>>                         ].else_
>
> This could almost be plain Python, if Python allowed code blocks.  

I don't understand; it already *is* plain Python.  While Donovan
Preston and I were designing the syntax, his friend Bob was sitting on
the couch coding it up.

> The use of []'s isn't really necessary, and actually the expressions
> could maybe be eval'd as well...

You've lost me here.  Lots of the "syntax" choices aren't based on
what's neccessary, but on what's clear and expressive.  As far as
"eval" is concerned, well, I'm not sure if you realize this but those
if_ and for_each elements are lazily evaluated.  The expression
creates a callable object, "template", which generates new XML.

> td(id='room-contents', valign='bottom', contents=[
>   strong(contents=['Stuff you have']),

Uh, I don't know much XML but I'd be shocked if you weren't already
allowed to have an attribute called "contents".  I'd have used
single quotes, too, but double quotes are idiomatic for XML.


>   div(model='playerInventory', view="List", contents=[
>     if_(<not arg1>, 
>         <[div(_class="item", contents=['Nothing'])]>,
>         <[for_each("i", <arg1>,
>                   <[div(style=['color: red', 'color: blue', None],
>                         view="item", controller="look", 
>                         contents=[arg1])]>)]>)])])
>
> <> is the only nice nesting braces I could think of

Yikes!  One reason I use Python is to relieve my brain from having to
read angle brackets in C++ (and X/HTml ;-)). Plus, now you have two
levels of parenthesis nesting for each sub-element.  To my eye, your
version is *way* less-manageable than what we've already done in Pure
Python.

>  (`` wouldn't really
> do, since it doesn't nest, not to mention it looks much to much like
> '').  Of course, closing that nested structure is kind of crazy.
>
> And actually <> wouldn't work, because of if <a > 10> -- it's not
> entirely ambiguous, but it certainly *looks* ambiguous.  And << and >>
> are also used, even if it's only for one of the more stupid operators. 
> Maybe <()>...
>
> It's all pipe-dreams, but code blocks sure would be nice.  I don't think
> the argument that they allow New Control Structures is correct --
> iterators and the traditional control structures actually cover
> everything I've ever thought of.  Though now that I think about it, is
> the thing above a new control structure?  A sort of inside-out control
> structure, but also asort of lazy evaluation, like a callback.

I guess you do realize it's lazy.  The thing above is just a callable
object.  Nothing fancy <wink>

> I've played around with Python objects that allow lazily evaluated
> expressions, by overloading all the comparison operators (though you
> can't overload and, or, not, or is).  In some ways it pleases me, though
> it also feels fragile -- and it's not really compatible with normal
> Python, i.e., you couldn't call a normal function with a lazy argument
> (though you could call a method on a lazy object easily enough).

I think that if you're not concerned with object identity, you can
make lazy objects fairly transparent.  Python already hands you
proxies in a few places and you hardly ever even know the difference.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From aahz at pythoncraft.com  Fri Apr 18 01:49:38 2003
From: aahz at pythoncraft.com (Aahz)
Date: 18 Apr 2003 01:49:38 -0400
Subject: How many of you are Extreme Programmers?
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com> <mailman.1050504987.840.python-list@python.org>
Message-ID: <b7o3li$2dk$1@panix1.panix.com>

In article <mailman.1050504987.840.python-list at python.org>,
Christopher Blunck  <blunck at gst.com> wrote:
>
>Developing software
>is not all that difficult - it's always so confusing why people make
>software development more complex than it actually is.

Actually, I think that developing software *is* difficult, as difficult
as any other human endeavor that requires coordination between diverse
groups of people.  I think in a very real way claims such as yours are
precisely what hurts good development, because any yahoo who can program
thinks that zie doesn't need to learn *good* programming and social
skills.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From staschuk at telusplanet.net  Wed Apr 16 14:24:07 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 12:24:07 -0600
Subject: How many of you are Extreme Programmers?
In-Reply-To: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com>; from sismex01@hebmex.com on Wed, Apr 16, 2003 at 09:36:14AM -0500
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com>
Message-ID: <20030416122407.C4266@tibia.amotlpaa.bogus>

Quoth sismex01 at hebmex.com:
  [...]
> Many of us have used XP techniques without actually knowing
> that our methodology was labelled as "XP", it comes natural
> in the scope of our tools and goals.
> 
> Tool:   Python.
> Goal:   Write specs-conforming programs and modules.
> Method: Create tests which verify the code.
> 
> I don't know; "OOP", "XP", "Structured Programming", etc.. are
> all monickers, buzzwords, fads, which come and go; [...]

To be fair, there's a great deal more to XP than just writing
tests (as two minutes on the web will show you).  And verifying
conformance to specs is only a small part of the function of tests
in XP, or in TDD generally.

The hype may well be annoying -- it annoys me, certainly -- but
that doesn't mean the methodologies are in fact substanceless,
or not worth learning about.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"What I find most baffling about that song is that it was not a hit."
                                          -- Tony Dylan Davis (CKUA)




From gradha at titanium.sabren.com  Mon Apr 21 06:24:13 2003
From: gradha at titanium.sabren.com (Grzegorz Adam Hankiewicz)
Date: Mon, 21 Apr 2003 12:24:13 +0200
Subject: python dynamic scoping question
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DE011@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE011@admin56.narex.com>
Message-ID: <20030421102413.GA30849@pedos.es>

On 2003-04-20, Bjorn Pettersen <BPettersen at NAREX.com> wrote:
> > I have a question regarding scoping in python.  I wish to have
> > some mechanism which puts variables at a function level scope
> > into the global scope, dynamically.  I cannot alter my design.
> 
> Good luck, given what you're saying, as opposed to only the
> fragment of code, you would need perfect knowledge of all
> assignments that have dynamically happened in a program, i.e. you
> can't find these variables by static/lexical analysis. I'm almost
> positive there doesn't exists a language where you can perform
> this operation in this direction unless you consider making _all_
> variables global (dynamic scoping does something similar, but in
> the other direction :-)

Hmmm... maybe I don't understand the terminology well, but what
you say seems to be done with TCL's upvar/uplevel keywords, where
you can "overwrite" variables from the global scope, or even a
specific caller level up the stack. I think it's quite funky, and
can keep you from redesigning many things under tight deadlines.
Abusing it of course is bad, as many other things in life.

-- 
 Please don't send me private copies of your public answers. Thanks.




From duncan at NOSPAMrcp.co.uk  Wed Apr  9 09:22:51 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 9 Apr 2003 13:22:51 +0000 (UTC)
Subject: Adding item in front of a list
References: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de>
Message-ID: <Xns9358919606560duncanrcpcouk@127.0.0.1>

Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote in 
news:pan.2003.04.09.14.42.36.965880.811 at thomas-guettler.de:

> I think it would be nice if
>  l=[2, 3, 4]
>  l.insert(1)
> 
> would insert in front of the list, as opposed to append() which adds new
> elements at the end.
> 
> l.insert(1) looks better than l.insert(-1,1)
> 

If you want to do this a lot you can always subclass list:

>>> class fifo(list):
	def push(self, obj):
		self.insert(0, obj)

		
>>> f = fifo()
>>> f.push(1)
>>> f.push(2)
>>> f.push(3)
>>> f
[3, 2, 1]
>>> f.pop()
1
>>> f.pop()
2
>>> f.pop()
3


-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From gregory_r_warnes at groton.pfizer.com  Thu Apr  3 10:56:35 2003
From: gregory_r_warnes at groton.pfizer.com (Warnes, Gregory R)
Date: Thu, 3 Apr 2003 10:56:35 -0500 
Subject: Precidence of unary plus/minus relative to exponentiation
Message-ID: <D7A3CFD7825BD6119B880002A58F06C202F2C774@groexmb02.pfizer.com>

Oops.  Again it appears that I was typing with my brain off....  

-G

> -----Original Message-----
> From: Mikael Olofsson [mailto:mikael at isy.liu.se]
> Sent: Wednesday, April 02, 2003 2:45 AM
> To: python-list at python.org
> Subject: Re: Precidence of unary plus/minus relative to exponentiation
> 
> 
> On Tue, 1 Apr 2003 12:40:21 -0500
> "Warnes, Gregory R" <gregory_r_warnes at groton.pfizer.com> wrote:
> > The current operator precedence places unary plus and minus higher
> > than exponentiation.  This leads the counterintuitive result:
> > 
> > 	>>> -2.0**2.0
> > 	-4.0
> > 	
> > Python's operator predicence is causing this to be interpreted as:
> > 
> > 	>>> -(2.0**2.0)
> > 
> > rather than the normal mathematical precidence, which would give
> > 
> > 	>>> (-2.0)**2.0
> > 
> > Is there a good reason for the Python order of precedence?
> 
> I would say that the behaviour is what I expect. I would never expect
> -2.0**2.0 to mean (-2.0)**2.0. Nobody I've met in a 
> mathematical context
> (engineers and mathematicians) have ever assumed your interpretation.
> Where have you seen that interpretation?
> 
> /Mikael
> 
> --------------------------------------------------------------
> ---------
> E-Mail:  mikael at isy.liu.se
> WWW:     http://www.dtr.isy.liu.se/dtr/staff/mikael               
> Phone:   +46 - (0)13 - 28 1343
> Telefax: +46 - (0)13 - 28 1339
> 
>          /"\
>          \ /     ASCII Ribbon Campaign
>           X      Against HTML Mail
>          / \
> 
> This message was sent by Sylpheed.
> --------------------------------------------------------------
> ---------
> Link?pings kammark?r: www.kammarkoren.com
> 
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 


LEGAL NOTICE
Unless expressly stated otherwise, this message is confidential and may be privileged. It is intended for the addressee(s) only. Access to this E-mail by anyone else is unauthorized. If you are not an addressee, any disclosure or copying of the contents of this E-mail or any action taken (or not taken) in reliance on it is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately.




From anthony at interlink.com.au  Wed Apr 23 11:08:05 2003
From: anthony at interlink.com.au (Anthony Baxter)
Date: Thu, 24 Apr 2003 01:08:05 +1000
Subject: True/False 
In-Reply-To: <16037.36122.922342.405340@montanaro.dyndns.org> 
Message-ID: <200304231508.h3NF86C07880@localhost.localdomain>

>>> Skip Montanaro wrote
>     Ian> Is there any reason why reassigning True and False will be a
>     Ian> problem in the future?
> 
> The builtins may get "tightened up" in the future.  This particular idiom
> probably won't be affected, but you may not be able to (easily) shadow a
> builtin in another module's global namespace, e.g.:

Wasn't there also some talk of making None, True and False keywords (for
possible performance gains)? That would probably cause the reassignment
of them to fail.

Anthony




From brian094 at sympatico.ca  Fri Apr 11 16:06:34 2003
From: brian094 at sympatico.ca (Brian Alexander)
Date: Fri, 11 Apr 2003 16:06:34 -0400
Subject: Is curses thread-safe?
Message-ID: <3E97204A.5030106@sympatico.ca>

Hello;

Is the curses module thread-safe? Can I do this:

class windowThread(threadClass):
  def __init__(self):
   threadClass.__init__()
   self.thisWindow = curses.newwin()
   ...

... and expect every thread created to behave well during calls the the 
curses I/O routines?

Brian.




From gh at ghaering.de  Fri Apr 25 11:19:14 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 17:19:14 +0200
Subject: Linux SUID
In-Reply-To: <3ea94eae$0$23154$ba620e4c@reader0.news.skynet.be>
References: <3ea94eae$0$23154$ba620e4c@reader0.news.skynet.be>
Message-ID: <3EA951F2.10207@ghaering.de>

Helmut Jarausch wrote:
> Hi,
> upto now my favourite scripting language is (was?) Perl.
> There one can turn on the SUID-exec permission bit which
> is honoured by Perl.
> In this situation Perl enters a special "taint" mode where
> anything from the outside world like the environment or anything
> read in is tainted. If one tries to use a tained variable in
> critical operation like 'exec' or 'open' Perl refuses to do so.
> So this mechanism is safer than write an SUID program in C/C++ .
> 
> Is there something similar in Python?

Short answer: No.

There's a demo suid wrapper in the source distribution that you could 
use to get a suidperl equivalent at all. To minimize security risks, you 
can then perhaps divide your program into two processes, a suid one and 
one running with default privileges. These could communicate via a Unix 
domain socket or via pipes.

As far as I know that's how systems like Postfix work.

-- Gerhard




From bdesth at removethis.free.fr  Wed Apr 16 19:24:36 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 16 Apr 2003 23:24:36 +0000
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com>
Message-ID: <3E9DE634.4030102@removethis.free.fr>

Danra wrote:
> Here's an idea I've been playing with for a syntax for type checking
> in python. Instead of trying to explain it, I'll just start off with
> an example:
> 
(snip the example)

I think I remember having heard of something like interface or protocol 
checking in Zope (last version, probably). You may want to look at this?

Laotseu




From ckea25d02 at sneakemail.com  Mon Apr  7 23:23:59 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 15:23:59 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1049731219.27404.python-list@python.org>
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049731219.27404.python-list@python.org>
Message-ID: <b6teqm$956ai$1@ID-169208.news.dfncis.de>

Beni Cherniavsky wrote:
> Emacs has quite a lot of modes that are not about editing flat text at
> all (dired, buffer-list, custom, rmail's inbox, speedbar, etc.).
> They could well benefit from an being able to manipulate other
> structures.   To be able to use page-up/down and any other bindings
> that are not overriden, any such structure will need to provide view
> of itself in simpler terms, all down to flat text...

Now *that's* an interesting idea! We would have, not a text
editor, but a general hierarchical structure editor. Editing
plain text would just be a special case, when we're editing
an object of type "string".

This would solve a number of problems at once. Syntax
dependent editing is done by parsing the source into
a tree of objects that know how to display and edit
themselves. Embedding of "foreign" editing modes is done
simply by inserting an object of the appropriate type into
the tree. Want an reSt comment in your Python code? Just hit
M-x insert_py_rest_docstring and off you go!

The more I think about this, the more it seems like a
REALLY REALLY interesting idea...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mathieu.rohon at bull.net  Thu Apr 10 07:51:31 2003
From: mathieu.rohon at bull.net (mat)
Date: 10 Apr 2003 04:51:31 -0700
Subject: Python & Remote Access Database
References: <a6b6f195.0304090408.21003c3a@posting.google.com> <kjYka.2914$kd1.2980829@newssrv26.news.prodigy.com>
Message-ID: <a6b6f195.0304100351.54a1bef9@posting.google.com>

"drs" <drs at ecp.cc> wrote in message news:<kjYka.2914$kd1.2980829 at newssrv26.news.prodigy.com>...
> "mat" <mathieu.rohon at bull.net> wrote in message
> news:a6b6f195.0304090408.21003c3a at posting.google.com...
> > hi everyone!
> >
> > I have some trouble to find a way to connect and find informations in
> > a Remote Access Database, from my Python program which running under
> > linux Red Hat 7.3.
> >
> > the fact are:
> > -the access Database is on a WinNT4 server
> > -my python script is under Linux Red Hat 7.3, with Python 1.5
> > -both pc are ine the same domain
> >
> > I saw different module : ODBCmx, win32com... but i don't know which
> > one to use, and how to use it.
> >
> Assuming that by access you mean MS Access, I think there are 3 solutions.
> The real problem is that Access is not a db server, so using it remotely,
> even in an all windows environment, is not a simple thing (or at least not a
> good thing).
> 
> The first is solution is to use something like pyro to construct a remote
> object that runs on the windows box but is controled by the linux box. (i
> use a similar solution to this in places but between 2 win2k computers using
> dcom and so far it seems okay for processes that are not too resource
> intensive).  If you do this look at http://ecp.cc/pyado.html for
> instructions on using Python win32com/ADO.
> 
> A second similar solution is to use medusa, or something similar (twisted,
> mod_python, iis/asp, simplehttpserver, whatever) to send pickled stuff back
> and forth from the remote computer.  this is probably the most work in terms
> of hours, but perhaps the least in terms of thinking.
> 
> The last solution, which is politically often the most difficult, but which
> is probably the technically correct one, is to switch databases and use
> MySQL or Postgres or whatever you like that can actually act as a db server.
> (There is also a baby version of sql server which comes with some versions
> of visual studio that might work -- people's results have been mixed.)
> 
> -d


thanks for your answer -d

I think I'll use the first solution, it seems to be obvious and well
documented.

but in a first time i'll just use a python script to get a remote file
of data, and to store the data in my MySQL database.

do you know a way to get some file through a network, with a python
script?

thanks 

mat



From roy at panix.com  Fri Apr 25 18:46:21 2003
From: roy at panix.com (Roy Smith)
Date: Fri, 25 Apr 2003 18:46:21 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <3EA9B64D.3208D730@alcyone.com>
Message-ID: <roy-D5BE38.18462125042003@reader1.panix.com>

Erik Max Francis <max at alcyone.com> wrote:
> One thing that PyChecker, at least, is more likely to catch is "code has
> no effect" statements like
> 
> 	sys.exit
> 
> Without the parentheses, this clearly can't do anything useful

In a language as dynamic as Python, you have to be really careful with 
words like "clearly".

Invoking the name of an object to see if you raise a NameError could be 
useful in some situations.  Within the past week or so, there was a 
serious discussion on this group where code along the lines of:

   try:
      True
   except NameError:
      do stuff

was suggested as a way to tell if the version of Python being run had 
boolean support.

Not to mention that clever use of __getattr__ can make almost anything 
do something useful.  Or, if not useful, at least unexpected :-)



From exarkun at intarweb.us  Tue Apr 15 20:45:37 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 20:45:37 -0400
Subject: Why functional Python matters
In-Reply-To: <slrnb9p3vo.vg8.ramen@lackingtalent.com>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <slrnb9p3vo.vg8.ramen@lackingtalent.com>
Message-ID: <20030416004537.GB7550@meson.dyndns.org>

On Tue, Apr 15, 2003 at 11:04:06PM -0000, Dave Benjamin wrote:
> In article <3E9C9A91.8090209 at removethis.free.fr>, laotseu wrote:
> > Even for OO programmers, functionnal features in Python are IMHO a great 
> > plus, and BTW functionnal and OO paradigm does not have to conflict 
> > (that would be functionnal vs imperative and object vs procedural). CLOS 
> > is one of the great system objects out here, and it's been implemented 
> > on top of a functionnal language, so...
> 
> This is a good point. O'Caml comes to mind as well--a functional language
> with object support right out of the box. There is no reason the two
> methodologies need to be mutually exclusive, or step on each other's toes.
> 
> But, there is of course the risk of setting precedent that there are more
> than one ways to do it, and this is why I brought up Perl. I think that
> there is a subtle fear in the Python community of letting Python become too
> much like Perl, especially since both languages have similar application
> domains. There are other options besides TMTOWTDI and military rigor. This
> is precisely why I like Python, because it seems to be a good compromise
> between the two.
> 
> I think that list comprehensions bridge the gap somewhat between FP and OO
> methods, and this is one thing I really like them for. For instance:
> 
>   [obj.method(x) for x in data]
> 
> looks nicer to me than:
> 
>   map(lambda x: obj.method(x), data)
> 

  This reminds me of something I've long desired.  A "message" type.  I
think one might be implementable with descriptors, but I haven't gotten
around to figuring out just how yet.

  As many of us know, you can do this in 2.2:

    reduce(int.__add__, [1, 2, 3])

  Unfortunately, this breaks when ints get promoted to longs, because you
can't pass longs to int.__add__!  Similarly:

    map(ClassObj.method, [x, y, z])

  works, so long as x, y, and z are instances of ClassObj.  But if they are
instances of different types, all of which merely happen to implement
method, you're out of luck!  This is resolved with list comprehensions:

    [i.method(foo) for i in [x, y, z]]

  But what I'd like to see is a way to do:

    map(message("method", foo), [x, y, z])

  Of course, you could easily implement this as:

    def message(name, *args, **kwarg):
        return lambda o: getattr(o, name)(*args, **kwargs)

  But this is stupendously inefficient.  It'd be really great if there were
some built-in MessageType that could be used this way and also managed to
carry out the work in an efficient manner.

  And I think this is both a highly functional and a highly OO feature :)

  Jp

-- 
#!/bin/bash
( LIST=(~/.sigs/*.sig)
  cat ${LIST[$(($RANDOM % ${#LIST[*]}))]}
  echo -- $'\n' `uptime | sed -e 's/.*m//'` ) > ~/.signature
-- 
 up 26 days, 21:02, 4 users, load average: 0.25, 0.31, 0.63




From spam at magnetic-ink.dk  Tue Apr  8 03:07:27 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Tue, 8 Apr 2003 07:07:27 +0000 (UTC)
Subject: Tools for refactoring code
References: <4gr49v0pepuj0p4vj4t31s3mmgbo017tej@4ax.com>
Message-ID: <6f2870ef-0d7b-4b4e-93d8-86c8a0b75567@news.szn.dk>

Edmund wrote:

> In general, are there any tools to help the refactoring process?

<untested>

http://bicyclerepair.sf.net/

</untested>


  // Klaus

-- 
 ><> 	/JOIN #linux.dk @ irc.oftc.net



From deja-google at wotevah.com  Mon Apr 14 01:00:21 2003
From: deja-google at wotevah.com (L.C.)
Date: 13 Apr 2003 22:00:21 -0700
Subject: more informative stack traces
Message-ID: <e0bf36b5.0304132100.4324c3b8@posting.google.com>

I am proposing a different style of stack trace which I found more 
useful in debugging problems with live code. The main change involves
showing the actual parameter values in function calls, gdb-style.
It sounds simple because it really is, but it saves lots of guesswork
when navigating through the stack to see what caused a problem.

An implementation is available here: http://www.wotevah.com/code/log.py
To enable, just import it in your main program.

It requires syslog (thus Unix-only) because it seemed to be the natural
mechanism for reporting errors and warnings, although it will also use 
stderr if you are on a terminal.

Other useful features include methods to send debug, warning and error
messages directly to syslog, as well as a method for printing a trace
after catching an exception (or all, especially).

--
L.C. (Laurentiu C. Badea)



From andrew-pythonlist at puzzling.org  Wed Apr 23 04:40:30 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 23 Apr 2003 18:40:30 +1000
Subject: capturing output from os.system() in Windows?
In-Reply-To: <f405d9e7.0304230003.63baab05@posting.google.com>
References: <f405d9e7.0304230003.63baab05@posting.google.com>
Message-ID: <20030423084030.GH10754@frobozz.local>

On Wed, Apr 23, 2003 at 01:03:33AM -0700, David wrote:
> Being a complete novice at python, I thought I'd try a simple script
> to run a NET VIEW (which runs in a DOS box) command under Windows and
> then manipulate the resulting data.
> 
> The first problem I have encountered is how to capture the output from
> the 'NET VIEW' command.
> 
> os.system('NET VIEW') returns 0 under windows, although the DOS box
> pops up and the NET VIEW executes correctly.

Try this:

    import commands
    status, output = commands.getstatusoutput('NET VIEW')

-Andrew.





From ramen at lackingtalent.com  Tue Apr 15 15:22:47 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 15 Apr 2003 19:22:47 -0000
Subject: Why functional Python matters
Message-ID: <slrnb9on0q.sho.ramen@lackingtalent.com>

It has been my increasing concern that Python is headed in a decidedly
object-oriented direction, ignoring the benefits it currently offers to
functional programmers. Before you respond, "but Python *is*
object-oriented!", please allow me to explain. I don't know how many people
I speak for, but I feel a need to speak up and hopefully influence this
direction.

When I first discovered Python, it was with "mailman", which to this day is
the best mailing list software I have ever used. Before I was even a
competent PHP programmer, I knew of one high-quality web application written
in Python. Thus, in the back of my mind, I was interested in the language
for web-releated work (which is more or less what I do). However, at the
time, nobody I knew was writing web applications in Python. They were using
either Perl, PHP, ASP, or JSP, and being a rebellious GNU/Linux lover
frustrated with my awkwardness with Perl, I naturally chose PHP.

PHP turned out to be a very fine language for the tasks I had assigned it
to. It was fast, database-friendly, and provided massive libraries of
functionality catered toward typical web needs. As I became more advanced
with PHP, I began to uncover the beauty of its array functions. To clarify
some terminology here, an "array" in PHP is like a "dictionary" in Python
except that its keys are ordered, and it is sometimes treated like a "list".
One of the nicest things about built-in data structures like these is that
they can be dumped to the console or browser very easily. The presence of
fast, built-in collection types combined with a set of useful utility
functions and an easy method to observe their state at any point have made
them invaluable to my programming projects so far.

I did not do a lot of object-oriented programming in PHP, mainly because it
didn't buy me much, and techniques that I later discovered to be more
functional in nature were more helpful. For example, I tried looking at EJB
and modeling database records with dynamically-generated classes like Entity
Beans, but once I compared this with what I had been used to using
(typically lists of dictionaries representing tables of rows), I saw few
benefits for this method of modeling data. Representing result sets as lists
of dictionaries allowed me to generalize my algorithms and streamline the
data path from the database to the HTML template. Representing rows as
objects made my code bigger, more awkward, and harder to debug.

I should also mention that PHP's support for objects is not very robust, and
I admit in advance that this undoubtedly has biased me away from their use
in general. However, I will also note that I have a CS degree from a college
that promoted object-oriented programming very heavily, and was in the midst
of settling on Java as a standard. All feelings about Java set aside, I
think I have a fair amount of experience with--and well-rounded
understanding of--the object-oriented methodology.

My interest in functional programming exploded when I discovered the
following library of functional tools for PHP by Ian Kjos:

http://functional-php.sourceforge.net/

You might say that this library is to PHP what the Xoltar toolkit is to
Python, although the Xoltar toolkit is far more advanced. Nonetheless, this
source code and accompanying site opened my eyes to the functional style of
programming, and I saw immediate benefits which I was able to apply to my
everyday work. Learning about the power of higher-order functions and the
elimination of side-effects helped me write more concise, more reliable
software than ever before. This was a huge epiphany for me.

The next thing I read will probably not surprise you if you have ever done a
Google search on functional programming. It was David Mertz's "Charming
Python" series on IBM's DeveloperWorks. His papers were not easy to digest
in one sitting, but I made a great effort to understand them and apply the
techniques he described. At this point, I should mention that I was still
programming only in PHP!

By using Ian's functional.php library, I was satisfied to find that I could
follow many of David's examples, and fill in the gaps as I went along. Those
of you who balk at Python's lambda syntax might be less harsh if you knew
the kind of syntax we have to put up with PHP:

Python: x = lambda y, z: y + z
PHP:    $x = lam('$y, $z', '$y + $z');

The more I read of David's series, the less I could effectively translate to
PHP, mostly because it ventures into constructs like list comprehensions and
generators that PHP did not support. Eventually, Python's clean, simple
syntax and straightforward approach won me over, and I began migrating my
knowledge to the mod_python platform where I could free myself from all of
those dollar signs and semicolons and play with some really powerful
functional tools. As long as I'm naming names, let me just say that Gregory
Trubetskoy is my hero for what he did with mod_python and the publisher
module. I am greatly indebted to his work.

Now, I write web applications in PHP, mod_python, Zope, and even Java. Of
all of these platforms, my preference is mod_python because it lets me
program in Python--real Python, with no restrictions, no extension modules,
no Jython-stuck-at-Python-2.1, but Python proper. I absolutely adore Python
as a web language.

In fact, as I was originally learning Python, I began to consider it the
Perfect Language(tm). I know that this is a silly thing to do, but I had
learned over a dozen languages previously, and I honestly couldn't find a
single thing wrong with Python, and instead found myself trying to recreate
Python in every other language I used (often with positive results). Now,
I'd say I'm a lot more experienced with Python, and understand its strengths
and weaknesses better, but I still see Python as approaching perfection for
reasonable values of "perfect". =)

Guido's recent musings in the newsgroups and on his presentation on "Python
Regrets" have made me a bit uneasy. In particular, they brought to light the
fact that he generally prefers the object-oriented style to the functional
style and would like to "deprecate" certain features:

 - apply, to be replaced with *-notation
  - map and filter, to be replaced with list comprehensions
   - lambda
   
   Lambda has had the thorniest history of all of these, I think. Lambda has
its roots in LISP and the Lambda calculus, and can be used to create
"anonymous functions" at runtime that can be passed around as needed. This
is a very powerful and useful feature for functional programming, as well as
for event-driven programming. It baffles me that anyone would want to remove
it.

The first problem the "lambda" keyword had was that it couldn't be nested
easily or bound to a particular local variable namespace. This resulted in a
(minor) abuse of keyword arguments, and much criticism from misguided
functional advocates for its lack of "real closures"--I heavily debated this
because a) classes and closures are strikingly similar, and b) the
keyword-argument trick wasn't really that bad, once you got used to it. In
any case, with dynamic scoping, this problem has completely disappeared.

The second "problem" that everyone seems to keep bringing up is that lambda
limits you to expressions, not statements. This makes me laugh every time,
because this is one of the things that I *loved* about Python when I was
first learning it! If you get used to writing functions that only return
expressions, you are on the road to understanding the functional style! One
thing that I recommend that anyone try as a learning experience is this: try
writing a program where every single function starts with "return". Better
yet, write your whole program with lambdas. <screams from the back row> Of
course, I don't recommend doing this sort of thing "out in industry", but I
guarantee that you will learn to write code that is more reusable,
side-effect free, and finely decomposed once you have gained some practice
writing code this way.

I would not like to see lambda support "statements". I think that would ruin
a good thing. I would not like it removed from the language. I want it the
way it is. The only thing between "lambda" and perfection is the name. It
should be "fun" or something. If I had a lambda key on my keyboard, I would
press it. <grin>

"apply" is less of a big deal to me, but I'd still like to keep it because
it allows for greater portability. I feel so more strongly about map and
filter. Yes, I know about list comprehensions. I think they are wonderful, I
thank Guido for adding them, I thank Haskell for inspiring them, and I don't
want them to ever go away. But just because you can do a similar set of
things with list comprehensions than with map and filter doesn't mean that
they are completely redundant. In fact, I believe that it is advantageous to
design functions specifically for use with map and filter. If you design
this way, you will benefit from cleaner syntax than list comprehensions can
offer. For example:

 - map(process_record, records)
 
compared with:
 
 - [process_record(record) for record in records]
  
  I'm aware that adding lambdas, nesting, multiple filters, etc. can make
map/filter much less readable than list comprehensions, but functional
decomposition is another solution. If you have to nest that much, perhaps
your functions are too complicated and need to be broken up. I'm not saying
that this is always the case, but for the sake of clarity - in certain
situations - I greatly prefer map/filter to list comprehensions. For
starters, they eliminate the need to name the temporary variable.

Rather than remove these important functional built-ins, I would like to see
two last built-ins be added: unzip (to reverse the behavior of zip), and
curry (see the example by Alex Martelli, et. al. in the Python Cookbook,
sec. 15.7). Anything else can be thrown into a "functional" standard module.
The global namespace doesn't need any more clutter, and I sympathize with
those who wish to reduce it, but--please!--not at the expense of the
functions that form the foundation of functional programming!

I have observed Guido making several comments about Python being
object-oriented from the core, and that the natural way of expressing
solutions is object-oriented. I think that Paul Prescod feels even more
strongly about this. It is true that many common functional techniques could
be re-implemented elegantly through the use of object-oriented techniques,
and the opposite is also true. You can program o-o without side effects, and
you can emulate classes with closures. However, please do not draw the
conclusion that supporting both methodologies would be redundant. 

Please resist the temptation to oversimplify the language and force users
who enjoy Python's functional-friendly features to go back to the drawing
board, especially when Python's ability to do FP was a prime reason for
people like me to migrate to Python in the first place! I don't think that
objects are the way of the future. I think they are part of the future, but
I think that they focus too much on state, where I tend to concentrate on
data flow and functionality. If Python evolves to be a language that is
increasingly hostile toward functional programming, I will (*very
reluctantly*) have to look at other languages to focus my primary work on).

Finally, I'd like to make one comment about Perl. I think that this is
important and relevant. There is clearly a difference in philosophy between
Perl culture and Python culture. Perl says There's More Than One Way To Do
It, and Python begs to differ. Perl is all-inclusive, Python is more
exclusive. However, let's not get carried away with the exclusivity! There
are clearly multiple of ways of doing many things in Python, and that's not
a bad thing. Sometimes you want a list, other times you want an iterator.
Sometimes you want a list comprehension, other times you want map/filter.
Closures? Objects! Dispatches? Polymorphism! Inheritance? Composition? Just
because Perl says you can do anything any-which-way doesn't mean that to
compete Python should force you to a single method of problem-solving.

Perl is a beautiful thing. It encourages more creativity in the sense of
language design than any other language I have ever used, including Python.
There are many things to learn from Perl, and open-mindedness is one of
them. Java, on the other hand, is extremely rigid about its way of doing
things--trying to do functional programming in Java requires the use of
bulky wrapper classes instead of true, lightweight higher-order functions,
and its lack of parametric polymorphism makes the entire type system a pain
in the butt.

Python is a nice happy medium between regularity and freedom. I'd like to
see it stay that way.

Thanks for listening,
Dave



From paul at boddie.net  Tue Apr 15 06:24:12 2003
From: paul at boddie.net (Paul Boddie)
Date: 15 Apr 2003 03:24:12 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>	 <mailman.1049530929.9225.python-list@python.org>	 <a3b5b608.0304051049.7048d76@posting.google.com>	 <mailman.1049571010.25625.python-list@python.org>	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>	 <mailman.1049653811.11491.python-list@python.org>	 <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
Message-ID: <23891c90.0304150224.17c2fdb6@posting.google.com>

Andy Robinson <andy at reportlab.com> wrote in message news:<d5mm9v8em9j41kus2tvj2raipd7dagv18k at 4ax.com>...
> 

[Super-frameworks]

> I think it's sad that Python has this slogan "there's only one way
> to do it", but that nowhere is this less true than web development.
> At Python-UK I opined that this might be holding us back in comparison
> to other languages and tools.  Anyone coming into Java or .NET
> development gets a very clear feeling of "one way to do it".

At the lowest level, Java (and presumably .NET) does indeed have one
strongly standardised way of doing things, and there's little point in
people going against the flow at that level, especially since most
people don't do a lot of work at that level when writing their
applications. However, the same could almost be true for Python: many
frameworks seem to adopt a request/response style and reuse parts of
the CGI implementation, although it's true that not all frameworks
expose request data in similar ways.

It's also worth noting that whilst the Java platform has various
favoured frameworks for Web programming, the "framework space" is
arguably as fragmented as it is for Python. Choose your framework from
"straight JSP", Struts, Cocoon, Turbine, Expresso... (See
http://www.waferproject.org/index.html for a long list.) The
consequence of all this is that Java development projects can tend to
play it safe and select a fairly low common denominator requiring lots
of tedious legwork and wheel reinvention.

> We have a bunch of competing frameworks, but if we analyze them
> into layers then they compete less, and a more compatible approach
> must be possible.
> 
> First, let's note the half-a-dozen different templating schemes
> available.  DTML, Woven, Quixote, YAPTU, preppy.

Modularisation at the templating level is essential for generic
frameworks. Thus, I find hard-wired Python-in-HTML solutions
uninteresting.

[...]

> Wouldn't it be nice if one day we could have some standardized
> interfaces, so that one could switch between these without undue pain
> and reduce the confusion people must face?

Indeed.

Paul



From dave at pythonapocrypha.com  Wed Apr 23 16:21:27 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Wed, 23 Apr 2003 13:21:27 -0700 (PDT)
Subject: Data-driven testing
In-Reply-To: <b867d7$f53$1@panix2.panix.com>
References: <b867d7$f53$1@panix2.panix.com>
Message-ID: <Pine.LNX.4.51.0304231307230.3497@ns0.tbye.com>

On Wed, 23 Apr 2003, Aahz wrote:

> As I just mentioned in another post, I'm writing an ASCII database for
> my address book.  Actually, the main reason for writing it was because
> I've been maintaining multiple lists for various kinds of party invites,
> and the update problem was killing me.
>
> While writing it, it occurred to me that I *ought* to be using unit
> tests to make sure that it was working correctly, but two things stopped
> me.  Secondly, I was in a hurry.
>
> But firstly, I ran into the same problem I had when writing the BCD
> module: how the heck does one write tests when the results are
> hand-calculated?  There were several times when I wrote what I thought
> were correct tests for the BCD module, and my code was right and my test
> was wrong.

Hmm... seems like you'd have this problem regardless of whether or not
you write tests, right? In many ways your tests are just an automated
playback of an interactive debug/test session, so either way at some point
you have to bite the bullet and do a detailed check that the output is
right. Often you can intelligently pick a few good manual test cases that
exercize a good portion of the functionality and then do purely automated
testing of the rest (e.g. randomly choose a number, convert to BCD,
convert back, compare to original - that sort of thing).

Also, tests are a huge asset in that they detect regression in behavior,
so even if you have to double or triple check a few very manual tests it's
worth it because from then on they'll immediately notify you when
something has broken.

> In the case of my ASCII database, I'm going to need a big matrix to
> properly test the query language.  How do I build that matrix?

Exhaustive testing is sometimes economically infeasible unless you're a
government agency. :) If breaking the problem down and testing
subcomponents (in addition to end-to-end tests) still leaves too much,
then at least identify the most critical areas to test and focus on those.

The reason we don't all do formal verification of our code is that the
cost is too high. Likewise, in testing you test until the cost of a bug in
deployed code is lower than the cost of finding the bug through more
testing. For some applications it's worth it to test every code path,
every border case, etc. because it really is that critical. On the other
end of the spectrum (e.g. homegrown stuff where I'm my own customer) the
cost of a bug in "deployed" code is sometimes so small that I don't write
_any_ tests (I don't agree that writing tests _always_ saves you time -
although it _often_ or even _usually_ does).

-Dave




From reinhard at proceryon.at  Wed Apr  9 05:48:30 2003
From: reinhard at proceryon.at (reinhard)
Date: 9 Apr 2003 02:48:30 -0700
Subject: killing child processes
Message-ID: <39ee6fa9.0304090148.745fdc46@posting.google.com>

I've created a process (some batchfile) which in turn creates some
long running child processes using the win32api CreateProcess call
(declaring them as a group).

Is there a way to terminate (kill) all the child processes in the
group?

- reinhard nadrchal



From pan-newsreader at thomas-guettler.de  Thu Apr 10 07:07:11 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Thu, 10 Apr 2003 13:07:11 +0200
Subject: Adding item in front of a list
References: <mailman.1049902101.26281.python-list@python.org> <Xns93596C6F1EA0Eduncanrcpcouk@127.0.0.1>
Message-ID: <pan.2003.04.10.13.07.10.593583.815@thomas-guettler.de>

On Thu, 10 Apr 2003 11:40:38 +0200, Duncan Booth wrote:

> Tim Peters <tim.one at comcast.net> wrote in
> news:mailman.1049902101.26281.python-list at python.org:
> 
>> Guido says he also wishes list.insert() had been defined with the
>> arguments in the opposite order, so that list.insert(object) could have
>> a natural default index argument of 0.  I'd like to change that too,
>> but it's clearly too late for that one.
>> 
> I don't actually see the problem with this. list.insert() currently
> takes exactly two arguments, therefore there would be no backwards
> compatibility issues with allowing it to take one or two arguments where
> a single argument acts as though you had given two but the first
> defaulted to 0. Just like 'range' does in fact!

I newer looked at the source of python, but I don't think that it would
be difficult. Should this become a PEP or can someone with write access to the
CSV just do it?

I am sorry if I don't respond in the next days. I am moving.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From aahz at pythoncraft.com  Tue Apr 29 00:24:53 2003
From: aahz at pythoncraft.com (Aahz)
Date: 29 Apr 2003 00:24:53 -0400
Subject: POD vs. reST for standalone writing?
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
Message-ID: <b8kuql$ho3$1@panix1.panix.com>

In article <7gd6j6kpq0.fsf at pikespeak.metacarta.com>,
 <eichin at metacarta.com> wrote:
>
>So, after a recent bad experience with LyX and version skew, I was
>looking at alternatives for software end-user and administrator
>documentation.  The constraints are pretty simple:
>
>* produce good html and PDF
>  ** via linux tools, and highly preferably free ones
>* be editable on multiple platforms
>  ** linux first, but also, osx, windows
>* support "literal computer output" and screen shots

Aside from the screen shots issue, reST does reasonably well (reST may
in fact handle screen shots fine; I just can't say it does from personal
experience).

>Now I start seeing mention of reStructuredText (especially in comments
>about pyCon.)  reST files look very "ascii art" on the input side, but
>after a little experiment, that seems to make them *harder* to write,
>if somewhat (and I think only a little) easier to read. 
>
>Am I missing something?  Is there actually a reST clear short hand,
>and a tool that expands it to full-blown ascii-art?  Or is at least
>fuzzy in the counting of dashes and other punctuation for marking
>different levels?

You certainly don't need to use all of reST's power; you can teach your
users the minimum they need to get their work done.  But there's no
getting around the fact that reST does provide all that power, and users
who play around will run into it.  I wouldn't say that reST is intended
to be easy to write; it's intended to provide a balance between
writability and readability.  It's a text format that's human readable,
providing much (but far from all) the power of DocBook or LaTex.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From staschuk at telusplanet.net  Thu Apr 17 17:59:00 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 15:59:00 -0600
Subject: Teaching : Python, Scheme, Java...
In-Reply-To: <roy-8E37D1.09554617042003@malibu.unice.fr>; from roy@unice.fr on Thu, Apr 17, 2003 at 09:55:46AM +0200
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr>
Message-ID: <20030417155900.A10011@tibia.amotlpaa.bogus>

Quoth Jean-Paul Roy:
> I'm teaching Scheme in our Univ. to 2nd year students. The cursus is now:
  [...]
> Now, is it a good move to swap from Java to Python in 1st year ? Some of 
> us find Java a bit heavy for algorithmics. The probable discussion will 
> be about:
  [...]
> - Pyhton is bad : recursion limit (I am astonished at that one)

Although it's a bit annoying in principle, I've never found the
recursion limit to be a problem in practice.  (But then, I don't
usually implement Ackermann's function.)

> - Python is bad : tail recursion is not iterative (also astonished at 
> that, I can understand with an interpreter, but compiler ?)

For teaching purposes, imho tail recursion is a Bad Thing.

Writing an iterative process using tail recursion has always
struck me as a clever but obfuscated technique.  Write what you
mean: if you want an iterative process, write a loop; if you want
a recursive process, make recursive calls.

Conversely, it's easier to explain that a call always involves
stack growth, while a loop does not, instead of having to explain
that the compiler is under some circumstances able to do a call
without stack growth.

Talk about tail recursion when you start teaching the wee ones
about more advanced topics to which it is directly relevant, such
as compiler implementation, how loops might be implemented under
the hood, and so forth.

> - Python is bad : we are obliged to use message passing paradigm to add 
> an element to a list (it's ok for me but some of us get angry with 
> objects for beginners)

I'd like to hear more about this objection.  Is it that those who
object to objects (heh) feel they should be a more advanced topic?

  [...]

-- 
Steven Taschuk                                        staschuk at telusplanet.net
"Study this book; read a word then ponder on it.  If you interpret the meaning
 loosely you will mistake the Way."         -- Musashi, _A Book of Five Rings_




From theller at python.net  Tue Apr 29 03:38:39 2003
From: theller at python.net (Thomas Heller)
Date: 29 Apr 2003 09:38:39 +0200
Subject: Determining the volume ID for a file/drive?
References: <uv7savkms6lsgchcv6jh8ahtccu9lon3oi@4ax.com>
Message-ID: <8ytt4utc.fsf@python.net>

carroll at tjc.com writes:

> Is there any way to determine the volume serial number for a disk,
> given the filename (or drive, since I can determine the drive from the
> filename with os.path.splitdrive(os.path.abspath(file))?
> 
> My environment is Windows & Python 2.2.2, and I don't need for my code
> to be portable. 

win32api.GetVolumeInformation("c:\\")



From bokr at oz.net  Mon Apr  7 12:34:13 2003
From: bokr at oz.net (Bengt Richter)
Date: 7 Apr 2003 16:34:13 GMT
Subject: doing hundreds of re.subs efficiently on large strings
References: <3E80CE1D.1000200@NOmyrealCAPSbox.com> <b5qu69$t98$0@216.39.172.122> <3E8370F2.2030806@NOmyrealCAPSbox.com> <b65cq3$mi5$0@216.39.172.122> <3E90C634.1030900@NOmyrealCAPSbox.com>
Message-ID: <b6s9a5$525$0@216.39.172.122>

On Mon, 07 Apr 2003 00:28:13 GMT, nihilo <exnihilo at NOmyrealCAPSbox.com> wrote:

>Bengt Richter wrote:
>> 
>> Thanks for testing, that is interesting. It would be interesting to see
>> the test code you ran, so we could see the reasons for .25 vs .43 (and
>> I could revise my reality-model as necessary ;-)
>> 
>> Regards,
>> Bengt Richter
>
>Hi Bengt,
>
>Here is some code that I created to test the two versions.
>
>Original:
>
[...]
>
>print time.clock() - start
>
>And the modified code is:
>
>import time, re
>start = time.clock()
>data = '' # same data
>dict = { 'US and Britain' : 'Oceania', 'US and British' : 'Oceanian', 
[...~ 135 items ...]


>pat = re.compile('(Canada|the U\.S\.A\.|air force|International Herald 
[... ditto ...]
>strings = pat.split(data)
>for i in xrange(1,len(strings),2):
>     strings[i] = dict[strings[i]]
>data = ''.join(strings)
>
>I just checked these again, and the results were .052 on average for the 
>first version, and .080 on average for the second. The second would 
>obviously be a lot faster if I pre-compiled the pattern and cPickled it, 
I had assumed pre-compiled regex, but I learned something anyway:
The split is a _lot_ slower than I would have thought. I had guessed that
it would have compiled to a fast state machine like flex might generate
for or-ed static string matches, but something else must be happening.
Maybe it was not worth special casing static ors.

The loop to substitute into the list plus joining the result is quite fast,
but the split is 2+ orders of magnitude slower than the latter, and the
regex compilation is somewhat faster than the latter, but not much, apparently.
Something like 175:150:1 for split:regexcomp:subst-in-list-using-dict-and-join
in a test I ran on a 63.5k html file with 100 matching substrings in the data.

>but I wanted to include compilation time in the test.  The fastest 
Why? Just curious -- would the patterns be changing as often as the data?

>version that I have tested is using Alex's replacement suggestion in 
What version of python are you running? Mine wouldn't compile the regex
in the form Alex suggested (with every '|' term separately parenthesized,
like '(x)|(y)' vs '(x|y)' ). Maybe I typoed when I typed the change, but
I got

      File "D:\python22\lib\sre_compile.py", line 442, in compile
        assert p.pattern.groups <= 100,\
    AssertionError: sorry, but this version only supports 100 named groups

I was hoping to compare regex speed using the two forms. Can't pursue it now though...

>combination with pre-compilation and cPickle, but the naive data = 
>data.replace is surprisingly fast.
>
Yes.
BTW, neither of the regex-based methods are equivalent in semantics
to the naive sequence of data.replace statements, because each data.replace
has the possibility (in general, even if maybe not in the actual example)
of matching something created by a previous data.replace.

Thanks for posting the code.

Regards,
Bengt Richter



From aahz at pythoncraft.com  Wed Apr 30 09:38:50 2003
From: aahz at pythoncraft.com (Aahz)
Date: 30 Apr 2003 09:38:50 -0400
Subject: Question about accessing class-attributes.
References: <mailman.1051637056.20383.python-list@python.org> <jBCra.33970$K35.1085354@news2.tin.it>
Message-ID: <b8ojla$md8$1@panix1.panix.com>

In article <jBCra.33970$K35.1085354 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>
>There is probably some sound implementation reason for the current
>behavior, but if so it should be better documented, I think.

Don't bet on it.  :-/  (The sound reason, I mean.)
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From hdiwan at localhost.  Thu Apr 10 14:57:28 2003
From: hdiwan at localhost. (Hasan Diwan)
Date: Thu, 10 Apr 2003 13:57:28 -0500
Subject: Writing a TCP/IP chat server
References: <Pine.LNX.4.44.0304101906590.8243-100000@servalan.ecs.soton.ac.uk>
Message-ID: <Jaednbn65rEFIwijXTWcrg@speakeasy.net>

Ian:
	I've used method #1 and method #2. Your first possibility is the 
most flexible, but requires the most code and the most care, while the 
second is the simpler way. I haven't heard of the asyncore module. 
Contact me at hdiwan-at-mac-dot-com if you want an example of using 
SocketServer.py. 
Ian Partridge <iop199 at ecs.soton.ac.uk> writes:

>What is the best way to do this?

>I have searched the web and found three different approaches:

>1. Use the socket libraries and select to multiplex between the 
>connections.

>2. Use the SocketServer module (although I can't find a decent tutorial on 
>how to use this module)

>3. Use the asyncore module (I only found one example doing this, and I 
>couldn't understand it).
-- 
Hasan Diwan <hdiwan at mac.com> OpenPGP keyID: 0x7EE3855B
http://ibn.com/~hdiwan 
Fingerprint: 42F0 5758 C3EB BA1F ABD2  ED49 3390 CCF0 7EE3 855B



From aleax at aleax.it  Sat Apr 19 16:55:06 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 20:55:06 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it>
Message-ID: <KIioa.58844$T34.1577116@news2.tin.it>

Alex Martelli wrote:

> Steven Taschuk wrote:
> 
>> Quoth Alex Martelli:
>>   [...]
>>> "sum([y[1] for y in x])" would be very simple and direct, IMHO; the
>>> "same" thing expressed with "reduce(operator.add,[y[1] for y in x])"
>>> may only feel "simple and direct" to PhD's in maths -- it feels
>>> _abstruse_ to most people, in my opinion.
>> 
>> What if operator.add were changed to take any number of arguments?
>> Then you could write
>>     operator.add(*[y[1] for y in x])
>> Clear?  Abstruse?
> 
> That would be quite acceptable.  After all, it would then take
> nothing more than a simple:
> 
>   from operator import add as sum
> 
> to get my favourite syntax sugar for the operation:-).

Followup: actually, it would -- the mandatory * is somewhat of
a bother.  I'd rather have an operator.sum taking a sequence
(any *non-empty* sequence, see below -- or we could support
empty sequences with a second optional parameter to use as the
starting value, like reduce's third optional parameter).

> What would you have operator.add return when called with no
> arguments, though?  Either 0 or 0.0 or '' or ... have problems.
> I think it should be callable with ONE OR MORE arguments,
> not with ANY number.

I tried implementing this (as a simple patch to operator.c
in the current CVS sources) but performance is quite a
disappointment due to the mandatory * (which must build a
tuple from the supplied sequence, alas).  Using timeit.py
as usual, operator.add(*range(999)) turns out to be by far
the slowest way to sum numbers -- 967 microseconds on my 
box vs the 625 of reduce with lambda, 331 of reduce with
operator.add, and 255 of the simple for loop.  So, I think
that extending the number of arguments for operator.add
just wouldn't help here.

So, I tried implementing sum in the simplest way, as a
separate patch to operator.c (leaving add alone) -- THAT
gives me 124 microseconds, a 50% speedup with respect to
the simple for loop and clearly sufficient to justify
sum's existence (together with its simplicity, clarity,
and likely frequency of use).  As I was at it, I also
ensured that operator.sum(x) would immediately delegate
to ''.join(x) if x's first item is a string -- this way
it gets the same top performance as ''.join(x), avoiding
the trap into which the "for xx in x: t+=xx" falls here
(for x=map(str,range(999)), sum(x) clocks in at 79
microseconds, vs 78 for ''.join(x) and 1260 for the
simple loop with += ...).

Interesting... I have my doubts on whether Guido would
accept such a patch just a week before a beta, but maybe
I should try submitting it anyway...


Alex






From woodsplitter at rocketmail.com  Wed Apr  2 02:30:59 2003
From: woodsplitter at rocketmail.com (David Rushby)
Date: 1 Apr 2003 23:30:59 -0800
Subject: PyObject_CallFunction with keywords?
References: <mailman.1049251384.31007.python-list@python.org>
Message-ID: <7876a8ea.0304012330.63f405b9@posting.google.com>

Richard Cook <rcook at llnl.gov> wrote in message news:<mailman.1049251384.31007.python-list at python.org>...
> Long question:
> I am using an embedded Python interpreter in our C program to execute 
> a method from a Python module (pexpect, to be exact).  At a certain 
> point, I wish to do the equivalent of the following Python code in my 
> C code:
> 
> result =  obj.expect("hello", timeout=50);
> 
> So I have a PyObject *f which points to obj.expect, and a PyObject *s 
> which is a list containing the string "hello".  I figure I have to do:
> PyObject_CallObject(f, s, k)
> where k is the timeout parameter, but this does not match the param 
> list for PyObject_CallObject().
> 
> My question is, how can I pass keyword parameters such as the timeout 
> parameter in my example to a function which accepts them?

With Python 2.2 and later, there's PyObject_Call(callable, args,
kwargs), but strangely, it seems to have been omitted from section
"6.1 - Object Protocol" of the standard docs.

PyObject *PyObject_Call(PyObject *callable_object, PyObject *args,
PyObject *kw);



From greg at brondo.com  Tue Apr 15 15:07:27 2003
From: greg at brondo.com (Greg Brondo)
Date: Tue, 15 Apr 2003 19:07:27 GMT
Subject: Python and Schools
In-Reply-To: <b7grsc$63p$1@panix1.panix.com>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
Message-ID: <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>

Aahz wrote:
> In article <v9e21mhemm8c2 at corp.supernews.com>,
> Paul Watson <pwatson at redlinec.com> wrote:
> 
>>We need to teach students correct design priciples to build something
>>greater than what already exists.  We will never get very far if we
>>require everyone to start with a quark or atom.  Yes, of course we need
>>some people who design silicon and create microcode.  They will learn
>>the low-level details what they need to know as they need it.  Knowing
>>great design and organization principles will enable them to make the
>>most of it.
> 
> 
> While there's some truth to that, try explaining why the following code
> is a Bad Idea to someone who has no basic understanding of CS:
> 
>     s = ''
>     for i in range(1000000):
>         s += str(i)
> 
> Knowing the difference between O(N) and O(N^2) is critical to writing
> even the simplest programs.  OTOH, I do agree with you that focusing on
> programming as a craft is more important to being a programmer than
> learning CS.

I understand why the code is bad (copying the string each time in
memory as it grows).  However, I've never understood the notation's
O(N) and O(N2)?

Thanks,
Greg Brondo




From alanmk at hotmail.com  Tue Apr  1 03:07:07 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Tue, 01 Apr 2003 09:07:07 +0100
Subject: Parsing C
References: <Pine.GSO.4.44.0303311743400.19083-100000@amy4.Stanford.EDU>
Message-ID: <3E8948AB.5745C26C@hotmail.com>

Daniel Timothy Bentley wrote:
> 
> Hey,
> I find myself having to write a C interpreter in Jython.  Are there any
> useful C compilers that are open-source in python?

Is it coincidence that your post appeared in the newsgroup on April 1st
;-) ?

> Or compiler-compilers that have files for a C compiler available?
> 
> Or, compiler-compiler-compilers that have files for a compiler-compiler
> that have files for a C compiler available available?
> 
> Really, I'm looking for a way to get a C compiler and mildly painlessly
> build it to create the AST that my program will look for.
> 
> Oh, and the lexer in addition to the parser would be nice, but I bet
> it's all easy and fully of re-goodness in {p|j}ython.

I think one possible path for you is to use ANTLR, which is written in
Java and thus usable with Jython. There are ANTLR grammar files for C,
C++, and a host of other languages. Also ANTLR is very sophisticated in
terms of the representations that it builds, e.g. AST's, etc, and in
terms of the ways in which you can manipulate these representations, for
optimisation, etc. 

More from

http://www.antlr.org/

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From ianb at colorstudy.com  Wed Apr 16 01:09:09 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 16 Apr 2003 00:09:09 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <23891c90.0304140221.20e9e7fb@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
	 <mailman.1049571010.25625.python-list@python.org>
	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
	 <mailman.1049653811.11491.python-list@python.org>
	 <cQ_la.5597$LB6.158861@news1.tin.it>
	 <mailman.1050210449.24337.python-list@python.org>
	 <23891c90.0304140221.20e9e7fb@posting.google.com>
Message-ID: <1050469748.30726.104.camel@lothlorien>

On Mon, 2003-04-14 at 05:21, Paul Boddie wrote:
> Ian Bicking <ianb at colorstudy.com> wrote in message news:<mailman.1050210449.24337.python-list at python.org>...
> > Using the Twisted CGI code, it would be a quick change to make it
> > dispatch to a thread instead of a process (as with CGI).
> 
> One of the main concerns about making Twisted the "Pythonic standard"
> is that people who demand a non-threaded, non-CGI handler model for
> scalability reasons (and this did come up recently on
> comp.lang.python, so it is a valid concern) are very likely to fall
> outside the resulting audience. Unfortunately, especially given the
> success of mod_python and other process-forking frameworks, that could
> amount to a lot of people.

It seems like Twisted could fork as well...?  Its internal paradigm
(async) doesn't seem like it needs to be reflected in the things that
are built ontop of it.

> > I'm still thinking about the strategy around it, which is why I haven't
> > actually tried implementing it.  As an experiment it's not that
> > interesting to me, but maybe it could be more than that.
> 
> Well, I've been experimenting with implementing a small framework on
> top of WebKit and mod_python (with Zope being another likely
> candidate), and it has to be said that some frameworks share enough
> similarity and do provide enough flexibility to make
> "super-frameworks" possible. Moreover, with developments like POSH...
> 
>   http://poshmodule.sourceforge.net
> 
> ...it might even be possible for such "super-frameworks" to paper over
> the fundamental but desirable differences between the underlying
> frameworks. For example, some people might want a threaded model,
> whilst others might prefer a process-forking model together with
> similarly convenient data sharing semantics.

I'm increasingly drawn towards forking models as well, as they offer
more guarentees about stability, and some potential for using more
operating-system-level security (e.g., when run as root and dropping to
another user when executing a particular page).

Of course they don't port to Windows, which is why Apache 2 is so
wishy-washy about the execution model compared to 1.3.  On an operating
system like Linux, where threads are simulated with processes, I'm not
sure what the real advantage of threads are -- except for the easy data
sharing, which can be a flaw as well (since there's less partitioning). 
Maybe something like POSH would introduce the same problems.

Some partitioning can be achieved in a threading model by running
separate applications as long-running multi-threaded processes (which is
what I'd generally recommend for Webware/WebKit deployments).


Anyway, the particulars of the choice aside, there would be the
possibility of providing alternatives with a common framework ontop. 
I'm a little wary of this, though... my own perception of Webware (which
just happens to be what I'm more aware of) is that the choices provided
didn't pan out like was hoped (though I wasn't around at the point those
basic pieces were being developed).  For instance, there are a bunch of
different "adapters" meant to connect a web server to the application
server, but only a couple are actually useful (mostly mod_webkit and
wkcgi, and I suppose LWRPAdapter if you're using Xitami).  The choice is
a false one -- offering a diversity of options where most are inferior
isn't very helpful.

Even the AppServer in Webware was built with this in mind, with the
threaded option being one among many... but only the threaded option was
ever really developed, and I think it would be just as valid to add more
features to that AppServer as it would be to create an alternate
implementation with different (say forking) behavior.  No, it would be
better, because code wouldn't get as fragmented and the community
wouldn't be as fragmented.

So, choices are good, but only the best choices.  Too many choices is
just confusing. 

> As for the issue of standard APIs, I believe that such APIs really
> only serve a useful purpose as being something to build on that is
> widely agreed upon, recognised and stable; they don't necessarily
> serve the purpose of being something developers really want to write
> to directly. For example, Java Servlets have arguably served as a
> foundation for more high-level toolkits, whereas writing directly to
> the Servlet API is something that the most productive developers
> stopped doing years ago.

I think there's some basis for an API related to web stuff (on several
levels).  We just need to identify those areas that are ripe.

> Paul
> 
> P.S. The Web frameworks shootout paper is most informative and serves
> as an interesting guide as to the most accessible frameworks that
> could be subverted (or "leveraged" depending on your mindset) by such
> "super-frameworks".

Indeed, that has been on my mind too.

  Ian






From imbosol-1050446750 at aerojockey.com  Tue Apr 15 19:12:58 2003
From: imbosol-1050446750 at aerojockey.com (Carl Banks)
Date: Tue, 15 Apr 2003 23:12:58 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com>   <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <m2fzokfgni.fsf@mycroft.actrix.gen.nz> <slrnb9oajn.493.rmunn@localhost.localdomain> <yfs3ckjcsi8.fsf@black132.ex.ac.uk>
Message-ID: <_l0na.16003$883.6891@nwrdny02.gnilink.net>

Alexander Schmolck wrote:
> Robin Munn <rmunn at pobox.com> writes:
> 
>> Paul Foley <see at below.invalid> wrote:
>> > On Mon, 14 Apr 2003 16:50:40 GMT, Robin Munn wrote:
>> > 
>> >> What I really want to know is why you would consider sample #1 "broken
>> >> and fragile code". I can't think of any circumstances where sample #1
>> >> would fail to achieve its desired effect. Can you?
>> > 
>> > What if another thread (B) does the same thing, and changes sys.stdout
>> > while this thread (A) is still inside the "try" block?  And if thread
>> > A runs its "finally" clause and resets sys.stdout while thread B is
>> > still in its "try" block, they screw each other.
>> 
>> This is the classic problem that arises when two threads need to use any
>> shared resource, isn't it? I wasn't considering threads at all when I
>> wrote those two code samples, so of course I didn't use locks to ensure
>> good behavior.
>> 
>> Actually, now that I start thinking about it, any kind of
>> save-and-restore mechanism would start to get pretty hairy around
>> threads, wouldn't it? 
> 
> Yes and that's why dynamic variables *do* make a real difference: different
> threads don't have to share the same global variable.


Dynamic variables are unnecessary.  Since rebinding sys.stdout is
utterly stupid, dynamic variables or not, I will use your logging
level example:

    oldlevel = log_level[thread.get_ident()]
    log_level[thread.get_ident()] = 10
    try:
        ...
    finally:
        log_level[thread.get_ident()] = oldlevel


Encapsulate this into a few functions, and you get something easy like
this:

    oldlevel = logging.set_level(10)
    try:
        ...
    finally:
        logging.set_level(oldlevel)


There.  Without the use of dynamic globals, there is a simple system
in place to temporarily change something.  No dynamically-scoped
variables necessary.

Seeing that temporarily changing a global variable like this is such a
fringe concern, and can easily be handled *in a thread-safe manner*
with a try ... finally, I don't see any justification whatsoever for
the infamously bug-prone dynamically scoped variables.


-- 
CARL BANKS



From shalehperry at attbi.com  Fri Apr  4 23:56:59 2003
From: shalehperry at attbi.com (Sean 'Shaleh' Perry)
Date: Fri, 4 Apr 2003 20:56:59 -0800
Subject: I'm new to programming
In-Reply-To: <oprm4ogjb9acxezr@news.datafast.net.au>
References: <oprm4ogjb9acxezr@news.datafast.net.au>
Message-ID: <200304042056.59901.shalehperry@attbi.com>

>
> I know html I can look at Javascript and know what to look for when editing
> for roll overs etc but get
> me to program something from scratch huh no can do I'm at a complete loss.
>
> It's like I guess. "I know the moves in chess but ask me to checkmate a
> chessmaster in three moves forget it".
>
> Inspiration would be appreciated.

join the python tutor mailing list (linked from the python.org site).  
Interesting people, lots of interesting discussions and you may even find 
inspiration there.

Another item some find helpful is to take a program you use and try to 
reimplement it in Python.




From staschuk at telusplanet.net  Fri Apr 25 17:01:14 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 15:01:14 -0600
Subject: pythong referencing/aliasing
In-Reply-To: <ylcqa.9116$h%2.963225@read1.cgocable.net>; from khowe@perfnet.ca on Fri, Apr 25, 2003 at 11:13:21AM -0400
References: <ylcqa.9116$h%2.963225@read1.cgocable.net>
Message-ID: <20030425150114.A1403@tibia.amotlpaa.bogus>

About that subject line... PyThong?  Is this some new-fangled
scriptable underwear?

Sounds, er, a little exciting, actually.  <blush>

-- 
Steven Taschuk                                        staschuk at telusplanet.net
"Study this book; read a word then ponder on it.  If you interpret the meaning
 loosely you will mistake the Way."         -- Musashi, _A Book of Five Rings_




From sismex01 at hebmex.com  Mon Apr 14 15:54:38 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Mon, 14 Apr 2003 14:54:38 -0500
Subject: Complexity of standard Python data structures
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F0623304597605@mail.hebmex.com>

> From: marcus at infa.abo.fi [mailto:marcus at infa.abo.fi]
> Sent: Monday, April 14, 2003 3:28 PM
> 
> 
> Hmm. So the Python language is more about specifying _what_ will
> happen, instead of _how_, with no concern for actual speed or memory?
> 
> Marcus
>

Yes!

"He scores!!!"

Given that there's Python implementations in Java and C, and there's
(eventually) going to be a Python implementation in Python, then,
it's a safe thing to say (the "what" instead of the "how").

-gus




From aleax at aleax.it  Sun Apr 27 04:31:18 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 08:31:18 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <U0Mqa.13860$K35.457281@news2.tin.it> <3EAB932B.A5B42BE0@alcyone.com>
Message-ID: <qzMqa.9926$3M4.268144@news1.tin.it>

Erik Max Francis wrote:

> Alex Martelli wrote:
> 
>> I think "if f is None:" and even more "if callable(f):" are
>> better ways to express this test -- explicit is better than
>> implicit.
> 
> I agree wholeheartedly.  I was just giving an real-world example of use
> where the "Booleanness" of a function-or-maybe-not-a-function would be
> useful.

Yes, I understood that -- sorry if I wasn't explicit(:-) enough
in expressing my understanding.


> I personally explicitly do "x is None" tests for anything when I'm
> checking for Noneness, this was just as an example.

More importantly, it MAY be a widespread usage pattern (perhaps by
mistaken analogy with other "if x:" usages).  If so, then breaking
it would be unconscionable.  _However_ -- *warnings* can still be
optionally emitted, even though "if f:" were to be supported
forevermore -- newbies-who-forget-() might enable those warnings
and so still get some help, having their error diagnosed ASAP.
Hmmm...


Alex




From aleax at aleax.it  Tue Apr 15 04:28:36 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 08:28:36 GMT
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org>
Message-ID: <UoPma.20694$T34.658757@news2.tin.it>

Lulu of the Lotus-Eaters wrote:

> The topic of the O() problems of the "s += 'string'" operation has been
> discussed in several recent threads.  The aspect that seems not to have
> really come up is "why not FIX it?"
> 
> This particular slowness is not merely academic.  It's bitten me.  In
> one of his posts, I think the Martellibot even said it bit him (in the
> mists of past time). 

Yes (not in Python, the first time, but yes).

> Probably most Python programmers fell into this
> trap at least once.  I got past it, and now usually use the
> "lst.append(string); "".join(lst)" solution.  But the idea of making a
> string longer by... well, adding more to it... seems really obvious.
> Almost the "one obvious way to do it."
> 
> It seems to me that the behavior CAN be improved.  If strings were to
> use a preallocation strategy--similar to what lists do--MANY string

...then such strings would use more memory than they need to.

> appends could avoid a memory copy on the original string.  The intial
> value could stay where it was, and the extra characters could live in
> the already-available contiguous space.  Obviously, the utilized length
> would need to be stored somewhere.  Once in a while, of course, the
> preallocation would fill up; in that case you'd need to do a costly
> strcopy()... but this could be avoided MOST of the time, ISTM.

If you want s1+=s2 to be amortized O(1) for very long s1 and short s2,
you need s1's spare capacity to increase _exponentially_ when it does
need to increase.  Can be quite costly in terms of memory.


> I quite confess to not *really* understanding optimization issues.
> Doing the preallocation right could be complicated.  It's not like you
> want to preallocate a megabyte for every one-char string.  I suppose
> you'd want to determine how often a particular string already underwent
> appends, and what size strings were typically appended.  And since
> strings are immutable, it's not the self-same *objects* that underwent
> the change, but the underlying contiguous memory areas.  Still... it
> doesn't seem undoable.

Once you did it, it would be nearly impossible to retract it -- thus,
it would likely NOT be "undoable".  Which isn't what you MEAN, I think;
you probably mean it's not "unfeasible", or something like that:-).

If I wanted to pursue this kind of things, I think I'd try putting
a Boost Python wrapper on top of STL's "rope" implementation of
strings (note: STL and the Standard C++ Library are DIFFERENT things).

This would give mutable strings that allow, not just s1+=s2, but all 
sort of modifications, with very different performance characteristics
from what one normally thinks of.  Measuring performance of text
processing with those vs normal Python strings might be interesting...


Alex




From sholden at holdenweb.com  Tue Apr 15 10:06:02 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 15 Apr 2003 14:06:02 GMT
Subject: hex -> 16bit signed int (newbie)
References: <TTRma.1502$KH1.321574@news20.bellglobal.com>
Message-ID: <elUma.99378$It5.47898@news2.central.cox.net>

"zif" <zf_acs at DELETETHIShotmail.com> wrote in message
news:TTRma.1502$KH1.321574 at news20.bellglobal.com...
> How can I get 16bit signed int from a hex string?
> Eval('0x...') gives me 16bit unsigned,
> struct.pack('i', '\x..\x..') works fine, but
> Python doesn't allow me to concatenate strings
> with '\' (str(92) gives me '\\', which doesn't
> work as well).
>

Here I assume your "str(92)" was a typo/thinko for "chr(92)". If this is
correct then you are clearly being misled by the way the interpreter
presents strings. If you think about it, it doesn't make sense for a call on
chr() to give anything other than a single-character string. Sure enough, a
closer check reveals exactly that:

>>> chr(92)
'\\'
>>> len(chr(92))
1

So the reason you *see* two slashes is because the interpreter escapes the
slash to give you a value that could be used in a Python program.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From aleax at aleax.it  Sat Apr 12 16:39:36 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 20:39:36 GMT
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <mailman.1049653811.11491.python-list@python.org>
Message-ID: <cQ_la.5597$LB6.158861@news1.tin.it>

Ian Bicking wrote:

> On Sat, 2003-04-05 at 22:32, Neil Schemenauer wrote:
>> Ian Bicking <ianb at colorstudy.com> wrote:
>> > The next step I can envision would be for Twisted to become the
>> > canonical infrastructure, where higher-level web frameworks are all
>> > built ontop of it.
>> 
>> That's not feasible.  Some applications would require extensive
>> modification to work on top of an event driven framework.
> 
> That was my initial impression too, but you don't have to program in the
> Twisted/async style in order to use Twisted as a base -- Twisted can
> spawn threads and wait for responses from those threads.  I've been
> thinking about this for Webware, and I don't think there would need to
> be any change to user's code (and actually only very minimal changes to
> Webware's code -- mostly just deleting the parts that are redundant with
> Twisted and making a few wrappers).

A Webkit version running on top of Twisted would really be a wonderful
thing to have... I have a hard time conceiving of exactly how you'd fit
them together (maybe I'm not familiar enough with the _internals_ of
each, though I've used them both), but I'll accept your assessment that
it would be a pretty easy job for you, so, what about trying it out...?


> It does change the installation procedure and some of the management --
> which is okay in Twisted but not great.  However, none of the frameworks
> are great, and if we consolidate we have a better chance at improvement.

A wise strategy, that would be...


Alex




From dave at boost-consulting.com  Mon Apr 21 14:56:08 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Mon, 21 Apr 2003 14:56:08 -0400
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org>
 <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de>
 <mailman.1050725072.21332.python-list@python.org>
 <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl>
 <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl>
 <erk5avkh423eh8bc13ctvtbjqt9f6kt4kv@4ax.com>
 <84smscm5df.fsf@boost-consulting.com>
 <pf58avcm8u8r84183tia6alfo8gctc4fdu@4ax.com>
Message-ID: <84el3vptmf.fsf@boost-consulting.com>

Courageous <jkraska at san.rr.com> writes:

>>Really?  I thought longjmp could only go back up the call stack,...
>
> One *can* on some systems, although this isn't portable, and not
> supported on many, and guaranteed on none. Hence in the part you
> deleted: "although certain patterns of setjmp/longjmp aren't
> always safe." :)

References, please?
-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From Janssen at rz.uni-frankfurt.de  Thu Apr 10 15:02:46 2003
From: Janssen at rz.uni-frankfurt.de (Michael Janssen)
Date: Thu, 10 Apr 2003 21:02:46 +0200
Subject: newbie: strange python problem
References: <pan.2003.04.10.12.27.06.600319@somedomain.com>
Message-ID: <b74f4n$ap272$1@ID-190765.news.dfncis.de>

maxx wrote:

> Simple code (I think):
> 
> ---
> #!/usr/bin/python
> 
> import os
> 
> for line in os.popen('ls -la /').readlines():
>     direntry = line.split()
>     print direntry
>     print direntry[0]
>     print direntry[1]
>     print direntry[2]
> ---

beside to not trust the output of "ls" you should also write:
for field in line.split()[:3]:
    print field

this way you don't need to hardcode the slices and a "range-slice" like 
[:3] can't fail. The for-loop produce three fields of output or less if 
less is given (including no output at all).

Michael



From exarkun at intarweb.us  Mon Apr 21 13:38:13 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 21 Apr 2003 13:38:13 -0400
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
In-Reply-To: <25a672ee.0304210914.1cc2049c@posting.google.com>
References: <25a672ee.0304210914.1cc2049c@posting.google.com>
Message-ID: <20030421173813.GA16777@meson.dyndns.org>

On Mon, Apr 21, 2003 at 10:14:39AM -0700, Denis Agar wrote:
> Hi, I'm extremely new to python. I've spent the majority of my waking
> hours this easter reseaching Python GUIs, compilers, and the language
> in general. It's been arduous trying to piece together knowledge from
> different places but I'm slowly getting the hang of it. A little about
> me: My only other (major) experience with programming was a semester
> of Visual Basic in school. I really wanted to get away from the
> hodgepodge of bad vibes coming from Microsoft, and was suggested
> almost unanimously Python by a forum (http://www.shareaza.com/forum)
> 
> I'm blabbering! Anyway, my problem is this. I tried to compile a
> simple Hello World program into something runnable on non-python
> computers, and it came out being 7MB in size. Is there something I'm
> doing wrong? Just so you know, I do know somewhat what I'm doing. I
> know that it's incorporating all of the GUI. So is there a way to not
> incoporate all of the gui, and still have a functional program? I'm
> thinkin in the 800k range for a simple program? The reason I ask is
> because I'm thinking of porting something I made in VB to Python
> before I develop it any further.
> 

  Though 7MB is slightly larger than the numbers I've seen, it is in the
right ballpark.

  It's easy to see why.  You're distributing an entire runtime environment,
instead of just your application.  When you write VB code, you're only
distributing your application.

  It is unlikely you'll be able to get your app down to 800kB, though 3MB or
4MB may be likely.  If size is your main concern, I'd recommend simply *not*
bundling the Python runtime with your application.  This will give you a
"Hello world" that is 1kB or 2kB.  It will run on all machined with the
required runtime libraries, and you can simply point everyone else towards
http://www.python.org/ for a free download of what they're missing.

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 32 days, 13:02, 2 users, load average: 0.23, 0.21, 0.13




From staschuk at telusplanet.net  Tue Apr  8 21:47:38 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 8 Apr 2003 19:47:38 -0600
Subject: usage of assert and -O
In-Reply-To: <005e01c2fe26$608868c0$0ba587d9@jaws>; from matt@teaphoo.fsnet.co.uk on Wed, Apr 09, 2003 at 12:27:09AM +0100
References: <005e01c2fe26$608868c0$0ba587d9@jaws>
Message-ID: <20030408194738.B1973@tibia.amotlpaa.bogus>

Quoth Matthew Russell:
  [...]
> I would be interested to know what others think about usage of
> assert in context with optimised python code, in particular in
> releation to unittesting, as this is where i think it would be
> useful - if not critical to have one way of doing it (i.e not
> useing assert or -O [ one or the other ])

The assert statement is not well-suited for testing, precisely
because, as you point out, it gets optimized away.  For testing
you want a testing framework such as unittest or doctest.

They are also not useful for checking input or arguments, again
because they get optimized away.  For this you want an explicit
"if spam: raise eggs".

As Netzer pointed out, assert *is* useful for internal consistency
checks; such checks are handy when the code is under development,
to help pin down problems as they arise.  When the code ships,
presumably these checks are no longer necessary (because the code
works) and can be freely optimized into oblivion.

Asserts can also be somewhat useful as documentation; one might,
for example, assert an invariant of a data structure or a loop.
Here it is desirable that these assertions get optimized away
except when debugging.  (A comment can serve this purpose almost
as well; but assertions have the advantage of being automatically
checked for correctness.)

-- 
Steven Taschuk                                7\ 7'Z {&~         .
staschuk at telusplanet.net                        Y r          --/hG-
                                            (__/ )_             1^1`




From tim.one at comcast.net  Thu Apr 17 14:24:44 2003
From: tim.one at comcast.net (Tim Peters)
Date: Thu, 17 Apr 2003 14:24:44 -0400
Subject: Python and Schools
In-Reply-To: <eppstein-F15082.09415717042003@news.service.uci.edu>
Message-ID: <BIEJKCLHCIOIHAGOKOLHKEFAFGAA.tim.one@comcast.net>

[David Eppstein]
> ...
> A simple example would be an algorithm that takes as input a sequence of
> values that are all (for whatever reason) guaranteed to be integers in
> range(n**2), and tests whether any two of them are the same:
>
> def anysame(seq):
>     allocate array A, of dimension n**2, WITHOUT INITIALIZING IT
>         (is this possible in pure python?)

It shouldn't be possible:  if you find a way to do it, it would be
considered a security bug.  No memory is really uninitialized -- if Python
gave you read access to n**2 bytes in constant time, it would be revealing
to you the state left behind by some other function (or user, or process).
If you created a file with at least n**2 bytes beforehand, then you could
mmap the file in constant time, and that acts like an array object
initialized to whatever bytes happen to be sitting in the file.  But the O()
behavior of access to an mmap'ed region depends on the platform
implementation of mmap'ed regions.

>     for i,x in enumerate(seq):
>         if A[x] is an integer and 0 <= A[x] < i and A[A[x]] == x:
>             return True
>         A[x] = i
>     return False

Something must be missing here, right?  For example, suppose A just happens
to be uninitialized to all zeroes, and we pass in seq=[8, 8].  Then A[8] is
set to 0 in the first loop trip, and in the second

    A[8] is 0 so 0 <= A[x] < i is true
    A[A[x]] == A[A[8]] == A[0] == 0 != 8 so the if guard is false
    A[8] is set to 1

Then the loop ends and the function returns False.





From aleax at aleax.it  Wed Apr 30 18:31:05 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 22:31:05 GMT
Subject: Question about accessing class-attributes.
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE192@admin56.narex.com> <mailman.1051734465.1816.python-list@python.org>
Message-ID: <J8Yra.27200$3M4.754377@news1.tin.it>

<posted & mailed>

Michele Simionato wrote:

> Bjorn Pettersen wrote:
>> > Michele Simionato wrote:
>> > I must say, however, that len(S) is not *literally*
>> > type(S).__len__(S):
>> 
>> Yes, that was one of my problems. Where did you find this rule? I can't
>> find a real reference to it anywhere. For classic classes lookup is
>> o.__class__.__dict__['attr'].__get__(o.__class__, o) for regular
>> methods, which sort of looks like the above, but...
> 
> Never found any official reference except postings from Alex Martelli,

Well, my posts aren't *official* references, of course.

> and
> maybe I have read that in the Nutshell, too (Alex, correct me if I am
> wrong!)

Actually this is a tidbit I chose to explain textually instead (p. 87):
"In the new-style object model, implicit use of special methods always
relies on the class-level binding of the special method, if any", and
then a short example of the difference between classic and new-style
in this regard.

In the earlier part about the *classic* object model, p. 78, I do give:

    x.__class__.__dict__['name'](x, arg)

as being "just like" x.name(arg), but the previous paragraph does say
I'm talking of "a typical method call" and that 'name' is "naming one
of x's methods (a function-valued attribute of x's class)", so the
issue of __get__ [which wouldn't work in 2.1 -- I'm trying to cover
BOTH 2.1 and 2.2 in the part about the classic object model!] does
not need to enter the picture.

> I do agree with you that the rule does not work in the case you pointed
> out;
> not sure if it is a bug or intended behaviour.

Me neither.  It MIGHT be an "unavoidable [or too-hard-to-avoid:-)]
implementation restriction", too.  But it does need to be documented
even in that case, at least as long as it holds, I believe.


> With meta-metaclasses you may implement type(S).__len__(S) but
> I see now that your point is more with len(S) that does not
> recognizes '__len__' methods provided indirectly via __getattr__ .
> You can always do that
> 
> class Super(super):
>     def __getattr__(self,somename):
>         return somename
>       
> or even overrides __getattribute__ and implement an entirely
> different version of super. Anyway, this does not solve your
> problem.

Right.  The problem is that the special-method names must be
in the class dictionary at the time type.__new__ runs for it,
in order for the slots in the type object to be set.  They may
be set in various ways, but won't be set at all if the name
isn't in the class dictionary [or the name or slot gets
inherited from one of the class's bases] at the right time.


> It seems to me more a problem of len that of super,
> in the sense that len(x) is not *literally* type(x).__len__(x)
> (idem for __getitem__ in the other thread; I would expect the
> same for str and repr, I should check ...)

I think all of these built-ins go right for the jugular, i.e.,
the appropriate slot in the type object.  And I do NOT think that
is "a problem"... as long as the slot is set "correctly".  I do
wonder if the current non-setting in such advanced cases CAN be
deemed "correct", though.


Alex




From mac at die.spammer.die.dgp.toronto.edu  Mon Apr  7 14:18:04 2003
From: mac at die.spammer.die.dgp.toronto.edu (Maciej Kalisiak)
Date: Mon, 07 Apr 2003 18:18:04 GMT
Subject: subclassing "object" class in C?
Message-ID: <slrnb93cj7.193.mac@mac.dgp.toronto.edu>

How does one subclass the "object" class in C?  A lot of examples I've seen of
writing picklable subclasses in Python first subclass "object" (e.g. PEP 307).
I'd like to do this in C, as that's where I'm writing an extension class/type.




From staschuk at telusplanet.net  Wed Apr 16 14:19:15 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 12:19:15 -0600
Subject: Typing arguments in python
In-Reply-To: <fd1c664e.0304160524.295bbdc1@posting.google.com>; from danra@walla.co.il on Wed, Apr 16, 2003 at 06:24:15AM -0700
References: <fd1c664e.0304160524.295bbdc1@posting.google.com>
Message-ID: <20030416121915.B4266@tibia.amotlpaa.bogus>

Quoth Danra:
  [...]
> The advantages of the second definition are obvious - The fact that
> the argument passed doesn't have the write method would be detected
> immediately, rather than when trying to call the it.
> 
> Why is that so important? [...]
> 1) The stuff you do before calling write takes a lot of time, which
> could be saved if the error is detected when the function is called.

This is a really puzzling comment.  What kind of time are you
saving?  Running time of the program?  You can hardly have
discovered through profiling that the case which needs to be
optimized is the one in which your function is called with a
nonwritable object.

> 2) The object gets passed around from function to function, and maybe
> changes its name on the way. In this case it would be more difficult
> to trace the problem from the errorneous call to the original function
> call.

*boggle*  Read the traceback.  Use a debugger.  Look again at the
code you just changed.

...

Your line of thinking leads to things like this:

    def foo(writable f):
        # ...
        bar(f)
        # ...

    def bar(writable f):
        # ...
        f.write('spam')
        # ...

Here foo makes no use of f except to pass it to bar; bar needs f
to have a .write method.  For "early detection of problems", you
would, I surmise, want to have foo's argument declared as
writable, as above.  (After all, if only bar declares it writable,
a call to foo with a nonwritable object won't fail until bar is
called.)

Then you change bar to use a different method of f, and it is no
longer required to be writable.  Now you have to go find all calls
to bar and make sure that the functions making those calls don't
needlessly require their arguments to be writable.

This is tight coupling in action, and a fancy way to lose the
advantages of dynamic typing.

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From staschuk at telusplanet.net  Fri Apr 25 14:51:51 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 12:51:51 -0600
Subject: Data-driven testing
In-Reply-To: <nJGdnZlE070sxTSjXTWcog@comcast.com>; from tjreedy@udel.edu on Fri, Apr 25, 2003 at 11:46:34AM -0400
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com> <058qa.30797$y3.2515229@news010.worldonline.dk> <nJGdnZlE070sxTSjXTWcog@comcast.com>
Message-ID: <20030425125151.C1034@tibia.amotlpaa.bogus>

Quoth Terry Reedy:
  [...]
> *To paraphrase Socrates/Plato: "Who tests the testers?"

I thought that was Juvenal.

-- 
Steven Taschuk                staschuk at telusplanet.net
"I tried to be pleasant and accommodating, but my head
 began to hurt from his banality."   -- _Seven_ (1996)




From a.schmolck at gmx.net  Fri Apr 11 20:49:15 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 12 Apr 2003 01:49:15 +0100
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com>
Message-ID: <yfs8yugfsl0.fsf@black132.ex.ac.uk>

Erik Max Francis <max at alcyone.com> writes:

> My point was that if you find yourself wanting to make a truly unique
> copy of a string (an immutable object), then there is almost certainly
> something wrong with your design.

Hmm, do you mean if you want ``x is not y`` but ``x == y`` for interned string
x and some string y?

But I'm not sure that that's what the OP's up to. E.g. you could have a
container mapping symbols to other things -- possibly also symbols or further
containers. You then might want to make a deepcopy of it. You presumably still
want the symbols to be the same symbols in the copy (so that you can make use
of the identity), so asking whether ``deepcopy(symbol) is symbol`` doesn't
seem that unreasonable to me (is it?).

'as



From aleax at aleax.it  Thu Apr 17 09:55:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 13:55:52 GMT
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com> <hr9s9v8pgqr154h15lsjbk1pjs5u5n6n65@4ax.com>
Message-ID: <Inyna.27603$LB6.635831@news1.tin.it>

Courageous wrote:

>>Note that Guido deeply regrets permitting one-line blocks; this is
>>something he may try fixing for Python 3.0.
> 
> Eh? Rationale?

let me try to channel GvR (even though I'm quite a fan of one-line
blocks myself, particularly for such idioms as "if <xxx>: break")...:

Having MTOWTDI ain't Pythonic.  The ability to write either:

    if finished: break

OR quite equivalently

    if finished:
        break

is therefore a wart.  If only one of the two must survive, it's
clearly the second one (more general).  This will also ease the
task of semi-bright editors (right now, my GVIM tries to help
me by autoindenting right after each such "one-line block"...).


Alex




From aahz at pythoncraft.com  Thu Apr 17 10:40:42 2003
From: aahz at pythoncraft.com (Aahz)
Date: 17 Apr 2003 10:40:42 -0400
Subject: variable in Python
References: <b7m1g4$qqb$2@news-reader11.wanadoo.fr> <b7m3cq$r4k$1@news-reader12.wanadoo.fr>
Message-ID: <b7meda$lpp$1@panix2.panix.com>

In article <b7m3cq$r4k$1 at news-reader12.wanadoo.fr>,
Salvatore  <artyprog at wanadoo.fr> wrote:
>Salvatore wrote:
>> 
>> I have a little doubt concerning variables.
>> When one declare x = 3
>> does x contains the value 3 or
>> does x acts like pointer in C ?
> 
>I have the answer thanks:
>x is an implicit pointer

That's a good way of looking at it, but following that analogy too
closely can lead you into error.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From haribeau at gmx.de  Wed Apr 23 10:23:22 2003
From: haribeau at gmx.de (Clemens Hermann)
Date: 23 Apr 2003 07:23:22 -0700
Subject: dbm concurrency
Message-ID: <6c492956.0304230623.74b92d7f@posting.google.com>

Hi,

I want to use a standard unix dbm database to store user->regexp
combinations for squid ACLs. The dbm is written from a python cgi and
it is assured that only one write-access at a time occurs.
On the other side the same dbm is accessed read-only from a
C-programm. There may be many concurrent read-only accessess and also
requests for read-access while python writes to the dbm.

Might this be a problem when the C-program requests read-only access
while python is writing to the dbm or vice versa? If so, how can this
be solved?

thanks in advance for any hint,

/ch



From hal_wine at yahoo.com  Fri Apr  4 19:08:35 2003
From: hal_wine at yahoo.com (Hal Wine)
Date: Fri, 04 Apr 2003 16:08:35 -0800
Subject: win98 results (was Re: Cygwin Python regression testers wanted)
In-Reply-To: <mailman.1049385789.4709.python-list@python.org>
References: <mailman.1049385789.4709.python-list@python.org>
Message-ID: <3E8E1E83.4020209@yahoo.com>

Jason Tishler wrote:
> If you are willing to help, then perform the following steps:
> 
> 1. rebase your system [1]
> 
>    a. install the rebase package via Cygwin's setup.exe (if not installed)
>    b. shutdown all Cygwin processes
>    c. start bash (do not use rxvt)
>    d. execute rebaseall (in the bash window)
> 
> 2. run the Python regression test [2]
> 
>    python -E -tt /usr/lib/python2.2/test/regrtest.py -l 2>&1 | tee regrtest.out
> 
> 3. post any failures and/or hangs other than test_commands
> 
> Notes:
> 
> [1] Sorry, but Python's regression test requires a fully rebased system
>     to prevent Cygwin fork() problems.
> 
> [2] If a test hangs, then use regrtest.py's "-x" option to exclude this
>     test.  For example, in the past test_select has hung under 9X/Me, so
>     this test can be skipped as follows:

Final command:
#!/bin/bash

# test_fork1 -- system hang, reboot required
# test_hotshot  -- test hang > 1 min, kill cygwin window required
# test_popen2 -- mouse freeze & win abend dialog

python -E -tt /usr/lib/python2.2/test/regrtest.py -l 2>&1 \
     -x test_fork1		\
     -x test_hotshot		\
     -x test_popen2		\
     | tee regrtest_3.txt

test_pty requires 2 "Illegal program operation" for Python2 
dialogs answered 'debug'. test_socket takes 1

Output (test_XXXX lines removed) attached.


-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: regrtest_3.txt
URL: <http://mail.python.org/pipermail/python-list/attachments/20030404/7d50a9d6/attachment.txt>

From mcfletch at rogers.com  Wed Apr  9 16:46:10 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Wed, 09 Apr 2003 16:46:10 -0400
Subject: BEGINNER: converting integer to short month name?
In-Reply-To: <OF47A94545.30289BFA-ON85256D03.006E4C80-85256D03.006FA823@seminole.fl.us>
References: <OF47A94545.30289BFA-ON85256D03.006E4C80-85256D03.006FA823@seminole.fl.us>
Message-ID: <3E948692.9070605@rogers.com>

MCollins at co.seminole.fl.us wrote:

>hi,
>  
>
...

>that works fine for the numerical month names.  now i'm trying to convert
>the numerical month into it's short month name equivalent to grab those
>files.
>
>i could always use a long if , elif, elif to do the job, but i was
>wondering if there was built-in function already capable of that.  i was
>just looking for something a little more elegant.
>
>any suggestions?
>

    [ 'Dec','Jan','Feb','Mar', ... ] [ the_month ]

or

    import calendar
    calendar.month_abbr[ the_month ]

HTH,
Mike

 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From warlock at eskimo.com  Fri Apr 25 03:23:32 2003
From: warlock at eskimo.com (Jim Richardson)
Date: Fri, 25 Apr 2003 00:23:32 -0700
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth> <1051243370.926621@yasure>
Message-ID: <kc4nn-i04.ln1@grendel.myth>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Fri, 25 Apr 2003 04:02:52 -0000,
 Donn Cave <donn at drizzle.com> wrote:
> Quoth Jim Richardson <warlock at eskimo.com>:
> 
>| I have come across a weirdness. I have a number of floats, (lat long
>| pairs) which I stuff into a dict, but when I take them out, my nice %2f4
>| number, (say, 47.5894) is something long and *almost* right, like
>| 47.589333333333333339. What gives? I am not doing anything special as
>| far as I can tell, just stuffing the numbers into a dict, and recalling
>| it later. This is with Python 2.2 on Linux, and no, it's not an original
>| Pentium :)
> 
> They are floats, they were born weird.
> 
> You're looking at the repr() output, which for reasons we won't
> go into here renders the value to its ultimate precision.  If
> you run the same numbers through str(), they'll look different.
> 
> Maybe the big question here is whether your notion of "right"
> has any computational significance, that is, does it matter to
> your application which of these renditions more accurately
> describes the value.  Because in that respect, repr is right
> and str is almost right.
> 


It doesn't really matter to me in the application case, since one degree
to 4 or 6 points of precision, is plenty for most navigational purposes :) 


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+qOJ0d90bcYOAWPYRAj51AJ4rd84tO69/oK2ww2fsokBglVP5TwCgl2WW
KavjvS2/38mv+xPBX0h0MBg=
=GBMJ
-----END PGP SIGNATURE-----

-- 
Jim Richardson         http://www.eskimo.com/~warlock

Linux, because eventually, you grow up enough to be trusted with a fork()



From gh at ghaering.de  Fri Apr 25 13:24:11 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 19:24:11 +0200
Subject: pythong referencing/aliasing
In-Reply-To: <b8bpf5$hmj$1@ginger.mathworks.com>
References: <ylcqa.9116$h%2.963225@read1.cgocable.net>
	<b8bpf5$hmj$1@ginger.mathworks.com>
Message-ID: <3EA96F3B.3000103@ghaering.de>

Joshua Marshall wrote:
> Kevin Howe <khowe at perfnet.ca> wrote:
> 
>>How does Python handles variable references/aliases?
> 
> 
>>Example:
> 
> 
>>a = 1
>>b = (REF)a
>>b = 2
>>print a
> 
> 
>>In this case, since "b" is just a reference to "a", changes to "b"
>>automatically change "a".
> 
> 
>>Does python have this ability?
> 
> You're basically asking two questions, and both answers are "no".
> There is no way to associate two variables the same address, and there
> is no REF/address-of operator in Python.

Such an operator isn't necessary because creation of references is the 
default behaviour in Python.

Note that the assignment operator never changes an existing object. All 
it does is bind a name to an object. This object might already exist, 
like in:

#v+
a = 5	# Create the int object with the value 5 and bind the name 'a' to it
b = a	# bind the name 'b' to the object 'a' refers to
#v-

As  you seem to have a C background (wild guess), here's the C version:

#v+
PyObject* a;
PyObject* b;

a = PyInt_FromLong(5L);
b = a;	/* reference semantic, i. e. *not* *b = *a */
#v-

For the full glory, read the language specification: 
http://python.org/doc/current/ref/objects.html
It makes it very clear how things work.

Alternatively, you might create a new object in the second call, like in:

#v+
a = 5	# Create the int object with the value 5 and bind the name 'a' to it
a = 6	
# <-- Create the int object with the value 6 and bind the name 'a' to
# it. The previously created object int(5) isn't referenced any longer,
# so it is garbage-collected.
#v-

Here's the C function for you pointer-aficionados:

PyObject* a;

#v+
a = PyInt_FromLong(5L); /* reference count of 'a' is now 1 */

/* derease reference count of a. it is 0 now. Python can then 
garbage-collect it eventually. */
Py_DECREF(a);

a = PyInt_FromLong(6L);
#v-

> As others have mentioned, you might be able to use lists to give you
> the behavior you're looking for.

For assignments, Python doesn't treat immutable (such as ints) and 
mutable (such as lists) objects differently. The reference semantics are 
the same, no matter which type of object is involved.

-- Gerhard




From peter at engcorp.com  Wed Apr 16 12:53:20 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 12:53:20 -0400
Subject: How many of you are Extreme Programmers?
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com> <mailman.1050504987.840.python-list@python.org>
Message-ID: <3E9D8A80.FD76BDE9@engcorp.com>

Christopher Blunck wrote:
> 
> What's discouraging is that it often takes books, buzzwords,
> and fads to get the attention of higher-ups.  

Thankfully, in the case of XP all it took (at first) was 
a Wiki, not a book.  The book came later...

http://c2.com/cgi/wiki?ExtremeProgramming

> Developing software
> is not all that difficult - it's always so confusing why people make
> software development more complex than it actually is.

Perhaps one reason some people think it's complex or difficult is
because so many (90%?) of software projects *fail* in various ways,
including by being so far over their scheduled release date that 
their customer vanishes.

There *is* a reason why agile methodologies are being developed,
although perhaps some people aren't ready for them yet. ;-)

-Peter



From dale at riverhall.NOTHANKS.co.uk  Wed Apr 16 08:25:39 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 16 Apr 2003 13:25:39 +0100
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <d7aq9v0a5ecoue90gccfos9jtkkqcr73di@4ax.com> <Xns935F7BC1FC59duncanrcpcouk@127.0.0.1>
Message-ID: <1siq9v8kfnsnga7jr1o9jfbn6697nhv8lg@4ax.com>

Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote:

>I'm slightly surprised that your class doesn't include:
>
>    def lower(self):
>        return self.____lowerCaseMe
>
>A quick timing test indicates that this does make a difference.

I'm slightly surprised too.

Although, I'm more surpried that you haven't got anything better to do
than time other people's code. ;-)

Condiser it added.

-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From mhammond at skippinet.com.au  Sun Apr 13 09:18:42 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Sun, 13 Apr 2003 13:18:42 GMT
Subject: win32 COM test question
In-Reply-To: <cb035744.0304130224.6323b4a0@posting.google.com>
References: <mailman.1050209239.15961.python-list@python.org> <cb035744.0304130224.6323b4a0@posting.google.com>
Message-ID: <Ssdma.2913$hF.24090@news-server.bigpond.net.au>

Will Stuyvesant wrote:

> You know managers...

Yup - that's me ;)

Mark.




From peter at engcorp.com  Wed Apr  2 10:23:12 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 10:23:12 -0500
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org> <3e8afeb9.1125578@news-server>
Message-ID: <3E8B0060.D9C449A3@engcorp.com>

BBFoto wrote:
> 
> execfile('c:/program files/pysol/pysol-4.80/data/PySol_22.pyc')
> 
> invalid syntax

Why would you pass it the .pyc file when the example showed it
being passed the name of the .py file?  The two are not the
same thing; you can't use execfile() with a .pyc file!

-Peter



From snag at jool.po  Thu Apr 17 13:32:21 2003
From: snag at jool.po (kpop)
Date: Thu, 17 Apr 2003 18:32:21 +0100
Subject: More
References: <teBna.9109$pK2.11130@news.indigo.ie> <LSBna.28781$LB6.660244@news1.tin.it>
Message-ID: <sYGna.9164$pK2.11093@news.indigo.ie>

thanks
thats just what i needed





From duncan at NOSPAMrcp.co.uk  Tue Apr 29 06:19:20 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 29 Apr 2003 10:19:20 +0000 (UTC)
Subject: _winreg newbie question
References: <a5b86892.0304282156.208b81b1@posting.google.com> <a5b86892.0304290139.6870fb2b@posting.google.com>
Message-ID: <Xns936C72D87CE5Dduncanrcpcouk@127.0.0.1>

the_3_project at yahoo.com (Joe Green) wrote in 
news:a5b86892.0304290139.6870fb2b at posting.google.com:

>    if (key is "LOCAL_MACHINE"):

What do you think this line does? It doesn't compare two strings to see if 
they are equal.

Perhaps you meant:

     if key == 'LOCAL_MACHINE':

The 'is' operator checks whether two expressions both result in exactly the 
same object. This is not something you want to do very often.
Usually objects can be equal without being the same object. (You can use 
'is' to check whether an object 'is None' but that is because there is 
guaranteed only to be one None object in the system.)

> 
> -------------
> Doesn't make sense to me, I've been banging my head against the wall
> for several hours now! I can't figure it out.. and I know it's
> something very simple that I'm overlooking. I just don't see staring
> at it any longer to be productive :)
> 
> 
Some other style points (feel free to disagree/ignore):

Never use 'from <module> import *' unless the module documentation insists 
you should (_tkinter does). It is likely to come back and bite you later.

Get rid of the unnecessary parentheses in your 'if' statements. Python is 
not C.

There are several ways to avoid that unsightly list of if...elif 
statements. Perhaps the easiest would be:

    def IterKeys(self, key, subkey=None):
        regkey = getattr(self, key, None)
        ...

and then use regkey in place of 'self.reg[mykey]'.
 
Calling sys.exit when you cannot create the class seems a bit drastic. You 
might prefer to throw a more specific exception instead, or even just 
remove your try...except block and handle the exception at an outer level. 
As a general point an except block that catches every exception should make 
sure it does something useful (such as logging or displaying the error that 
actually occurred).

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From delete.this.part.joelj at attbi.com  Wed Apr 30 13:25:26 2003
From: delete.this.part.joelj at attbi.com (Joel Jensen)
Date: Wed, 30 Apr 2003 17:25:26 GMT
Subject: Python / Zope Flash Remoting
References: <MPG.1919beb84d09f671989680@netnews.attbi.com>
Message-ID: <MPG.1919c2ffeb65bf40989681@netnews.attbi.com>

Sorry about the repost-- my email address was wrong.


I am trying to convert some perl/php code to python for use in zope.
I have never done anything like this. 

If anyone would like to help, please post a response.


Flash Remoting is a way to share datastructures between Cold Fusion and 
flash,make a dataset, varible, or array in cold fusion/java, retreve it 
as a map,variable,array in flash via flashs' OO language actionscript.
http://www.macromedia.com/software/flashremoting/


This protocol is referred to by macromedia as AMF
These guys:
http://sourceforge.net/projects/amfphp
http://www.simonf.com/flap/

have hacked the protocol and written scripts to send datastructures from 
perl and php.
Make an array in PHP, call it from flash..

It would be nice if there was a version of this code available in 
python.

Zope is a Python based Object Oriented web application platform.
www.zope.org 
If flash could share datasets with ZOPE, a developer could make a Zope 
site to contain the data structure, searches, and navigation structure, 
and flash designers could access these datastructures to easily build 
dynamic flash sites.

Zope and flash already support xml-rpc, however AMF is far better.
http://www.zope.org/Members/teyc/howxmlrpc

thanks

/joel




From mk at net.mail.penguinista  Mon Apr  7 07:10:27 2003
From: mk at net.mail.penguinista (=?UTF-8?B?0LTQsNC80ZjQsNC9INCzLg==?=)
Date: Mon, 07 Apr 2003 13:10:27 +0200
Subject: urllib and binding to a interface on a multihomed system
Message-ID: <3e915ca3@news.mt.net.mk>

I have a computer that's connected to three networks. 
On that computer I have a python script that uses urllib or urllib2 to fetch
urls and POST data... 

The problem is, I want the outgoing IP address of the connections made by
urllib[2] to be one of the interfaces and not the default selected by the
kernel. I can do this to low-level sockets with the "bind" method... Can I
do the same thing to urllib?





-- 
??????                                  (jabberid:damjan at bagra.net.mk)

   "One world, one web, one program"  -- Microsoft promotional ad
         "Ein Volk, ein Reich, ein Fuehrer"  -- Adolf Hitler



From theller at python.net  Sat Apr  5 11:35:00 2003
From: theller at python.net (Thomas Heller)
Date: 05 Apr 2003 18:35:00 +0200
Subject: Calling C libraries from a Python script
References: <e81e0f3d.0304050639.37b59655@posting.google.com>
Message-ID: <r88gc37v.fsf@python.net>

joshua at goodish.org (Joshua Goodlett) writes:

> I'm trying to find examples of calling C library functions from within
> a Python script.  Any advice or pointers on where to look for examples
> would be appreciated.

If the C library is a dll, you could use ctypes.

Thomas



From b.maryniuk at forbis.lt  Thu Apr 24 13:17:47 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Thu, 24 Apr 2003 19:17:47 +0200
Subject: What means exactly "Memory error"?
In-Reply-To: <20030424155413.GA88893@hishome.net>
References: <200304241406.20833.b.maryniuk@forbis.lt> <200304241836.38456.b.maryniuk@forbis.lt> <20030424155413.GA88893@hishome.net>
Message-ID: <200304241917.47911.b.maryniuk@forbis.lt>

On Thursday 24 April 2003 17:54, Oren Tirosh wrote:
> Try to use a streaming XML API like SAX instead of DOM. For the splitting
> operation you describe there shouldn't be any need to store the entire file
> in memory. You may not even need and XML parser at all - looking for the
> HR markers could probably be done by a few regular expressions without full
> parsing of the XML stream.

Well, I only split it by <HR> and it surely causes no problem. But for the 
rest (a lot of changings in the document to adapt it for Zope) I probably 
need DOM -- regular expressions outside this job.

-- 
Regards, Bogdan

Linux -- the OS for the Renaissance Man





From a.schmolck at gmx.net  Thu Apr 10 09:24:11 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 10 Apr 2003 14:24:11 +0100
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com>
Message-ID: <yfsistmh4ec.fsf@black132.ex.ac.uk>

Alex Martelli <aleaxit at yahoo.com> writes:

> I suspect that any confusion may be caused by seeing __slots__
> as something OTHER than an optimization -- for example, as a
> way to reproduce the behavior of other languages where you
> can, or must, specify possible attributes for every class.  It
> isn't, really -- it's a memory-consumption optimization.

This confusion seems sort of understandable, given "What's new to python2.2"
introduced it explicitly as a bug avoidance feature.

__slots__ is the one addition in python2.2 that really annoys me. It seems an
ill-conceived jumble of conceptually unrelated things (bug avoidance,
optimization, safety, encapsulation) that screws up reflection big time for
little apparent gain. As Knuth said ...

'as



From gherron at islandtraining.com  Wed Apr 23 04:32:28 2003
From: gherron at islandtraining.com (Gary Herron)
Date: Wed, 23 Apr 2003 01:32:28 -0700
Subject: capturing output from os.system() in Windows?
In-Reply-To: <f405d9e7.0304230003.63baab05@posting.google.com>
References: <f405d9e7.0304230003.63baab05@posting.google.com>
Message-ID: <200304230132.28800.gherron@islandtraining.com>

On Wednesday 23 April 2003 01:03 am, David wrote:
> Being a complete novice at python, I thought I'd try a simple script
> to run a NET VIEW (which runs in a DOS box) command under Windows and
> then manipulate the resulting data.
>
> The first problem I have encountered is how to capture the output from
> the 'NET VIEW' command.
>
> os.system('NET VIEW') returns 0 under windows, although the DOS box
> pops up and the NET VIEW executes correctly.
>
> Is there some other command I should be using instead? I guess an
> alternative might be to run the 'NET VIEW' under dos and redirect the
> output into a python script using a BAT file?
>
> Any ideas?

Redirect works, but here's a better idea.

Use one of the functions in the popen2 module. These functions execute
a command with output going to a pipe (or several pipes) which can be
read by your program.

Gary Herron








From martin at v.loewis.de  Sun Apr 13 16:45:01 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 22:45:01 +0200
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk> <m38yuepdva.fsf@mira.informatik.hu-berlin.de> <LtFzMJA4ZZm+EwDh@jessikat.fsnet.co.uk> <3kjma.11793$T34.382463@news2.tin.it>
Message-ID: <m33ckm2kky.fsf@mira.informatik.hu-berlin.de>

Alex Martelli <aleax at aleax.it> writes:

> > resolve and break any loop. As you say tuples (and the None object)
> > aren't weakrefable (for some reason).
> 
> Micro-optimization...?

Backwards compatibility. To make tuples support weak references, one
has to define where the weakref slot goes into the tuple object. It
can't go at the end (which would be backwards compatible), as a tuple
is variably-sized, and it can't go in the beginning, as that
"dramatically" breaks ABI compatibility.

Performance is certainly another concern.

Regards,
Martin



From tdelaney at avaya.com  Sun Apr 13 21:00:33 2003
From: tdelaney at avaya.com (Delaney, Timothy C (Timothy))
Date: Mon, 14 Apr 2003 11:00:33 +1000
Subject: OT: Python application deployment?
Message-ID: <338366A6D2E2CA4C9DAEAE652E12A1DE4AB895@au3010avexu1.global.avaya.com>

> From: Nagy L?szl? Zsolt [mailto:nagylzs at freemail.hu]
> >
> Try NSIS (installer of winamp, made by NullSoft). It is free, makes a 
> compact installer and it has support for displaying eulas 
> (and many more).

I've recently evaluated a number of installers (to eventually replace Wise, but for now to wrap a couple of existing Wise installers).

My criteria were (in no particular order):

1. Size of final executable (esp. installer overhead);

2. Ease of development;

2a. Text-based installer script for ease of using version control

3. Standard Windows look & feel;

4. Speed and versatility of compilation (compiling custom installers as needed);

5. Ability to do everything else I needed (this one's a bit amorphous ;)

My two main options appeared to be InnoSetup and NSIS. InnoSetup was eliminated early by reading the FAQ, esp.

http://www.jrsoftware.org/isfaq.php#conditional

NSIS OTOH appears to meet all of my criteria. For example, it was quite simple to write an installer which was able to find older versions of the software and their (Wise) uninstallation script and uninstall it without user interaction.

After doing my evaluation, I later discovered that another project had independently decided on NSIS for many of the same reasons.

Tim Delaney




From daniele.trovato at obtree.com  Wed Apr 16 02:13:09 2003
From: daniele.trovato at obtree.com (Daniele Trovato)
Date: Wed, 16 Apr 2003 08:13:09 +0200
Subject: PyQT and Eric
Message-ID: <b7is7g$i1d$1@rex.ip-plus.net>

Hi

I tried to install Eric, but it won't work and I don't no why.
I'm using SuSE 8.2 and I installed Python, PyQT and PyQT Develop,
QScintilla, SIP ..... packages from the SuSE distribution.
My problem is, when I'm trying to install Eric with the following command
"python install.py", the shell returns me an error: "Sorry, Please install
PyQT"  but I have PyQT already installed.
Does anybody now where the problem is or how i can solve the problem ?

Thanks for any help

Daniele






From buzzard at urubu.freeserve.co.uk  Thu Apr 10 11:03:57 2003
From: buzzard at urubu.freeserve.co.uk (Duncan Smith)
Date: Thu, 10 Apr 2003 16:03:57 +0100
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <b7416g$fai$1@news7.svr.pol.co.uk>

"Alex Martelli" <aleaxit at yahoo.com> wrote in message
news:b73ql8011te at enews4.newsguy.com...
> I _know_ there's a decent algorithm to solve the following problem,
> but I don't recall its name and thus can't rapidly google for the
> details... can somebody help?
>
>
> The problem: I need to design a "decision criterion" to classify
> "observations".  For each observation I measure the values of a
> number N of features, x1, x2, ... xN; the desired design criterion
> is a set of weights w1, w2, ... wN such that for any observation I
> will then just compute a weighted sum
>   S = w1*x1 + w2*x2 + ... + wN*xN
> and classify the observation as Black if S<=1, White if S>1.  To
> train my classifier I have a large corpus of observations already
> made and annotated with "ground-truth" data about whether each
> given observation should have been classified as B or W, and an
> error-cost value for each kind of classification (Ebw is the cost
> of erroneously classifying a feature as W when it should be B,
> Ewb is that of classifying it as B when it should be W).  So,
> what's the algorithm to estimate the weights given the corpus of
> observation and ground-truth data, and the error-costs?
>
> Responses welcome either here or directly to my mailbox -- if I
> do get responses to my mailbox, I'll eventually summarize here --
> as a bonus, I promise to make available a Python implementation of
> whatever algorithm I end up with, if one's not already around...;-).
>
>
> TIA,
>
> Alex
>

Alex,
        I'm not sure exactly which algorithm it is that you've seen before.
But linear discriminant analysis and logistic regression would give you
functions of that form.  Is there any reason you couldn't use a Naive Bayes
classifier?  You could always drop the naive bit and use a Bayesian network
;-).  Then there's a number of tree-based classifiers (eg Chow-Liu).

Duncan





From exarkun at intarweb.us  Wed Apr  2 21:02:35 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 21:02:35 -0500
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
In-Reply-To: <b6fvei$3bu$1@panix1.panix.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <nFKia.141$Sd.25@fed1read01> <mailman.1049328065.20513.python-list@python.org> <3E8B7D49.829034FF@engcorp.com> <b6fvei$3bu$1@panix1.panix.com>
Message-ID: <20030403020234.GA29556@meson.dyndns.org>

On Wed, Apr 02, 2003 at 07:32:18PM -0500, Aahz wrote:
> In article <3E8B7D49.829034FF at engcorp.com>,
> Peter Hansen  <peter at engcorp.com> wrote:
> >
> >I hope Greg's comments about being able to disable this are on the
> >mark, because otherwise this would be a clear case of a desire for
> >(unnecessary, IMHO) optimization being given higher priority than
> >maintaining one of Python's strongest advantages.  Speed will *never*
> >be Python's strongest suit, but its dynamicism clearly is; let's not
> >bugger that up!
> 
> Maybe I'm misunderstanding something, but I was under the impression
> that disabling the ability to munge another module's namespace was at
> least partly driven by the desire to enable secure code.

  I don't think this was a consideration.

  Jp

-- 
        "I quite agree with you," said the Duchess; "and the moral of
that is -- Be what you would seem to be' -- or, if you'd like it put
more simply -- Never imagine yourself not to be otherwise than what it
might appear to others that what you were or might have been was not 
otherwise than what you had been would have appeared to them to be
otherwise.'"       -- Lewis Carrol, "Alice in Wonderland"
-- 
 up 13 days, 22:00, 4 users, load average: 1.56, 1.51, 1.46




From aleax at aleax.it  Thu Apr 17 05:49:35 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 09:49:35 GMT
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr>
Message-ID: <PMuna.38511$T34.1097150@news2.tin.it>

<posted & mailed>

Jean-Paul Roy wrote:

> Hi, snakes,
> 
> I'm teaching Scheme in our Univ. to 2nd year students. The cursus is now:
> - Java in 1st year
> - Scheme in 2nd year
> - Java, C++... in 3rd year and above (+ some Scheme to build an
> interpreter).

OK.  I might suggest adding some more-rigorously-functional language
to just Scheme (O'CAML has the advantage, for you, of being French and
having excellent French books about it...), but overall the picture
seems not bad.


> Now, is it a good move to swap from Java to Python in 1st year ? Some of

Yes, I think it WOULD be a good move, in general -- *however*, in terms
of the overall plan of study, I would then be a bit concerned if the
students were ONLY exposed to dynamically typed languages (Python first
year, Scheme second year) and never to statically typed ones until "the
third year and above".  Adding (perhaps some subset of) O'CAML would be
very good for this concern, too, because it would show students a *GOOD*
static typing scheme [not *EXCELLENT* -- that's reserved for Haskell!-) --
but, seriously, pretty good indeed, and heads and shoulders above the
kludgey typesystems of both Java and C++].

> us find Java a bit heavy for algorithmics. The probable discussion will
> be about:

Yes, Java IS indeed quite heavy for algorithmics -- Python "gets out
of your way", when you're teaching algorithms and data structures, in
a way Java never will.


> - Python is bad : no static types (it's ok for me)

It's perfectly OK *in general* that Python's typing is dynamic, just
as it's perfectly OK for Scheme, *BUT* you should consider the need
to expose students to SOME kind of static typing, preferably within
the first 2 years.  Doing it with a good, mathematically consistent
model such as ML's (any version -- O'CAML is a bit complicated by a
lot of twists and additions, but still the underlying model is quite
sound) or Haskell's is didactically VERY preferable to doing it with
the kludgey approaches of Java and the like, of course.

> - Pyhton is bad : recursion limit (I am astonished at that one)

Yeah, one spot where implementation and pragmatical issues have
taken priority over didactical and theoretical ones.  I suspect you
could use _Stackless_ Python and free yourself from the recursion
limit, if you could accept the limitation of running on Intel and
Intel-compatible CPU's only, but I don't know it for a fact.

> - Python is bad : tail recursion is not iterative (also astonished at
> that, I can understand with an interpreter, but compiler ?)

Implementation detail -- nobody's ever cared deeply about it.  I'm
sure a patch to _optionally_ deal with that would be well-received
(*optionally* because you want to be able to have the full stack
trace, normally, when a chain of calls including tail calls ends
up propagating an exception -- so, optimizing tail recursion would
need an explicit renunciation to see a complete stack trace in
such cases, and might be triggered by -O or some other dedicated
switch on the Python commandline, for example).

> - Python is bad : we are obliged to use message passing paradigm to add
> an element to a list (it's ok for me but some of us get angry with
> objects for beginners)

I'm sorry, but I just don't see this one.  E.g.:

onelist = [1, 2, 3]
anotherone = onelist + [4]

yes, the "+" operator, "under the covers", *IS* calling the
__add__ method of object onelist, but, why is that a problem?
If you want to modify the onelist in place,

onelist += [4]

does it (internally by calling __iadd__ on onelist, but,
again, why would that be a problem???).

Sure, calling onelist.append is more idiomatic, but if you don't
want to teach that right at the start, where's the problem with
onelist+=[...] or just onelist+[item] ...?  Perhaps I'm missign
something obvious -- I _would_ like you to clarify, thanks!

> - Python is bad : where are classical arrays ? Does:
>      from array import array
>   actually imports contiguous arrays ? How can they stay contiguous if
>   they support adding an element ? Is the access only O(1) amortized ?
>   Comes from : wanting to play with big arrays of bitmap images.

Python's so-called lists are actually "contiguous arrays" (of any
arbitrary objects, natch).  They stay contiguous when you add
elements by having a carefully tailored overallocation strategy
(and, in the case of .insert, ALSO shifting in memory all the
items that follow the insertion point -- and similarly for slice
assignments), and reallocating when necessary.  So, yes, appending
is amortized O(1) [inserting or deleting at an arbitrary point is O(N), 
and _accessing_ anylist[index] is "hard" O(1), not just amortized).

The type array.array is used for *homogeneous* arrays -- contiguous
arrays (just like builtin lists) where all items of one array are of 
the same primitive type (numbers or characters).  As for the type's
overallocation strategy, that may vary by release [as it did for
lists - they didn't have .append as O(1) amortized a few releases
ago, since they did only bounded overallocation back then] -- if
such implementation details are crucial to your didactical purposes,
they can be studied in the Python sources (and I'll be glad to help
you with that, if needed).  But I doubt you want to introduce the
array.array type to beginners -- it's really a memory optimization
for some common special cases, that's all!


> Aside from that points, a personal question as a Schemer who reads the
> Python primer:
> - how do you ask if a list is empty ? Is the following ok ?
>       if L == [] :

It works, though it's not idiomatic.  Other alternatives:

    if len(L) == 0: print 'L is empty'

    if not len(L): print 'L is empty'

    if not L: print 'L is empty'

the latter is idiomatic: any container is false IFF it's empty, so,
the boolean 'not' operator does all you need.

> - how do you ask if the list has only one element, with time 0(1) ? The
>   Scheme/Lisp analog is (null? (cdr L))

    if len(L) == 1: print 'L has exactly 1 element'

is the only construct that would come to most pythonistas' minds (and
yes, it's O(1)).  Other approaches you'd probably only see in some
kinds of "obfuscated Python" -- somebody trying hard to be clever --
and they might e.g. exploit slicing:

    if not L[1:2]: print 'L has 0 or 1 elements'

Slicing can be handy because it's tolerant of indices out of bounds
(differently from indexing, which raises exceptions when you index
out of bounds!).  But it makes no sense to "get too cute about it";
when you're interested in the length of L, len(L) is the way to go.

> - more generally, is the length of a list a builtin ? Is the L[k] access
>   time O(1) with a list ? Probably not. And with a tuple ?

len is a built-in function.  len(L) is O(1) for any list L, and len(T)
is O(1) for any tuple T.  Indexing, L[k] and T[k], are also both O(1).


> Sorry for so many questions and your time.

The root of half of your problems is the strange use of the term
"list" in Python -- QUITE different from most uses of the same term
in all of CS and programming literature -- so it's WE, the Python
community, who should be apologizing to you for wasting some of
your time with it.  It's too late of course to eradicate the Python
use of the term "list" (the word 'list' itself is the name of the
built-in type, among other considerations) but at least the tutorial
should definitely mention it up front that a 'list' is really a
(potentially heterogeneous) vector (able to grow and shrink, but
contiguous in memory at all times -- this is somewhat more than an
implementation detail, as it governs the O()-behavior of many of
the aspects of Python lists...).


> But thanks a lot, I guess that Python is a good guy :-)

You're welcome!!!  And, yes, you'll do your students a world of
good if you can teach them Python as their first language -- it
can give them a very useful tool that will help them in a lot of
ways throughout their careers, both conceptually and practically
(e.g., when they later get around to studying sockets, or Java
libraries, etc, etc, being able to play interactively with them
with Python [or Jython -- same language, but implemented on top
of a JVM] will perhaps ease their learning enormously).


Alex




From staschuk at telusplanet.net  Thu Apr 17 15:29:19 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 13:29:19 -0600
Subject: Python and Schools
In-Reply-To: <BIEJKCLHCIOIHAGOKOLHKEFAFGAA.tim.one@comcast.net>; from tim.one@comcast.net on Thu, Apr 17, 2003 at 02:24:44PM -0400
References: <eppstein-F15082.09415717042003@news.service.uci.edu> <BIEJKCLHCIOIHAGOKOLHKEFAFGAA.tim.one@comcast.net>
Message-ID: <20030417132919.E7968@tibia.amotlpaa.bogus>

Quoth Tim Peters:
> [David Eppstein]
  [...]
> >     allocate array A, of dimension n**2, WITHOUT INITIALIZING IT
> >         (is this possible in pure python?)
> 
> It shouldn't be possible:  if you find a way to do it, it would be
> considered a security bug.  No memory is really uninitialized -- if Python
> gave you read access to n**2 bytes in constant time, it would be revealing
> to you the state left behind by some other function (or user, or process).

Hm... it doesn't have to let us *read* from the uninitialized
memory.  Just as long as it lets us write to it, and later read
from the spots we've written to.

This is of dubious usefulness, of course.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From ramen at lackingtalent.com  Tue Apr  1 10:27:47 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 15:27:47 -0000
Subject: Terminal-based text editor with color syntax hilighting
References: <slrnb8hvka.pki.ramen@lackingtalent.com> <mailman.1049169002.28920.python-list@python.org>
Message-ID: <slrnb8j93q.421.ramen@lackingtalent.com>

In article <mailman.1049169002.28920.python-list at python.org>,
Sean 'Shaleh' Perry wrote:
> 
> new gnu emacs (around 8 months old I think) and vim (a super vi clone) both 
> support color coding.
> 

Ahh, that's right, I forgot about vim. I remember seeing vim in color over a
terminal, and if I recall correctly, it can write colored source to HTML
too... Unfortunately, I have such a hard time with vi - I feel like I have
to RTFM every time I use it again. ;)

Found some information about text terminal colors for emacs here:
http://www.gnu.org/manual/elisp-manual-21-2.8/html_node/elisp_469.html

Are there any others? What about Fte, I wonder... is that editor/IDE still
being actively maintained? According to their website it supports Python
color hilighting, but I'm not sure if the s-lang version does. I'll check
into it.

Thanks,
Dave



From andrew at acooke.org  Sat Apr 26 13:40:29 2003
From: andrew at acooke.org (andrew cooke)
Date: Sat, 26 Apr 2003 13:40:29 -0400 (CLT)
Subject: Some Python Code
In-Reply-To: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com>
References: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com>
Message-ID: <48823.127.0.0.1.1051378829.squirrel@127.0.0.1>

why not put the code on a web page somewhere?  that way it will eventually
get googled and people will find it when they need something like that and
search for it, and they can download it.

cheers,
andrew

Sean Ross said:

> Hi.
> I've just finished a semester of university where I've been fortunate
> enough
> to be able to use my language of choice, Python, for all of my programming
> projects/assignments. I've implemented a general purpose genetic algorithm
> package and a little circuit description programming language, plus a few
> other AI related items of differing quality. I don't really want to do any
> maintenance on these projects, so I'll not be bothering to put them up in
> some place like sourceforge or pypi. However, the code may be useful to
> someone else, perhaps as a starting point for their own projects, so I'd
> like to make it available. For anyone who may be interested, I can be
> contacted at sross at connectmail.carleton.ca or
> frobozz_electric at hotmail.com.
> The code is, of course, free and open source although it has not been
> licensed - except the programming language which contains a SPARK license.
> OK. Thanks for your time.
> Sean

-- 
http://www.acooke.org/andrew




From peter at engcorp.com  Thu Apr 10 15:42:14 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 10 Apr 2003 15:42:14 -0400
Subject: Python compared toJava
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com> <3E95A2A7.238BFA6C@engcorp.com> <3e95b7d6$1@rmcnt03a.rmc.ca>
Message-ID: <3E95C916.F8F019C7@engcorp.com>

bourassa wrote (using top-posting, which I've fixed):
> 
> Peter Hansen wrote:
> > Blaktyger wrote:
> >
> >>I recently read that the Python interpreter is like the Java Virtual Machine.
> >>Is that true?
> >
> > Yes!  In fact, it's exactly the same thing!  We just use a different name
> > for it to avoid the wrath of Sun Microsoftsystems, and because we kind of like
> > snakes.
> >
> > Note: see http://www.catb.org/~esr/faqs/smart-questions.html before posting
> > again please.

> Would it have bee so hard to either ignore the post or to give a useful
> url?  Learn some manners before _your_ next post.

A "useful" URL?  That's *exactly* what I did give.  I'm uncertain how you
expect one to define "useful" any better than that when the question 
asked is so incredibly vague and meaningless.  

Yes, I could have ignored the post, but then the OP wouldn't have had
the opportunity to learn to ask better questions, would he?  Seems like
I've actually helped him out in the long term, if he's willing to listen 
to good advice.  ;-)

-Peter



From max at alcyone.com  Wed Apr 16 18:43:13 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 15:43:13 -0700
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com>
Message-ID: <3E9DDC81.932475AF@alcyone.com>

Peter van Kampen wrote:

> I think I almost 'get it'. Except who or what decides what Big-O a
> certain algorithm has. Is that by 'hunch' or by running benchmarks or
> are there still other ways to recognize 'Big-O's (except for the
> formal computations that I understand are not very common)?

It's just mathematics.  You look at the behavior of the algorithm or
data structure as the number of elements gets very large (in order to
drown out fixed-cost or second-order effects).

> for example
> 
> bigO_N(i):
>     for n in xrange(i):
>         print n
> 
> Would be O(N) right?

That's right.  It iterates over each element, and that takes O(n).

> bigO_N2(i,j):
>     for n in xrange(i):
>         for o in xrange(j):
>             print i,j
> 
> This would be O(N**2)? Actually O(i*j) I suppose but if i==j N**2
> applies.

That's right also.  You're looking at the case where the number of
elements being processed -- even if there are multiple different sets --
grows large.  In this case you're iterating over two lists, but for each
list you're iterating over the other, so as i, j -> oo the behavior is
O(n^2).

> But these are not really 'algorithms'. Most algorithms are a lot more
> complex even when reduced to it's essentials. I understand why
quicksort is better than bubblesort but I don't see an obvious way to
label quicksort O(??). 

But there is always some set of behaviors that are core to the
algorithm, and that's what you look at as the number of elements goes to
infinity.  Bubble sort is O(n^2), but quicksort is O(n ln n).

One also often speaks of worst case and average cases (this ties into
what people mean by "amortized O(f(n))" behavior).  For instance, take
finding an element in a sorted binary tree.  The average case is O(ln
n), but the worst case (if the tree is maximally uneven) is O(n).

Adding an element to the end of a dynamically resizing vector is average
case O(1), but worst case O(n), but since this worst case happens on
average n times, we call it "amortized O(1)" behavior.

The big-O notation is one of those things that's really highly
mathematical (particularly if you want to talk about things rigorously),
but it's also something that's vitally important for even novice
programmers to understand very well when they're using algorithms to
solve complex problems that need to be scalable.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The only source of knowledge is experience.
\__/ Albert Einstein
    Bosskey.net / http://www.bosskey.net/
 A personal guide to online multiplayer first person shooters.



From ckea25d02 at sneakemail.com  Wed Apr 16 22:56:52 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 17 Apr 2003 14:56:52 +1200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <b7jv2v$1vd4$1@nntp6.u.washington.edu>
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org> <b7di0n$dd6gj$1@ID-169208.news.dfncis.de> <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net> <1050390062.742825@yasure> <b7i1mm$17cpu$1@ID-169208.news.dfncis.de> <b7jv2v$1vd4$1@nntp6.u.washington.edu>
Message-ID: <b7l4ge$1u76i$1@ID-169208.news.dfncis.de>

Donn Cave wrote:
> But the idea that repr should be the function that conveys that
> "whole truth" is suspiciously consistent with the marshalling idea.

Just to clarify, I wasn't suggesting that the "whole
truth" idea should appy to __repr__ in general --
only that it seems to have been the motivation in
the particular case of floats.

In fact, for debugging output, the "whole truth"
is often far too MUCH information! An object with
a big complex data structure underneath is, I feel,
best __repr__'ed simply as <someobject at 0x12876ad>
rather than trying to print it all out somehow.

If I need to know more, I'll insert a debugging
print that prints out a bit more. I find print
someobj.__dict__ to be quite useful at times like
this.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From hs2l2ATyahoodotcom at yahoo.com  Fri Apr 18 00:22:31 2003
From: hs2l2ATyahoodotcom at yahoo.com (Haran Shivanan)
Date: 18 Apr 2003 04:22:31 GMT
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl> <3E9EAA36.BFB1018D@engcorp.com> <IQmdnQg6xagDdQOjXTWcpA@august.net>
Message-ID: <b7nui7$2tk16$1@ID-148859.news.dfncis.de>

module "Terry Hancock" core dumped with the following output:
> 
> I know there's an X screensaver that does something like this to retrieve 
> random images.  I doubt they were too concerned about even distribution, 
> though.
I was actually going to suggest that X screensaver (webcollage).
It actually goes through some pains to get a random distribution from web-sites.
I forgot the exact heuristics used, but it basically discards any page that contains
content that might have lead to a bias in the search results.
The code's in perl and is documented enough to get a feel for it.


-- 
India and Pakistan may be the next countries to use nukes in war
but no one should forget who was first! "USA!, USA!, USA!"



From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Wed Apr 16 15:11:03 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Wed, 16 Apr 2003 21:11:03 +0200
Subject: Degree symbol (UTF-8 > ASCII)
In-Reply-To: <f7199550.0304160930.6cfeca65@posting.google.com>
References: <f7199550.0304160930.6cfeca65@posting.google.com>
Message-ID: <3e9daad2$0$49115$e4fe514c@news.xs4all.nl>

Peter Clark wrote:
> I'm working with a xml document which doesn't include an encoding, so
> it defaults to UTF-8. Of course, all of the text is ASCII, and likely
> to remain so. I would like to insert the degree symbol (chr(176)), but
> because this is outside the bounds (chr(128) is the limit), Python
> raises an XML error. What's the simplest way of getting including an
> unadorned degree symbol? Again, it's not necessary to preserve the
> UTF-8 encoding, but I'm not quite certain as to how to tell Python
> that the XML document is plain ASCII (or I guess in the case of the
> degree symbol, latin-1).
>     Thanks,
>     :Peter

The "degree" symbol is chr(176) in what character encoding?
Certainly not in UTF-8. Perhaps in windows cp-1252 or ISO-8859-1.

Are you creating the UTF-8 encoded XML document using Python?
Then try the following to add a "degree" character to your
output file:

To get the (unicode) character you want:
	degreeChar = u'\N{DEGREE SIGN}'

(this is unicode character 00b0, so you could also use:
	degreeChar = u'\u00b0'

To write it using UTF-8 encoding to a file object 'output':
	output.write(degreeChar.encode('UTF-8'))

(if you look what is written you'll see that the character
is encoded in two bytes when using UTF-8: 0xc2,0xb0)

--Irmen




From staschuk at telusplanet.net  Wed Apr  9 17:08:12 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 9 Apr 2003 15:08:12 -0600
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <20030408194700.A1973@tibia.amotlpaa.bogus>; from staschuk@telusplanet.net on Tue, Apr 08, 2003 at 07:47:00PM -0600
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org> <BAB8B22D.2022%dogen@mindspring.com> <20030408194700.A1973@tibia.amotlpaa.bogus>
Message-ID: <20030409150812.A1262@tibia.amotlpaa.bogus>

Quoth I:
  [...]
> My suggestion is that, if the word list is, say,
> 	abandon
> 	major
> 	wombat
> 	articulate
> 	orange
> 	royal
> then the program could start with 'abandon', notice that no word
> in the word list ends with 'a', and immediately, er, abandon
> trying to find palindromes starting with this word, and move on to
> finding palindromes starting with 'major'.  [...]

Having too much time on my hands today, I went ahead and
implemented this idea.  The results are encouraging; times in
seconds to find all n-word palindromic phrases using a list of
(allegedly) the thousand most common English words:
    n      original     pruning
    1        < 1            2
    2        140            3
    3     > 1000           15
    4          ?           45
    5          ?          500
(I gave up on running the original for n >= 3.)

I think this shows that, in this case, changing the algorithm is a
better first step than changing languages.

(Caveat: The two programs are strictly speaking not comparable.
For example, the original implementation requires all words in the
palindrome to be distinct, while mine does not.  But this
introduces a bias against my version, which searches a larger
space.)

I'm a bit reluctant to post my pruning search module, since it's
475 lines long.  I'll email it to interested posters on request.
(I'll probably submit it to Useless Python shortly too, but no
doubt it'll take a while to appear on the site.)

-- 
Steven Taschuk                 Go draw a war dog.
staschuk at telusplanet.net       Won't safe be fast now?




From gh at ghaering.de  Wed Apr 16 09:44:56 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 16 Apr 2003 15:44:56 +0200
Subject: variable visiblity
In-Reply-To: <20030416133500.GB2518@swordfish.capgemini.hu>
References: <20030416133500.GB2518@swordfish.capgemini.hu>
Message-ID: <3E9D5E58.4090706@ghaering.de>

Gabor Nagy wrote:
> Hi!
> 
> How can I access the "parent" function's local variable from a nested
> function?
> 
> example:
> def foo(par):
> 	def f(x): print x, par
> 	f(0)
> 
> SyntaxWarning: local name 'par' in 'foo' shadows use of 'par' as global in
> nested scope 'f'
> when run, prints x, then
> ...in f NameError: global name 'par' is not defined

You're using Python 2.1. Nested scopes were introduced with Python 2.1, 
but not enabled by default, so you'll need to put a

	from __future__ import nested_scopes

on the top of your module.

Or just upgrade to Python 2.2, where nested scopes are enabled by 
default :-)

-- Gerhard





From mis6 at pitt.edu  Tue Apr 29 17:46:08 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 29 Apr 2003 14:46:08 -0700
Subject: Problem subclassing tuple
References: <F7DB8D13DB61D511B6FF00B0D0F062330489D361@mail.hebmex.com> <mailman.1051634628.4864.python-list@python.org>
Message-ID: <2259b0e2.0304291346.64241ad@posting.google.com>

"John Wilson" <tug at wilson.co.uk> wrote in message news:<mailman.1051634628.4864.python-list at python.org>...
> <snip> I want an object which has all the characteristics of a tuple with
> additional methods.

What about this ?

class Holder(tuple):
    def additionalmethod(self):
        print "ciao ciao"

t=Holder((1,2))

print t #=> (1,2)

t.additionalmethod() #=> ciao ciao

If you want to avoid the parenthesis, here is the solution:

class Holder(tuple):
    def __new__(cls,*args):
        return tuple.__new__(cls,args)
    def additionalmethod(self):
        print "ciao ciao"

t=Holder(1,2)

print t # => (1,2)

t.additionalmethod() #=> ciao ciao

Easy, isn't it ?

                                        Michele



From ramen at lackingtalent.com  Tue Apr  1 15:04:41 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 20:04:41 -0000
Subject: Anything like Mailman's LockFile.py in std lib?
References: <20030401144641.L28887-100000@eden.ispol.com>
Message-ID: <slrnb8jpb1.5u1.ramen@lackingtalent.com>

In article <20030401144641.L28887-100000 at eden.ispol.com>,
Gregory (Grisha) Trubetskoy wrote:
> 
> Is there anything being planned for the standard lib that provides locking
> a la LockFile.py in Mailman?
> 
> I find it extremely useful, but just recently came across a project where
> it can't be used becuase it's GNU licensed - thus the question.
> 

Have you looked at portalocker? It's here at the ASPN Python Cookbook:
http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/65203

It doesn't support win95/98, and doesn't seem to have nearly as much
functionality as LockFile.py, but it might help you with the licensing
issue. I think that a library like LockFile should be part of the os module.

HTH,
Dave



From exarkun at intarweb.us  Wed Apr  2 12:51:56 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 12:51:56 -0500
Subject: distributed computing implementations
In-Reply-To: <b6f66v$28k$1@slb3.atl.mindspring.net>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <b6f66v$28k$1@slb3.atl.mindspring.net>
Message-ID: <20030402175156.GA7752@meson.dyndns.org>

On Wed, Apr 02, 2003 at 10:25:20AM -0700, Andrew Dalke wrote:
> robin:
> > If you want the biggest, boldest approach and don't care about
> > overhead, use CORBA.
> 
> What overhead would this be?  From what I see of omniORB,
> there isn't really that much.  Also, CORBA is the most complete,
> eg, it allows callbacks and passing around object references
> the others don't have.
> 

  I'm not sure about XML-RPC, SOAP, or Dopy, but I know Twisted Spread can
pass references around like this (and you can tell it just -how- you want
them passed), and I think Pyro can too.

> I've wanted to do something in CORBA for years.  I've
> never gotten there.  One problem is that I'm used to Python,
> where I don't need to describe the interface beforhand.
> CORBA wants that IDL, and a change in the object's interface
>  must be reflected in the IDL.  That just seems tedious to
> me now.
> 

  Right.  I think this is at least part of the overhead Robin was talking
about.  Several of the other schemes don't require this.

> I also do nearly everything in Python, so don't need the
> ability for different langauges to interoperate.  I just pass
> around Python objects.
> 
> I keep hoping that one of the component systems (like
> in GNOME or KDE) takes off, so that "scripting" a la
> COM takes off in unix systems, but I've been hoping
> for that for the last 5 years.

  That'd be nice.

> 
> > If you want a simpler approach but with performance and
> > implementation difficulties, use SOAP or XML-RPC.
> 
> I wouldn't put SOAP as simple, and I've had problems with
> interoperability between various packages.  If I needed to pass
> simple data around, I would use XML-RPC.

  Personally, I think SOAP is worth ignoring.  ;)  XML-RPC is what I would
choose if I were going for that sort of solution, too.

> 
> > If you want the ultimate in simplicity and are willing to foresake
> > multi-language support, use Dopy, Pyro, or Twisted Spread.
> 
> One reason I'm looking at Twisted is because it handles other
> interfaces as well.  I need to talk to SQL databases, SOAP and
> XML-RPC servers, straight HTTP, and spawned off external
> processes.

  This is definitely a benefit.  (Hooray, integration).

  From what you've said here and in your original post, I think Spread will
probably be a pretty good fit for you.

  HTH,

  Jp

-- 
"The problem is, of course, that not only is economics bankrupt but it has
 always been nothing more than politics in disguise ... economics is a form
 of brain damage."  -- Hazel Henderson
-- 
 up 13 days, 14:00, 7 users, load average: 0.15, 0.14, 0.07




From wweexxsseessssaa at telusplanet.net  Wed Apr  9 06:21:49 2003
From: wweexxsseessssaa at telusplanet.net (John Hall)
Date: Wed, 09 Apr 2003 10:21:49 GMT
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de>
Message-ID: <r7s79vghrcufpsostmv8avr7ncu2985ja7@4ax.com>

On Wed, 09 Apr 2003 09:55:00 GMT, skywalkerpackage at hotmail.com
(Johannes Eble) wrote:

>is there a library available for Python that behaves similar to MS
>Visio? 

The OpenOffice.org Draw module has a lot of functionality with which
you could perhaps emulate Visio. IIRC, there are some
reasonably-priced commercial Template kits available, tailored to
particular areas of interest (though, when I think about it, perhaps
that was not for OO0/Draw).

Controlling it from another program has until recently been difficult,
primarily because the documentation was incomplete and fragmented.
But OO recently released an SDK which address all or most of those
concerns.

I don't know if this will help with Python though. Someone has written
Py-UNO Bridge:
http://www.budischewski.com/pyuno/python-bridge.html
(UNO - Universal Network Objects is the OOo component model)

OpenOffice developer's guide:
http://api.openoffice.org/DevelopersGuide/DevelopersGuide.html

"""
Currently exist bindings for Java, C, C++ (compiler dependent, please
see http://porting.openoffice.org for a list of supported platforms),
OLE Automation and Python(in a alpha state). 
"""
-- 
John W Hall <wweexxsseessssaa at telusplanet.net>
Calgary, Alberta, Canada.
"Helping People Prosper in the Information Age"



From achrist at easystreet.com  Mon Apr 28 21:30:42 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Mon, 28 Apr 2003 18:30:42 -0700
Subject: Installer Problem WinNT to Win 98
Message-ID: <3EADD5C2.ABC48D5E@easystreet.com>

I built an installer for a wxPython/pysqlite app with the McMillan
installer (as downloaded last week), and carried it on a trip to a
friend's house over the weekend.  Because I wanted the app to look
polished, I made it with the noconsole option so that the UI would be a
gui with no messages or warnings showing up in a command window.
I tested this on my development machine here (Windows NT) and it worked
fine.  I tested it with the path set to nothing and my python directory
(which includes wxPython in site-packages, etc) renamed so that even
if there were some registry settings pointing to stuff installed on
my machine, the executable wouldn't be able to find them.  Even so, the
executable ran fine on my machine.  The executable crashed on friend's
Windows 98 machine.  Because of the noconsole option, I didn't get
any messages to figure out where.

My guess is that there is a needed DLL in my Windows or WinNT directory
that is not on the Win98 machine, but I've got no idea what it might
be or if it is possible to get it into the installer.

Any ideas what might be missing and/or how to fix?  The program
crashes in the middle of displaying a  wxHtmlCtrl that has some
widgets embedded.  It looks like it shows the html text fine, but
none of the widgets appear and it hangs right there.

Thanks for any suggestions.


Al



From tismer at tismer.com  Sun Apr 13 05:45:41 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sun, 13 Apr 2003 11:45:41 +0200
Subject: Stackless python
In-Reply-To: <221d8dbe.0304120516.4d543fb@posting.google.com>
References: <221d8dbe.0304050531.107bc603@posting.google.com> <mailman.1049558408.987.python-list@python.org> <221d8dbe.0304120516.4d543fb@posting.google.com>
Message-ID: <3E9931C5.3040200@tismer.com>

srijit at yahoo.com wrote:
...

>>This is temporary, the new version which I'm about to upload
>>tomorrow does not have that external file, since it is all moved into
>>the core dll, for simplicity.
> 
> You mention about a new version to be uploaded on 06th April, 2003.
> But I do not see any new version in www.stackless.com

Sorry, I forgot to do the upload, and meanwhile there were
many changes, again.

> I am not comfortable with CVS. At least can there be a small tutorial
> and some examples as HTTP/FTP download?
> 
> Without good documentation and examples it is very difficult to take
> off on a new topic like Stackless Python. One of the most important
> reasons for Python being so popular is excellent documentation with
> every release.

Please be patient. At the moment I'm under some pressure
of my clients to produce a stable new version. After
that, I can work on documentation and examples.
For now, I cannot produce more than snapshots.

> After going through StacklessEuroPy.ppt, I feel it is worth learning
> to use Stackless Python. But (I repeat!) I am afraid without good
> documentation my attempt cannot last long.

Sorry, I can't help it right now. Will upload another snapshot,
soon, but maybe you should come back when 3.0 is out of
alpha state.

cheers - chris

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From tylere at hotpop.com  Sun Apr 13 11:06:32 2003
From: tylere at hotpop.com (Tyler Eaves)
Date: Sun, 13 Apr 2003 15:06:32 GMT
Subject: Tutorial back online!
Message-ID: <Y1fma.181$8c4.175@tornadotest1.news.pas.earthlink.net>

After several months in limbo, my python tutorial / book-in-progress is back
on line! 

You can see it by going to http://cg1.org . It's on the software page.

(Note: Cg1.org is currently under HEAVY development, so there are liekly to
be lots of blank pages ;)





From jeremy at jdyallop.freeserve.co.uk  Tue Apr 22 04:55:14 2003
From: jeremy at jdyallop.freeserve.co.uk (Jeremy Yallop)
Date: 22 Apr 2003 08:55:14 GMT
Subject: problem starting python interpreter in emacs (win98)
References: <V65pa.4289$AB5.573029@news02.tsnz.net>
Message-ID: <b8301i$5nl2c$1@ID-114079.news.dfncis.de>

selwyn wrote:
> I am having diffculty getting the python interpreter to run under emacs,
> either from the menu or with C-c ! Syntax highlighting and .py file 
> recognition works fine.
> 
> When I try strating the interpreter with the above Emacs complains:
> "Searching for program: No such  file or directory, python"

Perhaps you need to add C:\Python22 to the PATH environment variable.
Alternatively, you could tell emacs where the Python interpreter lives
by setting py-python-command to the full path to the executable
("C:\Python22\python.exe").

> However I am slightly confused about when Emacs reloads .emacs. Do I
> have to restart windows everytime I make a change, or is there a method
> for reloading while Emacs is running?

M-x eval-buffer.  Also try M-x ielm for an interactive elisp shell.

Jeremy.



From sbaptista at yahoo.com  Mon Apr 21 18:20:56 2003
From: sbaptista at yahoo.com (Stan Baptista)
Date: Mon, 21 Apr 2003 15:20:56 -0700 (PDT)
Subject: IDLE question
Message-ID: <20030421222056.5204.qmail@web11602.mail.yahoo.com>

Hi all,

My apologies if this is not the correct forum for IDLE
questions.

I'm running IDLE on Win2K.  Is it possible to change
the default subdir on a file open (menu
File->Open...)?

It currently points to the install subdir for Python. 
However, I'm working in another subdir which requires
that I change to it every time I open another file.

Better yet, can IDLE be configured to point to the
last-opened subdir?

Thanks,
Stan


__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com




From exarkun at intarweb.us  Sun Apr  6 18:41:11 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 6 Apr 2003 18:41:11 -0400
Subject: Pythonic way of web-programming
In-Reply-To: <jQ_ja.12641$7w2.9369@nwrddc01.gnilink.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <e9570f37.0304060944.79564f88@posting.google.com> <jQ_ja.12641$7w2.9369@nwrddc01.gnilink.net>
Message-ID: <20030406224111.GA27038@meson.dyndns.org>

On Sun, Apr 06, 2003 at 07:01:35PM +0000, Neil Schemenauer wrote:
> Graham Fawcett <graham__fawcett at hotmail.com> wrote:
> > [...] what kinds of applications?  We are talking only about
> > Web application frameworks, which are request handlers and are
> > therefore event-driven by definition.
> 
> For example, we have a web application that interacts with
> FedEx's servers in order to create new shipping labels, track
> packages, etc.  With an event driven system, waiting for a
> response for FedEx would freeze the entire server.  Obviously
> that can be fixed the hooking into the event loop.  It's just a
> matter of programming, right?
> 

  Hmm, you point this out as a problem, then say it is easily fixable.  I'm
not sure what point you're trying to make here.


> Another example: we have an application that does some CPU
> intensive operations.  We have multiple CPUs in our server.  An
> event driven system would not (normally) take full use of the
> processors.  Also, clients who have IO bound requests would be
> blocked by one CPU bound request.

  Threads and multiple processes solve this easily.  So do event-based IO
APIs, such as kqueue and AIO (Yes, I know AIO isn't ready for use in a
production system yet).

> 
> I have nothing against the event driven model.  However, it is
> not the solution to every network server problem.

  No doubt.  "Canonical" doesn't mean "universal" or "solitary".

  Jp

-- 
No, `Eureka' is Greek for `This bath is too hot.'
                -- Dr. Who
-- 
 up 17 days, 19:01, 5 users, load average: 0.00, 0.20, 0.23




From staschuk at telusplanet.net  Sat Apr 26 14:30:39 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 26 Apr 2003 12:30:39 -0600
Subject: serveur TFTP
In-Reply-To: <timo-BF381D.13191926042003@news-central.giganews.com>; from timo@alum.mit.edu.invalid on Sat, Apr 26, 2003 at 01:19:20PM -0400
References: <e4d8f90f.0304240622.64cf7f8e@posting.google.com> <timo-BF381D.13191926042003@news-central.giganews.com>
Message-ID: <20030426123039.C30093@tibia.amotlpaa.bogus>

Quoth Timothy O'Malley:
> In article <e4d8f90f.0304240622.64cf7f8e at posting.google.com>,
>  naoaqualis at aol.com (alex) wrote:
> > j'ai recup?r? une librairie TFTP ecrit en python mais j'avoue avoir un
  [...]
> The TFTP code you cite appears to be the module I posted
> to comp.lang.python a few years ago.  I'd be willing to
> answer the question ... if I could understand it.  My
> Python is much better than my French.  (Which isn't to
> say my Python is good, rather that my French is bad. :-)

According to my almost but not quite complete ignorance of French:

    I have [received?] a TFTP library written in Python but
    [I have had some trouble?] using it.  How [can I] make a
    TFTP server with [it]?  I tried to identify the methods
    which I could apply to an object my_server, for example:

        import tftplib
        my_server = tftplib.TFTPServer()

    ...and after [that] I am a little lost.

    Here's the library tftplib.py:
    [...]

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From drs at ecp.cc  Tue Apr 29 22:12:44 2003
From: drs at ecp.cc (drs)
Date: Wed, 30 Apr 2003 02:12:44 GMT
Subject: set cookie problem (slightly OT)
Message-ID: <wiGra.5646$%_3.3813575@newssrv26.news.prodigy.com>

I have a web directory (apache2/FreeBSD) with ssi enabled but not cgi.  in
order to run simple python scripts i use something like

<!--#exec cmd="/usr/local/bin/python /path/to/script.py" -->

which works surprisingly well.  I cannot figure a way to set a cookie,
however, as i am not able to add to the header. Anyone have any suggestions?

So long as python scripts that are run in this way can read cookies (which
might be the next question), i am happy to set the cookie some other way.

thanks for suggestions -d





From jkraska at san.rr.com  Wed Apr 16 11:35:17 2003
From: jkraska at san.rr.com (Courageous)
Date: Wed, 16 Apr 2003 15:35:17 GMT
Subject: Big-O notation
References: <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <roy-92A049.08504216042003@reader1.panix.com> <Xns935F94F66A99Eduncanrcpcouk@127.0.0.1>
Message-ID: <1utq9v4mcjkhicpicdq1pbba5rc9mu2ads@4ax.com>

>I believe you are making some unreasonable assumptions here. Remember that 
>if I have an algorithm that is O(N^2) that is really just a shorthand for 
>saying that it will have a running time a*N^2 + b*N * c where a, b, and c 
>are constants but for sufficiently large N only the N^2 term matters.
>Now imagine that you have a program that runs in O(N^2) time and you have 
>10 items. If a is 1 and b is 1000, ...

Ah, the practical optimizer. We encounter so few of these in this day
and age. :-)

C//




From mertz at gnosis.cx  Tue Apr 15 16:55:33 2003
From: mertz at gnosis.cx (David Mertz)
Date: Tue, 15 Apr 2003 16:55:33 -0400
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
 <v9oq71itajkh39@news.supernews.com> <slrnb9orhr.toh.ramen@lackingtalent.com>
Message-ID: <FHHn+kKkXwvG092yn@gnosis.cx>                       

|> However, losing apply() isn't a big deal because there is
|> an exactly equivalent construct.

Not quite.  For Appendix A of my book, I discuss one (realistic) example
where extended call syntax and apply() are not directly substitutable:

  In most cases, the extended call syntax is more readable, since
  the call closely resembles the -declaration- syntax of generic
  positional and keyword arguments.  But in a few
  cases--particularly in higher-order functions--the older
  `apply()` built-in function is still useful.  For example,
  suppose that you have an application that will either perform
  an action immediately or defer it for later, depending on some
  condition.  You might program this application as:

      #*----------- apply() as first-class function -----------#
      defer_list = []
      if some_runtime_condition():
          doIt = apply
      else:
          doIt = lambda *x: defer_list.append(x)
      #...do stuff like read records and options...
      doIt(operation, args, keywds)
      #...do more stuff...
      #...carry out deferred actions...
      map(lambda (f,args,kw): f(*args,**kw), defer_list)

  Since `apply()` is itself a first-class function rather than a
  syntactic form, you can pass it around--or in the example,
  bind it to a name.

Of course, you could define your own apply() in one line of code if the
function were not built-in.

|..."itertools" module....
|I think it'd be good to have a module called "functional" where the
|more unusual functional tools can live

I definitely agree with this idea.  A [functional] module in the
standard library would be GREAT.  Xoltar had a number of useful ideas,
although other people have proposed different implementations of some of
them... so the best version would need to be decided.

But another collection of functions that I would REALLY like to see in
such a hypothetical standard library would be a good set of Higher Order
Functions (HOFs).  My Gnosis Utilities contains gnosis.utils.combinators,
which has at least the ideas for the most useful few.  The
implementations are overly concise and under documented though.  But
it's short enough to list here:

  from operator import mul, add, truth
  apply_each = lambda fns, args=[]: map(apply, fns, [args]*len(fns))
  bools = lambda lst: map(truth, lst)
  bool_each = lambda fns, args=[]: bools(apply_each(fns, args))
  conjoin = lambda fns, args=[]: reduce(mul, bool_each(fns, args))
  all = lambda fns: lambda arg, fns=fns: conjoin(fns, (arg,))
  both = lambda f,g: all((f,g))
  all3 = lambda f,g,h: all((f,g,h))
  and_ = lambda f,g: lambda x, f=f, g=g: f(x) and g(x)
  disjoin = lambda fns, args=[]: reduce(add, bool_each(fns, args))
  some = lambda fns: lambda arg, fns=fns: disjoin(fns, (arg,))
  either = lambda f,g: some((f,g))
  anyof3 = lambda f,g,h: some((f,g,h))
  or_ = lambda f,g: lambda x, f=f, g=g: f(x) or g(x)
  compose = lambda f,g: lambda x, f=f, g=g: f(g(x))
  compose3 = lambda f,g,h: lambda x, f=f, g=g, h=h: f(g(h(x)))
  ident = lambda x: x
  not_ = lambda f: lambda x, f=f: not f(x)

  def shortcut_all(*fns):
      accum = fns[0]
      for fn in fns[1:]:
          accum = and_(accum, fn)
      return accum
  lazy_all = shortcut_all

  def shortcut_any(*fns):
      accum = fns[0]
      for fn in fns[1:]:
          accum = or_(accum, fn)
      return accum
  lazy_any = shortcut_any


Yours, David...

--
Keeping medicines from the bloodstreams of the sick; food from the bellies
of the hungry; books from the hands of the uneducated; technology from the
underdeveloped; and putting advocates of freedom in prisons.  Intellectual
property is to the 21st century what the slave trade was to the 16th.




From LogiplexSoftware at earthlink.net  Thu Apr 10 13:40:37 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 10 Apr 2003 10:40:37 -0700
Subject: Python compared toJava
In-Reply-To: <20030410172428.GA4764@isis.gerrietts.net>
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
	 <20030410172428.GA4764@isis.gerrietts.net>
Message-ID: <1049996437.13013.57.camel@software1.logiplex.internal>

On Thu, 2003-04-10 at 10:24, Geoff Gerrietts wrote:
> Quoting Blaktyger (blaktyger at hotmail.com):
> > Hi!
> > 
> > I recently read that the Python interpreter is like the Java Virtual Machine.
> > Is that true?
> 
> It's hard to tell exactly what you mean by this, but since you seem to
> have received an unusually high proportion of flip or catty responses,
> I'll try to reply.

[snip]

> Answering your question really requires a better understanding of what
> you want to know: in what way have you heard that these two are alike?
> You should probably interpret the rash of snide replies to mean just
> that: we don't really know what you're asking, but we'd answer you if
> we did.

Geoff is right.  Sorry.  My comment was meant as humorous.  Please don't
take it otherwise.

Regards,

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From kseehof at neuralintegrator.com  Mon Apr 14 17:14:57 2003
From: kseehof at neuralintegrator.com (Ken Seehof)
Date: Mon, 14 Apr 2003 14:14:57 -0700
Subject: Variable inside module 
In-Reply-To: <000a01c302c1$aaa464f0$0d00000a@esss.com.br>
Message-ID: <5.2.0.9.0.20030414132932.00af3640@pop.hurrah.com>

At 01:08 PM 4/14/2003 Monday, Marcelo A. Camelo wrote:

>Hi!
>
>Can someone please explain-me what is
>happening here:
>
>Given the followin bar.py file (placed
>inside a foo/ directory):
>
>         var = 'foo'

Adds 'var':'foo' into the module bar namespace

>         def SetVar():
>             global var
>             var = 'bar'
>
>         def PrintVar():
>             global var
>             print 'var:', var
>
>In the same foo/ directory, I've placed
>the following __init__.py:
>
>         from bar import *

Copies 'var':'foo' from module bar into the foo namespace.

>Then, in the interactive python interpreter:
>
>         >>> import foo
>         >>> print foo.var
>         foo
>         >>> foo.PrintVar()
>         var: foo
>         >>> foo.SetVar()

Sets 'var':'bar' in the global namespace.

>         >>> foo.PrintVar()
>         var: bar

Prints 'var':'bar' in the global namespace.

>         >>> print foo.var
>         foo

Prints 'var':'foo' in the foo namespace.

>         >>>
>
>What I don't understand is that foo.var seems
>to be two different variables: one accessed
>inside the bar.py file and one accessed from
>within the python interpreter. What is going on
>here?
>Thank you,
>
>Marcelo A. Camelo

Not exactly.  It's just that the foo module is not
the same as the global namespace.

The statement "var = 'foo'" is executed in the
bar module, then the value is copied into the foo
module.

- Ken Seehof
www.neuralintegrator.com
kseehof @ neuralintegrator.com






From tjreedy at udel.edu  Wed Apr 16 13:48:14 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Wed, 16 Apr 2003 13:48:14 -0400
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com>
Message-ID: <RoacnSAeGeS6CACjXTWcoQ@comcast.com>

"Tj" <tj_scarlet at yahoo.com> wrote in message
> fold-left) that explain nothing.  Lambda could also stand renaming;
> Pythonistas probably think it's an arrogant name.  (Or exotic?  I
> can't tell.)

For people without knowledge of Lisp, and maybe lacking complete
knowledge of the Greek 'alpha-beta' and math people's habit of using
such letters, try 'obscure'.  To me (who had such knowledge) it is
just a bit jarring.  All other Python keywords are English words (or
abbreviations thereof) with Python meanings similar to their everyday
meanings.  'Lambda' sticks out like a sore thumb.  'Func' would have
been better.

Terry J. Reedy





From amk at amk.ca  Wed Apr 16 08:06:37 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Wed, 16 Apr 2003 07:06:37 -0500
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <mailman.1049653811.11491.python-list@python.org> <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net> <mailman.1050470255.12198.python-list@python.org>
Message-ID: <7kCdnQb9MqHQ2gCjXTWcow@speakeasy.net>

On 16 Apr 2003 00:16:49 -0500, 
	Ian Bicking <ianb at colorstudy.com> wrote:
> One option would be to use an interface with adapters, where if say the
> Quixote interface was used, Webware would have an adapter from that
> interface to the Webware HTTP{Request,Response} interface... though I'm
> not sure how interesting request and response object are.  

Perhaps they're not very interesting, but that seems the only level that
could be feasibly standardized.

I'm willing to host a mailing list for web discussion on my DSL-connected
machine.  That might be easier than jumping through the SIG hoops.  If
you're interested, just suggest a name ('web'?) for the list and I'll create
it.  I can't do CVS, but maybe we could just use a sandbox/ directory in the
Python CVS tree.

> I disagree.  Unifying with one model for building applications is not
> likely, but there are a lot of places where the frameworks duplicate
> effort without significant differentiation.  By identifying those we
> could make a more robust, more easily distributed, more compelling
> foundation for web frameworks.  

Yes, but usually those areas of duplication don't actually require a lot of
maintenance effort after the initial implementation.  We're four months into
2003, and Quixote's HTTPRequest and Response haven't had a significant
change yet.  The changes have all been to higher-level things like the
publishing logic.

--amk                                                             (www.amk.ca)
OTHELLO: Put out the light, and then put out the light.
      -- _Othello_, V, ii



From peter at engcorp.com  Tue Apr 29 13:41:37 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 29 Apr 2003 13:41:37 -0400
Subject: how to parse numeric data files
References: <20030429132519.4e92271e.gry@ll.mit.edu>
Message-ID: <3EAEB951.5DC25694@engcorp.com>

george young wrote:
> 
> We have several electronic testing machines(of various ages and
> manufacturers) that spew out testing data files in various ascii formats.
> Currently we have a nasty mess of awk/shell/C/fortran programs that
> extract and process some data from these files.  I have a dream of
> a suite of simple, clear, maintainable python programs to do these tasks.
> 
> The trick is I hope to come up with something that our hardware
> engineers can understand and maintain easily without studying
> things like BNF, LALR etc. (they won't).
> 
> [Below is a sample of one of the worst formats, shortened from a 40MB file!]

The format is quite amenable to parsing with re, if rather large...
the real question is how much of that data do you need, and what do
you need to do with it?  What do your current scripts actually do?
Also, how much of the content you showed is *fixed* format, and how
much of the format can vary?  Is anything optional?

If you want the hardware engineers to be able to maintain it, you might
want to support a kind of "template" specification, where you provide the
names of various tags which are recognized (e.g. "PH_lot_id:") and you
automatically extract the appropriate value found thereater.

What you're trying to do is not really that complex, I think, so I
do think you should be able to find a good simple solution for it.
As you said, this _is_ Python...

-Peter



From roy at panix.com  Tue Apr 22 15:07:15 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 22 Apr 2003 15:07:15 -0400
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net> <nCUoa.41676$LB6.941312@news1.tin.it> <pan.2003.04.22.03.08.58.526973@mn.rr.com> <3japa.44301$LB6.997057@news1.tin.it> <roy-6ECAE1.08483422042003@reader1.panix.com> <b840q9$ggs$1@slb5.atl.mindspring.net>
Message-ID: <roy-CE134F.15071522042003@reader1.panix.com>

In article <b840q9$ggs$1 at slb5.atl.mindspring.net>,
 "Andrew Dalke" <adalke at mindspring.com> wrote:

> Roy Smith:
> > I'm with Guido j^2 on this one.
> 
> Don't you mean -j**2?

Sorry, yes.



From aahz at pythoncraft.com  Tue Apr 15 07:53:48 2003
From: aahz at pythoncraft.com (Aahz)
Date: 15 Apr 2003 07:53:48 -0400
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com>
Message-ID: <b7grsc$63p$1@panix1.panix.com>

In article <v9e21mhemm8c2 at corp.supernews.com>,
Paul Watson <pwatson at redlinec.com> wrote:
>
>We need to teach students correct design priciples to build something
>greater than what already exists.  We will never get very far if we
>require everyone to start with a quark or atom.  Yes, of course we need
>some people who design silicon and create microcode.  They will learn
>the low-level details what they need to know as they need it.  Knowing
>great design and organization principles will enable them to make the
>most of it.

While there's some truth to that, try explaining why the following code
is a Bad Idea to someone who has no basic understanding of CS:

    s = ''
    for i in range(1000000):
        s += str(i)

Knowing the difference between O(N) and O(N^2) is critical to writing
even the simplest programs.  OTOH, I do agree with you that focusing on
programming as a craft is more important to being a programmer than
learning CS.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From mcfletch at rogers.com  Fri Apr 11 13:16:18 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Fri, 11 Apr 2003 13:16:18 -0400
Subject: Confusion: assignments create object references.
In-Reply-To: <0C642BE4EE085149B52839125D4FC9D422BDA0@idc1exc0002.corp.global.level3.com>
References: <0C642BE4EE085149B52839125D4FC9D422BDA0@idc1exc0002.corp.global.level3.com>
Message-ID: <3E96F862.6080004@rogers.com>

Havens, Peter wrote:

>// -----Original Message-----
>  
>
...

>//   The above never happened, I assure you. >:)    Python 
>// doesn't work this way.
>
>I'm a silly rabbit. I hand copied the test incorrectly. The real test was:
>
>  
>
>>>>L = 1
>>>>M = L
>>>>L
>>>>        
>>>>
>1
>  
>
>>>>M = 0
>>>>L
>>>>        
>>>>
>1
>  
>
>
>Those are the real results (I double checked), but now I'm really confused. Why does the former test (the one I lied about) above give expected results, and the later doesn't?
>
>  
>
Because Python isn't C ;) .  Seriously, what you want to think about is 
namespaces and names.  A name is a simple mapping from a string to a 
(smart (i.e. reference-counted)) PyObject * (under the covers).  That 
is, a name holds a reference (pointer) to an object.  That's all it ever 
holds a reference to.  It never holds a reference to another name  
"slot"(though it can hold the same string value as a name, but that's 
still just a string object).

When you do:

    M = L

what's happening is that the name M in the current namespace is given a 
pointer to the object currently pointed to by L, so if L is pointing to 
a list, M now points to that list.  M knows nothing about the name L, it 
has *its own* pointer to the list, and there's no dependence on L at 
all.  Complementarily, L knows nothing about M, so assigning a new 
object to L won't affect M either.

Your sequence of actions:

    L = 1 # L now points to the integer object 1
    M = L # M now points to the integer object 1
    M = 0 # M now points to the integer object 0

never changed the object to which L pointed, so L still points to the 
original object.

HTH,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From skip at pobox.com  Wed Apr  2 18:59:52 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 2 Apr 2003 17:59:52 -0600
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
In-Reply-To: <nFKia.141$Sd.25@fed1read01>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
        <b68jti$2mj66$1@ID-169208.news.dfncis.de>
        <RbCia.6782$9I.555309@news20.bellglobal.com>
        <mailman.1049298845.7655.python-list@python.org>
        <3E8B4094.F61515B3@engcorp.com>
        <ZpHia.23$Sd.4@fed1read01>
        <3E8B48E0.C786D192@engcorp.com>
        <mailman.1049316186.31356.python-list@python.org>
        <3E8B55B2.D8D7D2F2@engcorp.com>
        <mailman.1049321730.7572.python-list@python.org>
        <nFKia.141$Sd.25@fed1read01>
Message-ID: <16011.31096.680721.1124@montanaro.dyndns.org>

    Tim> I think that nearly this exact use case was discussed here:
    Tim> http://mail.python.org/pipermail/python-dev/2003-March/034272.html
    Tim> . If you go forward a few messages, GVR suggests adding "from
    Tim> __builtins__ import open" to modules that you might want to fiddle
    Tim> with.

Yeah, but that's quite different from Peter's use case, where he wants to
unit test a module more-or-less from the outside.  He shouldn't have to add
hooks in the module itself to unit test it.

    Tim> This made me recall that the proposal was slightly more lenient
    Tim> than I initially remembered:

    Tim> some_random_module.open = my_open

    Tim> would be OK if open is already shadowed in some_random_module. The
    Tim> compiler then can figure out that optimizing access to open is not
    Tim> going to work.

That won't help Peter either.  I suspect all of the modules he wants to test
don't actually shadow open().

Skip




From duncan at NOSPAMrcp.co.uk  Tue Apr 15 11:13:55 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 15 Apr 2003 15:13:55 +0000 (UTC)
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1> <3e9c1413$1@news.012.net.il> <jJUma.23762$T34.707989@news2.tin.it>
Message-ID: <Xns935EA3DC45768duncanrcpcouk@127.0.0.1>

Alex Martelli <aleax at aleax.it> wrote in 
news:jJUma.23762$T34.707989 at news2.tin.it:

> import re
listofnames = [re.escape(n) for n in listofnames]
> namesre = re.compile('(%s)'%')|('.join(listofnames), re.I)
> 
> for m in namesre.finditer(astring):
>     dosomething(astring, listofnames[m.lastindex-1])
> 
> 

I suggest adding the line above to avoid any nasty surprises (or use a 
different variable if you don't want to clobber listofnames).


-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From cpl.19.ghum at spamgourmet.com  Tue Apr 29 06:53:08 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Tue, 29 Apr 2003 12:53:08 +0200
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org>
Message-ID: <Xns936C82C71A292cpl19ghumspamgourmet@62.153.159.134>

Roman,

just one addition:

>     Python could work with COM.

Python is the best COM-Client development tool I met. Together with PyCrust 
(Part of wxPython) developing COM-Client Application becomes a charme. 
Pythons interacte shell (duplicated in PyCrust) allows you to test loops, 
exceptions, Lists (=Arrays) communicating with COM-Servers like Excel or 
Word.

Pythons COMability really SAVED MY ASS in a big COM-project.

Harald



From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr  3 13:43:36 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 3 Apr 2003 18:43:36 GMT
Subject: windows-cmd shell doesn't reflect _winreg changes
References: <b6e56f7a.0304021609.521700e6@posting.google.com>
Message-ID: <slrnb8p06l.174.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <b6e56f7a.0304021609.521700e6 at posting.google.com>, haughki wrote:
> hi
> 
> i'm using _winreg to 'hard-change' and environmental variable in
> windows.  the change is 'successful': regedit correctly reflects the
> changes, and crtl panel > system > properties envronmental vars have
> been correctly changed.  but, when i start a cmd shell (dos), the
> shell evironment still reflects the old values until i manually open
> the env vars cntrl panel and click 'ok'.  anyone?

This is really a Win32 programming issue rather than a Python one.
However, I do vaguely remember this one.

Environmental variables are normally passed to a new process by the
process that creates it, and they can never be updated from outside
a process.  The exception to this is that Windows Explorer (and any
other program that wishes to) will read those initial environment
variables out of the registry when notified of a change in settings.
The changes will then be passed on to any program started by Explorer
afterwards.

Here's the Python code:

    import win32con, win32gui
    win32gui.SendMessage(
        win32con.HWND_BROADCAST, win32con.WM_SETTINGCHANGE, 0, 0)

You might want to use SendMessageTimeout instead of SendMessage.



From eppstein at ics.uci.edu  Tue Apr 15 22:10:40 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Tue, 15 Apr 2003 19:10:40 -0700
Subject: Big-O notation (was: Re: Python and Schools)
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <3E9C6C62.8010808@rogers.com> <mailman.1050441947.6447.python-list@python.org> <b7id3q$3s4$1@news.hccnet.nl>
Message-ID: <eppstein-3018A9.19104015042003@news.service.uci.edu>

In article <b7id3q$3s4$1 at news.hccnet.nl>,
 anton at vredegoor.doge.nl (Anton Vredegoor) wrote:

> >Also, one further class that is import to know is O( 2**N ), such as the
> >brute-force solution to the "travelling salesman problem".  These are
> 
> Although it's important to not set too severe limitations I have the
> impression this salesman is underestimating the problem.
> 
> >>> def fac(n):
> ...     return reduce(operator.mul,range(2,n+1),1)
> ... 
> >>> fac(10)
> 3628800
> >>> 2**10
> 1024

Actually, it's possible to solve the traveling salesman problem in time 
O(n 2^n), much better than factorial.  But the solution involves dynamic 
programming and is probably not what Chad N. meant by "brute-force".

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From rnd at onego.ru  Fri Apr  4 08:09:41 2003
From: rnd at onego.ru (Roman Suzi)
Date: Fri, 4 Apr 2003 17:09:41 +0400 (MSD)
Subject: server cluster support in Python?
In-Reply-To: <a3b5b608.0304040454.736b680b@posting.google.com>
Message-ID: <Pine.LNX.4.44.0304041709160.2287-100000@suzi.com.onego.ru>

On 4 Apr 2003, Giorgi wrote:

> Does Python have the one?

Probably Amoeba version of Python does ;-)
 
> GRTZ,
> Giorgi
> 

Sincerely yours, Roman A.Suzi
-- 
 - Petrozavodsk - Karelia - Russia - mailto:rnd at onego.ru -
 





From km-news1 at onlinehome.de  Sat Apr 26 15:18:37 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Sat, 26 Apr 2003 21:18:37 +0200
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <mailman.1051376980.15852.python-list@python.org>
Message-ID: <oprn83xbstmu3s0q@news.cis.dfn.de>

> Hm.  Yes and no.  In many cases such a mistake will make the code
> fail dramatically with an informative exception.

Yeah, fine -  we have no problem to notice this bug! :-)

My example shout only show, that we have a mix of values and functions, 
returning values.
An example of an unnoticed, critical error-case gave the OP.

> and just makes your program behave strangely.  I don't think
> there's any good technological solution here (see below); my
> feeling is that one just needs good troubleshooting skills.

You are right. I had made this error maybe 2 or 3 times and found the bug 
fast and without problems.
But because it's a simple "typing error" or so, i am not save to make this 
error again... :-(

-- 
Gru? - regards Klaus :-)



From me at somewhere.org  Sun Apr 20 13:41:15 2003
From: me at somewhere.org (made some)
Date: Sun, 20 Apr 2003 12:41:15 -0500
Subject: !! PAID $5 made $356 sofar, resell e-book, give opinions...check
 out: http://fly.to/stoneh
Message-ID: <3ea35e7f_29@News.100ProofNews.com>

100+ e-books to resell
Get paid for your opinion database
cost $5
I already made $356 giving my opinion,it takes some effort to sign up
but it's worth it.
got to http://fly.to/stoneh and sign up.
No envelopes or stamps required.





-x-   100 Proof News - http://www.100ProofNews.com
-x-   3,500+ Binary NewsGroups, and over 90,000 other groups
-x-   Access to over 700 Gigs/Day - $8.95/Month
-x-   UNLIMITED DOWNLOAD




From tjreedy at udel.edu  Fri Apr 11 13:23:34 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Fri, 11 Apr 2003 13:23:34 -0400
Subject: Surprise: del weakref.WeakKeyDictionary[ item ] raises RuntimeError
References: <mailman.1049836994.31575.python-list@python.org> <b701jn$9ls3d$1@ID-169208.news.dfncis.de> <mailman.1050074899.23901.python-list@python.org>
Message-ID: <2-mdna5QK4MZZQujXTWcoQ@comcast.com>

"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
news:mailman.1050074899.23901.python-list at python.org...
> Greg Ewing (using news.cis.dfn.de) wrote:
>
> > Mike C. Fletcher wrote:
> >
> >>        for ref in self.data.iterkeys():
> >>            o = ref()
> >>            if o == key:
> >>                del self.data[ref]
> >>                return
> >>
> >> Using keys seems to avoid the problem, but with the obvious
> >> disadvantages of keys vs. iterkeys.
> >
> >
> > Maybe a separate list of keys could be maintained
> > alongside the dict?
>
>
> It would have a pretty significant performance penalty if you always
> maintained a list.  Although I guess that doing two dictionaries:
>
>     ref : object
>     id : ref
>
> And then doing all the logic to make sure they keep in synch would
give
> you acceptable performance (and much faster than the current version
for
> del (at the expense of lots of memory used and slower performance
> elsewhere)).
>
> An "atomic" threading mechanism applied to the iterator-based del
method
> might work just as well, I suppose.
>
> In the meantime, using keys() *is* actually "maintaining" the
seperate
> key-list for as long as it's needed.
>
> Oh well, has to be some downside to iterators or we wouldn't have
any
> fun :) ,
> Mike
>
> _______________________________________
>   Mike C. Fletcher
>   Designer, VR Plumber, Coder
>   http://members.rogers.com/mcfletch/
>
>
>
>





From markscottwright at hotmail.com  Mon Apr  7 05:02:18 2003
From: markscottwright at hotmail.com (msw)
Date: Mon, 07 Apr 2003 09:02:18 GMT
Subject: How to protect python code ?
In-Reply-To: <mailman.1049389032.4904.python-list@python.org>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org>
Message-ID: <u8bka.651$001.591065@monger.newsread.com>

Jp Calderone wrote:
> On Thu, Apr 03, 2003 at 09:35:49AM +0200, christophe grimault wrote:
> 
>>Hello,
>>
>>Before trying my own solutions, I wonder if someone already has a solution
>>to this. We have been programming a quite large application in python and
>>it contains many original algorithm and other things. The application will
>>be sent to customers in a few month and since it is a commercial app, we
>>need some protection against copy (illegal copies) and we need to make
>>sure that critical parts of the code cannot be viewed and studied.
>>

I'm really interested in this as well, and I think that it's an 
important issue for python.  Unfortunately, as I'm sure you've 
discovered by now, it's a question that comp.lang.python is 
constitionally incapable of answering.  The only answers you will get in 
this group are:

1. intellectual property law is sufficient
2. no protection scheme is perfect

People who answer #1 don't sell their product in China.  People who 
answer #2 don't understand how security works - a security scheme 
doesn't have to be perfect, it just has to be more expensive to break it 
than to purchase your product.  For most of us, distributing code in 
binary format is sufficient, even if people can theoretically de-compile 
it.  Distibuting code in .py format is not.

I think it ought to be relatively easy to distribute a hacked version of 
the python interpreter with your application that substitutes the 
standard "import" code with code that first looks for .pys files (or, 
what the hell, .tcl - let's not make this any easier) that are 
des-encrypted .py files.  Compilation to .pyo would have to be disabled, 
which would increase start-up time.  The hacked interpreter (it's pretty 
easy to embed the python interpreter in your application), could also 
check some sort of hardware token before starting.  I have no idea if 
this would violate python's license.

But unfortunately, if anyone's done this, you'll never find out about it 
on c.l.p.  :-)

Mark




From timo at alum.mit.edu.invalid  Sat Apr 26 13:19:20 2003
From: timo at alum.mit.edu.invalid (Timothy O'Malley)
Date: Sat, 26 Apr 2003 13:19:20 -0400
Subject: serveur TFTP
References: <e4d8f90f.0304240622.64cf7f8e@posting.google.com>
Message-ID: <timo-BF381D.13191926042003@news-central.giganews.com>

In article <e4d8f90f.0304240622.64cf7f8e at posting.google.com>,
 naoaqualis at aol.com (alex) wrote:

> j'ai recup?r? une librairie TFTP ecrit en python mais j'avoue avoir un
> peu de mal ? l'utiliser.comment faire un serveur TFTP avec? j'essai
> d'identifier les methodes que je pourrais appliquer a un objet
> mon_serveur par exemple :


The TFTP code you cite appears to be the module I posted
to comp.lang.python a few years ago.  I'd be willing to
answer the question ... if I could understand it.  My
Python is much better than my French.  (Which isn't to
say my Python is good, rather that my French is bad. :-)

TimO



From martin at v.loewis.de  Sun Apr 13 12:09:36 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 18:09:36 +0200
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk> <3E977928.B3A6DC7E@engcorp.com> <yfsznmueec1.fsf@black132.ex.ac.uk>
Message-ID: <m3he92pef3.fsf@mira.informatik.hu-berlin.de>

Alexander Schmolck <a.schmolck at gmx.net> writes:

> Well, unless python has a radically different idea from anyone else on what
> interning a string (and "returning *the* interned string" string) means, I'd
> really think that you can rely on:
> 
> >>> type(a) == type(b) == str and b == a and inter(b) is intern(a)

To hold for arbitrary a,b? Certainly not:

>>> a=1
>>> b=2
>>> type(a) == type(b) == str and b == a and inter(b) is intern(a)
False

If you meant that 

type(a) == type(b) == str and b == a 'implies' intern(b) is intern(a)

then yes: Python does guarantee that, for arbitrary a,b.

Notice that Python does *not* (anymore) guarantee

type(a) == type(b) == str and b == a 'implies' id(intern(b)) == id(intern(a))

With Python 2.3, interned strings become mortal, i.e. they may get
uninterned if they are not referred-to anymore (from places other
than the interning dictionary itself).

> I agree the wording there is not particularly clear (and python
> doesn't much encourage "symbolic computation" -- so the intent is
> not to have people write intern("foo") all over the place where a
> lisper would write e.g. :foo), but I'd be really suprised if the
> intern semantics were intended to be unspecified w.r.t. the identity
> property above.

The intention is that intern gives you a unique distinct
representative from the set of equal strings.

Regards,
Martin




From warrenpstma at _______.com.hotmail  Wed Apr  2 09:14:24 2003
From: warrenpstma at _______.com.hotmail (WP)
Date: Wed, 02 Apr 2003 09:14:24 -0500
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to Python.
 Good idea? Bad Idea? PEP?
In-Reply-To: <b68jti$2mj66$1@ID-169208.news.dfncis.de>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de>
Message-ID: <RbCia.6782$9I.555309@news20.bellglobal.com>

Greg Ewing wrote the following Pyrex snippet:
>     for i from 0 <= i < n:
>       result.append(a[i])

This pyrex snippet reminds me (because of its orthogonality to C) of one 
thing, more than tab indents, that I have always thought was bogus in Python, 
which is the handling of integer loops.  range(0,n) sucks, xrange(0,n) is an 
interesting hack which obscures the readability of code, and post 2.x versions 
of python, with generators, and other 'inverted' ways of accomplishing the 
obvious procedural effect of an integer for loop, are all just ways of 
obfuscating otherwise perfectly readable python code. Okay, so I'm being
opinionated. I know that. But does anyone else thing that for the extremely 
common (and therefore bytecode optimizeable, syntax customizeable) case where 
you need to do a for loop over a range (possibly large range) of integers, 
that a language level keyword construct (such as the above) would be 
preferable to Python's choices of range, xrange, generators, custom 
class-based iterators, and whatever else you can come up with, that the lack 
of the most common idiom, being expressed in a bytecode optimized fashion 
would be good?

Of course, it all comes I suppose from integers being "first class objects" 
and thus, not actually integers, they are actually immutable objects which 
contain an integer, and so we get to the ugly condition where "i++" actually 
decref's one object, and either looks up or creates another object, and 
increfs it.  Hardly a fast operation to implement, but certainly one of the 
most common idioms in all of programming.

Going back (dizzyingly) from the bytecode/c implementation, to the syntax, for 
a moment. Imagine if Python had a pyrex-like first-class iteration case, where 
i (in the above example) is not an integer, but rather a reference to the 
for-integer-bytecode emitted by the compiler, an object which knows (in C) how 
to increment i's value quickly, and jump out of a loop when the limit n is 
reached.

Ever since Python 1.5.2, I have been hacking around on my own to see if I 
could get huge speed improvements out of Python.  I have settled down to the 
reality that Python+Pyrex, with syntactical speedbumps in between them, are 
probably the way things are going to stay. But I can dream, can't I?

Just a thought, no intent to start another PEP-flamewar.  ;-)

Warren




From dave at pythonapocrypha.com  Tue Apr  1 16:01:14 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Tue, 1 Apr 2003 13:01:14 -0800 (PST)
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <x7y92uhzpw.fsf@guru.mired.org>
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <Pine.LNX.4.51.0304011257400.15952@ns0.tbye.com>

On Tue, 1 Apr 2003, Mike Meyer wrote:

> PEP: 313
> Title: Adding Roman Numeral Literals to Python
> Version: $Revision: 1.2 $
> Last-Modified: $Date: 2003/04/01 17:41:34 $
> Author: Mike Meyer <mwm at mired.org>
> Status: Draft
> Type: Standards Track
> Content-Type: text/plain
> Created: 01-Apr-2003
> Python-Version: 2.4
> Post-History:

+1, well done!

Before we submit this to the BDFL, I propose that we wait a short while to
give the community an opportunity to debate its merits, followed by a
vote, just so Guido can get idea of where the community stands. Further, I
propose that Aahz be the official vote counter. Does anybody have any
preferences as to the voting method we should use?

-Dave




From aleax at aleax.it  Sat Apr 12 14:27:54 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 18:27:54 GMT
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk>
Message-ID: <KUYla.7409$T34.201025@news2.tin.it>

Matt Bergin wrote:

> I am trying to write a class which gives out objects and keeps track of
> them so they can be returned to the class and given out to other code
> when the code they were given out to no longer has any reference to the
> objects (i.e. has finished using it).

OK, makes sense.

> The trouble is, the manual says not to try and stop objects from being
> killed by creating new references in __del__. My original thoughts were
> to write some code in __del__ of the given out object that returned it
> to the original class. Is this is a good way of going about this
> problem? Any ideas would be welcome!

That might cause you problems at termination time, I think.  Also,
is there any risk that your objects could end up as parts of cycles
of mutual references?  If so, defining a __del__ makes it impossible
for the garbage collector to ever reclaim them.

Rather, you might ensure that all objects "given out" by the class
are kept alive anyway -- by simply keeping a reference to them in a
class variable, e.g. a list -- and, periodically, checking which
objects have no other references extant except the one in the list
(sys.getrefcount returning 2) so you can move them to a separate
list of "objects ready to be recycled".

You can try to avoid the periodic scanning overhead by using a
tricky __del__, or a non-tricky __del__ or weak-reference callback
for some object that's "packed together with" the object you give
out (keeping the latter alive in a per-class container anyway).  A
radically different and far cleaner approach would be to specify
a way (e.g. a method) for the using-code to say "I'm finished with
this object, reuse it freely from now on" -- no black magic at all
(of course, perhaps marginally smaller convenience for the using
code, but it's not all that bad to ensure a finalization method
gets explicitly called, IMHO).


Alex




From mboedick at mboedick.org  Fri Apr  4 13:08:02 2003
From: mboedick at mboedick.org (Matthew M. Boedicker)
Date: Fri, 4 Apr 2003 13:08:02 -0500
Subject: negative values in profile output
Message-ID: <20030404130802.F8345@gurthang.twcny.rr.com>

When using the profile module to profile a long-running program,
I am getting some negative values:

6753628 -4036.027   -0.001 -4036.027   -0.001 gobject_sql.py:36(<lambda>)

Am I correct to assume these don't have special meaning and are the
result of something wrapping?

Thanks.

-- 
Matthew M. Boedicker http://mboedick.org




From geoff at gerrietts.net  Thu Apr 10 13:24:28 2003
From: geoff at gerrietts.net (Geoff Gerrietts)
Date: Thu, 10 Apr 2003 10:24:28 -0700
Subject: Python compared toJava
In-Reply-To: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
Message-ID: <20030410172428.GA4764@isis.gerrietts.net>

Quoting Blaktyger (blaktyger at hotmail.com):
> Hi!
> 
> I recently read that the Python interpreter is like the Java Virtual Machine.
> Is that true?

It's hard to tell exactly what you mean by this, but since you seem to
have received an unusually high proportion of flip or catty responses,
I'll try to reply.

A Java program runs inside the context of the JVM. A Python program
runs inside the context of the Python interpreter. In this way they
are similar: the actual process observed in the operating system's
process table will be the JVM or interpreter, not the program itself.

Similarly, the JVM provides a standard interface for Java programs.
The idea is that the JVM will be standard across multiple different
platforms. The Python interpreter tries to offer the same kind of
portability. Neither is completely successful, but a careful developer
can achieve portability on either.

In other ways, these two environments are different. The JVM will not
compile code into bytecode for you; you need a Java compiler for that.
The Python interpreter will compile code into bytecode as required.
The JVM runs Java bytecode; the Python interpreter runs Python code.
Etcetera.

Answering your question really requires a better understanding of what
you want to know: in what way have you heard that these two are alike?
You should probably interpret the rash of snide replies to mean just
that: we don't really know what you're asking, but we'd answer you if
we did.

--G.

-- 
Geoff Gerrietts             "Me and my homies, we tag O.D.."
<geoff at gerrietts net>        --Unknown grafitti artist at a party




From aleax at aleax.it  Tue Apr 15 11:17:31 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 15:17:31 GMT
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1> <3e9c1413$1@news.012.net.il> <jJUma.23762$T34.707989@news2.tin.it> <Xns935EA3DC45768duncanrcpcouk@127.0.0.1>
Message-ID: <foVma.17305$LB6.443056@news1.tin.it>

Duncan Booth wrote:

> Alex Martelli <aleax at aleax.it> wrote in
> news:jJUma.23762$T34.707989 at news2.tin.it:
> 
>> import re
> listofnames = [re.escape(n) for n in listofnames]
>> namesre = re.compile('(%s)'%')|('.join(listofnames), re.I)
>> 
>> for m in namesre.finditer(astring):
>>     dosomething(astring, listofnames[m.lastindex-1])
>> 
>> 
> 
> I suggest adding the line above to avoid any nasty surprises (or use a
> different variable if you don't want to clobber listofnames).

If the "names" can include punctuation, that's definitely wise.  I
had taken "names" to be basically the same as "identifiers", but if
e.g. "c++" is a ``name'', then Duncan's precaution is warranted.


Alex




From johnroth at ameritech.net  Wed Apr 23 07:14:51 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 23 Apr 2003 07:14:51 -0400
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <3EA5B173.9420324C@alcyone.com> <mailman.1051056037.8077.python-list@python.org> <3EA5DE1E.22704437@alcyone.com> <m3r87tdbo3.fsf@mira.informatik.hu-berlin.de> <mailman.1051081416.21801.python-list@python.org>
Message-ID: <vacteqa3mi8ta9@news.supernews.com>

"Ian Bicking" <ianb at colorstudy.com> wrote in message
news:mailman.1051081416.21801.python-list at python.org...
On Wed, 2003-04-23 at 00:29, Martin v. L?wis wrote:
> Erik Max Francis <max at alcyone.com> writes:
>
> > but at this point it's looking awfully ugly.
>
> That's why it will take a long time until Python warns about
> assignments to True/False. Essentially, when Python 2.3 is considered
> an old Python version (i.e. at Python 4.0), the warning may be
> enabled. If the warning hasn't been seen for some time (say, at Python
> 6.2), it becomes an error.
>
> So the typical process is that people will have to actively break
> backwards compatibility at some time by removing the assignment.
> There is still a point in writing
>
> try:
>   True
> except NameError:
>   True, False = 1, 0
>
> as this gives you the "proper" boolean objects in 2.3+.

The other point in writing it this way is that it allows the compiler to
recognize
the idiom and remove it without issuing a warning. Whether or not
the compiler will ever do this is another issue, but I would consider
it a definite plus in providing a clean, hassle free migration path.

John Roth

Ian








From b.hall at irl.cri.nz  Thu Apr 10 23:08:37 2003
From: b.hall at irl.cri.nz (Blair Hall)
Date: Fri, 11 Apr 2003 15:08:37 +1200
Subject: __slots_ and inheritance
Message-ID: <3E9631B5.5070505@irl.cri.nz>

I'm having trouble understanding exactly how __slots__
is supposed to behave when one class is derived from
another. For example:

	# First case: works as I would expect
	class A(object):
     		__slots__ = ('x','y')
     		def __init__(self):
         		pass

	class B(A):
     		__slots__ = ('z',)
     		def __init__(self):
         		A.__init__(self)

	b = B()
	b.x = 1
	b.z = 3
	# b.c = 3 # fails

This script would fail at the last line as expected, because
there is no attribute 'c' allowed. However,


	# Second case: class B has no slots
	class A(object):
     		__slots__ = ('x','y')
     		def __init__(self):
         		pass

	class B(A):
     		def __init__(self):
         		A.__init__(self)

	b = B()
	b.x = 1
	b.c = 3 # passes

In this case the attribute 'c' can be created. Perhaps a __slots__
in the base class does not prevent attributes from being added
to the derived class (I guess this is good). But,

	# Third case: class A has no slots
	class A(object):
     		def __init__(self):
         		pass

	class B(A):
     		__slots__ = ('z',)
     		def __init__(self):
         		A.__init__(self)

	b = B()
	b.c = 3 # passes

in this case, although the class B has slots its base class
does not. This seems to allow instances of B to ignore their
own __slots__ requirements. This, to my mind, is a problem:
it seems to require knowledge of the __slots__ in all
base classes.

Can anyone steer me towards a better understanding of __slots__?




From km-news1 at onlinehome.de  Mon Apr 21 15:56:28 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Mon, 21 Apr 2003 21:56:28 +0200
Subject: Pythonwin don't show the complete list in interactiv window!?
Message-ID: <oprnzwcechmu3s0q@news.cis.dfn.de>

PythonWin 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32.
Portions Copyright 1994-2001 Mark Hammond (mhammond at skippinet.com.au) - see 
'Help/About PythonWin' for further copyright information.
>>> x=range(1000)
>>> x
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
(lines removed)
762, 763, 764, 765, 766, 767, 768, 769, 770, 771,
>>>

The interactive windows shows only a part of the list (last value 771). 
(Win2000)

A bug?

-- 
Gru? - regards Klaus :-)



From max at alcyone.com  Thu Apr  3 21:57:57 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 03 Apr 2003 18:57:57 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122>
Message-ID: <3E8CF4B5.1E221DD7@alcyone.com>

Bengt Richter wrote:

> So how do you feel about properties? Seems like the OP's proposal
> would
> open up another way to implement at least the assignment part. Perhaps
> there should be an analogous __getvalue__ method for the right hand
> side.
> (Already ducking ;-)

I really should have said "rebinding" instead of "rebinding/attribute
assignment."  I misspoke, sorry.

I know

	something = somethingElse

is always going to mean rebinding, and I like that.  I don't mind if

	something.subSomething = somethingElse

can mean different things; that seems very reasonable, in fact.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Listen to my song
\__/ Chante Moore
    Lsystem / http://www.alcyone.com/pyos/lsystem/
 A Lindenmayer systems explorer in Python.



From sdieselil at yahoo.com  Fri Apr  4 12:36:57 2003
From: sdieselil at yahoo.com (sdieselil)
Date: 4 Apr 2003 09:36:57 -0800
Subject: Why deleting large variables doesn't free up any memory?
Message-ID: <169a7283.0304040936.69e2b9a1@posting.google.com>

Hi

I'm using Python 2.2.2 under FreeBSD. When I create a large variable
(using for example "a = range(1000000)") 'top' command indicates that
Python ate 12MB of swap space. But when I execute "del a" nothing
happens, it doesn't release these 12MB! Using "gc.collect()" also
doesn't help. What's the problem? What happened to Python's garbage
collector?



From aleax at aleax.it  Sun Apr 20 03:47:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 20 Apr 2003 07:47:39 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <mailman.1050821313.12531.python-list@python.org>
Message-ID: <vgsoa.37954$LB6.862283@news1.tin.it>

Steven Taschuk wrote:

> Quoth Alex Martelli:
>   [...]
>> So, I tried implementing sum in the simplest way, as a
>> separate patch to operator.c (leaving add alone) -- THAT
>> gives me 124 microseconds, a 50% speedup with respect to
>> the simple for loop and clearly sufficient to justify
>> sum's existence (together with its simplicity, clarity,
>> and likely frequency of use).  [...]
> 
> +1.  Optimization of the most common case.  Excellent.
> 
> I'm not sure such a sum function belongs in the operator module,
> though.  At the moment I think it might fit best as a built-in; it
> goes nicely with min and max.

Yes, this seems to be the general feeling of those who have
responded on python-dev so far, with, however, one important
exception -- Raymond Hettinger thinks there are already too
many built-ins, so he'd rather add a new module instead.


Alex




From dale at riverhall.NOTHANKS.co.uk  Thu Apr 17 17:26:29 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Thu, 17 Apr 2003 22:26:29 +0100
Subject: Zope Sequences and Python Scripting
References: <pan.2003.04.17.12.34.15.777976@fraley.de>
Message-ID: <qt6u9vs13houri99i5fmsen20oof07ile0@4ax.com>

"Patrick W. Fraley" <patrick at fraley.de> wrote:

>Hi,
>
>first off:  I am pretty new to python and have just started with Zope, so
>please excuse and fopars on my side :)
>
Does that count as a faux pas?

  ;-)


-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From gh at ghaering.de  Wed Apr 16 07:39:40 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 16 Apr 2003 13:39:40 +0200
Subject: Problem Cron-Job using Python-Script
In-Reply-To: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de>
References: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de>
Message-ID: <3E9D40FC.3040006@ghaering.de>

Andreas Penzel wrote:
> Hello!
> 
> I have a Python-Script to create Log-Files und work with them.
> Very well.
> But if I start the Script with Cron it works not correct.
> I often use os.system() to send commands to the shell including paths from
> the os.
> Could this be the problem?

Yes. Cron normally doesn't set any environment variables.

> And if, how to solve?

a) Use full paths.

or

b) Set the PATH in the crontab:

PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin

-- Gerhard





From aleaxit at yahoo.com  Thu Apr 10 08:36:06 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: Thu, 10 Apr 2003 14:36:06 +0200
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <7xwui295xq.fsf@ruckus.brouhaha.com>
Message-ID: <b73ofo0ae1@enews3.newsguy.com>

Paul Rubin wrote:

> Alex Martelli <aleaxit at yahoo.com> writes:
>> __slots__ is, essentially, an optimization (in terms of memory
>> consumption) intended for classes that are likely to be created
>> in very large numbers: you give up the normal flexibility of
>> being able to add arbitrary attributes to an instance, and what
>> you gain in return is that you save one dictionary per instance
>> (dozens and dozens of bytes per instance -- that will, of course,
>> not be all that relevant unless the number of instances that are
>> around is at the very least in the tens of thousands).
> 
> I generally use __slots__ as a way to "declare" instance variables,
> so I get an exception if I accidentally assign to an undeclared one.
> I'd hope that was one of __slots__'s intended uses.

I don't see anything in __slots__'s well-documented design intent 
and behavior, http://www.python.org/2.2/descrintro.html &c, that
justifies such a "hope".  It may be feasible to have __slots__ do
double duty for that, if you don't need all the flexibility it's
going to lose you (e.g., no weak references, no "default values"
for instance variables through class variables...) AND you're in a 
very, VERY special case where no inheritance, __setattr__, access
to __dict__, or the like, can ever possibly impact your "hoped-for" 
functionality.  But, to me, such attempted "abuse" of __slots__ is
akin to using, say,
    hash(x)==x
as a way of testing whether "x is an integer-oid number" (handier
than int(x)==x because the latter fails when x is e.g. 23+0j) --
yeah, well, it works, sort of, but it's NOT what hash is meant for,
and you have no cause to complain if it doesn't work for your
purposes in any subtle or marginal case, or in the future.

If you really want to get an exception when an attribute is set
that you don't want to be set, you can use __setattr__, or a
custom metaclass.  __slots__ doesn't really help much here, IMHO.


Alex




From andrew.markebo at telia.com  Sat Apr  5 12:53:03 2003
From: andrew.markebo at telia.com (Andrew Markebo)
Date: Sat, 05 Apr 2003 17:53:03 GMT
Subject: nntplib: how to access NNTP response codes on exceptions?
References: <cirn8vsv4bhtp82n1tfil3e2kkounpgvav@4ax.com> <m3brzn27z3.fsf@telia.com>
 <hn2s8vs4p97bl4jcskrn34mgi7eartj4h6@4ax.com>
Message-ID: <m3he9crfum.fsf@telia.com>

/ carroll at tjc.com wrote:
|
|>You might have missed that parameters can be thrown with the
|>exceptions, and thats what e is.. an instansiation of a
|>NNTPTemporaryError class.
|
| No, I knew that, but the inclusion of the response code in
| NNTPTemporaryError is not in the docs.

Oh.. you read the docs?? Eh? ;-) I do the code ;-) (To be honest I do
both :-) )

     /Andy

-- 
 The eye of the compiler rests on the code!



From tim.one at comcast.net  Mon Apr 14 14:04:57 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 14 Apr 2003 14:04:57 -0400
Subject: Complexity of standard Python data structures (Re: Speed of string
 += string)
In-Reply-To: <slrnb9lsbe.4pa.marcus@infa.abo.fi>
Message-ID: <BIEJKCLHCIOIHAGOKOLHGEFPFFAA.tim.one@comcast.net>

[Marcus Alanen]
> ...
> As an example of "unfinished" interfaces: there is no list.clear()
> function, even though dict.clear()

If you want to clear a list,

   del list[:]

is the usual way to spell it;

   list[:] = []

is also sometimes used.  So there isn't a need for a list.clear() method.
dict.clear() can't be spelled efficiently in any other way.


> ...
> If somebody would say "python's list is a really a vector/array",
> that's more or less the only documentation one will ever need,
> help(list) tells the rest. The key characteristics are obvious from
> that one single sentence, right?

I believe they would be to you, but I also believe you're in a minority
there.

Note that it's not true that Python's list is really a vector/array -- the
language doesn't define its implementation, and doesn't want to.  It is true
that Python's lists have been implemented as contiguous vectors in the C
implementations of Python so far.  I don't know how Jython implements Python
lists, or how Vyper did (a Python derivative implemented in the functional
Ocaml language).





From vjb at vuit.com  Tue Apr 29 12:51:47 2003
From: vjb at vuit.com (Vince Buonassisi)
Date: Tue, 29 Apr 2003 09:51:47 -0700
Subject: import problem
In-Reply-To: <b8m82u$52d$1@news1.wdf.sap-ag.de>
Message-ID: <DD7D46ED-7A62-11D7-B1FA-000A95682CD2@vuit.com>

daniel,

thanks for the response.  i am new to python.  i inherited a project 
that was written in python from someone who left the company.  he was 
also the only python programmer here.

i have been trying to find information about those __init__.py/pyc 
files that i see laying around and how they relate to running the 
python scripts.  all of them are empty except for the compiled ones.  
can you point me to a source that will give me an explanation about the 
purpose of this file, __init__.py.  i have been unable to find an 
explanation for this file.

the test directory was missing the __init__.py file.  i put an empty 
one in there and my script ran just fine.

thanks,

On Marted?, apr 29, 2003, at 09:09 US/Pacific, Daniel Dittmar wrote:

> Vince Buonassisi wrote:
>> from WVWService.util import XMLUtil
>> from WVWService.test import HTTPClient
>>
>> the first import works just fine.  but, the second one i get the error
>> "ImportError: No module named test".  i have checked for mispellings;
>> i have checked to see if the directory test exists; i have checked the
>> system path by printing out sys.path from w/i the script to see if the
>> directory path is in sys.path (/home/vjb/workspace/wvwBase3/src/wvw).
>
> Maybe a missing __init__.py in WVWService/test?
>
> Daniel
>
>
>
> -- 
> http://mail.python.org/mailman/listinfo/python-list
>





From 'tert at pncen.hf'.decode  Sun Apr 20 20:49:55 2003
From: 'tert at pncen.hf'.decode (Greg Krohn ('rot-13'))
Date: Sun, 20 Apr 2003 19:49:55 -0500
Subject: python dynamic scoping question
References: <79ad5955.0304201513.6ffd4f83@posting.google.com>
Message-ID: <b7vg9t01mdo@enews2.newsguy.com>

"Julia Goolia" <juliagoolia301 at hotmail.com> wrote in message
news:79ad5955.0304201513.6ffd4f83 at posting.google.com...
> Hi,
>
> I have a question regarding scoping in python.  I wish to have some
> mechanism which puts variables at a function level scope into the
> global scope, dynamically.  I cannot alter my design.
>
> For example, I want this to work:
>
> def f():
>   def g():
>     x = 5
>   g()
>   print x
> f()
>
> I want to be able to see x in f, event though it was assigned in a
> lower scope, g.
>
> Thank you for your time.

>From the docs:

globals() - Return a dictionary representing the current global symbol
table. This is always the dictionary of the current module (inside a
function or method, this is the module where it is defined, not the module
from which it is called).

>>> def f():
...   def g():
...     x = 5
...     globals()['x'] = 5
...   g()
...   print x
...
>>> f()
5


Cool?

greg





From peter at engcorp.com  Wed Apr 16 15:09:26 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 15:09:26 -0400
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com> <v9r6m7864jmic6@news.supernews.com>
Message-ID: <3E9DAA66.171CAC9@engcorp.com>

John Roth wrote:
> 
> > So how many of you guys use XP processes?
> 
> I do. Writing tests before code is something that I've thought would
> be a good idea for literally decades, but it took both XP and having
> the test tool delivered with the language to get me to start.

I know you knows this, but for the benefit of others who aren't
"into" this approach yet, I think a clarification might be in order.

With TDD and XP, the tests are not _all_ written before the code.
In fact, the tests and the code are developed together, with 
rapid alternation back and forth between writing tests (which should
fail at first, because of course the code doesn't exist) and code
(which then passes the tests), tests and code.

And on this note, I'd also like to point out Kent Beck's recent
book "Test Driven Development" http://www.amazon.com/exec/obidos/tg/detail/-/0321146530/qid=1050520075/sr=8-1/ref=sr_8_1/102-3162235-5348141?v=glance&s=books&n=507846) which is a good introduction to the topic.  Also note
that it uses Python in one section to demonstrate using test-driven
development to bootstrap a *unit-testing framework* using itself!

-Peter



From and-google at doxdesk.com  Sat Apr 26 03:47:55 2003
From: and-google at doxdesk.com (Andrew Clover)
Date: 26 Apr 2003 00:47:55 -0700
Subject: True/False
References: <mailman.1051110756.23235.python-list@python.org> <2c60a528.0304240517.25f2fe29@posting.google.com> <a8b7f07a.0304251249.ad24819@posting.google.com>
Message-ID: <2c60a528.0304252347.142f0548@posting.google.com>

A. Lloyd Flanagan <alloydflanagan at attbi.com> wrote:

> This has the advantage of making True and False look more like builtins.

Interesting. I'm not sure about this; it seems to have the potential to
confuse other, unrelated modules. For example some package might check for
the existence of 'True' to see whether it needs to provide its own boolean
type that's distinguishable from integers. If the builtin is fiddled it
would look like there was a real boolean type but attempts to use it could
fall over. Modules that affect unrelated modules merely by being imported
can be tricky to debug.

-- 
Andrew Clover
mailto:and at doxdesk.com
http://www.doxdesk.com/



From mads at troest.NEVERMORE.dk  Sat Apr 12 14:37:29 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Sat, 12 Apr 2003 20:37:29 +0200
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it>
Message-ID: <1c8pot8sof9m3$.1a39huke99f0m.dlg@40tude.net>

On Sat, 12 Apr 2003 18:45:26 GMT, Alex Martelli wrote:

> Building up lists of strings with mylist.append(piece), then joining
> them together at the end with ''.join(mylist), is the canonical solution;
> others, which I've seen already pointed out to you, include using cStringIO.

Alex, thanks for your response. I had a hunch the += approach would be a 
performance killer. I'll try the list.append approach.

Regards,
   /\/\\ads Orbesen Troest



From martin at v.loewis.de  Fri Apr  4 19:33:47 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Sat, 05 Apr 2003 02:33:47 +0200
Subject: cgi.py -- anybody using it?
In-Reply-To: <Pine.LNX.4.44.0304050051160.1193-100000@alice>
References: <Pine.LNX.4.44.0304050051160.1193-100000@alice>
Message-ID: <b6l896$961$07$1@news.t-online.com>

John J Lee schrieb:
> I just tried testing some HTML form upload code using cgi.py, and
> discovered that it wants lower-case HTTP headers :-/

How did you discover that? This is simply not true.

Regards,
Martin




From pan-newsreader at thomas-guettler.de  Fri Apr 11 03:17:47 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Fri, 11 Apr 2003 09:17:47 +0200
Subject: Promoting Python
References: <qGqla.61$Tj7.91655@news.uswest.net>
Message-ID: <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>

On Fri, 11 Apr 2003 05:34:35 +0200, Kevin Altis wrote:

> One of the hot topics at PyCon was Promoting Python.
> I would like to invite you to join the marketing-python list if you're
> interested in contributing ideas
> ...

[cut]

I think python is a very good language for education. You could get in
contact with local schools.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From mh at pixar.com  Fri Apr 11 14:24:03 2003
From: mh at pixar.com (Mark Harrison)
Date: Fri, 11 Apr 2003 18:24:03 GMT
Subject: module for doing unix "find"-like things?
Message-ID: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com>

Is there such a thing?  I didn't see such a thing, but
"find" is not the easist thing to search for.

Specifically, I would like to replace an external call to
    for i in `find $top -type l -maxdepth 1`; ...

Many TIA,
Mark

-- 
Mark Harrison
Pixar Animation Studios



From jjl at pobox.com  Mon Apr  7 21:06:25 2003
From: jjl at pobox.com (John J Lee)
Date: Tue, 8 Apr 2003 02:06:25 +0100
Subject: exceptions with spaces don't work??
In-Reply-To: <b6t48i$mvd$1@news.doit.wisc.edu>
References: <b6t48i$mvd$1@news.doit.wisc.edu>
Message-ID: <Pine.LNX.4.44.0304080156480.1201-100000@alice>

On Mon, 7 Apr 2003, Paul Probert wrote:
> As a python newbie, I came across a problem:
> If I raise an exception with a string, as in:
>     raise 'somestring'
> then my statement
>     except 'somestring':
> will catch it unless there are spaces. For instance
> suppose I have a file ext.py:
[...]

1. Don't Do That.  String exceptions are deprecated.  Use a real exception
object instead:

class SomeError(Exception): pass

2. I think your theory about spaces in strings wouldn't stand up to some
experimentation with longer strings.  IIRC, catching a string exception
requires identity of the string objects, not mere equality -- they must be
the same object.  So:

someerror = "somestring"
try:
    raise someerror
except someerror:
    assert 1

is guaranteed to work (at least until string exceptions finally get
prohibited), while:

try:
    raise "someerror"
except "someerror":
    assert 1

is not.  Maybe it happens to work sometimes because Python is caching
short strings, so the two occurrences of "someerror" happen to be the same
object in memory.


John




From shalehperry at attbi.com  Tue Apr 29 11:36:42 2003
From: shalehperry at attbi.com (Sean 'Shaleh' Perry)
Date: Tue, 29 Apr 2003 08:36:42 -0700
Subject: Where did you learn to unit test?  How did you learn?
In-Reply-To: <1147e466.0304290638.287687f6@posting.google.com>
References: <1147e466.0304290638.287687f6@posting.google.com>
Message-ID: <200304290836.42804.shalehperry@attbi.com>

On Tuesday 29 April 2003 07:38, Christopher Blunck wrote:
> Hi all-
>
> This is my second generic question posted to c.l.p (the first was 'How
> many of you are Extreme Programmers?') but it seems like a somewhat
> appropriate discussion for the group.
>
> I've noticed in my development that Python programmers (in general)
> write more unit tests than their java counterparts.  If I download a
> project off of jakarta.apache.org, it very rarely has unit tests.
> However, when I download a python module, it frequently has unit
> tests.
>

I believe this is due in part to Python being an interpreted language.  With 
Java you have a compiler there to catch many problems but with Python most 
bugs, misfeatures, etc. can only be found by exercising the code.  Python 
coding is all about handling the runtime which makes for rapid writing of 
code but almost requires stronger testing.

As Skip said, I got into unit testing by writing test cases to prove bugs were 
fixed.  When more bugs were found either new tests were written or old ones 
improved.

After that I got into test writing as a way to actually write my applications.  
One thing I find hard for me is interface design.  So now I write the code 
that uses my class or function before I write the function.  This helps me 
identify the ways I want the code to work and also provides me with a test 
infrastructure.




From aleax at aleax.it  Mon Apr 14 18:05:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 22:05:46 GMT
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it> <pan.2003.04.14.18.33.53.118906@effectif.com>
Message-ID: <_gGma.14411$LB6.389752@news1.tin.it>

Graham Ashton wrote:

> On Sat, 12 Apr 2003 15:53:38 +0000, Alex Martelli wrote:
> 
>>> Alexander Schmolck <a.schmolck at gmx.net> wrote in message
>>> news:<yfsistmh4ec.fsf at black132.ex.ac.uk>...
>>>  
>>>> __slots__ is the one addition in python2.2 that really annoys me. It
>>>> seems an ill-conceived jumble of conceptually unrelated things (bug
>>>> avoidance, optimization, safety, encapsulation) that screws up
>>>> reflection big time for little apparent gain. As Knuth said ...
>> 
>> I think it's an important memory optimization, PERIOD.  The other
>> conceptually unrelated things (bug avoidance -- I can't even SEE the
>> others!) aren't in __slots__' own design (even though a commentator
>> who's usually quite perceptive seems to have seen the "bug avoidance"
>> feature that wasn't really there).
> 
> From http://www.python.org/doc/2.2.1/whatsnew/sect-rellinks.html:
> 
>     Finally, it's possible to constrain the list of attributes that can be
>     referenced on an object using the new __slots__ class attribute.
>     Python objects are usually very dynamic; at any time it's possible to
>     define a new attribute on an instance by just doing obj.new_attr=1.
>     This is flexible and convenient, but this flexibility can also lead to
>     bugs, as when you meant to write obj.template = 'a' but made a typo
>     and wrote obj.templtae by accident.
> 
>     A new-style class can define a class attribute named __slots__ to
>     constrain the list of legal attribute names.
> 
> This thread has pointed out rather well that __slots__ don't really behave
> as suggested above in anything but the simplest of situations. It wasn't
> until I read the __slots__ section in the nutshell book that I realised
> that this wasn't the case. Consequently, I think the above paragraphs are
> misleading. Shouldn't they be fixed?

Yes, I would agree they should be.  I don't know how (and indeed, if)
one can submit patches for AMK's "what's new" documents, though -- I
don't normally think of them as part of "Python's docs" (I don't even
think they are in the CVS...).


Alex




From brian094 at sympatico.ca  Mon Apr  7 16:48:13 2003
From: brian094 at sympatico.ca (Brian Alexander)
Date: Mon, 07 Apr 2003 16:48:13 -0400
Subject: curses: xterm vs. linux terminal
References: <3E9065FE.1090108@sympatico.ca> <b6qd2p$ib9$1@news1.radix.net>
Message-ID: <3E91E40D.8050409@sympatico.ca>

I think you are right. The terminal comes up as gnome-terminal when ps 
is run. I guess it is supposed to provide xterm-like behaviour.

I tried running the xterm from gnome-terminal and after it loaded, I ran 
my box-drawing program. The line characters (ACS_HLINE etc.) still worked.

What is the difference between these two terminal programs? Does the one 
just interact with Gnome a little better?

Many thanks,

Brian.


Thomas Dickey wrote:
> Brian Alexander <brian094 at sympatico.ca> wrote:
> 
>>Hello;
> 
> 
>>Why is it that the curses special (alt) characters do not display on xterms?
> 
> 
>>Ex:
> 
> 
>>theScreenObj.box(curses.ACS_VLINE, curses.ACS_HLINE)
>>theScreenObj.refresh()
> 
> 
>>This code puts a box with graphical characters around the window when 
>>the terminal is of type linux. When the terminal is of type xterm, 
>>blanks appear where the line characters should be.
> 
> 
>>Is there anyway to fix this behaviour? I find it hard to believe an 
>>xterm does not support the drawing of these simple graphical characters.
> 
> 
> normally it doesn't (linux console doesn't draw lines either).
> 
> I modified XFree86 xterm a few years ago to draw lines when the font
> does not contain them; you're apparently not using XFree86 (or else
> thinking that "rxvt" is "xterm" ;-).
> 





From alloydflanagan at attbi.com  Fri Apr 11 19:59:41 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 11 Apr 2003 16:59:41 -0700
Subject: module for doing unix "find"-like things?
References: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com>
Message-ID: <a8b7f07a.0304111559.6707284b@posting.google.com>

Mark Harrison <mh at pixar.com> wrote in message news:<7LDla.1483$Pq2.135830754 at newssvr14.news.prodigy.com>...
> Is there such a thing?  I didn't see such a thing, but
> "find" is not the easist thing to search for.
> 
> Specifically, I would like to replace an external call to
>     for i in `find $top -type l -maxdepth 1`; ...
> 
> Many TIA,
> Mark

If such a module doesn't exist I might write one.  A friendlier (more
pythonic?) interface to os.stat() would be an obvious starting point.



From chasm at rift.sytes.net  Tue Apr  1 08:41:37 2003
From: chasm at rift.sytes.net (Julian Tibble)
Date: Tue, 1 Apr 2003 13:41:37 +0000 (UTC)
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift> <3E890322.4B003C9C@engcorp.com>
Message-ID: <slrnb8j5og.o6a.chasm@galileo.rift>

In article <3E890322.4B003C9C at engcorp.com>, Peter Hansen wrote:
> Julian Tibble wrote:
>> 
>> p.s. maybe my python version is not very good, feel free to make
>> it go faster:
> 
> I'd love to try improving it, and there are certainly many
> possibilities for improvement, but the requirements are 
> inadequately specified.  Specifically, what is the nature of
> the input?  Is it supposed to work with any kind of file,
> including binaries, or only with files containing, for example,
> text with basic punctuation and numbers? (In other words, 
> exactly what character set is involved, and what is the nature
> of the input file?)
>

If some of the design decisions seem a bit strange, it's because I wrote
it to produce exactly the same output as my C program so that it was a
fair test.

It is supposed to work with all binary files, although in practise it
would be used on text files - documentation, etc.


>> # Regular expression to match a word
>> re_word = re.compile('[a-z]+')
> 
> Regular expressions are often not the fastest way to approach
> a problem.

Well, I thought they might not be, but I couldn't think of any
alternatives.

> 
>> # Read in whole file and split into words
>> try:
>>         while 1:
>>                 line = raw_input().lower()
> 
> Dealing with data one line at a time is often not the 
> fastest approach.  It does suggest that your data is
> line-oriented, however, but is it necessarily so?  For
> example, could newlines be replaced with spaces and the
> program would still work correctly?

Yes, I did try reading the whole file into memory (that's what the C
version does) but the process kept getting killed for using too much
memory, even though I have 256MB of RAM and the input file is ~80MB


>>                 for word in line:
>>                         if wordcount.has_key(word):
>>                                 wordcount[word] += 1
>>                         else:
>>                                 wordcount[word] = 1
> 
> Among other things, it's possible that catching a KeyError
> for the case where the key does not yet exist is faster,
> and there is also dict.setdefault() which might be of help.

setdefault() actually makes the running time worse, however catching
the KeyError did not.  This is probably due to the fact that the input
file is the same file repeated a number of times - so you should only
get exceptions during the first "section".  This shouldn't make too much
difference, as the C version should also degrades in speed when there
are more unique words due to hash collisions.

Also, the program would probably only prove useful on text files, and
there are only ~45,000 words in /usr/dict/words IIRC, so large numbers
of KeyErrors shouldn't occur.


>> for n,w in sortable:
>>         print w,n
> 
> Since you're not using the output here, why include it in the
> timing?  Seems like it could skew the result under certain 
> conditions...

As I said, I wanted it to be identical to the C program.
Also, by _far_ the majority of the time is spent in the loop.


Julian



From majbritt37 at hotmail.com  Fri Apr 18 16:22:40 2003
From: majbritt37 at hotmail.com (=?ISO-8859-1?Q?M=E5rten_Ask?=)
Date: 18 Apr 2003 13:22:40 -0700
Subject: PIL binary for windows?
References: <fe1cb5be.0304120417.5b756718@posting.google.com> <o4yma.19580$c42.5968@fe01.atl2.webusenet.com>
Message-ID: <fe1cb5be.0304181222.1879d35a@posting.google.com>

"David Rushby" <woodsplitter at rocketmail.com> wrote in message news:<o4yma.19580$c42.5968 at fe01.atl2.webusenet.com>...
> "M?rten Ask" <majbritt37 at hotmail.com> wrote in message
> news:fe1cb5be.0304120417.5b756718 at posting.google.com...
> > Does anyone know where I can find a PIL binary for windows and Python
> > 2.3? Or maybe some help in building it myself?
> 
> http://thor.prohosting.com/~drushby/files/PIL_113_for_py23_custom_compiled.z
> ip


Thanks! :)



From mis6 at pitt.edu  Fri Apr 11 15:40:03 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 11 Apr 2003 12:40:03 -0700
Subject: Getting source code from interpreter-defined function
References: <mailman.1050055156.23007.python-list@python.org>
Message-ID: <2259b0e2.0304111140.3170a48f@posting.google.com>

"POYEN OP Olivier (DCL)" <Olivier.POYEN at clf-dexia.com> wrote in message news:<mailman.1050055156.23007.python-list at python.org>...
> Hi all,
> 
> I'm playing with inspect module, which, BTW, is great for finding useful =
> info on saved object, written object.
> With inspect, I can get source code from a function.
> 
> Is there anyway I could do the same with "alive" object, without having =
> to wrap the whole python interpreter and create tmpfile for writing =
> session ? (an "alive" object would be an object defined at the =
> interpreter prompt)
> 
 --

This thread maybe of interest to you:

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&threadm=2259b0e2.0301030725.93d18c9%40posting.google.com&rnum=1&prev=/groups%3Fhl%3Den%26lr%3D%26ie%3DISO-8859-1%26q%3Dsimionato%2Bdecompyle%26meta%3D

In short, you cannot extract the source from live objects; you can
decompile the bytecode, though, by using the decompyle module.


                             Michele



From warlock at eskimo.com  Thu Apr 24 22:14:01 2003
From: warlock at eskimo.com (Jim Richardson)
Date: Thu, 24 Apr 2003 19:14:01 -0700
Subject: Dict handling of floats
Message-ID: <98imn-ust.ln1@grendel.myth>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I have come across a weirdness. I have a number of floats, (lat long
pairs) which I stuff into a dict, but when I take them out, my nice %2f4
number, (say, 47.5894) is something long and *almost* right, like
47.589333333333333339. What gives? I am not doing anything special as
far as I can tell, just stuffing the numbers into a dict, and recalling
it later. This is with Python 2.2 on Linux, and no, it's not an original
Pentium :)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+qJnnd90bcYOAWPYRAly3AKCXn4yiCCh+q1J9Pp8/ZR0GlbwYywCeKbZG
O/FWrpY8wXYvjU+OIu62wpc=
=NUpX
-----END PGP SIGNATURE-----

-- 
Jim Richardson         http://www.eskimo.com/~warlock

Linux, because eventually, you grow up enough to be trusted with a fork()



From jdhunter at ace.bsd.uchicago.edu  Tue Apr  1 09:28:03 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Tue, 01 Apr 2003 08:28:03 -0600
Subject: Mimic a web surfer ?
In-Reply-To: <3e894b7d$0$29158$626a54ce@news.free.fr> ("shagshag13"'s
 message of "Tue, 1 Apr 2003 10:22:25 +0200")
References: <3e894b7d$0$29158$626a54ce@news.free.fr>
Message-ID: <m2r88m2sy4.fsf@mother.paradise.lost>

>>>>> "shagshag13" == shagshag13  <shagshag13 at yahoo.fr> writes:

    shagshag13> hello, i would like to develop a script which will
    shagshag13> mimic a web surfer, that is :

    shagshag13> - follow links as a human and thus always come from
    shagshag13> previous page (i think there is something to do with
    shagshag13> http 'referer' and cookies) - be able to log if
    shagshag13> necessary

    shagshag13> is there anything already existing ? i'm looking for
    shagshag13> any tips which will help me to start...

    shagshag13> note : this could be used to have "automatic" access
    shagshag13> to free webmails not providing pop protocol.

Yes, this is possible.  As you say, you generally need to set referer
and or cookies.  This can be done with httplib.  I often find the best
way to do this is to install a sniffer which monitors traffic on port
80.  If you are on linux, you may want to try sniffit.  On windows, I
have had luck with proxomitron.  With one of these sniffers running,
browse the site you want to mimic and capture all the traffic.  Pore
through the logs and figure out how the site wants you to handle
cookies and referers, and then whip up your python version.  Here is
an example script using httplib  that sets referer and cookie headers

import httplib
import socket
host = 'slashdot.org'
pathn = '/science/01/12/03/1630212.shtml'
try:
    h = httplib.HTTP(host)
    h.putrequest('GET', pathn)
    h.putheader('Accept', 'text/html')
    h.putheader('Accept', 'text/plain')
    h.putheader('User-Agent', 'Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.4) Gecko/20010913')
    h.putheader('From', 'marshall at migley.zko.dec.com')
    h.putheader('Cookie', 'mycookieval')
    h.putrequest('Referer', 'http://myserver.com')

    h.endheaders()
    errcode, errmsg, headers = h.getreply()
except socket.error, er:
    print 'socket error ', er

print errcode

hkeys = headers.keys()
#for key in hkeys:
#    print key, ' = ', headers.getrawheader(key)

John Hunter




From jason at tishler.net  Thu Apr  3 11:02:05 2003
From: jason at tishler.net (Jason Tishler)
Date: Thu, 03 Apr 2003 11:02:05 -0500
Subject: Cygwin Python regression testers wanted
Message-ID: <20030403160205.GA2896@tishler.net>

I'm in the process of updating the Cygwin section of the official Python
README.  Since I only have access to Windows NT 4.0 and 2000, I am
seeking Cygwin Python regression testers for the following Windows
platforms:

    95
    95 OSR2
    98
    98 Second Edition
    Me
    XP Home
    XP Professional

If you are willing to help, then perform the following steps:

1. rebase your system [1]

   a. install the rebase package via Cygwin's setup.exe (if not installed)
   b. shutdown all Cygwin processes
   c. start bash (do not use rxvt)
   d. execute rebaseall (in the bash window)

2. run the Python regression test [2]

   python -E -tt /usr/lib/python2.2/test/regrtest.py -l 2>&1 | tee regrtest.out

3. post any failures and/or hangs other than test_commands

Notes:

[1] Sorry, but Python's regression test requires a fully rebased system
    to prevent Cygwin fork() problems.

[2] If a test hangs, then use regrtest.py's "-x" option to exclude this
    test.  For example, in the past test_select has hung under 9X/Me, so
    this test can be skipped as follows:

        python -E -tt /usr/lib/python2.2/test/regrtest.py -l -x test_select ...

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From h_schneider at marketmix.com  Thu Apr 17 05:11:12 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Thu, 17 Apr 2003 11:11:12 +0200
Subject: py2exe Registry Access from NT Service ...
Message-ID: <b7lr3g$b3q$06$1@news.t-online.com>

Hi,

I created a NT service under with py2exe under Python 2.1
The service ran without any problems.

Then I upgraded to Python 2.2 and py2exe for 2.2 and now I have a big
problem:
All functions that access registry keys from within the service fail.
When I run my application outside the Service class, everything works well.

Any idea?

Thnanks in advance,
Harald





From ChuckEsterbrook at yahoo.com  Mon Apr 14 02:58:47 2003
From: ChuckEsterbrook at yahoo.com (Chuck Esterbrook)
Date: Sun, 13 Apr 2003 23:58:47 -0700
Subject: sys.excepthook and threads
Message-ID: <200304132358.47933.ChuckEsterbrook@yahoo.com>

Is there a reason why sys.excepthook does not get invoked by threads 
(other than the main thread)?

The program below does not output 'caught exception', but will if 
Raise() is called directly. Tested on Python 2.2.2 and 2.3a2. A quick 
search of c.l.p showed mention of this, but no explanation.  I'm just 
curious.

BTW the docs for sys.excepthook don't mention this phenomena at all. 
They say "When an exception is raised and uncaught, the interpreter 
calls sys.excepthook with three arguments" so that's what I expected, 
threads or not.

-----------------------------------------------
def MyHandler(typ, exc, tb):
    print 'caught exception'

import sys
sys.excepthook = MyHandler

from threading import Thread

def Raise():
    assert 0

if 1:
    Thread(target=Raise).start()
else:
    Raise()
-----------------------------------------------


-- 
Chuck
http://ChuckEsterbrook.com





From jepler at unpythonic.net  Fri Apr 18 15:46:23 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 18 Apr 2003 14:46:23 -0500
Subject: A new and very robust method for doing file locking over NFS?
In-Reply-To: <5j65pbai03.fsf@xoxy.net>
References: <lcu1cw525a.fsf@gaffa.mit.edu> <5j65pbai03.fsf@xoxy.net>
Message-ID: <20030418194620.GB18062@unpythonic.net>

The other problem you run into with NFS is that the client might see
stale data.

In old-fashioned NFS client/server setups, a client does an attribute
fetch to determine whether cached data (file contents) is still current.
Time stamps have 1-second granularity.  If you have an operation that
does not require very long to complete, a client can see out-of-date
information.  Example (lock/unlock correspond to calls to a custom lock
daemon over TCP, just as you suggest):
    Client 1        Client 2
    Lock file
    ---------------------
    Write something
    Unlock

    Lock file
    Read it
    Unlock
                    Lock file
                    Write something
    ---------------------
                    Unlock
    Lock file
    read it 

If the actions between the dashed lines all took place within the same
wall second, client 1 may read its own cached data instead of the data
written by client 2.  Our eventual solution was a kernel patch which let
us evict any cached data for a particular file from memory, and we did
it every time we performed a locking operation.  You can bet that
performance suffered, but at least it was correct.

I don't know if any of these problems have been corrected.  I think that
some NFS servers have timestamps that have finer resolution than one
second.

Jeff




From johnroth at ameritech.net  Wed Apr 16 14:18:59 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 16 Apr 2003 14:18:59 -0400
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512847.28150.python-list@python.org>
Message-ID: <v9r7c9obu9153c@news.supernews.com>

"Michael Chermside" <mcherm at mcherm.com> wrote in message
news:mailman.1050512847.28150.python-list at python.org...
> Christopher writes:
> > So how many of you guys use XP processes?
>
> Well, I'm not sure if I qualify or not. I really LIKE most of the
points
> that XP makes, but I find many of its proponants overly exuberant. I
> think exhaustive, automated tests are a wonderful idea. I believe in
> refactoring. I think pair programming works very well most of the
time.
> I think super-short release cycles (with code that grows feature-by-
> feature) is a great idea. I think having a customer on-site is
extremely
> helpful.
>
> I DON'T think that it's wise to just start coding without a bit of
design
> up front (although one must keep YAGNI in mind). I don't think that
using
> the code itself as the only documentation (no design docs) works well.
> Most of all, I don't like mandating any particular methodology... the
> needs vary from project to project and team to team.

At the risk of being overly exuberant, most XP practitioners do
some design up front (enough to get a handle on the problem, at least.)
They just don't treat that design as cast in stone. And I've never seen
a project that wouldn't be helped by a few pages of carefully thought
out introduction to the basic concepts of the thinking behind it. With
UML if it is helpful.

> But none of that really matters, because I don't get to set the policy
> where I work. We don't do unit tests (except for the ones that _I_
write)
> because "they would take too long". We deliver functionality in a
single
> big release. We don't do pair programming except for very brief
periods.
> Perhaps someday I can change those things, but for now, I don't get to
> practice much of this, except in my at-home-on-my-own-time projects
> (basically non-existant since my son was born 4 months ago).
>
> -- Michael Chermside
>
>





From rodrigob at elo.utfsm.cl  Mon Apr 28 01:12:33 2003
From: rodrigob at elo.utfsm.cl (Rodrigo Benenson)
Date: 27 Apr 2003 22:12:33 -0700
Subject: Pyrex Hard Case
Message-ID: <607d8200.0304272112.2f02bcf5@posting.google.com>

Hi!

I'm a very happy python user.
I had develloped script, web pages and some research software using
python.
When I needed to connect Python to more efficient C code I use only
the best: Pyrex, it's easy, it's pythonic and it's works.

BUT I'm in front of a hard case. I'm finishing the so long imagined
Scilab <-> Python interface (if you do not know what is Scilab, look
at the end of mail).

Some day's ago talking with another Scilab fan I had found the door to
this conection.

After three days of works I had got my first matrix addition !

All the job is done using pyrex and disutils except one step. I'm
mailling this list to ask to you how do you think I could solve this
problem.

Basically there is a Macro function that I can't map to Pyrex, so I
create an equivalent dummy function and hand edit the .c file erasing
the definition. This is sufficient because Pyrex use the dummy
function every time and the next the macro do the correct C
translation.

The problem became rough because the macro itself maps a Fortran call.
And the Fortran code correspond to an virtual structure ised to match
a memory stack.

So, I admit, is a hard case.

I don't want to bore every one with the technical details, if some one
ask for them I can describe the code in more detail.

Basically I thinks that there are many case where some strange macro
call will make hard to map the code into Pyrex.
There should be way to define some "macro functions" or something that
like,

that allow you to define a prototype without including it into the
definitions of the C code,

it should be only used to check the consistency of the code and to
create automatic type translations.


I hope that the Pyrex users/devellopers read this news group (as the
web page said).

I think that this is a True Feature and they are many possible case
where things like this are usefull (specially when Fortran enter in
the scene).

The "tell me the C name" feature has saved the other half of the
difficulties, using things like 'cdef int hello "C2F(hello)" (char
*)'.

I suppose that the Scilab <-> Python interface is beauty example of
software integration.
Using the Blas/Atlas and Lapack Fortran code for low level number
crunching, using Fortran and C to devellop the core of Scilab, using
Tk and Gtk to provide X graphics outputs and using Python to connect
the power of number crushing to the heterogeneous data ambients.


Many thanks for your support.
RodrigoB.

Scilab: it is the best open source matrix computation software
aviable.



From andrew-pythonlist at puzzling.org  Fri Apr 25 04:06:58 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Fri, 25 Apr 2003 18:06:58 +1000
Subject: Date validation
In-Reply-To: <b8ap7f$4go$1@main.gmane.org>
References: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au> <b8ap7f$4go$1@main.gmane.org>
Message-ID: <20030425080658.GG15695@frobozz.local>

On Fri, Apr 25, 2003 at 09:48:56AM +0200, Fredrik Lundh wrote:
> James Gregory wrote:
> 
> > I mean I *could* write this myself, but I had to ask the computer how
> > many days in April - someone with a better grasp of the date must have
> > written this stuff before :)
> 
> don't you have any knuckles on your left hand?

And which knuckles do you use to remember which years are leap years?  ;)

> if you want your program to check how many days there are in
> a given month, use the calendar module:

He does.  In his words: "how can I check whether or not such strings
represent a valid date?"
    
His problem is verifying that the entire date is correct.  By the time you
factor in the varying days of the months and leap years, I can understand
wanting to use a prewritten and tested module rather than an ad-hoc solution
for a problem that many people must have encountered before.

-Andrew.





From cben at techunix.technion.ac.il  Mon Apr  7 09:58:57 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Mon, 7 Apr 2003 16:58:57 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m28yum76rt.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
 <b6qs04$7eqi1$1@ID-169208.news.dfncis.de> <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
 <mailman.1049693590.23386.python-list@python.org> <m28yum76rt.fsf@localhost.localdomain>
Message-ID: <Pine.GSO.4.51.0304071626360.8306@techunix.technion.ac.il>

Patrick K. O'Brien wrote on 2003-04-07:

> Ian Bicking <ianb at colorstudy.com> writes:
>
> > On Mon, 2003-04-07 at 00:23, Carl Banks wrote:
> > > Third, there are a *ton* of functions in Emacs that operate on a
> > > certain point in a certain buffer (delete-char, kill-region, etc.).
> > > These kinds of functions typically get bound to keystrokes and other
> > > sorts of commands, which a user expects to work in the current buffer.
> > > If all these functions had to be passed the current buffer as an
> > > argument, it would just complicate things.
> >
> > Maybe instead they should be methods of the buffer, and a keystroke
> > automatically calls a method on the current buffer.
>
> That should be easy enough.
>
> > Managing so many methods from diverse sources might be difficult.  I
> > want generic methods, I keep thinking of places where they'd be mighty
> > useful... like Mixins but better.
>
> This I don't understand.  Could you elaborate?
>
I'm not Carl but I had similar thoughts, so I'll try :-).  Tons of
things in emacs are dependant on the "current something" and they are
usually inherited from more basic places in some ways.

So it would be nice to manage those things with classes / subclasses
and classes/instances.  Subclassing and multiple inheritance can be
used to merge them in appropriate ways.  Attribute access is used to
access everything.  With python's dynamism (think metaclasses), there
is enough flexibility to extend this scheme.

Each buffer will be an object.  It will have a lot of methods, like
`forward_sentence`, that come from different places, including the
current mode(s) of this buffer.

Cooperative super calls will be needed for things like keymap lookup
ot accomodate the need to look in more than one keymap.
Alternatively, you acn create a single default method to do keymap
lookup and it will dispatch to all appropriate methods / dicts of each
class involved that can then be written without bothering with
cooperation.

Hooks are just methods that can be implemented by subclasses.  Again,
you need multiple dispatch.  I really think that python's standard
behavior of attribute lookup won't fit all cases.  Emacs is full of
these things.  Every second library has some multiple dispatch
mechanism implemented.

Frequently the best approach is to define the cooperation policy in
the top class and write multiple plugins managed by this policy in the
subclasses (a different name is needed to avoid the shadowin, unless
you want to do metaclasses which will not play nice with dynamic class
changes I list below.  Some library of common policies might be
useful.

Now the issue is that the class - nay, the whole mro - of a buffer has
to be adjustable dynamically, e.g. when you switch modes.
Fortunately, Python allows class and basess changes.  This bringing of
many classes together is probably what Carl Banks talked about.  E.g.
minor modes are mixins - ones that you turn on and off on the fly!

If you want to closely mimick emacs, you can have *all* classes merged
into one emacs object.  Then you have one namespace and all knids of
buffer-local, etc things can be implemented.  Internally there aree
separate buffer, window, etc. objects but they all are exposed and
should be used through the same object.  So you access a property or
method (`self.paragraph_start`, `self.mark_function`) without knowing
whether it's hadled by the main global class, the current buffer's
class of anything else...

This sounds like a fun abuse of Python classes but I'm not sure it's
the right direction.  I think I'd rather see many objects.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>
When I grow up I want to write an Emacs.




From dwelch91 at no.spam.attbi.com  Sun Apr 13 18:24:38 2003
From: dwelch91 at no.spam.attbi.com (djw)
Date: Sun, 13 Apr 2003 22:24:38 GMT
Subject: resizeable windows in pyqt
In-Reply-To: <B8jma.1829$WY.308@newsfep3-gui.server.ntli.net>
References: <B8jma.1829$WY.308@newsfep3-gui.server.ntli.net>
Message-ID: <Gslma.469390$S_4.529751@rwcrnsc53>

Tom Chance wrote:
> Hullo,
> 
> I know this sounds really daft, but I can't get my windows to be nicely
> resizeable in PyQt. I'm using Qt Designer to make the GUI, and I start with
> a MainWindow project, without any of the toolbars/menus (so a blank slate).
> I then add the various widgets in grids, and when I preview it, the window
> resizes but none of the widgets resize with it. I've set the sizePolicy as
> "preferred", and played around with other likely looking options, but I
> can't work it out! How do I do it?
> 
> Thanks,
> Tom
You need to layout your controls in the window using a layout control. 
Try this: with your window with all of its controls showing, click on 
the window itself (that is, not on any control). Then, click on the grid 
layout (block of 9 blue squares). Hit Ctrl-T and try it out: it should 
now resize the controls in the window as the window changes size. 
Understanding the layout controls (there are several others), is 
critical to successfully creating usable GUIs in Qt/PyQt. I would 
recommend that you take a look at one of the Qt programming books and/or 
online help.

/d/




From cl.venturini at virgilio.it  Wed Apr  9 16:28:31 2003
From: cl.venturini at virgilio.it (Claudio Venturini)
Date: 9 Apr 2003 13:28:31 -0700
Subject: wxPython: wxWindow $ LoadFromResource method
Message-ID: <9f93ec15.0304091228.66159416@posting.google.com>

I don't understand how I must use the LoadFromResource method of a
wxWindow window. The wxWindows reference seems very confused.

Someone can help me??

Claudio



From amk at amk.ca  Mon Apr 14 13:56:00 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Mon, 14 Apr 2003 12:56:00 -0500
Subject: Cryptography (was Re: How to protect python code ?)
References: <3E8BE455.4000904@novagrid.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com> <OIqdnV1p3aljSw-jXTWcpg@comcast.com> <b7bvft$os4$1@panix1.panix.com> <b7ecno$n7m@newton.cc.rl.ac.uk>
Message-ID: <R7qdnYcmzPGtawejXTWcrg@speakeasy.net>

On Mon, 14 Apr 2003 14:23:03 +0100, 
	Richard Brodie <R.Brodie at rl.ac.uk> wrote:
> In general, yes. Whether experts working in classified establishments can't
> do better than the public state of the art is much harder to quantify.
> Consider the history of PKCS, for example.

To expand on this, someone working for the UK's GCHQ invented the idea of
public-key cryptography in 1973 but couldn't publish it; some of his
co-workers soon invented actual algorithms for implementing it, including
the one now called RSA.  Diffie and Hellman independently invented the idea
and published it in 1977.  (See Steven Levy's book _Crypto_ for a readable 
discussion of this history.)

I doubt that government establishments such as the NSA or GCHQ are miles
ahead of the public sector, though.  Note that the public invention of
public-key was only four years behind its classified invention.  In 1998
three cryptographers found an attack against a version of NSA-designed
Skipjack algorithm that used one fewer round of scrambling; it's unlikely
the NSA knew about this technique, or the cipher would have had more rounds.  

Classified establishments certainly have more money to spend, so I expect
they have big computers and custom-designed chips for trying keys, but I
doubt they have a super factoring algorithm or powerful attacks against 
block ciphers.

--amk                                                    (www.amk.ca)
FLUTE: Nay, faith, let me not play a woman; I have a beard coming.
      -- _A Midsummer Night's Dream_, I, ii



From anton at vredegoor.doge.nl  Mon Apr 21 09:02:31 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 21 Apr 2003 15:02:31 +0200
Subject: __setitem__ puzzle
Message-ID: <b80q7c$cmd$1@news.hccnet.nl>

I have subclassed the builtin list type to accept tuples for indexing.
I am just experimenting with the code to see what I can do with it, so
asking for what the code will be used for in order to provide a better
way to do it will be answered with 'ni' :-)

There's a problem with __setitem__ however. What's the proper way to
write this class?

Anton

class tlist(list):
    
    def __getitem__(self,x):
        y = self[:]
        for z in x:
            y = y[z]
        return y

    def __setitem__(self,x,val):
        y = self[:]
        for z in x[:-1]:
            y = y[z]
        y[x[-1]] = val

def test():
    L = tlist(['aa',['bb',['cc','dd'],'ee'],'ff',['gg']])
    indices = [(0,), (1,), (1, 0), (1, 1), (1, 1, 0),
                (1, 1, 1), (1, 2), (2,), (3,), (3, 0)]
    for x in indices:
        print L[x]
    print
    L[(1,1,0)] = 0
    print L
    L[(1,1)] = 0
    print L
    #this doesn't work
    L[(1,)] = 0
    print L

if __name__=='__main__':
    test()




From grahamd at dscpl.com.au  Wed Apr  2 22:16:59 2003
From: grahamd at dscpl.com.au (Graham Dumpleton)
Date: 2 Apr 2003 19:16:59 -0800
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
Message-ID: <dc6f5c99.0304021916.15213208@posting.google.com>

escalation746 at yahoo.com (robin) wrote in message news:<a626bbd7.0304020829.1e5cac2f at posting.google.com>...
> I am attempting to summarise the distributed computing implementations
> available to Python programmers. My conclusions so far are as follows:
> 
> If you want the biggest, boldest approach and don't care about
> overhead, use CORBA.
> 
> If you want a simpler approach but with performance and implementation
> difficulties, use SOAP or XML-RPC.
> 
> If you want the ultimate in simplicity and are willing to foresake
> multi-language support, use Dopy, Pyro, or Twisted Spread.
> 
> Comments? Specifically, what are the relative advantages and
> disadvantages of the last three mentioned products?
> 
> -- robin

You might want to consider looking at OSE as well. It isn't pure Python
but then that can be a good thing depending on what you are doing.

The web site for OSE is:

  http://ose.sourceforge.net

Have a look through the "Python Manual" link on the web site.



From lost at soul.net  Mon Apr 14 13:32:02 2003
From: lost at soul.net (lost)
Date: Mon, 14 Apr 2003 18:32:02 +0100
Subject: Reading/writing files
Message-ID: <pan.2003.04.14.17.32.01.209427@soul.net>

Hi there,
I'm a bit lost in terms of reading and writing files.
I've been reading the python docs that came with my distibution, and they
aren't too clear.
I need something like this

for line in a file
	print line
print done

and for writing, I've tried doing file.write(blah) (blah is a list), but
it wont write.
I need to write lists into a file, but i cant.  are there other methods
for printing to a file, other than file.write?
Also, do i need to take into account the \n at the end of a line?
for example, if I have
hello:there
and
hello:there\n
how could i ignore the \n?

cheers!
	



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Sat Apr 26 08:03:44 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Sat, 26 Apr 2003 14:03:44 +0200
Subject: RuntimeError at end of test.autotest in Python 2.3b on windows, tmpfile
 also fails
Message-ID: <3eaa75a0$0$49106$e4fe514c@news.xs4all.nl>

Hello
when running "import test.autotest" in Python2.3b on windows xp,
I get a RuntimeError at the end:

209 tests OK.
2 tests failed:
     test_imp test_os
42 tests skipped:
     test_aepack test_al test_bsddb3 test_cd test_cl test_commands
     test_crypt test_curses test_dbm test_dl test_email_codecs
     test_fcntl test_fork1 test_gdbm test_gl test_grp test_imgfile
     test_ioctl test_largefile test_linuxaudiodev test_macfs
     test_macostools test_mhlib test_mpz test_nis test_normalization
     test_openpty test_ossaudiodev test_plistlib test_poll test_posix
     test_pty test_pwd test_resource test_scriptpackages test_signal
     test_socket_ssl test_socketserver test_sunaudiodev test_timeout
     test_timing test_urllibnet
Those skips are all expected on win32.
Traceback (most recent call last):
   File "<string>", line 1, in ?
RuntimeError: not holding the import lock


What's up?


test_imp failed like this:
test test_imp failed -- Traceback (most recent call last):
   File "C:\Python23\lib\test\test_imp.py", line 22, in testLock
     raise TestFailed, \
TestFailed: release_lock() without lock should raise RuntimeError

Also, test_os fails:
ERROR: test_tmpfile (__main__.TemporaryFileTests)
-----------------------------------------------------------
Traceback (most recent call last):
   File "lib\test\test_os.py", line 49, in test_tmpfile
     fp = os.tmpfile()
OSError: [Errno 13] Permission denied



I'm running this on a regular user account on Windows XP.

--Irmen de Jong




From chasm at rift.sytes.net  Tue Apr  1 08:58:27 2003
From: chasm at rift.sytes.net (Julian Tibble)
Date: Tue, 1 Apr 2003 13:58:27 +0000 (UTC)
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift> <mailman.1049181423.2644.python-list@python.org>
Message-ID: <slrnb8j6o1.o6a.chasm@galileo.rift>

In article <mailman.1049181423.2644.python-list at python.org>, Jp Calderone wrote:
>   It looks like this could easily be improved with the application of a
> slightly smarter searching algorithm, such as Rabin-Karp search or
> Boyer-Moore find.
> 
>   Both of these use a technique of doing a little extra work up front to
> make each pass through the loop much less expensive.  Google for details.

I can't see why.  Maybe for more complex patterns, but not for just finding
words.

All you have to do to match words is go through the file character by
character with a simple finite state machine.  The C snippet I used in
my comparison program was as follows:

/* C extract to count occurences of words
 * ht = hash table handle
 * t is one past the last character (set to '\0' in case there is a
 *                                   word right at the end)
 */

for (i = 0; i <= t; i++) {
        if (isalpha(buf[i])) {
                if (!in_word) {
                        in_word = 1;
                        start = &buf[i];
                }
                buf[i] = tolower(buf[i]);
        } else {
               if (in_word) {
                       buf[i] = '\0';
                       in_word = 0;
                       hash_countword(ht, start);
               }
        }
}

This is clearly O(n)

Rabin-Karp pattern search algorithm is an alternative to trying to match
a pattern at every position, which I'm not doing.


Julian



From aleax at aleax.it  Mon Apr 14 03:12:05 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 07:12:05 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <3E98D281.8C0CE87D@easystreet.com> <b7djfa$deo4s$3@ID-169208.news.dfncis.de>
Message-ID: <9btma.13658$T34.441487@news2.tin.it>

Greg Ewing (using news.cis.dfn.de) wrote:

> achrist at easystreet.com wrote:
>> Now, which is preferred:
>> 
>> if astring == "":
>> 
>> or
>> 
>> if len(astring) == 0:
> 
> The first ought to be faster, since it avoids a
> global name lookup and a Python function call,
> both of which are fairly expensive.
> 
> But better still would be simply
> 
>    if astring:
>      ...

...which, however, has exactly the opposite truth value.
Sometimes that will be OK (e.g., whenever you also have
an else clause and switching the two bodies does not
diminish readability), but not always.  You can use
"if not astring:" to get the same truth value as
"if len(astring)==0:" more directly.  I doubt speed
considerations are very significant at all here, though.


Alex




From cpl.19.ghum at spamgourmet.com  Tue Apr  1 13:11:13 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Tue, 1 Apr 2003 20:11:13 +0200
Subject: Problem: 1040, too many connections
References: <3e899a43$0$21745$9b622d9e@news.freenet.de>
Message-ID: <Xns9350CD870CCD9cpl19ghumspamgourmet@62.153.159.134>

Susanne,

> we do a mysql-Connection (Mysql 3.23.41, MySQL-python-0.9.2) from
> python scripts in ZOPE under SUSE Linux 7.3.
> And we have no problems.

> Now we set up a new server with Mysql 4.0.12 instead of 3.23.41. And
> since that we get the problem "1040, too many connections" when we
> start a python script which perform a lot of SQL-Statements.

Maybe the default number of connections with 4.0.12 is smaller? Look in 
mysql.conf

Greetings

Harald



From wijgerden at hetnet.nl  Wed Apr  2 06:52:42 2003
From: wijgerden at hetnet.nl (vwa)
Date: Wed, 02 Apr 2003 13:52:42 +0200
Subject: Mimic a web surfer ?
References: <3e894b7d$0$29158$626a54ce@news.free.fr>
Message-ID: <3E8ACF0A.62CEB94A@hetnet.nl>

On http://fvw.freezope.org/wiki/Testing you can find some of my links,
one of which is MaxQ (see http://maxq.tigris.org/).

[from the MaxQ site:] MaxQ is a free web functional testing tool. It
includes an HTTP proxy that records your test script, and a command line
utility that can be used to playback tests. The proxy recorder
automatically stores variables posted to forms, so you don't have to
write that stuff by hand. The paradigm of MaxQ is similar to commercial
web testing tools like Astra QuickTest or Empirix e-Test. These products
are quite expensive. MaxQ hopes to provide the essential features: HTTP
test recording, scripting, and playback without the huge cost. It uses
Java for the HTTP proxy and recording, Python for the script, and
Java/Jython/Python for the playback.



From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Fri Apr 11 12:59:45 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Fri, 11 Apr 2003 18:59:45 +0200
Subject: distributed computing implementations
In-Reply-To: <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <d116fbae.0304100704.68ef4833@posting.google.com> <c48$3e960655$516049d2$3498@nf1.news-service.com> <3e96efac$0$49107$e4fe514c@news.xs4all.nl> <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk>
Message-ID: <3e96f481$0$49099$e4fe514c@news.xs4all.nl>

Robin Becker wrote:
> In article <3e96efac$0$49107$e4fe514c at news.xs4all.nl>, Irmen de Jong
> <irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> writes
>>To add some more to the mix, I benchmarked Pyro (3.2) :
[...]
> .....was this on the same machine? I find things seem to be awful off
> the machine even if I pass in a dotted numeric address. I'm getting the
> idea that I'm doing name lookups even if there's no apparent need. 

Yep that was on the same machine. Running the Pyro server on a different
machine on my 100 Mbit lan gives:

Connecting to server
Time to connect to server, 0.024740

Sending a long string to the server
Time to send a string of 21000 chars, 0.003866

Recieving a long stirng from the server
Time to receive a string of 22000 chars, 0.003579

Sending lots of ints to the server
Time to send 5000 ints, 4.586772 (0.000917 per call)

So you see, it's even faster when the server is running on another machine. 
This can be explained because the pickling/unpickling is split across two 
CPUs, when you run client+server on a single CPU they are fighting for cycles.

--Irmen de Jong




From selwyn at home.net.nz  Tue Apr 22 06:11:03 2003
From: selwyn at home.net.nz (selwyn)
Date: Tue, 22 Apr 2003 22:11:03 +1200
Subject: problem starting python interpreter in emacs (win98)
In-Reply-To: <b831pv$2d65$1@serv.vrn.ru>
References: <V65pa.4289$AB5.573029@news02.tsnz.net> <b831pv$2d65$1@serv.vrn.ru>
Message-ID: <Kq8pa.4362$AB5.585892@news02.tsnz.net>

Lexy Zhitenev wrote:
> "selwyn" <selwyn at home.net.nz> wrote in message:
> news:V65pa.4289$AB5.573029 at news02.tsnz.net...
> 
>>system: win98
>>python version 2.2.2
>>emacs  21.2.1
>>..........
>>hi all,
> 
> 
> <snip>
> 
>>My .emacs file is as follows:
>>
>>(setq load-path (cons "C:\Python22" load-path))
> 
> 
> 
> Not sure about ELisp, but try to use escaped backslash:
> 
> (setq load-path (cons "C:\\Python22" load-path))
> 
> 
> Regards, Lexy
> 
> 

cheers for the advice.
however it doesn't seem to help my problem- Emacs seems to be able to 
find the python-mode.elc ok, as evidenced by the syntax highlighting 
working...

cheers again,
Selwyn




From duncan at NOSPAMrcp.co.uk  Thu Apr 24 09:19:02 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 13:19:02 +0000 (UTC)
Subject: sys.argv problem in Win 2k
References: <bd993a2f.0304240121.7c795cf0@posting.google.com> <mailman.1051177117.30168.python-list@python.org> <bd993a2f.0304240506.5b206f3d@posting.google.com>
Message-ID: <Xns936790B52180Eduncanrcpcouk@127.0.0.1>

abpillai at lycos.com (Anand B Pillai) wrote in 
news:bd993a2f.0304240506.5b206f3d at posting.google.com:

>  As you see the arguments I pass are not being printed!
> This is the case with all my scripts now. They were working
> well when I was on a WinNT workstation :-?

What does the command 'assoc .py' print?
What does the command 'ftype Python.File' print?

Running a .py file depends on the file association which the assoc command 
will show, and if it is Python.File the ftype command will show how the 
command line is passed.

They should look something like this:

D:\temp>ftype Python.File
Python.File=D:\Python22\python.exe "%1" %*

D:\temp>assoc .py
.py=Python.File

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From gh at ghaering.de  Tue Apr 15 21:33:28 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: Wed, 16 Apr 2003 03:33:28 +0200
Subject: =?iso-8859-1?Q?R=E9ception_et_d=E9coda?=
	=?iso-8859-1?Q?ge?= d'une trame UDP
In-Reply-To: <3E9CA7C2.2C9730EA@engcorp.com>
References: <3e9c9b1e$0$1509$7a628cd7@news.club-internet.fr>
	<3E9CA7C2.2C9730EA@engcorp.com>
Message-ID: <20030416013328.GA29065@mephisto.ghaering.test>

* Peter Hansen <peter at engcorp.com> [2003-04-15 20:45 -0400]:
> Benoit BESSE wrote:
> > 
> > Bonsoir tout le monde,
> > 
> > qui a d?j? ?crit un module qui ?coute, re?oit, d?code et insert dans une BD
> > une trame UDP.
> > Je voudrais d?coder la trame UDP ainsi et l'ins?rer dans une table :
> >     - Port Source
> >     - Port Destination
> >     - Longueur
> >     - Somme de contr?le
> >     - Donn?es
> > 
> > Merci par avance.
> 
> Hmm... almost looks like another "Twisted has that" but I'm 
> not sure what "BD" is so I can't be sure...  ;-)

BD = base donn?es = database.

Of course, twisted has that (twisted.enterprise.adbapi,
twisted.enterprise.row) ;-)

The original poster might also want to have a look at a packet sniffer
like tcpdump - looks like tcpdump can be talked to do pretty much all of
that, except the database interface.

Btw. the usual language on this group is English, but there *is* a
french Python newsgroup: fr.comp.lang.python.

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/




From ianb at colorstudy.com  Sun Apr  6 14:30:05 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 06 Apr 2003 13:30:05 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
	 <mailman.1049571010.25625.python-list@python.org>
	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
Message-ID: <1049653805.659.45.camel@lothlorien>

On Sat, 2003-04-05 at 22:32, Neil Schemenauer wrote:
> Ian Bicking <ianb at colorstudy.com> wrote:
> > The next step I can envision would be for Twisted to become the
> > canonical infrastructure, where higher-level web frameworks are all
> > built ontop of it.
> 
> That's not feasible.  Some applications would require extensive
> modification to work on top of an event driven framework.

That was my initial impression too, but you don't have to program in the
Twisted/async style in order to use Twisted as a base -- Twisted can
spawn threads and wait for responses from those threads.  I've been
thinking about this for Webware, and I don't think there would need to
be any change to user's code (and actually only very minimal changes to
Webware's code -- mostly just deleting the parts that are redundant with
Twisted and making a few wrappers).

It does change the installation procedure and some of the management --
which is okay in Twisted but not great.  However, none of the frameworks
are great, and if we consolidate we have a better chance at improvement.

  Ian






From news at datatailors.xs4all.nl  Thu Apr 17 03:21:41 2003
From: news at datatailors.xs4all.nl (Peter van Kampen)
Date: Thu, 17 Apr 2003 09:21:41 +0200
Subject: HTML form pre-populating functionality or hints on building my own
References: <mailman.1050550654.3607.python-list@python.org>
Message-ID: <slrnb9slg5.ruu.news@datatailors.com>

In article <mailman.1050550654.3607.python-list at python.org>, Scott
Chapman wrote:
> Hi!
> I'm wanting to find (or make) a module that will take a HTML file, locate the 
> input-related form tags (<INPUT, <TEXTAREA, <SELECT, <OPTION) and modify them 
> by putting in values from a dictionary (i.e. pre-populate the form). I want a 
> module that is implemented 100% in python and does _not_ use any special tags 
> to do this (such as webware, albatross, spyce, etc.)

100% pure python...but it doesn't really 'locate' form-tags...

--------- cut here ---------------
#!/usr/bin/env python

templatevars = {}
templatevars['title'] = 'Demo'
templatevars['meta'] = ''
templatevars['stylesheet'] = 'demo.css'

formvars = {}
formvars['username'] = 'Scott'
formvars['password'] = ''

templatevars['body'] = """
    <div class="fancy">Python rocks!</div>
    <br />
    <form>
    <input type="text" name="username" value="%(username)s" />
    <input type="password" name="password" value="%(password)s" />
    <input type="submit" name="submit" value="submit" />
    </form>
""" % (formvars)
            
template = """
    <html>
    <head>
    <title>%(title)s</title>
    %(meta)s
    <link rel="stylesheet" href="%(stylesheet)s">
    </head>
    <body>
    %(body)s
    </body>
    </html>
""" % (templatevars)
            
print template

--------- cut here ---------------

> The only thing I've seen that does this is EmbPerl (I'm tired of Perl) and 
> Integratis' DFP module (implemented in C and not a straight-forward install).

I don't know EmbPerl (nor do I want to ;-)) but if the above isn't
what you mean maybe you can try to subclass the new HTMLParser module
that comes with python 2.2. It's surprisingly (to me anyway) easy to
use. You can feed it plain HTML (ie. read from a file or from urlopen)
and it will call handle_starttag functions where you can check for
input tags. It passes the attributes as a list of tuples so you have
to do little (or no) parsing yourself.

Hth,

PterK

-- 
Peter van Kampen
pterk -- at -- datatailors.com



From jarausch at skynet.be  Fri Apr 25 11:05:18 2003
From: jarausch at skynet.be (Helmut Jarausch)
Date: Fri, 25 Apr 2003 17:05:18 +0200
Subject: Linux SUID
Message-ID: <3ea94eae$0$23154$ba620e4c@reader0.news.skynet.be>

Hi,
upto now my favourite scripting language is (was?) Perl.
There one can turn on the SUID-exec permission bit which
is honoured by Perl.
In this situation Perl enters a special "taint" mode where
anything from the outside world like the environment or anything
read in is tainted. If one tries to use a tained variable in
critical operation like 'exec' or 'open' Perl refuses to do so.
So this mechanism is safer than write an SUID program in C/C++ .

Is there something similar in Python?

Please no lectures on the riscs of using suid programs.

Many thanks for a hint,

Helmut Jarausch

Lehrstuhl fuer Numerische Mathematik
RWTH - Aachen University
D 52056 Aachen, Germany




From mathias-usenet at valpo.de  Wed Apr 23 05:35:11 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 23 Apr 2003 09:35:11 GMT
Subject: Problem Cron-Job using Python-Script
References: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de> <b85lrk$62480$1@ID-69142.news.dfncis.de>
Message-ID: <b85mof$6ksni$1@ID-15245.news.dfncis.de>

"Andreas Penzel" <mail at andreas-penzel.de> wrote:

> os.system("iptables -L -v > /path/logfile.txt")

Is iptables in the PATH? Try calling iptables with the full path. And 
maybe try capturing stderr.

> The file logfile.txt in /path is after running the script empty.
> Is there a other way to become the result from iptables -L -v in
> a file? 

Yes, look at the popen functions. 

Mathias



From aahz at pythoncraft.com  Sun Apr 27 02:22:44 2003
From: aahz at pythoncraft.com (Aahz)
Date: 27 Apr 2003 02:22:44 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com>
Message-ID: <b8fsvk$kkg$1@panix2.panix.com>

In article <698f09f8.0304262115.386c2571 at posting.google.com>,
Jeremy Fincher <tweedgeezer at hotmail.com> wrote:
>
>Perhaps the answer is to remove the __nonzero__ method on
>functions/methods.  Logically, what purpose does it serve?

"I see no __nonzero__ method here."

Anyone wanna guess what quote I'm mangling?  ;-)

Anyway, all objects that do not implement a __nonzero__ method are true:

>>> class C: pass
...
>>> x = C()
>>> print not(not x)
1
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From imbosol-1050185957 at aerojockey.com  Sat Apr 12 18:35:37 2003
From: imbosol-1050185957 at aerojockey.com (Carl Banks)
Date: Sat, 12 Apr 2003 22:35:37 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <65pjlhdx.fsf@morpheus.demon.co.uk>
Message-ID: <Zw0ma.4690$883.4089@nwrdny02.gnilink.net>

Paul Moore wrote:
> Mads Orbesen Troest <mads at troest.NEVERMORE.dk> writes:
> 
>> Given that strings are immutable, is it exceedingly slow (and memory 
>> spiking) to do a lot of "string += string" operations in one's code? IOW, 
>> does each and every += evaluation lead to a new concatenated copy of the 
>> previous string (now freed for garbage collection) and the new string, so 
>> that the longer the string to which stuff is appended is, the longer times 
>> each += operation takes?
> 
> I don't know the underlying details too well (and I'm not sure they
> are all that important) but yes, string += string does construct a new
> string each time, and hence is relatively expensive.
> 
> The normal idiom for building a string in chunks is to create a list
> of the chunks, and then join them once at the end. So you do:
> 
>    chunks = []
>    while whatever():
>        chunks.append(next_bit_of_string())
>    result = ''.join(chunks)
> 
> You could also use StringIO. However, all this does is basically what
> I describe above, plus extra complexity to handle all the methods you
> don't need. So you're better sticking with the above idiom.


According to my benchmark, cStringIO is a bit faster than joining a
list of strings when appending one character at a time (as opposed to,
say, one line at a time, where it appears to be slightly slower).
Also, it certainly uses less memory than a list of strings, so I'd say
sometimes it is better to use cStringIO.

Surprisingly, += doesn't appear to be a major time waster for small
(meaning about 10000 bytes) strings.  I'm guessing it's because the
malloc my system (Debian Linux) is optimized for small memory
allocations.

    from cStringIO import StringIO
    from time import time
    import sys
    
    def benchmark(n,m):
        c = 'c'*m
        
        start = time()
        s = ''        
        for i in xrange(n):
                s += c
        end = time()
        print "+=:         %.3f seconds" % (end-start)

        start = time()
        l = []
        for i in xrange(n):
                l.append(c)
        end = time()
        s = ''.join(l)
        print "list:       %.3f seconds" % (end-start)

        start = time()
        f = StringIO()
        for i in xrange(n):
                f.write(c)
        s = f.getvalue()
        end = time()
        print "cStringIO:  %.3f seconds" % (end-start)
    
    benchmark(int(sys.argv[1]),int(sys.argv[2]))


On my system, with n=10000, m=1:
+=:         0.041 seconds
list:       0.023 seconds
cStringIO:  0.019 seconds

With n=100000, m=1:
+=:         8.002 seconds
list:       0.239 seconds
cStringIO:  0.187 seconds

With n=10000, m=40:
+=:         19.681 seconds
list:       0.022 seconds
cStringIO:  0.026 seconds

I'm running Python 2.1.3 on Linux.


-- 
CARL BANKS





From anthony at interlink.com.au  Tue Apr 29 00:18:12 2003
From: anthony at interlink.com.au (Anthony Baxter)
Date: Tue, 29 Apr 2003 14:18:12 +1000
Subject: Dynamic repeat counts in struct format strings? 
In-Reply-To: <varrqma1u6bocb@corp.supernews.com> 
References: <varrqma1u6bocb@corp.supernews.com>  
Message-ID: <200304290418.h3T4ICv09527@localhost.localdomain>

>>> "Francis Avila" wrote
> I suppose this message is just wondering aloud, but:
> 
> Why hasn't the struct module grown a feature to specify the repeat count of
> an element based on the value of the previous element, like perl's
> (http://www.perldoc.com/perl5.6/pod/func/pack.html)?
> 
> What I mean is, why can't I do something like this (borrowing perl syntax of
> '/'):
> 
> >>> struct.unpack('!h/s', '\x00\x05Hello')
> ('Hello',)

I like this, but I have to wonder if something a bit more general
purpose would be better? What about the cases where the length isn't
immediately before the string? Why not instead have something like 
the named groups in REs? 

Anthony
-- 
Anthony Baxter     <anthony at interlink.com.au>   
It's never too late to have a happy childhood.





From member27735 at dbforums.com  Wed Apr  9 00:28:36 2003
From: member27735 at dbforums.com (nitintz)
Date: Wed, 09 Apr 2003 04:28:36 +0000
Subject: Importing problem
References: <2737271.1049712504@dbforums.com>
Message-ID: <2745665.1049862516@dbforums.com>

Actually i need to import a Qt project (Its a PyQt Project where i need
to execute a Qt project and do some manipulations on it).

--
Posted via http://dbforums.com



From ckea25d02 at sneakemail.com  Tue Apr  1 20:02:58 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 13:02:58 +1200
Subject: Questions on linked lists
In-Reply-To: <3E8A1189.3F806B4B@alcyone.com>
References: <kl8ia.27315$OV.127379@rwcrnsc54> <5vonl-vm3.ln1@beastie.ix.netcom.com> <slrnb8ja7a.421.ramen@lackingtalent.com> <3E8A1189.3F806B4B@alcyone.com>
Message-ID: <b6dccl$49i5m$1@ID-169208.news.dfncis.de>

Erik Max Francis wrote:
> A linked list, although it's
> expensive to index [O(n)], is very inexpensive to insert or delete
> objects at _any_ location [O(1) for either].
> 
> If you find you need a data structure where you are doing a lot more
> inserting and deleting than indexing, then a linked list might well make
> a great deal more sense.

Be careful, though -- chances are, in Python you'd
need quite a large list before the overhead of the
Python operations needed to create a node and insert
it became less than the C-speed operation of inserting
into a Python list.

The O(x) statements above are true, but there are
some pretty hefty constant factors involved here!

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ckea25d02 at sneakemail.com  Wed Apr 16 20:57:51 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 17 Apr 2003 12:57:51 +1200
Subject: ANN:  Pyrex 0.7.2
Message-ID: <mailman.1050619204.11008.clpa-moderators@python.org>

Pyrex 0.7.2 is now available:

    http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/

This release fixes another couple of bugs, and also
supplies the Demo/callback/cheesefinder.c file that got
missed out before.

What is Pyrex?
--------------

Pyrex is a new language for writing Python extension modules.
It lets you freely mix operations on Python and C data, with
all Python reference counting and error checking handled
automatically.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg





From martin at v.loewis.de  Mon Apr 21 12:24:25 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Mon, 21 Apr 2003 18:24:25 +0200
Subject: _socket problem
In-Reply-To: <b80ves$52ova$1@ID-181029.news.dfncis.de>
References: <b80ves$52ova$1@ID-181029.news.dfncis.de>
Message-ID: <3EA41B39.5060204@v.loewis.de>

Jorl Shefner wrote:

>    Is there any known workaround for this problem, or does it require a 
> re-installation of python.

A reinstallation of Python is necessary. The administator should 
carefully watch the messages that are produced when the socket module is 
compiled. If the messages are inconclusive, the administrator should 
edit Modules/Setup to incorporate the socket module into Python, making
an explicit choice as to whether and how to include SSL support.

Regards,
Martin

P.S. The above is, strictly speaking, not true. You can build just
the socket module separately. But if you knew how to do that, you
probably wouldn't have asked, and it is not trivial to do so.




From mads at troest.NEVERMORE.dk  Fri Apr 11 09:52:38 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Fri, 11 Apr 2003 15:52:38 +0200
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net>

On Fri, 11 Apr 2003 10:48:49 +0200, Harald Massa wrote:

> Yes, I love Python, and it is the most elegant programming language I 
> ever met.

I agree it's mostly elegant, but the OO features are clumsy in my oppinion 
(coming as a professional developer from other languages). Not as bad as 
Perl's bless-a-dictionary-into-being-a-class retro-fitting hack, of course, 
but really; all the self me here and self me there is - to me at last - 
exceedingly clumsy and a constant source of frustration.

Regards,
   /\/\\ads Orbesen Troest



From mhammond at skippinet.com.au  Tue Apr 29 18:56:43 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Tue, 29 Apr 2003 22:56:43 GMT
Subject: Calling dotNet components from Python.
In-Reply-To: <924a9f9c.0304282138.15aa7635@posting.google.com>
References: <924a9f9c.0304282138.15aa7635@posting.google.com>
Message-ID: <LqDra.7014$lD4.74951@news-server.bigpond.net.au>

Paradox wrote:

> I was wondering if it is possible to call a dotNet dll written in C#
> from within a Python script. If anyone knows of any documentation or
> example code that accomplishes something like this. It seems that you
> can call a com component but all my searching on dotNet has not found
> anything about using dotNet components in Python. Thanks

.NET objects can be accessed via win32com.  You may need to run 
"tlbexport" to create a typelib that COM can see.

Mark.




From nhodgson at eb2.net.au  Tue Apr 15 19:47:19 2003
From: nhodgson at eb2.net.au (Neil Hodgson)
Date: Tue, 15 Apr 2003 23:47:19 GMT
Subject: Using property in classic class may not work
In-Reply-To: <hRUma.23848$T34.709378@news2.tin.it>
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it> <a8b7f07a.0304150527.5b0b1187@posting.google.com> <hRUma.23848$T34.709378@news2.tin.it>
Message-ID: <3E9C99E9.9000201@eb2.net.au>

[With some editing to isolate the points]

[Neil Hodgson]
  The class in question is derived from a wxWindows class, so I can't
make it derive just from object. The "What's New in Python 2.2"
document only mentions property wrt new-style clases.

[A. Lloyd Flanagan]
However, the last time I tried to subclass from both wxFrame and
object, wxPython freaked out (that's a technical term for "I don't
remember what happened, but boy, it sure didn't work").
Not sure about the latest version of wxPython, though -- there's
binaries for python 2.3 out by now.  I'll have to try it again..

[Alex Martelli]
If wxPython keeps freaking, you may have to give up on properties
if you HAVE to inherit its classes (can't just wrap with automatic
delegation) and go back to "good old" (yeah right) __setattr__.  Or,
gently pressure wxPython's maintainers to have them move to
new-style classes... whatever aspects of old-style classes they
may be taking for granted at present, migrating to new-style can
hardly be a major job, in my experience.

[Me again]
Robin may feel constrained to keep wxPython compatible with multiple 
Python major versions.
Here is the "freaking out" with wxPython+wxGTK 2.4.0.7 on Python 2.2 
caused by adding 'object' as a second base class to the most derived 
frame class, that is, the class that has the property:

Traceback (most recent call last):
   File "DemonTerm.py", line 1343, in ?
     app = DemonTermApp(sys.argv)
   File "DemonTerm.py", line 1290, in __init__
     wxApp.__init__(self, 0)
   File 
"/usr/local/ActivePython-2.2/lib/python2.2/site-packages/wxPython/wx.py", 
line 1802, in __init__
     _wxStart(self.OnInit)
   File "DemonTerm.py", line 1327, in OnInit
     self.frame = DemonFrame(BasicScreenDefs)
   File "DemonTerm.py", line 577, in __init__
     TerminalGUI.TerminalFrame.__init__(self, screenDefs)
   File "TerminalGUI.py", line 762, in __init__
     style=wxDEFAULT_FRAME_STYLE | wxNO_FULL_REPAINT_ON_RESIZE)
   File 
"/usr/local/ActivePython-2.2/lib/python2.2/site-packages/wxPython/frames.py", 
line 150, in __init__
     self._setOORInfo(self)
   File 
"/usr/local/ActivePython-2.2/lib/python2.2/site-packages/wxPython/windows.py", 
line 60, in _setOORInfo
     val = apply(windowsc.wxEvtHandler__setOORInfo,(self,) + _args, _kwargs)
TypeError: Type error in argument 1 of wxEvtHandler__setOORInfo. 
Expected _wxEvtHandler_p.

    OOR is wxPythonese for 'Original Object Return' but I don't know 
what that really means.

    Neil




From achim.domma at syynx.de  Mon Apr  7 02:14:23 2003
From: achim.domma at syynx.de (Achim Domma)
Date: Mon, 7 Apr 2003 08:14:23 +0200
Subject: PythonMagick 0.4
Message-ID: <b6r4n6$d7i$03$1@news.t-online.com>

Hello,

I have uploaded a new version of PythonMagick to
http://www.procoders.net/pythonmagick. PythonMagick is a wrapper around
GraphicsMagick (http://www.graphicsmagick.org). I have tested it on windows
and linux, and several bugs are fixed. Please let me know, if there are any
problems.

regards,
Achim





From romany at actimize.com  Mon Apr  7 07:38:10 2003
From: romany at actimize.com (Roman Yakovenko)
Date: Mon, 7 Apr 2003 14:38:10 +0300
Subject: Python & COM
Message-ID: <91BFE89EFFA2904E9A4C3ACB4E5F2DF5027AD0@exchange.adrembi.com>

	Hi. I built COM object with win32com library( excellent tool,
thank Mark ).
I have one problem: what I have to do in order to use this class from
VB?

Also an other question how can I distribute this COM object? 

Any references to the documentation will be appreciated.

Roman




From andrew-pythonlist at puzzling.org  Tue Apr 22 22:17:41 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 23 Apr 2003 12:17:41 +1000
Subject: Twisted 1.0.4
In-Reply-To: <1051031635.18957.24.camel@software1.logiplex.internal>
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local> <341410000.1051007646@leeloo.intern.geht.de> <20030422121836.GA8567@frobozz.local> <1051031635.18957.24.camel@software1.logiplex.internal>
Message-ID: <20030423021741.GA10754@frobozz.local>

On Tue, Apr 22, 2003 at 10:13:55AM -0700, Cliff Wells wrote:
> 
> On Tue, 2003-04-22 at 05:18, Andrew Bennetts wrote:
> > On Tue, Apr 22, 2003 at 12:34:06PM +0200, Marc Recht wrote:
> > > >to Twisted and Twisted's users.  Similarly, Twisted has some C extensions
> > > >which are typically built with gcc, but gcc's licence (GPL) also has no
> > > >relevance to Twisted.
> > > IIRC then gcc libs are LGPL'ed, so it doesn't affect Twisted. I wouldn't be 
> > > so sure if they were GPL'ed.
> > 
> > No, gcc is GPLed.  I'm not sure what you mean by "gcc libs"?  I'm pretty
> > sure libgcc1 is part of gcc and thus GPLed also.
> 
> No, it's LGPL'd.  GCC itself is GPL, the libraries are LGPL'd.

I stand corrected.  Thanks.  I couldn't find obvious evidence of this
browsing the docs on my local system, so I assumed it was under gcc's
licence.

> > The point is that a C program or library compiled with gcc is not a derived
> > work of gcc, anymore than a thesis written in and printed with MS Word is a
> > derived work of MS Word.  Python isn't GPLed, despite often being compiled
> > with gcc.
> 
> You are forgetting that by compiling your source with gcc, you are most
> likely linking against LGPL'd libraries (even in rather trivial
> programs).  That means that your executable will *contain* LGPL'd code. 
> The LGPL was created just for this purpose: to allow non-GPL'd software
> to link against (L)GPL'd software.

Using gcc doesn't require you to link against anything, let alone LGPLed
libraries -- although of course that is the most common situation; most
systems using gcc will also be using glibc.  But the libraries linked against
are a seperate issue to the point I was making about gcc; using gcc doesn't
necessitate using LGPLed libraries, and its the same situation as applies if
I were to use MS Visual C++ to compile a program linked against an LGPLed
library.

-Andrew.





From skip at pobox.com  Tue Apr 29 15:45:07 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 29 Apr 2003 14:45:07 -0500
Subject: Problem subclassing tuple
In-Reply-To: <F7DB8D13DB61D511B6FF00B0D0F062330489D915@mail.hebmex.com>
References: <F7DB8D13DB61D511B6FF00B0D0F062330489D915@mail.hebmex.com>
Message-ID: <16046.54851.525610.356475@montanaro.dyndns.org>

    >> You might want to make it "cooperative", that is, instead of using
    >> "tuple.__new__", use "super(Holder,cls).__new__"

    >> This delegates the search for the superclass to the function super();
    >> and that gives you the ability to have multiple inheritance, since
    >> super() will "do the right thing" when getting your current class'
    >> superclass; so every class which defines a __new__ method has it
    >> called.

I don't believe it matters here.  As I recall, builtins don't support
multiple inheritance very well (if at all).  The most you can do is add
mixin classes.

Skip




From bdesth at removethis.free.fr  Mon Apr 21 21:02:58 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 22 Apr 2003 01:02:58 +0000
Subject: *Newbie Question* --> Is it viable to distribute Python apps
 over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com>
Message-ID: <3EA494C2.5010701@removethis.free.fr>

Denis Agar wrote:
(snip)
> 
> I'm blabbering! Anyway, my problem is this. I tried to compile a
> simple Hello World program into something runnable on non-python
> computers, and it came out being 7MB in size. Is there something I'm
> doing wrong? Just so you know, I do know somewhat what I'm doing. I
> know that it's incorporating all of the GUI. So is there a way to not
> incoporate all of the gui, and still have a functional program? I'm
> thinkin in the 800k range for a simple program? The reason I ask is
> because I'm thinking of porting something I made in VB to Python
> before I develop it any further.
> 
> PS: The program is 92k in Visual Basic.

<OT>
Just for your information : your 92k VB program won't run without the 
vbrun.dll, which is at least 1.x Mo... To be sure to have a correct 
setup, you must package a VB program with the runtime (vbrun.dll) and 
any 'external' ActiveX component, type libs etcs used by the application 
(and be sure you won't write many real-life VB apps without ActiveX 
components and type libs etc)
</OT>

Laotseu




From skip at pobox.com  Tue Apr  8 12:27:44 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 8 Apr 2003 11:27:44 -0500
Subject: fdopen() question
In-Reply-To: <slrnb95s3m.u9b.news@cipw08.physik.uni-bonn.de>
References: <slrnb95s3m.u9b.news@cipw08.physik.uni-bonn.de>
Message-ID: <16018.63616.597967.470042@montanaro.dyndns.org>

    Carsten> Traceback (most recent call last):
    Carsten>   File "./listener.py", line 27, in ?
    Carsten>     f = os.fdopen(fd, 'r')
    Carsten> OSError: [Errno 9] Bad file descriptor

    Carsten> What am I missing?

If you print fd, is it a small integer?  Is there some place in the code
where os.close(fd) might have been called before you got ahold of fd?

Skip




From ben at hal.rescomp.berkeley.edu  Thu Apr 24 19:18:57 2003
From: ben at hal.rescomp.berkeley.edu (Benjamin Hartshorne)
Date: Thu, 24 Apr 2003 23:18:57 +0000 (UTC)
Subject: Exception and finally question
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com>
Message-ID: <b89rd1$2oij$2@agate.berkeley.edu>

Tung Wai Yip <tungwaiyip at yahoo.com> wrote:
> Hello,

> How to I access the exception object in a handler?

> try:
>     raise NameError, 'HiThere'
> except NameError:
>     #how to print the message 'Hi There' here?

try:
  raise NameError, "HiThere'
except NameError, whoshere:
  print whoshere, "how are you?"

-ben

-- 


-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Ben Hartshorne



From daniel.dittmar at sap.com  Tue Apr 15 05:16:01 2003
From: daniel.dittmar at sap.com (Daniel Dittmar)
Date: Tue, 15 Apr 2003 11:16:01 +0200
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org>
Message-ID: <b7gikh$s66$1@news1.wdf.sap-ag.de>

Lulu of the Lotus-Eaters wrote:
> This particular slowness is not merely academic.  It's bitten me.  In
> one of his posts, I think the Martellibot even said it bit him (in the
> mists of past time).  Probably most Python programmers fell into this
> trap at least once.  I got past it, and now usually use the
> "lst.append(string); "".join(lst)" solution.  But the idea of making a
> string longer by... well, adding more to it... seems really obvious.
> Almost the "one obvious way to do it."

Doesn't cStringIO exist for this purpose? Now that extension classes can be
subclassed, adding += and str should be enough for most purposes. Perhaps
it's even possible to add a general __getattr__ to forward all string
methods to the inner string (which doesn't exist and has to be created for
each such request).

Daniel






From boud at valdyas.org  Sun Apr  6 16:14:54 2003
From: boud at valdyas.org (Boudewijn Rempt)
Date: Sun, 06 Apr 2003 22:14:54 +0200
Subject: Python & Music: Performance question
References: <1759581e.0304060218.164d4cd4@posting.google.com> <V%Wja.5267$y3.728616@news010.worldonline.dk>
Message-ID: <3e908b18$0$211$e4fe514c@dreader3.news.xs4all.nl>

Max M wrote:

> It is premature optimisation. First you should just get it to work, and
> then you should optimize if there are any problems.
> 

Based on his examples, yes -- but if the possibility of using simple tuples
is taken into account, it becomes a design decision that should be made
fairly early. Once you've got a whole system that depends on notes to be
objects with lots of behaviour, it's not easy to go back. Conversely, once
you've got a system that's spread note behaviour all over the place like
nutella you've got a sticky mess that's not easy to clean up.

> No matter how many notes you have, it will make no problems for you on a
> reasonably moderne computer.

There's that too. I'd go for reasonably intelligent notes, myself.

-- 
Boudewijn Rempt | http://www.valdyas.org



From phil at river-bank.demon.co.uk  Fri Apr  4 18:01:19 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Sat, 5 Apr 2003 00:01:19 +0100
Subject: What are the differences between SWIG and Boost.Python?
In-Reply-To: <b6l0eq$tf0$1@main.gmane.org>
References: <169a7283.0304040821.9ad0976@posting.google.com> <b6l0eq$tf0$1@main.gmane.org>
Message-ID: <200304050001.19200.phil@river-bank.demon.co.uk>

SIP is used to wrap PyQt which is C++ and 300+ classes, just under 6000 
methods.

Phil

On Friday 04 April 2003 11:20 pm, Mike Rovner wrote:
> FWIW, I decided for myself:
>
> - wrap small C libs with Pyrex,
> - wrap big C libs with SWIG,
> - wrap small C++ libs with boost.python,
> - wrap big C++ with (I still have to do that ;)
>
> Mike
>
> PS.
> For me small is tens, big is hundreds and more
>
> "sdieselil" <sdieselil at yahoo.com> wrote in message
> news:169a7283.0304040821.9ad0976 at posting.google.com...
>
> > Hi
> >
> > I want to write C extension for Python. I know that there are 2 most
> > popular libraris for this purpose: SWIG and Boost.Python. What are the
> > differences between them? Which one is better for little extensions
> > and which is better for large-scale extensions?
> > --
> > http://mail.python.org/mailman/listinfo/python-list





From ramen at lackingtalent.com  Fri Apr 25 17:42:25 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Fri, 25 Apr 2003 21:42:25 -0000
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com>
Message-ID: <slrnbajav2.e5n.ramen@lackingtalent.com>

In article <3EA80827.6277D27C at engcorp.com>, Peter Hansen wrote:
> Dave Benjamin wrote:
>> Well, I think Dijkstra's basic point was that you can't really "unlearn"
>> anything. You can ignore what you've learned, but it's still there. Once
>> you've learned some BASIC, it's a part of your brain, and that imprint will
>> remain there FOR THE REST OF YOUR LIFE. But it's okay, it hasn't stopped me
>> either. Maybe someday we'll start a therapy group.
> 
> After having spent the last few months getting on average something like 
> five hours of sleep (my need is for more like 7.5 or 8), I'd have to say
> I'm a living rebuttal to your argument.  I have completely forgotten a
> variety of interesting things which I used to know.  More than that, I've
> forgotten *that* I used to know these things, and am going on the say-so
> of others whose memories are more intact.  And I strongly doubt I will 
> ever recover those memories, or even the memory of the memories.

Hehe... well, first of all, it's not necessarily my argument, but
Dijkstra's, or at least my rough approximation thereof. But I tend to agree
with Dijkstra on this one, so I'll let you argue with me anyway. ;)

Speaking as one who experiences memory loss for various reasons, I will
admit that there are certain memories which I can't seem to conjour up, even
though I have "learned" them in the past. However, that doesn't mean they're
gone. Perhaps they just weren't indexed properly. I have yet to learn how to
do a linear scan on my brain (anyone have a code snippet for that?)

Anyway, even if the *memory* is gone, the *effect* of learning is still a
part of us. What we know today is the direct result of a chain of causes and
effects which we'll never be able to retrace.

> I think if pathways in the brain can be built up, through repetition of
> stimuli, dreaming, etc., then they can likewise be dissolved, whether
> through drug use, pyschological problems or chemical imbalances caused 
> by lack of sleep, and so on.  Perhaps simple disuse leads only to an
> asymptotic decay which never quite reaches zero, as your argument might
> suggest, but I think other processes could act in different ways.

You might be right, but if I'm still suffering from BASIC when I'm 80, I
think I ought to be able to sue someone. =)

Peace,
Dave



From antonmuhin at sendmail.ru  Tue Apr  8 14:44:17 2003
From: antonmuhin at sendmail.ru (Anton Muhin)
Date: Tue, 08 Apr 2003 22:44:17 +0400
Subject: STRING TO LIST
In-Reply-To: <mailman.1049824753.10292.python-list@python.org>
References: <mailman.1049824753.10292.python-list@python.org>
Message-ID: <b6v5a0$27bh$1@news.peterlink.ru>

jubafre at brturbo.com wrote:
> Hello, 
> 
> I need to change a variable from string to list. 
> 
>>>>>x=str("[[12,13],[14,15]]") 
>>>>>x 
>>
>>'[[12,13],[14,15]]' 
>>
>>i have x and i want to change to a list, but not in this way: 
>>
>>
>>>>>x=list(x) 
>>>>>x 
>>
>>['[', '[', '1', '2', ',', '1', '3', ']', ',', '[', '1', '4', ',', '1', '5', ']', ']'] 
>>
>>must be a list like this: 
>>
>>
>>>>>x=[[12,13],[14,15]] 
>>>>>x 
>>
>>[[12, 13], [14, 15]] 
>>
>>someone know??? 
> 
> 
> 
> Juliano Freitas

Use eval:
PythonWin 2.2.1 (#34, Apr 15 2002, 09:51:39) [MSC 32 bit (Intel)] on win32.
Portions Copyright 1994-2001 Mark Hammond (mhammond at skippinet.com.au) - 
see 'Help/About PythonWin' for further copyright information.
 >>> s = str([[1, 2], [3, 4]])
 >>> s
'[[1, 2], [3, 4]]'
 >>> l = eval(s)
 >>> l
[[1, 2], [3, 4]]
 >>>

hth,
Anton.




From pauf at mtu-net.ru  Wed Apr 16 20:01:52 2003
From: pauf at mtu-net.ru (=?windows-1251?b?1uXt8vAg7uHz9+Xt6P8=?=)
Date: Thu, 17 Apr 2003 00:01:52 +0000
Subject: =?windows-1251?b?wu3o7ODt6P4g8PPq7uLu5Ojy5evl6Swg4fP14+Dr8uXw7uIsIP7w6PHy7uIhIGVi?=
Message-ID: <200304170001.ZTML71HX@pmxmdkpiwf.com>

?????????? ??? ???????? ?????????? ????????:

22 ?????? 2003 ???? - <????????????? ????????: ????????????? ???? ?
    ???????????????>,
23 ?????? 2003 ???? - <???????????? ?????????????>, ??????? ?????????  ?
    ?????????-???? ????????? ??????? ?????????? ???????? ????. 

??????????? ????????? ????????????? ????? ????????. ?????????? ????????
207-26-21, 789-81-90, 772-92-46.

????????? ????????  22 ?????? 2003 ????  - <????????????? ????????:
????????????? ???? ? ???????????????>:

1.  ????? ? ?????????? ??????????-???????? ??????. 
    ???????? ????? ?????? ? ?? ????????????? ??????????. 
    ???????????????? ?????? ? ?? ????????? ???????????
    ????????? ? ????????????? ????? ?????????????? ??????????  ???
???????? ??????.
  ???????????? ??????? ?????????.
2. ??????????? ?????????????? ????? ? ??????????????? ????????? ?????
  ?????????:
  * ?????-??????? (?????? ???????? ???? ?????????????;  ??????? ??????;
  ??????? ????? ????????? ? ??????? ?????????; ???????????? ??????  ?.?.);
  * ????  (??????????? ????????? ? ?????  ???? ? ??????? ??   ??? 9/99 ?
  ??? 10/99);
  * ?????? (??????? ?? ???????, ??????????? ? ?????????? ?????????????
  ????????? (??????? ???????????? ???????), ?????? ????????? ? ??????????
  ???);
  * ????????????? ??????????? ?????????? (?????); 
  * ?????? (??? ??????????? ?????? ?????????? ?? ????????????? ?
  ?????????????);
  * ?????????? ???????? ????? (??????????? ????? ???????? ?? ????????, ??
  ??????, ?? ??????? ? ?? ?????? ????? ?????);
  * ?????????????? ?????? (?????????, ????????, ?????????????, ???????? ?
  ????? ????????  ??? ?????????);
  * ???????????;
  * ?????? ???? ?????? (?? ??????? ?????????? ????? ???????? ????? ????
  ???????????  ????? ???? ?????? ? ?????????? ????????????? ?????????).

    ? ???? ????????  ????????????? ???????? ??????????????? ? ???? ???????:
????????; ?????????????; ?????????.
    ????????? ??????? ??????? ?????????????? ????????? ?? ???????-???????????
???????? ??? ??????? ???????????? ???? ??, ??? ? ??????????? ???????????
????? ??????? ??????.

    ?? ???????? ?????????:
    ???????? ????????? ??????????,  ?.?.?, ??????? ??????? ??? "????? ??",
??????  40 ???? ? ????? 200  ?????????? ??????  ?? ???????? ??????????????
?????,  ???????????????  ? ?????????????? ?????.

????????? ????????  23 ?????? 2003 ???? <???????????? ?????????????>:

1. ?????? ? ???????????????? ????????. 
2. ??????? ?? ???????????? ? ?????????? ??? ????? ?????????????? ????? 
  - ???????????? ???????
  - ??????? ?? ?????? ?????
  - ?????????? ?? ?????????? ?????
  - ???????????
  - ?????? ???????.
3. ???????, ????????? ? ????????????? ? ??????????? ??? ?????
??????????????? ???????
  - ???????????? ???????
  - ??????? ?? ?????? ?????
  - ???????????
  - ?????? ???????.
4. ???????, ??????? ? ???? ????????? ???????? ? ????????????? ?  ????????? ?????.
5. ??????????????? ????????????? ????????? (?????, ?????).
6. ??????? ???????? ??????????? ? ??????? ???????????? ?????????????.
7. ??????? ???????  ???????? ?????? ??? ??????????????? ?????????????.

?? ???????? ????????? ??????? ?????? ??????????, ?.?.?., ????????????????
?????????, ????? ???? "????? ?? ?????? ?????????? ???",  "??? ?????????
???????? ??????? ???????? ???????????", "??????????????? ? ?????????????
?????", "???????? ?????? ??: ??????????? ???????????",  "?????????? ?????:
???????????? ???????????", <?????????? ??????? ???????????????> ? ??????
??????.

??? ??????? ?????? ???????? ????????? ?????? ???????? ?????????? 3?900
???., ? ?????? ???. ??? ??????? ? ???? ????????? (??? ???? ??????? ? ?????
????????) ??????????????? ?????? 10%. ??? ??????? ? ???? ????????? (???
???? ???????), ????????? - ? ???????. ? ????????? ???????? ??????
??????????? ????????, ????-????? ? ???? ? ?????????. ?????? ???????? ?
10.00 ?????, ????????? -  ? 17.00 ?????. ?????? ??????????? - ? 09-30.

????? ?????????? ???????? - ? 3-? ??????? ?????? ?? ????? <?????????????>.

lbjnbzjn




From aahz at pythoncraft.com  Wed Apr 23 10:19:19 2003
From: aahz at pythoncraft.com (Aahz)
Date: 23 Apr 2003 10:19:19 -0400
Subject: Data-driven testing
Message-ID: <b867d7$f53$1@panix2.panix.com>

As I just mentioned in another post, I'm writing an ASCII database for
my address book.  Actually, the main reason for writing it was because
I've been maintaining multiple lists for various kinds of party invites,
and the update problem was killing me.

While writing it, it occurred to me that I *ought* to be using unit
tests to make sure that it was working correctly, but two things stopped
me.  Secondly, I was in a hurry.

But firstly, I ran into the same problem I had when writing the BCD
module: how the heck does one write tests when the results are
hand-calculated?  There were several times when I wrote what I thought
were correct tests for the BCD module, and my code was right and my test
was wrong.

In the case of my ASCII database, I'm going to need a big matrix to
properly test the query language.  How do I build that matrix?
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From jepler at unpythonic.net  Tue Apr 15 15:48:11 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Tue, 15 Apr 2003 14:48:11 -0500
Subject: file read : feature or bug ?
In-Reply-To: <00af01c3036f$50b0c0c0$82fea8c0@pcspiderman>
References: <3e9c2eea$0$6528$afc38c87@sisyphus.news.be.easynet.net> <20030415162616.GB30739@unpythonic.net> <00af01c3036f$50b0c0c0$82fea8c0@pcspiderman>
Message-ID: <20030415194811.GE30739@unpythonic.net>

On Tue, Apr 15, 2003 at 06:51:47PM +0200, Asbuilt Easynet wrote:
> But is there a workaround ?

Yes, kinda.

> > When you're using text files, seek will not operate as you expect.  The
                      ^^^^
If you open your file in binary (f = open(filename, "rb")) you'll be
able to use seek() in the way you expect.  *BUT* "newline translation"
is also not performed.  This means that if you wrote a text file on
Windows (opened in "w" mode) you will see the byte sequence '\r\n' where
you wrote a newline before.  On platforms with even more bizarre ideas
of what makes a "text file", you could get even stranger things, but
nobody uses those anymore, anyway.  Remember, "all the world's a vax".

I don't exactly remember how your program worked, but you can also use
the advice to only use a value returned from the tell() method when
seeking in a file.  This is actually guaranteed (by the C language
standard) to work and do what you'd expect.

Jeff




From revanna at mn.rr.com  Sat Apr 26 00:20:57 2003
From: revanna at mn.rr.com (Anna)
Date: Sat, 26 Apr 2003 04:20:57 GMT
Subject: a new pythonista?
Message-ID: <pan.2003.04.26.04.22.16.562041@mn.rr.com>

My daughter just asked me about this "python stuph". So I pulled up
livewires.org.uk and showed her some basics.

She just coded her first while loop and for loop... she thought it was
kewl! In fact, we had to kick her off the computer and send her to bed cuz
she's got so much to do tomorrow...

FWIW, she's 12...

Anna



From aleax at aleax.it  Mon Apr 28 03:08:43 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 07:08:43 GMT
Subject: Newbie: ( if good in self: print "Foobar" )
References: <uN1ra.15662$kc1.13753@nwrddc02.gnilink.net>
Message-ID: <%r4ra.19022$K35.649255@news2.tin.it>

<posted & mailed>

David Broadwell wrote:
   ...
> class foo:
>     bar = ['B','A','R']
> 
>     def baz():
>         for item in bar: print item
>         boggle()
>         for item in bar: print item

This baz function just won't work -- its unqualified references to 'bar'
and 'boggle' can't be solved, and the "signature" (pattern of arguments,
namely none) is only compatible with a staticmethod or a free function,
not with an ordinary method as this appears to be.

>     def boggle(self):
>         self.bar[1] = 'a'
>         self.bar[2] = 'r'

This one, on the other hand, is fine.


> Question 1: Do I read it right that; a classes method called with self can
> change the 'global' value without having to type the bloddy global for
> every one?!? (ie: the above code not the below code)

No, it's NOT a *global* attribute -- it's an attribute of the CLASS.  You
can access an attribute of a class through any instance of that class, as
long as the instance hasn't "overridden" the attribute (i.e., bound an
attribute of the same name as a per-instance attribute on itself).

> Question 2: If that is true, what do we call this notation? (anyone got
> links?)

"Class attribute" is the normal way to call a class attribute.  In C++
or Java, those attributes are normally known as "static attributes".

>>>> bar = ['B','A','R']
>>>> def baz():
>         for item in bar: print item
>         boggle()
>         for item in bar: print item

Now this is quite fine.

>>>> def boggle(self):
>         self.bar[1] = 'a'
>         self.bar[2] = 'r'

However, this function needs an argument.  You're not passing that
argument when you call it, thus, boom.


>>>> baz()
> B
> A
> R
> Traceback (most recent call last):
>   File "<pyshell#4>", line 1, in ?
>     baz()
>   File "<pyshell#2>", line 3, in baz
>     boggle()
> TypeError: boggle() takes exactly 1 argument (0 given)
>>>>

Exactly: it needs exactly 1 argument and you didn't give it.


> Question 3: So what exactly does my code have to look like to work?
> Question 4: Is this notation only valid inside a class?

You can use the "notation" wherever you want, but you cannot call a
free function without arguments if it wants an argument.

To make the above-exemplified free-standing boogle function work,
you would have to pass it an object having an attribute named bar.
In this case, the object in question is the module in which the
functions baz and boggle live.  You can get at the module object
in the caller, though it's a rather fishy usage IMHO:

>>> def baz():
         for item in bar: print item
         import sys
         boggle(sys.modules[__name__])
         for item in bar: print item




Alex




From tim.one at comcast.net  Mon Apr  7 21:45:24 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 07 Apr 2003 21:45:24 -0400
Subject: [Slightly OT]: More on ints and floats
In-Reply-To: <xdpka.23011$It5.19472@news2.central.cox.net>
Message-ID: <LNBBLJKPBEHFEDALKOLCAEMCECAB.tim.one@comcast.net>

[Steve Holden]
> ...
> As a matter of pure mathematics the integers are a subset of the reals,
> which are a subset of the complex numbers.
>
> Unfortunately pure mathematics and floating-point arithmetic only tend to
> intersect in the brains of people like Tim Peters.

The relationship can be formalized easily enough, but it's not so useful to
do so unless you work in that field:  computer floats are a peculiarly lumpy
and finite subset of the rational numbers, but the computer-float basic
operations (+ - * /) "usually" deliver different results than the same-named
operations on rationals.  The float operations can be defined precisely, but
the definitions are difficult to work with and reason about.  Knuth wasn't
the only one to note that many a good mathematician has given up in
frustration after trying to rigorously analyze just a few lines of
floating-point code.

For that reason I don't think there's much value in conflating computer ints
and computer floats:  the identities the latter fail to satisfy are
endlessly surprising to people.  Like addition and multiplication aren't
associative for floats, cancellation laws are shot all to hell for floats
(x+y == x+z doesn't imply y==z for floats), and so on.  This makes reasoning
hard.

OTOH, 754 float arithmetic has an inexact flag, set whenever the result of a
basic float operation differs from what the rational operation on the same
inputs would deliver.  Computer languages (including Python) overwhelmingly
failed to give users sane ways to get at this flag (among others).  If you
could get at it, it would be a lot easier to use floats for speed and
automatically detect when results became approximations.

Note that the IBM proposal for decimal arithmetic:

    http://www2.hursley.ibm.com/decimal/

doesn't distinguish between floats and integers.  It's an interesting mix
where the number of significant digits isn't unbounded, but where the
maximum is user-specified.  Make that big enough, and you can safely use the
same arithmetic for floating and integer calculations.  Figuring out how big
"big enough" is in advance can be tricky, though.  Setting the max # of
digits to 999999999 is a good trick -- provided you don't care how long
division takes <wink>.





From sde at recombinant.demon.co.uk  Tue Apr 29 14:15:37 2003
From: sde at recombinant.demon.co.uk (Stephen D Evans)
Date: Tue, 29 Apr 2003 19:15:37 +0100
Subject: 2.3b1 sys.path anomaly on W2K
Message-ID: <b8mfgc$ll8$1$830fa795@news.demon.co.uk>

With a fresh installation of Python 2.3b1 on Windows 2000 (having
'uninstalled' earlier installations).
The c:\python23 is on PATH.

Where does the 'C:\\WINNT\\System32\\python23.zip' come from in sys.path?
Does anyone else get this? Is it meant to be there?

C:\>python
Python 2.3b1 (#40, Apr 25 2003, 19:06:24) [MSC v.1200 32 bit (Intel)] on
win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.path
['', 'C:\\WINNT\\System32\\python23.zip', 'C:\\', 'c:\\python23\\DLLs',
'c:\\python23\\lib', 'c:\\python23\\lib\\plat-win',
'c:\\python23\\lib\\lib-tk',
'c:\\python23', 'c:\\python23\\lib\\site-packages']
>>>

Stephen D Evans





From g_will at cyberus.ca  Wed Apr  9 16:01:56 2003
From: g_will at cyberus.ca (Gordon Williams)
Date: Wed, 9 Apr 2003 16:01:56 -0400
Subject: [ANN] wxPyPlot V1.0 Beta 2
Message-ID: <001901c2fed2$df855640$6c5ce640@wnt20337>

Hi All,

I have released Version 1.0 beta 2 of wxPyPlot for producing simple plots
in wxPython.  This is a fix for Linux users because of some problems with
fonts and clipping area.

Users will require wxPython 2.4.0.6 or later and Numeric.

For your viewing pleasure, the website now has some html docs as well as a
few more screen shots of things.  Check out:
http://www.cyberus.ca/~g_will/wxPython/wxpyplot.html

Comments and suggestions welcome.

Regards,

Gordon Williams

-------------------------------------------------------------
What is wxPyPlot??

wxPyPlot is an enhanced derivative version of wxPlotCanvas to provide simple
lightweight plotting in wxPython. The emphasis is both on simple and
lightweight.

Before starting this work I looked at two plotting packages that could be
adapted to work with wxPython. SciPy Plt and Chaco are sophisticated and
powerful, but at the expense of consuming huge amounts of computer resources
for simple plots. They can be found at http://scipy.com. I was looking for
something that could be incorporated into a Boa application and had the
ability to plot lines with large numbers of points quickly for data capture
applications. wxPlotCanvas had the speed and some of the functionality that
I was looking for, but lacked a lot of features that I needed. This effort
has filled the gap.

Features:
- Automatic Scaling of Axes
- Legends for lines and markers
- Plot Title and Axis Labels
-  Zoom Region with Mouse (Rubber Band) or Programmatically
- Scroll Up/Down, Left/Right
- Plot Grid (Graticule) on Graph
- Printing, Preview, and Page Set-up (margins) -WYSIWYG
- Cursor X-Y User Axis Values
- Simple, Lightweight and Fast
- Can be Used with Boa Applications if Desired
- Lots of Comments and Doc Strings along with Example Usage

I have tested this out on Windows using Python 2.2 and wxPython 2.4.0.6. If
you have the chance to test it on Linux, please let me know how it works.
Suggestions and improvements welcome.

Screen shots, source and instructions for Boa can be found at

http://www.cyberus.ca/~g_will/wxPython/wxpyplot.html









From dale at riverhall.NOTHANKS.co.uk  Tue Apr 22 11:35:27 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Tue, 22 Apr 2003 16:35:27 +0100
Subject: Paintshop Pro 8
Message-ID: <hmnaavcdkebqaeq3e3alq07nubu80a1jts@4ax.com>

You may already know this but the latest version of Paintshop Pro
(public beta available from www.jasc.com) supports scripting.

Guess what the scripting language is?

 :-)

This is interesting given that - for the moment at least - Paintshop
Pro only runs on Windows so a COM based 'any language' scripting
approach might have been more natural.

So why is this?

* Possible Linux/Mac versions in the pipelines?
* Don't like COM?
* Want to promote Python?

Anyone any insight?
-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From martin.vonloewis at hpi.uni-potsdam.de  Thu Apr 10 04:34:38 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Thu, 10 Apr 2003 10:34:38 +0200
Subject: Python/C API
References: <mailman.1049956515.20986.python-list@python.org>
Message-ID: <b73aat$l7e$1@zeppelin.rz.uni-potsdam.de>

>   How does one write generators in C, or does one not?

One does not. Instead, just implement a function that returns an iterator.
Python applications cannot tell the difference between an iterator and the
result of calling a generator.

Regards,
Martin





From dkuhlman at rexx.com  Sat Apr  5 12:48:40 2003
From: dkuhlman at rexx.com (Dave Kuhlman)
Date: Sat, 05 Apr 2003 09:48:40 -0800
Subject: Calling C libraries from a Python script
References: <e81e0f3d.0304050639.37b59655@posting.google.com> <qmCja.4148$Du.14013@newsc.telia.net>
Message-ID: <b6n4rn$6uitb$1@ID-139865.news.dfncis.de>

Tim Gahnstr?m /Bladerman wrote:

> "Joshua Goodlett" <joshua at goodish.org> wrote in message
>> I'm trying to find examples of calling C library functions from within
>> a Python script.  Any advice or pointers on where to look for examples
>> would be appreciated.
> 
> 
> There is a whole chapter in the docs dedicated to the subject,
> 
> Have a look at
> 
> http://www.python.org/doc/current/ext/ext.html

Also look at SWIG.  It attempts to generate for you what the above doc tells 
you how to do.  If you use SWIG, you may be able to write/generate that 
code without learning anything, except how to use SWIG of course.  Feed 
SWIG a C header file (or a SWIG interface definition file) that describes 
the functions you want to call and SWIG will generate the C wrappers that 
enable you to call those functions from Python.

    http://www.swig.org/

  - Dave

-- 
Dave Kuhlman
dkuhlman at rexx.com
http://www.rexx.com/~dkuhlman



From aahz at pythoncraft.com  Sun Apr 27 19:57:40 2003
From: aahz at pythoncraft.com (Aahz)
Date: 27 Apr 2003 19:57:40 -0400
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <3EAC2962.2050609@soraia.com>
Message-ID: <b8hqpk$rn3$1@panix2.panix.com>

In article <3EAC2962.2050609 at soraia.com>,
Joe Francia  <usenet at soraia.com> wrote:
>
>The Deitels have a textbook called _Python How to Program_ that is aimed 
>at beginning programmers.  I'm not sure how good/bad it is.  Three out 
>of four reviewers on Amazon seem to like it (the fourth guy gave it one 
>star based on its lack of Perl coverage).

If you Google through the history of c.l.py, you'll find many people
giving it a thumbs-down.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From max at alcyone.com  Wed Apr 23 20:26:00 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 23 Apr 2003 17:26:00 -0700
Subject: Curses on Win32
References: <b86r8p$27n$1@gateway.northgrum.com> <vae8s19g0aug3f@corp.supernews.com>
Message-ID: <3EA72F18.C2DC9BE8@alcyone.com>

Adam Vandenberg wrote:

> You can take a look at:
> http://flangy.com/dev/python/curses/
> 
> So far I have a binary distribution for Python 2.2.2, but I'll be
> releasing
> the source soonish.

Looks pretty good so far!  I have a Windows machine around for playing
games only, and do all my real work on Linux or Solaris.  I have some
Python projects that use curses, so I installed your curses module on my
Windows machine and tried them out.  Only two things came up:

1.  It would be good if the README.txt in the curses.zip package were in
the curses subdirectory, not in the root.  Otherwise it tries to replace
the README.txt that's in site-packages (at least with the standard
Windows installer), which isn't what you want.

2.  I wasn't able to get stdscr.subwin to perform properly.  In calling
it with the .subwin(begin_y, begin_x) form, I couldn't manage to get
anything other than

	ValueError: Child's extent not in bounds of parent

even with something reasonable like (IDLE):

	import curses
	stdscr = curses.initscr()
	w = stdscr.subwin(10, 10)

Perhaps that modality (documented in the Python Library documentation
under 6.13.2) is just not supported yet?

Nevertheless, great work!  Only one or two of my projects actually
attempt to create subwindows; all those which didn't ran flawlessly. 
Specifically, my cellular automata library CAGE, many of whose examples
use curses (and which quite a few Windows users have complained about),
worked fine:

	http://www.alcyone.com/pyos/cage/

Congratulations on the good work!

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Stars are the golden fruit of a tree beyond reach.
\__/ George Eliot
    Blackgirl International / http://www.blackgirl.org/
 The Internet resource for black women.



From tim at bladerman.com  Wed Apr 16 16:16:03 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Wed, 16 Apr 2003 20:16:03 GMT
Subject: flickering in wxPython with bliting and memmoryDC
References: <ZJ7na.241$dP1.739@newsc.telia.net>
Message-ID: <7Sina.337$dP1.1344@newsc.telia.net>

"Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote

> The panel is not large or anything and refreshes arent done at a
particular
> high speed, but it flickers everytime. Just like if the dc gets cleared
> before the blitting but I can't se that I do that.

Replying to myself here again... I just got the answer to my prayers from
comp.soft_sys.wxwindows and thougt that I could just as well put the info in
here to fill up the google archives in case somene else is experiencing the
same problem.

For some reason there is a wxEraseEvent issued every time the  panel is
repainted. This must be caught to avoid flickering.

Catch it with a callback to an empty function.

EVT_ERASE_BACKGROUND(fromPanel,useEmptyFunc)

Tim





From Dialtone.NOSPAM.ME at aruba.it  Wed Apr 16 04:12:03 2003
From: Dialtone.NOSPAM.ME at aruba.it (Dialtone)
Date: Wed, 16 Apr 2003 10:12:03 +0200
Subject: Any good donothing?
References: <slrnb9p2vf.t3.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix> <mailman.1050453505.22283.python-list@python.org>
Message-ID: <slrnb9q42j.10n.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix>

* Gerhard H?ring <gh at ghaering.de> [160403 00:37]:
> From what I remember other such solutions just use pipes to pipe the
> output of cdda2wav into the mp3 encoder. No synchronization problems,
> then.

Uhmmm... I tried to use pipes but I didn't managed it since it didn't
take the right input file to encode.

>> Is there any other way to make a do_nothing_for_some_time(n) function?
> 
> The sleep function from the time module does just this. I doubt it
> solves your problems, though. I'd look into pipes and try to use them
> using os.popen*. If you want to display the progress in a GUI, you might
> also want to select.select or select.poll on the file descriptor in
> order to keep the GUI smooth.

I didn't know select.select select.poll, I'll give them a try.

thanks for the infos.

-- 
try: troll.uses(Brain)
except TypeError, data: 
   troll.plonk()
Co-Responsabile Hardware Information & Technology http://hit.edengames.net



From jepler at unpythonic.net  Wed Apr 30 15:50:03 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Wed, 30 Apr 2003 14:50:03 -0500
Subject: weird behavior in 'for line in sys.stdin'
In-Reply-To: <0SDra.65118$xB4.1544318@news2.east.cox.net>
References: <0SDra.65118$xB4.1544318@news2.east.cox.net>
Message-ID: <20030430195001.GD22792@unpythonic.net>

The way that xreadlines works is something like this (but expressed as a
generator):
    def xreadlines(f):
        while 1:
            lines = f.readlines(500) # I'm not sure what number
            if not lines: break
            for l in lines:
                yield l
so you only exit when 'f.readlines()' returns zero lines.  As you
discover, this means hitting ^D twice---the first one ends the lines
returned by readlines prematurely (dropping into the 'yield' loop) 
and the second makes readlines return an empty list.

Unfortunately, since the parameter to readlines is only a *hint*,
reading fewer lines can't be treated as an EOF condition.  If that were
true you could write
    def xreadlines2(f):
        while 1:
            lines = f.readlines(500)
            for l in lines:
                yield l
            if len(lines) != 500: break
or so.  Other times that readlines will return fewer lines than expected
would be if the file is a socket or a pipe, for instance.

Jeff




From hupp at cs.wisc.edu  Thu Apr 24 23:19:25 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Thu, 24 Apr 2003 22:19:25 -0500
Subject: Dict handling of floats
In-Reply-To: <98imn-ust.ln1@grendel.myth>
References: <98imn-ust.ln1@grendel.myth>
Message-ID: <20030425031925.GA32131@upl.cs.wisc.edu>

On Thu, Apr 24, 2003 at 07:14:01PM -0700, Jim Richardson wrote:
> I have come across a weirdness. I have a number of floats, (lat long
> pairs) which I stuff into a dict, but when I take them out, my nice %2f4
> number, (say, 47.5894) is something long and *almost* right, like
> 47.589333333333333339. What gives? I am not doing anything special as
> far as I can tell, just stuffing the numbers into a dict, and recalling
> it later. This is with Python 2.2 on Linux, and no, it's not an original
> Pentium :)

http://www.python.org/doc/current/tut/node14.html

-Adam
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 240 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030424/9c1c19eb/attachment.sig>

From riccardo_togli_questo at _e_questo_sideralis.net  Sat Apr 19 09:28:18 2003
From: riccardo_togli_questo at _e_questo_sideralis.net (Riccardo Galli)
Date: Sat, 19 Apr 2003 15:28:18 +0200
Subject: getting an item's list index
References: <mailman.1050757111.12171.python-list@python.org>
Message-ID: <pan.2003.04.19.13.28.16.939747@_e_questo_sideralis.net>

On Sat, 19 Apr 2003 14:57:45 +0200, Michael Anckaert wrote:

> Hello all,
> 
> I have a list called contents. If one of the elements has the value
> "item1" for example, how can I get that elements index?
> 
> 
> for item in contents:
> 	# here I need to get that items index

Hi Michael.

You're speaking about 'index' method.

contents.index(item) gives you what you want.

Maybe suits more your needs the range function

for i in range(len(contents)):
	contents[i] ...

Ciao,
Riccardo

-- 
-=Riccardo Galli=-

 _,e.
s~  ``
 ~@.   ideralis Programs
.   ol 
 `**~  http://www.sideralis.net




From max at alcyone.com  Sat Apr 19 17:15:07 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sat, 19 Apr 2003 14:15:07 -0700
Subject: arg-passing style question
References: <A4ioa.87812$Vzu.36931@news02.bloor.is.net.cable.rogers.com> <mailman.1050784411.15301.python-list@python.org>
Message-ID: <3EA1BC5B.DC5CE92E@alcyone.com>

Chad Netzer wrote:

> For more discussion, see these threads:
> 
> http://mail.python.org/pipermail/python-list/2003-March/152761.html
> http://mail.python.org/pipermail/python-list/2003-March/152768.html

I would have responded to the question, but you quoted one of my
responses to a similar question in another thread ... :-).

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Never had very much to say / Laugh last, laugh longest
\__/ Des'ree
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From sdieselil at yahoo.com  Fri Apr  4 12:01:44 2003
From: sdieselil at yahoo.com (sdieselil)
Date: 4 Apr 2003 09:01:44 -0800
Subject: What are the differences between _weak_ references and "normal" references?
Message-ID: <169a7283.0304040901.4bc3531b@posting.google.com>

Hi

I noticed a module weakref in Python standard library. What is weak
reference and in what way it differs from "normal" reference? By the
way, how can I implement usual reference in Python (I mean C++
reference)?



From jmv-lists at linux-ink.ru  Tue Apr  1 04:05:31 2003
From: jmv-lists at linux-ink.ru (Eugene Morozov)
Date: 1 Apr 2003 01:05:31 -0800
Subject: Locale and threads in python 2.1.3 and python 2.2.1
Message-ID: <e59f7389.0304010105.1853730f@posting.google.com>

Hello!
Here's a simple example which works correctly when using 2.2.1
interpreter and fails under 2.1.3 (text.lower() result isn't in
accordance with locale):
#######################################
import time, locale
from threading import Thread
from thread import start_new_thread

LOCALE='ru_RU.KOI8-R'
TEXT='????? ? ????'

class TestThread(Thread):
    def __init__(self):
        Thread.__init__(self)

    def run(self):
        print 'In thread: %s' % TEXT.lower()

        locale.setlocale(locale.LC_ALL, LOCALE)
        print 'After setlocale in thread: %s' % TEXT.lower()

        
print 'Before setlocale: %s' % TEXT.lower()
locale.setlocale(locale.LC_ALL, LOCALE)
print 'After setlocale: %s' % TEXT.lower()

t = TestThread()
t.start()
#######################################

I cannot switch to python 2.2 because 2.1 is still required by Zope.
Does anyone know how to fix that without patching Zope to call
setlocale in every thread?
Eugene



From richie at entrian.com  Tue Apr 29 06:01:59 2003
From: richie at entrian.com (Richie Hindle)
Date: Tue, 29 Apr 2003 11:01:59 +0100
Subject: Python-2.3b1.exe install error on W98se
References: <mailman.1051538624.9122.python-list@python.org> <b8k9rp$9fn$1@news.hccnet.nl>
Message-ID: <p0jsavcfgt346tad93i4hnnsdr90cdmfji@4ax.com>

Hi Anton,

> Please run this program and show me the output and I will be
> able to complete Python-2.3b1.exe myself. 
> [...]

Here you go:

>d:\Python23\pythonw -u anton.py
  96   71   60  109   61  134   90  177  137  222
 243  242  153  210   22  174  126  201  218  105
  61   56   35   30  117  116    6   85  216  195
  39   88  156  111  222  168  104   75  103   35
 157  173  125  152  238   91   82   81  130  218
  17   78  216  248   74  165  115  237  145  242
 147  140    9   90   65   59  166  222   26   82
   2   30   11  214  219    4   20   21  118  204
  34  151   19   63  115   89  168  226  120  206
 180   97  190  227  234  249   86  112  159  249
  37  227  132   53  187   48   87  154   61  179
  99   42   62   58   31  225    6  240    4   99
 211  238   14   44  215    4   95  155  128  154
 234  224  160   27   48  212  148   98   95  251
 152  120  151  105   10  202   12  229   14   41
  59   86  113  178  207  197  142  221  151  232
 104  177  133   36  143    7  184   56  104    7
 108   17   61  200  197   42   92  147  166   54
 228  129  132   28  241  104   10  126   25  198
  24  112  179  172   56  179  139   11   75  212
  31  197  142  251   73  117  148   84  132  185
 216   59   45   66  224    6   84   97  217   67
  98  111  209   37  123  139  207  149  162  219
 207  168  207   61   31  219  238   71   85   26
 201  214  100   56   19   77  178  223  164  185
  60  118  162  119  241  251  117  133   84  181
 205  211   98   69  119  157  240  234  253  152
 210  141    8  116   52   13  134  232  104  247
 240  121   52  232   17   87  224  110  251   41
 142  146   58  156  166  144    2   34  240  183
  11  233   62  247  159   39  181  227   90  208
  44  219   55  203   80  220   44   21  231   31
  69    7  228  200  160  184   27   55   76   20
 162  137  156   13   31  178  103    7    2  103
  51  173  205  230  204  137   12   63  192  217
 170   80  155  107   28   75  130   95  235  123
  50   65  145  173  232   80   35  224   96  217
 135  241   15   61   51  226   49  224  234  248
 211  128  242  167  147   51  235   68  154  180
  29   36  142  198  100  140  109  114  106  161

>Exit code: 0

-- 
Richie Hindle
richie at entrian.com




From aleax at aleax.it  Sat Apr 12 18:37:34 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 22:37:34 GMT
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <Z7nla.45$W4.8475@news.uchicago.edu> <7oZla.5396$LB6.150919@news1.tin.it> <mailman.1050180741.23110.python-list@python.org>
Message-ID: <Oy0ma.5908$LB6.168653@news1.tin.it>

Norman Vine wrote:
        ...
>> > Have you considered using R (www.r-project.com) and RS/Python
>> 
>> Hmmmm -- this site's most succint homepage's in English, but all
>> links from it seem to be trying to show me lots and lots of Japanese
>> text.  What's the right way to get R docs & other downloads these
>> days for the non-Japanese-speaker...?
> 
> http://www.r-project.org/

ah yes, MUCH better, thanks -- not sure what the .com site is
about, but the .org one is most assuredly the one that should
have been suggested in the first place.


Alex





From BPettersen at NAREX.com  Wed Apr 16 17:15:51 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 16 Apr 2003 15:15:51 -0600
Subject: Why functional Python matters
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFA2@admin56.narex.com>

> From: laotseu [mailto:bdesth at removethis.free.fr] 
> 
> Paul Foley wrote:
> > On Tue, 15 Apr 2003 23:49:37 +0000, laotseu  wrote:
> > 
> > 
> >>Even for OO programmers, functionnal features in Python are IMHO a
> >>great plus, and BTW functionnal and OO paradigm does not have to
> >>conflict (that would be functionnal vs imperative and object vs
> >>procedural). CLOS is one of the great system objects out here, and
> >>it's been implemented on top of a functionnal language,
> > 
> > Oh yes?  Which functional language would that be?
> > 
> 
> CLOS means Common Lisp Object System.

"Houston: we have a claim that CL is a functional language, please
proceed with crosspost to c.l.sml, c.l.haskell, ..." <grin>.

-- bjorn




From duncan at NOSPAMrcp.co.uk  Thu Apr 24 08:25:10 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 12:25:10 +0000 (UTC)
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com>
Message-ID: <Xns9367885737E62duncanrcpcouk@127.0.0.1>

Nikolai Kirsebom <nikolai.kirsebom.NOJUNK at siemens.no> wrote in 
news:0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062 at 4ax.com:

> class A(object):
>      InstCount = 0
>      def __init__(self):
>           self.__class__.__dict__['InstCount'] += 1
> 
> class B(A):
>      InstCount = 0
> 
> class C(A):
>      InstCount = 0
> 
> Is the syntax used for incrementing the class-attribute for every
> instanciation 'the way' it should be done ?

It would be more normal just to write:

		self.__class__.InstCount += 1

Alternatively you could create a class method to do the increment, but that 
is probably needlessly complex here.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From kylnas at tiscalinet.it  Wed Apr 23 12:23:55 2003
From: kylnas at tiscalinet.it (Giacomo)
Date: Wed, 23 Apr 2003 18:23:55 +0200
Subject: execute a file
Message-ID: <jcfdav4pr3ebi52pf5cdt1lfjmipgjuh7n@4ax.com>

Hi everybody, i would like to know how can i execute a generic file.
I mean execute a file like *.txt, *.jpg launching the application to
see that file (an editor for txt file, for example paint shop pro for
jpg file, and so on).
bye and thanks



From ulope at gmx.de  Fri Apr  4 07:32:55 2003
From: ulope at gmx.de (Ulrich Petri)
Date: Fri, 4 Apr 2003 14:32:55 +0200
Subject: Is there a python realaudio player? For windows?
References: <b6i6uq$rcs$0@216.39.172.122> <yH4ja.7115$ey1.562521@newsread1.prod.itd.earthlink.net>
Message-ID: <b6ju1n$68cvf$1@ID-67890.news.dfncis.de>

"Tipton Bandy" <tipton51 at subdimension.moc> schrieb im Newsbeitrag
news:yH4ja.7115$ey1.562521 at newsread1.prod.itd.earthlink.net...
>
>
> A *.ram file is a "real audio metafile". It does not contain any
> sound at all.  If you save the file and open it up in a text
> editor, you will find a line that looks like this:
> rtsp://server/path/file.rm
> This is a streaming media file.  I was hoping to be able to copy
> the *.rm file to my hard disk so that I could listen offline, but
> so far I haven't been able to do that.  Google hasn't helped.
> Tipton
>

There has been a tool named "Streambox VCR" which could save Real streams to
HD.
Ii think the makers have been sued by Real to drop it, but i'm sure, you
still can find it on the net.

Cioa Ulrich





From skip at pobox.com  Tue Apr 29 12:30:51 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 29 Apr 2003 11:30:51 -0500
Subject: Where did you learn to unit test?  How did you learn?
In-Reply-To: <1147e466.0304290753.5cbf3c8c@posting.google.com>
References: <1147e466.0304290638.287687f6@posting.google.com>
        <b8m3cj$8vu$1@ginger.mathworks.com>
        <1147e466.0304290753.5cbf3c8c@posting.google.com>
Message-ID: <16046.43195.485094.904228@montanaro.dyndns.org>

>>>>> "-c" == Christopher Blunck <blunck at gst.com> writes:

    >> people to go back and add unit tests for all functionality.  One good
    >> way to ease people into it is to get them to write a unit test for
    >> every bug-fix they make.


    c> We do that already.  It works to a certain extent.  But usually
    c> requires brow beating of the bug fixer.  I guess we really just need
    c> some time to pass.  Unit testing is an investment and by definition
    c> investments don't pay off immediately.

When people who get serious about unit testing start getting larger pay
raises than those who don't, the others will follow. ;-)

Skip




From Yu.Wang at synopsys.com  Thu Apr 17 04:04:51 2003
From: Yu.Wang at synopsys.com (Yu Wang)
Date: Thu, 17 Apr 2003 16:04:51 +0800
Subject: get pid from popen3 or get redirection in execl
In-Reply-To: <3E9E44A8.3080203@synopsys.com>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <bDSma.22753$T34.687158@news2.tin.it> <slrnb9o8s8.iea.news@datatailors.com> <mailman.1050427771.31620.python-list@python.org> <3E9E0FBA.6050503@indy.rr.com> <3E9E44A8.3080203@synopsys.com>
Message-ID: <3E9E6023.3020603@synopsys.com>

Sorry for this stupid question. I've found at least two ways to resolve it.
use sys.stderr, and dup2() function both work for it.


Yu Wang wrote:
> Hi,all
>   Does "execl", or whatever this family, support redirection of output?
>   Say, I have a script: ( on my Linux7.2,C shell environment,with 
> python2.2.1)
> 
> command="somecommand"
> arg="-i somefile > output file"
> execl(command,arg)
> 
> what's wrong with this?
> 
> And,
>    how could I get child pid from "popen3(somecommand)".
> 
> Thanks to you.
> Regards.
> 
> 





From news at saarpfalz.de  Tue Apr  1 08:53:44 2003
From: news at saarpfalz.de (Susanne Ziemerle)
Date: Tue, 1 Apr 2003 15:53:44 +0200
Subject: Problem: 1040, too many connections
Message-ID: <3e899a43$0$21745$9b622d9e@news.freenet.de>

Hello,

we do a mysql-Connection (Mysql 3.23.41, MySQL-python-0.9.2) from python
scripts in ZOPE under SUSE Linux 7.3.
And we have no problems.

Now we set up a new server with Mysql 4.0.12 instead of 3.23.41. And since
that we get the problem "1040, too many connections" when we start a python
script which perform a lot of SQL-Statements.

Any hints?

Greetings,
 Susanne






From staschuk at telusplanet.net  Tue Apr 29 14:04:28 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 18:04:28 -0000
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 29)
Message-ID: <vatflcjn7dum95@corp.supernews.com>

QOTW:  "But I like to [polish] the code while it does not yet exist, or
make hundreds of new options DURING the programming. That's why many of
my programs never see the daylight :)" -- Dino Levy

"For a long time, the industry labored under the illusion that if we
could all agree on One Big API then we'd have interoperability.
Examples have included Posix, X11, OLE/COM/DCOM, CORBA, DCE, OpenDoc,
and  the list goes on; and it's never, ever, ever worked (with the
single exception of the "sockets" library for IP networking)."  -- 
Tim Bray  http://www.tbray.org/ongoing/When/200x/2003/03/24/XMLisOK


Discussion:
    Techniques for making *small* stand-alone versions of Python
    programs.
        <http://groups.google.com/groups?th=3dac0b6a85d8c13a>

    Jack Diederich exhibits two real-life uses for metaclasses:
    registering classes with a factory, and named prototypes.
        <http://groups.google.com/groups?th=296a0181247593f3>

    Jeff Epler describes some common uses of C++ references, and shows
    how the same problems are usually solved in Python.
        <http://groups.google.com/groups?th=5c54ac4ec68aeafa#link6>

    Tim Peters comments on reverse-engineering a mysterious 48-bit
    floating-point number format, illustrating the complexity of the
    subject.
        <http://groups.google.com/groups?th=6a220801e92e14e#link8>

    In the middle of a tongue-in-cheek discussion of overloadable
    short-circuiting operators, a speculative notion for implementing
    microthreads (inter alia) by turning all functions into generators.
        <http://groups.google.com/groups?selm=mailman.1051284157.17376.python-list%40python.org>

Announcements:
    Python 2.3b1: The first beta release of Python 2.3.
        <http://www.python.org/2.3/>

    Matplotlib 0.1: Matplotlib's goal is to make publication-quality
    plotting easy in Python, with a syntax familiar to Matlab users.
        <http://nitace.bsd.uchicago.edu:8080/matplotlib>

    mxODBC Zope Database Adapter 1.0.4: Allows you to easily connect
    your Zope installation to just about any database backend on the
    market today.
        <http://www.egenix.com/>

    Pyro 3.2: Pyro is an advanced and powerful Distributed Object
    Technology system written entirely in Python, that is designed to
    be very easy to use.
        <http://pyro.sourceforge.net/>

    Sybase DB-API module 0.36: The Sybase module provides a Python
    interface to the Sybase relational database system.  It supports
    all of the Python Database API, version 2.0 with extensions.
        <http://www.object-craft.com.au/projects/sybase/>

    Twisted 1.0.4: Twisted is an event-driven framework for building
    networked clients and servers.  It contains a powerful and simple
    networking core, a full-featured suite of interoperable protocols,
    among them a powerful web server and applications framework.
        <http://www.twistedmatrix.com/>


========================================================================
Everything you want is probably one or two clicks away in these pages:

    Python.org's Python Language Website is the traditional
    center of Pythonia
        http://www.python.org
    Notice especially the master FAQ
        http://www.python.org/doc/FAQ.html

    PythonWare complements the digest you're reading with the
    daily python url
         http://www.pythonware.com/daily  
    Mygale is a news-gathering webcrawler that specializes in (new)
    World-Wide Web articles related to Python.
         http://www.awaretek.com/nowak/mygale.html 
    While cosmetically similar, Mygale and the Daily Python-URL
    are utterly different in their technologies and generally in
    their results.

    comp.lang.python.announce announces new Python software.  Be
    sure to scan this newly-revitalized newsgroup at least weekly.
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python.announce

    Brett Cannon continues the marvelous tradition established by 
    Andrew Kuchling and Michael Hudson of summarizing action on the
    python-dev mailing list once every other week.
	http://www.python.org/dev/summary/

    The Python Package Index catalogues packages.
        http://www.python.org/pypi/

    The somewhat older Vaults of Parnassus ambitiously collects references
    to all sorts of Python resources.
        http://www.vex.net/~x/parnassus/   

    Much of Python's real work takes place on Special-Interest Group
    mailing lists
        http://www.python.org/sigs/

    The Python Business Forum "further[s] the interests of companies
    that base their business on ... Python."
	http://www.python-in-business.org

    The Python Software Foundation has replaced the Python Consortium
    as an independent nexus of activity
        http://www.python.org/psf/

    Cetus does much of the same
        http://www.cetus-links.org/oo_python.html

    Python FAQTS
        http://python.faqts.com/

    The old Python "To-Do List" now lives principally in a
    SourceForge reincarnation.
        http://sourceforge.net/tracker/?atid=355470&group_id=5470&func=browse
        http://python.sourceforge.net/peps/pep-0042.html
     
    The online Python Journal is posted at pythonjournal.cognizor.com.
    editor at pythonjournal.com and editor at pythonjournal.cognizor.com
    welcome submission of material that helps people's understanding
    of Python use, and offer Web presentation of your work.

    *Py: the Journal of the Python Language*
        http://www.pyzine.com

    Archive probing tricks of the trade:
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python&num=100
        http://groups.google.com/groups?meta=site%3Dgroups%26group%3Dcomp.lang.python.*

Previous - (U)se the (R)esource, (L)uke! - messages are listed here:
  http://www.ddj.com/topics/pythonurl/         
  http://purl.org/thecliff/python/url.html (dormant)
or
  http://groups.google.com/groups?oi=djq&as_q=+Python-URL!&as_ugroup=comp.lang.python


Suggestions/corrections for next week's posting are always welcome.
E-mail to <Python-URL at phaseit.net> should get through.

To receive a new issue of this posting in e-mail each Monday morning
(approximately), ask <claird at phaseit.net> to subscribe.  Mention
"Python-URL!".


-- The Python-URL! Team--

Dr. Dobb's Journal (http://www.ddj.com) is pleased to participate in and
sponsor the "Python-URL!" project.




From avner at skilldesign.com  Tue Apr 15 09:52:57 2003
From: avner at skilldesign.com (Avner Ben)
Date: Tue, 15 Apr 2003 15:52:57 +0200
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de> <Xns9358723026D8Cduncanrcpcouk@127.0.0.1> <3e9bdd3d.100422968@news.online.de>
Message-ID: <3e9c0ebd$1@news.012.net.il>

"Johannes Eble" <skywalkerpackage at hotmail.com> wrote in message
news:3e9bdd3d.100422968 at news.online.de...
> Hi Duncan,
>
> On Wed, 9 Apr 2003 10:20:01 +0000 (UTC), Duncan Booth
> <duncan at NOSPAMrcp.co.uk> wrote:
>
> >skywalkerpackage at hotmail.com (Johannes Eble) wrote in
> >news:3e93e50f.676072541 at news.online.de:
> >
> >> is there a library available for Python that behaves similar to MS
> >> Visio? For example, I would like to create vector graphic objects and
> >> connect the with a line. If I move, flip or rotate a connected object,
> >> the line should follow the connection.

If you are looking for using a class library from within a Python program,
take a look at OGL that comes bundled with wxPython.

A sample open-source UML editor that is already using wxOGL is PyUT - see
http://pyut.sourceforge.net .

    Avner.





From wcmartin at vnet.net  Fri Apr 18 11:32:43 2003
From: wcmartin at vnet.net (Bill Martin)
Date: Fri, 18 Apr 2003 10:32:43 -0500
Subject: Classes in Python
Message-ID: <3EA01A9B.4030209@vnet.net>

I'm wondering about the value of allowing a class definition like this:

             class C:
               pass

Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or 
some such. If I try to print a.x2 or b.x1, I get an exception message 
basically saying those member variables don't exist. It seems to me this 
defeats one of the basic ideas behind OOP, that is to assign members to 
the class definition, thus ensuring that all instantiations of a class 
have the same member variables and methods. I know there must be a 
reason that python allows this kind of thing that I described, but I 
can't figure out what that reason could be. To me, it looks like a very 
dangerous thing to do. Can someone help me understand why we'd like to 
do it? Thanks,

Bill Martin




From garryknight at gmx.net  Mon Apr 21 18:00:21 2003
From: garryknight at gmx.net (Garry Knight)
Date: Mon, 21 Apr 2003 23:00:21 +0100
Subject: regular expressions
References: <b81lq4$17n8$1@galaxy.us.dell.com>
Message-ID: <1050961701.8396.0@despina.uk.clara.net>

junk at aol.com wrote:

> Can anybody recommend some good resources for learning regular
> expressions.  Primarly web sites.

The best resource (IMO) is Jeffrey Friedl's "Mastering Regular Expressions",
published by O'Reilly, ISBN 1-56592-257-3.

-- 
Garry Knight
garryknight at gmx.net   ICQ: 126351135
Linux registered user 182025



From peter at engcorp.com  Tue Apr 15 14:32:21 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 15 Apr 2003 14:32:21 -0400
Subject: Printing Columns
References: <mailman.1050426466.11198.python-list@python.org>
Message-ID: <3E9C5035.F52E63BB@engcorp.com>

Lindstrom Greg - glinds wrote:
> 
> myList = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H']
> [...]
> but I would like
> 
>         A       D       G
>         B       E       H
>         C       F
> 
> problem is, the code (all 12 lines) is a bit convoluted and would not be
> easy for the next person using it to figure out (and in my mind,
> maintainability is priority one; that's why I use Python).  I would like to
> know if there is a "pythonic" way to do this. So, here are the specs....
> 
> Given a list of length N with 0<=N<=infinity and a number of columns, C (C>0
> and may be less than N), how do we print out the elements of the list in the
> above (column major?) form?

Unmaintainable, but with the "virtue" of being shorter than 12 lines:  ;-)

c = 3
print '\n'.join([' '.join([myList[i] for i in range(s, len(myList), c)]) 
    for s in range(len(myList)/c+1)])

A  D  G
B  E  H
C  F


Note: relies on current integer division, otherwise you need to add an
int() call in the last bit.

Disclaimer: may not work with *all* values across the range specified for N...

(I'm not in any way suggesting the above is maintainable, but on the other
hand it's a compact way of showing you want I would do: use a list 
comprehension which chooses a subset of the items using the step parameter 
on range(), and iterating over the starting offset for that list.)

(I'd also write tests first, to drive the code and make sure it actually
works in all cases.  The above might not.)

-Peter



From lbates at swamisoft.com  Mon Apr 28 14:53:12 2003
From: lbates at swamisoft.com (Larry Bates)
Date: Mon, 28 Apr 2003 13:53:12 -0500
Subject: Fix for datetime conversion to integer and back to string; was 1 hour off
References: <4f705b08.0304281031.ddd8f75@posting.google.com>
Message-ID: <I5OcnQDmyq4b5DCjXTWcpQ@comcast.com>

mx.DateTime has function DateTimeFromCOMDate to
go from COM date (seconds since 1970) to mx.DateTime
instance.  DateTime class has a method COMDate to go
the other direction.

Larry Bates

=--------------------

"La Uruguaya" <lacuidadanadelmundo at yahoo.com> wrote in message
news:4f705b08.0304281031.ddd8f75 at posting.google.com...
> I searched this group for a solution to a problem I had converting
> dates from a string format to integer (in seconds since 1970) and back
> to strings.  During some times of the year, the conversion didn't do
> a perfect round trip and I ended up with a datetime that differed from
> the original datetime by one hour.  I found 2 postings with variations
> on the problem, but not a solution.  I implemented the following
> solution and wanted to share it in case somebody needs it later.
>
>
> import time
>
> DATE_FORMAT = "%Y/%m/%d %H:%M:%S"
>
> def date2int(strdate):
>     result = 0
>     try:
>        tuple = time.strptime(strdate, DATE_FORMAT)
>        result = time.mktime(time.strptime(strdate, DATE_FORMAT))
>     except:
>        pass
>     return (result, tuple)
>
> def correctForDaytime(tup):
>     lst = list(tup)
>     lst[3] -= tup[-1] # if this is zero, then this is a no op
>     lst[-1] = 0
>     return tuple(lst)
>
> def int2date(intdate):
>     result = ''
>     if intdate > 0:
>        try:
>           tuple = time.localtime(intdate)
>           tuple = correctForDaytime(tuple) # commenting this line out
> will show you the bug
>           result = time.strftime(DATE_FORMAT, tuple)
>        except:
>           pass
>     return (result, tuple)
>
> def f(t1str):
>     (t1int, tup1) = date2int(t1str)
>     (t1str2, tup2) = int2date(t1int)
>     print t1str, "->", t1int
>     print t1int, "->", t1str2
>     if tup1 == tup2:
>        print "SAME", tup1, tup2
>     else:
>        print "DIFF", tup1, tup2
>     if t1str != t1str2:
>        print "FAILED", t1str, t1str2
>     else:
>        print "OK"
>
> f("2003/01/15 11:15:00")
> f("2003/02/15 11:15:00")
> f("2003/03/15 11:15:00")
> f("2003/04/15 11:15:00")
> f("2003/05/15 11:15:00")
> f("2003/06/15 11:15:00")
> f("2003/07/15 11:15:00")
> f("2003/08/15 11:15:00")
> f("2003/09/15 11:15:00")
> f("2003/10/15 11:15:00")
> f("2003/11/15 11:15:00")
> f("2003/12/15 11:15:00")
>
> f("2003/01/03 11:15:00")
> f("2003/02/03 11:15:00")
> f("2003/03/03 11:15:00")
> f("2003/04/03 11:15:00")
> f("2003/05/03 11:15:00")
> f("2003/06/03 11:15:00")
> f("2003/07/03 11:15:00")
> f("2003/08/03 11:15:00")
> f("2003/09/03 11:15:00")
> f("2003/10/03 11:15:00")
> f("2003/11/03 11:15:00")
> f("2003/12/03 11:15:00")
>
> f("2003/01/27 11:15:00")
> f("2003/02/27 11:15:00")
> f("2003/03/27 11:15:00")
> f("2003/04/27 11:15:00")
> f("2003/05/27 11:15:00")
> f("2003/06/27 11:15:00")
> f("2003/07/27 11:15:00")
> f("2003/08/27 11:15:00")
> f("2003/09/27 11:15:00")
> f("2003/10/27 11:15:00")
> f("2003/11/27 11:15:00")
> f("2003/12/27 11:15:00")





From gh at ghaering.de  Thu Apr 24 09:51:15 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Thu, 24 Apr 2003 15:51:15 +0200
Subject: Vim Newbie =?windows-1252?Q?=96_Python_Edit/Compile/Run_?=
 =?windows-1252?Q?cycle_and_integrating_py_help_questions?=
In-Reply-To: <a791ab19.0304240532.5a025b81@posting.google.com>
References: <a791ab19.0304240532.5a025b81@posting.google.com>
Message-ID: <3EA7EBD3.3010605@ghaering.de>

Daniel K wrote:
> I can't seem to find how to ask Vim to "compile" the current buffer
> the way I could in emacs (using c-c, c-c.)  What I'm hoping for is to
> run the current buffer, show the output in a split window
> (horizontal?) and move the cursor in the buffer to the first error. 
> Is there a Vim way of doing things?

Sure, you can define custom errorformats. This (untested) is perhaps a 
start:

http://www.vim.org/scripts/script.php?script_id=477

> Is there any way to integrate some python language reference to be
> accessed from within Vim?  Something like hot-key keyword lookup to
> the docs.

Not sure, but should be easy to implement if you have an external 
command that can be called with a Python keywords as argument. Perhaps 
the info version of the Python documentation?

> Is asking for keyword completion in Vim going against the grain of
> what Vim stands for?  (Simplicity)

Hmm. For starters, you could define a few abbreviations (:h 
abbreviations) for common snippets. I can type ok, so I don't need these ;-)

> Any suggestions or pointers to where I can find these answers would be
> greatly appreciated.

a) http://vim.org/ - "Scripts" and "Tips" - you'll also find 
Python-related stuff there.

b) The vim mailing list (also available via the news.gmane.org newsserver).

>  I'm really enjoying the power of Vim!!

I like vim a lot, too :-)

-- Gerhard





From stuffduff at cox.net  Sat Apr 26 22:21:32 2003
From: stuffduff at cox.net (stuffduff)
Date: Sun, 27 Apr 2003 02:21:32 GMT
Subject: Class problem: loosing my mind
Message-ID: <v0fmavkm9iknef7hit5n17d4m0qi6tpnia@4ax.com>

I'm 'doing the class pickle thing.'

I understand that both the module doing the pickling and the module
doing the unpickling need to reference the same class definition.

So I go and make a directory under site-packages named gV and in the
directory i add a file gV.py class gO and a completely empty
__init__.py

Now I have an importable module for the class, as much as I
understand.

So I fire up python, import sys and print sys.path, and the last path
is pointing to the correct site-packages.

At the python prompt I enter import gV, no error.

Next I try and create an instance:

x=gV.gO()

and I get AttributeError: 'gV' module has no attribute 'gO'

What am I missing here?

Sean
home: stuffduff at cox.net
work: swduffy at unmc.edu



From max at alcyone.com  Thu Apr 17 13:38:27 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 17 Apr 2003 10:38:27 -0700
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net> <3E9ED903.79B5F164@sympatico.ca>
Message-ID: <3E9EE693.A152E5B9@alcyone.com>

"Colin J. Williams" wrote:

> The back ticks, as in `self.value`,  are equivalent to
> repr(`self.value`).

Actually, they're equivalent to repr(self.value).  If `x` -> repr(`x`),
then repr(`x`) -> repr(repr(`x`)) -> repr(repr(repr(`x`))) -> etc.  :-)

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Freedom is a road seldom travelled by the multitudes
\__/ Public Enemy
    The laws list / http://www.alcyone.com/max/physics/laws/
 Laws, rules, principles, effects, paradoxes, etc. in physics.



From cben at techunix.technion.ac.il  Sun Apr 13 13:10:32 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Sun, 13 Apr 2003 20:10:32 +0300 (IDT)
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <mailman.1050200792.6949.python-list@python.org>
References: <mailman.1050200792.6949.python-list@python.org>
Message-ID: <Pine.GSO.4.51.0304132002000.26349@techunix.technion.ac.il>

Tim Peters wrote on 2003-04-12:

> OTOH, no type other than a string type seems to feel a pressing need for
> more than two convert-to-string gimmicks, and the basic container types
> don't even distinguish repr from str (str(list) and repr(list) act the same
> today because they are the same today).
>
Well, it is true that things will get confusing with strings.  But
when you do an str() of a list, how much do you care for it?  Perhaps
it's worth to apply str() in list.__str__, keeping list.__repr__ for
the cases when you need an unambiguos representation?  I.e. leve it up
to the user to decide whether he will be confused.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>

Never spend several weeks on a single homework - or you will miss a
nice Linux conference ;-(.




From skip at pobox.com  Wed Apr 16 08:57:37 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 16 Apr 2003 07:57:37 -0500
Subject: reading Unix mailbox
In-Reply-To: <hZbna.17434$DN.480144@tornado.fastwebnet.it>
References: <hZbna.17434$DN.480144@tornado.fastwebnet.it>
Message-ID: <16029.21313.428517.5071@montanaro.dyndns.org>

    Alessio> I am using Python2.3 alpha and I thought there was something
    Alessio> updated to iterate over the messages of a mailbox in mbox
    Alessio> format, but the next() method still return a deprecated RFC822
    Alessio> message object reference, am I wrong? Is there a way to read
    Alessio> mailboxes in mbox format but iterating over MIMEMessages?

Try:

    from spambayes import mboxutils
    mbox = mboxutils.getmbox("somefile")
    for msg in mbox:
        do_stuff(msg)

Skip




From rmunn at pobox.com  Wed Apr  2 13:53:45 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 18:53:45 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <3E88C1D8.3070305@motorola.com> <3e8a57d0.55986091@news-server> <3E8B1602.72BB1A37@juno.com>
Message-ID: <slrnb8lne2.d4t.rmunn@localhost.localdomain>

Jeff Sandys <sandysj at juno.com> wrote:
> try:
> 
> C:\python22\pythonw.exe "C:\program files\pysol\pysol.py"
> 
> or: do a google search for "download pysol.exe" 
> and download the windows executable.
> 
> 
> Re: top-posting: Blind people with text to speech converters 
> prefer top-posting so they don't have to listen to the same 
> stuff over and over again.

This article disagrees:

    http://www.blakjak.demon.co.uk/gey_chr0.htm

Summary: Usenet readers for blind people are well aware of Usenet
conventions and can handle bottom-posted messages very well, but are
thrown into confusion by top-posted messages.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From ianb at colorstudy.com  Wed Apr 16 01:16:49 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 16 Apr 2003 00:16:49 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
	 <mailman.1049571010.25625.python-list@python.org>
	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
	 <mailman.1049653811.11491.python-list@python.org>
	 <cQ_la.5597$LB6.158861@news1.tin.it>
	 <mailman.1050210449.24337.python-list@python.org>
	 <23891c90.0304140221.20e9e7fb@posting.google.com>
	 <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
	 <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net>
Message-ID: <1050470209.30727.113.camel@lothlorien>

On Tue, 2003-04-15 at 07:01, A.M. Kuchling wrote:
> On Tue, 15 Apr 2003 02:30:49 +0100, 
> 	Andy Robinson <andy at reportlab.com> wrote:
> > Second, it has always seemed to me that there is a very well
> > understood model of a web app consisting of a controller, Request,
> > Response and a few other things, handled in very similar ways by Java
> > servlets, ASP and WebKit.  
> 
> I've speculated about adding a standard HTTP{Request,Response} type to the
> standard library, but even this seems unlikely because I doubt we can agree
> on what should be in it, or what naming convention to use.  (Quixote uses
> underscore_names, WebWare uses camelCase; whose users have to change all
> their code.)

One option would be to use an interface with adapters, where if say the
Quixote interface was used, Webware would have an adapter from that
interface to the Webware HTTP{Request,Response} interface... though I'm
not sure how interesting request and response object are.  Defining the
adapter may not be much less work than reimplementing it from scratch. 
But maybe not...

> There's also the inflexibility that comes with putting something in the
> standard library; the interfaces can't be changed without breaking existing
> users, so it's impossible to fix things.  cgi.py has lots of cruft in it,
> but I've never tried to clean it up because I have no idea what applications
> out there might be broken, so cgi.py stays crufty.  It's noteworthy that
> Quixote's upload handling only got really reliable when Greg Ward wrote his
> own upload parsing code and jettisoned use of cgi.py's code.

I don't think development of the standard library would do well at
syncing with development of something like this, at least not for a
while.  It seems to me like there's some questions about batteries
included anyway, that it makes development of new functionality
unwieldy.  It's more important in making something the canonical
implementation, and maybe that's achievable through other techniques.

> All this talk of unifying Web frameworks strikes me as pointless.  Different
> toolkits have different design principles and different goals; hoping for a
> unified one is like hoping Python and Perl will settle their differences and
> merge.  

I disagree.  Unifying with one model for building applications is not
likely, but there are a lot of places where the frameworks duplicate
effort without significant differentiation.  By identifying those we
could make a more robust, more easily distributed, more compelling
foundation for web frameworks.  

  Ian






From oktaysafak at ixir.com  Mon Apr 21 11:41:42 2003
From: oktaysafak at ixir.com (Oktay Safak)
Date: Mon, 21 Apr 2003 18:41:42 +0300
Subject: Twisted 1.0.4
References: <20030421110438.4991f713.itamar@itamarst.org>
Message-ID: <006901c3081c$8240d7e0$0200a8c0@b>

I see that Buildbot is GPL'ed. I am not very knowledgable in these licensing
stuff but I wonder whether this would be a problem for Twisted users.

Oktay


----- Original Message -----
From: "Itamar Shtull-Trauring" <itamar at itamarst.org>
To: <python-list at python.org>
Cc: <python-announce-list at python.org>
Sent: Monday, April 21, 2003 6:04 PM
Subject: ANN: Twisted 1.0.4


> Twisted is an event-driven networking framework for server and client
> applications. More detailed information can be found after the "What's
> New" section.
>
> For more information, visit http://www.twistedmatrix.com, join the list
> http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python or
> visit us on #twisted at irc.freenode.net.
>
>
> What's New in 1.0.4
> ===================
>
> This release is mainly a bug fix release. 1.0.4 includes large numbers
> of bug fixes, small feature enhancements, additional tests and so on.
> Additional reliability is assured by our deployment of Brian Warner's
> Buildbot (http://buildbot.sf.net) for real-time testing
> (http://twistedmatrix.com/~warner.twistd/ -- note that certain
> functionality is experimental and therefore the tests will be failing,
> but this does not affect the stability or performance of supported,
> stable packages).
>
>
> What is Twisted?
> ================
>
> Twisted is an event-driven framework for building networked clients and
> servers.  It contains a powerful and simple networking core, a
> full-featured suite of interoperable protocols, among them a powerful
> web server and applications framework.
>
> Twisted supports many event loops for both server apps and GUI
> integration on the client side, including:
>
>  - Win32 events, including GUI support
>  - GTK+
>  - GTK+ 2
>  - Qt
>  - wxPython
>  - Tkinter
>  - Java
>
> Twisted can run protocols over TCP, SSL, UDP, multicast, Unix sockets
> and subprocesses. It also includes scheduling support, threading
> integration, RDBMS event loop integration and other basic requirements
> for networked applications.
>
> Also included are implementations of many protocols. In some cases this
> includes complete frameworks providing facilities on top of the base
> protocol:
>
>  - SSH
>  - FTP
>  - HTTP, including a complete web framework
>  - XML-RPC
>  - SOAP server framework
>  - NNTP and complete NNTP server framework
>  - SOCKSv4 (server only)
>  - SMTP
>  - IRC
>  - telnet
>  - POP3
>  - AOL's instant messaging TOC
>  - OSCAR, used by AOL-IM as well as ICQ (client only)
>  - DNS
>  - MouseMan serial mice, and GPS devices
>  - Twisted Perspective Broker, a remote object protocol
>
> --
> http://mail.python.org/mailman/listinfo/python-list
>





From hst at empolis.co.uk  Tue Apr  1 03:08:29 2003
From: hst at empolis.co.uk (Harvey Thomas)
Date: Tue, 1 Apr 2003 09:08:29 +0100
Subject: Regular expression help
Message-ID: <8FC4E7C302A6A64AAD5DB1FA0E825DEB53BC97@hendrix.empolisuk.com>

Stephen Boulet wrote:
> How do I use the sub function to replace my match with my match 
> prepended by a newline?
> 
> Thanks.
> 
> -- Stephen
> 
Take advantage of the fact that the first argument of the sub method can be a function whose single argument is a match object:

(Untested)

def mysub(x):
    return '\n' + x.group(0)

myregex.sub(mysub, mystring)


or

myregex.sub(lambda x: '\n' + x.group(0), mystring)


For something as simple as this, I'd probably prefer the lambda form.

Harvey

_____________________________________________________________________
This message has been checked for all known viruses by the MessageLabs Virus Scanning Service.




From withers_jp at yahoo.com  Mon Apr 28 14:01:25 2003
From: withers_jp at yahoo.com (John Withers)
Date: 28 Apr 2003 11:01:25 -0700
Subject: Zope, Relative Object Calls
Message-ID: <f1844358.0304281001.41a2d4b2@posting.google.com>

I am trying to create object calls in Zope that are relative and can't
figure out how to do it.

In the application I am working on, tags like the following appear:
<dtml-var "apps.profile.forms.rpt_fetchallrefs(_.None,_,aboutnum=_['serialnum'])">

And this rev of my application lives at /apps/profile.

Now, I am trying to create multiple versions of this application in an
itterative development processs, so the next version lives at
/apps/profile_01, then /apps/profile_02 and so on. The problem is,
this call hardwires me to /apps/profile/forms/rpt_fetchallrefs. There
are quite a few of these calls and I don't want to ahve to change them
manually for each version we are doing.

I am trying to get something like a relative anchor, where:
<a href="forms/rpt_fetchallrefs">.</a> would work if I were copying
everything into a new directory or renaming directories further up the
structure.

I suspect that this is a pretty simple thing I am somehow missing,
since it seems an obvious and utilitarian feature.

Thanks for your help.

John P. Withers



From tjg at craigelachie.org  Thu Apr 17 00:42:27 2003
From: tjg at craigelachie.org (Timothy Grant)
Date: Wed, 16 Apr 2003 21:42:27 -0700
Subject: How many of you are Extreme Programmers?
In-Reply-To: <1147e466.0304160630.5b510724@posting.google.com>
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <200304162142.27375.tjg@craigelachie.org>

On Wednesday 16 April 2003 07:30 am, Christopher Blunck wrote:
>
> So how many of you guys use XP processes?

I work for a place that uses XP. Unfortunately, all our stuff is done in Perl 
(with a home-rolled testing engine that actually works extraordinarily well).

I'm gradually doing some utility stuff in Python (without a pair, but with 
extraordinarily good tests--I love Beck's new book: 
_Test_Driven_Development_.)

-- 
Stand Fast,
    tjg.

Timothy Grant
www.craigelachie.org





From escalation746 at yahoo.com  Thu Apr  3 14:46:39 2003
From: escalation746 at yahoo.com (robin)
Date: 3 Apr 2003 11:46:39 -0800
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
Message-ID: <a626bbd7.0304031146.6f98fc28@posting.google.com>

I wrote:
> If you want the biggest, boldest approach and don't care about
> overhead, use CORBA.

martin at v.loewis.de (Martin v. L?wis) wrote:
> I can't agree with that analysis: CORBA has, of all your alternatives,
> the least network bandwidth requirements.

I was referring to overhead like:
a) writing the interface definition and dealing with ID strings
b) more code 
c) size of the "library"

I am quite sure the other implementations of DC are lighter in these
areas, though for some that may not be an issue. I was not considering
network bandwidth, but am glad to hear CORBA is efficient in that
respect.

-- robin



From nirina at mail.blueline.mg  Fri Apr 11 06:46:19 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Fri, 11 Apr 2003 14:46:19 +0400
Subject: Python 2.2.2 & Tix on win32 - What must I do?
References: <b7666e$oto$1@newstree.wise.edt.ericsson.se>
Message-ID: <00a801c30017$efc4b640$fba2383e@raseliar>

"Frithiof Andreas Jensen" wrote:

>   >>> import Tix
>   >>> root = Tix.Tk()
>   Traceback (most recent call last):
>     File "<pyshell#1>", line 1, in ?
>       root = Tix.Tk()
>     File "C:\Python22\lib\lib-tk\Tix.py", line 213, in __init__
>       self.tk.eval('package require Tix')
>   TclError: can't find package Tix
>   >>>
> "
> What must I *do* exactly to fix this?
> 

You must have Tix widgets installed.
http://tix.sourceforge.net/
and maybe also the Tix library

nirinA





From martin at v.loewis.de  Fri Apr  4 18:59:06 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Sat, 05 Apr 2003 01:59:06 +0200
Subject: unicode bug in turkish characters?
In-Reply-To: <mailman.1049401867.14209.python-list@python.org>
References: <mailman.1049401867.14209.python-list@python.org>
Message-ID: <b6l685$14j$05$1@news.t-online.com>

oktaysafak at ixir.com schrieb:
>>>>locale.setlocale(locale.LC_ALL, "tr_tr")
> 
> Traceback (most recent call last):
>   File "<pyshell#3>", line 1, in ?
>     locale.setlocale(locale.LC_ALL, "tr_tr")
>   File "C:\PYTHON23\lib\locale.py", line 381, in setlocale
>     return _setlocale(category, locale)
> Error: locale setting not supported
> 
>>>>locale.setlocale(locale.LC_ALL, "turkish")
> 
> 'Turkish_Turkey.1254'

Ah, I see. Windows. My example was from Linux, where tr_TR is a
valid locale name.

Python delegates the toupper call to the operating system. It does not,
in itself, include a database of lower/upper case conversions for all 
locales.

So there is nothing we can do about that; ask Microsoft.

> Locale defaults as determined by getdefaultlocale():
> ------------------------------------------------------------------------
> Language:  tr_TR
> Encoding:  cp1254

Please understand that getdefaultlocale does *not* determine
the locale used by the program. Instead, it determines (sometimes)
the locale that Python would use if you had invoked setlocale
with "", i.e. asking for the user's preference (sometimes,
getdefaultlocale just doesn't work).

Regards,
Martin




From Patrick.Bussi at space.alcatel.fr  Tue Apr  8 04:49:41 2003
From: Patrick.Bussi at space.alcatel.fr (Patrick.Bussi at space.alcatel.fr)
Date: Tue, 8 Apr 2003 10:49:41 +0200
Subject: Headers issue with urllib2 and ClientCookie (was : ClientCookie
	.read() failing on some servers )
Message-ID: <C1256D02.00309E78.00@vzmta01.netfr.alcatel.fr>



I did not receive any answer, but, if any, it could have been "RTFM", because
the excellent ClientCookie.__doc __ addresses the issue. Answers to my own
questions are :

A/ Yes definitely. Some server  are sensitive to (un)correct header, while
others are not.
B/ My previous program attempted to add User-agent header as an ordinary header,
and it omitted the 'referer'.

As a record, here is an extract of the docstring :

------snip------
Sometimes, a server wants particular headers set to the values it expects, or it
won't play nicely.  The most  frequent offenders here are the Referer [sic] and
/ or User-Agent HTTP headers.
[...]
urllib2.OpenerDirector automatically adds a User-Agent header to every request.
Since ClientCookie.urlopen uses an OpenerDirector instance, you need to install
your own OpenerDirector using the ClientCookie.install_opener function to change
this behaviour :
>> opener.addheaders = [("User-agent",
"Mozilla/5.5.(X11;.U;.Linux.2.4;.en-US;.0.8).Gecko/20010409")]
[..]
If things don't seem to be working as expected, the first thing to try is
to switch off RFC 2965 handling, using the netscape_only argument to the
Cookies constructor.  This is because few browsers implement it, so it is
likely that some servers incorrectly implement it.  This switch is also
useful because ClientCookie does not yet fully implement redirects with
RFC 2965 cookies.
------snip------

Here is the program working :

------snip------
#! /usr/bin/env python
'''usage:
         $ python test111-03.py www.python.org  > result.log 2>&1

'''

def test(h):
    '''h is the host name. Caution : no protection against mistakes
    '''
    import ClientCookie, urllib2
    ClientCookie.HTTP_DEBUG = 1
    ClientCookie.CLIENTCOOKIE_DEBUG = 1
    cookies = ClientCookie.Cookies(netscape_only=1)
    hh = ClientCookie.HTTPHandler(cookies, handle_refresh=1)
    opener = ClientCookie.build_opener(hh)
    opener.addheaders = [("User-agent",
"Mozilla/5.5.(X11;.U;.Linux.2.4;.en-US;.0.8).Gecko/20010409")]
    ClientCookie.install_opener(opener)

    req = urllib2.Request('http://'+h)
    req.add_header("Referer", 'http://'+h+'/') # ugly. I know
    req.add_header('Accept','*/*')
    req.add_header('Accept-Language','en')
    req.add_header('Accept-Encoding','gzip,deflate,compress,identity')
    req.add_header('Keep-Alive','300')
    req.add_header('Connection','keep-alive')

    response=ClientCookie.urlopen(req)
    print '\nResponse from server %s :' % response.geturl()
    print response.info()
    for line in response.readlines(): print line

if __name__ == '__main__':
    import os, sys
    print 'Linux', ''.join(['%s'*3 %os.uname()[2:]])
    print 'Python', sys.version
    try: h=sys.argv[1]
    except: sys.exit(__doc__)
    test(h)
------snip------


 More, there was a mail in this list-archive
(http://mail.python.org/pipermail/python-list/2003-February/150438.html) which
addressed issue on http_error_302, while my error message was :
urllib2.HTTPError: HTTP Error 302: The HTTP server returned a redirect error
that wouldlead to an infinite loop.
May be the issue raised by Akai would be solved by using ClientCookie  which
seems to manage redirections well.

Thanks to the list anyway.

Patrick.


---------------------- Envoy? par Patrick Bussi/ALCATEL-SPACE le 08/04/2003
09:34 ---------------------------

Patrick Bussi   07/04/2003 11:03
(Embedded image moved to file: pic06482.pcx)

Pour :    python-list at python.org
cc :
Objet :   Headers issue with urllib2 and ClientCookie (was : ClientCookie.read()
      failing on some servers )



A/ does the wrong headers justify the failure in server response ?
B/ what is wrong in my test program headers ?



---
Patrick Bussi
patrick.bussi at space.alcatel.fr


Any opinions expressed are my own and not necessarily those of my Company.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pic06482.pcx
Type: application/octet-stream
Size: 4787 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030408/b688d9f7/attachment.obj>

From tipton51 at subdimension.moc  Thu Apr  3 13:18:08 2003
From: tipton51 at subdimension.moc (Tipton Bandy)
Date: Thu, 03 Apr 2003 18:18:08 GMT
Subject: newbie os.stat question
Message-ID: <AV_ia.7382$4P1.570905@newsread2.prod.itd.earthlink.net>

I've got my first python utility working pretty good, but
I just need help on one point.

I need to test a filename and return true or false based on 
whether the file is readable and writable by the user running
my utility. Oh, yes, this is on Linux:)
I'm used to writing things like, 'if [ -w $filename ]' in bash.

I think I should use os.stat() somehow, but it seems like a lot
of code would be required to check the appropriate field (user,
group, or others) based on the ownership of the file and the
uid of the user.

This seems like such a common thing.  Is there a pre-written function
to handle it?

Thanks in advance,
Tipton



From max at alcyone.com  Sat Apr  5 04:27:51 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sat, 05 Apr 2003 01:27:51 -0800
Subject: In memory File
References: <b6m741$dqo$1@news-reader12.wanadoo.fr>
Message-ID: <3E8EA197.3AE807EB@alcyone.com>

Salvatore wrote:

> How can I create file-like objects ?
> I want to work in memory only.

You want the StringIO module.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The graveyards are full of indispensible men.
\__/ Charles de Gaulle
    Fauxident / http://www.alcyone.com/pyos/fauxident/
 A "faux" ident daemon in Python.



From peter at engcorp.com  Fri Apr 18 23:09:11 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 18 Apr 2003 23:09:11 -0400
Subject: Are reference cycles bad form?
References: <3o11avop6vkh3ake8dgbtst93t1or771ct@4ax.com>
Message-ID: <3EA0BDD7.2E08771D@engcorp.com>

Manuel Garcia wrote:
> 
> But as a general pricipal, is a design without reference cycles better
> than a design with them?

Roy is right, of course, but putting on blinders and answering the
question asked above, out of context, I would say that coupling is
_as a general rule_ something that worsens a design the more of it
there is, and based solely on that premise a reference cycle is 
worse than something that gets by without the cycle.  On the other
hand, if the code has to jump through hoops to avoid them, go
back to Roy's answer. :-)

-Peter



From duncan at NOSPAMrcp.co.uk  Tue Apr  8 11:38:48 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 8 Apr 2003 15:38:48 +0000 (UTC)
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com> <b6soav$7p2$1@panix1.panix.com> <Qdpka.2889$Zu2.2627@news01.roc.ny.frontiernet.net> <a8b7f07a.0304080603.473a8d90@posting.google.com> <WRBka.134$jS3.16@news02.roc.ny.frontiernet.net>
Message-ID: <Xns9357A93AF1C62duncanrcpcouk@127.0.0.1>

Lee Harr <missive at frontiernet.net> wrote in 
news:WRBka.134$jS3.16 at news02.roc.ny.frontiernet.net:

> Should this be showing
> Af
> Bf
> ?
> 
> How would I even get that output?  I know I can do:

Try this:
>>> class A(object):
	def f(self):
		print 'Af'
		if hasattr(super(A, self), 'f'):
			super(A, self).f()

			
>>> class B(object):
	def f(self):
		print 'Bf'
		if hasattr(super(B, self), 'f'):
			super(B, self).f()

			
>>> class C(A,B):
	def f(self):
		print 'Cf'
		super(C, self).f()

		
>>> c = C()
>>> c.f()
Cf
Af
Bf
>>> a = A()
>>> a.f()
Af
>>> 

Note that A.f() calls B.f() if self is a C, but doesn't call anything if 
self is an A.

Its easier if you want to propogate a function such as __init__ which you 
know is in the base class, because then you don't have to check for 
existence. Another way might be to define a class that is object with an 
empty f and use that as the base for both A and B.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From mwm at mired.org  Sun Apr  6 13:14:11 2003
From: mwm at mired.org (Mike Meyer)
Date: Sun, 06 Apr 2003 17:14:11 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain>
Message-ID: <x7smsvh7kr.fsf@guru.mired.org>

Robin Munn <rmunn at pobox.com> writes:
> carroll at tjc.com <carroll at tjc.com> wrote:
>     # Is it a file-like object?
>     try:
>         output_file.write
>         print "output_file is a file-like object"
>     except AttributeError:
>         # No: it doesn't have a write() method. Is it a string?
>         try:
>             output_file.lstrip
>             print "output_file is a string-like object"
>         except AttributeError:
>             # Neither a file nor a string
>             raise ValueError, "output_file should be a file or a string"

If you're just going to check the attribute and not use it, why not
use hasattr?

        if hasattr(output_file, 'write'):
            print "output_file is a file-like object"
        elif hasattr(output_file, 'lstrip'):
            print "output file is a string-like object"
        else:
            raise ValueError, "output_file should be a file or a string"
          

> Third, I'm not sure I like the functionality of passing a string
> filename. It sounds convenient, but all it really does is *add*
> ambiguity to your function which could be avoided. Just let that
> parameter be a file-like object that's already been opened for writing,
> and do *not* close it when you're done. If the user really wants the
> functionality you provide (open filename, write, close it when done),
> they can do it as easily as this:

That's excellent advice, but...

>     xover(start, end, open('foo.txt', 'w'))
> 
> See? File-like object gets created, passed to your function, and then
> automatically garbage-collected (and thus closed) when your function
> returns.

The garbage collection happening when your function returns is
implementation specific. It works that way in CPython, but not
Jython. You should always explicitly close files to make sure they get
closed - and the output flushed, in this case - when you're done with
them.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From news at datatailors.xs4all.nl  Wed Apr 16 05:36:11 2003
From: news at datatailors.xs4all.nl (Peter van Kampen)
Date: Wed, 16 Apr 2003 11:36:11 +0200
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org>
Message-ID: <slrnb9q90a.mhg.news@datatailors.com>

In article <mailman.1050457621.2105.python-list at python.org>, Mike C.
Fletcher wrote:
> Andrew Koenig wrote:
> 
> Okay, so now we have "The Practical Coder's Guide to big-O Notation" :) 
> .  We rock ;) ,
> Mike

You do, you know...c.l.p really does.

I think I almost 'get it'. Except who or what decides what Big-O a
certain algorithm has. Is that by 'hunch' or by running benchmarks or
are there still other ways to recognize 'Big-O's (except for the
formal computations that I understand are not very common)? 

for example

bigO_N(i):
    for n in xrange(i):
        print n

Would be O(N) right?

bigO_N2(i,j):
    for n in xrange(i):
        for o in xrange(j):
            print i,j

This would be O(N**2)? Actually O(i*j) I suppose but if i==j N**2
applies.

But these are not really 'algorithms'. Most algorithms are a lot more
complex even when reduced to it's essentials. I understand why
quicksort is better than bubblesort but I don't see an obvious way to
label quicksort O(??). 

PterK

-- 
Peter van Kampen
pterk -- at -- datatailors.com



From http  Wed Apr  9 12:13:07 2003
From: http (Paul Rubin)
Date: 09 Apr 2003 09:13:07 -0700
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <7x7ka3ejjg.fsf@ruckus.brouhaha.com>

iamshady at rediffmail.com (Jim Shady) writes:
> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string. For
> example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?

def longest_str(s):
  lengths = [len(x) for x in s.split('/')]
  return max(lengths)



From ianb at colorstudy.com  Sun Apr 13 01:06:54 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 13 Apr 2003 00:06:54 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <cQ_la.5597$LB6.158861@news1.tin.it>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <mailman.1049530929.9225.python-list@python.org>
	 <a3b5b608.0304051049.7048d76@posting.google.com>
	 <mailman.1049571010.25625.python-list@python.org>
	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
	 <mailman.1049653811.11491.python-list@python.org>
	 <cQ_la.5597$LB6.158861@news1.tin.it>
Message-ID: <1050210413.16374.609.camel@lothlorien>

On Sat, 2003-04-12 at 15:39, Alex Martelli wrote:
> A Webkit version running on top of Twisted would really be a wonderful
> thing to have... I have a hard time conceiving of exactly how you'd fit
> them together (maybe I'm not familiar enough with the _internals_ of
> each, though I've used them both), but I'll accept your assessment that
> it would be a pretty easy job for you, so, what about trying it out...?

It should be pretty simple, really.  WebKit has a single entry point
that takes an argument with the CGI-like variables for the request, an
input stream (e.g. for the POST data), and an output stream (with a
write method) that accepts CGI-like output (i.e., it may add headers,
translate Status into a response code, etc. -- but simple stuff).  This
entry point is expected to run in a thread of its own.

Using the Twisted CGI code, it would be a quick change to make it
dispatch to a thread instead of a process (as with CGI).

The other detail is initializing the WebKit environment -- the
persistent Application object, plugins, etc.  That would be easy to do
in a simplistic way, maybe more subtle if I was to use a more Twisted
way.

I'm still thinking about the strategy around it, which is why I haven't
actually tried implementing it.  As an experiment it's not that
interesting to me, but maybe it could be more than that.

  Ian






From mertz at gnosis.cx  Tue Apr 29 16:28:00 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 29 Apr 2003 16:28:00 -0400
Subject: Python 2.3b1: RuntimeError using rexec
References: <b8mh4l$2oo0$1@nntp6.u.washington.edu>
 <m31xzl6pac.fsf@mira.informatik.hu-berlin.de> <45tr+kKkXstQ092yn@gnosis.cx>
Message-ID: <QBur+kKkXUXP092yn@gnosis.cx>         

|"Russell E. Owen" <no at spam.invalid> writes:
|> I stumbled across this when testing some existing code. This code uses
|> rexec to convert a string representation of a dictionary to a dictionary
|> as part of reading in font preferences.
|> Any suggestion on how to do that without rexec would also be
|> appreciated. (I was quite disappointed that "dict" didn't do it!)

I'm not sure exactly what string representation of a dictionary you use.
But it's not too hard to use dict().  E.g.:

  >>> d = '''
  ... a:b
  ... c:d
  ... e:f
  ... '''
  >>> dict([x.split(':') for x in d.split()])
  {'a': 'b', 'c': 'd', 'e': 'f'}

Or:

  >>> dct = "{'a': 'b', 'c': 'd', 'e': 'f'}"
  >>> dict([map(str.strip, x.split(':'))
  ...       for x in dct[1:-1].replace("'",'').split(',')])
  {'a': 'b', 'c': 'd', 'e': 'f'}


--
 mertz@  _/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: \_\_\_\_    n o
gnosis  _/_/             Postmodern Enterprises            \_\_
.cx    _/_/                                                 \_\_  d o
      _/_/_/ IN A WORLD W/O WALLS, THERE WOULD BE NO GATES \_\_\_ z e





From lafou at wanadoo.fr  Wed Apr  2 15:59:59 2003
From: lafou at wanadoo.fr (Philippe =?ISO-8859-1?Q?Lafoucri=E8re?=)
Date: Wed, 02 Apr 2003 22:59:59 +0200
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org>
Message-ID: <b6fj57$626$1@news-reader11.wanadoo.fr>

danny at iamlearning.com wrote:

> I'm new on this, please bear with we.
> 
> Let's say I have this script:
> 
> --------------------------
> #!/usr/bin/env python
> x = 'Hello Python'
> 
> print x,
> 
> --------------------------
> 
> How can I compile this script to *.pyo?
> 
> Thanks in advance,
> 
> Danny


python -O yourscript.py



From aleax at aleax.it  Thu Apr 24 13:31:10 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 19:31:10 +0200
Subject: Question about accessing class-attributes.
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DE0E9@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE0E9@admin56.narex.com>
Message-ID: <200304241931.10969.aleax@aleax.it>

On Thursday 24 April 2003 05:35 pm, Bjorn Pettersen wrote:
   ...
> > Personally, I'd define a custom metaclass for this purpose.  Maybe
> > I'm getting TOO familiar with those, but, it doesn't sound all that
> > complex at all to me...:
>
> [...]
>
> I would too, but just for fun I thought "how about an abstract base
> class?". Due to a 'thinko' I didn't consider that B wasn't calling its
> superclass __init__, so I came up with:
>
>   class AbstractBaseCounter(object):
>       __InstCount = {}
>       def InstCount(cls):
>           return AbstractBaseCounter.__InstCount[cls]
>       InstCount = classmethod(InstCount)
>
>       def __init__(self):
>           cls = self.__class__
>           if cls == AbstractBaseCounter:
>               raise TypeError("Attempt to instantiate abstract base
> class")
>           d = AbstractBaseCounter.__InstCount
>           if cls in d:
>               d[cls] += 1
>           else:
>               d[cls] = 1
>
>   class A(AbstractBaseCounter):
>       def __init__(self):
>           AbstractBaseCounter.__init__(self)
>
>   class B(A): pass
>
>   a1=A()
>   a2=A()
>   b1=B()
>   b2=B()
>   print B.InstCount()

comparing the complication of this one, and the simplicity of the custom
metaclass I posted and you snipped, PLUS the minor risks such as
"B overriding __init__ and forgetting to call its superclass's"(:-), I think
the custom metaclass is vindicated;-).

> which happens to work...? I've verified that
> AbstractBaseCounter.__init__ gets called for B(). Is this
> correct/documented/guaranteed/works with MI/multiple args/old-style
> classes, etc., etc.?

well, you COULD get dual calls to that __init__ in presence of
"inheritance diamonds" -- using super may be safer.


Alex





From scott_list at mischko.com  Tue Apr 22 12:34:37 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Tue, 22 Apr 2003 09:34:37 -0700
Subject: Can someone send me the latest Gadfly?
Message-ID: <200304220934.38013.scott_list@mischko.com>

www.chordate.com appears to be seriously down.  Their DNS is resolving to 
ns1.rackshack.net.  I called Rackshack and they don't have a zone file for 
chordate.com.

Chordate hosts the Gadfly database and I would like a copy of it on my RH8 
system but can't download it.  Google isn't helpful in finding another 
location to get it from.

Can someone help out here?

Thanks!
Scott





From johnwadeunderwood at yahoo.com  Tue Apr 29 16:18:45 2003
From: johnwadeunderwood at yahoo.com (John Underwood)
Date: 29 Apr 2003 13:18:45 -0700
Subject: python , activex control , vba
Message-ID: <f004e9c2.0304291218.864fbdb@posting.google.com>

I have an activex control to an application that opens files that can
have vba code in them.. the ActiveX Container supports vba.. When I
use the control in python to open the file, it works if I strip the
file of any vba code, but causes an error if the vba code exists.. I'm
trying to figure out if I need support for vba for my controls that
support them... Anyone else run into anything like this with a activex
control to an app that has VBA in it? I'm hoping I 'm just missing
something I don't know about, and it's not the control.. ha

John Underwood..

thanks



From gilles.celli at ecgs.lu  Thu Apr 10 04:08:38 2003
From: gilles.celli at ecgs.lu (Gilles)
Date: Thu, 10 Apr 2003 10:08:38 +0200
Subject: pyserial newbie questions
Message-ID: <gilles.celli-943DC4.10083810042003@webmail.restena.lu>

Hello folks,

Since I am very new to Python I have the following question:

I try to write a program in Python (on RHat 8.0 Linux) with the 
"pyserial" module which communicates to a device via RS232 serial cable.

The program should do:

1. Set Com port  to 300,n,8,1
2. Send 20 NULL characters Chr(0)
3. Wait 10 ms ( may vary in your case)
4. Send a string "HELLO"
5. Wait to receive ?READY? string  from device
6. If above string OK SEND   ?START?

I figured out how to set the port, send the 20 NULL characters :)

But I have no idea how to "wait for 10ms", send the HELLO string and 
read the READY string...

Anyone has a clue or source code sample ?

Thanks a lot.

Gilles



From aleax at aleax.it  Sat Apr 12 06:32:05 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 10:32:05 GMT
Subject: readline clear buffer
References: <slrnb9edf0.tds.luc@grizzly.dont-panic.info>
Message-ID: <FWRla.4654$T34.116245@news2.tin.it>

luc wastiaux wrote:

> Hello,
> when using readline, I can retrieve the contents of the line buffer using
> get_line_buffer()
> 
> but is there a way to clear this buffer ? none of the provided functions
> seem to allow that: http://www.python.org/doc/lib/module-readline.html

Right: it seems that Python's readline.c does not expose interfaces to
rl_delete_text, rl_kill_text, or other GNU Readline functions that would
allow such behavior.  I suspect the simplest solution would be to patch
readline.c to add such interfaces (also patch the docs accordingly), and
submit the patch to sourceforge, where I think it would be welcome (and
still in time to get into Python 2.3, for a modest addition, I think).


Alex




From martin.vonloewis at hpi.uni-potsdam.de  Tue Apr 29 07:36:34 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Tue, 29 Apr 2003 13:36:34 +0200
Subject: Tkinter exit problem in Win98
References: <Pine.GSO.4.21.0304291934400.25385-100000@supreme.pcug.org.au>
Message-ID: <b8lo47$nc$1@zeppelin.rz.uni-potsdam.de>

> Has anyone encountered a problem with Tkinter apps not shutting down
> properly in Wi98?

That is the oldes Python-Bug, see www.python.org/sf/216289
Nobody has been able so far to study in detail what is causing the problem
and how it could be fixed.

Most likely, the bug will be fixed the day Python stops supporting W9x.

Regards,
Martin





From mwh at python.net  Fri Apr  4 08:09:17 2003
From: mwh at python.net (Michael Hudson)
Date: Fri, 4 Apr 2003 13:09:17 GMT
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122> <3E8CF4B5.1E221DD7@alcyone.com> <9383481e.0304040445.a3752dd@posting.google.com>
Message-ID: <7h3d6k2jtoi.fsf@pc150.maths.bris.ac.uk>

drew at astro.pas.rochester.edu (Drew Moore) writes:

> I'd like this PEP to go one of two ways..
> either implementation,
> or a good solid technical reason why its a bad idea.

How about: it involves changing the language into something it's not.

For a more technical sounding reason, how do you cope with the fact
that `=' can both establish and change a binding?

Cheers,
M.

-- 
  If you don't have friends with whom to share links and conversation,
  you have social problems and you should confront them instead of
  joining a cultlike pseudo-community.    -- "Quit Slashdot.org Today!"
   (http://www.cs.washington.edu/homes/klee/misc/slashdot.html#faq)



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Wed Apr 30 19:19:22 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Thu, 01 May 2003 01:19:22 +0200
Subject: How to make a repeatedly timer with 5ms precision?
In-Reply-To: <mailman.1051736974.9481.python-list@python.org>
References: <mailman.1051736974.9481.python-list@python.org>
Message-ID: <3eb05ae2$0$49098$e4fe514c@news.xs4all.nl>

??? wrote:
> I am trying to make a repeatedly timer with high precision, the following
> program 'countdown.py' is an example which shows the way how I tried to
> make it. The problem is that it doesn't work properly with tick set to 5ms
> or smaller.

You might want to try your luck with select:

import select
select.select([],[],[],timeout)

where timeout can be small, 0.005 in your case (5 ms).
Don't know about select's precision, and the above almost
certainly doesn't work on Windows.

--Irmen




From t.evans at paradise.net.nz  Wed Apr  9 07:43:59 2003
From: t.evans at paradise.net.nz (Tim Evans)
Date: Wed, 9 Apr 2003 11:43:59 GMT
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
Message-ID: <87znmzrj40.fsf@cassandra.evansnet>

tweedgeezer at hotmail.com (Jeremy Fincher) writes:

> The topic really says it all -- I'd like to *force* repr(aString) to
> use double quotes.  Is that possible?
> 
> If it's not, then is it possible to achieve the same thing (everything
> escaped properly, etc.) through other means?
> 
> Thanks,
> Jeremy

def drepr(s):
    return '"' + s.encode('unicode_escape').replace('"', '\\"') + '"'

-- 
Tim Evans



From mike at bindkey.com  Mon Apr 21 19:04:40 2003
From: mike at bindkey.com (Mike Rovner)
Date: Mon, 21 Apr 2003 16:04:40 -0700
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <mailman.1050953314.21065.python-list@python.org> <b81od0$gcc$1@news.hccnet.nl>
Message-ID: <b81tg9$hhs$1@main.gmane.org>

The Python problem with continuation is IMHO they downgrade the language
level instead of upgrading it. They are so 'basic' that, like macros, tend
to make Python too fluid and thus unmaintainable and bad for understanding.

>From the other hand most agree that they are a brilliant idea which may lead
to some good extensions and abilities.

So from recent mentioned "On Lisp" book "The best programming language
abstractions save not just typing, but thought."

In my opinion continuations are good not per se but for their abilities to
build better abstractions that leverage our thoughts, which may be state
machine or a tree structure as a language primitives.

So the whole discussion may be split to implementation part (and moved to
comp.python.devel) and concept abstraction part which unfortunately was not
even started (that shows the subject immaturity).

my 0.02

Mike







From staschuk at telusplanet.net  Sat Apr  5 13:50:18 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 5 Apr 2003 11:50:18 -0700
Subject: constructor inquiry
In-Reply-To: <59e5b87.0304040917.12c15907@posting.google.com>; from mirandacascade@yahoo.com on Fri, Apr 04, 2003 at 09:17:00AM -0800
References: <59e5b87.0304040917.12c15907@posting.google.com>
Message-ID: <20030405115018.B371@tibia.amotlpaa.bogus>

Quoth Miranda Evans:
  [...]
> My questions are:
> 1) is it correct that the constructor for the xlEv class never got
> invoked?  if so, is it correct to infer that in the sample code below,
> at no time does an object of class xlEv get instantiated?

Certainly your code doesn't instantiate one.  I'm not sure about
this line:

> 	xl = win32com.client.DispatchWithEvents("Excel.Application", xlEv)

Since I know nothing about win32com, I don't know whether this can
be expected to instantiate the class you've passed in.

If you wanted to create an instance yourself here, you need xlEv()
instead of xlEv.

> 2) is it correct to infer that a class method (e.g.
> OnSheetBeforeDoubleClick) can get invoked without an instance of the
> class to which that method belongs getting instantiated?

Class methods can indeed be invoked without an instance; that's
the whole point of class methods.  But I think you wanted
OnSheetBeforeDoubleClick to be an instance method.

> 3) is the 'xl' object in the demo() method an object of class xlEv? 
> if so, how could the xl object come into existence without the
> constructor getting invoked?  if not, how could the xl object have
> visibility to the type attribute?

You can start learning the answers to these questions by adding
	print type(xl)
	print isinstance(xl, xlEv)
at an appropriate point.  What do these lines produce?

-- 
Steven Taschuk                  staschuk at telusplanet.net
"Telekinesis would be worth patenting."  -- James Gleick




From drs at ecp.cc  Wed Apr  9 12:58:56 2003
From: drs at ecp.cc (drs)
Date: Wed, 09 Apr 2003 16:58:56 GMT
Subject: Python & Remote Access Database
References: <a6b6f195.0304090408.21003c3a@posting.google.com>
Message-ID: <kjYka.2914$kd1.2980829@newssrv26.news.prodigy.com>

"mat" <mathieu.rohon at bull.net> wrote in message
news:a6b6f195.0304090408.21003c3a at posting.google.com...
> hi everyone!
>
> I have some trouble to find a way to connect and find informations in
> a Remote Access Database, from my Python program which running under
> linux Red Hat 7.3.
>
> the fact are:
> -the access Database is on a WinNT4 server
> -my python script is under Linux Red Hat 7.3, with Python 1.5
> -both pc are ine the same domain
>
> I saw different module : ODBCmx, win32com... but i don't know which
> one to use, and how to use it.
>
Assuming that by access you mean MS Access, I think there are 3 solutions.
The real problem is that Access is not a db server, so using it remotely,
even in an all windows environment, is not a simple thing (or at least not a
good thing).

The first is solution is to use something like pyro to construct a remote
object that runs on the windows box but is controled by the linux box. (i
use a similar solution to this in places but between 2 win2k computers using
dcom and so far it seems okay for processes that are not too resource
intensive).  If you do this look at http://ecp.cc/pyado.html for
instructions on using Python win32com/ADO.

A second similar solution is to use medusa, or something similar (twisted,
mod_python, iis/asp, simplehttpserver, whatever) to send pickled stuff back
and forth from the remote computer.  this is probably the most work in terms
of hours, but perhaps the least in terms of thinking.

The last solution, which is politically often the most difficult, but which
is probably the technically correct one, is to switch databases and use
MySQL or Postgres or whatever you like that can actually act as a db server.
(There is also a baby version of sql server which comes with some versions
of visual studio that might work -- people's results have been mixed.)

-d





From andersjm at inbound.dk  Thu Apr  3 17:11:20 2003
From: andersjm at inbound.dk (Anders J. Munch)
Date: Fri, 4 Apr 2003 00:11:20 +0200
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org> <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk> <x7el4jibzl.fsf@guru.mired.org>
Message-ID: <Kp2ja.2589$y3.426276@news010.worldonline.dk>

"Mike Meyer" <mwm at mired.org> skrev:
> "Anders J. Munch" <andersjm at dancontrol.dk> writes:
> > "Mike Meyer" <mwm at mired.org> wrote:
> > > I disagree with that statement. Clearly, *someone* knows all the
> > > coding tricks used by the compiler - because they put it into the
> > > compiler. So they can achieve the same level of performance.
> > And chess program authors can beat their own inventions?  Not so.
>
> Go program authors can beat their own inventions, so clearly this
> depends on the problem domain. In the chess (and go) problem domains,
> the time the human spends on the problem is artificially limited by
> the rules of the game. For programming, limiting the programmer time
> is a goal, not an axiom.

Here's a brilliant algorithm for register allocation that I just stole
from the PSU labs.

1. Generate all possible combinations.
2. Score each of them according to estimated performance.
3. Pick the best scoring combination.

Now try with 8 registers and 10 variables.  1,073,741,824 variations.

No limit on programmer time?

please-share-the-secret-of-eternal-life-with-the-rest-of-us-ly y'rs,
Anders






From clj at acme.com  Mon Apr 14 21:27:13 2003
From: clj at acme.com (Chris Jones)
Date: 14 Apr 2003 21:27:13 -0400
Subject: Complexity of standard Python data structures
References: <mailman.1050354988.8671.python-list@python.org>
Message-ID: <tdnk7dwlfda.fsf@shell01.TheWorld.com>

Tim Peters <tim.one at comcast.net> writes:

> [Tim]
> > Note that it's not true that Python's list is really a vector/array --
> > thelanguage doesn't define its implementation, and doesn't want to.
> 
> [Marcus Alanen]
> > Hmm.  So the Python language is more about specifying _what_ will
> > happen, instead of _how_, with no concern for actual speed or memory?
> 
> The Python language reference manual, and the Python standard library
> reference manual, are like that, and that's par for the course for language
> specs:  they specify syntax ("how is it spelled?") and semantics ("what does
> it mean?").  Everything else falls under pragmatics, often called "quality
> of implementation" issues.  The STL is close to unique, outside the
> real-time world, in making promises related to performance.  For the other
> side, I spent the first 15 years of my career working largely on optimizing
> Fortran compilers and runtime libraries:  no user base was more concerned
> about speed than Fortran's, yet the Fortran standard didn't breathe a word
> about how fast or slow any construct or standard function might be, could
> be, or should be.  Similarly, the IEEE-754 standard for floating-point
> arithmetic doesn't say anthing about speed either.  Standards bodies are
> generally loathe to constrain implementations.
> 
> Of course the Python language itself is distinct from (albeit related to)
> its implementations.  People working on implementations are typically quite
> concerned about speed and memory.  The pragmatic tradeoffs can (and do) vary
> across implementations, though (same as for C, C++, Fortran, Java, etc).

People wanting to program should nail a copy of this posting above their
keyboard.



From rmunn at pobox.com  Fri Apr  4 10:21:12 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 04 Apr 2003 15:21:12 GMT
Subject: Cygwin Python regression testers wanted
References: <mailman.1049385789.4709.python-list@python.org>
Message-ID: <slrnb8qjnc.i81.rmunn@localhost.localdomain>

Jason Tishler <jason at tishler.net> wrote:
> I'm in the process of updating the Cygwin section of the official Python
> README.  Since I only have access to Windows NT 4.0 and 2000, I am
> seeking Cygwin Python regression testers for the following Windows
> platforms:
> 
>     95
>     95 OSR2
>     98
>     98 Second Edition
>     Me
>     XP Home
>     XP Professional
> 
> If you are willing to help, then perform the following steps:

I'd be willing to help with 95 OSR2, 98, and 98SE -- but it will have to
wait until this weekend, I'm afraid.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From aahz at pythoncraft.com  Sun Apr 20 21:23:21 2003
From: aahz at pythoncraft.com (Aahz)
Date: 20 Apr 2003 21:23:21 -0400
Subject: python dynamic scoping question
References: <79ad5955.0304201513.6ffd4f83@posting.google.com>
Message-ID: <b7vh69$pco$1@panix3.panix.com>

In article <79ad5955.0304201513.6ffd4f83 at posting.google.com>,
Julia Goolia <juliagoolia301 at hotmail.com> wrote:
>
>I have a question regarding scoping in python.  I wish to have some
>mechanism which puts variables at a function level scope into the
>global scope, dynamically.  I cannot alter my design.

Why not?  Is this a homework assignment?  Either way, please provide the
precise specification so we can find appropriate solutions.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From timh at zute.net  Thu Apr  3 06:29:02 2003
From: timh at zute.net (Tim Hoffman)
Date: Thu, 03 Apr 2003 19:29:02 +0800
Subject: distributed computing implementations
In-Reply-To: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
Message-ID: <3e8c19fe$1@news>

For simple cases you might like to look at ZODB

Tim

robin wrote:
> I am attempting to summarise the distributed computing implementations
> available to Python programmers. My conclusions so far are as follows:
> 
> If you want the biggest, boldest approach and don't care about
> overhead, use CORBA.
> 
> If you want a simpler approach but with performance and implementation
> difficulties, use SOAP or XML-RPC.
> 
> If you want the ultimate in simplicity and are willing to foresake
> multi-language support, use Dopy, Pyro, or Twisted Spread.
> 
> Comments? Specifically, what are the relative advantages and
> disadvantages of the last three mentioned products?
> 
> -- robin




From nirina at mail.blueline.mg  Sun Apr  6 15:25:42 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Sun, 6 Apr 2003 23:25:42 +0400
Subject: I'm new to programming
References: <oprm4ogjb9acxezr@news.datafast.net.au>
Message-ID: <007301c2fc8c$92e69080$f9a2383e@raseliar>

----- Original Message ----- 
From: "Casey Jones" <cajones at datafast.net.au>
Newsgroups: comp.lang.python
To: <python-list at python.org>
Sent: Saturday, April 05, 2003 3:20 AM
Subject: I'm new to programming

> Inspiration would be appreciated. 

follow the thread 'need help with game'
i've posted some food for your imagination.

nirinA





From syver-en+usenet at online.no  Sun Apr  6 14:43:01 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 06 Apr 2003 20:43:01 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain>
Message-ID: <uel4fh3gq.fsf@online.no>

pobrien at orbtech.com (Patrick K. O'Brien) writes:

> No, I really want something very similar to Emacs, not just
> "comparable in power."  The reason is that there is a ton of elisp
> code out there, including most of Emacs itself, that could be
> translated to work with PyAlaMode.  My goal isn't to be able to
> convert in a completely automated fashion, but close enough to be
> attractive to both an elisp programmer and a Python programmer.
> 
> That means all the Emacs primitives need to exist and be spelled the
> same (other than using underscores instead of dashes in the names),
> and the ability to register functions and bind keystrokes needs to be
> there, including support for define-key, global-set-key,
> local-set-key, and commands bound to multiple keystrokes using
> ctl-x-map and esc-map.

Maybe you could embed an elisp interpreter in PyAlaMode, that way we
could use Emacs modules, right away. 

> And if we can remove
> some Emacs warts, that would be even better.

First on my list would be multithreading and better
extension/embedding capabilities, but that would probably be rather
automatic as Python provides this out of the box.




-- 

Vennlig hilsen 

Syver Enstad



From staschuk at telusplanet.net  Wed Apr 16 14:13:44 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 12:13:44 -0600
Subject: A better unittest
In-Reply-To: <3ckier9a.fsf@python.net>; from theller@python.net on Wed, Apr 16, 2003 at 11:16:49AM +0200
References: <3ckier9a.fsf@python.net>
Message-ID: <20030416121344.A4266@tibia.amotlpaa.bogus>

Quoth Thomas Heller:
  [...changing output of unittest...]
> The big advantage is that if I execute the tests from within XEmacs with
> C-c C-c, the output ocurrs in a *Python Output* buffer, and the cursor
> in the source buffer jumps to the real cause of the error: the
> self.failxxx() call.
> 
> What do people think? Is this a useful change, and should I submit a
> patch?

+0, since I don't use XEmacs and the change would have negligible
effect for me.

(Making XEmacs let you navigate the traceback would imho be a
better solution.)

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From brian at sweetapp.com  Mon Apr 28 20:56:05 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Mon, 28 Apr 2003 17:56:05 -0700
Subject: In which versions is list.sort stable?
In-Reply-To: <b8kg8g$1d10$1@ns.felk.cvut.cz>
Message-ID: <007401c30dea$1dbc1680$21795418@dell1700>


Cheers,
Brian 

> -----Original Message-----
> From: python-list-admin at python.org
[mailto:python-list-admin at python.org]
> On Behalf Of Frantisek Fuka
> Sent: Monday, April 28, 2003 17:15
> To: python-list at python.org
> Subject: Re: In which versions is list.sort stable?
> 
> Can you please explain to a newbie what this thread means?

A stable sort is a sorting algorithm that does not change the relative
order of elements that have equal values. It has nothing to do with
crashing. Here is an example:

Python 2.2.1 Build...
>>> stuff = [("grape", 5.50), ("banana", 5.50), ("cherry", 3)]
>>> stuff.sort(lambda x,y: cmp(x[1], y[1]))
>>> print stuff
[('cherry', 3), ('banana', 5.5), ('grape', 5.5)]

As you can see, the banana and grape tuples have been swapped even
though swapping them was not necessary to satisfy the sorting condition.
So the Python 2.2 sorting algorithm is not stable. The Python 2.3
sorting algorithm is i.e. if would return:
[('cherry', 3), ('grape', 5.5), ('banana', 5.5)]

Cheers,
Brian





From david.bear at asu.edu  Thu Apr  3 13:41:26 2003
From: david.bear at asu.edu (David Bear)
Date: 03 Apr 2003 11:41:26 -0700
Subject: computer science with python
Message-ID: <m3istvtodl.fsf@moroni.pp.asu.edu>

I purchased the OReilly book "Computer Science with Perl" (or
something similar and found it very fun -- even though I really hate
Perl.  It made me think that python would be a terrific language to
demonstrate some key computer science concepts.

Anyone know if a book has already been done?  I find that many CS
language concepts are debated/discussed here and have to be careful
when I look at this group because I become too interested in reading
all the conceptual stuff rather than getting a specific question
answered.  Seems this list would provide a good starting point for
such a book.  I image it could be many volumes...



From juneaftn at REMOVETHIShanmail.net  Mon Apr 28 15:47:30 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Tue, 29 Apr 2003 04:47:30 +0900
Subject: Slice confusion : a[n:p] is a list exclude the last  element p (thus spoke Dijkstra)
References: <mailman.1051532982.25352.python-list@python.org>
Message-ID: <b8k0fl$dsa$1@news.hananet.net>

A bit related to this is,

"Why Numbering Should Start At Zero" by Edsger Dijkstra at
http://www.cs.utexas.edu/users/EWD/ewd08xx/EWD831.PDF

He explains why a) is preferred in the following list:

 a. 2<= i < 13
 b. 1< i <=12
 c. 2<= i <= 12
 d. 1< i < 13


And then continues on to explain why b) is preferred in the following list:

 a. 1<= s < N+1
 b. 0<= s < N

With warmest regards,

June




From drew at astro.pas.rochester.edu  Fri Apr  4 14:56:03 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 4 Apr 2003 11:56:03 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122> <3E8CF4B5.1E221DD7@alcyone.com> <9383481e.0304040445.a3752dd@posting.google.com> <7h3d6k2jtoi.fsf@pc150.maths.bris.ac.uk>
Message-ID: <9383481e.0304041156.2a0041a2@posting.google.com>

Michael Hudson <mwh at python.net> wrote in message news:<7h3d6k2jtoi.fsf at pc150.maths.bris.ac.uk>...
> drew at astro.pas.rochester.edu (Drew Moore) writes:
> 
> > I'd like this PEP to go one of two ways..
> > either implementation,
> > or a good solid technical reason why its a bad idea.
> 
> How about: it involves changing the language into something it's not.
> 
> For a more technical sounding reason, how do you cope with the fact
> that `=' can both establish and change a binding?
> 
> Cheers,
> M.

I think I came up with a reason.

when the overloading object is always on the left side of the assign,
I think things stay pretty benign. 

When they appear on the right side, however, you get trouble..
The overloading is sticky.. contagious.. eek!

for s in L : # loop thru list of things..
    p = s    # hey, what if an item in the list overloads assignment?
            # it overloads it for the remainder of the loop!

Sounds like an easy way to shoot yourself in the foot.
I can't think of a sensible way around this.

Drew



From nhodgson at bigpond.net.au  Mon Apr  7 05:47:55 2003
From: nhodgson at bigpond.net.au (Neil Hodgson)
Date: Mon, 07 Apr 2003 09:47:55 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <7xn0j3fsdu.fsf@ruckus.brouhaha.com> <m21y0fa5kq.fsf@localhost.localdomain>
Message-ID: <fPbka.1053$1h6.6248@news-server.bigpond.net.au>

Patrick K. O'Brien:

> The Scintilla text control (written in C and wrapped in wxPython by a
> control called wxStyledTextCtrl) handles the on-screen representation
> and user-manipulation of the text.  So that part is actually easy.
> Just read in a file, feed it to Scintilla, and go.  The next step is
> wrapping all this in an Emacs-like framework.

   What Scintilla doesn't give you though, is simple direct access to the
document (buffer) objects and their text - instead you have to go through a
view object. This was a design decision based on the styles of API that
could be easily exposed from the target widget technologies. A better
approach would be to provide an "object model" API that provided access to
the sub-objects, particularly the document objects. An object model API can
be implemented as a facade over the existing flat API although there will be
a cost in complexity and possibly speed. Some of the techniques used here
include invisible view objects playing the part of document objects or the
use of accessor objects (like the WindowAccessor and DocumentAccessor clases
inside Scintilla) to provide higher bandwidth interaction with document
text.

   The kinds of operation where you want fast access to the buffer contents
include regular expression search and replace where it would be best to use
Python's builtin RE library. Don't use Scintilla's very primitive RE code
except as a temporary measure to get the editor running.

   Neil





From paul at prescod.net  Mon Apr  7 14:39:14 2003
From: paul at prescod.net (Paul Prescod)
Date: Mon, 07 Apr 2003 11:39:14 -0700
Subject: Problem setting COM property using win32com
In-Reply-To: <E1B4F80EB080194BBAD3092392F511B3DADEEA@XCH-NW-13.nw.nos.boeing.com>
References: <E1B4F80EB080194BBAD3092392F511B3DADEEA@XCH-NW-13.nw.nos.boeing.com>
Message-ID: <3E91C5D2.9070003@prescod.net>

Green, Gregory P wrote:
> I am creating the object like so:
> 
>    attrs = win32com.client.Dispatch("SmRecList.SmRecordList")
> 
> I hadn't noticed the dynamic. I guess I can try to force early binding.

Do some introspection on the objects to see whether they are makepy 
objects or dynamic objects. Probably if you print out the __class__ or 
dir() of the objects you'll get a strong hint.

  Paul Prescod





From tjreedy at udel.edu  Fri Apr  4 10:23:26 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Fri, 4 Apr 2003 10:23:26 -0500
Subject: writing to windows memory
References: <3E8C01CE.6070601@netscape.net> <h49q8v4aoduebq0n0o9h08ask5o3e8d2fl@4ax.com>
Message-ID: <hgudnYO57805PBCjXTWc3g@comcast.com>

> >newbie:
> >pythonwin 2.2.2
> >Id like to bind a button.widget that simply writes a one to windows
> >address 440h. any ideas on where to start.(the button stuf is
trivial) i
> >need to know how to access the address space.

I believe that you mean 440h above to be an Intel i/o port number,
analogous to a socket port number, rather than a (windows) memory
address.  If so, you need a machine-specific C- or assembler-code
module, analogous to the socket module, to access it.

Terry J. Reedy





From claird at lairds.com  Mon Apr  7 18:25:58 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 07 Apr 2003 22:25:58 -0000
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049722991.18682.python-list@python.org> <v938h26mdl9s8f@corp.supernews.com> <h469m-4et.ln1@news.lairds.org>
Message-ID: <v93unmeqt8q515@corp.supernews.com>

In article <h469m-4et.ln1 at news.lairds.org>,
Kyler Laird  <Kyler at news.Lairds.org> wrote:
			.
			.
			.
>software for which I can not get the source code.  I
>certainly wouldn't ever want to *depend* on software
>that I can't control.
>
>What kind of business *pays* to become dependent on
>software they don't control?!
>
>--kyler

Most.

And probably a majority of businesses perceive that they
have more control over purchased proprietary products than
they do over what their own programmers generate.

Is that, to phrase it delicately, a misperception?  Given
standard management practices, perhaps not.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From spam04 at kuwest.de  Mon Apr  7 04:33:52 2003
From: spam04 at kuwest.de (Jens Kubieziel)
Date: Mon, 7 Apr 2003 08:33:52 +0000 (UTC)
Subject: PYTHON EDITOR!!
References: <mailman.1049690233.24009.python-list@python.org>
Message-ID: <b6rd5g$67s$1@QBI050102.kubieziel.de>

Pathmenanthan RamaKrishna <nanthan at siswa.utm.my> wrote:
> have been working in windows, but my development actually is on 
> Linux, so can u suggest me the best editor(i mean use friendly 
> and graphical,i dont want command line editor)for python?

Use gVim.

-- 
Jens Kubieziel                                   http://www.kubieziel.de
Freude ist ein Zeichen, da? man dem Licht nahe ist.
Stein, Edith



From Vector180W at netscape.net  Thu Apr  3 19:36:58 2003
From: Vector180W at netscape.net (vector)
Date: Fri, 04 Apr 2003 10:36:58 +1000
Subject: pyparallel errors
Message-ID: <3E8CD3AA.7070704@netscape.net>

Thanks chris but I get this error any ideas?
im running xp
i wonder if I have to tell xp to release the printer

PythonWin 2.2.2 (#37, Nov 26 2002, 10:24:37) [MSC 32 bit (Intel)] on win32.
Portions Copyright 1994-2001 Mark Hammond (mhammond at skippinet.com.au) - 
see 'Help/About PythonWin' for further copyright information.


 >>> import parallel
Traceback (most recent call last):
   File "<interactive input>", line 1, in ?
   File "G:\Development\Python\Lib\site-packages\parallel\__init__.py", 
line 13, in ?
     from parallelwin32 import *
   File 
"G:\Development\Python\Lib\site-packages\parallel\parallelwin32.py", 
line 53, in ?
     import _pyparallel
ImportError: Couldn't access giveio device




From member27604 at dbforums.com  Thu Apr 17 21:42:32 2003
From: member27604 at dbforums.com (dolephi9080)
Date: Fri, 18 Apr 2003 01:42:32 +0000
Subject: [Que]Do any SGML lib good for Python?
References: <2778697.1050595278@dbforums.com> <m3of35ni8d.fsf@mira.informatik.hu-berlin.de>
Message-ID: <2781052.1050630152@dbforums.com>

Originally posted by Martin V. L?Wis 
> dolephi9080  writes:
>
> > I search  the SGML Lib for python, but it's hard to find. even
>     the
> > python contain a sgml parser itself. But that's a limite
>     parser.
> > Does some one know it? pls tell me. Thx
>
> I find that parsing the output of nsgmls is quite powerful and
> effective.
>
> Regards,
> Martin 

What is your mean parsing the output of nsgmls?

--
Posted via http://dbforums.com



From snowball3 at bigfoot.com  Mon Apr 21 17:48:38 2003
From: snowball3 at bigfoot.com (Mark E.)
Date: Mon, 21 Apr 2003 17:48:38 -0400
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com> <a3b5b608.0304080302.c26ad68@posting.google.com> <af6b5bdd.0304181001.558da6f@posting.google.com> <84of317cd6.fsf@boost-consulting.com> <b7vlfb$27c$0@216.39.172.122> <84of30m4cp.fsf@boost-consulting.com>
Message-ID: <p0p8avsd7nlqu52p4rnjobuemctd3foil8@4ax.com>

David Abrahams <dave at boost-consulting.com> wrote:

>    xml.sometag(attr1="value1", attr2="value2", ...)[ "contents" ]

Of course there are always corner cases to deal with. For example, what's a good way
to handle XHTML's 'class' attribute?

xml.sometag(class="css_class_name")["contents"]
won't work.

One could map 'foo_' to 'foo' to handle conflicting names used in xml/html, or allow
a dictionary to serve the same purpose:

xml.sometag({"attr1" : "value1", "class" : "value2"})["contents"]

Mark

-- 
Mark E.: snowball3#&0064;softhome.net



From ramen at lackingtalent.com  Thu Apr 24 04:05:19 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Thu, 24 Apr 2003 08:05:19 -0000
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com> <gcjna.35728$T34.931514@news2.tin.it>
Message-ID: <slrnbaf6n1.gb9.ramen@lackingtalent.com>

In article <gcjna.35728$T34.931514 at news2.tin.it>, Alex Martelli wrote:
> Aahz wrote:
>> Note that Guido deeply regrets permitting one-line blocks; this is
>> something he may try fixing for Python 3.0.
> 
> Forcing linebreak after each & every colon?  Oh well -- he's probably
> getting bribes from large-screen and vertical-aspect-ratio monitor
> manufacturers, is my guess...;-)

A certain colleague of mine insists on putting a newline after every
function call argument, with the final parenthesis on its own line. His
aspect ration looks about normal; I think it has more to do with the
lines-of-code metric. ;)

Peace,
Dave



From usenet at rob.dqd.com  Mon Apr 14 16:21:00 2003
From: usenet at rob.dqd.com (Rob Mayoff)
Date: 14 Apr 2003 13:21:00 -0700
Subject: isearch, LEAP and patents  was: Re: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org> <yfsr88fkppr.fsf@black132.ex.ac.uk> <eaeka.1333$1h6.8933@news-server.bigpond.net.au> <yfs1y0el9xc.fsf_-_@black132.ex.ac.uk> <e0caaad7.0304100951.57cb4572@posting.google.com> <yfsadeudlud.fsf@black132.ex.ac.uk>
Message-ID: <e0caaad7.0304141221.1f045e29@posting.google.com>

Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfsadeudlud.fsf at black132.ex.ac.uk>...
> Have you got a publically available reference at hand by any chance? I'd be
> really curious to learn how the patent claim is justified.

<http://tinyurl.com/9imo>



From kp at kyborg.dk  Wed Apr  9 12:12:17 2003
From: kp at kyborg.dk (Kim Petersen)
Date: Wed, 09 Apr 2003 18:12:17 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E9267F6.9020605@kyborg.dk> <yfsu1d9hqw6.fsf@black132.ex.ac.uk> <3E9444E8.4070805@kyborg.dk>
Message-ID: <3E944661.9080503@kyborg.dk>

Kim Petersen wrote:

> implementation of let:
> 
> def let(cons,namespace):
>    # of course a scheme to reasonably map lisp functions to python
>    # implementations is needed - i just assume that (f x y z) is called
>    # here as f(lisplist(x,y,z),namespace)
>    #
>    context=namespace.getcontext()
>    defs=cons.car()
>    body=cons.cdr()
>    if iscons(defs.car()):
>       namespace.localvar(defs.car().car(),defs.car().cdr().car())
>    else:
>       namespace.localvar(defs.car())
>    lispeval(body,namespace)

forgot to do:
      namespace.restorecontext(context)

-- 
Med Venlig Hilsen / Regards

Kim Petersen - Kyborg A/S
IT - Innovationshuset
Havneparken 2
7100 Vejle
Tlf. +4576408183 || Fax. +4576408188




From andersjm at dancontrol.dk  Thu Apr  3 02:42:44 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Thu, 3 Apr 2003 09:42:44 +0200
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org>
Message-ID: <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk>

"Mike Meyer" <mwm at mired.org> wrote:
> I disagree with that statement. Clearly, *someone* knows all the
> coding tricks used by the compiler - because they put it into the
> compiler. So they can achieve the same level of performance. 

And chess program authors can beat their own inventions?  Not so.

- Anders






From andy at wild-flower.co.uk  Fri Apr 18 14:58:06 2003
From: andy at wild-flower.co.uk (Andy Jewell)
Date: Fri, 18 Apr 2003 19:58:06 +0100
Subject: Classes in Python
In-Reply-To: <s1Vna.126364$It5.107742@news2.central.cox.net>
References: <3EA01A9B.4030209@vnet.net> <s1Vna.126364$It5.107742@news2.central.cox.net>
Message-ID: <200304181958.06870.andy@wild-flower.co.uk>

On Friday 18 Apr 2003 4:42 pm, Steve Holden wrote:
> "Bill Martin" <wcmartin at vnet.net> wrote ...
>
> > I'm wondering about the value of allowing a class definition like this:
> >
> >              class C:
> >                pass
> >
> > Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or
> > some such. If I try to print a.x2 or b.x1, I get an exception message
> > basically saying those member variables don't exist. It seems to me this
> > defeats one of the basic ideas behind OOP, that is to assign members to
> > the class definition, thus ensuring that all instantiations of a class
> > have the same member variables and methods. I know there must be a
> > reason that python allows this kind of thing that I described, but I
> > can't figure out what that reason could be. To me, it looks like a very
> > dangerous thing to do. Can someone help me understand why we'd like to
> > do it? Thanks,
>
> Well, if you think it's dangerous you probably *don't* want to do it, but
> it's the easiest way to set up a data structure with named components - you
> are just taking advantage of the fact that each instance has its own
> namespace. It's a firly standard idiom, though it isn't required.
>
> Now, other object-oriented languages can go to great pains to make sure
> that an instance's attributes can only be accessed in very controlled ways,
> such as using accessor methods (getthis(), setthat(), etc.). Python, on the
> other hand, pays you the compliment of assuming you know what you are
> doing, and "gives you enough rope to shoot yourself in the foot"!
>
> If you choose to abuse that freedom, you end up with a limp. [Please,
> nobody ask "a limp what?"].
>
> regards

In all my days playing with rope, I've never managed to shoot myself with it.  
Hang myself, maybe, but not shoot myself.  :-)

I think, however, that Bill was trying to say that it is dangerous for Python 
not to behave in the manner he expected...

Shuffling off to the gallows pole...

-andyj




From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Mon Apr 28 12:49:36 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Mon, 28 Apr 2003 18:49:36 +0200
Subject: RuntimeError at end of test.autotest in Python 2.3b on windows,
 tmpfile also fails
In-Reply-To: <7h3llxukgx5.fsf@pc150.maths.bris.ac.uk>
References: <mailman.1051367863.3218.python-list@python.org> <3eaaa408$0$49109$e4fe514c@news.xs4all.nl> <3eac5942$0$49098$e4fe514c@news.xs4all.nl> <7h3llxukgx5.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3ead5ba1$0$49109$e4fe514c@news.xs4all.nl>

Michael Hudson wrote:
> Irmen de Jong <irmen at -NOSPAM-REMOVETHIS-xs4all.nl> writes:
> 
> 
>>PS if developers don't run "import test.autotest" what *do* they run?
> 
> 
> "make test" (or at least, that's what I do).
> 
> Cheers,
> M.
> 

Ofcourse. This runs "./python -E -tt ./Lib/test/regrtest.py -l"
so I typed in my windows command prompt:
c:\python23> python -E -tt Lib\test\regrtest.py -l

and this works nicely, no RuntimeError at all.
It was only a problem when doing "import test.autotest"
(and has been fixed already).

--Irmen





From haris.bogdanovic at zg.hinet.hr  Sun Apr 13 16:36:32 2003
From: haris.bogdanovic at zg.hinet.hr (Haris Bogdanovic)
Date: Sun, 13 Apr 2003 22:36:32 +0200
Subject: Pyrex - undefined struct/union
Message-ID: <b7b7ht$jga$2@as201.hinet.hr>

I have a problem with this piece of code :

cdef extern from "portaudio.h":
    struct PaDeviceInfo:
        int structVersion
        char *name
        int maxInputChannels
        int maxOutputChannels
        int numSampleRates
        double *sampleRates
        PaSampleFormat nativeSampleFormats

def GetDeviceInfo(device):
    cdef PaDeviceInfo *info
    info=Pa_GetDeviceInfo(device)

    for i from 0 <= i < info.numSampleRates:
        sampleRates[i]=info.sampleRates[i]

    return dict([('structVersion',info.structVersion),
                       ('name',info.name),
                       ('maxInputChannels',info.maxInputChannels),
                       ('maxOutputChannels',info.maxOutputChannels),
                       ('numSampleRates',info.numSampleRates),
                       ('sampleRates',sampleRates),
                       ('nativeSampleFormats',info.nativeSampleFormats),
                       ('structVersion',info.structVersion)])

This part compiles successfully with pyrexc but when compiling C file
(generated by pyrexc) I get errors that left from all this "info" struct
members is an undefined struct/union.
What does that mean and how can I fix it ?

Thanks







From jon+usenet at unequivocal.co.uk  Mon Apr 28 20:23:07 2003
From: jon+usenet at unequivocal.co.uk (Jon Ribbens)
Date: 29 Apr 2003 00:23:07 GMT
Subject: Web programming and a different "type" problem (was: Slice confusion : a[n:p] is a list exclude the last  element p)
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com>
Message-ID: <slrnbarhfb.67t.jon+usenet@snowy.squish.net>

In article <vaqvrb3ipqblf2 at corp.supernews.com>, Cameron Laird wrote:
> 'Doesn't work, though.  Here's where the special case
> enters:  if only one datum with the indicated name is
> present as input, Zope and the others return it as a
> simple string, rather than a list with that string as 
> its sole member.

jonpy (jonpy.sf.net) avoids this problem. You name your multiple field
with a '*' suffix (e.g. <input type="checkbox" name="foo*">). All
fields with a '*' suffix are *always* lists when examined in your
Python program. All fields without a '*' suffix are *always* strings
(if there is more than one field with the same name but no '*' suffix
then all but one are discarded). e.g.:

  for foo in req.params.get("foo*", ()):
    process(foo)

It seems to me that allowing the Evil User to manipulate the type of
Python variables, as everything except jonpy seems to allow, is
fundamentally a bad idea.



From light at soton.ac.uk  Tue Apr  1 07:58:46 2003
From: light at soton.ac.uk (Mark Light)
Date: Tue, 1 Apr 2003 13:58:46 +0100
Subject: spawning a separate process
Message-ID: <b6c2c6$lr7$1@aspen.sucs.soton.ac.uk>

Hi,
       I have a python program from which I wish to spawn an os.system call
(a bash script) that may take several minutes to run - whilst the rest of
the script carries on. I have tried using bits of code from books - but it
never seem to work quite right

Any help would be greatfully recieved

Many thanks


Mark







From alloydflanagan at attbi.com  Tue Apr  8 09:39:09 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 8 Apr 2003 06:39:09 -0700
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com>
Message-ID: <a8b7f07a.0304080539.465049c8@posting.google.com>

msw <markscottwright at hotmail.com> wrote in message news:<3E921A4C.1020908 at hotmail.com>...
> Dave Brueck wrote:
> > On Mon, 7 Apr 2003, msw wrote:
...
> >>
> >>I'm really interested in this as well, and I think that it's an
> >>important issue for python.  Unfortunately, as I'm sure you've
> >>discovered by now, it's a question that comp.lang.python is
> >>constitionally incapable of answering.  The only answers you will get in
> >>this group are:
> >>
> >>1. intellectual property law is sufficient
> >>2. no protection scheme is perfect
> > 
> > 

I'm going to make a suggestion you may not hear from anybody else in
this group.  If hiding the code is important to you, don't use Python.
 Use C++ or something similiar.

Python was deliberately NOT designed to make it easy to obfuscate or
hide the code.  It's designed for power, readability, and fairly open
access to the internals of the class implementation.  I'd suggest that
the reasons you don't find more on code hiding are two-fold:  1) it's
somewhat technically difficult, and 2) most python users have no need
for it.

So quite aside from points 1 and 2 above, my suggestion is that you're
trying to insert a screw with a socket wrench -- you've just plain got
the wrong tool for the job.  Of course, if the program is written
already, it's a little late to be discovering this.

What about coding the sensitive algorithms in C and calling them from
python?  There are of course decompilers for C as well, but it's at
least harder.



From lx1978 at 263.net  Sun Apr  6 17:30:01 2003
From: lx1978 at 263.net (Xin Liu)
Date: Mon, 7 Apr 2003 05:30:01 +0800
Subject: Thank you very much.
References: <b6q0n6$1c55$1@mail.cn99.com> <mailman.1049659510.9004.python-list@python.org>
Message-ID: <b6q679$1f3i$1@mail.cn99.com>

>
> Do you mean, to swap two variables?  You do it like:
>
> a, b = b, a
>
>
>
>





From gh at ghaering.de  Wed Apr 30 12:48:50 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 30 Apr 2003 18:48:50 +0200
Subject: Calling dotNet components from Python.
In-Reply-To: <924a9f9c.0304300546.2cb5617d@posting.google.com>
References: <924a9f9c.0304282138.15aa7635@posting.google.com>
	<LqDra.7014$lD4.74951@news-server.bigpond.net.au>
	<924a9f9c.0304300546.2cb5617d@posting.google.com>
Message-ID: <3EAFFE72.4000106@ghaering.de>

Paradox wrote:
> Thank you all for your replies. I was hoping for something better than
> just putting a COM wrapper on my components. I would like to be able
> to script my notNet components without having to take an extra step
> every time I change something. I will give it a try though and see how
> tedious it is. [...]

I'd be surprised if Visual Studio .NET couldn't be persuaded to 
automatically register a .NET class as a COM component on each recompile.

-- Gerhard




From ckea25d02 at sneakemail.com  Wed Apr 16 22:06:02 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 17 Apr 2003 14:06:02 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <slrnb9qk41.7eb.rmunn@localhost.localdomain>
References: <b6pgcm$1t3$1@panix3.panix.com> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it> <slrnb9o9ob.493.rmunn@localhost.localdomain> <eX%ma.18570$Gb1.6139@nwrdny01.gnilink.net> <slrnb9qk41.7eb.rmunn@localhost.localdomain>
Message-ID: <b7l1h5$222st$1@ID-169208.news.dfncis.de>

Robin Munn wrote:
> I would call using Unix pipes
> "rebinding stdout", handled by the shell.

That's not quite the same thing, since the shell passes
the new stdin/out/err to the invoked process as parameters.
It doesn't change its *own* stdin/out/err, invoke the
process, and then change them back.

> Specifically,
> I've got this function that prints something to sys.stdout. And this
> function is not under my control.

If this is a library function, or anything else intended
to be called in a variety of circumstances, then I would
say it is badly designed. It should *not* be writing to
sys.stdout, or at least an alternative interface should
be provided which lets you specify where the output goes.

> How would you suggest capturing output *without* rebinding sys.stdout?

In that case, you have no choice. But it's not your
fault, it's the fault of whoever wrote the function.
The bad design is in the function, not the code
which is calling it.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From tjreedy at udel.edu  Sun Apr  6 12:27:47 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Sun, 6 Apr 2003 12:27:47 -0400
Subject: properties and get/set methods
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org>
Message-ID: <lA6dnc_HzLQazg2jXTWcqg@comcast.com>

"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
news:mailman.1049604615.22077.python-list at python.org...
> In these cases, you're using properties to simplify the modelling of
> domain objects, or to provide some *generic* functionality over and
> above a regular value get/set.  You don't want to have to specify
that
> functionality for every property, you just want to say "this is one
of
> _those_ properties".  The code bloat (not to mention mindless
> duplication of code) when you get up to 15 or 20 properties for an
> object would be ridiculous.

Perhaps what you want are __get/setattr__ methods that check if the
attribute is one of 'those-properties' (in a set/dict lookup) and if
so, does the generically different thing.  My impression is that
properties are meant for doing attribute specific customization
without loading up/slowing down __get/setattr__ with a long list of
if/elif blocks.

Terry





From dan at grassi.org  Tue Apr  1 17:18:57 2003
From: dan at grassi.org (Dan Grassi)
Date: Tue, 1 Apr 2003 22:18:57 +0000 (UTC)
Subject: How to obtain python file path
References: <20030401163212631-0500@news.covad.net> 
		<slrnb8jusc.5u1.ramen@lackingtalent.com> 
	<20030401165546102-0500@news.covad.net> 
	<b6d260$lg$1@news.eusc.inter.net>
Message-ID: <20030401171633237-0500@news.covad.net>

Not Quite,

I need b to know it's path, not the script that imports b.

Dan

In <b6d260$lg$1 at news.eusc.inter.net> Daniel Dittmar  wrote:
> Dan Grassi wrote:
>> Not quite, that only provides the path to the initial .py file.
>> 
>> With a directory structure:
>> fa
>>     a.py
>>     fb
>>         b.py
>> 
>> cd .../fa
>> execute a.py
>> where a adds fb to sys.path
>> a imports b.py
>> 
>> I need b.py to report it's path which would be .../fa/fb
> 
>>>>  import string  string.__file__
> 'D:\\Python22\\lib\\string.py'
> 
> Daniel
> 
> 



From mwh at python.net  Tue Apr 29 04:54:54 2003
From: mwh at python.net (Michael Hudson)
Date: Tue, 29 Apr 2003 08:54:54 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304251030.30fd1f94@posting.google.com>
Message-ID: <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk>

mis6 at pitt.edu (Michele Simionato) writes:

> Michael Hudson <mwh at python.net> wrote in message news:<7h37k9ilm5v.fsf at pc150.maths.bris.ac.uk>...
> > Thomas Heller <theller at python.net> writes:
> > 
> > > Wouldn't a caching class factory be able to do this?
> > 
> > Yes.  I only said "tricky", not impossible.
> > 
> > class MetaGit(type):
> >     def __eq__(self, other):
> >         # aha! this makes class objects unhashable, cackle!
> >         return self is other
> > 
> > I don't know whether you want to cope with things like that.
> > 
> > Cheers,
> > M.
> 
> Two questions:
> 
> 1) I don't see how this solve the problem of avoiding duplications of
> metaclasses;

Well, I imagine a caching class factory would store base classes in a
dict...

> 2) why you say this makes classes unhashable ?

Well, I thought it would.  I was wrong:

> It seems I can still use A and B as keys for a dictionary: 
> 
> class A: __metaclass__ = MetaGit
> class B: __metaclass__ = MetaGit
> 
> dic={A:1, B:2}
> 
> print dic[A]  #=>1 okay
> print dic[B]  #=>2 okay

Then you get this kind of weirdness:

/>> class C(type):
|..     def __eq__(self, other):
|..         return 1
\__ 
/>> class A:
|..     __metaclass__ = C
\__ 
/>> class B:
|..     __metaclass__ = C
\__ 
->> d = {}
->> d[A] = 1
->> d[B] 
Traceback (most recent call last):
  File "<input>", line 1, in ?
KeyError: <class '__main__.B'>
->> A == B
1

Then I suspect whether a instance of C gets A out of the dictionary
depends on accidents of memory allocation and hash table
implementation.  That's not a good thing.

CHeers,
M.

-- 
  ARTHUR:  Yes.  It was on display in the bottom of a locked filing
           cabinet stuck in a disused lavatory with a sign on the door
           saying "Beware of the Leopard".
                    -- The Hitch-Hikers Guide to the Galaxy, Episode 1



From bgporter at artlogic.com  Wed Apr  9 15:57:47 2003
From: bgporter at artlogic.com (Brett g Porter)
Date: Wed, 09 Apr 2003 15:57:47 -0400
Subject: string() - Is this some kind of sick joke?
In-Reply-To: <36dbf824.0304091142.527af834@posting.google.com>
References: <36dbf824.0304091142.527af834@posting.google.com>
Message-ID: <3iOdnWZk6dYT5wmjXTWcqw@comcast.com>

Daniel Gowans wrote:
> 
> What gives?  Did I get the bogus version?  Is this wishful thinking
> documentation?  I also tried  y.strip("'")  with the same results.
> 

Hm -- works here, running ActiveState 2.2.2 on Win2K:

 >>> y = "'hello'"
 >>> y.strip("'")
'hello'
 >>>




From antoni3 at otenet.gr  Mon Apr  7 12:44:05 2003
From: antoni3 at otenet.gr (Antonis Kaklis)
Date: Mon, 7 Apr 2003 19:44:05 +0300
Subject: Parsing data
References: <b6s5lt$5c1$1@usenet.otenet.gr> <b6s71b$9j3$1@news-reader12.wanadoo.fr>
Message-ID: <b6s9lu$a9i$1@usenet.otenet.gr>

> Was that what you wanted? What you want to do with your data is not quite
clear
> from your post...
> --
> - Eric Brunel <eric.brunel at pragmadev.com> -
> PragmaDev : Real Time Software Development Tools -
http://www.pragmadev.com
>
Thanks very much for your response...
but  now  I have lists like this:
[2.1614729999999999e-006, 2.183854e-006, 2.1778039999999998e-006,
2.1740989999999998e-006, 2.2548290000000002e-006]
[2.1451629999999998e-006, 2.322683e-006, 2.1120490000000001e-006,
2.0925849999999999e-006, 2.0696180000000001e-006]
[2.05246e-006, 2.5724020000000002e-006, 2.6257250000000002e-006,
2.7019989999999999e-006, 2.7715390000000001e-006]
[1.88422e-006, 1.8648680000000001e-006, 1.8359719999999999e-006,
3.1508029999999998e-006, 3.2457659999999998e-006]

What i want is to process each column. First the values of the first column,
continue with the second column and so on...,
in order to plot the values with time step of 0.02sec. (f(t)).
I don't know if I make myself clear now(I hope so).

Antonis






From aleax at aleax.it  Sun Apr 13 12:15:59 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 16:15:59 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <Wl9ma.6864$LB6.194823@news1.tin.it> <m3adeuww2q.fsf@mira.informatik.hu-berlin.de>
Message-ID: <33gma.10651$T34.347631@news2.tin.it>

Martin v. L?wis wrote:

> Alex Martelli <aleax at aleax.it> writes:
> 
>> The crucial difference is that Python dicts are hashtables, while the
>> std::map in the standard library of C++ can't be, because it's based
>> on an order relationship, NOT on hashing.  So, std::map has an order,
>> but its O(...) performance just can't be as good as a dict's.
> 
> Nit-pickers (like myself) will point out that the complexity of a
> hash-table depends on the distribution of the hash function (in
> addition to depending on the complexity of the hash function).  With a

"To a nitpicker, a nitpicker squared": nothing stops you (in C++) from
writing an operator< whose complexity is just as high as you can have
for a __hash__ method in Python.  In this sense, dependency on the
complexity of some auxiliary function is just the same in the two
languages (actually, the dependency is stronger in C++, since for each
insertion or memebership test operator< will typically be called O(log N)
times, while, in Python, for such an operation hash is called O(1) time).

> carefully-chosen hash O(1) function, all dictionary operations will be
> O(n). For std::map, O(log n) is the worst-case performance.

Yes, it's technically true that (once some fairy godmother ensures
that the auxiliary function is O(1)) worst-case behavior is potentially
worse for a hash table than it is for a sufficently smart order-based
container (and the C++ standard library does constrain its container
to be sufficiently smart in this sense).  I think Tim's post has already
been quite eloquent on the relative care Python places on
worst-case-guaranteed vs typical-case-expected behavior.


Alex




From LogiplexSoftware at earthlink.net  Tue Apr  8 10:58:53 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 08 Apr 2003 07:58:53 -0700
Subject: Got a kick out of this
In-Reply-To: <b6tic1$8os9s$1@ID-169208.news.dfncis.de>
References: <237247fb.0304071335.619a468b@posting.google.com>
	 <b6tic1$8os9s$1@ID-169208.news.dfncis.de>
Message-ID: <1049813933.3724.17.camel@software1.logiplex.internal>

On Mon, 2003-04-07 at 21:24, Greg Ewing (using news.cis.dfn.de) wrote:
> Terin wrote:
> > 
> > http://vil.nai.com/vil/content/v_994.htm
> 
> Well, some people call Windows a boot virus... so
> I suppose Python could be regarded as a programmer's
> brain virus...
> 
> Name:           Python
> Preferred Host: Perl, C, C++, PHP, sh, csh programmers
> Symptoms:       The host's programs suddenly become shorter,
>                  more readable and easier to maintain.
> Spread By:      Usenet/email contact, word of mouth
> Treatment:      No effective treatment known. Once infected,
>                  the host will never willingly use any other
>                  programming language.

Also causes sporadic <wink>ing. <wink>


But-no-other-symptoms-that-I-know-of'ly yrs,

-- 
Cliff-Wells, Software Engineer <wink>
Logiplex Corporation'ly yrs (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From claird at lairds.com  Tue Apr  8 13:14:31 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 08 Apr 2003 17:14:31 -0000
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org> <kauu6b.f2d1.ln@boundary.tundraware.com>
Message-ID: <v960rni1esvo94@corp.supernews.com>

In article <kauu6b.f2d1.ln at boundary.tundraware.com>,
Tim Daneliuk  <tundra at tundraware.com> wrote:
			.
			.
			.
>As a side note, this question was ultimately motivated by an entirely
>different sort of "pure math" question: Does (integer) 1 = 1.0?
>Engineers, physicists, and computer scientists all would
>say "no" because of the precision problem with the latter number.
>I am unclear, though, on whether a mathematician understands
>1.0 to mean 1.0000000... (a single number) or the _neighborhood_
>around 1 with one digit precision.
			.
			.
			.
I'm a (former) mathematician.  I've always taken "1.0"
to be a local, linguistic act, requiring me to infer
the speaker's context.  "1.0" means different things in
the mouths of different people.

The open interval (0.95, 1.05) is one of the plausible
interpretations.  It's not among the first half-dozen
likely to come to my mind, though; maybe in the top 
ten ...  
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From alloydflanagan at attbi.com  Thu Apr  3 17:32:41 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 3 Apr 2003 14:32:41 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk>
Message-ID: <a8b7f07a.0304031432.3c585657@posting.google.com>

Michael Hudson <mwh at python.net> wrote in message news:<7h3znn7k308.fsf at pc150.maths.bris.ac.uk>...
> drew at astro.pas.rochester.edu (Drew Moore) writes:
> 
> > Thanks for responding.. Maybe posting on April 1st was a mistake,
> > people might have thought I was joking around!!
> 
> I certainly did.
> 
> Cheers,
> M.

Oh, good, if this had turned out to be a joke, I'd be afraid I won't
get my Roman numeral literals after all.

MMCCXXXLIII forever!!!



From rajarshi at presidency.com  Mon Apr  7 10:44:41 2003
From: rajarshi at presidency.com (Rajarshi Guha)
Date: Mon, 07 Apr 2003 10:44:41 -0400
Subject: packaging Python files
Message-ID: <pan.2003.04.07.14.44.41.50125.1629@presidency.com>

Hi,
  sometimes when I write Python code I'll split up the code into several
files and import the modules when I need them. I know that when
deploying the program I can place everyhting into a directory under
PYTHONPATH and it should run. 

However, I was wondering whether there was any framework to create a
single Python file that would contain all the required files of the
project so that a user would simply run the file. Is this a good idea?
Has it been implemented?

Thanks,



From bdesth at removethis.free.fr  Fri Apr 18 19:44:13 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Fri, 18 Apr 2003 23:44:13 +0000
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <3EA08DCD.8090906@removethis.free.fr>

Bill Martin wrote:
> I'm wondering about the value of allowing a class definition like this:
> 
>             class C:
>               pass
> 
> Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or 
> some such. If I try to print a.x2 or b.x1, I get an exception message 
> basically saying those member variables don't exist. It seems to me this 
> defeats one of the basic ideas behind OOP, that is to assign members to 
> the class definition, thus ensuring that all instantiations of a class 
> have the same member variables and methods. I know there must be a 
> reason that python allows this kind of thing that I described, but I 
> can't figure out what that reason could be. To me, it looks like a very 
> dangerous thing to do. Can someone help me understand why we'd like to 
> do it? Thanks,

Why we'd like to do it? Well, it seems that some peoples would like to 
do it and need a whole lot of extra work for the same result. It's 
called the Decorator pattern.

Laotseu




From aleax at aleax.it  Fri Apr 25 19:29:08 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 23:29:08 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <8xjqa.5807$3M4.157577@news1.tin.it>

<posted & mailed>

Frantisek Fuka wrote:

> When I try to call methods, I sometimes forget to include the
> parentheses. Instead of:
> 
> if object.isGreen():
> do something...

I would suggest not using 'object' (and other names of built-in
types) as a name for your own variables, AND not using tabs for
formatting (some tools don't show those reliably -- use spaces,
which get shown reliably everywhere).


> i sometimes write:
> 
> if object.isGreen:
> do something...
> 
> If I understand it correctly, the if statement in this case tests if
> pointer to hasParent method is non-zero, which is of course always True

Well, no "of course" about it -- you might perfectly well have
executed at any previous time

    object.isGreen = 0

and then the if would fail.  Determining whether it's possible that
such an assignment has been done is a hellishly hard problem in the
general case.  (Here you mention "hasParent", but your examples show
"isGreen", so that's what I am using).

> so no error is reported and "do something..." always gets executed, so
> the application behaves in quite different way than I expected.
> 
> You can say to me "Don't forget to always include the parenteses" but

Oh, good, because that's very much the crucial issue;-).

> I'm still curious if this cannot be somehow configured, so that I get
> error when I try to just access the method pointer (".isGreen") instead
> of calling the method (".isGreen()"). I understand that there has to be
> the possibility of accessing the method pointers but I usually don't
> need it and for a beginner like me this makes the applications very hard
> to debug.

I guess one might develop a special version of Python which, when
asked to assess a bound-method object in boolean context, raises an
exception rather than accepting it as 'true'.  That, of course,
would still leave you with lots of problems (I'm sure boolean contexts
don't exhaust the list of places where you forget mandatory parentheses,
and functions and other callables have just the same issues as methods --
in general, you need parentheses to call any callable).
 
It's unlikely that such a special version of Python will be developed,
but you can make a case for it, and write a PEP (Python Enhancement
Proposal).  If Guido van Rossum gets convinced that this would indeed
be an enhancement to Python, then it will get implemented.

In Python as it stands, one COULD design a custom metaclass that,
instead of bound-methods, has its classes' instances return special
objects that raise an exception if assessed in a Boolean context and
otherwise delegate to the underlying bound-method.  For example:

import types

class SpecialWrapper(object):
    def __init__(self, func, inst, cls):
        self.func = func
        self.inst = inst
        self.cls = cls
    def __call__(self, *args, **kwds):
        self.func.__get__(self.inst, self.cls)(*args, **kwds)
    def __nonzero__(self):
        raise TypeError, 'Method used in boolean context'
        
class SpecialWrapDesc(object):
    def __init__(self, func):
        self.func = func
    def __get__(self, inst, cls):
        return SpecialWrapper(self.func, inst, cls)
    
class FF(type):
    def __new__(cls, name, bases, clasdict):
        for atnam, atval in clasdict.iteritems():
            if type(atval) is types.FunctionType:
                clasdict[atnam] = SpecialWrapDesc(atval)
        return type.__new__(cls, name, bases, clasdict)
__metaclass__ = FF

if __name__ == '__main__':
    class X:
        def ciao(self): print 'ciao!'
    x=X()

    # show normal method use works
    x.ciao()

    # show unbound-method use works
    X.ciao(x)

    # show use in boolean context raises an exception:
    if x.ciao:
       print 'oops...'
    else:
        print 'weird...?'


So, if you're willing to copy these three classes ito some
module (say FF.py), and start your own modules with:

import FF
__metaclass__ = FF.FF

then you could have the semantics you require.  Since I'm
typing this in at 1:30 AM with minimal testing, it's quite
likely that I have forgotten some doodad or other, but I
think the general concept is sound and any minor issues that
may arise might be dealt with.

Then, once you've finally learned that nothing gets called
until you provide parentheses, you can pension off these
crutches (or, you might stash them somewhere for the time
when you're ready to study custom metaclasses, descriptors,
and the like -- they might then be interesting examples).


Hope this helps...!

Alex




From andros at no_mail.com  Fri Apr 18 19:18:25 2003
From: andros at no_mail.com (Andros)
Date: Sat, 19 Apr 2003 01:18:25 +0200
Subject: hexadecimal calculations
Message-ID: <b7q0uv$hco$1@aphrodite.grec.isp.9tel.net>

Hello,
how to make these hexadecimal calculations in python?
A0 XOR 79 = D9
D9 - A8   = 31

Thanks ;)






From adalke at mindspring.com  Sun Apr 13 19:20:47 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Sun, 13 Apr 2003 17:20:47 -0600
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050261920.3205.python-list@python.org>
Message-ID: <b7cr51$qoi$1@slb2.atl.mindspring.net>

Grzegorz Adam Hankiewicz
> An idiom I prefer over that which works with every python version:
>
>         line = file.readline()
>         while line:
>                 do_stuff()
>                 line = file.readline()
>
> But it also makes for an infinite loop if you happen to forget the
> last line of the loop. Life is tough, you have to accept that or
> move to newer python versions, as suggested.

The fastest implementation to work across all versions of
Python since .. 1.3?  Or perhaps 1.4 ... is

while 1:
    lines = file.readlines(100000)
    if not lines:
        break
    for line in lines:
        do_stuff()

One line longer, faster, doesn't have the "if you happen to
forget the last line of the loop" problem (which is more likely
then forgetting the "if not lines: break" lines).

But seriously, upgrade.

                    Andrew
                    dalke at dalkescientific.com





From camel at oasis.com  Wed Apr  9 18:37:23 2003
From: camel at oasis.com (dromedary)
Date: Wed, 09 Apr 2003 18:37:23 -0400
Subject: Palindrome finder: help find ways to speed up?
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org> <BAB8B22D.2022%dogen@mindspring.com> <20030408194700.A1973@tibia.amotlpaa.bogus> <mailman.1049922136.32568.python-list@python.org>
Message-ID: <090420031837238086%camel@oasis.com>

Thanks to all for the excellent suggestions. Below is a rewritten
version of the palindrome finder. My results pretty much match
Steven's. For two-word combinations, the time went from about 150
seconds to five seconds. (That's on a G4/400 running OS 10.2.4.) That's
a huge improvement. 

And if I were less a dunderhead, I'd see how to generalize to
combinations above n=2. Times like this I wish I'd studied math or CS.
Perhaps in another life. In this one, I'll happily accept help from
Steven or others.

_________________________________________________

##### Find all unique combinations
def combinations(alist, numb, blist=[]):
   if not numb: 
      return rPal(''.join(blist))
   for i in range(len(alist)):
      blist.append(alist.pop(i))
      combinations(alist, numb-1, blist)
      alist.insert(i, blist.pop())

##### Look for symmetry
def rPal(w):
   c=len(w)/2
   if w[:c]==w[::-1][:c]:
      print w


words = open(SOME_WORDS_DOC).read().splitlines()
i=0
dict = {}
n=2

for word in words:
   i=0
   j=1
   while i < len(word):
      if dict.has_key( word[i:][::-1] ):
         if word in dict[ word[i:][::-1] ]:   #skip if in list
            i=i+1
         else:
            dict[ word[i:][::-1] ].append(word)
            i=i+1
      else:
         dict.update( { word[i:][::-1] : [word] } )
         i=i+1
         
      if dict.has_key( word[0:j][::-1] ):
         if word in dict[ word[0:j][::-1] ]: #skip if in list
            i=i+1
         else:
            dict[ word[0:j][::-1] ].append(word)
            j=j+1 

      else:
         dict.update( {  word[0:j][::-1] : [word] } )
         j=j+1 

      
for word in words:
   if dict.has_key(word):
      if word in dict[word]:
         dict[word].remove(word) #delete repeats
      dict[word].append(word)
      combinations(dict[word], n)



From rdsteph at earthlink.net  Sat Apr 12 20:52:24 2003
From: rdsteph at earthlink.net (Ron Stephens)
Date: Sun, 13 Apr 2003 00:52:24 GMT
Subject: Python in a Nutshell book review
Message-ID: <3E98B564.8030207@earthlink.net>

Python in a Nutshell, by Alex Martelli, 2003 O'Reilly, 636 pages.

Perhaps the best book about Python ever written, this book is the 
perfect capstone to anyone's library of Pythonic books, and also the 
perfect introduction to Python for anyone well versed in other 
programming languages. For newbies to programming, this would still be a 
good second book after a good introductory book on Python, such as 
Learning Python by Mark Lutz.

Written by my favorite author and Pythonista, Alex Martelli, this book 
manages to fill three roles in extremely pleasing fashion. First and 
foremost to me, it is a great read, straight through. Mr. Martelli's 
prose is always sparkling and always keeps the reader interested. No 
matter how many Python books you have read, you will learn new nuances 
from this book, and it is about the best review of the whole Pythonic 
subject matter that I can imagine. While there is absolutely no fluff 
whatsoever in these 636 pages, it still makes for rather easy reading 
because the explanations are so clearly thought out and explored as to 
lead one gently to understanding, without in any way being verbose. It 
is obvious that Alex Martelli took his time and put in sufficient 
thought, effort, and intellectual elbow-grease to make this work a 
classic for all time.

Secondly, this book is the ultimate Pythonic reference book, the best 
fit to this role I have yet seen. You will keep this book in the most 
cherished spot on your book shelf, or else right at your side on your 
computer desk, because you can almost instantly find any topic on which 
you need to brush up, in the midst of a programminng project.

Third, Python in a Nutshell is the most up-to-date book on Python (as of 
April 2003) and includes the best and most complete expositions yet on 
the new features introduced in Python 2.2 and 2.3. These topics are not 
only covered in depth, they are integrated into the text in their proper 
positions and relationships to the language as a whole. They are 
explained better here than I have seen anywhere else, so much so as to 
make them not only understandable to me (a duffer), but indeed so that 
they appear seamlessly Pythonic, as if they had been a part of the 
language since version 1.0. Topics explored in depth include new style 
classes, static methods, namespaces, iterators, generators, and new 
style division. List comprehensions are made not only comprehesible but 
indeed intuitive.

The book is surprisingly complete. It covers the core language as well 
as the most popular libraries and extension modules. It is difficult to 
choose any one portion of the book to highlight for extra praise, as all 
topics are treated so well. It is a complete book, the new definitive 
book about Python.

Everything about this book speaks of quality. In addition to the top 
notch writing and editing, O'Reilly really did the right thing and 
published this book printed on the highest quality paper, paper so thin 
that the 636 pages are encompassed in a book much thinner than one would 
expect for such a size, but strong enough to resist wear and tear. The 
text is most pleasing to the eye. Holding the book, and turning its 
pages, gives one  a feeling of satisfaction.

Any job worth doing is worth doing well. Alex Martelli and O'Reilly have 
done justice to a topic dear to our hearts, the Python programming 
language. Perhaps, in years to come, the passing time may make this book 
be no longer the most up-to-date reference on the newest features added 
to Python. But time can not erase the quality craftsmanship and the 
shear joy of reading such a well thought out masterpiece of Pythonic 
literature.

Ron Stephens

http://www.awaretek.com/plf.html

27+ reviews of books about Python

67+ links to online tutorials about Python and related subjects

Daily newsfeed of Pythonic web articles, new sourceforge projects, etc.




From adalke at mindspring.com  Tue Apr 29 14:41:07 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Tue, 29 Apr 2003 12:41:07 -0600
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com>
Message-ID: <b8mgo9$nbu$1@slb2.atl.mindspring.net>

Aahz
> I'm a little curious why you [that's me!] didn't write this as
>
>     i += 1
>
> Any reason (i.e. still need to support 1.5.2) or just habit?  If the
> latter, why haven't you changed your habit?

Learning BASIC in my formative years has rotted my brain?

I have a plodding brain.  I like using the same constructs over
and over.  I see "i = i + 1" much more easily than "i += 1".
(Though I see "i++" quite nicely.)

It's easier for others to see the intent as well, even without
experience in a language which has "+=".

The downside of making a typo just does not exist for "i".
(BTW, for long names I often copy&paste, making typos
less likely.)  You'll need to compare those odds with the
off-chance of making a typo for "i ++ 1" or "i == 1" (slipping
on the shift key).

The only time I've used += in Python is when I really was
mutating an object in-place.  Eg, I was throwing a "parse
error at character 12" where the parse only had part of a
string so didn't know the absolute location.  Each higher
block catches the exception, += offset to it, and reraises it.

What's wrong with habits?  It's a habit that I brush my
teeth.  It's a habit that I look both ways before crossing
the street.  It's a habit that I post to c.l.py.  :)

                    Andrew
                    dalke at dalkescientific.com





From aahz at pythoncraft.com  Wed Apr 30 12:17:09 2003
From: aahz at pythoncraft.com (Aahz)
Date: 30 Apr 2003 12:17:09 -0400
Subject: How to call built in function 'open' from inside object that has 'open' method?
References: <db81c492.0304300526.7b8d3c86@posting.google.com> <mailman.1051711257.436.python-list@python.org> <b8onf8$b7d$1@panix2.panix.com>
Message-ID: <b8osu5$rdv$1@panix1.panix.com>

In article <b8onf8$b7d$1 at panix2.panix.com>, Aahz <aahz at pythoncraft.com> wrote:
>
>  [...]

Uh, never mind.  Blame pre-breakfast blood sugar.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From ccosse at asymptopia.com  Mon Apr 14 02:30:11 2003
From: ccosse at asymptopia.com (Charlie Cosse)
Date: 13 Apr 2003 23:30:11 -0700
Subject: Tux Math Scrabble v2.0 released under GPL
Message-ID: <58329000.0304132230.428803b@posting.google.com>

Tux Math Scrabble is a math version of the popular board game for ages
4-40. Highly entertaining as well as great educational value.
Challenges young people to construct compound equations and consider
multiple abstract possibilities. Three skill-levels for practice from
basic addition and subtraction through multiplication and division.
Great for use in schools and at home. Runs on Windows and Linux.
Written in Python with PyGame as multimedia toolkit. Kid tested,
grownup approved. Windows/Linux Install Scripts Included.

Download at: http://www.asymptopia.com



From rpluim at grandNOSPAMpied.com  Sat Apr  5 10:43:45 2003
From: rpluim at grandNOSPAMpied.com (Robert Pluim)
Date: 05 Apr 2003 17:43:45 +0200
Subject: POP3 and selective download/deletion
References: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com> <pan.2003.04.05.16.22.31.832230.878@thomas-guettler.de> <luCja.11925$It5.5735@news2.central.cox.net>
Message-ID: <86d6k1ar0u.fsf@fibble.nospam.dev.null>

"Steve Holden" <sholden at holdenweb.com> writes:

> "Thomas Guettler" <pan-newsreader at thomas-guettler.de> wrote in message
> news:pan.2003.04.05.16.22.31.832230.878 at thomas-guettler.de...
> > On Sat, 05 Apr 2003 12:12:26 +0200, Dino Levy wrote:
> >
> > > Hello!
> > > I'm making a mail program and I have a little problem: I don't know how
> > > to achieve the selective downloads/deletion from POP3 servers.
> >
> > Maybe this recipe from the python cookbook does help you to get started:
> >   http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/134945
> >
> > It prints all messages which are biger then N bytes and ask if it should
> > be deleted.
> >
> 
> Unfortunately the POP3 protocol doesn't allow you download the headers of a
> message without downloading the whole thing. IMAP4 is much more capable, but
> it's also a very much more complex protocol.

Que?  You've never heard of the TOP command?
-- 
For mail, remove NOSPAM and translate from French.



From ckea25d02 at sneakemail.com  Tue Apr 15 20:12:29 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 12:12:29 +1200
Subject: Complexity of standard Python data structures
In-Reply-To: <A6Uma.99232$It5.2059@news2.central.cox.net>
References: <mailman.1050354988.8671.python-list@python.org> <tdnk7dwlfda.fsf@shell01.TheWorld.com> <A6Uma.99232$It5.2059@news2.central.cox.net>
Message-ID: <b7i6gl$172lm$1@ID-169208.news.dfncis.de>

Steve Holden wrote:
> "Chris Jones" <clj at acme.com> wrote in message
>
>>People wanting to program should nail a copy of this posting above their
>>keyboard.
> 
> READERS: Please DO NOT FOLLOW THIS ADVICE.
> 
> Following these instructions, I printed the posting out and nailed it
> directly above my keyboard. As a result my laptop no longer works -

He forgot to mention that you need the new experimental
quantum de-interference nails being prototyped by the PSU,
which are capable of penetrating normal matter without
otherwise disturbing it. Unfortunately, it's very difficult
to obtain advance samples of these, and even mentioning
their existence in a public newsgroup tends to result in
the PSU cutting off your




From sholden at holdenweb.com  Thu Apr 17 11:22:18 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 15:22:18 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com> <3e9d9bb9$0$151$a1866201@newsreader.visi.com> <a8b7f07a.0304161314.361a80fa@posting.google.com> <3e9dca74$0$170$a1866201@newsreader.visi.com> <dSxna.114901$It5.77194@news2.central.cox.net> <3e9ec5c7$0$178$a1866201@newsreader.visi.com>
Message-ID: <KEzna.115586$It5.56408@news2.central.cox.net>

"Grant Edwards" <grante at visi.com> wrote in message
news:3e9ec5c7$0$178$a1866201 at newsreader.visi.com...
> In article <dSxna.114901$It5.77194 at news2.central.cox.net>, Steve Holden
wrote:
> > "Grant Edwards" <grante at visi.com> wrote in message
> > news:3e9dca74$0$170$a1866201 at newsreader.visi.com...
> >> In article <a8b7f07a.0304161314.361a80fa at posting.google.com>, A. Lloyd
> > Flanagan wrote:
> >>
> >> >> > python << DONE
> >> >> > print 2
> >> >> > for i in (1,4):
> >> >> >      print i
> >> >> > DONE
> >> >>
> >> >> But how do you do it in a Makefile?
> >> >
> >> > Err.  Put it in a script file, and call the script file from make?
> >>
> >> That's one solution, but now you've got another file to keep
> >> track of.
> >>
> >> > I tried sticking it in a Makefile, can't get it to work.  So far.
> >>
> >> How 'bout this:
> >>
> >> all:
> >>         python2 -c $$'for i in range(5):\n print i\n\n'
> >>
> >>
> >> I think that the $'<string>' notation is a bash-ism.  If that's
> >> a problem, then
> >>
> >>         echo -e "for i in range(5):\n print i\n\n" | python2
> >>
> >> should also work.  Adjust the echo options appropriately to get
> >> interpretation of \n to work.
> >>
> >> There may be other ways to get Make to interpret \n in a
> >> string...
> >>
> > Now try putting an import before the "for" statement ...
>
> OK...
>
>   echo -e 'import sys\nfor i in range(5):\n
sys.stdout.write(str(i)+"\\n")\n\n' | python2
>
> Works for me.  Was it supposed to fail?
>

We=ll, you might say I was just too lazy to test it for myself :-)

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From bokr at oz.net  Thu Apr  3 19:29:01 2003
From: bokr at oz.net (Bengt Richter)
Date: 4 Apr 2003 00:29:01 GMT
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <mailman.1049398267.20730.python-list@python.org> <3E8CBD4F.23406A07@alcyone.com>
Message-ID: <b6ijkd$q6i$0@216.39.172.122>

On Thu, 03 Apr 2003 15:01:35 -0800, Erik Max Francis <max at alcyone.com> wrote:

>Lulu of the Lotus-Eaters wrote:
>
>> Then again, I wouldn't mind adding the new operator ":=" which was
>> "generic augmentation"... the operator could do -absolutely nothing-
>> for
>> standard types, but could be available for custom meaning in custom
>> types.
>
>That strikes me as a better approach; having a special "augmentation"
>operator which has no predefined meaning but can be overridden as
>desired, and thereby leaving usual assignment alone.
>
>The only reservation I might have is whether or not := is a good choice,
>simply because of the :=/= and =/== language differences out there.
>
I feel a bit reserved re the name "generic augmentation" -- how 'bout
something with "update" or "mutate" in it, so x:=y reads "update/mutate x with y"
or "x updated/mutated with y". The latter would read better if  x:=y were
an _expression_ whose value was x after the side effect.

Regards,
Bengt Richter



From pobrien at orbtech.com  Mon Apr  7 10:15:34 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 07 Apr 2003 09:15:34 -0500
Subject: Could Emacs be rewritten in Python?
References: <mailman.1049702963.15645.python-list@python.org>
Message-ID: <m24r5a75rt.fsf@localhost.localdomain>

tanzer at swing.co.at (Christian Tanzer) writes:

> martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=) wrote:
> 
> > pobrien at orbtech.com (Patrick K. O'Brien) writes:
> >
> > > Are there translators out there?  I looked and couldn't find any.
> >
> > Parsing S-Expressions is fairly trivial - they are designed to be
> > readily parsable. For a full translator, you need to define the
> > language mapping first, i.e. how do you want to represent atoms,
> > function names, etc.
> 
> I expect the main difficulty to be the difference in scoping
> models -- the dynamic scoping of Elisp is probably pretty hard to
> simulate in Python without incurring major performance problems. And
> that's used by lots of Elisp code, so skipping it is not an option if
> you want to have comparable behavior.

I'm starting to think this will be the reason *not* to mimick the
Emacs API as closely as I originally thought.  I have neither the
time, energy, nor interest in making an elisp to Python translator.
So I think a better approach will be to simply create a new, Pythonic
API that allows the same kind of programmability as Emacs, but without
trying to adhere to the Emacs syntax or "object model."

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From missive at frontiernet.net  Thu Apr  3 20:44:57 2003
From: missive at frontiernet.net (Lee Harr)
Date: Fri, 04 Apr 2003 01:44:57 GMT
Subject: Persistent server with GUI interactive functions ?
References: <mgqn8v01utuhhioq5vq32rm2r2oadtrm8i@4ax.com>
Message-ID: <ts5ja.90$lJ2.84@news01.roc.ny.frontiernet.net>

In article <mgqn8v01utuhhioq5vq32rm2r2oadtrm8i at 4ax.com>, Frank wrote:
> I need to write a GUI app that will serve as a peer to peer
> client/server.  At the highest level it would be similar to a chat
> client, which has a server component (waiting for any incoming
> connections) as well as the client part (initiating a connection,
> executing user commands).
> 
> Each of those function appears to involve entering an "endless loop"
> as the last logical step to initiate processing (i.e. "mainloop()" for
> the user interface and "serve_forever()" for the server).  I haven't
> yet grasped the logical structure that will allow me to have two
> non-returning calls and maintain control of the program.
> 
> Is there an example of this around somewhere, or does anyone know of
> documentation that will explain this?
> (Note:  I will need to have the app run under Win, Linux, and
> Solaris.)
> 


Have you looked at Twisted?




From ckea25d02 at sneakemail.com  Tue Apr  8 00:47:51 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 16:47:51 +1200
Subject: Lazy string with re.match?
In-Reply-To: <slrnb93ubu.esr.mlh@furu.idi.ntnu.no>
References: <slrnb93ubu.esr.mlh@furu.idi.ntnu.no>
Message-ID: <b6tjnu$8rpnv$1@ID-169208.news.dfncis.de>

Magnus Lie Hetland wrote:
> I first tried
> simply implementing __getitem__, but was told in no uncertain terms
> that a string or buffer was expected by the re object.

What it's really telling you is that it needs all the chars
available all at once in one block of memory (that's what
a "buffer" provides). So, unfortunately, what you want is
impossible to do with the re module.

> Any ideas?

Yes - use Plex instead:

   http://www.cosc.canterbury.ac.nz/~greg/python/Plex/

Plex reads its data sequentially from a stream, just as
you want. The only downside is that it's currently not as
fast as re, being implemented in pure Python. But, even
as we speak, I'm applying Pyrex to this problem, and very
soon I hope to release a much faster version...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ajsiegel at optonline.net  Tue Apr  1 17:54:29 2003
From: ajsiegel at optonline.net (ajsiegel at optonline.net)
Date: Tue, 01 Apr 2003 17:54:29 -0500
Subject: PEP 313 - Roman numeral literals
Message-ID: <1f61ee1f9ce7.1f9ce71f61ee@optonline.net>

>But what does roman(0) do?


Now or in Python3.0?

I // II  perhaps?

Just a convention, of course.

Art





From bokr at oz.net  Sat Apr 26 21:55:07 2003
From: bokr at oz.net (Bengt Richter)
Date: 27 Apr 2003 01:55:07 GMT
Subject: Reading a non-standard floating point format
References: <mailman.1051400681.24630.python-list@python.org>
Message-ID: <b8fd9r$6fo$0@216.39.172.122>

On Sat, 26 Apr 2003 19:41:43 -0400, "Tim Peters" <tim_one at email.msn.com> wrote:
[...]
>There's no way to guess without examples, though.  If you can, show us the
>48-bit byte strings for these numbers:
>
>    0.0
>    1.0
>    1.25
>    2.9e-39
>    1.7e38
>
>That should be enough to fill in the missing pieces.
>
Maybe include
    -1.25
also?
And can minus zero be generated?
Are numbers sign/magnitude or?

Regards,
Bengt Richter



From hupp at cs.wisc.edu  Tue Apr 29 13:23:55 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Tue, 29 Apr 2003 12:23:55 -0500
Subject: import problem
In-Reply-To: <DD7D46ED-7A62-11D7-B1FA-000A95682CD2@vuit.com>
References: <b8m82u$52d$1@news1.wdf.sap-ag.de> <DD7D46ED-7A62-11D7-B1FA-000A95682CD2@vuit.com>
Message-ID: <20030429172355.GA30175@upl.cs.wisc.edu>

On Tue, Apr 29, 2003 at 09:51:47AM -0700, Vince Buonassisi wrote:
> daniel,
> 
> thanks for the response.  i am new to python.  i inherited a project 
> that was written in python from someone who left the company.  he was 
> also the only python programmer here.
> 
> i have been trying to find information about those __init__.py/pyc 
> files that i see laying around and how they relate to running the 
> python scripts.  all of them are empty except for the compiled ones.  
> can you point me to a source that will give me an explanation about the 
> purpose of this file, __init__.py.  i have been unable to find an 
> explanation for this file.
> 
> the test directory was missing the __init__.py file.  i put an empty 
> one in there and my script ran just fine.

See http://www.python.org/doc/2.2.1/tut/node8.html#SECTION008400000000000000000

-Adam




From donn at u.washington.edu  Mon Apr 28 12:22:26 2003
From: donn at u.washington.edu (Donn Cave)
Date: 28 Apr 2003 16:22:26 GMT
Subject: Class problem: loosing my mind
References: <v0fmavkm9iknef7hit5n17d4m0qi6tpnia@4ax.com> <mailman.1051411241.22289.python-list@python.org> <diimavsd60audlnjm80m6dcan0e8b67hae@4ax.com> <3EAB5922.A4F8DDA5@alcyone.com> <gaqnavcfl36jfa2m1k1gre429m5vts5od9@4ax.com>
Message-ID: <b8jkg2$2gnk$1@nntp6.u.washington.edu>

Quoth "<stuffduff>" <stuffduff at cox.net>:
| I think I'm getting this now, and my zope example works.

Excellent.  Now maybe it would be OK to say, I hope you have in
fact been loosing your mind, too, because that's good.  Losing
it would not be so good.

	Donn



From seppi at seppi.de  Tue Apr 15 04:02:59 2003
From: seppi at seppi.de (Sebastian Wilhelmi)
Date: 15 Apr 2003 01:02:59 -0700
Subject: accessing a functions var from a subfunction.
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it>
Message-ID: <a3cae241.0304150002.a04e4ad@posting.google.com>

Hi Alex,

Alex Martelli <aleax at aleax.it> wrote:
> Sebastian Wilhelmi wrote:
> 
> > Hi,
> > 
> > I would like to do the following:
> > 
> > -------8<-------8<-------8<-------8<-------
> > def test ():
> >     count = 0
> >     def inc_count ():
> >         count += 1
> >     inc_count ()
> >     inc_count ()
> >     print count
> > 
> > test ()
> > -------8<-------8<-------8<-------8<-------
> > 
> > This doesn't work (and I even understand, why ;-)
>
> Depends on your tastes in syntax, e.g.:
> 
> def test():
>     class Bunch: pass
>     loc = Bunch()
>     loc.count = 0
>     def inc_count():
>         loc.count += 1
>     inc_count()
>     inc_count()
>     print loc.count

Not my taste ;-)

> def test():
>     test.count = 0
>     def inc_count():
>         test.count += 1
>     inc_count()
>     inc_count()
>     print test.count

Ah, thats great. I tested this, but without the first 'test.'
qualifier, so it didn't work, but this solution is exactly, what I am
looking for.

> Class Bunch (or some highly refined version thereof) IS typically
> around whenever I program, see for example:
> 
> http://mail.python.org/pipermail/python-list/2002-July/112007.html
> 
> for a typical presentation.  Having available the MetaBunch there
> described, I'd start the function as follows:
> 
> def test():
>     class Bunch(MetaBunch): count=0
>     loc = MetaBunch()
>     def inc_count():
>         loc.count += 1

Nice construct. Thanks for your help.
 
Bye,
Sebastian



From abelikov72 at hotmail.com  Thu Apr 17 04:03:25 2003
From: abelikov72 at hotmail.com (Afanasiy)
Date: Thu, 17 Apr 2003 08:03:25 GMT
Subject: Web Hosting
References: <2rjna.171$hT2.182473@news2.news.adelphia.net> <6Bjna.111823$It5.95226@news2.central.cox.net>
Message-ID: <e5os9vg7aj3gfek0abt05s1bjh9pfvfvb3@4ax.com>

On Wed, 16 Apr 2003 21:06:10 GMT, "Steve Holden" <sholden at holdenweb.com>
wrote:

>"Bob X" <bobx at linuxmail.org> wrote in message
>news:2rjna.171$hT2.182473 at news2.news.adelphia.net...
>> Can you name a good (reasonably priced) Web Hosting company that supports
>> Python (either CGI or mod_python)?
>>
>
>See http://www.python.org/cgi-bin/moinmoin/PythonHosting
>
>regards

I was under the impression mod_python was not safe for shared
environments. If this is true, I don't think you will find a
reasonably price host which is also a good host.

If it is not true, I wonder why pair.com will not run mod_python.

CGI, on the other hand, is CGI. :p




From aldo at nullcube.COM  Wed Apr  2 21:18:22 2003
From: aldo at nullcube.COM (Aldo Cortesi)
Date: Thu, 3 Apr 2003 12:18:22 +1000
Subject: ANNOUNCE: pyGDChart2 alpha
Message-ID: <20030403021822.GA25702@nullcube.com>

What is it?
===========

pyGDChart is a Python interface to the popular GDChart
graphing library. This release, pyGDChart2, is a complete
rewrite, featuring:

	- Support for GDChart 0.11.x
	- An OO interface
	- A cleaner overall architecture

For examples of what can be done using pyGDChart2, see the
manual at:

	http://www.nullcube.com/software/pygdchart2/doc

Note that this is an Alpha release, with all the usual
caveats. 


Where can it be found?
======================

pyGDChart2 itself can be found at:
	
	http://www.nullcube.com/software/pygdchart2.html







-- 
Aldo Cortesi
aldo at nullcube.com
http://www.nullcube.com




From rmunn at pobox.com  Wed Apr 30 14:08:58 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 30 Apr 2003 18:08:58 GMT
Subject: Fix for datetime conversion to integer and back to string; was 1 hour off
References: <4f705b08.0304281031.ddd8f75@posting.google.com>
Message-ID: <slrnbavfhh.ltg.rmunn@localhost.localdomain>

La Uruguaya <lacuidadanadelmundo at yahoo.com> wrote:
> I searched this group for a solution to a problem I had converting
> dates from a string format to integer (in seconds since 1970) and back
> to strings.  During some times of the year, the conversion didn't do
> a perfect round trip and I ended up with a datetime that differed from
> the original datetime by one hour.  I found 2 postings with variations
> on the problem, but not a solution.  I implemented the following
> solution and wanted to share it in case somebody needs it later.
[snip code and test]

Interesting: "During some times of the year ...", "... differed from the
original datetime by one hour." Sounds to me like you're running into
daylight savings time issues.

Yes, looking at your code confirms it: you're subtracting the value of
the last element of the tuple (the DST flag, which is 1 if daylight
savings time is in effect) from the third element (the hours).

Did you figure this out from looking at the docs for the time module, or
did you do it by trial and error? Also, it would have been nice if you
had explained more clearly that daylight savings time issues were the
root of your problem; at first glance, your code looked like a "magic"
solution to a strange kind of problem.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From linux42 at freemail.c3.hu  Wed Apr 16 10:53:11 2003
From: linux42 at freemail.c3.hu (Gabor Nagy)
Date: Wed, 16 Apr 2003 16:53:11 +0200
Subject: pattern match
Message-ID: <20030416145311.GA4031@swordfish.capgemini.hu>

I have a string, like '(FOO(1) BAR("2"))'

I'd like to have a { 'FOO':1, 'BAR':'2' }

Is there an easy way to do this, or do I have to write parsing, maybe use
regexps, or something like that?

Regards,
Gee




From eric.brunel at pragmadev.com  Fri Apr 25 05:44:14 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Fri, 25 Apr 2003 11:44:14 +0200
Subject: Date validation
References: <mailman.1051255479.31840.python-list@python.org>
Message-ID: <b8avlm$eeh$1@news-reader13.wanadoo.fr>

James Gregory wrote:
> Hi all,
> 
> I'm writing a killer web app in python and I want to validate all the
> input it gets. Amongst said input are dates represented as text strings
> of the format 'YYYY-MM-DD' - how can I check whether or not such strings
> represent a valid date? I had thought that
> 
>     time.strptime('2003-4-31', '%Y-%m-%d')
> 
> would throw an exception (since gnome-something tells me that april only
> has 30 days), but it just gives me back
> 
>     (2003, 4, 31, 0, 0, 0, 3, 121, 0)
> 
> indeed I'm even able to use mktime on that without an exception.

And you even can to

 >>> tt = time.strptime('2003-4-31', '%Y-%m-%d')
 >>> t = time.mktime(tt)
 >>> print time.localtime(t)
(2003, 5, 1, 1, 0, 0, 3, 121, 1)

So just testing if you get the same day and month in your original one and the 
one returned by localtime may be enough.

BTW, allowing invalid dates in time tuples is in fact quite useful. For example, 
you can easily get the day after a given date by doing:

time.localtime(time.mktime((year, month, day+1, 0, 0, 0, 0, 0, -1)))

See:

 >>> time.localtime(time.mktime((2003, 2, 28+1, 0, 0, 0, 0, 0, -1)))
(2003, 3, 1, 0, 0, 0, 5, 60, 0)
 >>> time.localtime(time.mktime((2000, 2, 28+1, 0, 0, 0, 0, 0, -1)))
(2000, 2, 29, 0, 0, 0, 1, 60, 0)

HTH
-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From dave at pythonapocrypha.com  Wed Apr  2 12:21:34 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Wed, 2 Apr 2003 09:21:34 -0800 (PST)
Subject: Socket timeouts for HTTP connections (httplib)?
In-Reply-To: <9df3d134.0304020550.7e32bf6b@posting.google.com>
References: <9df3d134.0304020550.7e32bf6b@posting.google.com>
Message-ID: <Pine.LNX.4.51.0304020916520.17427@ns0.tbye.com>

On Wed, 2 Apr 2003, nick wrote:

> I have neither an expert understanding of socket programming,
> threading or asyncronous IO, yet can get by using threading.Thread and
> httplib.  I am therefore looking for a timeout at a fairly high level
> (ie the httplib api).
>
> There were rumblings about Guido implementing this at a socket level
> in June 2002, but I've not been able to find anything since.

In newer versions of Python sockets can have timeouts, but someone
recently posted that they don't work for the high-level APIs.

> Has/is someone involved with this?  timeoutsocket simply doesn't work
> for me and throws continuous "connection reset by peer" exceptions.
> So I'm completely stuck for options.

Maybe the best thing to do is to post some code that exhibits this
problem, because many people have used timeoutsocket with the higher level
APIs and had lots of success with it. You _should_ be able to just drop it
into your code and catch the appropriate timeout exception, so getting it
to work is probably your best bet.

-Dave




From mikeb at mitre.org  Wed Apr 16 13:10:49 2003
From: mikeb at mitre.org (Mike Brenner)
Date: Wed, 16 Apr 2003 13:10:49 -0400
Subject: How many of you are Extreme Programmers?
Message-ID: <3E9D8E99.59342555@mitre.org>

Most of the previous answers stated that they
do the continual and thorough testing portion of XP.

Continual and thorough testing was not new to 
XP, and XP did not invent that. We, too, do a lot of that.

To be Xtreme, you also need to overcome the social barriers, 
quite a difficult feat. Getting people to see the advantage,
for example, of two people working together at all times
is challenging (to the territorial ego of almost all
developers).





From tismer at tismer.com  Sat Apr 19 00:17:00 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sat, 19 Apr 2003 06:17:00 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <0191avo4391fumu3rg6k0ok6lrbde52hro@4ax.com>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <b7pchr$ig0$1@main.gmane.org> <mailman.1050690871.4874.python-list@python.org> <0191avo4391fumu3rg6k0ok6lrbde52hro@4ax.com>
Message-ID: <3EA0CDBC.9020505@tismer.com>

Courageous wrote:
>>AFAIK, generator functions have turned out to be quite difficult in Java
>>too, but that didn't seem to stop their acceptance.
> 
> 
> I'll chime in here.
> 
> I am emphatically _against_ targeting subsets of features that cater
> to a virtual machine. The Java guys can get it together, or fall into
> a hell of conformity, for all I care. Same with .NET.

I have to say that I never took the Jython argument for serious.
This might come from that fact that I never took Java for serious,
btw.
Without tring to offend anybody, I have the feeling that the main
motivation to bring that argument up was to put continuations
down, finally, which I now understand to some extent.
Stackless 3.0 does not support them, it has a clean and mean
concept of tasklets, it can switch them at blindingly speed,
and it will exist as a software-only (meaning without C stack
mangling) alternative, soon.
I'm kind of waiting for the new set of arguments why this
cannot be included into the distribution. I am obviously
solving a problem that nobody has. Funny that I'm paid by
companies, just to continue my work.
This suggests that "nobody" doesn't eval to 0, but at least 3 :-)

expect-3.0.1-shortly-after-easter -- sincerely -- chris

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From ckea25d02 at sneakemail.com  Tue Apr  8 00:24:26 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 16:24:26 +1200
Subject: Got a kick out of this
In-Reply-To: <237247fb.0304071335.619a468b@posting.google.com>
References: <237247fb.0304071335.619a468b@posting.google.com>
Message-ID: <b6tic1$8os9s$1@ID-169208.news.dfncis.de>

Terin wrote:
> 
> http://vil.nai.com/vil/content/v_994.htm

Well, some people call Windows a boot virus... so
I suppose Python could be regarded as a programmer's
brain virus...

Name:           Python
Preferred Host: Perl, C, C++, PHP, sh, csh programmers
Symptoms:       The host's programs suddenly become shorter,
                 more readable and easier to maintain.
Spread By:      Usenet/email contact, word of mouth
Treatment:      No effective treatment known. Once infected,
                 the host will never willingly use any other
                 programming language.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From evan at 4-am.com  Thu Apr  3 10:54:55 2003
From: evan at 4-am.com (Evan Simpson)
Date: Thu, 03 Apr 2003 09:54:55 -0600
Subject: Zope style (was: tuning our site
In-Reply-To: <v8mjffpuph1oc2@corp.supernews.com>
References: <mailman.1046819836.25704.python-list@python.org> <mailman.1047414442.16521.python-list@python.org> <v81d26sb3vf170@corp.supernews.com> <irdbl-1ds.ln1@news.lairds.org> <v8mjffpuph1oc2@corp.supernews.com>
Message-ID: <3E8C594F.8050708@4-am.com>

Cameron Laird wrote:
> There are a couple of pertinent facts.  First, much of the
> passion "Web designers" are supposed to have for "templates"
> comes from the supposition that they want stuff to look like 
> HTML, so they can use their "HTML editors".  Is *anyone* doing
> that with Zope, though?  It would surprise me.

Be surprised :-)  This was a major motivator behind the design of Zope 
Page Templates (ZPTs).  Unlike DTML, ZPT makes it fairly easy to create 
a template from a Web designer's mockup that the designer can continue 
to maintain, with a small amount of care and training.  ZPTs don't 
replace all of DTML's uses in Zope (e.g. SQL generation), but they are 
now the preferred tool for XML/HTML.

 > Interesting!  I want to be clear on the work-flow; you'd edit
 > and preview in HomeSite, then switch back to the Zope manage-
 > ment screens, and use the "Upload File" selection to update
 > the site itself--do I have that right?  Have I just demon-
 > strated my ignorance of some far slicker interface?

Yep:  The exquisitely excellent External Editor, by caseman.  It puts a 
pencil icon next to most Zope objects in a Folder listing.  When clicked 
on, this icon interacts with a local Python helper script (associated in 
your browser via mime-type) to open the object in the editor of your 
choice.  It WebDAV-locks the object, then unlocks it when the editor 
releases the file.  It sends changes to Zope whenever it detects that 
the editor has modified the file.

Between them, ZPTs and External Editor have made me several times more 
productive, and an order of magnitude happier to use Zope.

Cheers,

Evan @ 4-am






From max at alcyone.com  Mon Apr 21 15:13:54 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 21 Apr 2003 12:13:54 -0700
Subject: "pythonic"
References: <mailman.1050921572.30147.python-list@python.org>
Message-ID: <3EA442F2.5F44982F@alcyone.com>

soporte_cali wrote:

> I am quite new to python so please forgive me if my question is too
> newby. I am running a phyton exe file, linux red hat, it is an
> indexing
> system that takes a flat file and index it. the py program once runned
> is giving  me this:
> 
>  ./index.py:57: SyntaxWarning: import * only allowed at module level
>   def createIndex(dbname,idtokenNum=1,keySize=10,idtokenChar=' '):
> error: unknown library " <nameoflibrary.py> " does not exit
> Library  nameoflibrary.py  Does not exist
> 
> so where should I place the name of this library? I mean is there any
> special place to put it? libnames? well thanks in advance..

You haven't shown us the code that generates this warning and error, so
it's very difficult to say what the resolution will be.  The first
warning is generated by something like:

	def f():
	    from someModule import *

and is simply telling you that using the "from ... import *" syntax is
only officially allowd at the toplevel of a module, e.g.:

	from someModule import *

	def f():
	    ...

(But there are lots of good reasons not to use "from ... import *" in
the first place; it decontructs the very namespaces which modules were
meant to create, and can encourage name collisions.)

As for the second error, I can't say.  My only guess from the error
would be you wrote something like:

	import <someModule.py>

when you really meant

	import someModule

(no angle brackets, no .py extension).  It may well be the very same
line where you do from ... import *.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The revolution will not televised
\__/ Public Enemy
    Maths reference / http://www.alcyone.com/max/reference/maths/
 A mathematics reference.



From andersjm at inbound.dk  Tue Apr 15 10:53:39 2003
From: andersjm at inbound.dk (Anders J. Munch)
Date: Tue, 15 Apr 2003 16:53:39 +0200
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1049901964.20840.clpa-moderators@python.org> <b72fo5$bq6$1@panix1.panix.com> <87ptnuxryk.fsf@theasylum.dyndns.org> <b7g2fk$g1u$1@panix3.panix.com> <87smsj4z4t.fsf@theasylum.dyndns.org>
Message-ID: <h2Vma.17871$y3.1620573@news010.worldonline.dk>

"Tim Lavoie" <tlavoie at acm.org> wrote:
>     aahz> So how does Capon differ from SCons?
>     >>  Well, the name has funnier connotations, for one.  :)
>
>     aahz> You deserve to be roasted for that comment.
>
> Well, sometimes you stick your neck out...
> "Capon: Batteries included, nuts aren't."   :)

<g> actually the name comes from the boring "Cascading property
notation".  Gives me an idea for a slogan though: "Less is more".

nah-maybe-not-ly y'rs, Anders






From pinard at iro.umontreal.ca  Sat Apr 19 08:44:33 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 19 Apr 2003 08:44:33 -0400
Subject: hexadecimal calculations
In-Reply-To: <b7q0uv$hco$1@aphrodite.grec.isp.9tel.net>
References: <b7q0uv$hco$1@aphrodite.grec.isp.9tel.net>
Message-ID: <oqfzoe3bd9.fsf@titan.progiciels-bpi.ca>

[Andros]

> how to make these hexadecimal calculations in python?
> A0 XOR 79 = D9
> D9 - A8   = 31

hex(0xa0 ^ 0x79)
hex(0xd9 - 0xa8)

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From theller at python.net  Wed Apr  9 02:16:25 2003
From: theller at python.net (Thomas Heller)
Date: 09 Apr 2003 08:16:25 +0200
Subject: How to build Numeric in Debug mode
References: <99457f2.0304081215.214a5a9c@posting.google.com>
Message-ID: <4r58w5za.fsf@python.net>

edo4 at emich.edu (Ngoc) writes:

> I want to use Numeric.array from Numeric module.  I can run python
> with import Numeric in release but not debug mode. What is the command
> to build Numeric module in debug version ?
> >>python_d setup.py install doesn't work??
> Thanks in advance.
> Eve
Assuming Numeric uses standard distutils, you can build a debug
version giving the '-g' flag to build:

python setup.py build -g install

Thomas



From bob at redivi.com  Tue Apr 22 17:50:46 2003
From: bob at redivi.com (Bob Ippolito)
Date: 22 Apr 2003 14:50:46 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com> <a3b5b608.0304080302.c26ad68@posting.google.com> <af6b5bdd.0304181001.558da6f@posting.google.com> <84of317cd6.fsf@boost-consulting.com> <b7vlfb$27c$0@216.39.172.122> <84of30m4cp.fsf@boost-consulting.com> <p0p8avsd7nlqu52p4rnjobuemctd3foil8@4ax.com> <848yu3o5o6.fsf@boost-consulting.com>
Message-ID: <6a36e729.0304221350.5d30c3bb@posting.google.com>

David Abrahams <dave at boost-consulting.com> wrote in message news:<848yu3o5o6.fsf at boost-consulting.com>...
> Mark E. <snowball3 at bigfoot.com> writes:
> 
> > David Abrahams <dave at boost-consulting.com> wrote:
> >
> >>    xml.sometag(attr1="value1", attr2="value2", ...)[ "contents" ]
> >
> > Of course there are always corner cases to deal with. For example, what's a good way
> > to handle XHTML's 'class' attribute?
> >
> > xml.sometag(class="css_class_name")["contents"]
> > won't work.
> >
> > One could map 'foo_' to 'foo' to handle conflicting names used in xml/html, or allow
> > a dictionary to serve the same purpose:
> >
> > xml.sometag({"attr1" : "value1", "class" : "value2"})["contents"]
> 
> Sure, you have lots of options. Also:
> 
>     xml.sometag(class_("value2"), attr1 = "value1")["contents"]
> 
>     xml.sometag({'class':"value2", 'while':"value3"}, attr1 = "value1")["contents"]
> 
>     xml.sometag(attr1 = "value1", _ = {'class':"value2", 'while':"value3"})["contents"]
> 
> it's all in the fun of designing domain-specific sublanguages.

well, in our implementation (off the top of my head) these are valid:

xml.sometag(_class = 'value2', attr1='value1')['contents']
xml.sometag({'class':'value2'}, attr1='value1')['contents']
xml.sometag('class="value2"', attr1='value1')['contents']
xml.sometag(('class', 'value2'),), attr1='value1')['contents']

The preferred is the first, it's easiest.  It's assumed that you will
never want an attribute that actually starts with an underscore.  If
you do (probably never), you'll have to use one of the other three
options.

-bob



From mirko-lists at zeibig.net  Tue Apr  1 04:24:34 2003
From: mirko-lists at zeibig.net (Mirko Zeibig)
Date: Tue, 01 Apr 2003 11:24:34 +0200
Subject: Catching webbrowser error in RedHat
References: <v8edjq9cc4o13@corp.supernews.com>
Message-ID: <b6blsh$mea$1@news.web.de>

Lance wrote:
> I'm running Python 2.2 under RedHat8. I use webbrowser to open an html
> file in a browser window. This is my code:
> 
>         try:
>             webbrowser.open(fname)
>         except:
>             print "ERROR: "+fname+" not found."
> 
> 1. A Redhat "splash" screen shows for a few seconds prior to my desired
> screen when try block is successful. Can this splash screen be disabled?
Hello Lance,

is this really a "splash" screen? Maybe the webbrowser is just showing the
predefined homepage. Try to start Mozilla, set the homepage to blank and
see what happens.

Regards
Mirko
-- 
"Ein, zwei, drei im Sauseschritt, saust die Zeit, wir sausen mit!"
(Busch sein Willi, und nit sein Georg)



From alloydflanagan at attbi.com  Tue Apr  8 10:22:20 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 8 Apr 2003 07:22:20 -0700
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com>
Message-ID: <a8b7f07a.0304080622.d7cb15b@posting.google.com>

msw <markscottwright at hotmail.com> wrote in message news:<u8bka.651$001.591065 at monger.newsread.com>...
> 
> I'm really interested in this as well, and I think that it's an 
> important issue for python.  Unfortunately, as I'm sure you've 
> discovered by now, it's a question that comp.lang.python is 
> constitionally incapable of answering.  The only answers you will get in 
> this group are:
> 
> 1. intellectual property law is sufficient
> 2. no protection scheme is perfect
> 

I suspect you're seeing the effect of selective response here.  In
other words, there are two kinds of people on c.l.py:
1.  Those who think any kind of code hiding is immoral, unhealthy, and
probably causes cancer; and
2.  Those who have no need to hide code but could care less whether
you do or not.

I suspect the great majority (including myself) are in group 2, but
generally only group 1 cares enough to actually respond.  So it looks
like everybody's in group 1.  Come to think of it, I guess I'm in
group 3) people who belong to group 2, but can't resist adding their
two cent's worth to the discussion anway. :)



From max at cNOvSisiPonAtecMh.com  Sun Apr  6 12:10:46 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Sun, 06 Apr 2003 16:10:46 GMT
Subject: Novelwriting 0.3: rule-based random document generation
References: <mailman.1049563988.5389.python-list@python.org>
Message-ID: <akYja.6653$n95.2301201@twister.nyc.rr.com>

Oh, goody! I was looking for some assistance with my next novel, :)

--
========================================
Max Khesin, software developer -
max at cNvOiSsPiAoMntech.com
[check out our image compression software at www.cvisiontech.com, JBIG2-PDF
compression @
www.cvisiontech.com/cvistapdf.html]


"Jeff Epler" <jepler at unpythonic.net> wrote in message
news:mailman.1049563988.5389.python-list at python.org...
> http://unpythonic.net/~jepler/novelwriting/index.html
>
> Description
> -----------
>
> Novelwriting[1] generates random text based on a user-defined grammar.
> It is similar to the Dada Engine[2], which is used by the Postmondernism
> Generator[3].  The name was chosen based on the Monty Python sketch,
> "Thomas Hardy / Novel Writing"[4].  It requires Python 2.2.
>
>
> Example
> -------
> Start: folk-saying;
> folk-saying: person " once said that " folk-saying ".\n";
> person: "My " relative | "The President" | "Cardinal Fang";
> relative: "mother" | "aunt" | "grandmother" ;
> folk-saying: "a rolling stone gathers no moss"
>     | "a stitch in time saves nine"
>     | "no-one expects the Spanish Inquisition";
> ;;
>
>
> Advantages compared to the Dada Engine
> --------------------------------------
> The Dada Engine has only a limited degree of programmability.  In
> Novelwriting, you can use the full power of Python at any point in a rule.
> For example:
> Start: product-claim
> product-name: "Novelwriting" | "Dada Engine";
> product-claim:
> @set(this-product, product-name)
> " is twice as good as "
> @alternative(product-name, @get(this-product));
> ;;
> from novelwriting.calls import get, set, alternative
> (The facility provided by @get/@set is available in the Dada Engine,
> but @alternative is impossible to write)
>
>
> Limitations compared to the Dada Engine
> ---------------------------------------
> Novelwriting does not have a formatting library like the Dada Engine.
> Novelwriting does not have a nice syntax for text substitution rules or
> maps.  Novelwriting does not have parameterized rules.
>
>
> [1] http://unpythonic.net/~jepler/novelwriting/index.html
> [2] http://dev.null.org/dadaengine/
> [3] http://www.csse.monash.edu.au/cgi-bin/postmodern
> [4]
http://www.stone-dead.asn.au/albums-cds/sketches/matching-tie-and-handkerchi
ef/thomas-hardy-novel-writing.html
>





From cjw at sympatico.ca  Tue Apr  1 09:16:38 2003
From: cjw at sympatico.ca (Colin J. Williams)
Date: Tue, 01 Apr 2003 09:16:38 -0500
Subject: The Missing Manual? (Re: Finding Python help...)
References: <3E88310F.6090709@mq.edu.au> <b6aanj$39skv$1@ID-71831.news.dfncis.de> <b6am14$3f4rk$1@ID-169208.news.dfncis.de>
Message-ID: <3E899F46.E71AC153@sympatico.ca>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030401/eac04d19/attachment.html>

From Vector180W at netscape.net  Wed Apr  9 01:51:36 2003
From: Vector180W at netscape.net (vector)
Date: Wed, 09 Apr 2003 15:51:36 +1000
Subject: PythonWin Open(file) path bug?
Message-ID: <3E93B4E8.6020003@netscape.net>

newbie,pythonwin,2.2.2
im simply trying to open a file
script snip...
cfg_file = open("lptctrl.cfg", "rb")     #open  config text file
end snip..

  File "E:\Md_prj\md_422\Code\Python\LptCtrl_104.py", line 20, in __init__
     cfg_file = open("lptctrl.cfg", "rb")     #open  config text file
IOError: [Errno 2] No such file or directory: 'lptctrl.cfg'

the file is in the same folder as the python script
i get the same error if opened from the interactive window
import sys
print sys.path
indeed shows the script dir is on the path so why cant it find the 
config file?

If i call the script direct from explorer (no IDE) it works fine
if I open lptctrl.cfg within pythonWin and save As it then runs and 
finds the file(it must be save as, even modifing and just save dosent 
change the error)

any ideas
a bug or am i dong something wrong
I dont want absolute addressing, the file will always be in the "current 
dir"




From list-python at ccraig.org  Fri Apr  4 10:51:47 2003
From: list-python at ccraig.org (Christopher A. Craig)
Date: 04 Apr 2003 10:51:47 -0500
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304040445.a3752dd@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com>
	<3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>
	<9383481e.0304030705.524cf0fa@posting.google.com>
	<7h3znn7k308.fsf@pc150.maths.bris.ac.uk>
	<3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122>
	<3E8CF4B5.1E221DD7@alcyone.com>
	<9383481e.0304040445.a3752dd@posting.google.com>
Message-ID: <t1whe9ejm5o.fsf@kermit.wreck.org>

drew at astro.pas.rochester.edu (Drew Moore) writes:

> so far, the response seems to be:
> "it is technically possible, but my scripts might
> run slower.. and it just makes me nervous."

Not might run slower, would run not just slower but a ton slower.
You're basically replacing the C code
a=b

with at least one, and possibly several dict lookups.  Your idea about
caching the value is a decent one, but you have the problem that class
dicts are mutable, so what if someone does

>>> class foo:
...  pass
...
>>> t = foo()
>>> t.__dict__.__assign__ = lambda t, a: t.elem=a

If __assign__ is allowed, that will be expected to work.

A bigger problem to me is that '=' currently means rebinding names.
If I see 'a=b' in Python, I immediately presume that 'a' doesn't exist
after the assignment.  If you change this then it will sometimes mean
rebinding and sometimes mean something else.

If you want a way to set something's value, I imagine you'll have a
lot more support for adding an operator rather than making '=' one.
I'm all for Lulu's suggestion of adding a ':=' operator for that sort
of situation.


-- 
Christopher A. Craig <list-python at ccraig.org>
"The world would be a better place if Larry Wall had been born in Iceland, 
 or any other country where the native language actually has syntax." 
                -- Peter da Silva (in alt.sysadmin.recovery) 




From phil at river-bank.demon.co.uk  Fri Apr  4 04:22:06 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Fri, 4 Apr 2003 10:22:06 +0100
Subject: PyQt QTableItem inheritance...
In-Reply-To: <x7istwiwzm.fsf@guru.mired.org>
References: <x77kaejf9b.fsf@guru.mired.org> <mailman.1049223123.616.python-list@python.org> <x7istwiwzm.fsf@guru.mired.org>
Message-ID: <200304041022.06728.phil@river-bank.demon.co.uk>

On Wednesday 02 April 2003 7:18 pm, Mike Meyer wrote:
> Phil Thompson <phil at river-bank.demon.co.uk> writes:
> > On Tuesday 01 April 2003 5:31 pm, Mike Meyer wrote:
> > > I've already tried posting this to the pyKDE mail list, and gotten no
> > > response.
> > >
> > > I'm trying to create an item in a QTable that uses the
> > > QDoubleValidator on the input edit field. The only way I can see to do
> > > this is subclass QTableItem, and override createEditor with a method
> > > that sets the validator, like so:
>
> [code elided]
>
> > > While I get the output from __init__, I never see the output from
> > > createEditor. Any help would be appreciated.
> >
> > A small but complete example would help.
>
> The rest of it is biolerplate, but here it is.
>
>         <mike
>
> import sys
> from qt import *
> from qttable import *
>
> class Float_Field(QTableItem):
>
>     def __init__(self, table, data):
>         print "Creating float field."
>         QTableItem.__init__(self, table, QTableItem.OnTyping, str(data))
>
>     def createEditor(self):
>         print "createEditor"
>         editor = QLineEdit(self.text(), self.table().viewport())
>         self.connect(editor, SIGNAL("textChanged(QString &)"),
>                      self.table(), SLOT("doValueChanged"))
>         editor.setValidator(QDoubleValidator(self))
>         return editor
>
>
> app = QApplication(sys.argv)
> table = QTable(1, 1)
> table.setItem(1, 1, Float_Field(table, "12.34"))
> app.setMainWidget(table)
> table.show()
> app.exec_loop()

There are a number of problems here...

connect() is a method of QObject derived classes and QTableItem isn't derived 
from QObject. Use either QObject.connect() or editor.connect().

"textChanged(QString &)" is not a QTable signal - it's "textChanged(const 
QString &)".

QTable doesn't have a slot called "doValueChanged". If you are trying to 
handle the signal in Python then the call to connect() should be something 
like...

	editor.connect(editor, SIGNAL("textChanged(const QString &)"),
			self.doValueChanged)

...and implement doValueChanged() as a method of Float_Field...

	def doValueChanged(self, value):
		print "Changed value:", str(value)

The validator will not work as QLineEdit.setValidator() doesn't take ownership 
of it and it will be garbage collected immediately afterwards. Do something 
like...

	self.validator = QDoubleValidator(editor)
	editor.setValidator(self.validator)

Finally, the call to QTable.setItem() is wrong. The only valid coordinates for 
this table are (0, 0).

With those changes double-clicking on the cell will create the editor and 
validate any changes.

Attached is a corrected script.

Phil

-------------- next part --------------
A non-text attachment was scrubbed...
Name: table.py
Type: text/x-python
Size: 859 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030404/d36892bb/attachment.py>

From aleax at aleax.it  Tue Apr  1 04:23:24 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 09:23:24 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com>
Message-ID: <gUcia.53479$i26.1635692@news2.tin.it>

Max Khesin wrote:

> I personally question such beliefs 'a priori' from language design
> perspective but justify them retrospectively in terms of soft
> engineering/economis. What I mean is the following: a language that gives
> you more of an access to the machine architecture (whether you like it or
> not) is capable of no less optimization (and most likely more) than a
> language that is further removed from it. Python vs. C is a particularly
> poigniant example since Python is a C program. On a minute level such C++

No more poignant than C vs machine code.  The tradeoff is similar.  Yes,
in theory, if you have unbounded development time available, you can
always write machine-code programs that will outperform C or C++ ones,
and can always wite C or C++ programs that will outperform Python ones
(unless, perhaps, the Python ones are using machine code that your C
or C++ compiler cannot generate).  If there are any bounds on the time
you can spend developing your programs, then the theoretical situation
is not so clear and sharp any more (partly depending on what set of
already-written code you can be thought to have at hand in each case).

The theory is only important up to a point.  In _practice_, a mix of
SOME Python, SOME C/C++, and, yes, even some machine code when that's
appropriate (check out GMP implemented in C vs machine code, most
particularly on an Athlon... if you do heavy multiprecision arithmetic,
the performance difference is important!), is almost invariably the
optimal choice -- main exception being when the target machine just
doesn't have the needed resources (e.g. some embedded development),
mostly memory, in which case using just C+machine code, or even just
machine code in some cases, CAN be pragmatically optimal.


Alex




From nessus at mit.edu  Thu Apr 17 15:56:17 2003
From: nessus at mit.edu (Douglas Alan)
Date: Thu, 17 Apr 2003 15:56:17 -0400
Subject: A new and very robust method for doing file locking over NFS?
Message-ID: <lcu1cw525a.fsf@gaffa.mit.edu>

I'd like to do file locking over NFS without using lockd.  The reason
I want to avoid using lockd is because many lockd implementations are
too buggy.

It is fairly easy to avoid using lockd -- just avoid using lockf() to
lock a file.  Instead of using lockf(), lock a file by creating a lock
file that you open with the O_CREAT | O_EXCL flags.  To unlock the
file, you merely unlink the lock file.  This method is fairly
reliable, except that there is a small chance with every file lock or
unlock that something will go wrong.

The two failure symptoms, as I understand things, are as follows:

(1) The lock file might be created without the client realizing that
it has been created if the file creation acknowledgement is lost due
to severe network problems.  The file being locked would then remain
locked forever (until someone manually deletes the lock) because no
process would take responsibility for having locked the file.  This
failure symptom is relatively benign for my purposes and if needed it
can be fixed via the approach described in the Red Hat man page for
the open() system call.

(2) When a process goes to remove its file lock, the acknowledgement
for the unlink() could be lost.  If this happens, then the NFS driver
on the client could accidentally unlink a lock file created by another
process when it retries the unlink() request.  This failure symptom is
pretty bad for my purposes, since it could cause a structured file to
become corrupt.

I have an idea for a slightly different way of doing file locking that
I think solves problem #2 (and also solves problem #1).  What if,
instead of using a lock file to lock a file, we rename the file to
something like "filename.locked.hostname.pid"?  If the rename()
acknowledgement gets lost, the client will see the rename() system
call as having failed due to the file not existing.  But in this case
it can then check for the existence of "filename.locked.hostname.pid".
If this file exists, then the process knows that the rename() system
call didn't actually fail--the acknowledgement just got lost.  Later,
when the process goes to unlock the file, it will rename the file back
to "filename".  Again, if the rename system call appears to fail, the
process can check for the existance of "filename.locked.hostname.pid".
If the file no longer exists, then it knows the rename call really did
succeed, and again the acknowledgement just got lost.

How does this sound?  Is this close to foolproof, or am missing
something?

I'm not much of an NFS expert, so I am a bit worried that there are
details of NFS client-side caching that I don't understand that would
prevent this scheme from working without some modification.

|>oug



From nhodgson at bigpond.net.au  Mon Apr 14 15:54:38 2003
From: nhodgson at bigpond.net.au (Neil Hodgson)
Date: Mon, 14 Apr 2003 19:54:38 GMT
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il>
Message-ID: <2mEma.5369$hF.43953@news-server.bigpond.net.au>

Avner Ben:

>     I have tried the lower them all approach, but at the rate of
processing
> I need, it has turned out to reduce application speed by a factor of four.
> What I need is an efficient built-in way of doing this.

   Could you explain your processing? I have found the case-standardization
and then compare much faster than in-place case-insensitive comparison when
for example, matching words in a program against a list of case-insensitive
keywords. Or when sorting.

   Neil





From iop199 at ecs.soton.ac.uk  Thu Apr 10 14:10:38 2003
From: iop199 at ecs.soton.ac.uk (Ian Partridge)
Date: Thu, 10 Apr 2003 19:10:38 +0100
Subject: Writing a TCP/IP chat server
Message-ID: <Pine.LNX.4.44.0304101906590.8243-100000@servalan.ecs.soton.ac.uk>

Hi,

I am trying to write a chat server, which users will telnet to in order to 
interact with each other.

What is the best way to do this?

I have searched the web and found three different approaches:

1. Use the socket libraries and select to multiplex between the 
connections.

2. Use the SocketServer module (although I can't find a decent tutorial on 
how to use this module)

3. Use the asyncore module (I only found one example doing this, and I 
couldn't understand it).

What's the the simplest way to approach the problem?

Any tips gratefully received.

TIA,
Ian




From anton at vredegoor.doge.nl  Sat Apr 26 14:36:11 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Sat, 26 Apr 2003 20:36:11 +0200
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com> <b8b7kj$g74$1@news.hccnet.nl> <slrnbal5ap.jc7.news@datatailors.com>
Message-ID: <b8eluh$kna$1@news.hccnet.nl>

Peter van Kampen <news at datatailors.xs4all.nl> wrote:

>> This discussion about whether throwaway scripts tend to get thrown
>> away or not is too heavily based on the wrong Dutch quotes. A more
>> fitting quote would be "ieder voordeel heb zijn nadeel" from a famous
>> Dutch soccer player (and later soccer trainer in Italy). It means as
>> much as "every advantage has its disadvantage".
>
>Just nitpicking: He (Johan Cruyff) went to Barcelona, Spain. And the
>exact quote is: "Elk nadeel heeft zijn voordeel" (Elluk nadeel hep se
>fordeel)

Thanks for correcting me, that's the way science is supposed to work,
after all :-) Here's a link (Dutch) quoting "Kruif" inexactly, as I
did:

http://www.abcta.nl/webapp/abcta/abctanewsitem.jsp?newsitemid=287

Anton



From mwm at mired.org  Fri Apr  4 13:57:15 2003
From: mwm at mired.org (Mike Meyer)
Date: Fri, 04 Apr 2003 18:57:15 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org> <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk> <x7el4jibzl.fsf@guru.mired.org> <Kp2ja.2589$y3.426276@news010.worldonline.dk>
Message-ID: <x7znn6gkfn.fsf@guru.mired.org>

"Anders J. Munch" <andersjm at inbound.dk> writes:
> "Mike Meyer" <mwm at mired.org> skrev:
> > "Anders J. Munch" <andersjm at dancontrol.dk> writes:
> > > "Mike Meyer" <mwm at mired.org> wrote:
> > > > I disagree with that statement. Clearly, *someone* knows all the
> > > > coding tricks used by the compiler - because they put it into the
> > > > compiler. So they can achieve the same level of performance.
> > > And chess program authors can beat their own inventions?  Not so.
> > Go program authors can beat their own inventions, so clearly this
> > depends on the problem domain. In the chess (and go) problem domains,
> > the time the human spends on the problem is artificially limited by
> > the rules of the game. For programming, limiting the programmer time
> > is a goal, not an axiom.
> Here's a brilliant algorithm for register allocation that I just stole
> from the PSU labs.
> 1. Generate all possible combinations.
> 2. Score each of them according to estimated performance.
> 3. Pick the best scoring combination.
> Now try with 8 registers and 10 variables.  1,073,741,824 variations.
> 
> No limit on programmer time?

You just have to hire more programmers.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From Sorry at ButNo.com  Thu Apr 24 14:56:15 2003
From: Sorry at ButNo.com (Hyperion)
Date: Thu, 24 Apr 2003 19:56:15 +0100
Subject: ActivePython Scripts help with ?
Message-ID: <b89bu8$5bp$1@newsg2.svr.pol.co.uk>

Hi all

 I am new to python, and although I am learning at a good rate, things are
still a little shacky.

 Take today for instance, being new to linux also, I caused a problem for
myself.

 You see i wantedm idle, but on slackware 9.0 idle wasnt there with the 2.2
python, although it is supposedly to be shiped with it, anyway, with coming
from ActivePython on my windows xp box, I decided to install ActivePython
on my slackware box and take out the original install.

Wll I maniged to install it fine, but now I haven't got the slightest clue
what to put at the top of my scripts so the interpreter or what ever can
find them.
 You no the little bit at the top that starts with,  #!

The ActivePython Install is setup here as follows:

---------------
/home/me/ActivePython-2.2/bin
---------------

. I added that line to me PATH also in my /etc/profile  file as follows:

--------------------
# Set the default system $PATH:
PATH="/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/usr/games:/home/me/ActivePython-2.2/bin"

# For root users, ensure that /usr/local/sbin, /usr/sbin, and /sbin are in
# the $PATH.  Some means of connection don't add these by default (sshd
comes
# to mind).
if [ "`id -u`" = "0" ]; then
  echo $PATH | grep /usr/local/sbin 1> /dev/null 2> /dev/null
  if [ ! $? = 0 ]; then
    PATH=/usr/local/sbin:/usr/sbin:/sbin:/home/me/ActivePython-2.2/bin:$PATH
  fi
fi
------------------

So anyways can someone please help me with what I put in the top of my
scripts after the , #! ,  so I can run them.
And if there are any linux guru's out there that could check that PATH thing
out and see if its ok, I would be greatfull.

Thanks for your time people.
Hyperion





From hs2l2ATyahoodotcom at yahoo.com  Wed Apr 23 09:22:47 2003
From: hs2l2ATyahoodotcom at yahoo.com (Haran Shivanan)
Date: 23 Apr 2003 13:22:47 GMT
Subject: Python implementation for a web spider \ site mirroring
References: <662cede5.0304220114.54e4dd10@posting.google.com> <S8WdnR2SKryGsTijXTWcrg@speakeasy.net>
Message-ID: <b86436$6hsdv$1@ID-148859.news.dfncis.de>

module "A.M. Kuchling" core dumped with the following output:
> On 22 Apr 2003 02:14:34 -0700, 
> 	Tzahi <tzahij at hotmail.com> wrote:
>> I need to copy complete web sites form the web.
> 
> Try GNU wget (http://www.gnu.org/software/wget/wget.html); not written in
> Python, but quite featureful.
> 
> --amk

wget -k -r http://base-site-to-crawl will do the trick.
-- 
Freedom is a state of mind



From skvamme at as4-2-8.an.g.bonet.se  Wed Apr 30 02:42:07 2003
From: skvamme at as4-2-8.an.g.bonet.se (Sten Kvamme)
Date: Wed, 30 Apr 2003 08:42:07 +0200
Subject: Self modifying systems question
Message-ID: <DCB542A4-7AD6-11D7-93CD-003065DC4D52@as4-2-8.an.g.bonet.se>

Hello,

Self modifying code is a well known technology. Has anyone done this in
Python? What I'm seeking for is examples of software that adds to
itself (methods and variables) in runtime. I'd like an instantiated 
object to
be able to add a method to itself.

Thanks,
Sten Kvamme





From gupt_vive at hotmail.com  Tue Apr 22 07:09:45 2003
From: gupt_vive at hotmail.com (vivek kumar)
Date: Tue, 22 Apr 2003 16:39:45 +0530
Subject: Python implementation for a web spider \ site mirroring
Message-ID: <Law15-F74CaeFDX3K4T0000c2f8@hotmail.com>

I have written a very basic crawler (just for fun :-)). Originally i have 
downloaded it from some site and made some changes according to my need. 
This may/may not be of some use for you. Find it attached in the mail


>From: tzahij at hotmail.com (Tzahi)
>Hi
>
>I need to copy complete web sites form the web.
>
>I think Python is a natural fit for this task, with the intensive HTML
>and HTTP support. Being lazy, I would rather have someone else's
>implementation, so that I do not start from scratch.
>
>Anyone knows of such a free implementation?
>
>Thanks in advance,
>
>Tzahi Matar


_________________________________________________________________
Gizmo freak? See what's the hottest. http://www.msn.co.in/Computing/Gizmos/ 
Read product reviews.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: spider.py
URL: <http://mail.python.org/pipermail/python-list/attachments/20030422/0cfb431b/attachment.ksh>

From phuoc-khanh.nguyen at bull.net  Mon Apr 28 08:25:26 2003
From: phuoc-khanh.nguyen at bull.net (nguyen khanh)
Date: Mon, 28 Apr 2003 14:25:26 +0200
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
Message-ID: <000801c30d81$407b98d0$5718b681@frcls1011>

a=range(10)
b=a[3:7]
b=[3,4,5,6]
Why the design exclude the last element ?
I's because the distinction of "element and list-element" ?
a[3]=3        # element
a[3:4]=[3]   # list-element

Why the designer d'nt use a[3:3]=[3] instead of a[p:p]=[] ?
The result a[p:p]=[] for all p d'nt add new information ,
The design a[n:p] clarify the mind to include the last element p .
and a[p:p]=[p] is very clear and dinctint from a[p]=p .

Thank to advise me your idea about the circus-slice notion .
Faithfull .
A Pythonian construct .
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030428/5a32ed6b/attachment.html>

From user at domain.invalid  Wed Apr 16 06:27:19 2003
From: user at domain.invalid (user at domain.invalid)
Date: Wed, 16 Apr 2003 12:27:19 +0200
Subject: Tutorial - section 8.5 -- User Defined Execptions?
Message-ID: <b7jauh$arj$1@ctb-nnrp2.saix.net>

 From the python documentation tutorial, section 8.5
+++++++++++++++++++++++++++++++++++=
 >>> class MyError:
...     def __init__(self, value):
...         self.value = value
...     def __str__(self):
...         return `self.value`
...
 >>> try:
...     raise MyError(2*2)
... except MyError, e:
...     print 'My exception occurred, value:', e.value
...
My exception occurred, value: 4
 >>> raise MyError, 1
Traceback (innermost last):
   File "<stdin>", line 1
__main__.MyError: 1

+++++++++++++++++++++++++++++++++++=

What do the back ticks in method "__str__" mean?




From maxm at mxm.dk  Sun Apr  6 10:47:31 2003
From: maxm at mxm.dk (Max M)
Date: Sun, 06 Apr 2003 16:47:31 +0200
Subject: List of Numbers
In-Reply-To: <m6009vst7snfmo645pd9h8clbgio2dp0lp@4ax.com>
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com> <3E8F8C10.A7F921CF@engcorp.com> <m6009vst7snfmo645pd9h8clbgio2dp0lp@4ax.com>
Message-ID: <3E903E03.7050203@mxm.dk>

Simon Faulkner wrote:
> Thanks for all the help on this chaps,
> 
> The file has changed and I am now importing 25,000 15 digit numbers!
> 
> I am compiling a report as I go on each number, wether is is duped
> within this batch and then looking it up in a MySQL database.
> 
> It seems to slow down as it goes and I guess that this is the overhead
> of checking each number against the dictionary?

More likely it is due to MySQL

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From oktaysafak at ixir.com  Tue Apr 22 04:50:52 2003
From: oktaysafak at ixir.com (Oktay Safak)
Date: Tue, 22 Apr 2003 11:50:52 +0300
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org> <QxAoa.540832$3D1.296696@sccrnsc01> <a5415304.0304210544.1342b436@posting.google.com> <3EA490F5.30606@removethis.free.fr>
Message-ID: <005c01c308ac$4a02cb00$0200a8c0@b>

Very rude of you, Laotseu ;)

> (oh shit, I'm sure I've just said another stoopidity that will end up in 
> a long cross-post with c.l.smalltlak now)






From Padraig at Linux.ie  Mon Apr  7 12:38:51 2003
From: Padraig at Linux.ie (Padraig at Linux.ie)
Date: Mon, 07 Apr 2003 17:38:51 +0100
Subject: Parsing data
In-Reply-To: <b6s9lu$a9i$1@usenet.otenet.gr>
References: <b6s5lt$5c1$1@usenet.otenet.gr> <b6s71b$9j3$1@news-reader12.wanadoo.fr> <b6s9lu$a9i$1@usenet.otenet.gr>
Message-ID: <3E91A99B.2020305@Linux.ie>

Antonis Kaklis wrote:
>>Was that what you wanted? What you want to do with your data is not quite
> 
> clear
> 
>>from your post...
>>--
>>- Eric Brunel <eric.brunel at pragmadev.com> -
>>PragmaDev : Real Time Software Development Tools -
> 
> http://www.pragmadev.com
> 
> Thanks very much for your response...
> but  now  I have lists like this:
> [2.1614729999999999e-006, 2.183854e-006, 2.1778039999999998e-006,
> 2.1740989999999998e-006, 2.2548290000000002e-006]
> [2.1451629999999998e-006, 2.322683e-006, 2.1120490000000001e-006,
> 2.0925849999999999e-006, 2.0696180000000001e-006]
> [2.05246e-006, 2.5724020000000002e-006, 2.6257250000000002e-006,
> 2.7019989999999999e-006, 2.7715390000000001e-006]
> [1.88422e-006, 1.8648680000000001e-006, 1.8359719999999999e-006,
> 3.1508029999999998e-006, 3.2457659999999998e-006]
> 
> What i want is to process each column. First the values of the first column,
> continue with the second column and so on...,
> in order to plot the values with time step of 0.02sec. (f(t)).
> I don't know if I make myself clear now(I hope so).

This should give you a clue
http://aspn.activestate.com/ASPN/Mail/Message/python-list/831566




From juneaftn at REMOVETHIShanmail.net  Wed Apr 16 15:14:31 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Thu, 17 Apr 2003 04:14:31 +0900
Subject: How many of you are Extreme Programmers?
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com> <mailman.1050504987.840.python-list@python.org> <3E9D8A80.FD76BDE9@engcorp.com> <b7k3kj$4gj$1@news.hananet.net> <3E9DA883.CE0C38B4@engcorp.com>
Message-ID: <b7ka1p$8sv$1@news.hananet.net>

"Peter Hansen" <peter at engcorp.com> wrote in message
news:3E9DA883.CE0C38B4 at engcorp.com...
> Changjune Kim wrote:
> >
> > Check out the Standish Group CHAOS report.
> >
> > http://www.softwaremag.com/archive/2001feb/CollaborativeMgt.html
(refers to
> > 2001 report)
> > http://www.pm2go.com/sample_research/chaos_1994_1.php
>
> The first link redirects to http://www.softwaremag.com/magazine.css,
> which is not found (404).
>
> -Peter

I can read the page without any problem, but if you can't read it,
Internet Archive will answer you:
http://web.archive.org/web/*/http://www.softwaremag.com/archive/2001feb/Col
laborativeMgt.html




From dale at riverhall.NOTHANKS.co.uk  Wed Apr 16 05:13:17 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 16 Apr 2003 10:13:17 +0100
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 14)
References: <v9lcekr9vkl734@corp.supernews.com>
Message-ID: <ui7q9v4k6ufcn7ssvgjm05q43cm8mo3f3l@4ax.com>

I love these little digests.

I don't get time to read this group thoroughly and they often point me
at things I've missed.

Thanks and keep up the good work.

--
Dale Strickland-Clark
Riverhall Systems Ltd



From brian at sweetapp.com  Fri Apr 25 22:39:30 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Fri, 25 Apr 2003 19:39:30 -0700
Subject: Forgetting "()" when calling methods
In-Reply-To: <3EA9B64D.3208D730@alcyone.com>
Message-ID: <004301c30b9d$1006fbd0$21795418@dell1700>

Erik Max Francis wrote:
> One thing that PyChecker, at least, is more likely to catch is "code
has
> no effect" statements like
> 
> 	sys.exit

How do you know that the code has no effect?

> Without the parentheses, this clearly can't do anything useful, 

class Sys:
    def __getattr__(self, attr):
        if attr == 'exit'
            print 'cya'

sys = Sys()





From exarkun at intarweb.us  Fri Apr  4 15:45:14 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 4 Apr 2003 15:45:14 -0500
Subject: [Newbie] Opinions on GUI and distribution
In-Reply-To: <17nr8vgmif2df7dr8dd5saut781b1fr044@4ax.com>
References: <17nr8vgmif2df7dr8dd5saut781b1fr044@4ax.com>
Message-ID: <20030404204514.GA11174@meson.dyndns.org>

On Fri, Apr 04, 2003 at 09:32:20PM +0200, Jane Doe wrote:
> Hi,
> 
> 	I'd like to check if Python is a viable alternative to
> VisualBasic to develop stand-alone apps for Windows, and I did read
> the archives and stuff on the Web about those issues but... I'd like
> some feedback from people experienced with this kind of developement.
> 
> 1. How easy is it to build and deploy Python apps on hosts that either
> have nothing pre-installed (ei. Python itself, the GUI framework, and
> my scripts), or need some updates (ie. my scripts in a newer version)?

  Easy.  Distutils, py2exe, McMillian Installer, and several others give you
a wide variety of installation approaches.

> 
> 2. Are the GUI frameworks currently available (win32all as part of
> ActivePython, wxWindows, PyGTK, PythonQT, etc.) really good enough to
> build professional applications?
> 

  Some of them are, certainly.  I think that this is an area where the
differences tend to come down mainly to personal taste.

> 3. What about performance as compared to VB, VC++ or PowerBasic?

  Hard to answer this one.  Python is "fast enough" for almost all
programming tasks.  When it's not, extension modules which are
native-compiled can be written and used.  Bottom line, I doubt you'll find
an application for which Python's runtime performance disqualifies it as a
possible implementation language.

  Jp

-- 
http://catandgirl.com/view.cgi?44
-- 
 up 15 days, 16:00, 9 users, load average: 0.00, 0.02, 0.07




From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Mon Apr 21 17:27:12 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Mon, 21 Apr 2003 23:27:12 +0200
Subject: regular expressions
In-Reply-To: <b81lq4$17n8$1@galaxy.us.dell.com>
References: <b81lq4$17n8$1@galaxy.us.dell.com>
Message-ID: <3ea46230$0$49114$e4fe514c@news.xs4all.nl>

junk at aol.com wrote:
> Can anybody recommend some good resources for learning regular
> expressions.  Primarly web sites.

   http://www.python.org/doc/current/lib/module-re.html

together with

   http://www.google.com

should help you a lot.
For instance,

   http://www.amk.ca/python/howto/regex/

turned up as the first result when googling for "python regular expressions".

--Irmen




From max at alcyone.com  Mon Apr 14 05:09:37 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 02:09:37 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <mailman.1050156199.10180.python-list@python.org> <Q0Wla.4427$LB6.127243@news1.tin.it>
Message-ID: <3E9A7AD1.34492946@alcyone.com>

Alex Martelli wrote:

> There's supposed to be a good reason for this to print [rep],
> rather than [str] as common sense would suggest, but I keep
> forgetting it (so it's probably nor very compelling;-).

I have to agree.  It makes more sense to me that the str of a collection
prints the str of its elements (and on down), and that the repr of a
collection prints the repr of its elements.  That isn't the way things
work, and I accept that and it's not a huge issue for me, but I
personally consider it one of Python's minor quirks.  I can see the
argument (which was put forth in this thread) that printing the str
might be misleading in some cases, but that doesn't seem to me a very
compelling argument; since both the str and repr are user-selectable, it
should be easy for an inquisitive soul who's confused about what he's
seeing to figure out what's going on with little or no trouble.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Bachelors have consciences, married men have wives.
\__/ H.L. Mencken
    WebVal / http://www.alcyone.com/pyos/webval/
 URL scanner, maintainer, and validator in Python.



From staschuk at telusplanet.net  Thu Apr 10 11:41:33 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 10 Apr 2003 09:41:33 -0600
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <090420031837238086%camel@oasis.com>; from camel@oasis.com on Wed, Apr 09, 2003 at 06:37:23PM -0400
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org> <BAB8B22D.2022%dogen@mindspring.com> <20030408194700.A1973@tibia.amotlpaa.bogus> <mailman.1049922136.32568.python-list@python.org> <090420031837238086%camel@oasis.com>
Message-ID: <20030410094133.B196@tibia.amotlpaa.bogus>

Quoth dromedary:
  [...]
> And if I were less a dunderhead, I'd see how to generalize to
> combinations above n=2. Times like this I wish I'd studied math or CS.

You've made a good start.  If I understand correctly, your
variable 'dict' (which, btw, could use a different name -- 'dict'
is a built-in) holds a map from 'foo' to a list of words starting
or ending with 'oof'.  Thus, you only try to find two-word
palindromes using pairs of words such that the second contains the
first reversed, which much reduces the search space.  Quite a win.

The code skips combinations in which the first word occurs
reversed in the middle of the second; this is not important for
two-word palindromes, of course.  For larger palindromes one might
imagine making a dict which maps from strings to words in which
those strings occur anywhere as a substring... at least you'd be
able to say your code has a big dict.

But there's something better.  If the palindrome is to start with
'foo', we already know where the word ending with 'oof' has to go:
at the end of the phrase.  So we can construct a phrase with a gap:
    foo ... aloof
This is partway towards a palindrome: the leading 'foo' matches
the trailing 'oof'.  We also know that it needs a word after
'foo', a word which starts with 'la', to match the unmatched 'al'
of 'aloof'.  Supposing the word list contains 'latent', we can
build
    foo latent ... aloof .
Now we need 'tnet' on the right.  Note that we want to be able to
find the word 'net' here; that is, we have to look not just for
words ending with 'tnet', but also for the complete words 'net',
'et', and 't'.  Such words would partially match what we're trying
to match, and we'd continue trying to add words to the same side
as before.  With 'net', in fact,
    foo latent net aloof
is a palindrome, because the 't' of 'tnet' which 'net' doesn't
match can match itself.  The search can continue from
    foo latent ... net aloof ,
seeking words ending with 't', to find such things as
    foo latent endorse tar rates rod net net aloof .

This approach is illustrated in detail by the following trace from
a slightly hacked version of the aforementioned code.  It's doing
a depth-first search looking for palindromes of up to seven words,
using words from ['is', 'it', 'no', 'one', 'open', 'position'].
Parentheses indicate letters whose mirror image has not yet been
put in the palindrome being built.

  "(is) ... "
  "(it) ... "
  "(no) ... "
   -  "no ... (positi)on"
   -   -  "no it ... (posi)tion"
   -   -   -  "no it is ... (po)sition"
   -   -   -   -  "no it is op(en) ... position"
   -   -   -   -   -  "no it is open ... (o)ne position"
                              palindrome: no it is open one position
   -   -   -   -   -   -  "no it is open o(ne) ... one position"
   -   -   -   -   -   -  "no it is open o(pen) ... one position"
  "(one) ... "
   -  "on(e) ... no"                              palindrome: one no
   -   -  "one ... (on)e no"
   -   -   -  "one no ... one no"          palindrome: one no one no
   -   -   -   -  "one no ... (one) one no"
   -   -   -   -  "one no ... (open) one no"
   -   -   -   -  "one no ... (position) one no"
   -   -   -   -   -  "one no no ... (positi)on one no"
   -   -   -   -   -   -  "one no no it ... (posi)tion one no"
   -   -   -   -  "one no ... (no) one no"
   -   -   -   -   -  "one no on(e) ... no one no"
                                    palindrome: one no one no one no
   -   -   -   -   -   -  "one no one ... (on)e no one no"
   -   -   -   -  "one no ... (is) one no"
   -   -   -   -  "one no ... (it) one no"
  "(open) ... "
  "(position) ... "

Note that this search strategy has no need for the combinations()
function; the right order of words is found by construction.

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From exarkun at intarweb.us  Fri Apr 25 13:08:45 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 25 Apr 2003 13:08:45 -0400
Subject: embedding python2.2 in C++/Qt
In-Reply-To: <23049.1051269931@www49.gmx.net>
References: <23049.1051269931@www49.gmx.net>
Message-ID: <20030425170845.GA883@meson.dyndns.org>

On Fri, Apr 25, 2003 at 01:25:31PM +0200, jdominion at gmx.de wrote:
> hi!
> 
> Who has up-to-date experience in embedding python into C++?  All examples i
> found in the internet or in the (not recommendable) book of Himstedt/M?tzel
> are for python version 1.5.

  Did you read the adequate reference at http://www.python.org/doc/?

> 
> I include the Python.h and cannot perform the command Py_Initialize(). what
> is the problem?

  This isn't a good way to ask questions.  Please read this:

  http://www.catb.org/~esr/faqs/smart-questions.html

> I know, this is more a C++-question. but  everybody promise that it is "very
> easy" to embed/extend python.
> 

  It is very easy to extend Python.  Embedding takes a little more work, and
gives you slightly less useful results.  For a more detailed comparison,

  http://www.twistedmatrix.com/users/glyph/rant/extendit.html

  Jp 

-- 
#!/bin/bash
( LIST=(~/.sigs/*.sig)
  cat ${LIST[$(($RANDOM % ${#LIST[*]}))]}
  echo -- $'\n' `uptime | sed -e 's/.*m//'` ) > ~/.signature
-- 
 up 36 days, 13:03, 6 users, load average: 0.47, 0.37, 0.35




From missive at frontiernet.net  Wed Apr 30 21:03:25 2003
From: missive at frontiernet.net (Lee Harr)
Date: Thu, 01 May 2003 01:03:25 GMT
Subject: plain object()
Message-ID: <slrnbb0el9.2iks.missive@localhost.my.domain>


Hi;

Can someone explain this?


>>> a = object()
>>> a
<object object at 0x8149bc0>
>>> a.x = 3
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
AttributeError: 'object' object has no attribute 'x'
>>> class B(object):
...   pass
...
>>> b = B()
>>> b
<__main__.B object at 0x81194cc>
>>> b.x = 3
>>> b.x
3
>>>





From a.schmolck at gmx.net  Sun Apr  6 22:46:01 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 07 Apr 2003 03:46:01 +0100
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com>
Message-ID: <yfsn0j3kot2.fsf@black132.ex.ac.uk>

llothar at web.de (Lothar Scholz) writes:

> Then you have to present the same old and ugly user interface with a
> mini  buffer as input line and so on. I don't see any reason at the
> moment for cloning EMACS. You should ask people why they like emacs as
> an editor.
> There are only a handful of features that are normally the reason to
> keep
> this as your primary editor. For me its the window handling C-X 1, the
> way to prefix commands with a repeat command, the macro recording and
> a the autoindentation with TAB.
> 
> To make clear, i'm using XEMACS 8 hours a day for years now. But it's
> absolutely losing ground compared to modern editors. And i don't think
> that
> extensibility in the emacs sense is really necessary today or a plus,
> because
> people want out of the box solutions that are specialized to their
> domain.

The plausibility of your suggestion that the majority of emacs's long time
user base is formed by mindless drones depresses me.

alex



From roy at panix.com  Sun Apr 13 14:38:28 2003
From: roy at panix.com (Roy Smith)
Date: Sun, 13 Apr 2003 14:38:28 -0400
Subject: How to do per-line profiling?
References: <roy-7A3D8E.11045713042003@reader1.panix.com> <zegma.8647$LB6.234768@news1.tin.it>
Message-ID: <roy-FD4515.14382813042003@reader1.panix.com>

Alex Martelli <aleax at aleax.it> wrote:
> Yes, package hotshot (part of the standard Python library) is just 
> such a tool.  Unfortunately, it's somewhat under-documented!

Thanks for the pointer -- at least it got me looking in the right place.  
Unfortunately, when I try to follow the examples from

http://www.python.org/dev/doc/devel//////lib/hotshot-example.html

I get:

Traceback (most recent call last):
  File "./rulex-diff.py", line 604, in ?
    stats = hotshot.stats.load ('profile-stats')
  File "/usr/local/lib/python2.2/hotshot/stats.py", line 12, in load
    return StatsLoader(filename).load()
  File "/usr/local/lib/python2.2/hotshot/stats.py", line 51, in load
    assert not self._stack
AssertionError

Here's what I tried:

    import hotshot, hotshot.stats
    import pstats
    app = rulexDiffApp()

    prof = hotshot.Profile ('profile-stats', 1, 1)
    prof.run ('app.main()')
    prof.close()

    stats = hotshot.stats.load ('profile-stats')
    stats.strip_dirs()
    stats.sort_stats('time', 'calls')
    stats.print_stats(20)

This is Python-2.2.2 on a RedHat 6.2 box.



From bdesth at removethis.free.fr  Fri Apr 18 19:33:12 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Fri, 18 Apr 2003 23:33:12 +0000
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net> <s1Vna.126364$It5.107742@news2.central.cox.net>
Message-ID: <3EA08B38.3050507@removethis.free.fr>

Steve Holden wrote:
> "Bill Martin" <wcmartin at vnet.net> wrote ...
> 
>>I'm wondering about the value of allowing a class definition like this:
>>
>>             class C:
>>               pass
>>
>>Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or
>>some such. If I try to print a.x2 or b.x1, I get an exception message
>>basically saying those member variables don't exist. It seems to me this
>>defeats one of the basic ideas behind OOP, that is to assign members to
>>the class definition, thus ensuring that all instantiations of a class
>>have the same member variables and methods. I know there must be a
>>reason that python allows this kind of thing that I described, but I
>>can't figure out what that reason could be. To me, it looks like a very
>>dangerous thing to do. Can someone help me understand why we'd like to
>>do it? Thanks,
>>
> 
> 
> Well, if you think it's dangerous you probably *don't* want to do it, but
> it's the easiest way to set up a data structure with named components - you
> are just taking advantage of the fact that each instance has its own
> namespace. It's a firly standard idiom, though it isn't required.
> 
> Now, other object-oriented languages can go to great pains to make sure that
> an instance's attributes can only be accessed in very controlled ways, such
> as using accessor methods (getthis(), setthat(), etc.). Python, on the other
> hand, pays you the compliment of assuming you know what you are doing, and
> "gives you enough rope to shoot yourself in the foot"!
> 
> If you choose to abuse that freedom, you end up with a limp. [Please, nobody
> ask "a limp what?"].
> 

A limp what ?

[sorry, could not resist]
Laotseu




From danra at walla.co.il  Mon Apr  7 14:56:42 2003
From: danra at walla.co.il (Danra)
Date: 7 Apr 2003 11:56:42 -0700
Subject: super question
Message-ID: <fd1c664e.0304071056.31a0d377@posting.google.com>

Hi,

I must be doing something wrong, but it seems that the typical use for
the 'super' function, suggested in its help and in the python library
refernce, does not work for me.

Here's my experience with 'super':

>>> class b:
	pass

>>> class c(b):
	pass

>>> super(c)
Traceback (most recent call last):
  File "<pyshell#6>", line 1, in ?
    super(c)
TypeError: super() argument 1 must be type, not classobj

>>> super(c,c())
Traceback (most recent call last):
  File "<pyshell#7>", line 1, in ?
    super(c,c())
TypeError: super() argument 1 must be type, not classobj

What am I doing wrong?

P.S. pardon the pun in the question's subject :)



From ryomanga at aol.com  Sun Apr  6 07:42:08 2003
From: ryomanga at aol.com (Ryomanga)
Date: 06 Apr 2003 11:42:08 GMT
Subject: pb py32exe and PIL
Message-ID: <20030406074208.01628.00000159@mb-cb.aol.com>

As in the title I got a pb with py2exe and the PIL library.
When I use the exe i've got the following error message :

  File "Image.pyc", line 960, in open
IOError: cannot identify image file

It seems it can't read my pictures. I've try to change their format and
directory but nothing changes. Is there something special to do with PIL like
with PMW??

Thanks for your help
     Guilh?m




From patrickdlogan at attbi.com  Thu Apr 24 01:13:52 2003
From: patrickdlogan at attbi.com (Patrick Logan)
Date: 23 Apr 2003 22:13:52 -0700
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050569128.32513.python-list@python.org> <27cc4d4a.0304222142.199ba7c9@posting.google.com> <mailman.1051105417.17659.python-list@python.org>
Message-ID: <27cc4d4a.0304232113.57bfc2e7@posting.google.com>

> Can you give me some pointers to timings nd maybe comparisons?
> I'm not familiar with those langauge's web sites.

Not exactly an apples-to-apples with your numbers, but it gives you an idea...

http://www.erlang.org/ml-archive/erlang-questions/200212/msg00058.html

-Patrick



From borco at go.ro  Mon Apr 28 12:30:14 2003
From: borco at go.ro (Ionutz Borcoman)
Date: 28 Apr 2003 09:30:14 -0700
Subject: simple extension modules with GNU's C#
Message-ID: <80dac0f4.0304280830.756ed8c@posting.google.com>

Hi,

What's the simplest way to make an extension in C# ? 

For the beginig it would be enough for me to make some simple
functions that can be called from python and are writen in C#, so they
don't need to be recompiled to run the project on Linux or Win32.

Tia,

Ionutz



From ianb at colorstudy.com  Thu Apr 17 15:54:36 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 17 Apr 2003 14:54:36 -0500
Subject: Teaching : Python, Scheme, Java...
In-Reply-To: <roy-8E37D1.09554617042003@malibu.unice.fr>
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
	 <3E9C9A91.8090209@removethis.free.fr>
	 <m27k9vfcz3.fsf@mycroft.actrix.gen.nz>
	 <3E9DE024.3020807@removethis.free.fr>
	 <roy-8E37D1.09554617042003@malibu.unice.fr>
Message-ID: <1050609276.718.303.camel@lothlorien>

On Thu, 2003-04-17 at 02:55, Jean-Paul Roy wrote:
> - Python is bad : tail recursion is not iterative (also astonished at 
> that, I can understand with an interpreter, but compiler ?)

Some would argue that tail recursion is not good for teaching and
understanding, because you loose the call stack.  In the likely event of
a traceback, it can be very confusing if you don't fully indicate how
you got there, or if points upon the path were lost in an optimization.

  Ian






From gansevle at ewi.utwente.nl  Mon Apr  7 15:59:35 2003
From: gansevle at ewi.utwente.nl (Fred Gansevles)
Date: Mon, 07 Apr 2003 21:59:35 +0200
Subject: Wrapper for _winreg
Message-ID: <b6slau$c0l$1@netlx020.civ.utwente.nl>

Hi,

I've written a wrapper for _winreg which, IMHO, makes registry access in
Windows trivial.
I looked at other approaches and wrote several of my own (only for internal
use) but this time I think this approach is usable for more than just my
own tinkering.

The main goal is to simplify both the substring notation for registry keys
an the access to registry data.

For example, to list the nameservers on a WindowsXP machine, the following
code snippet can be used:

        def NameServers():
            from reg import Keys, Value, HKLM
            Services = ["SYSTEM", "CurrentControlSet", "Services"]
            Interfaces = [Services, "Tcpip", "Parameters", "Interfaces"]
            return [Value(HKLM, [Interfaces, interface], "NameServer")
                        for interface in Keys(HKLM, Interfaces)
                   ]

As you can see, you don't need to open a number of keys, this is done in the
module for you so 'one-shot' access to a part of the registry is simple.

Another example: find the proxyserver in the registry (from urllib.py)

            Settings = Key(HKCU, ["Software", "Microsoft", "Windows",
                                   "CurrentVersion", "Internet Settings"]
            if Value(Settings, name = "ProxyEnable"):
                proxyServer = str(Value(Settings, name = "ProxyServer"))

Here it is more efficient to pre-open a key since more than one value in the
key is referenced.


The complete module is located at 

        http://wwwhome.cs.utwente.nl/~gansevle/python/reg.py

The module is 'mildly' documented so it should be usable right away.

If anyone finds this module usefull, please drop me a line at
gansevle at ewi.utwente.nl







From duncan at NOSPAMrcp.co.uk  Tue Apr 15 10:56:20 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 15 Apr 2003 14:56:20 +0000 (UTC)
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1> <3e9c1413$1@news.012.net.il>
Message-ID: <Xns935EA1AE09A15duncanrcpcouk@127.0.0.1>

"Avner Ben" <avner at skilldesign.com> wrote in
news:3e9c1413$1 at news.012.net.il: 

> I am searching for one or more members of a list of names (that
> may get 
> near 1000 pieces in a normal application and may average some 40
> characters) in a string (that may average 100 characters). For each
> name found in the string, I do something to the string, using the
> original name (in its proper case). At present, I am looking for the
> exact name as stated, with agreeable performance. However, this makes
> me miss many cases where the author of the string did not care for the
> casing (which happens frequently). I have little problem with either
> lowering or uppering the entire list once, because it is not expected
> to change often and when that happens once in a while, a subscription
> mechanism can take care of that. However, this requires holding two
> lists (I still need the original names as are) and lowering or
> uppering each string for comparison (and keep its original version for
> restructuring).

I take it from your description that the names are not in any way
delimited by spaces or punctuation in the strings. If they were, then
splitting the strings on the delimiters might lead to a nice solution. 

My initial gut feel would be to approach this by building a dictionary
mapping the lowercase version of each name to its original casing. You
only talk of 1000 names of about 40 characters, so keeping two copies
isn't going to be a concern. 

Then you have to find the best way to locate which of the lowercase
names occur in the lowercase version of the string. Three ways to try
would be to iterate over all the names using str.find to check whether
it occurs in the string, or to try concatenating all the names into one
humungous regular expression, or to try something based on your C++
method and lookup substrings in a dictionary. 

The last of these would be the most complex, but might be pretty fast.
What I'm thinking is that you could split your 100 character string into
each 5 character substring in turn, and look them up in a dictionary of
names keyed by the first 5 characters. That way you do perhaps 95
dictionary lookups instead of 1000 string find operations. Of course you
would have to handle very short names as a special case (i.e. find or
regex).

-- 
Duncan Booth                                            
duncan at rcp.co.uk int month(char
*p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3" 
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure? 



From davecook at nowhere.net  Wed Apr 16 01:42:16 2003
From: davecook at nowhere.net (David M. Cook)
Date: Wed, 16 Apr 2003 05:42:16 GMT
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <Y26na.108494$It5.3942@news2.central.cox.net>

In article <12240-3E9C5F83-187 at storefull-2371.public.lawson.webtv.net>, W S
wrote:

> Hello,
> 
> I find the material, as presented, in the book Programming Python
> frustrating and extremely difficult to understand.
> 
> Are the nutshell and cookbook O'Reilly Python books easier to read?
> 
> I'm fairly new to programming but trying to learn. I have read Learning
> Python, Core Python, Python Bible and a bunch of perl books

Well, if you have those, you'll probably get more out of the cookbook than
anything else; it's one of the most useful Python books I've seen.

Dave Cook



From duncan at NOSPAMrcp.co.uk  Mon Apr 28 10:46:31 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Mon, 28 Apr 2003 14:46:31 +0000 (UTC)
Subject: Recursive Algorithm getting stuck (because of Python Limits?)
References: <236da4ed.0304280616.349e8bdd@posting.google.com>
Message-ID: <Xns936B9F5E6265Fduncanrcpcouk@127.0.0.1>

gildemere at hotmail.com (Marc Floessel) wrote in 
news:236da4ed.0304280616.349e8bdd at posting.google.com:

> This works. However increasing the Depth or Breadth to 8 seems to
> break Python. Python rises to about 100 MB memory consumption (as
> indicated by the Windows Task Manager) and then seems to drop and rise
> randomly. The algorithm never finishes. Am I breaking the stack?

7 wide and deep you are trying to create 137,257 nodes. 8 wide and deep you 
are trying to create 2,396,745 nodes so it should take 17 times as much 
memory. It will take much more than 17 times longer if you are using Python 
2.2.x as Microsoft's memory allocation routines grind to a halt when you 
have that many objects. Also you are probably running out of memory. When 
the memory figure in task manager starts fluctuating it is because youre 
process is thrashing. My machine started thrashing like that around 400Mb 
used. 

Adding a __slots__ member to the Node class removes the need to create a 
dictionary for each instance. On Python 2.3 this reduces the total memory 
requirement for 8 deep to about 170Mb, and the runtime to a couple of 
minutes.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From junk at aol.com  Fri Apr 18 11:04:40 2003
From: junk at aol.com (junk at aol.com)
Date: Fri, 18 Apr 2003 10:04:40 -0500
Subject: How can I read event logs with python?
Message-ID: <b7p48d$15pt$1@galaxy.us.dell.com>

I just want python to look through my event logs and print warnings and
errors.  I know that win32evtlog should allow me to but I'm just not
able to get it to work.



From oblivious at web.de  Thu Apr 10 05:57:04 2003
From: oblivious at web.de (Dominic)
Date: Thu, 10 Apr 2003 11:57:04 +0200
Subject: Macros in Python?
In-Reply-To: <l_0la.1989$d8.141@nwrdny02.gnilink.net>
References: <b71qde$q71$07$1@news.t-online.com> <l_0la.1989$d8.141@nwrdny02.gnilink.net>
Message-ID: <b73f31$r95$02$1@news.t-online.com>

Carl Banks wrote:
> Dominic wrote:
> 
>>
>>(defmacro with-root-privs (() &body body)
>>  (let ((oldidsym (gensym)))
>>    `(let ((,oldidsym (geteuid)))
>>       (seteuid 0)
>>       (unwind-protect
>>          (progn , at body)
>>        (seteuid ,oldidsym)))))
>>
>>Is there an easy way to produce similar code
>>in Python which gurantees that certain things
>>happen around some code? (Maybe by using
>>generators?)
>>
>>The normal approach would be to enclose everything
>>with try: finally: but this would scatter the
>>code all over.
> 
> 
> 
> It doesn't have to.  First, define:
> 
>     def get_root_privledges():
>         euid = os.geteuid()
>         os.seteuid(0)
>         return euid
> 
>     def release_root_privledges(euid):
>         os.seteuid(euid)
> 
> 
> Then you can do:
> 
>     old_euid = get_root_privledges()
>     try:
>         if x in user:
>             active.append(x)
>             ...
>     finally:
>         relase_root_privledges(old_euid)
> 
> 
> I hope this isn't what you consider "scattering the code all over."
> There is a certain niceness to the with macros (the ones that don't
> rebind any symbols, anyways), but try ... finally does the job just
> fine.
This solution is very straightforward but I have been curious wether
there is a "nicer" way to do the job. Alexander and Jp's postings
showed me that there are two very different approaches which could
also be used. Certainly all 3 solutions have their pros and cons
I'll consider all three.
So, thank you very much for all your answers :)

Ciao,
  Dominic

P.S. It would be nice to have a database for python usage patterns.


> 
> 




From tylere at hotpop.com  Sun Apr 13 16:06:13 2003
From: tylere at hotpop.com (Tyler Eaves)
Date: Sun, 13 Apr 2003 16:06:13 -0400
Subject: Help with CGI
In-Reply-To: <pan.2003.04.13.19.41.42.420658@soul.net>
References: <pan.2003.04.13.19.41.42.420658@soul.net>
Message-ID: <200304131606.13498.tylere@hotpop.com>

On Sunday 13 April 2003 15:41, lost wrote:

> form = cgi.FieldStorage

There's your problem. 
That line should read: 

form = cg1.FieldStorage()

The () is the difference between what you want, and actually making a copy of 
the class.




From donn at u.washington.edu  Wed Apr 16 18:39:56 2003
From: donn at u.washington.edu (Donn Cave)
Date: 16 Apr 2003 22:39:56 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <mailman.1050525028.7854.python-list@python.org>
Message-ID: <b7km3s$238m$1@nntp6.u.washington.edu>

Quoth Tim Peters <tim.one at comcast.net>:
| [Donn Cave]
|> But the idea that repr should be the function that conveys that
|> "whole truth" is suspiciously consistent with the marshalling idea.
|
| Indeed, the marshal module uses repr() to encode floats in code objects.
| The idea behind __repr__ seems prettly clear in the docs:
|
|     http://www.python.org/doc/current/ref/customization.html
|
|     __repr__(self)
|     Called by the repr() built-in function and by string conversions
|     (reverse quotes) to compute the ``official'' string representation of
|     an object.  If at all possible, this should look like a valid Python
|     expression that could be used to recreate an object with the same
|     value (given an appropriate environment).  If this is not possible, a
|     string of the form "<...some useful description...>" should be returned.
|     The return value must be a string object.   This is typically used for
|     debugging, so it is important that the representation is information-
|     rich and unambiguous.

It's clear that the docs promote this idea, but they also hint at
its half-baked quality - repr() evals to the original value, unless
it doesn't.  I don't think the idea is clear at all.

Taken together with the __str__ documentation, the original poster saw
this stuff and did not find it clear.  He learned that repr is ``official'',
vs. the ``informal'' str.  This is not explained.  The requirements for
repr conflict, and the primary one is conceded to perhaps not be practicable.
The requirements for str are on the other hand silent on its intended
application.  The only kind of clarity this documentation possesses is
when it lets you read into it what you already thought.  Not the fault
of the documentation, though, but of the unsound ideas behind it.

	Donn Cave, donn at u.washington.edu



From hupp at cs.wisc.edu  Fri Apr  4 17:42:20 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Fri, 4 Apr 2003 16:42:20 -0600
Subject: Invalid format character in string
In-Reply-To: <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com> <mailman.1049493668.26508.python-list@python.org> <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>
Message-ID: <20030404224220.GA8439@upl.cs.wisc.edu>

On Sat, Apr 05, 2003 at 12:17:52AM +0200, Nikolai Kirsebom wrote:
> Hi,
> The problems arise where the HTML editor used (to produce the file in
> the first place) has put various expressions of %; and %>.  These are
> HTML codings the creator of the HTML file does not really know about.
> As an example, I've just told them (the users creating the HTML file)
> that they should write the text "%(name)s" where they want the name
> generated and "%(age)s" where they want the age generated.  Then my
> script is run for a set of person objects producing a set of HTML
> files with the correct names and corresponding ages. 

You could apply this regexp to the string before you do the insertion:

>> re.sub('%(?!\(\w+\)s)','%%','%foo %(bar)s')
<< '%%foo %(bar)s'

The '(?!...)' is a negative lookahead assertion.

-Adam




From jack at performancedrivers.com  Sat Apr 12 02:47:19 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Sat, 12 Apr 2003 02:47:19 -0400
Subject: Singletons
In-Reply-To: <roy-20E369.23282211042003@reader1.panix.com>; from roy@panix.com on Fri, Apr 11, 2003 at 11:28:23PM -0400
References: <roy-20E369.23282211042003@reader1.panix.com>
Message-ID: <20030412024719.F1039@localhost.localdomain>

On Fri, Apr 11, 2003 at 11:28:23PM -0400, Roy Smith wrote:
> Hunting around for ideas on implementing the singleton pattern in 
> Python, I found
> 

If you are using python 2.2 or above writing a singleton is much easier

class World(object):
  real_world = None
  def __new__(cls, *args, **opts):
    if (World.real_world is None):
      World.real_world = object.__new__(cls)
      World.first_init(World.real_world, *args, **opts)
    return World.real_world

  def first_init(self, *args, **opts):
    """this will only be called the first time"""
    # do stuff
    return

-jack




From jkraska at san.rr.com  Sun Apr 13 13:26:31 2003
From: jkraska at san.rr.com (Courageous)
Date: Sun, 13 Apr 2003 17:26:31 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <bJ2ma.5447$883.4511@nwrdny02.gnilink.net> <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net>
Message-ID: <5d7j9vch642sjfipee9ul333iafkfoimgb@4ax.com>

>But is that /guaranteed/ (as it is with STL), or does it just happen to be 
>that way with the current implementation (however unlikely it is to 
>change)?

It's amortized. It's a highly tuned hash table. Python doesn't have
implementation _specs_ (requirements) in the nature of STL.

C//




From gregory.p.green at boeing.com  Mon Apr  7 17:16:10 2003
From: gregory.p.green at boeing.com (Green, Gregory P)
Date: Mon, 7 Apr 2003 14:16:10 -0700
Subject: Problem setting COM property using win32com
Message-ID: <E1B4F80EB080194BBAD3092392F511B3DADEF2@XCH-NW-13.nw.nos.boeing.com>

I missed the last line in your example code:

so if I do (using the EnsureModule line):

prmidattrs = MyModule.ISmRecordList()

I get:
Traceback (most recent call last):   
	File "example.py", line 43, in ?     
		prmidattrs = MyModule.ISmRecordList()   
	File "C:\PYTHON22\lib\site-packages\win32com\client\__init__.py", line 310, in __init__     
		oobj = pythoncom.new(self.CLSID) pywintypes.com_error: (-2147221164, 'Class not registered', None, None) 
pywintypes.com_error: (-2147221164, 'Class not registered', None, None) 

If I use 
prmidattrs = MyModule.SmRecordList()

I get:
Traceback (most recent call last):   
	File "example.py", line 43, in ?     
		prmidattrs = MyModule.SmRecordList()   
	File "win32com\gen_py\C85E7012-8B4F-11D1-8E20-00A02498EA3Cx0x1x0.py", line 1217, in __init__ 
TypeError: 'str' object is not callable 

So either way, that isn't working either. I don't know enough about this stuff to figure it out.

-----Original Message-----
From: Paul Prescod [mailto:paul at prescod.net]
Sent: Monday, April 07, 2003 1:34 PM
To: Green, Gregory P
Cc: python-list at python.org
Subject: Re: Problem setting COM property using win32com


I don't know why that's happening, but I would suggest you take more 
control of the instantiation by doing something like this:

from win32com.client.gencache import EnsureModule
MyModule = EnsureModule('{3BBEE982-742B-11D2-BD0A-00A024C36143}', 0, 1, 0)

reclist = MyModule.SmRecList()

In my recent COM project, I used the "dispatch" mechanism and pythoncom 
was smart enough to pick up my gen_py'd module. But the mechanism above 
is fine too. In fact, it is probably better by virtue of being more 
explicit...

Green, Gregory P wrote:
> I don't think the gen_py class is being used:
> 
> print obj: <COMObject SmRecList.SmRecordList>
> 
> print obj.__class__: win32com.client.CDispatch
> 
> print dir(obj): ['_ApplyTypes_', '_LazyAddAttr_', '_NewEnum', '_Release_', '__AttrToID__', '__LazyMap__', '__call__', '__cmp__', '__doc__', '__getattr__', '__getitem__', '__init__', '__int__', '__len__', '__module__', '__nonzero__', '__repr__', '__setattr__', '__setitem__', '__str__', '_builtMethods_', '_enum_', '_find_dispatch_type_', '_get_good_object_', '_get_good_single_object_', '_lazydata_', '_make_method_', '_mapCachedItems_', '_oleobj_', '_olerepr_', '_print_details_', '_proc_', '_unicode_to_string_', '_username_', '_wrap_dispatch_']






From aleaxit at yahoo.com  Wed Apr  9 10:44:29 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: 9 Apr 2003 07:44:29 -0700
Subject: Nutshell on Safari (was Re: OT: Re: Prime number)
References: <pfZea.36203$7y3.1184215@twister1.libero.it> <C2EC1B511DA9194C.FE8C0AF04D64344A.DE1987687F0858E8@lp.airnews.net> <joIga.12$Xf4.3@twister.nyc.rr.com> <mailman.1048795126.3028.python-list@python.org> <_iJga.16$Xf4.13@twister.nyc.rr.com> <3E836408.72F35838@engcorp.com> <3QWga.32114$i26.828378@news2.tin.it> <bMZga.1757$yY.1013003@twister.nyc.rr.com> <W3_ga.33354$i26.869082@news2.tin.it> <3E84CCB9.3080608@NOmyrealCAPSbox.com> <Z7lha.40309$i26.1110464@news2.tin.it>
Message-ID: <f39543cf.0304090644.4b337940@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<Z7lha.40309$i26.1110464 at news2.tin.it>...
   ...
> > do you know when O' Reilly
> > will make the book available through Safari, their online book service
> 
> I don't know, but I just asked my O'Reilly editor and perhaps she'll
> be able to give me some estimate for that time.

I hear that, according to the schedule, Python in a Nutshell should
be available on Safari within April.  Hope this helps...!


Alex



From jimmy at retzlaff.com  Thu Apr  3 12:28:39 2003
From: jimmy at retzlaff.com (Jimmy Retzlaff)
Date: Thu, 3 Apr 2003 09:28:39 -0800
Subject: Cygwin Python regression testers wanted
Message-ID: <A7B5C22F20EE3C4E95744E2F03F1887F0F1D1B@MAYOR.retzlaff.com>

Jason Tishler (jason at tishler.net) wrote:
>Please post instead of sending private email.

OK

>On Thu, Apr 03, 2003 at 08:48:24AM -0800, Jimmy Retzlaff wrote:
>> Python 2.2.2 (#1, Mar  9 2003, 08:18:26)
>> [GCC 3.2 20020927 (prerelease)] on cygwin
>> 
>> The output is attached.
>>
>> ...
>> 172 tests OK.
>> 1 test failed:
>>    test_commands
>> ...
>
>XP Pro is the same as NT 4.0 and 2000 (as expected).
>
>Did you rebase your system before running the regression test?

I did.

Jimmy




From martin at v.loewis.de  Wed Apr 23 01:20:43 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 23 Apr 2003 07:20:43 +0200
Subject: Locale settings
References: <3ea437a1_1@dns.sd54.bc.ca> <b81k5m$822$06$1@news.t-online.com> <fc3b2087.0304221550.5ebdf985@posting.google.com>
Message-ID: <m3vfx5dc2c.fsf@mira.informatik.hu-berlin.de>

bvdpoel at kootenay.com (Bob van der Poel) writes:

> The funny think about all this is that my user didn't have any
> problems when he was using Mandrake Linux 9.0. Just the upgrade to 9.1
> which caused the problem. Which is why I was thinking that he'd not
> set the locale stuff up properly at his end.

This might well be the problem. However, it might also be that the
system treats "de" as an alias to, say, de_DE.ISO-8859-15. The C
library would know about this, but Python's getdefaultlocale can't
possibly know.

Regards,
Martin




From alloydflanagan at attbi.com  Tue Apr  8 14:05:02 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 8 Apr 2003 11:05:02 -0700
Subject: Intellectual property and management fashion (was: How to protect python code ?)
References: <mailman.1049758213.26256.python-list@python.org> <slrnb945tu.p4.bignose-hates-spam@rose.localdomain.fake> <5ppka.23067$It5.15751@news2.central.cox.net> <v95njdh72j0af9@corp.supernews.com>
Message-ID: <a8b7f07a.0304081005.7fafef2d@posting.google.com>

claird at lairds.com (Cameron Laird) wrote in message news:<v95njdh72j0af9 at corp.supernews.com>...
...
> As a non-lawyer, I understand this to support the proposition
> that, when I buy from Microsoft, I first need to investigate
> all of Microsoft's affairs, to be certain the product I intend
> to buy doesn't infringe on any patents.  So:  as a Microsoft
> customer, my liability is essentially unlimited, while Micro-
> soft's "is usually limited to the purchase price of the 
> software."
> 
> It's good to know that.

It's not entirely Microsoft's fault; the whole case is a classic
example of why software patents are a terrible idea.  See the cases
listed by The League for Programming Freedom (http://lpf.ai.mit.edu/).

Most of these patents are invalid due to prior art, it's virtually
impossible to determine if you've infringed on a patent, many are
riduculously broad -- it's a terrible mess.



From exarkun at intarweb.us  Mon Apr  7 15:09:24 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 7 Apr 2003 15:09:24 -0400
Subject: super question
In-Reply-To: <fd1c664e.0304071056.31a0d377@posting.google.com>
References: <fd1c664e.0304071056.31a0d377@posting.google.com>
Message-ID: <20030407190924.GB29780@meson.dyndns.org>

On Mon, Apr 07, 2003 at 11:56:42AM -0700, Danra wrote:
> [snip]
>
> TypeError: super() argument 1 must be type, not classobj
> 
> What am I doing wrong?
> 

  super() works on type objects, but what you have above are classic class
objects.

    class B(object):
        pass

    class C(B):
        pass

    super(C)  # <super: <class 'C'>, NULL>

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 18 days, 15:01, 12 users, load average: 0.12, 0.09, 0.06




From hygt2 at cam.ac.uk  Mon Apr 28 21:23:59 2003
From: hygt2 at cam.ac.uk (Garrison Tsang)
Date: Tue, 29 Apr 2003 02:23:59 +0100
Subject: Python code -> to C code
Message-ID: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk>

Dear all,

I just signed up and this is my first post.

Knowing nothing about Python at the moment, I need to convert a python-code
into compatible c-code.  I recall that were some program/utility that does
the transformation for you.

Can anyone help?

Best regards,
Garrison





From brian at sweetapp.com  Mon Apr  7 13:23:51 2003
From: brian at sweetapp.com (Brian Quinlan)
Date: Mon, 07 Apr 2003 10:23:51 -0700
Subject: selecting a class in execution time
In-Reply-To: <AE59EE7E-691B-11D7-8CA9-000393B10A78@mac.com>
Message-ID: <003401c2fd2a$75970340$21795418@dell1700>

I'd be nice to know the why of this to suggest the best idiom.

> Object x = Class.forName( "mypackage.TheClass" ).newInstace();

x = getattr(__import__("mypackage"), "TheClass")()

The last set of parens is the argument list for TheClass' constructor.

If you know the name of the package at compile time then:

import mypackage
x = getattr(mypackage, "TheClass")()

Cheers,
Brian 





From 'tert at pncen.hf'.decode  Sat Apr 12 23:00:43 2003
From: 'tert at pncen.hf'.decode (Greg Krohn ('rot-13'))
Date: Sat, 12 Apr 2003 22:00:43 -0500
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
Message-ID: <b7ajth02vi8@enews3.newsguy.com>

"Istvan Albert" <ialbert at mailblocks.com> wrote in message
news:OJ3ma.462340$sf5.784124 at rwcrnsc52.ops.asp.att.net...

[snip]

> while 1:
> line = file.readline()
> if not line: break
>
> Maybe I'm picky here but having to start an infinite loop
> then needing a conditional to  break out of it is
> anything but an elegant pattern.

Is this more elegant:

>>> f = file('somefile', 'r')
>>> for line in f.readlines():
...  processTheLine(line)
...
>>> f.close()


Or in newer Pythons (at least 2.2.1, maybe earlier, too - I can't remember):

>>> f = file('somefile', 'r')
>>> for line in f:
...  processTheLine(line)
...
>>> f.close()




greg





From paddy3118 at netscape.net  Wed Apr 23 16:58:00 2003
From: paddy3118 at netscape.net (Paddy McCarthy)
Date: 23 Apr 2003 13:58:00 -0700
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com>
Message-ID: <2ae25c6b.0304231258.27506543@posting.google.com>

aahz at pythoncraft.com (Aahz) wrote in message news:<b867d7$f53$1 at panix2.panix.com>...
> As I just mentioned in another post, I'm writing an ASCII database for
> my address book.  Actually, the main reason for writing it was because
> I've been maintaining multiple lists for various kinds of party invites,
> and the update problem was killing me.
> 
<<SNIP>>
> 
> But firstly, I ran into the same problem I had when writing the BCD
> module: how the heck does one write tests when the results are
> hand-calculated?  There were several times when I wrote what I thought
> were correct tests for the BCD module, and my code was right and my test
> was wrong.
> 
> In the case of my ASCII database, I'm going to need a big matrix to
> properly test the query language.  How do I build that matrix?

You could use another language, another algorithm or both, and check
one against the other.


Cheers,  Pad.



From dcengija_remove_ at inet.hr  Thu Apr  3 03:53:15 2003
From: dcengija_remove_ at inet.hr (Davor Cengija)
Date: Thu, 03 Apr 2003 10:53:15 +0200
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com>
Message-ID: <1140234.2iir6UR1sK@lothar.cengija.com>

christophe grimault wrote:

> 
> One way is to deliver only .pyc files, but I guess this is quite a poor
> protection, uh ?

There is at least one decompiler available on the net, so delivering .pyc 
only doesn't make any sense.

> 
> An other way is to move some parts to a C API ext module and deliver
> only the .so,
> but this is quite time consuming ...
> 

I'd try with python-to-exe compiler. Read this: 
http://www.mcmillan-inc.com/install1.html . I tried it on both Linux and 
Windows and it really works good, even with GUI intensive applications. No 
problems so far. I don't know, however, how easy is to decompile the 
binaries produced in such a way. Anyway, if somebody really really wants to 
steal your algorithm, he/she will probably find a way to do that. With 
delivering .exe (and .so) you make that process a little bit more difficult.

And as a side-effect, it turns your 10k python source into 15MB binary 
installation, which automatically makes a respectable developer from you 
:-))

-- 
Davor Cengija, dcengija_remove_ at inet.hr



From khowe at perfnet.ca  Thu Apr 24 10:35:19 2003
From: khowe at perfnet.ca (Kevin Howe)
Date: Thu, 24 Apr 2003 10:35:19 -0400
Subject: php equivalents?
Message-ID: <OHSpa.8996$h%2.915957@read1.cgocable.net>

Hi,

I'm switching from PHP to Python, and am wondering if there are Python
equivalents to the following two PHP functions:

> register_shutdown($func)

This is basically a "cleanup" function. Just before a script ends, it will
invoke all functions registered using register_shutdown(). This can be used
to close database connections, clear sessions, save data, etc.

> var_export($var)

This function converts a variable to a text string of valid code which can
be reevaluated at any time. Python has pickle/cPickle, and also has modules
that can dump/load to XML, however, none of these create actual code, but
instead create a formatted data string.

Any help is much appreciated.

- Kevin







From lx1978 at 263.net  Sun Apr  6 17:45:48 2003
From: lx1978 at 263.net (Xin Liu)
Date: Mon, 7 Apr 2003 05:45:48 +0800
Subject: Thank you,  all freind, for your help to me! :)
References: <b6q0n6$1c55$1@mail.cn99.com>
Message-ID: <b6q74s$1fa9$1@mail.cn99.com>





From peter at engcorp.com  Wed Apr  2 08:38:42 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 08:38:42 -0500
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com>
Message-ID: <3E8AE7E2.4F9B3E8B@engcorp.com>

sdieselil wrote:
> 
> How can I implement double-linked lists and other complex tree-like
> structures in Python?

In precisely the same way you would go about doing it with any
other language, really...  you could take a stab at it, say, using 
the interactive interpreter and see how far you get, then ask for
help when you've hit a wall.

For a start: (untested code)

>>> class Node:
...   def __init__(self):
...     self.prev = self.next = None
...
>>> class LinkedList:
...   def __init__(self):
...     self.head = self.tail = None
...   def append(self, node):
...     if self.tail:
...       self.tail.next = node
...       node.prev = self.tail
...     else:
...       self.head = self.tail = node

and other stuff like that...

Of course, you might also want to do a Google search and see
if there is an existing implementation.  Ask here again if you
need help using Google. ;-)

-Peter



From ckea25d02 at sneakemail.com  Fri Apr 18 00:29:37 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 18 Apr 2003 16:29:37 +1200
Subject: wxPython: Intercepting keyboard events from multiple widgets?
In-Reply-To: <mailman.1050572372.16930.python-list@python.org>
References: <b7le9j$266le$1@ID-169208.news.dfncis.de> <mailman.1050572372.16930.python-list@python.org>
Message-ID: <b7nu9u$2pbg0$1@ID-169208.news.dfncis.de>

Donnal Walter wrote:
> Would an accelerator table do the job?
> 
> http://www.wxwindows.org/manuals/2.4.0/wx28.htm#wxacceleratortable

That seems to work, thanks. I'm not sure it's exactly
what I want, because sometimes some controls might
need to override some of the accelerators. But it
will do for a start.

I have another problem now. I've set up an Edit
menu with wxID_CUT etc. in the appropriate places.
According to the docs wxTextCtrl is supposed to
come with default event bindings for these, but
they don't seem to do anything.

And if I try to
use EVT_MENU to explicitly attach a handler for
one of them to a wxTextCtrl, it never gets called,
although if I do that for the parent window of the
wxTextCtrl (which is a wxNotebook) it does get called.
It's as though the menu event is not being delivered to the
control itself, but to its parent.

Is there anything special I have to do to get menu
events delivered to text controls?

Thanks,

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ckea25d02 at sneakemail.com  Thu Apr 17 23:34:38 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 18 Apr 2003 15:34:38 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2brz514kb.fsf@orbtech.com>
References: <mailman.1050394215.20671.python-list@python.org> <b7i26g$17t5m$1@ID-169208.news.dfncis.de> <m24r4yh1l8.fsf@orbtech.com> <b7l1ov$22o4q$1@ID-169208.news.dfncis.de> <m2brz514kb.fsf@orbtech.com>
Message-ID: <b7nr2r$2t3nf$1@ID-169208.news.dfncis.de>

Patrick K. O'Brien wrote:
> Assuming I had the
> mechanism in place to bind keystrokes to a user-created command, how
> would I know what parameters that command expected?

I think there's a mistake here in thinking of commands
as being bound to keystrokes and nothing else. It
would be better to think in terms of handlers for
keystrokes being bound to particular objects, or
classes of objects.

For instance, a command that operated on the text
in a buffer would be attached to the Buffer class,
and one that operated on a window would be
attached to the Window class.

So the command for binding a command wouldn't
just say "bind this function to this key", it
would say "bind this function to an occurrence
of this key being received by this class of
object".

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From jkn at nicorp.f9.co.uk  Sun Apr  6 09:49:46 2003
From: jkn at nicorp.f9.co.uk (Jon Nicoll)
Date: 6 Apr 2003 06:49:46 -0700
Subject: "Automation" of OpenOffice apps via Python?
References: <mailman.1049616555.27075.python-list@python.org>
Message-ID: <8351bb33.0304060549.51f7e8a5@posting.google.com>

Hi Alessandro

Have a look here for details of the python bridge to the OpenOffice
'Uno' API.

http://www.budischewski.com/pyuno/python-bridge.html

I have been meaninig  to try this out but haven't got to it yet. I'm
sure other people here would be interested to learn of your
experiences with python-bridge

    Regards
    jon N


Alessandro Bottoni <alessandro.bottoni at infinito.it> wrote in message news:<mailman.1049616555.27075.python-list at python.org>...
> I'm in the need of remote-controlling OpenOfficeWriter and others OpenOffice 
> applications from a custom program and I wonder if there is any way to acces 
> the OpenOffice API from Python. The "at a glance" presentation of OpenOffice 
> SDK/API tells that that is possible but I'm unable to find any further 
> instruction.
> 
> Does anybody know if there is a specific Python binding for the OpenOffice 
> API? Or is the API accessible from the "standard"  versions of Python and 
> OpenOffice via... via what? CORBA? XPCOM? DCOP? Telnet?! What else?
> 
> Many thanks in advance for you attention.
>  
> ---------------------
> Alessandro Bottoni



From aleax at aleax.it  Sat Apr 26 06:31:28 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 10:31:28 GMT
Subject: MD5 to long?
References: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de>
Message-ID: <4etqa.9481$K35.309220@news2.tin.it>

Axel Grune wrote:

> Hi, is there any function which can convert a MD5 or SHA1 sum
> (e.g. fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8) to a long integer?

Just call the built-in type long, with a second argument of 16, and
you can easily turn any sequence of hex digits into a long.  E.g.:

>>> import md5
>>> mm = md5.new('four score and seven years ago')
>>> print mm.hexdigest()
6b1c88b5b09b86d22fb12b50f74a19a3
>>> long(mm.hexdigest(), 16)
142375552663685867188287505851356944803L
>>>


Alex




From Dries.Jan at skynet.be  Sun Apr 13 15:07:50 2003
From: Dries.Jan at skynet.be (Jan Dries)
Date: Sun, 13 Apr 2003 21:07:50 +0200
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com>
Message-ID: <3E99B586.5040107@skynet.be>

Don't worry too much about protected methods. Bjarne Stroustrup, creator 
of the C++ language, considers them to be a design mistake, and now 
regrets they were ever added to C++. See "The Design and evolution of 
C++" by Bjarne Stroustrup, page 302.
Java and I assume Ruby as well simply copied the concept from C++.

Regards,
Jan

srijit at yahoo.com wrote:
> I do not see a  concept of Protected Methods in Python 2.2 like C++ and Ruby.
> Is this a serious limitation for Object Oriented Programing with Python? 
> Or may be it is possible to emulated Protected Methods in Python.
> 
> I look forward to some discussion by OOP experts.
> 
> Regards,
> Srijit






From rcook at llnl.gov  Tue Apr  1 21:39:53 2003
From: rcook at llnl.gov (Richard Cook)
Date: Tue, 1 Apr 2003 18:39:53 -0800
Subject: PyObject_CallFunction with keywords?
Message-ID: <p05200f08baaff8e7aee8@[134.9.17.165]>

Hi,
Short question:
Is it possible to pass keywords to a function with PyObject_CallObject()?


Long question:
I am using an embedded Python interpreter in our C program to execute 
a method from a Python module (pexpect, to be exact).  At a certain 
point, I wish to do the equivalent of the following Python code in my 
C code:

result =  obj.expect("hello", timeout=50);

So I have a PyObject *f which points to obj.expect, and a PyObject *s 
which is a list containing the string "hello".  I figure I have to do:
PyObject_CallObject(f, s, k)
where k is the timeout parameter, but this does not match the param 
list for PyObject_CallObject().

My question is, how can I pass keyword parameters such as the timeout 
parameter in my example to a function which accepts them?  The 
documentation for PyObject_CallObject says it is equivalent to 
apply(o, args), but says nothing about the actual apply syntax:
apply(o, args, keywords)
where keywords is a dictionary.

-- 
Richard Cook
Lawrence Livermore National Laboratory
Bldg-451 Rm-2043, Mail Stop L-561
7000 East Avenue,  Livermore, CA, 94550, USA
phone (925) 423-9605 (work)    fax (925) 423-8704
---
Information Management & Graphics Grp., Services & Development Div., 
Integrated Computing & Communications Dept.
(opinions expressed herein are mine and not those of LLNL)




From kkto at csis.hku.hk  Tue Apr  1 07:56:30 2003
From: kkto at csis.hku.hk (Isaac To)
Date: 01 Apr 2003 20:56:30 +0800
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <IFcia.39458$Jg1.934636@news1.tin.it>
Message-ID: <7i1y0m8jgh.fsf@enark.csis.hku.hk>

>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:

    Alex> Jacek Generowicz wrote: ...
    >> Will you make this code public ?

    Alex> Sure, it's nothing special -- just a tiny, simple example.

    Alex> Here's the naivest Python I could come up with for the task
    Alex> (Python 2.3 given the use of enumerate, but that's easy to remove
    Alex> if you want, of course):

The algorithms have different time order.  The Python version is around O(n)
due to the use of hashed dictionary, the C++ version is O(n log n) due to
the use of tree maps.  This is a two-side sword, though: I usually want to
have a predecessor operation in Python but then remember that it is
impossible unless I use some custom dictionaries.  BTW, nearly all STL
implementation of STL does support hashed implementation, the problem is
just that the standard committee is waiting for C++ implementations to catch
up with the current standard before rolling out new things into the library.

Regards,
Isaac.



From jepler at unpythonic.net  Tue Apr  8 22:54:15 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Tue, 8 Apr 2003 21:54:15 -0500
Subject: It seems that some *.o file is not installed, but how can I get them?
In-Reply-To: <0HD100EE2YK4UG@mail.fudan.edu.cn>
References: <0HD100EE2YK4UG@mail.fudan.edu.cn>
Message-ID: <20030409025415.GB1488@unpythonic.net>

You'll need -l<libname>.  Based on the included error messages, you
probably need -lm -ldl -lpty

In the /usr/lib/python2.2/config/Makefile, there's a definition of
SYSLIBS.  It may be that this is the list you're looking for.

Jeff




From aleax at aleax.it  Tue Apr  1 04:07:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 09:07:52 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch>
Message-ID: <IFcia.39458$Jg1.934636@news1.tin.it>

Jacek Generowicz wrote:
   ...
> Will you make this code public ?

Sure, it's nothing special -- just a tiny, simple example.

Here's the naivest Python I could come up with for the
task (Python 2.3 given the use of enumerate, but that's
easy to remove if you want, of course):


# build a word -> line numbers mapping
import sys
idx = {}
for n,line in enumerate(sys.stdin):
    for word in line.split():
        idx.setdefault(word,[]).append(n)

# display by alphabetically-sorted word
words = idx.keys(); words.sort()
for word in words:
    print "%s:" % word,
    for n in idx[word]: print n,
    print


Note that "words" are defined in the naivest way (just
the way the >> input operator will build them in C++),
as whitespace-separated sequences of non-whitespace.

Here's the closest simple translation I could easily give 
for this in C++:


#include <string>
#include <iostream>
#include <sstream>
#include <map>
#include <vector>

int main()
{
    typedef std::map<std::string, std::vector<int> > index;
    index idx;

    std::string line;
    int n = 0;
    while(getline(std::cin, line)) {
        std::istringstream sline(line);
        std::string word;
        while(sline >> word) {
            idx[word].push_back(n);
        }
        n += 1;
    }

    for(index::iterator i = idx.begin(); i != idx.end(); ++i) {
        std::cout << i->first << ": ";
        for(std::vector<int>::iterator j = i->second.begin();
                j != i->second.end(); ++j) {
            std::cout << ' ' << *j;
        }
        std::cout << "\n";
    }

    return 0;
}


and here is a slightly optimized Python version:


def main():
        import sys
        idx = {}
        listofword = idx.setdefault

        for n,line in enumerate(sys.stdin):
            n = str(n)
            for word in line.split():
                listofword(word,[]).append(n)

        words = idx.keys()
        words.sort()
        emit = sys.stdout.writelines
        for word in words:
            emit((word,':',' '.join(idx[word]),'\n'))

main()


Alex




From jbauer at rubic.com  Sat Apr 26 21:21:44 2003
From: jbauer at rubic.com (Jeff Bauer)
Date: Sat, 26 Apr 2003 20:21:44 -0500
Subject: Python 2.3b1 + Tkinter on Solaris 2.6
Message-ID: <3EAB30A8.F16C6748@rubic.com>

Prior to filing a bug report, I wanted to check
if there's anything obvious I'm overlooking with
getting tk8.4.2 working with Python2.3b1 on
Solaris 2.6.

I've not modified Modules/Setup, but I have recompiled
and reinstalled tcl/tk after the first failure to
ensure that I'm not working with stale libraries.

running build
running build_ext
building '_tkinter' extension
gcc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -fPIC -DWITH_APPINIT=1 
  -I/usr/openwin/include -I.
-I/export/home/jbauer/Python-2.3b1/./Include
  -I/usr/local/include -I/export/home/jbauer/Python-2.3b1/Include
  -I/export/home/jbauer/Python-2.3b1
  -c /export/home/jbauer/Python-2.3b1/Modules/_tkinter.c 
  -o build/temp.solaris-2.6-sun4u-2.3/_tkinter.o

gcc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -fPIC -DWITH_APPINIT=1
  -I/usr/openwin/include -I.
-I/export/home/jbauer/Python-2.3b1/./Include
  -I/usr/local/include -I/export/home/jbauer/Python-2.3b1/Include
  -I/export/home/jbauer/Python-2.3b1
  -c /export/home/jbauer/Python-2.3b1/Modules/tkappinit.c
  -o build/temp.solaris-2.6-sun4u-2.3/tkappinit.o

gcc -shared build/temp.solaris-2.6-sun4u-2.3/_tkinter.o 
  build/temp.solaris-2.6-sun4u-2.3/tkappinit.o -L/usr/openwin/lib
  -L/usr/local/lib -ltk8.4 -ltcl8.4 -lX11
  -o build/lib.solaris-2.6-sun4u-2.3/_tkinter.so

*** WARNING: renaming "_tkinter" since importing it failed:
  ld.so.1: ./python: fatal: relocation error: file
  /usr/local/lib/libtk8.4.so: symbol __eprintf: referenced symbol not
found

--
Jeff Bauer
Rubicon Research




From jtiller at sjm.com  Tue Apr  1 20:58:30 2003
From: jtiller at sjm.com (Jason Tiller)
Date: Tue, 1 Apr 2003 17:58:30 -0800 (Pacific Standard Time)
Subject: converting from shell script to python
In-Reply-To: <oqsmt17k43.fsf@alcyon.progiciels-bpi.ca>
Message-ID: <Pine.WNT.4.44.0304011754180.1416-100000@WS-USSV-50244>

Hi, Shao & Francois, :)

On 1 Apr 2003, Francois Pinard wrote:
> [Shao Zhang]

> > I need to convert a very simple shell script to python in order
> > get it run on both windows and unix.

> Another possibility is to install a ported `bash' on Windows.  I do
> not use Windows often, but the last time I had to, a while ago,
> `bash' was working quite well.  I used the one from DJGPP -- Cygwin
> probably has one too.

Truly, Cygwin provides a very sophisticated *nix emulation
environment, including all of the popular shells (bash, csh, tcsh,
pdksh, and zsh), major script languages (tcl, perl, python, etc.), and
just about anything you'd need to recompile and run *nix software on
Windows.  There's even a full-fledged X server on which (insane)
people are running KDE.

I'm using Cygwin's Python 2.2.2 + Tkinter + Pmw to go through the
examples in John Gray's "Python and Tkinter Programming," and the
environment is *very* good.

www.cygwin.com

---Jason





From cnetzer at mail.arc.nasa.gov  Mon Apr 28 16:10:38 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 28 Apr 2003 13:10:38 -0700
Subject: Getting mouse position interms of canvas unit.
In-Reply-To: <20030428152318.80404.qmail@web41710.mail.yahoo.com>
References: <20030428152318.80404.qmail@web41710.mail.yahoo.com>
Message-ID: <1051560638.593.5.camel@sayge.arc.nasa.gov>

On Mon, 2003-04-28 at 08:23, perl lover wrote:
> hi,
>   iam new to python and tkinter. I have created a
> canvas of size 300m * 300m (in millimeter). I bind
> mouse move method to canvas. when i move the mouse
> over the canvas the mouse position gets printed in
> pixel unit. But i want to get mouse position values
> interms of canvas unit (ie, millimeter). How can i get
> mouse position values interms of canvas unit?

Below is a demonstration of the concept, based on your original code. 
It basically finds the screen dimensions in units of pixels and
millimeters, and from that, creates a conversion factor that can be
multiplied to pixel values.


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
from __future__ import division
from Tkinter import *

if __name__ == '__main__':
    root = Tk()

    # get screen width and height in both mm and pixels
    # use values to create factor for converting pixels to mm
    #
    # Note - left as globals for easy demonstration use in callback
    mm_x = root.winfo_screenmmwidth() / root.winfo_screenwidth()
    mm_y = root.winfo_screenmmheight() / root.winfo_screenheight()

    c = Canvas(root, width="300m", height="300m", background = 'gray')
    c.pack()

    # Note - Should change this to use currying rather than globals.
    def mouse_pos_in_mm(event):
        """Report mouse position in canvas coords, units of mm."""
        print c.canvasx(event.x) * mm_x, c.canvasy(event.y) * mm_y

    c.bind('<Motion>', mouse_pos_in_mm)

    c.create_rectangle('16m','10.5m','21m','15.5m',fill='blue')

    root.mainloop()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From tim.one at comcast.net  Sun Apr 27 22:33:55 2003
From: tim.one at comcast.net (Tim Peters)
Date: Sun, 27 Apr 2003 22:33:55 -0400
Subject: Forgetting "()" when calling methods
In-Reply-To: <g3qdncyq1Zx2FzGjXTWcrg@comcast.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCOEFJEEAB.tim.one@comcast.net>

[Terry Reedy]
> ...
> At one time, it was easy for anyone to get 100% on a quiz like 'how
> many of the following lists can Python sort?".

OTOH, how many would have scored 100% if also asked what the result of
sorting would be?  I'm not sure anyone could have.  For years, ints came out
"less than" lists, which in turn came out "less than" longs, so that

    [10, [9], 7L, 8]

was considered to be "in sorted order" as-is.  It was something like 7 years
before someone finally noticed how nuts that was.

> Now, even some experienced users, might miss the correct answer for
> the following pair (0, 1, or 2?):
>
> [1, 1j].sort()
> ['1', 1j].sort()
>
> (For anyone who is not sure, make a guess and then cut&paste to fine
> out.)

It may be even more valuable for people who are certain <wink>.

> One bad turn does not deserve another.

I'm pretty sure that if Python had never supported meaningless (although
consistent within a single run) comparisons, a compelling case for adding
them couldn't have been made.  Mixed-type comparisons are usually
meaningless (which is less, an xrange object or math.sin?  The answer is
math.sin.  So how does math.sin compare to math.cos?  The answer is that's
not defined, but that you'll get the same result each time you compare those
two within a single program run).  When that's so, sorting brings objects of
the same type next to each other, and that's about it.  But if that's "the
use case" (if it isn't, what is the use case?!), it's much more efficient
(O(N) vs O(N log N)) to segregate objects via a dict mapping type to a list
of objects with that type.






From aleax at aleax.it  Sat Apr 19 13:35:45 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 17:35:45 GMT
Subject: Parent references: hot or not?
References: <1050763980.14721.41.camel@ida.dylanreinhardt.com> <20030419145756.GA25916@meson.dyndns.org> <1050771656.14717.48.camel@ida.dylanreinhardt.com> <mailman.1050770791.27643.python-list@python.org>
Message-ID: <RNfoa.58223$T34.1557059@news2.tin.it>

Dylan Reinhardt wrote:

> On Sat, 2003-04-19 at 10:00, Dylan Reinhardt wrote:
> 
>> Sounds like I should be sure to write a __del__ method for Container
>> that ensures things work correctly.
> 
> And upon re-reading your post you say *not* to do that... :-)
> 
> I'd like to understand that a bit better.  What would be wrong with:
> 
> class Container:
> 
>    def __del__(self):
>       # ensure all Items are deleted first
>       for item in self.items.keys():
>          del self.items[item]

It would never be executed when self.items is non-empty: since
each item in self.items holds a reference to self, self can
never be destroyed in the normal way (by the number of references
on it going to zero), *AND*, the very existence of a __del__
STOPS the garbage collector from collecting reference-loops.


Alex




From aleax at aleax.it  Tue Apr 15 10:40:13 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 14:40:13 GMT
Subject: Using property in classic class may not work
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it> <a8b7f07a.0304150527.5b0b1187@posting.google.com>
Message-ID: <hRUma.23848$T34.709378@news2.tin.it>

A. Lloyd Flanagan wrote:

> Alex Martelli <aleax at aleax.it> wrote in message
> news:<ifPma.15344$LB6.409776 at news1.tin.it>...
>> Neil Hodgson wrote:
>> 
>> >     The class in question is derived from a wxWindows class, so I can't
>> > make it derive just from object. The "What's New in Python 2.2"
>> > document only mentions property wrt new-style clases. Can I depend on
>> > property working for classic classes? Is there a workaround such as
>> > deriving from object as well?
>> 
>> No, and yes, respectively -- deriving from object means your class
>> isn't classic any more, so properties do work as intended there.  Vide:
>> 
> ... (big chunck cut) ...
>> as you see, this does work as intended.
>> 
>> 
>> Alex
> 
> However, the last time I tried to subclass from both wxFrame and
> object, wxPython freaked out (that's a technical term for "I don't
> remember what happened, but boy, it sure didn't work").
> Not sure about the latest version of wxPython, though -- there's
> binaries for python 2.3 out by now.  I'll have to try it again..

If wxPython keeps freaking, you may have to give up on properties
if you HAVE to inherit its classes (can't just wrap with automatic
delegation) and go back to "good old" (yeah right) __setattr__.  Or,
gently pressure wxPython's maintainers to have them move to
new-style classes... whatever aspects of old-style classes they
may be taking for granted at present, migrating to new-style can
hardly be a major job, in my experience.


Alex




From gustav at morpheus.demon.co.uk  Sat Apr 19 09:39:44 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Sat, 19 Apr 2003 14:39:44 +0100
Subject: Simple wxPython Application
References: <7d22d8d6.0304190355.6098714c@posting.google.com>
Message-ID: <znmmr4gv.fsf@morpheus.demon.co.uk>

ekranawetter at bmw.co.at (ekranawetter_steyr) writes:

> from wxPython.wx import wxPySimpleApp, wxFrame
>   app = wxPySimpleApp()
>   frame = wxFrame(None, -1, "Hello World")
>   frame.Show(1)
>   app.MainLoop()
>
> I got a syntax error in line 2

If you typed it exactly like this, you'll get an error "SyntaxError:
invalid syntax", which admittedly is not very helpful. The problem is
the indentation. Indentation in Python is significant, and indicates
block structure. The first line (from...import) does not start a
block, so indenting the next line is wrong.

Sorry, I can't help with your second problem.

Paul.
-- 
This signature intentionally left blank



From staschuk at telusplanet.net  Mon Apr 21 14:36:39 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 21 Apr 2003 12:36:39 -0600
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <b7ufg4$o1d$1@news.hccnet.nl>; from anton@vredegoor.doge.nl on Sun, Apr 20, 2003 at 05:42:24PM +0200
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl>
Message-ID: <20030421123639.B596@tibia.amotlpaa.bogus>

Quoth Anton Vredegoor:
> "A.M. Kuchling" <amk at amk.ca> wrote:
> 
> >http://mail.python.org/pipermail/python-dev/1999-July/000467.html
> 
> Thanks. It seems Tim is making a heroic effort to explain the new in
> terms of the old. This necessarily must fail, but what other option is
> there instead of rebuilding all terminology out of gotos?

I am not sure it needs to fail, nor that it is a bad idea to
"rebuild" everything out of gotos.

One possible method for understanding continuations, for those
with C (or similar language) experience:  Assume you have only one
control flow construct, a conditional goto.  Show how if/else,
loops, and function call/return can be implemented using
conditional goto.  Then show how coroutines, generators, and
backtracking can be implemented using it.  Notice especially your
use of constructions which consist of "save state, goto there" and
"goto there, restore state" combined in useful ways, for various
values of "state".

Having thus prepared the conceptual ground, *then* go read
material on continuations.

-- 
Steven Taschuk                                                 o- @
staschuk at telusplanet.net                                      7O   )
                                                               "  (




From gh at ghaering.de  Fri Apr 25 13:32:16 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 19:32:16 +0200
Subject: Python Package Index @ python.org down?
In-Reply-To: <3ea96bce$0$49106$e4fe514c@news.xs4all.nl>
References: <3ea96bce$0$49106$e4fe514c@news.xs4all.nl>
Message-ID: <3EA97120.3000705@ghaering.de>

Irmen de Jong wrote:
> I get a Python stacktrace when trying to access the Python Package Index,
> http://www.python.org/pypi
> 
> "OperationalError: database is locked"
> 
> Tried three times. Anything wrong over there?

I did get these as well yesterday.

See http://mail.python.org/pipermail/catalog-sig/2003-April/000376.html

-- Gerhard




From aleax at aleax.it  Sun Apr 27 03:54:28 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 07:54:28 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com>
Message-ID: <U0Mqa.13860$K35.457281@news2.tin.it>

Erik Max Francis wrote:

> Jeremy Fincher wrote:
> 
>> Perhaps the answer is to remove the __nonzero__ method on
>> functions/methods.  Logically, what purpose does it serve?
> 
> # f holds a callable object or None
> if f:
> f(...)
> 

I think "if f is None:" and even more "if callable(f):" are
better ways to express this test -- explicit is better than
implicit.  However, if this (inferior) expression is indeed
widespread, then we can't break it until Python 3000 (well,
we COULD start giving warnings about it, etc, etc, but it's
unlikely that this can happen).


Alex




From nomad*** at ***absamail.co.za  Mon Apr  7 07:55:07 2003
From: nomad*** at ***absamail.co.za (Nomad)
Date: Mon, 07 Apr 2003 13:55:07 +0200
Subject: PYTHONPATH question
References: <E1927CG-0007k5-00@fast.aub.edu.lb> <mailman.1049632149.9888.python-list@python.org>
Message-ID: <5ai29vogfthpqvbmlic09pbllltun61d2l@4ax.com>

On Sun, 6 Apr 2003 07:27:49 -0500, Skip Montanaro <skip at pobox.com>
wrote:

>
>    Ali> shouldn't i be able to read PYTHONPATH directly like i read any other env 
>    Ali> variable? because i am saying
>    Ali> echo $PYTHONPATH
>    Ali> and nothing is being displayed.
>
>That just means PYTHONPATH isn't set currently.  In your ~/.profile, simply
>set
>
>    PYTHONPATH=<colon-separated list of directories>
>    export PYTHONPATH
>
>and you should be okay.

For future safety, it might be an idea to make it:

    PYTHONPATH=$PYTHONPATH:<colon-separated list of directories>
    export PYTHONPATH

just in case you add any 3rd party modules later that add themselves
to your `global' PYTHONPATH.  Of course, you could also reverse the
order, placing the global $PYTHONPATH at the end rather than the
beginning -- its a matter of taste I suppose.

HTH

-- 
Nomad

Wondering of the vast emptyness of the 'net
in search of something cool.



From bokr at oz.net  Thu Apr  3 16:59:17 2003
From: bokr at oz.net (Bengt Richter)
Date: 3 Apr 2003 21:59:17 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com> <wBIia.2000$y3.310439@news010.worldonline.dk> <b6fmcq$dh0$1@panix2.panix.com>
Message-ID: <b6iarl$3lb$0@216.39.172.122>

On 2 Apr 2003 16:57:46 -0500, roy at panix.com (Roy Smith) wrote:
[...]
>
>The dark side of the force is seductive.  It tries to make us
>efficient, even at the cost of clarity and maintainability.  In that
>way lies bit-twiddling and assembler language and soldering irons.  We
>must be diligent in our fight against it.

Or schedule it. If it's scheduled, it can't be called procrastination ;-)

Regards,
Bengt Richter



From do-not-spam-ben.hutchings at businesswebsoftware.com  Wed Apr 16 06:32:13 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 16 Apr 2003 10:32:13 GMT
Subject: inserting Python commands into HTML documents
References: <zpqdne0Kztcb0AejXTWcow@comcast.com> <6-CdnUrJoLKFkQajXTWcqA@comcast.com>
Message-ID: <slrnb9qc9d.1bc.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <6-CdnUrJoLKFkQajXTWcqA at comcast.com>, Em wrote:
> These options all seem to require me to install 3rd party software onto 
> the server of this company, and I believe that to be outside of the 
> scope of the test project.
> 
> The only issue is that I want it to be invisible that I'm pulling from a 
> CGI script (I don't want the URL bar to say 
> "www.mywebsite.com/cgi-bin/myProj.py?page=projects", I want it to simply 
> say "www.mywebsite.com/projects.html" or "www.mywebsite.com/projects.php").
> 
> Is there a way for me to send my output to /index.html, /projects.html, 
> etc.?

You'd need to configure the server explicitly to use a CGI script for
those URLs.  I do think it's a good idea to keep URLs short, and to avoid
letting your current implementation show in the name.  In fact I would go
further and drop the '.html', which is only really necessary in static
filenames to let the server know their type.



From tjreedy at udel.edu  Tue Apr  8 10:23:14 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 8 Apr 2003 10:23:14 -0400
Subject: Need help with Python porting
References: <78ada8a6.0304080156.6a780fe2@posting.google.com>
Message-ID: <Wa-cnYpAeMrERg-jXTWcpQ@comcast.com>

"Guruprasad Prahallad" <guruprasad_p at lycos.com> wrote in message
news:78ada8a6.0304080156.6a780fe2 at posting.google.com...
> Hi all,
>
> I am currently porting Python to Netware. I would really appreciate
if
> someone could tell me where I could find a reference implementation
> for Netware. I downloaded the 2.2.2 source from python.org for
Linux.
> Could someone give me a pointer as to how to go about modifying this
> code so that it works for NetWare.

As far as I know, 'the' Python source code is *the* source code for
all the more or less 'officially' supported platforms, not just Linux.
There are a number of platform specific #ifdefs (take a look).  Have
you actually tried compiling on your system yet?  If you do, and it
fails, you can ask specific questions.

TJR





From aleax at aleax.it  Tue Apr 22 08:32:33 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 12:32:33 GMT
Subject: instance creatiion as an "event"
References: <mailman.1050964534.23836.python-list@python.org>
Message-ID: <BDapa.72872$T34.1894522@news2.tin.it>

Arthur wrote:
   ...
> I could, of course, add self.redraw() to TestPoint.__init__.  But it
> seems, because of the overall "real world" design where TestPoint is
> somewhere down an inheritance chain, I need to call this explicitly on the
> top level class - and need to do some restructuring of the inheritance

I don't understand this.  When you call:

    p = TestPoint(args)
    p.redraw()

it's EXACTLY AS IF TestPoint.__init__'s very last statement was

    self.redraw()

and I don't see any way the latter can fail to do exactly the same
job as the former.  It cannot matter how deeply nested in whatever
inheritance DAG class TestPoint might be.


However, if you did put self.redraw() in TestPoint.__init__, then
the redraw would occur WHENEVER a TestPoint instance is created,
be that "interactively", or otherwise.

> Being stumped there, and lazy, I was looking for some interactive magic.
> so that p1=TestPoint(1,2) would somehow register as an event calling
> draw_all().

You could make the redraw conditional on module sys having an
attribute named 'ps1' -- that's what, I think, is normally used
to check whether one is in an interactive session, or not.  Still,
that wouldn't tell you if a SPECIFIC call was "at the interactive
prompt".

You could help by avoiding the from-star-import idiom and
doing instead some trick to ensure that the callables in your
module are wrapped (by wrappers of the same name) for use
at the interactive-prompt only (i.e. in module __main__ iff
you're in an interactive session).  E.g. at the end of your
module's code you might have [warning, untested code!]:

if __name__ == '__main__':
    "normal run-as-script code here"
else:
    import sys
    if hasattr(sys, 'ps1'):
        import __main__
        def makewrapper(func):
            def wrapper(*args, **kwds):
                result = func(*args, **kwds)
                draw_all()
                return resukt
            return wrapper
        for name, value in vars():
            if callable(value):
                setattr(__main__,name,makewrapper(value))

or something like that.  If you do this sort of thing often
you could factor it out into an helper module, of course.


Alex




From aleax at aleax.it  Wed Apr 16 03:53:29 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 07:53:29 GMT
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com> <mailman.1050087439.11600.python-list@python.org> <435d40e0.0304130212.5659bd4d@posting.google.com> <7Bgma.8749$LB6.237526@news1.tin.it> <435d40e0.0304152211.5a9511d4@posting.google.com>
Message-ID: <ZZ7na.27561$T34.822409@news2.tin.it>

Chris wrote:

> Alex Martelli <aleax at aleax.it> wrote in message
> news:<7Bgma.8749$LB6.237526 at news1.tin.it>...
>> Chris wrote:
>>    ...
>> > Thanks Inyeol. Be nice if they gave a small example of using finditer.
>> 
>> How small?  E.g., here's how you could rewrite (a simpler version of)
>> findall using finditer (warning, untested):
>> 
>> def findall(are, instring):
>>     return [mo.group(0) for mo in are.finditer(instring)]
>> 
>> Too small?  Then perhaps you could mention a larger one...?
>> 
>> 
>> Alex
> 
> I had some small exposure to "iterators" in C++, and I remember them
> being used in loops - iterating over the items in a container. The
> example you have given appears to create a list from an "iterator"
> with no loop. Something called an "iterator" creating a list with no
> loop is very perplexing to me. But maybe my problem is I don't
> recognize the syntax [ mo.group(0) for ... ]. It looks like a method
> call on an object, followed by an incomplete for loop.

the syntax:

    [ <expression> for <target> in <iterable> ]

is known in Python as a "List Comprehension".  There is no issue of
"with no loop" -- the "for" keyword inside the list comprehension
is indicating exactly the fact that the loop is taking place.  (List
Comprehensions may optionally have other clauses too -- for clauses
to indicate nested loops, if clauses to use only SOME of the items
in an iterable -- but I'm using the simplest form in the above).

If you're familiar with Python 1.5.2 but none of the Python changes
in the last 4 years or so, then one way to explain list comprehensions
is to say that

    return [ <expression> for <target> in <iterable> ]

is just the same as:

    templist = []
    for <target> in <iterable> :
        templist.append( <expression> )
    return templist

The list comprehension notation is a Pythonization of Haskell's, which
uses punctuation rather than the for and in keywords -- it would say:
    [ <expression> | <target> <- <iterable> ]
which reads just fine in Haskell but definitely not in Python; in turn,
Haskell's list comprehension notation is a simple adaptation of the
widespread "set comprehension" notation used in maths, e.g.
    { x*x | x <- S }
to mean "the set of the squares of the elements of set S" (where
the <- set-membership indicato is often indicated by some glyph
that's more reminiscent of the Greek letter epsilon).


Python iterators are rather different beasts from C++'s, though
the analogy between a Python iterator and a C++ "input iterator"
is reasonably good... where in C++ you might have *it++ to
indicate "fetch the current value of the iterator then advance
the iterator", in Python you'd have it.next() for just the same
purpose [the for loop makes that call intrinsically on your
behalf!]; where in C++ you need to find out whether an iterator
is done by comparing it with some "end marker", in Python an
iterator lets you know it's done by raising StopIteration when
you call its .next method [again, the for loop handles that for
you intrinsically, catching StopIteration and just taking it as
the indicator for normal termination of the iteration/loop].


So, back to our muttons: are.finditer(instring) returns an iterable
(which is actually already an iterator, but that's not very
important here) whose items are the match-objects for each of
the nonoverlapping matches of compiled regular expression object
are inside string instring.  So, a loop of the form:
    for mo in are.finditer(instring):
        ...
(whether written out like this, or inside a list comprehension,
makes no difference) makes mo assume, one after the other, the
values of RE matchobjects for each of those non-overlapping
matches, left to right inside instring.  So for example a call
to mo.group(0) gives the substring of instring that corresponds
to the specific nonoverlapping match we're currently at, inside
a loop such as the above.


Alex




From drew at astro.pas.rochester.edu  Sat Apr  5 08:32:14 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 5 Apr 2003 05:32:14 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com>	<3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>	<9383481e.0304030705.524cf0fa@posting.google.com>	<7h3znn7k308.fsf@pc150.maths.bris.ac.uk>	<3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122>	<3E8CF4B5.1E221DD7@alcyone.com>	<9383481e.0304040445.a3752dd@posting.google.com> <mailman.1049471534.24194.python-list@python.org>
Message-ID: <9383481e.0304050532.44ffa0de@posting.google.com>

list-python at ccraig.org (Christopher A. Craig) wrote in message news:<mailman.1049471534.24194.python-list at python.org>...
> drew at astro.pas.rochester.edu (Drew Moore) writes:
> 
> > so far, the response seems to be:
> > "it is technically possible, but my scripts might
> > run slower.. and it just makes me nervous."
> 
> Not might run slower, would run not just slower but a ton slower.
> You're basically replacing the C code
> a=b
> 
> with at least one, and possibly several dict lookups.  Your idea about
> caching the value is a decent one, but you have the problem that class
> dicts are mutable, so what if someone does
> 
> >>> class foo:
> ...  pass
> ...
> >>> t = foo()
> >>> t.__dict__.__assign__ = lambda t, a: t.elem=a
> 
> If __assign__ is allowed, that will be expected to work.
> 
> A bigger problem to me is that '=' currently means rebinding names.
> If I see 'a=b' in Python, I immediately presume that 'a' doesn't exist
> after the assignment.  If you change this then it will sometimes mean
> rebinding and sometimes mean something else.
> 
> If you want a way to set something's value, I imagine you'll have a
> lot more support for adding an operator rather than making '=' one.
> I'm all for Lulu's suggestion of adding a ':=' operator for that sort
> of situation.

Excellent observation about the multiple dict search, Christopher!

A cached bit would not work if you expected derived classes
to inherit the overload. A base class, if modified, would have
to go to all derived instances and set the bit.

That would definitely not fly! 

The most sensible implementation I can think of:
The __assign__ method belongs to the object.
BUT---
The "use the __assign__ method if present" flag
belongs to the name, and is only set under certain
circumstances.. such as if the name
is bound to a freshly instantiated instance
that has __assign__ method defined.
(got to think about this one some more)

a = myOAclass() # a overloads assignment
b = a # b does not. it has the __assign__ method but..
      # will not use it.
b = c # b has been reassigned!!

f(a) # reference passed to f does not overload assignment!
L[3] = a # list item does not overload assignment

### perhaps the := operator has a place here...
a = myOAclass() # a does not overload assignment..
b := a # but b does..??
# (set the AO flag in b if the := operator is used
# and a has __assign__ method.)

c := myOAclass() # c overloads assignment
d = c # d does not.

Hmm, I like this..

The OA flag sensibly belongs to the name, not the object.
(its really the name that wants this property, not the object.)
Now, performance hit is tiny, because it is just checking a bit.
reassignment does not propogate the bit in the name...
no ripples..

Drew



From p-abel at t-online.de  Wed Apr  2 06:06:04 2003
From: p-abel at t-online.de (Peter Abel)
Date: 2 Apr 2003 03:06:04 -0800
Subject: [Q]How to append paths in a platform-independent way?
References: <420ced91.0304011946.56a22d2a@posting.google.com>
Message-ID: <13a533e8.0304020306.4ca06b5d@posting.google.com>

sdhyok at yahoo.com (sdhyok) wrote in message news:<420ced91.0304011946.56a22d2a at posting.google.com>...
> How to do the following jobs in a platform-independent way?
> 
> path = /home/sdhyok
> file = data.txt
> 
> 1. Append the file name to the path => /home/sdhyok/data.txt
> 2. Extract only path from /home/sdhyok/data.txt => /home/sdhyok
> 
> Daehyok

>>> import os
>>> # Because I work under MSW I took backslashes instead of slashes
>>> # But you can take os.sep, which is os-independant
>>> path=r'\home\sdhyok'
>>> #file is not a good name, because it is a builtin-function
>>> lfn='data.txt'
>>> path_file=os.path.join(path,lfn)
>>> print path_file
\home\sdhyok\data.txt
>>> print os.path.dirname(path_file)
\home\sdhyok

Regards Peter



From gregory.p.green at boeing.com  Sat Apr  5 17:51:30 2003
From: gregory.p.green at boeing.com (Greg Green)
Date: Sat, 5 Apr 2003 22:51:30 GMT
Subject: Problem setting COM property using win32com
Message-ID: <he9ck171.fsf@boeing.com>

I have a commercial com api that I'm trying to use with python. I am
trying to set a property of the com object. The property is called
Value, and I know enough to use the SetValue function. But that
doesn't work.

Here is the code:
    attrs.SetValue("CN_ID", 0, "0000005")

Here is the error:
Traceback (most recent call last):
  File "example.py", line 49, in ?
    prmidattrs.SetValue("CN_ID", 0, "0000005")
  File "C:\PYTHON22\lib\site-packages\win32com\client\dynamic.py", line 438, in __getattr__
    raise AttributeError, "%s.%s" % (self._username_, attr)
AttributeError: SmRecList.SmRecordList.SetValue

If I look at the makepy generated module, I see the method.

	# The method SetValue is actually a property, but must be used as a method to correctly pass the arguments
	def SetValue(self, HeaderName=defaultNamedNotOptArg, RecordIndex=defaultNamedNotOptArg, arg2=defaultUnnamedArg):
		"""Retrieves, sets or adds a value to the object. Specified by header name and record index."""
		return self._oleobj_.InvokeTypes(0x1, LCID, 4, (24, 0), ((8, 1), (3, 1), (12, 1)),HeaderName, RecordIndex, arg2)

Any ideas?

-- 

Greg Green



From aleax at aleax.it  Sun Apr 13 17:56:25 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 21:56:25 GMT
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk> <m38yuepdva.fsf@mira.informatik.hu-berlin.de> <LtFzMJA4ZZm+EwDh@jessikat.fsnet.co.uk> <3kjma.11793$T34.382463@news2.tin.it> <m3vfxi15xg.fsf@mira.informatik.hu-berlin.de>
Message-ID: <d2lma.9993$LB6.267225@news1.tin.it>

Martin v. L?wis wrote:

> Alex Martelli <aleax at aleax.it> writes:
> 
>> The simplest solution is to have pyRXP return a tree of *LISTS*,
> 
> Why is that a solution?
> 
>>>> x=[]
>>>> import weakref
>>>> weakref.ref(x)
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> TypeError: cannot create weak reference to 'list' object

Sigh, you're right, it isn't -- lists aren't weakly referenceable
either!  Oh well...


Alex




From staschuk at telusplanet.net  Fri Apr 25 18:39:01 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 16:39:01 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <20030417205405.A10672@tibia.amotlpaa.bogus>; from staschuk@telusplanet.net on Thu, Apr 17, 2003 at 08:54:05PM -0600
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <20030417205405.A10672@tibia.amotlpaa.bogus>
Message-ID: <20030425163901.A1656@tibia.amotlpaa.bogus>

Quoth I:
  [...]
> ... How about this?  (Draft.)
> 
> __repr__(self)
> 
>     Called by the repr() built-in function and by string conversions
  [...]

I've submitted a documentation patch for the __str__ and __repr__
docs, following ideas discussed in this thread:
    <http://www.python.org/sf/727789>
I solicit comments.  (The __repr__ blurb is much the same as that
in the previous post; I've fiddled more extensively with the
__str__ blurb.)

__repr__(self)

    Called by the repr() built-in function and by string conversions
    (reverse quotes) to compute a string representation of an object.
    The return value must be a string object.

    The primary audience of this representation is programmers seeking
    detailed information about the object for debugging or inspection
    purposes; thus it is important that the representation be
    information-rich and unambiguous.  This representation should not
    usually be parsed or manipulated by other code.

    In some cases the best information-rich and unambiguous
    representation is a Python expression which evaluates to the
    object, that is, for eval(repr(x)) to return an object which is
    equal to x (given an appropriate environment).  For example, the
    built-in numeric types all have such __repr__() implementations,
    as do the built-in sequence and mapping types (except for cases in
    which a container contains itself).  Where such a representation
    is infeasible or undesirable, this method should return a string
    of the form <...some useful description...>.

    If a class defines __repr__() but not __str__(), then __repr__()
    is also used when __str__() would otherwise be called.

__str__(self)

    Called by the str() built-in function and by the print statement
    to compute a string representation of an object.  The return value
    must be a string object.

    This differs from __repr__() in that it does not specifically
    serve the purpose of providing a programmer with detailed
    information about the object; it may provide a string
    representation of the object for other purposes, as appropriate to
    the object.  Such purposes include display to an application
    end-user, reference in an application log for administrators, or
    providing a string to other code which requires one -- in general,
    purposes for which the programmer-oriented information of
    __repr__() is not appropriate.

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From mwilson at the-wire.com  Thu Apr 24 23:33:18 2003
From: mwilson at the-wire.com (Mel Wilson)
Date: Thu, 24 Apr 2003 23:33:18 -0400
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth>
Message-ID: <+xKq+ks/K/pQ089yn@the-wire.com>

In article <98imn-ust.ln1 at grendel.myth>,
Jim Richardson <warlock at eskimo.com> wrote:
>I have come across a weirdness. I have a number of floats, (lat long
>pairs) which I stuff into a dict, but when I take them out, my nice %2f4
>number, (say, 47.5894) is something long and *almost* right, like
>47.589333333333333339. What gives? I am not doing anything special as
>far as I can tell, just stuffing the numbers into a dict, and recalling
>it later. This is with Python 2.2 on Linux, and no, it's not an original
>Pentium :)

   Chances are that the float hasn't changed, but that
you're printing it in a different way.  You'll see this kind
of difference between `repr(float_value)` and
`str(float_value)` and `"%f"%(float_value,)" and
`"%g"%(float_value,)`.

        Regards.        Mel.



From aleax at aleax.it  Wed Apr 30 12:24:50 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 16:24:50 GMT
Subject: question on the compile built in function.  Filename ignored?
References: <7rSra.706428$S_4.744795@rwcrnsc53>
Message-ID: <mNSra.25571$3M4.711144@news1.tin.it>

Alex wrote:

> I am fairly to new to Python, so please bear with me.
> 
> I am trying to use the built in function compile.  The filename argument
> seems to be ignored no matter what I try.  According to the docs, it
> should
> be used in the output error message.  Here is a simple example:
> 
> Python 2.2.2 (#1, Apr 16 2003, 22:00:16)
> [GCC 3.2.1 20021207 (Gentoo Linux 3.2.1-20021207)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>>> compile(':\n', 'abc', 'exec')
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "<string>", line 1
>     :
>     ^
> SyntaxError: invalid syntax
> 
> 
> From the docs I would expect the line
> 
>   File "<string>", line 1
> 
> to read
> 
>   File "abc", line 1
> 
> 
> What am I missing?

You're missing nothing: this small bug is fixed in Python 2.3 beta 1:

Python 2.3b1 (#1, Apr 26 2003, 10:44:25)
[GCC 3.2 (Mandrake Linux 9.0 3.2-1mdk)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> compile(':\n', 'abc', 'exec')
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "abc", line 1
    :
    ^
SyntaxError: invalid syntax
>>>


I don't know if Python is the only software where a "beta" version
seems to have fewer bugs than the previous "stable" version, but my
personal and unofficial impression of Python 2.3 beta 1 is that the
choice to change the language proper very, VERY little (while allowing
some change in the library and built-ins), and focus on stability and
optimization instead, may indeed have had this unexpected result!-)


Alex




From trentm at ActiveState.com  Tue Apr  1 13:00:00 2003
From: trentm at ActiveState.com (Trent Mick)
Date: Tue, 1 Apr 2003 10:00:00 -0800
Subject: Visual Python without ActivePython
In-Reply-To: <ud6kw45ry.fsf@fitlinxx.com>; from db3l@fitlinxx.com on Wed, Mar 12, 2003 at 04:46:41PM -0500
References: <ud6kw45ry.fsf@fitlinxx.com>
Message-ID: <20030401100000.A9686@ActiveState.com>

[David Bolen wrote]
> I was looking to evaluate Visual Python, but was wondering if anyone
> knew a way to get it to operate with a regular distribution (not
> ActiveState's ActivePython) since it seems to check specifically for
> ActivePython on startup.
> 
> I've already got 5 versions of Python installed on my system (and have
> carefully picked which ones get involved in the few single-use
> registry settings), and would prefer not to need ActivePython just for
> an evaluation (not to mention I'm not sure what registry "adjustments"
> installing it might do for free).  I'd expect my standard 2.1 or 2.2
> installation, including win32all, should be more than sufficient to
> run Visual Python.
> 
> Thanks.

A new version should be coming out shortly that will deal with this
problem: Visual Python will work with ActivePython or python.org's
Python, versions 2.1 and up.

You can try the beta now if you would like to help:
    http://aspn.activestate.com/ASPN/Mail/Message/visualpython-discuss/1588507

Cheers,
Trent

-- 
Trent Mick
TrentM at ActiveState.com




From roy at panix.com  Tue Apr  1 20:54:59 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 01 Apr 2003 20:54:59 -0500
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <bd993a2f.0303310333.5a95ecb3@posting.google.com> <roy-D262FA.09230131032003@reader2.panix.com> <b6aam1$7o6$1@news.island.liu.se> <uk7ef3w3c.fsf@boost-consulting.com> <b6cobh$hck$1@news.island.liu.se> <b6dahn$40kd1$1@ID-169208.news.dfncis.de>
Message-ID: <roy-E744DA.20545901042003@reader2.panix.com>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> wrote:
> The real problem is that when you
> get something *out* of the collection, you want the compiler
> to know statically that it's of whatever type the collection
> is supposed to contain.
> 
> Interfaces don't help you with that at all -- you need a
> notion of parameterized types, which is what C++ templates
> are.

Or, you need a langauge with dynamic typing, like Python :-)



From tungwaiyip at yahoo.com  Thu Apr 24 20:30:30 2003
From: tungwaiyip at yahoo.com (Tung Wai Yip)
Date: Thu, 24 Apr 2003 17:30:30 -0700
Subject: variable scoping problem
Message-ID: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com>

Consider this

>>> x=1
>>> def foo():
...   print x
...
>>> def bar():
...   #print x   <- referenced before assignment error
...   x=2
...   print x
...
>>> foo()
1
>>> bar()
2
>>> print x
1
>>>

What rules in Python explain the scoping of x? Is there a name for the
scope of the first x? 'module' scope? Why is bar() not able to access
x but foo() can?

Wai Yip Tung




From a.schmolck at gmx.net  Fri Apr 25 16:55:03 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 25 Apr 2003 21:55:03 +0100
Subject: variable scoping problem
References: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com> <yfs65p3wf5e.fsf@black132.ex.ac.uk> <ilmiavspd10m814g2gcpfe14dpg4rnh1v4@4ax.com>
Message-ID: <yfsvfx2uwk8.fsf@black132.ex.ac.uk>

Tung Wai Yip <tungwaiyip at yahoo.com> writes:

> Bingo. The 'global' keyword solved my problem. All I'm trying to do is
> to have a global variable to shared value so that one function can
> read the value saved by a previous function.
> 
> For people with C background Python's way is rather different. It
> threats a name differently when it is used for read only versus it is
> written to. I understand your example. But I'm not sure which problem
> is worst, to threat a name differently depends on usage or to
> unintentional overwritten a global variable. 

The answer really has to be: the second one.

I think after brief consideration there can be little doubt: python's
semantics are not really likely to cause any problems *once* you understood
them[1]. If you want to change a global, you just explicitly say so (this has
the nice side effect that it also makes the code easier to read) -- simple[2].

Even if you accidently omit the global statement (not that likely, once you
know about it), the error should become obvious pretty immediately in most
cases (as in your example; in other cases pychecker is likely to find the
problem in static analysis).

But just think what would happen if assigning to a name that also happens to
exist in global scope just silently assigned to that global variable: you
would no longer know what the scope of variable is by just looking at the
function it occurs in! The net-effect would be pretty disastrous: you would be
forced to resort to idiotic naming schemes like "g_x" for "x in global scope"
and "l_x" for "local variable x" to minimize the chance of accidently
overwriting some other function/module/variable.


> C's way has its problem but at least I'm very used to it. The most
> unambigous way is to force people to declare variable first (but I guess
> this would be unpythonic :).

Yes, one could force users to declare local variables instead, but I guess
this would percuss further than one might think (consider how similar classes
and modules currently work, for example).

'as


[1] In respect to this question of alternatives.

[2] The explicitness of a global statement also provides a convenient
    opportunity for second (third and fourth) thoughts (should I *really* be
    using a global here?).




From aleax at aleax.it  Tue Apr 15 18:38:03 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 22:38:03 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it> <slrnb9o9ob.493.rmunn@localhost.localdomain>
Message-ID: <fR%ma.19741$LB6.496324@news1.tin.it>

Robin Munn wrote:
   ...
> I suppose I just illustrated how a try ... finally block might be
> considered "broken and fragile". I tried to use it, and missed some
> subtleties. I'm still not convinced that a "with"-like statement is
> necessary in Python, but I think I just went from -1 to -0 on it.

Hmmm, yes, good point.  I'm about a +0 on "with" or the like - I
_know_ how to write try/finally, but I also know it requires me
some care and attention to get it right, and a well-designed 'with'
might be a snap to use -- OTOH, any language addition is a burden,
intrinsically (more stuff to teach, to learn, &c)... so it's hard
for me to get above +0 for anything that CAN be done reasonably
well already, albeit with some care and fussing!-)


Alex




From peter at engcorp.com  Mon Apr 21 07:48:56 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 21 Apr 2003 07:48:56 -0400
Subject: python dynamic scoping question
References: <79ad5955.0304201513.6ffd4f83@posting.google.com> <b7vg9t01mdo@enews2.newsguy.com> <mailman.1050900693.8645.python-list@python.org> <b800h901rgl@enews3.newsguy.com>
Message-ID: <3EA3DAA8.17689EDE@engcorp.com>

Greg Krohn wrote:
> "Jp Calderone" <exarkun at intarweb.us> wrote:
> > On Sun, Apr 20, 2003 at 07:49:55PM -0500, Greg Krohn wrote:
> > > globals() - Return a dictionary representing the current global symbol
[snip]
> > > Cool?
> >
> >   No.  The right way to do this is:
[snip]
> 
> >   But [it] isn't the answer to the original poster's question, afaict.
> 
> Really? The only differnece I see is 'global x'.

If I understand you correctly, you misunderstood Jp.  He's saying
that *neither* reply, neither yours nor the corrected version, answers 
the original poster's question, for which AFAICT we still don't fully 
understand the rationale ...

-Peter



From duncan at NOSPAMrcp.co.uk  Thu Apr 17 10:06:04 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 17 Apr 2003 14:06:04 +0000 (UTC)
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk>
Message-ID: <Xns9360989556563duncanrcpcouk@127.0.0.1>

Robin Becker <robin at jessikat.fsnet.co.uk> wrote in 
news:XiCE$IAjFmn+Ew43 at jessikat.fsnet.co.uk:

> I get the impression that XP is intended as a design none/test all
> process where the code some grows organically into the requirement
> through bursts of coding and testing. It even seems as though the
> testing is primary.

XP ranks design very highly. Its just that it recognises that coding is an 
aspect of design. 'Big design up front' thinks design is something you do 
once before you start coding, XP thinks that design is continuous 
throughout the process.

> Can XP be reduced to choosing candidates from the pool of programs that
> best satisfy a set of tests? If so experience from GP suggests that the
> tests require care to avoid common pitfalls. As an example a program
> might complete all the tests perfectly, but fail on all other instances
> of data (this commonly happens to overspecialised solutions).

What makes test driven development so powerful is that you write the 
specification while thinking about WHAT the code has to achieve. If you 
write the test after the code you are inevitably thinking about HOW the 
code achieves its results. Writing the test first means your tests relate 
to the specification rather than the implementation.

You are correct to some extent, if the programmer didn't think of 
particular cases when writing the tests, then XP almost demands that the 
code doesn't support those cases. In particular, deliberately writing code 
that fails on all data other than the tested data is a common way to do 
TDD: Write a test, make sure it fails for the expected reason, then write 
minimal code to make it pass (which might involve hardwiring the result), 
make sure the test passes, refactor all the while making sure all tests 
still pass.

The refactoring stage is the one that (hopefully) renders the code as short 
and clean as possible, which includes removing special case results and 
making the code support the general case.

If later in the process you discover a case that you overlooked, then it is 
a simple matter to add a test for the new case and modify the code to 
support it while being sure you aren't breaking other cases.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From bdesth at removethis.free.fr  Wed Apr  2 10:22:13 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 02 Apr 2003 17:22:13 +0200
Subject: Best practice for object attributes?
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk> <3E8AEC44.2090909@removethis.free.fr> <b6eusn$avn$1@nntp0.reith.bbc.co.uk>
Message-ID: <3E8B0025.4090602@removethis.free.fr>

Michael Sparks wrote:
> laotseu wrote:
> 
> 
>>One outstanding advantage of the second 'idiom' is that it's the right
>>way to define *instance* attributes !-) (which is probably what your
>>looking for)
> 
> 
> Given the first idiom seems to create instance attributes by cloning the
> current state of the class attributes, I'm finding it useful is where
> certain instance attributes are unlikely to differ much between
> instances, or changes infrequently related to object creation. 
> 
> For example: if you're creating large numbers of objects which need a
> ballpark timestamp, you could timestamp the class periodically, and
> reduce your number of calls to gettimeofday.
> 
> Another example is in a custom logging class - it has default logging
> levels for most objects defined by the class attributes, but can be
> raised/lowered on a per instance basis, or globally for all new objects
> by changing the class attributes.
> 
> 
>>With the first form you get *class* attributes.  Make 2 instances of
>>Shrubbery1, change the name of first one to 'Brian', print the name of
>>the second one, and then try to understand what is the purpose of the
>>'self' argument in methods.
> 
> 
> I'm aware of the meaning of self :) and know that the first form gives
> you class attributes. The reason I was asking is because I'd noticed
> the following interesting behaviour, which if _intentional_ is useful.
> 
> 
>>>>a=shrubbery1()
>>>>b=shrubbery1()
>>>>a.name,b.name,shrubbery1.name
>>>
> (None, None, None)                  # a.name, b.name exist - presumed
>                                     # references to shrubbery1.name
> 
>>>>a.name="hello"
>>>>a.name,b.name,shrubbery1.name
>>>
> ('hello', None, None)               # Bzzzt! a.name seems to be
>                                     # independent
> 
>>>>b.name="brian"
>>>>a.name,b.name,shrubbery1.name
>>>
> ('hello', 'brian', None)            # Hmm. So is b.name
> 
>>>>shrubbery1.name="Surprise"
>>>>a.name,b.name,shrubbery1.name
>>>
> ('hello', 'brian', 'Surprise')      # Changing class leaves existing
>                                     # instances unaffected.
> 
>>>>a=shrubbery1()                  # Get 2 fresh copies.
>>>>b=shrubbery1()
>>>>a.name,b.name,shrubbery1.name
>>>
> ('Surprise', 'Surprise', 'Surprise') # a.name, b.name cloned from
>                                      # shrubbery1.name
> 
> As far as I can tell the behaviour seems to be that you automatically
> gain instance attributes that clone the state of the class attributes
> at object creation time, which are then independent and can be modified
> via the usual self.<attr> syntax later on. (They cannot however be
> modified via the self.__dict__[<attr>] syntax since they're not stored
> there)
> 
> For reference, this behaviour holds with the following python versions:
>    * python 2.3 alpha 2
>    * python 2.2.1
>    * python 2.2.2 stackless
> 
> Given it seems consistent across versions, and seems useful (changing
> default values for later objects just by updating a class attribute -
> eg an external timestamper perhaps) I was wondering if there's a good
> reason not to use this.
> 

oops ! I should have shut my mouth :(
Ok, sorry, next time I'll try first and only then answer.

I always thought of the first form as a way to define class attributes, 
and never tried to use it the way you did. And effectively, it works as 
you noticed (python 2.2.x), and may be useful sometimes.

Just one more thing I noticed :
 >>>class Toto:
 >>>    name = 'toto'
 >>>    def __init__(self, age)
 >>>        self.age = age
 >>>
 >>>a = Toto(42)
 >>>a.age
42
 >>>a.name
'toto'
 >>>a.__dict__
{'age' : 42}

Until there, ok. Now if we set a.name...
 >>>a.name = 'tutu'
 >>>a.__dict__
{'age' : 42, 'name' : 'tutu'}

So if you try to modify this [class ??? instance ??? or what ???] 
attribute, you in fact create a new instance attribute that hides the 
[class ??? instance ??? whatsoever ???] one.

Do we have a guru 'round here ?-) Could use one...


Laotseu







From sandysj at juno.com  Wed Apr  2 11:55:30 2003
From: sandysj at juno.com (Jeff Sandys)
Date: Wed, 2 Apr 2003 16:55:30 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <3E88C1D8.3070305@motorola.com> <3e8a57d0.55986091@news-server>
Message-ID: <3E8B1602.72BB1A37@juno.com>

try:

C:\python22\pythonw.exe "C:\program files\pysol\pysol.py"

or: do a google search for "download pysol.exe" 
and download the windows executable.


Re: top-posting: Blind people with text to speech converters 
prefer top-posting so they don't have to listen to the same 
stuff over and over again.


BBFoto wrote:
> 
> I copied your statement, changed the path and tried it.
> "Syntax error" was the result.

> <stephen.boulet at motorola.com> wrote:
> 
> >You could fire up a "cmd" console, and type in:
> >
> >   C:\python22\python "C:\program files\pysol\pysol.py"
> >



From Anish.Mehta at enst-bretagne.fr  Mon Apr 14 16:29:12 2003
From: Anish.Mehta at enst-bretagne.fr (Mehta, Anish)
Date: Mon, 14 Apr 2003 22:29:12 +0200
Subject: Viewing Opentype font tables
Message-ID: <3E9B1A18.6070707@antares.enst-bretagne.fr>

Hello Everyone,

Can you please let me know how can i view the tables in the opentype fonts.

Your response will be welcomed

Thanks

Regards,

Anish MEHTA





From skip at pobox.com  Mon Apr 21 12:24:34 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 21 Apr 2003 11:24:34 -0500
Subject: negative lookahead question
Message-ID: <16036.6978.270321.861442@montanaro.dyndns.org>

This re.sub call lives in Lib/stmplib.py as a way to make line endings
canonical: 

    re.sub(r'(?:\r\n|\n|\r(?!\n))', CRLF, data)

This certainly seems to do what's desired, however, it looks overly complex
to me.  First, the non-grouping parens are unnecessary.  Second, I don't
think the negative lookahead assertion is required.  This simpler function
call seems to do the trick:

    re.sub(r'\r\n|\n|\r', CRLF, data)

A simple test case containing a combination of different line endings seems
to yield identical results:

    >>> data = 'line 0\r\nline 1\nline 2\rline 3\r\r\nline 4\n'
    >>> re.sub(r'\r\n|\n|\r(?!\n)',CRLF,data) == re.sub(r'\r\n|\n|\r',CRLF,data)
    True

Is there a case where the negative lookahead assertion will produce correct
results but the simpler regular expression won't?

(FYI, this isn't a performance question, but a readability question.)

Thx,

Skip




From gherron at islandtraining.com  Mon Apr 28 12:53:30 2003
From: gherron at islandtraining.com (Gary Herron)
Date: Mon, 28 Apr 2003 09:53:30 -0700
Subject: Tkinter: Position of windows
In-Reply-To: <b8jllt$ei3$1@lacerta.tiscalinet.it>
References: <b8jllt$ei3$1@lacerta.tiscalinet.it>
Message-ID: <200304280953.30578.gherron@islandtraining.com>

On Monday 28 April 2003 09:43 am, Federico wrote:
> How can I control where a window will appear? I'd like to set the initial
> position of a window...
> Thanks

On the top level window (say root) do things like this:

  set size:      root.geometry('400x300')
  set position:  root.geometry('+100+100')
  set both:      root.geometry('400x300+100+100')

Gary Herron






From tug at wilson.co.uk  Tue Apr 29 15:22:11 2003
From: tug at wilson.co.uk (John Wilson)
Date: Tue, 29 Apr 2003 20:22:11 +0100
Subject: Problem subclassing tuple
References: <08e601c30e67$bda54090$640a0b0a@handel>
Message-ID: <09dd01c30e84$a2738ee0$640a0b0a@handel>

Success!!!

class Holder(tuple):
    def __new__(cls, *args, **kargs):
        return tuple.__new__(cls, args)


is the way to do it for a tuple :))))))

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From point14 at magma.ca  Sun Apr 27 00:45:16 2003
From: point14 at magma.ca (Martin d'Anjou)
Date: Sun, 27 Apr 2003 04:45:16 GMT
Subject: Looping around a simple list of tuples
Message-ID: <3EAB7213.4030209@magma.ca>

I have a list of tuples, each tuple is a (scalar , list)

a=[("r1",[1,2,3]),("r2",[8,9,10]),("r5",[12,14,16])]

for el in a:
    print el
    for r,l in el:
       print "R "+`r`
       print "L "+`l`

('r1', [1, 2, 3])
R 'r'
L '1'
Traceback (innermost last):
   File "<stdin>", line 3, in ?
ValueError: unpack list of wrong size

In my mind, the 'el' implicit variable is a tuple, and the for r,l loop should 
just work but it does not. How come?

Thanks,
Martin




From python at dylanreinhardt.com  Fri Apr  4 12:11:30 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 04 Apr 2003 09:11:30 -0800
Subject: Zope style (was: tuning our site (was: help! advocacy
	resources needed fast))
In-Reply-To: <1049466050.3e8d94c2d048b@mcherm.com>
References: <1049466050.3e8d94c2d048b@mcherm.com>
Message-ID: <1049476289.10845.23.camel@ida.dylanreinhardt.com>

On Fri, 2003-04-04 at 06:20, Michael Chermside wrote:
> I think you just need a minor style change in your python code. I find
> the following about as readable as your DHTML code:
> 
>   my_output = """Hello, %(fname)s.  This is demonstration of MySoft
>            version %(software_version)s.  Just think, %(fname)s, now 
>            you can shift paradigms and think outside the box at 
>            %(company)s without leaving your office.""" % nameMap
>   return my_output

An excellent suggestion.  Thank you.

> 
> Of course, PEP 292 makes this marginally better:
> 
>   my_output = """Hello, $fname.  This is demonstration of MySoft
>            version $software_version.  Just think, $fname, now 
>            you can shift paradigms and think outside the box at 
>            $company without leaving your office.""".sub(nameMap)
>   return my_output
> 
> Now that's really quite readable and maintainable.

Funny... I've been using Python pretty much exclusively for a couple
years now... just *seeing* $var_name syntax gave me a brief case of the
Perl Willies. :-)  But the PEP makes it perfectly obvious why this might
be a good idea anyway.

Dylan





From claird at lairds.com  Tue Apr  8 10:36:29 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 08 Apr 2003 14:36:29 -0000
Subject: Intellectual property and management fashion (was: How to protect python code ?)
References: <mailman.1049758213.26256.python-list@python.org> <slrnb945tu.p4.bignose-hates-spam@rose.localdomain.fake> <5ppka.23067$It5.15751@news2.central.cox.net>
Message-ID: <v95njdh72j0af9@corp.supernews.com>

In article <5ppka.23067$It5.15751 at news2.central.cox.net>,
Steve Holden <sholden at holdenweb.com> wrote:
			.
			.
			.
>>   "How could it possibly be of benefit to a business to *pay* to become
>>   dependent on software they don't control?"
>>
>> Thus, a rhetorical question perhaps.  The "What kind of business" was
>> not a query for what *industry* the business is in, but rather in a
>> spirit of "what kind of idiot".
>>
>
>Of course, the businesses that make a good living out of proprietary
>software encourage all sorts of FUD about open source. Consequently, large
>organizations implement rules forbidding the use of open source software
>products because "they aren't supported" when in fact the support for some
>open source products is actually superior to that of most proprietary
>products.
>
>The fact of the matter is that most companies would prefer to pay for a
>black-box proprietary product because ultimately there's assumed to be a
>responsible (and accountable) organization behind it. If they bothered to
>read the licenses, of course, they'd realize that the producer's liability
>is usually limited to the purchase price of the software.
			.
			.
			.
Hey, me, I only know what the mass media tell me; that, for
example, if I buy from Microsoft, I can expect to end up in
court <URL: http://www.eweek.com/article2/0,3959,985552,00.asp >:
"There's one thing Microsoft officials ... agree on:  SQL Server
users should comb through code to determine if they're at risk of
being dragged into court.  . . .  If companies determine there
may be grounds for patent infringement [from use of Microsoft
products as documented by Microsoft], Microsoft officials are
advising them to take it up with" the patent holder.

As a non-lawyer, I understand this to support the proposition
that, when I buy from Microsoft, I first need to investigate
all of Microsoft's affairs, to be certain the product I intend
to buy doesn't infringe on any patents.  So:  as a Microsoft
customer, my liability is essentially unlimited, while Micro-
soft's "is usually limited to the purchase price of the 
software."

It's good to know that.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From jason00taylor at hotmail.com  Wed Apr 16 18:21:45 2003
From: jason00taylor at hotmail.com (Jason Taylor)
Date: 16 Apr 2003 15:21:45 -0700
Subject: LinearAlgebra vs LinAlg.py (access to full LAPACK routines)
Message-ID: <1ada6ef5.0304161421.2c81eeea@posting.google.com>

In the documentation for the LinearAlgebra package it states:

    # This module is a lite version of LinAlg.py module which contains
    # high-level Python interface to the LAPACK library.  The lite
versioho
    # only accesses the following LAPACK functions: dgesv, zgesv,
dgeev,
    # zgeev, dgesdd, zgesdd, dgelsd, zgelsd, dsyevd, zheevd, dgetrf,
dpotrf.

Does LinAlg.py exist? If so, where can it be found?

The reason I want it, is so that I can have access to all LAPACK
routines (If it provides that).



From gregory.p.green at boeing.com  Mon Apr  7 15:54:17 2003
From: gregory.p.green at boeing.com (Green, Gregory P)
Date: Mon, 7 Apr 2003 12:54:17 -0700
Subject: Problem setting COM property using win32com
Message-ID: <E1B4F80EB080194BBAD3092392F511B317F2B7@XCH-NW-13.nw.nos.boeing.com>

I don't think the gen_py class is being used:

print obj: <COMObject SmRecList.SmRecordList>

print obj.__class__: win32com.client.CDispatch

print dir(obj): ['_ApplyTypes_', '_LazyAddAttr_', '_NewEnum', '_Release_', '__AttrToID__', '__LazyMap__', '__call__', '__cmp__', '__doc__', '__getattr__', '__getitem__', '__init__', '__int__', '__len__', '__module__', '__nonzero__', '__repr__', '__setattr__', '__setitem__', '__str__', '_builtMethods_', '_enum_', '_find_dispatch_type_', '_get_good_object_', '_get_good_single_object_', '_lazydata_', '_make_method_', '_mapCachedItems_', '_oleobj_', '_olerepr_', '_print_details_', '_proc_', '_unicode_to_string_', '_username_', '_wrap_dispatch_']


-----Original Message-----
From: Paul Prescod [mailto:paul at prescod.net]
Sent: Monday, April 07, 2003 11:39 AM
To: Green, Gregory P; python-list at python.org
Subject: Re: Problem setting COM property using win32com


Green, Gregory P wrote:
> I am creating the object like so:
> 
>    attrs = win32com.client.Dispatch("SmRecList.SmRecordList")
> 
> I hadn't noticed the dynamic. I guess I can try to force early binding.

Do some introspection on the objects to see whether they are makepy 
objects or dynamic objects. Probably if you print out the __class__ or 
dir() of the objects you'll get a strong hint.

  Paul Prescod





From tweedgeezer at hotmail.com  Tue Apr  8 14:57:03 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 8 Apr 2003 11:57:03 -0700
Subject: Is there any way to make repr(aString) use double quotes?
Message-ID: <698f09f8.0304081057.1ba7041f@posting.google.com>

The topic really says it all -- I'd like to *force* repr(aString) to
use double quotes.  Is that possible?

If it's not, then is it possible to achieve the same thing (everything
escaped properly, etc.) through other means?

Thanks,
Jeremy



From usenet_spam at janc.invalid  Tue Apr 22 18:10:33 2003
From: usenet_spam at janc.invalid (JanC)
Date: Tue, 22 Apr 2003 22:10:33 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com> <e9570f37.0304220625.11312687@posting.google.com>
Message-ID: <Xns93661CA5C015JanC@213.118.32.89>

graham__fawcett at hotmail.com (Graham Fawcett) schreef:

> Presumably a small C program could be written and made available as a
> bootstrap to be placed in binary executable installers (Inno Setup,
> etc) that could detect the absence of a Python environment and
> download/install it prior to final installation of the application.

This can already be done using NSIS and its download plugin ("NSISdl").
<http://nsis.sourceforge.net/>

-- 
JanC

"Be strict when sending and tolerant when receiving."
RFC 1958 - Architectural Principles of the Internet - section 3.9



From matt at teaphoo.fsnet.co.uk  Thu Apr  3 08:04:16 2003
From: matt at teaphoo.fsnet.co.uk (Matt Russell)
Date: 3 Apr 2003 05:04:16 -0800
Subject: generalize connect in High Level Network Classes ?
Message-ID: <459f9b68.0304030504.48f0a940@posting.google.com>

Hi all,

recently i have wanted a timeout on my ftp connection using ftplib,
and have submitted a patch on sf to allow me to do this.

However, I recently came across httplib that has (almost) duplicate
code, that also would benifit from using the new timeout features in
2.3

What does anyone think of a base for these two classes?

e.g

class NetworkConnection(object):
    def settimeout(self, t):
        """Proxy method to set the timeout on the underlying socket
object"""
        try:
            float(t)
        except ValueError, e:
            raise e
        else:
            try:
                self.sock.settimeout(t)
            except AttributeError, e:
                raise Exception("Connect has not been called. pass
timeout=n to connect instead")

    def gettimeout(self):
        """Get the  timeout if any of the underlying socket object"""
        return self.sock.gettimeout()
         
    def connect(self):
        """Connect to the host and port specified in __init__."""
        msg = "getaddrinfo returns an empty list"
        for res in socket.getaddrinfo(self.host, self.port, 0,
                                      socket.SOCK_STREAM):
            af, socktype, proto, canonname, sa = res
            try:
                self.sock = socket.socket(af, socktype, proto)
                if self.debuglevel > 0:
                    print "connect: (%s, %s)" % (self.host, self.port)
                self.sock.connect(sa)
            except socket.error, msg:
                if self.debuglevel > 0:
                    print 'connect fail:', (self.host, self.port)
                if self.sock:
                    self.sock.close()
                self.sock = None
                continue
            break
        if not self.sock:
            raise socket.error, msg
    


Removing connect from httplib and replacing the connect method in
ftplib with

def connect(self, host, port, timeout):
    NetworkConnection.connect((host,port), timeout=timeout)
    self.af = af
    self.file = self.sock.makefile('rb')
    self.welcome = self.getresp()
    return self.welcome

Any takers?
If so will submit/revise a patch

Matt Russell



From a.schmolck at gmx.net  Wed Apr 16 07:51:56 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 16 Apr 2003 12:51:56 +0100
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <yfsistfb5n9.fsf@black132.ex.ac.uk> <slrnb9pekp.vg8.ramen@lackingtalent.com>
Message-ID: <yfs3ckibqxv.fsf@black132.ex.ac.uk>

Dave Benjamin <ramen at lackingtalent.com> writes:

> > No argument here. I tend to think python would have been a better language if
> > it had included something like smalltalk-like blocks from day 1 (I still
> > really like python).
> 
> Are those anything like code blocks in Ruby? I think that's a cool idea...

Yes, in pretty much the same way that butter is like margarine.


'as



From zdys at zdys.de  Tue Apr 15 03:37:24 2003
From: zdys at zdys.de (luigi monaco)
Date: Tue, 15 Apr 2003 09:37:24 +0200
Subject: vim configuration help
References: <P4Oma.78974$rd4.2975295@twister.austin.rr.com>
Message-ID: <b7gcm8$go4$07$1@news.t-online.com>

On Tue, 15 Apr 2003 08:58:55 +0200, Randall Smith wrote:

> I'm learning Python and I've been looking for a good IDE to use with it.
>     I haven't found an IDE that has text editing capabilities close to
> Vim/Emacs.
> 
> Much of my previous programming is with Emacs/PHP and I love the
> auto-indenting, tab-indenting, well thought-out key strokes, and the
> ability to ssh into a server and run Emacs.
> 
> I'd like to try out Vim while writing Python code (I'm using Webware).
> Right now, the default settings using color coding, have an 8 space tab
> and no autoindenting.  I would like to have a tab space of 2 or 4 spaces
> and have auto-indenting.
> 
> I have no idea how to configure Vim.  I'm a total newbie.  Can someone
> please help me get started.  I'm looking to configure Vim for Python.
> Autoindenting and 4 space tabs are high priority now.
> 
> Thanks,
> 
> Randall
Hi Randall, 

I had the same problems when I started with 'vim'. So here some initial
stuff to faster do some finger practices ;-)

<ESC>set tabstop= 4<CR>
# set tabstop to 4 space length

<ESC>set ai<CR>
# set autoindent

<ESC>set noai<CR>
# set no autoindent

<ESC>:mkv<CR>
# write vim configuration file to actual directory
# copy .vimrc to you ~/ directory, so it will be used by default. 

<ESC>set nu<CR>
# enable line numbering

<ESC>syntax enable<CR>
# enable syntax highlighting

<ESC>:%s/<PATTERN>/<SUBST>/
# to substitute <PATTERN> with <SUBST>

<ESC>dd<CR>
# delete actual line

<ESC>start#,end#d<CR>
# to delete from line# start to line# end NOTE copied to buffer

<ESC>P
# insert buffer above actual position

<ESC>p
# insert buffer below actual position

hth

Luigi aka zdys



From kriley1 at hotmail.com  Wed Apr 16 09:58:34 2003
From: kriley1 at hotmail.com (Ken R.)
Date: 16 Apr 2003 06:58:34 -0700
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com>
Message-ID: <1e5ab84a.0304160558.2719521c@posting.google.com>

Peter van Kampen <news at datatailors.xs4all.nl> wrote in message news:<slrnb9q90a.mhg.news at datatailors.com>...
> In article <mailman.1050457621.2105.python-list at python.org>, Mike C.
> Fletcher wrote:
> > Andrew Koenig wrote:
> > 
> > Okay, so now we have "The Practical Coder's Guide to big-O Notation" :) 
> > .  We rock ;) ,
> > Mike
> 
> You do, you know...c.l.p really does.
> 
> I think I almost 'get it'. Except who or what decides what Big-O a
> certain algorithm has. Is that by 'hunch' or by running benchmarks or
> are there still other ways to recognize 'Big-O's (except for the
> formal computations that I understand are not very common)? 
> 
It's been nearly forever since my compsci algorithms class so be
gentle if I'm glaringly wrong :). Big O notation is a way to convey
the magnitude of work that it takes for an algorithm to do it's job
(ignoring any constant time). So, for a sufficiently large N, the work
required for an algorithm is something like N or NlogN etc.. No one
decides what order an algorithm is, it is an inherant property of the
algorithm.

If you execute quicksort on a wide variety of sample sets of different
sizes then measure and plot the processor time vs sample size, it will
approximate the graph of NlogN. Therefore the algorithm has order
NLogN. The same will apply to bubble sort (order N^2), etc.

I haven't seen the referenced Big-O posting but there is a decent
explanation of  Big-O and sorting algorithms here
http://www.autoobjects.com/Home/Teaching/CmpE_126/CmpE_126_Lectures/Sortings/sortings.html#Algorithm

Ken



From peter at engcorp.com  Wed Apr 16 08:58:26 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 08:58:26 -0400
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee>
Message-ID: <3E9D5372.494AB8D9@engcorp.com>

Vahur Lokk wrote:
> 
> I wanted to build a simplistic? document management system. For example,
> I've got folder for letters out as well as project-based folders. I would
> save the letter in Letters Out folder and a symlink created in respective
> project folder etc. Programs that create such files are different - office
> suite, mail client (attachments coming in), browser (files from web).
> 
> I have learned programming some 15 years ago (BASIC) and wanted to re-learn
> it and am looking for some small project for a start.
> File system drivers do not sound like a newbie project, though. Polling file
> system for changes did come to my mind as a solution, but it seems like
> very resource-hungry way to do things.

One of the most important things to learn as a programmer is this,
as stated by various people but most popularly attributed to Knuth:
"Premature optimization is the root of all evil in programming."
Basically, worrying about whether something is "resource-hungry" 
before you've actually tested it out is premature optimization.

Why not write a little program that *does* poll for changes 
(maybe just limiting it to one folder as you describe above), 
and see how bad it is.  I assure you, unless you try to poll on 
the order of every tenth of a second, it won't be too resource-
hungry unless you've implemented it very poorly.

Meanwhile, you've learned more about your problem, you've learned 
more about Python, you've learned more about programming, and 
you've quite possibly got yourself a working solution with
adequate performance, and *that's more than most software projects
ever produce*!

-Peter



From a.schmolck at gmx.net  Wed Apr  9 15:44:29 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 09 Apr 2003 20:44:29 +0100
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org>
Message-ID: <yfs3ckrihgi.fsf@black132.ex.ac.uk>

Beni Cherniavsky <cben at techunix.technion.ac.il> writes:

> It also describes some implementation details - not that I believe we
> want dynamic scope...  

Why not? Amongst other things dynamically scoped global variables seems the
only sane way to me to have global variables. Otherwise, everytime you screw
around with some global variable temporarily you have to go through all sorts
of pain making sure you clean it up after you're done (even if something went
wrong) and you'd better not be multithreaded. Compare:

# simple, single threaded case
try:
   global precision
   old = precision
   precision = 10e-10
   dosomething()
finally:
   precision = old

(let ((precision 10e-10)) (dosomething))

> Note that the listed motivation is mainly implemention simplicity and
> effeciency. 

Well yes, but emacs lisp is decades old. GC was also inefficient 20 years ago,
now memory handling by hand is not only a pain but also most likely slower.

> I read somewhere that this is no longer true but for a long time dynamic
> scope was considered faster even for compiled code because so much effort of
> LISP hackers went into optimizing it in clever ways :-)...

I think that's wrong. Common Lisp has both dynamic scope (for globals and
everything and locals your explicitly declare so) and lexical scope and I
think lookup for the dynamically scoped variables is slower.

'as



From aleax at aleax.it  Fri Apr 25 11:59:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 15:59:46 GMT
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com>
Message-ID: <SXcqa.6797$K35.186991@news2.tin.it>

Peter Hansen wrote:

> Alex Martelli wrote:
>> 
>> Peter Hansen wrote:
>> > throwaway than the ones I "know" are not.  If I'm wrong... no big deal:
>> > I rewrite as a real script with proper tests.  Since the script was
>> > only twenty or thirty lines of code (any more and it could not possibly
>> > be throwaway, right?) it isn't really a big deal.
>> 
>> Wrong.  "Throwaway" means "I [think I] am going to need to run this
>> only once".  Whether it's 30 or 60 lines has nothing to do with it.
> 
> This is getting us nowhere.

Quite possibly.  You seem to be interested in the ability to predict
"throw-away-ness" of a few specific individuals, which, to me, seems
quite a marginal issue.  I'm more interested in the general issue.


> I'm asserting that I have actually seen "throwaway" scripts, even
> predicted their imminent development, and subsequently have actually
> thrown them away.
> I'm not asserting I do this with 100% accuracy.  I'm asserting that
> I very infrequently do decide that a script will be throwaway, and
> that in those rare cases I often turn out to be right, and often
> enough that it seems worth continuing to make such predictions in
> order to save myself some time.  These I believe are facts about *me*.

I compliment you on your accuracy (assuming your perception of it
is accurate).  However, even if you, and/or Aahz, are in fact
exceptional individuals who are indeed able to make such specific
predictions reliably, I would still caution all readers against
drawing any conclusions from this, unless they have the good fortune
of working at your side and being able to benefit from your advice
about this specific matter.

I've met quite a few people who made specific claims about their
ability to predict short-term stock market movements, horse-race
results, whether a script is going to be thrown away, what parts of
a program never yet profiled are its bottlenecks, whether it will
rain next weekend, and other notoriously hard-to-predict events.

I have never been able to actually verify that any of these claims
could be substantiated.  I may suspect on general grounds that some
people DO have such unusual predictive abilities, but until and
unless I'm able to rely on specific advice by somebody proven to
be such an individual, I consider it wiser and more prudent to act
as if such predictions were as unreliable as they notoriously are
held to be in popular opinion.

So, I don't play the stock-market nor bet on horse racing, etc.
And in fields in which I think I can teach and advise others, I
suggest to others that they behave similarly.  In programming, for
example, I suggest NOT relying on one's intuition about where the
bottlenecks will be, nor on that about what's throw-away code
and what's going to be used again.



> You're asserting that you have never accurately predicted this, or
> at least not accurately enough to make prediction worthwhile, or
> at least that even when you predict correctly it you discount the
> prediction and proceed on the assumption that you are wrong.  (Or
> something like that... I'm unsure it's actually relevant to my point.)

If your point is strictly about YOUR predictive abilities, then any
assertion about MINE, or those of the general population, cannot be
relevant to it.


> Neither of us can assert anything about the other's real situation,
> nor presumably about Aahz' (though I admit I did... maybe I was wrong).
> 
> I can't see any point in continuing the discussion, can you?

As long as the general readership has followed our assertions and
can draw its own conclusions, sure.  I highly recommend to that
readership to keep a healthy skepticism about such predictions and
always consider the possible costs of being wrong vs the costs of
"covering" for possibly wrong predictions.

"rm throwaway.py" has clearly miniscule benefits in most cases, and
the potential cost is obvious -- if you're wrong, and next week
need that same functionality again, you're going to have to write
and possibly debug it again.  Leaving your hard-disk strewn with
dusty old code, however, is quite sub-optimal too -- any single
given script matters little, but cumulatively they may make things
untidy indeed.  Keeping your "probably throwaway scripts" in their
own folder[s] is a good move -- and so is writing SOME docs about
what it IS that they do, even just a couple lines worth of docstring.

And that is just the start.  The *mindset* of "I _think_ this will
be run once - but just possibly it will be run again" leads to
some documenting, testing, and a modicum of care about structure,
which collectively make it *more* likely that some part of the
script, if not the whole, DOES then later prove worth reusing.


Alex




From dsada at asd.com  Wed Apr  2 06:01:25 2003
From: dsada at asd.com (John Does)
Date: Wed, 02 Apr 2003 11:01:25 GMT
Subject: i cant remember that python regex viewer
Message-ID: <pan.2003.04.02.11.03.07.614824@asd.com>

i cant remember that python regex viewer:

i do remember:

- it being on sourceforge (which has a horrible search feature,
and its not on fm)

- it having a gui

- it being able to spit out usable regex's

- it was written in python

- it had a funny name


google has not helped.





From exarkun at intarweb.us  Tue Apr 15 20:35:24 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 20:35:24 -0400
Subject: Why functional Python matters
In-Reply-To: <slrnb9p4oo.vg8.ramen@lackingtalent.com>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com>
Message-ID: <20030416003524.GA7550@meson.dyndns.org>

On Tue, Apr 15, 2003 at 11:17:26PM -0000, Dave Benjamin wrote:
> In article <yfs7k9vcu2p.fsf at black132.ex.ac.uk>, Alexander Schmolck wrote:
> [snip]
> 
> >> keyword-argument trick wasn't really that bad, once you got used to it. In
> >> any case, with dynamic scoping, this problem has completely disappeared.
> >                  ^^^^^^^^^^^^^^^
> > not quite.
> 
> Am I using the wrong terminology here? I'm specifically talking about the
> named-argument hack to get around Python's previous lack of nested scopes.
> This specific problem, as far as I know, is gone now. Am I missing something?
> 

  I think you meant "partial lexical scoping" ;)  Dynamic scoping is what
Python used to do.  Since it isn't quite lexical scoping, it's best to just
stick with the term "nested scopes".

> [snip]

  Jp

-- 
#!/bin/bash
( LIST=(~/.sigs/*.sig)
  cat ${LIST[$(($RANDOM % ${#LIST[*]}))]}
  echo -- $'\n' `uptime | sed -e 's/.*m//'` ) > ~/.signature
-- 
 up 26 days, 21:02, 4 users, load average: 0.25, 0.31, 0.63




From claird at lairds.com  Sat Apr  5 09:01:53 2003
From: claird at lairds.com (Cameron Laird)
Date: Sat, 05 Apr 2003 14:01:53 -0000
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <b6f66v$28k$1@slb3.atl.mindspring.net> <v8m9a4nk4dpjf7@corp.supernews.com> <a626bbd7.0304031149.7abe84c0@posting.google.com>
Message-ID: <v8toehsbm8bob5@corp.supernews.com>

In article <a626bbd7.0304031149.7abe84c0 at posting.google.com>,
robin <escalation746 at yahoo.com> wrote:
>claird at lairds.com (Cameron Laird) wrote:
>
>> SOAP and such are just concessions to commercial misunder-
>> standings about what business needs.
>
>I would like to know more about what you mean by this.
			.
			.
			.
I'll do this in an abbreviated form.

SOAP's s'posed to be the "Simple Object Access Protocol".
It's defining document begins, "SOAP is a lightweight
protocol ..." 

It's a bad sign that it's fiction from the start.  SOAP
isn't lightweight or simple, and it doesn't particularly
access objects.

SOAP is an RPC implementation.  I'm fine with RPC, and I
like SOAP--'hope I get more jobs to do it during the next
year.  However, I think commercial experience has demon-
strated adequately that RPC isn't safe in the hands of
the programming fraternity at large.  It's something medi-
ocre programmers do wrong.

Businesses *think* they want their development crews to
standardize on an RPC, and XML is a good thing, isn't it?,
but they're wrong.  RPC across organizational boundaries
turns out to be somewhere between difficult and a disaster.

Businesses that are happy with SOAP are actually using it
as a messaging service for asynchronous transmission of
XMLified documents with business content.

I repeat:  for a mixture of correct and incorrect reasons,
XML, RPC, and so on are believed to be good things for
business.  People conclude that SOAP must be a super-
technology, solving whole layers of issues at once.  It's
not.  It's OK, and, with enough support from Microsoft,
IBM, Oracle, and a few others, it certainly can dominate.
In truth, though, it answers the wrong question.

Python's own Paul Prescod has plenty to say about SOAP's
technical flaws.  Check out <URL: http://
mail.python.org/pipermail/xml-sig/2002-February/007183.html >
and other references available through <URL: http://prescod.com >.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From skip at pobox.com  Sun Apr  6 08:27:49 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sun, 6 Apr 2003 07:27:49 -0500
Subject: PYTHONPATH question
In-Reply-To: <E1927CG-0007k5-00@fast.aub.edu.lb>
References: <E1927CG-0007k5-00@fast.aub.edu.lb>
Message-ID: <16016.7493.311731.974816@montanaro.dyndns.org>

    Ali> shouldn't i be able to read PYTHONPATH directly like i read any other env 
    Ali> variable? because i am saying
    Ali> echo $PYTHONPATH
    Ali> and nothing is being displayed.

That just means PYTHONPATH isn't set currently.  In your ~/.profile, simply
set

    PYTHONPATH=<colon-separated list of directories>
    export PYTHONPATH

and you should be okay.

Skip




From mwm at mired.org  Fri Apr  4 13:03:57 2003
From: mwm at mired.org (Mike Meyer)
Date: Fri, 4 Apr 2003 12:03:57 -0600
Subject: PyQt QTableItem inheritance...
In-Reply-To: <200304041022.06728.phil@river-bank.demon.co.uk>
References: <x77kaejf9b.fsf@guru.mired.org>
	<mailman.1049223123.616.python-list@python.org>
	<x7istwiwzm.fsf@guru.mired.org>
	<200304041022.06728.phil@river-bank.demon.co.uk>
Message-ID: <16013.51469.237864.683801@guru.mired.org>

In <200304041022.06728.phil at river-bank.demon.co.uk>, Phil Thompson <phil at river-bank.demon.co.uk> typed:
[ My code elided. ]

> There are a number of problems here...

I know that createEditor was probably badly broken. As far as I could
tell, it wasn't running. That's because I was expecting createEditor
to be called when the item was added to the table, not when you
clicked on it.

thank you very much for the corrected code.

	<mike
-- 
Mike Meyer <mwm at mired.org>		http://www.mired.org/consulting.html
Independent Network/Unix/Perforce consultant, email for more information.




From mcfletch at rogers.com  Tue Apr 15 14:17:02 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 15 Apr 2003 14:17:02 -0400
Subject: Printing Columns
In-Reply-To: <ABEE81BE08ADD311B6A1009027DE908D16A949D6@conmsx03.corp.acxiom.net>
References: <ABEE81BE08ADD311B6A1009027DE908D16A949D6@conmsx03.corp.acxiom.net>
Message-ID: <3E9C4C9E.5080303@rogers.com>

Straightforward approach (will be inefficient if you really do approach 
infinity though ;) )...

 >>> def indices( total, columns = 3 ):
...     rows, remainder = divmod( total, columns )
...     if remainder:
...         rows += 1
...     return [
...         range( row, total, rows )
...         for row in xrange(rows)
...     ]
...
 >>> def columnMajor( items, columns = 3, format = '%8d', joint="|"):
...     for row in indices(len(items),columns):
...         print joint.join([format%item for item in row])
...
 >>> columnMajor( range(23),4)
       0|       6|      12|      18
       1|       7|      13|      19
       2|       8|      14|      20
       3|       9|      15|      21
       4|      10|      16|      22
       5|      11|      17
 >>>

HTH,
Mike

Lindstrom Greg - glinds wrote:
...

>Given a list of length N with 0<=N<=infinity and a number of columns, C (C>0
>and may be less than N), how do we print out the elements of the list in the
>above (column major?) form?
>  
>
...

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From andrew-pythonlist at puzzling.org  Thu Apr 17 04:13:44 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Thu, 17 Apr 2003 18:13:44 +1000
Subject: Teaching : Python, Scheme, Java...
In-Reply-To: <roy-8E37D1.09554617042003@malibu.unice.fr>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr>
Message-ID: <20030417081344.GA21962@frobozz.local>

On Thu, Apr 17, 2003 at 09:55:46AM +0200, Jean-Paul Roy wrote:
[..]
> Aside from that points, a personal question as a Schemer who reads the 
> Python primer:
> - how do you ask if a list is empty ? Is the following ok ?
>       if L == [] :

Usually you'd just do
    if not L:

You could also do
    if len(L) == 0:

> - how do you ask if the list has only one element, with time 0(1) ? The
>   Scheme/Lisp analog is (null? (cdr L))

len(L) == 1

> - more generally, is the length of a list a builtin ? Is the L[k] access
>   time O(1) with a list ? Probably not. And with a tuple ?

Python lists are like C++ and Java vectors, they're stored contiguously in
memory, so indexing is O(1).  Same with tuples.  

-Andrew.





From michaels at rd.bbc.co.uk  Wed Apr  2 10:12:36 2003
From: michaels at rd.bbc.co.uk (Michael Sparks)
Date: Wed, 02 Apr 2003 16:12:36 +0100
Subject: Best practice for object attributes?
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk> <3E8AEC44.2090909@removethis.free.fr>
Message-ID: <b6eusn$avn$1@nntp0.reith.bbc.co.uk>

laotseu wrote:

> One outstanding advantage of the second 'idiom' is that it's the right
> way to define *instance* attributes !-) (which is probably what your
> looking for)

Given the first idiom seems to create instance attributes by cloning the
current state of the class attributes, I'm finding it useful is where
certain instance attributes are unlikely to differ much between
instances, or changes infrequently related to object creation. 

For example: if you're creating large numbers of objects which need a
ballpark timestamp, you could timestamp the class periodically, and
reduce your number of calls to gettimeofday.

Another example is in a custom logging class - it has default logging
levels for most objects defined by the class attributes, but can be
raised/lowered on a per instance basis, or globally for all new objects
by changing the class attributes.

> With the first form you get *class* attributes.  Make 2 instances of
> Shrubbery1, change the name of first one to 'Brian', print the name of
> the second one, and then try to understand what is the purpose of the
> 'self' argument in methods.

I'm aware of the meaning of self :) and know that the first form gives
you class attributes. The reason I was asking is because I'd noticed
the following interesting behaviour, which if _intentional_ is useful.

>>> a=shrubbery1()
>>> b=shrubbery1()
>>> a.name,b.name,shrubbery1.name
(None, None, None)                  # a.name, b.name exist - presumed
                                    # references to shrubbery1.name
>>> a.name="hello"
>>> a.name,b.name,shrubbery1.name
('hello', None, None)               # Bzzzt! a.name seems to be
                                    # independent
>>> b.name="brian"
>>> a.name,b.name,shrubbery1.name
('hello', 'brian', None)            # Hmm. So is b.name
>>> shrubbery1.name="Surprise"
>>> a.name,b.name,shrubbery1.name
('hello', 'brian', 'Surprise')      # Changing class leaves existing
                                    # instances unaffected.
>>> a=shrubbery1()                  # Get 2 fresh copies.
>>> b=shrubbery1()
>>> a.name,b.name,shrubbery1.name
('Surprise', 'Surprise', 'Surprise') # a.name, b.name cloned from
                                     # shrubbery1.name

As far as I can tell the behaviour seems to be that you automatically
gain instance attributes that clone the state of the class attributes
at object creation time, which are then independent and can be modified
via the usual self.<attr> syntax later on. (They cannot however be
modified via the self.__dict__[<attr>] syntax since they're not stored
there)

For reference, this behaviour holds with the following python versions:
   * python 2.3 alpha 2
   * python 2.2.1
   * python 2.2.2 stackless

Given it seems consistent across versions, and seems useful (changing
default values for later objects just by updating a class attribute -
eg an external timestamper perhaps) I was wondering if there's a good
reason not to use this.

Thanks,


Michael.
-- 
Michael.Sparks at rd.bbc.co.uk    
British Broadcasting Corporation, Research and Development
Kingswood Warren, Surrey KT20 6NP

This message (and any attachments) may contain personal views
which are not the views of the BBC unless specifically stated.





From max at alcyone.com  Mon Apr  7 20:10:00 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 07 Apr 2003 17:10:00 -0700
Subject: exceptions == errors?
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
Message-ID: <3E921358.64ADEFF5@alcyone.com>

Mark Harrison wrote:

> Are exceptions always considered errors, or there the same feeling
> as in C++ that exceptions can be used for non-error situations?

Even moreso than in C++; in Python, exceptions are often used internally
for flow control.  By default an iteration protocol will iterate over an
object with the subscription operator, using successively high indices,
until it hits an IndexError, at which point it stops:

>>> class C:
...  def __getitem__(self, index):
...   if index < 10:
...    return index**2
...   else:
...    raise IndexError
... 
>>> [x for x in C()]
[0, 1, 4, 9, 16, 25, 36, 49, 64, 81]

> Therefore, I'm thinking of structuring the code like this:
> 
>         while 1:
>                 try:
>                         read request
>                         process request
>                 catch MyException,e:
>                         send response contained in e
>                 catch Exception,e:
>                         send "internal error" response
> 
> Am I setting myself up for trouble if I follow this route?
> All advice gratefully received.

The only potential flaw that I see here is one with a standard "catch
all" phrase.  In C++ this generally isn't a problem, since you have a
compile phase and so syntax errors and typos will get caught before the
program runs, but in a dynamic language like Python you can find
yourself in headaches where something's failing and you have no idea
why, until you finally track it down and realize you spelled a variable
wrong, which was generating a NameError, but was being caught by your
catch Exception: ... clause.

Once you're cognizant of the problem, it's not as much of an issue, but
it might bite you the first time (or few times) before you get used to
it.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Blood is the god of war's rich livery.
\__/ Christopher Marlowe
    Computer science / http://www.alcyone.com/max/reference/compsci/
 A computer science reference.



From peter at engcorp.com  Wed Apr  2 15:01:07 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 15:01:07 -0500
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com> <b6esse$pge$1@heckle.cc.uregina.ca> <3E8B0030.E3B0DE00@engcorp.com> <b6f3o3$t1n$1@panix3.panix.com>
Message-ID: <3E8B4183.3DF3F57E@engcorp.com>

Aahz wrote:
> 
> In article <3E8B0030.E3B0DE00 at engcorp.com>,
> Peter Hansen  <peter at engcorp.com> wrote:
> >
> >I believe Google could provide more detail with a search using
> >"stainless roman numerals" or something like that.
> 
> You're using Roman numerals for underwear?

_What_ underwear?


-too-much-information-ly y'rs,
 Peter



From roy at panix.com  Wed Apr 23 10:58:07 2003
From: roy at panix.com (Roy Smith)
Date: 23 Apr 2003 10:58:07 -0400
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz>
Message-ID: <b869lv$aa2$1@panix2.panix.com>

Frantisek Fuka  <fuka at fuxoft.cz> wrote:
>I want a simple test to see if str contains any occurences of substr. I 
>am currently using this:
>
>import string
>if string.find(str,substr) + 1:
>     ...do something...
>
>But it somehow looks "ugly" to me. Is there more "standard" way to test 
>this?

My first comment would be that adding 1 to the result of string.find
is quite ugly.  The code is much more readable if you simply test for
equality to -1.

Next, the string module is rather out-dated.  It's nicer to use the
new (in 2.0, I think) string methods.  So you end up with:

if str.find (substr) == -1:
   do something

which I think is pretty clean.




From afriere at yahoo.co.uk  Tue Apr 15 06:22:56 2003
From: afriere at yahoo.co.uk (Asun Friere)
Date: 15 Apr 2003 03:22:56 -0700
Subject: A python server
References: <15f445fb.0304140751.170d410e@posting.google.com>
Message-ID: <38ec68a6.0304150222.65857897@posting.google.com>

santacruz at southern.edu (dansan) wrote in message news:<15f445fb.0304140751.170d410e at posting.google.com>...
Ooops, you're wanting to write a server, not a client.  Just can my
previous suggestion.



From ckea25d02 at sneakemail.com  Mon Apr 14 02:00:19 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 14 Apr 2003 18:00:19 +1200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>
Message-ID: <b7di5e$dd6gj$2@ID-169208.news.dfncis.de>

Mads Orbesen Troest wrote:
> The 
> documentation says __str__ is for the, quote, informal representation; 
> whereas __repr__ is for the, quote, official representation. Furthermore, 
> __reptr__ should, quote, look like an expression.

I don't find those guidelines particularly useful except
in fairly special cases. More generally, the rule of thumb
I use is that __str__ is for the "normal" output of the
program, whatever that might be, whereas __repr__ is for
debugging output.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From kosh at aesaeion.com  Wed Apr 16 04:53:56 2003
From: kosh at aesaeion.com (kosh)
Date: Wed, 16 Apr 2003 02:53:56 -0600
Subject: Why functional Python matters
In-Reply-To: <20030416004537.GB7550@meson.dyndns.org>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <slrnb9p3vo.vg8.ramen@lackingtalent.com> <20030416004537.GB7550@meson.dyndns.org>
Message-ID: <200304160253.56636.kosh@aesaeion.com>

On Tuesday 15 April 2003 06:45 pm, Jp Calderone wrote:
>   As many of us know, you can do this in 2.2:
>
>     reduce(int.__add__, [1, 2, 3])
>
>   Unfortunately, this breaks when ints get promoted to longs, because you
> can't pass longs to int.__add__!  Similarly:

Thankfully this is trivially solved. I just tested with python 2.3a2+

import operator
reduce(operator.__add__, range(6000000))

which gives

17999997000000L





From cmkleffner at gmx.de  Wed Apr 16 04:24:04 2003
From: cmkleffner at gmx.de (cmkl)
Date: 16 Apr 2003 01:24:04 -0700
Subject: book on Python for Fortran programmers
References: <3064b51d.0304150458.595afc30@posting.google.com> <vhbo9vgvsqjg8l7s0aluijrran56cg8mlm@4ax.com> <pan.2003.04.15.20.55.25.783394@mn.rr.com> <490na.26787$T34.799064@news2.tin.it>
Message-ID: <3b091a1c.0304160024.7a764356@posting.google.com>

I like Rick Muller's comprehensive tutorial:
http://www.wag.caltech.edu/home/rpm/python_course/
because it is targeted as a Python tutorial for
scientific programmers. Things like the numeric python
extension, extending python with C-modules, OpenGL
are treated with usuable and understandable examples.

Use this as an appetizer to get an idea what can be
done in Python in numeric computing.

Eric Jones presentation:
http://www.scipy.org/site_content/tutorials/scipytutorial.pdf
is an impressive overview over numeric and scientific
python modules.

But to 'learn' the python syntax and semantics from the 
beginning a good book like the mentioned 'Python in a Nutshell' 
is definitly the best idea. You find this book at: 
http://www.oreilly.com/catalog/pythonian/

Regards,

Carl

Alex Martelli <aleax at aleax.it> wrote in message news:<490na.26787$T34.799064 at news2.tin.it>...
> Anna wrote:
> 
> > On Tue, 15 Apr 2003 12:15:18 +0000, Michael Prager wrote:
> > 
> >> beliavsky at aol.com wrote:
> >> 
> >>>Python has been suggested by several posters to comp.lang.fortran as a
> >>>scripting language worth learning. What are some good book(s) and web
> >>>site(s) to get an experienced Fortran 95 programmer started in Python?
>  ...
> > Well, if you're looking for up-to-date, from the reviews I've seen,
> > _Python in a Nutshell_ would be worth looking at for this. I've gotten the
> > impression that it's particularly suited to experienced programmers.
> 
> I am, of course, quite biased in favour of the Nutshell (guess why;-),
> but I do think it might be particularly helpful to Fortran programmers.
> 
> Being a Fortran programmer IS, after all, a substantial part of my own
> background -- Fortran was my first language, the one I wrote my thesis in,
> the main one I taught when I served as a university professor, etc, etc).
> 
> If you're interested in numeric programming, as many Fortran programmers
> are likely to be, the chapter on Numeric might be helpful, for example.
> (Do also look at the feedback about it that was posted on this NG -- I
> haven't yet examined that feedback but it seems to be pointing out a
> couple of important points that have to be added to that chapter!!!).
> 
> I would suggest having a look at the sample chapter, and table of
> contents, that you can get from the O'Reilly site (or amazon.com, etc).
> 
> The chapter that the editor chose to make available as a sample is an
> early but crucial one, summarizing the essentials of the language
> itself (except for exceptions, modules/packages, and OO stuff, which
> each have their own chapter) -- so, I think it might help you judge
> whether the book can best meet your needs.
> 
> 
> Alex



From mwilson at the-wire.com  Wed Apr  2 10:52:19 2003
From: mwilson at the-wire.com (Mel Wilson)
Date: Wed, 02 Apr 2003 10:52:19 -0500
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org>
 <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <zcwi+ks/Kj3I089yn@the-wire.com>

In article <slrnb8lre3.15s.do-not-spam-ben.hutchings at tin.bwsint.com>,
Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> wrote:
>In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer wrote:
><snip>
>>     3.  At most three Is, Xs and Cs may appear in any given literal.
><snip>
>
>OK cleverclogs, how do we write 39?

XIL



in Roman-2003


        Regards.        Mel.




From altis at semi-retired.com  Fri Apr 11 11:15:36 2003
From: altis at semi-retired.com (Kevin Altis)
Date: Fri, 11 Apr 2003 08:15:36 -0700
Subject: Promoting Python
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
Message-ID: <EXAla.454$1A.20758@news.uswest.net>

"Thomas Guettler" <pan-newsreader at thomas-guettler.de> wrote in message
news:pan.2003.04.11.09.17.46.771369.890 at thomas-guettler.de...
> On Fri, 11 Apr 2003 05:34:35 +0200, Kevin Altis wrote:
>
> > One of the hot topics at PyCon was Promoting Python.
> > I would like to invite you to join the marketing-python list if you're
> > interested in contributing ideas
> > ...
>
> [cut]
>
> I think python is a very good language for education. You could get in
> contact with local schools.

Thanks for the suggestion Thomas. Kirby Urner, who is quite active on
edu-sig has already agreed to lead the charge in promoting Python in
education. Suggestions can be passed onto Kirby via the edu-sig mailing list
and as progress is made it will get posted to edu-sig and the
marketing-python lists.

  http://mail.python.org/mailman/listinfo/edu-sig

The first step is collecting information on what is already being done with
Python. We have a pretty big list of links to schools, educators, and
education-related projects using Python. Additional info is always welcome.

  http://mail.python.org/pipermail/edu-sig/2003-April/002765.html

Kirby and I will be organizing that into categories to make it easier to
use. Kirby is also editing the edu-sig page on python.org.

  http://www.python.org/sigs/edu-sig/

ka





From andy at reportlab.com  Mon Apr 14 20:54:27 2003
From: andy at reportlab.com (Andy Robinson)
Date: Tue, 15 Apr 2003 01:54:27 +0100
Subject: win32 COM test question
References: <mailman.1050209239.15961.python-list@python.org>
Message-ID: <ellm9vkm89172me07efgem5aaoc1grdbgn@4ax.com>

>http://www.hps1.demon.co.uk/users/andy/pyvb/spamsrv.htm
>
>And next time try to search for the answer with google ( python com
>visual basic )

Aaarrgggghhhh!

As the author, thanks for the polite feedback, but that is REALLY old.
I no longer work for the company that hosts it and I thought they took
it down at my request 3 years ago when PPW32 came out. Be 
warned that many things in it might not work any more...

The general concepts are fine but I would recommend anyone
reading it to get the examples from the book, which are only
3 years old instead of 5-6,. They are on Mark's page at
  http://starship.python.net/crew/mhammond/ppw32/

Thanks,

Andy Robinson





From pinard at iro.umontreal.ca  Fri Apr 11 19:00:08 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 11 Apr 2003 19:00:08 -0400
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <Pine.GSO.4.51.0304092053190.1222@techunix.technion.ac.il>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	<3E9040CF.2040606@removethis.free.fr>
	<m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com>
	<m27ka7aae0.fsf@localhost.localdomain>
	<6ee58e07.0304061722.2472b3d0@posting.google.com>
	<3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk>
	<3E92E00B.6000708@kyborg.dk>
	<mailman.1049814375.24609.python-list@python.org>
	<b6vq8r$9qev7$1@ID-169208.news.dfncis.de>
	<Pine.GSO.4.51.0304092053190.1222@techunix.technion.ac.il>
Message-ID: <oqhe947i87.fsf@titan.progiciels-bpi.ca>

[Beni Cherniavsky]

> To clarify the issue completely, the emacs lisp manual contains a
> must-read section:

> http://www.gnu.org/manual/elisp-manual-21-2.8/html_node/elisp_147.html#SEC147

> [...] It also describes some implementation details - not that I believe
> we want dynamic scope... [...]

Oh, I would not say that the above text is clarifying the issue completely.
It rather gives a flying glimpse over the problem, within the context of a
discussion about a project considering a rewrite of Emacs in Python.

For example, the shallow binding implementation is not really acceptable for
buffer local variables, even if most non buffer local variables use it.  Not
to speak about the complexity associated with various defaulting or
initialising mechanism for variables in buffers.

Bindings are also different, despite simultaneous, for functions and for
non-functions.  Over these two, properties allow one to simulate many
concurrent flavours of bindings to symbols.  Evaluation contexts may also be
fully switched through usage of alternate obstacks.

Now, key bindings are another kind of binding which does not work exactly
like variables, yet the project quoted above might be wise to consider as
well.  For example, key bindings may be associated with limited text areas
within buffers, and while the cursor moves between windows, or over text
within windows, key bindings may change quite often.  Emacs uses heaps (some
call them priority stacks) for tracking these changes.

So, all in all, in real practice, we have a flurry of different types and
qualities of bindings with Emacs, and a variety of implementation techniques
for them.  It would not be wise to underestimate the problem, or to quickly
dismiss this complexity as being gratuitous.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From uptonothing at yahoo.com  Mon Apr  7 14:05:05 2003
From: uptonothing at yahoo.com (Mike)
Date: Mon, 7 Apr 2003 14:05:05 -0400
Subject: Updating CVS
References: <5Nhka.139$N5.25540@mencken.net.nih.gov> <pan.2003.04.07.19.52.22.351156.788@thomas-guettler.de>
Message-ID: <l5jka.141$N5.25497@mencken.net.nih.gov>

"Thomas Guettler" <pan-newsreader at thomas-guettler.de> wrote in message
news:pan.2003.04.07.19.52.22.351156.788 at thomas-guettler.de...
> On Mon, 07 Apr 2003 18:35:13 +0200, Mike wrote:
>
> > Hello,
> >
> > I wrote the following python web script to update a CVS directory and I
> > don't receive any errors when executing it; however, the directory
> > doesn't update.
> >
> > #!/usr/bin/python
> >
> > import cgi, os, urllib, sys
>
> Since you import cgi, I assume you are writing a cgi script.
>
> Which web server do you use?
>
> Try running the script from the shell.
>
> Have a look at the error log of you web server.
>
> You can run the script in a wrapper like this. This
> result in error pages which are displayed in the browser:
>
> #!/usr/bin/python
> import cgi
>
>
> def main():
>     print "Content-Type: text/html\n"
>     try:
>                 import quizpy
>                 print quizpy.main()
>     except:
>                 print "<html><body><pre>"
>                 cgi.print_exception()
>                 print "</pre></body></html>"
>
> main()
>
>
>  thomas
>
> BTW: The subject of your post is not very good, since it is a cgi problem.
>
> --
> Thomas Guettler <guettli at thomas-guettler.de>
> http://www.thomas-guettler.de

The line "sys.stderr = sys.stdin" redirects any errors to the browser. My
question is less cgi and more cvs. I guess I didn't state it clear enough. I
just want to know why the previous script doesn't update my CVS directory.
Is it because I trying to run in from a web script and not as myself being
logged in? I'm using an Apace web server running on Solaris.

Mike





From martin.vonloewis at hpi.uni-potsdam.de  Tue Apr 22 15:14:40 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Tue, 22 Apr 2003 21:14:40 +0200
Subject: pygettext but no msgfmt
References: <b83snm$1ms$1@news.tu-darmstadt.de>
Message-ID: <b844b0$92e$1@zeppelin.rz.uni-potsdam.de>

> I tried to use gettext for internationalizing but i can't find the msgfmt
> tool.
>
> I have debian woody with python2.1 and python 2.2 installed.
> In the debian archive the file msgfmt.py isn't listed either.
>
> How can i convert the po files then?

I recommend to install the gettext package, and use msgfmt(1).

If you want to use msgfmt.py instead,
you'll find it in /usr/share/doc/python2.1/examples/Tools/i18n/msgfmt.py.

Regards,
Martin





From peter at engcorp.com  Fri Apr 18 09:28:02 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 18 Apr 2003 09:28:02 -0400
Subject: Testing databases/forms/etc (was Re: How many of you are Extreme 
 Programmers?)
References: <1147e466.0304160630.5b510724@posting.google.com> <8b2d5523.0304180059.722de25a@posting.google.com>
Message-ID: <3E9FFD62.138634FF@engcorp.com>

Qvx wrote:
> 
> All work is done in Oracle
> database and Forms and they don't fit nicely with TDD (at least to
> me). How do you test a form? How do you test a lengthy database job
> whose input depends on dozens of tables and which modifies/creates
> records across dozens of tables?
> 
> I might be in a position to influence some things here at my company
> so any help would be appreciated. Any thoughts on this? Books? Links?

You might check the archives of the extremeprogramming or 
testdrivendevelopment mailing lists at groups.yahoo.com.  These
issues have been discussed there in the past, although I don't think
I would say I saw clear, concise instructions on how to avoid all
the problems in this area.  You might get some good ideas to get
past your mental blockage about how this could work.

-Peter



From cben at techunix.technion.ac.il  Thu Apr 17 18:19:08 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Fri, 18 Apr 2003 01:19:08 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <b7fpmq$ht9r$1@ID-169208.news.dfncis.de>
Message-ID: <20030417221908.414AA22093@techunix.technion.ac.il>

In comp.lang.python, you wrote:
>Paul Foley wrote:
>> Don't forget the ~20000 other variables that you'd have to pass around
>> as explicit arguments to *every single function*.
>
>Paul, you're suffering from an extremely bad case of
>tunnel thinking.
>
>You wouldn't pass around 20,000 variables. You would
>pass around a very few objects (in most cases probably
>just one, i.e. a buffer) from which everything else you
>need can be gotten to if and when you need it.
>
It's hard to know which objects you need to pass.  Most commands only touch
the current buffer but some touch the frame configuration, the global
variables and just about anything in emacs.  You would pretty much need to
pass the whole emacs object :-).

Perhaps it's possible to specify the required objects with the interactive
call parameter specification mechanism but tracking that would be non-trivial
when manully doing calls.  Imagine that some function you call from your
command suddenly is changed to sometimes require access to the window
configuration, which it didn't do before.  Now you have to change the call to
it to include this and perhaps change your own signature to include it if you
don't currently.

It's the same issue as capabilities-based security with the "principle of
least authority" - nice idea but painful to track precisely...

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From beau at nyc-search.com  Tue Apr  1 18:22:59 2003
From: beau at nyc-search.com (Python Programmers, NYC)
Date: Tue, 01 Apr 2003 23:22:59 GMT
Subject: JOB: Python Programmers, NYC
References: <mailman.1049234284.8188.python-list@python.org>
Message-ID: <3E8A1F24.831B47A@nyc-search.com>

> Python Programmers, NYC
> http://www.nyc-search.com/jobs/python.html




From no at spam.invalid  Tue Apr 29 17:17:55 2003
From: no at spam.invalid (Russell E. Owen)
Date: Tue, 29 Apr 2003 14:17:55 -0700
Subject: Python 2.3b1: RuntimeError using rexec
References: <b8mh4l$2oo0$1@nntp6.u.washington.edu> <just-53D0EC.21532729042003@news1.news.xs4all.nl>
Message-ID: <b8mq64$23o6$1@nntp6.u.washington.edu>

Thank you very much. That sounds much better than just a bare "eval".

I could probably also get pickle to do the job, also. A brief inspection 
suggests it puts repr(dict) in some simple boilerplate (at least for 
trivial dicts such as mine -- consisting only of strings and numbers). 
But I'd sure have to learn more before I'd believe that. Your safe eval 
seems much safer from that angle.

-- Russell

In article <just-53D0EC.21532729042003 at news1.news.xs4all.nl>,
 Just <just at xs4all.nl> wrote:

>For ...(converting a string repr of  a simple dict back to a dict)... I often use
>
>def safeEval(data):
>   """A safe replacement for eval."""
>   return eval(data, {"__builtins__": {}}, {})
>
>Exactly _how_ safe it is I don't know, but it's safe enough for my 
>purposes...



From nirmalkannan at hotmail.com  Thu Apr 24 02:21:57 2003
From: nirmalkannan at hotmail.com (N.K)
Date: 23 Apr 2003 23:21:57 -0700
Subject: Python Thread Question
References: <bd993a2f.0304170534.245aa2d5@posting.google.com> <mailman.1050587609.6119.python-list@python.org> <b7meoo$mf6$1@panix2.panix.com> <5b04353b.0304220351.2034a26@posting.google.com> <mailman.1051039477.30652.python-list@python.org>
Message-ID: <5b04353b.0304232221.787d294e@posting.google.com>

So it seems a Optimal design for crawler is non-blocking sockets +
Ability to run more process/threads if required.i.e,

- Store all URLS to be fetched in queue, 
- One or more process/thread reading que and performing IO using non
blocking sockets
- Decide on number of process/threads is depending on situation 

Comments please,

Thanks in advance,
NKC

Jp Calderone <exarkun at intarweb.us> wrote in message news:<mailman.1051039477.30652.python-list at python.org>...
> On Tue, Apr 22, 2003 at 04:51:20AM -0700, N.K wrote:
> > aahz at pythoncraft.com (Aahz) wrote in message news:<b7meoo$mf6$1 at panix2.panix.com>...
> > > 
> > > This isn't quite true on a multi-CPU box.  Because I/O releases the GIL,
> > > the OS can schedule threads to run on all available CPUs.  So it's
> > > certainly possible for a threaded app to outrun non-blocking I/O.
> > 
> > 
> >   Heard that new Intel Processors support Hyper Threading ( HT)
> > Architecture from which multi-threaded apps can benefit from. Is that
> > true?
> 
>   Some (All?) P4s have this technology.  Practically speaking, this means
> applications see the system as having multiple processors.  More than one
> thread can run at once.  So yes, multi-threaded apps benefit from this in
> much the same way they benefit from traditional SMP systems.
> 
>   Jp
> 
> -- 
> Seduced, shaggy Samson snored.
> She scissored short.  Sorely shorn,
> Soon shackled slave, Samson sighed,
> Silently scheming,
> Sightlessly seeking
> Some savage, spectacular suicide.
>                 -- Stanislaw Lem, "Cyberiad"



From exarkun at intarweb.us  Wed Apr 30 11:47:50 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 30 Apr 2003 11:47:50 -0400
Subject: How to call built in function 'open' from inside object that has 'open' method?
In-Reply-To: <b8onf8$b7d$1@panix2.panix.com>
References: <db81c492.0304300526.7b8d3c86@posting.google.com> <mailman.1051711257.436.python-list@python.org> <b8onf8$b7d$1@panix2.panix.com>
Message-ID: <20030430154750.GB26204@meson.dyndns.org>

On Wed, Apr 30, 2003 at 10:43:52AM -0400, Aahz wrote:
> In article <mailman.1051711257.436.python-list at python.org>,
> Fredrik Lundh <fredrik at pythonware.com> wrote:
> >"vadim" <vf at interna.ru> wrote:
> >>
> >> I am new to Python and do not know the answer to very simple how to
> >> 'open' from inside another object that has 'open' method. I know that
> >> in C++ I should use following syntax ::open.
> >>
> >> More details on reasons. I have met this problem trying to write a
> >> simple script use ActiveScripting inside IE.  I am trying "file =
> >> open(...) but because context window object has 'open' method that
> >> just opens window it is called first.
> >
> >Unless you're typing "file = self.open(...)", the class you're in has
> >nothing to do with it.
> 
> Not quite true; what happens with
> 
>     class C:
>         def open(self, name):
>             f = open(name)
> 
> (Which is how I interpreted vadim, possibly incorrectly.)

  The builtin open is called.

  There would be a problem with this class, though:

    class C:
        def open(self, name):
            ...

        f = open('foo')

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 41 days, 11:04, 5 users, load average: 0.14, 0.11, 0.03




From tug at wilson.co.uk  Tue Apr 29 11:55:21 2003
From: tug at wilson.co.uk (John Wilson)
Date: Tue, 29 Apr 2003 16:55:21 +0100
Subject: Problem subclassing tuple
Message-ID: <08e601c30e67$bda54090$640a0b0a@handel>

Given this class:

class Holder(tuple):
    def __init__(self, *values):
        tuple.__init__(self, values)


When I try to create an instance:

Holder(1, 2)

I get:

TypeError: tuple() takes at most 1 argument (2 given)

It looks like the constructor for the superclass is being called rather than
the __init__ method on Holder.

I'm obviously missing something here. Do I have to supply a __new__ method?
Any help would be appreciated.

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From mlichten at harris.com  Tue Apr 22 14:31:30 2003
From: mlichten at harris.com (Lichtenstein, Mark (Contractor))
Date: Tue, 22 Apr 2003 14:31:30 -0400
Subject: FW: passing a C struct to python
Message-ID: <EFAA78F77E74D411A0F700508BDFEAD607D3E88B@mlbmx9.corp.harris.com>


-----Original Message-----
From: Lichtenstein, Mark (Contractor) 
Sent: Tuesday, April 22, 2003 2:13 PM
To: 'Rajarshi Guha'
Subject: RE: passing a C struct to python


I suggest you place the object in the interpreter class. I will attach some java
code to show how this is done. See interp.set statements:

/* %filename%.java
 * Copyright Harris Corporation, 2002. All rights reserved.
 */
package com.harris.netboss.shared.model.rules.commands;

import com.harris.netboss.shared.model.rules.ICommand;
import com.harris.netboss.shared.model.rules.IContext;
import com.harris.netboss.shared.model.rules.PythonFactory;
import com.harris.netboss.shared.model.rules.commands.SetTimer;
import javax.management.Notification;
import org.python.util.PythonInterpreter;
import org.python.core.PyDictionary;
import org.python.core.PyObject;
import java.io.InputStream;
import java.util.Map;
import org.python.core.PyList;
import org.python.core.PyTuple;
import java.util.HashMap;

/**
 * PythonAdaptor ilements the Adaptor pattern.  
 * It implements an ICommand interface, repackages the arguments 
 * and invokes a jython script.
 *
 * @author Mark Lichtenstein
 * @version %I%, %G%
 * @since NextGen
 * @see com.netboss.harris.shared.model.rules
 */
public class PythonAdaptor implements ICommand {

/** execute: executes the jython interpreter.
 * @param arg[] The input argument to this command. (i.e. Jython Script, URL...)
 * @param arg[0] - an InputStream or String of jython
 * @param arg[1...] - any argument that is a java class.
 * @param event The Indication that triggered this command to be invoked.
 */
    public void execute(Notification event, Object[] arg)
    {
        // Get a new or existing PythonInterpreter from PythonFactory
        IContext context = (IContext) arg[0]; // set context from object array
        PythonInterpreter interp = PythonFactory.getInstance(context);
        PyObject localsObject;
        SetTimer setTimer = new SetTimer(); // Create object for Jython
        // Get the local variables in a PyObject if there is a context map
        if (context != null) {
           localsObject = (PyObject) context.get("LOCALS");
           if (localsObject != null)
              interp.setLocals(localsObject);
        }  // end if
	// Pass variables to interpreter's local namespace so that 
	// they may be accessed in a Jython (.py) script
	interp.set("event",(Object) event); 
	interp.set("context",(Object) context); 
        interp.set("setTimer",(Object) setTimer);
	interp.set("args",(Object) arg); 
        if (arg[1] instanceof String)  { // Interpret a string
            interp.exec((String)arg[1]); // run the string
        }  // end if
        if (arg[1] instanceof InputStream) { // execute a file
           InputStream is = (InputStream) arg[1];
           interp.execfile(is);
        }  // end if
        // After executing the jython command(s) we need to restore the
        // context map using any local variables found in the local namespace.
        if (context != null) {
            localsObject = interp.getLocals(); // Get local namespace variables
            context.put("LOCALS",(PyObject)localsObject);
        }
    }  // end of execute method

    /** toMap: Extract a Python dictionary into a Java Map.
    @param interp The Python interpreter object
    @param pyName The id of the python dictionary
    */
    public static Map toMap(PythonInterpreter interp, String pyName){
      PyList pa = ((PyDictionary)interp.get(pyName)).items();
      Map map = new HashMap();
      while(pa.__len__() != 0) {
        PyTuple po = (PyTuple)pa.pop();
        Object first = po.__finditem__(0).__tojava__(Object.class);
        Object second = po.__finditem__(1).__tojava__(Object.class);
        map.put(first, second);
      } // end of while loop
      return map;
   }

}

-----Original Message-----
From: Rajarshi Guha [mailto:rajarshi at presidency.com]
Sent: Tuesday, April 22, 2003 1:57 PM
To: python-list at python.org
Subject: passing a C struct to python


Hi,
  I've recently looked into calling some Python functions from my C code.
I've worked out the ways of calling a Python function
(PyObject_CallMethodObjArgs etc). However these are OK if I want to pass
a list, tuple, string etc.

However in some cases I might want to pass  an 'object' (which would be a
C struct) to the Python function. Alternatively, the Python function
might require me to pass an instance of some class. Am I correct in
thinking that these two problems are equivilant?

And my question is: what do I need to do to pass structs/classes etc? I
tried looking for pointers in the Embedding section of the Python manual,
but I could'nt really find anything :(

Any pointers would be appreciated

Thanks,
-- 
http://mail.python.org/mailman/listinfo/python-list




From mertz at gnosis.cx  Tue Apr 15 13:31:48 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 15 Apr 2003 13:31:48 -0400
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org>
 <b7gikh$s66$1@news1.wdf.sap-ag.de>
Message-ID: <EIEn+kKkXc9B092yn@gnosis.cx>         

"Daniel Dittmar" <daniel.dittmar at sap.com> wrote previously:
|Doesn't cStringIO exist for this purpose?  Now that extension classes
|can be subclassed, adding += and str should be enough for most purposes.
|Perhaps it's even possible to add a general __getattr__ to forward all
|string methods to the inner string (which doesn't exist and has to be
|created for each such request).

cStringIO.StringIO cannot be subclassed (at least through 2.3a2).  Maybe
it could be implemented in a subclassable way, but it is not now.  The
Python version can be subclassed, but presumably it is much slower (not
in a big-O sense, but as practical times).

In the abstract, StringIO has much of the behavior I would want; but it
always feels kind cumbersome to use.  I usually find that building and
joining a list of strings feels less heavy.

Assuming StringIO were to get some new methods, there are quite a few
you'd want to worry about.  .__iadd__() is one.  .__repr__() and/or
.__str__() might be desireable to make "print sio" more friendly.  And
slicing.  And string methods like .upper() or .find().  (OK, Dittmar
mentions the __getattr__() approach, but again, that is a big time
cost).

Yours, Lulu...

--
 mertz@  _/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: \_\_\_\_    n o
gnosis  _/_/             Postmodern Enterprises            \_\_
.cx    _/_/                                                 \_\_  d o
      _/_/_/ IN A WORLD W/O WALLS, THERE WOULD BE NO GATES \_\_\_ z e





From andrew-pythonlist at puzzling.org  Sun Apr 27 01:56:43 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 15:56:43 +1000
Subject: Forgetting "()" when calling methods
In-Reply-To: <698f09f8.0304262115.386c2571@posting.google.com>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com>
Message-ID: <20030427055643.GL26530@frobozz.local>

On Sat, Apr 26, 2003 at 10:15:03PM -0700, Jeremy Fincher wrote:
> Frantisek Fuka <fuka at fuxoft.cz> wrote in message news:<b8cadp$1fj9$1 at ns.felk.cvut.cz>...
> > When I try to call methods, I sometimes forget to include the 
> > parentheses. Instead of:
> > 
> > if object.isGreen():
> > 	do something...
> > 
> > i sometimes write:
> > 
> > if object.isGreen:
> > 	do something...
> 
> Perhaps the answer is to remove the __nonzero__ method on
> functions/methods.  Logically, what purpose does it serve?

For that matter, removing __nonzero__ from functions still doesn't help this
case:

    sys.exit

So it's of pretty minimal benefit.

-Andrew.





From mlichten at harris.com  Tue Apr 22 09:53:05 2003
From: mlichten at harris.com (Lichtenstein, Mark (Contractor))
Date: Tue, 22 Apr 2003 09:53:05 -0400
Subject: python registry file
Message-ID: <EFAA78F77E74D411A0F700508BDFEAD607D3E884@mlbmx9.corp.harris.com>

I am using python 2.1 and I would like to find out the best way to set the
registry file for the python.cachedir property.
For users running NT I can set this property on a Unix shared drive to C:\temp
and this works fine. In this case each
user will use their personal drive and there would be no users step on each
other writing to the packages directory.
However, On Solaris (Unix), we are worried that Python may not be set up for
multi-user use for the python.cachedir
files. We can go to version 2.2 if there is a better multi-user solution in that
version.

What is everyone else out there doing with this setting?

Thank you,

Mark Lichtenstein
Harris Corporation




From fog at initd.org  Wed Apr  2 10:04:00 2003
From: fog at initd.org (Federico Di Gregorio)
Date: 02 Apr 2003 17:04:00 +0200
Subject: [Psycopg] ANNOUNCE: dropping 1.0.x series
Message-ID: <1049295839.974.138.camel@momo.initd.org>

with the release of psycopg 1.1.3 i am officially dropping the 1.0.x
series.

if enough people ask, i'll backport a couple of fixes from 1.1.3 to
1.0.15.3 and make available a tarball but there will be no announce or
further support.

1.1.3 is as much as stable as 1.0.15.2, please switch to it, if you can.

federico

-- 
Federico Di Gregorio
Debian GNU/Linux Developer                                fog at debian.org
INIT.D Developer                                           fog at initd.org
   Lasciate che i furetti vengano a me. -- Maria Luisa Benedetta Panzani
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030402/9c3fcfcb/attachment.sig>

From peter at engcorp.com  Thu Apr 17 09:07:19 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 17 Apr 2003 09:07:19 -0400
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk>
Message-ID: <3E9EA707.4DB53582@engcorp.com>

Robin Becker wrote:
> 
> I get the impression that XP is intended as a design none/test all

Definitely not "design none".  Programmers with greater experience,
with better design skills, are consistently better than lesser
programmers, even when using XP.  XPers believe the abilities of
both are enhanced by using XP, so that they are more effective
than when not using it, but if it were truly "design none" I believe
you would expect to see a monkey program as capably as a senior
programmer, except for the part about typing.  (Monkeys, as it's
well known, are good at typing Shakespeare, and little else.)

> process where the code some grows organically into the requirement
> through bursts of coding and testing. It even seems as though the
> testing is primary.

Yes, definitely.

> This sounds remarkably similar to what we commonly call genetic
> optimisation or evolution strategy selection. There a fitness criterion
> is applied to randomly selected elements of a pool of solutions, with

The key is that XP is _not_ applying "randomly selected elements".
The experience and abilities of the programmer are heavily involved
in the process.

> Do XPers combine differing solutions etc etc?

Although one might see a surface similarity in some aspects of XP
with evolutionary strategies (and because evolution probably shows through 
in many parts of the world), it might seem there are more links between
the two than there really are.  Maybe one could consider XP to be
sort of a *directed* evolution, but it's far from random.  The design
that arises from working in an XP style has been described as emergent,
but the meme of "emergent properties" is not inherently related to 
evolution.

Interesting food for thought though.

-Peter



From tdelaney at avaya.com  Tue Apr 15 17:45:26 2003
From: tdelaney at avaya.com (Delaney, Timothy C (Timothy))
Date: Wed, 16 Apr 2003 07:45:26 +1000
Subject: Complexity of standard Python data structures
Message-ID: <338366A6D2E2CA4C9DAEAE652E12A1DE4ABA26@au3010avexu1.global.avaya.com>

> From: Michael Hudson [mailto:mwh at python.net]
> 
> I'm afraid your line of questioning slightly confuses me.  The Python
> language is not specified, in any precise sense of the word, anywhere.
> For one thing, it's evolving.  If some super new implementation of
> hash tables appeared[1] that had somewhat different algorithmic
> behaviour but were considered superior, they might well turn up in
> Python 2.5 (say) and *if* the closest thing Python has to a spec (the
> language reference) described the algorithmic nature of various
> operations (which it doesn't) it would then be changed.

One of the big advantages of this is that is assists greatly in trying things out in different implementations, etc.

For example, PyPy can implement something in the simplest way that *works* (matches the CPython semantics) early in the development cycle.

After getting much more of the system to work, the PyPy developers can optimise those parts that need it.

Likewise, the CPython developers can put something in early which does what is needed, in order to build infrastructure for a new feature. It can later be reworked to give performance improvements, if needed.

Tim Delaney




From johnroth at ameritech.net  Tue Apr 15 16:22:01 2003
From: johnroth at ameritech.net (John Roth)
Date: Tue, 15 Apr 2003 16:22:01 -0400
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
Message-ID: <v9oq71itajkh39@news.supernews.com>

"Dave Benjamin" <ramen at lackingtalent.com> wrote in message
news:slrnb9on0q.sho.ramen at lackingtalent.com...

[snip long post about functional constructs]

I agree completely. I got turned on to the existing functional
constructs while I was doing a TDD challenge. I managed
to do that with exactly one if statement and no loops because
of the functional constructs.

However, losing apply() isn't a big deal because there is
an exactly equivalent construct. I'd hate to see the rest
of them go away, especially since the (possible) implementation
of PEP 308 will remove one of the restrictions on lambdas.

Also notice that there are a set of new functional constructs
in 2.3 alpha.

John Roth
>
> Thanks for listening,
> Dave





From staschuk at telusplanet.net  Thu Apr 17 15:06:57 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 13:06:57 -0600
Subject: Python and Schools
In-Reply-To: <_Fzna.27993$LB6.643892@news1.tin.it>; from aleax@aleax.it on Thu, Apr 17, 2003 at 03:23:38PM +0000
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050529915.4666.python-list@python.org> <_Fzna.27993$LB6.643892@news1.tin.it>
Message-ID: <20030417130657.C7968@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
> Steven Taschuk wrote:
>    ...
> > trade-off as being between, say, an algorithm which runs in O(n)
> > time but needs O(n^2) memory to do it, and an alternative which
> 
> I don't think there can be any such algorithm -- in O(n) time,
> the algoritm cannot USE more than O(n) memory, so how could it
> NEED it...?  The reverse, sure, but not this way, I think.

Eppstein's given a nice example here, assuming O(1) memory
allocation.  What I had in mind was something much less clever --
just not counting the time to set up ancillary data structures.

A silly example: comparing two strings for lexicographic order.
The obvious character-by-character algorithm is O(len(s)) (where s
is one of the strings -- doesn't matter which).  If, however, the
set of strings which might enter into comparisons during the
execution of the program is fixed and computable beforehand (in
particular, finite), you might precompute the results of all the
n^2 comparisons.  Then comparing two strings is just a lookup in
the table, which might be just O(1).

Of course, it does take O(n^2) time to set up such a table, and
from that point of view you are of course correct.  In practice
that might not matter.

-- 
Steven Taschuk                  staschuk at telusplanet.net
"Telekinesis would be worth patenting."  -- James Gleick




From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Sat Apr  5 14:42:32 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Sat, 05 Apr 2003 21:42:32 +0200
Subject: How to compile a python script to *.pyo
In-Reply-To: <mailman.1049424492.28480.python-list@python.org>
References: <2977.143.166.82.240.1049308819.squirrel@iamlearning.com>        <3E8C8A57.9080603@gatwick.westerngeco.slb.com> <mailman.1049424492.28480.python-list@python.org>
Message-ID: <3e8f31a5$0$49103$e4fe514c@news.xs4all.nl>

danny at iamlearning.com wrote:
> For Martin,
> Thanks for your advice. That's what I need.
> If this mailing list got more people like you, there are less junk
> messages to read.

Are you serious? If you are, I think you clearly
misunderstood the other replies you got in this thread.

Peter Hansen's remark:
"(My advice: use education of the users to solve your problem, not
fancy technical solutions.)"

is very valuable advice in my opinion, and not 'junk' at all!

--Irmen de Jong




From http  Sat Apr 26 17:45:39 2003
From: http (Paul Rubin)
Date: 26 Apr 2003 14:45:39 -0700
Subject: OO + Python Developers Wanted
References: <mailman.1051183297.29110.python-list@python.org> <yyyznmezekm.fsf@adams.patriot.net> <vl6rn-4f3.ln1@beastie.ix.netcom.com>
Message-ID: <7x1xzplypo.fsf@ruckus.brouhaha.com>

Dennis Lee Bieber <wlfraed at ix.netcom.com> writes:
>         Given those phone number formats, I suspect it isn't supposed to be in 
> the US... Though reformatting would give a US structure
>         0 (207) 333-1518
> but that implies operator intervention as direct dial long-distance 
> would be
>         1 (207) 333-1518
> 
>         The 207 area code (US of course) covers all of Maine...

Given the domain name and looking at the web site, I'd say London UK.



From tjreedy at udel.edu  Wed Apr 30 18:17:06 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Wed, 30 Apr 2003 18:17:06 -0400
Subject: How to make a repeatedly timer with 5ms precision?
References: <mailman.1051736974.9481.python-list@python.org>
Message-ID: <7IOcndHyCOFH1y2jXTWcqA@comcast.com>

"??????" <jxlu at fudan.edu.cn> wrote in message
news:mailman.1051736974.9481.python-list at python.org...
> I am trying to make a repeatedly timer with high precision, the
following program 'countdown.py' is an example which shows the way how
I tried to make it.
> The problem is that it doesn't work properly with tick set to 5ms or
smaller.
> Is there a better way to implement a more precise repeatedly timer?
...
>         print count,time.time()
...
> start=time.time()

Dependig on your system, tou may do better with time.clock() than
time.time().
See Lib Ref 6.9 time -- Time access and conversions

Terry





From db3l at fitlinxx.com  Tue Apr 29 14:40:57 2003
From: db3l at fitlinxx.com (David Bolen)
Date: 29 Apr 2003 14:40:57 -0400
Subject: stupid q-n (wxPy)
References: <Ukxra.33100$7M5.3263531@twister.nyc.rr.com> <3EAEB850.50D9DEB9@engcorp.com> <7hzra.27540$su3.3180311@twister.nyc.rr.com>
Message-ID: <ullxt6tae.fsf@fitlinxx.com>

"Max Khesin" <max at cNOvSisiPonAtecMh.com> writes:

> The only detail I have is that the script was copy-pasted out straight of
> wxPython demo app. There were no specific instructions in the readme for
> running the code so I just gave it a shot. Since wxPython is so popular I
> fugured someone would just tell me "didn't you bla-bla, (idiot)?".
> anyway, thanks.

Ah - it's the point about copying it from the "demo" that would have
helped mentioning in your first post (as opposed to just saying
"examples").  As it was, it's hard to know why the code snippet you
gave was referencing the run module, since it's not a standard module.

The wxPython demo is built around a framework that automates tying
each demo module into the overall demo.  If you look in the same demo
directory where you found this example, you'll find the run.py that
you probably didn't copy over to where you are experimenting.

Note that run.py is not part of standard Python, nor is it something
that is documented as part of wxPython.  It's just part of the
wxPython demo code base.

My guess is that you don't really want to use this code (nor the
run.py module) since that's geared specifically towards the way the
wxPython demo presents itself.  Instead, run the wxPython demo, find
the module that you're interested in, and then borrow the actual code
that implements the behavior to put in your own module, and ignore the
boilerplate framework that ties the module into the demo display.

--
-- David
-- 
/-----------------------------------------------------------------------\
 \               David Bolen            \   E-mail: db3l at fitlinxx.com  /
  |             FitLinxx, Inc.            \  Phone: (203) 708-5192    |
 /  860 Canal Street, Stamford, CT  06902   \  Fax: (203) 316-5150     \
\-----------------------------------------------------------------------/



From kemu at sdf-eu.org  Sat Apr 19 13:59:39 2003
From: kemu at sdf-eu.org (Geirgat Jonas)
Date: Sat, 19 Apr 2003 17:59:39 +0000
Subject: cgi parser question ?
Message-ID: <pan.2003.04.19.17.59.39.393264@sdf-eu.org>

Hello,
maybe this isn't the right place to ask this question I know....
I have written a basic HTTP server in python now I would like it 
to have python CGI support only I haven't found how CGI works
is it like this a user requests a page you just parse it "python page.py"
and send to output to the user ?

any help, links would also help on this subject
regards Jonas



From Olivier.POYEN at clf-dexia.com  Fri Apr 11 05:57:43 2003
From: Olivier.POYEN at clf-dexia.com (POYEN OP Olivier (DCL))
Date: Fri, 11 Apr 2003 11:57:43 +0200
Subject: Getting source code from interpreter-defined function
Message-ID: <8963D6370B323E4AA3241200F0EAF731235628@FCEXVEXM002.dcl.int.dexwired.net>

Hi all,

I'm playing with inspect module, which, BTW, is great for finding useful info on saved object, written object.
With inspect, I can get source code from a function.

Is there anyway I could do the same with "alive" object, without having to wrap the whole python interpreter and create tmpfile for writing session ? (an "alive" object would be an object defined at the interpreter prompt)

That was the question ;  now on to the details:

I'm trying to do some kind of interactive console IDE for PyGtk. I want people to try fast code, binding code to live widget, and then just do a "dump" in order to create the source file. 

So, in a way, I'd like to be able to do, at prompt:

>>> def foo():
...    print "bar"

And then , latter, after playing with it, after having tweaked it *whithout* killing the python interpreter

>>> dump_alive_code(foo):
def foo():
   print "bar"

or at least:
>>> dump_inner_code(foo):
   print "bar"

Insect do already this for written code, by reading the file and retrieving source code. But for object defined in the interpreter... I can't figure that out ! 

I know, I know, why on earth would I want to get that code: just write it down on file and fire 'python myfile.py'
But as I said, I want some kind of interactive console IDE, and I don't want to shutdown python interpreter. Plus, I want to be able to quickly test/bind widget with quick function. 

One solution I thought about is to wrap the whole python interpreter, try saving the code as it is feed, and then, when user asked for a dump, produce it. Overkill ! So much work for almost nothing ! There must be a more elegant solution ! 

Other solution: given any function, even defined in interpreter, I can get function.func_code.co_code, which is byte compiled code for this function. Perhaps I could use it and decompile it. But how....? 


Well, that was long, but good.

Any pointers, any RTFM, any URL will be well received.

Thanks in advance.

OPQ
-------------- next part --------------
-----------------------------------------------------------------

Ce message est confidentiel ; son contenu ne represente en aucun

cas un engagement de la part de Dexia Credit Local ou de Dexia

CLF Banque. Toute publication, utilisation ou diffusion, meme

partielle, doit etre autorisee prealablement par l'emetteur. Si

vous n'etes pas destinataire de ce message, merci d'en avertir

immediatement l'expediteur.



This message is confidential ; its contents do not constitute a

commitment by Dexia Credit Local or Dexia CLF Banque. Any

unauthorised disclosure, use or dissemination, either whole or

partial, is prohibited. If you are not the intended recipient of

the message, please notify the sender immediately.

-----------------------------------------------------------------

Consultez notre site internet www.dexia-clf.fr

La cle d'une gestion optimisee du secteur public local

-----------------------------------------------------------------

From tim_one at email.msn.com  Thu Apr 24 23:30:43 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Thu, 24 Apr 2003 23:30:43 -0400
Subject: Dict handling of floats
In-Reply-To: <98imn-ust.ln1@grendel.myth>
Message-ID: <LNBBLJKPBEHFEDALKOLCIEBAEIAB.tim_one@email.msn.com>

[Jim Richardson]
> I have come across a weirdness. I have a number of floats, (lat long
> pairs) which I stuff into a dict, but when I take them out, my nice %2f4
> number, (say, 47.5894) is something long and *almost* right, like
> 47.589333333333333339. What gives? I am not doing anything special as
> far as I can tell, just stuffing the numbers into a dict, and recalling
> it later. This is with Python 2.2 on Linux, and no, it's not an original
> Pentium :)

> ...
> Linux, because eventually, you grow up enough to be trusted with a fork()

OTOH, I'm not sure anyone grows up enough to be trusted with a float <wink>:

    http://www.python.org/doc/current/tut/node14.html





From mark at freelance-developer.com  Fri Apr 25 04:01:50 2003
From: mark at freelance-developer.com (Mark Nenadov)
Date: Fri, 25 Apr 2003 08:01:50 +0000
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
Message-ID: <Kq9qa.9083$h%2.957726@read1.cgocable.net>

Alessio Pace wrote:

> Hi, I'd like to make a GUI for a program written in python, meant to run
> on mac, win and linux: which is the preferable solution? I saw around
> tkinter, wxpython and pyqt....
> 

I'd strongly recommend wxPython. wxPython is quite a rich GUI library and I 
have found it to be more advanced/modern/feature-rich than tkinter.

-- 
~Mark Nenadov



From cben at techunix.technion.ac.il  Sun Apr 27 14:17:04 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Sun, 27 Apr 2003 21:17:04 +0300 (IDT)
Subject: In which versions is list.sort stable?
In-Reply-To: <b8en69$263$1@panix2.panix.com>
References: <slrnbalhft.bc1.mlh@furu.idi.ntnu.no> <yyAqa.8330$3M4.231032@news1.tin.it>
 <b8en69$263$1@panix2.panix.com>
Message-ID: <Pine.GSO.4.51.0304272112030.17285@techunix.technion.ac.il>

Aahz wrote on 2003-04-26:

> In article <yyAqa.8330$3M4.231032 at news1.tin.it>,
> Alex Martelli  <aleax at aleax.it> wrote:
> >Magnus Lie Hetland wrote:
> >>
> >> Has list.sort always been stable? Or, if not, in which Python versions
> >> is it not stable?
> >
> >list.sort has not been stable in 1.5.2, 2.0.*, 2.1.*, 2.2.*.  I do not
> >know about previous versions.  Note that list.sort may once again cease
> >being stable (in 2.4.* or further versions) if the timbot dreams up one
> >that's even better on _other_ criteria -- stability's not going to
> >constrain his algorithm choices.
>
> OTOH, Uncle Timmy won't gratuitously remove stability, either.  But I
> still wouldn't bet on Python's sort staying stable past 2.3.
>
Why don't we just grow a `list.stablesort` method that can be an alias
to `list.sort` currently but is guaranteed to be some stable sort
forever?  Explicit is better than implicit, so in the rare case that
you do rely on it, why not spell it out?

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From anthony.hawes at breathemail.net  Fri Apr 25 18:08:40 2003
From: anthony.hawes at breathemail.net (Anthony Hawes)
Date: 25 Apr 2003 15:08:40 -0700
Subject: OO + Python Developers Wanted
References: <mailman.1051183297.29110.python-list@python.org> <yyyznmezekm.fsf@adams.patriot.net>
Message-ID: <8ebbfec2.0304251408.11fb68cc@posting.google.com>

Nick Vargish <nav at adams.patriot.net> wrote in message news:<yyyznmezekm.fsf at adams.patriot.net>...
> "Neeta Ramchandani" <nramchandani at harveynash.com> writes:
> 
> > This is a role for an Investment Bank in the City, for duration of
> > 3-6 months. Rates are OPEN!!
> 
> Which "City" would that be?
> 
> Nick

The website (harveynash.com) mentions London stock exchange

Ant



From peter at engcorp.com  Wed Apr  2 16:50:38 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 16:50:38 -0500
Subject: Page ranking in Google (was Re: Can double-linked lists be implemented 
 in python?)
References: <169a7283.0304020432.5a824d1@posting.google.com> <slrnb8lbo0.ccp.rmunn@localhost.localdomain> <3E8AFFB7.9516A49F@engcorp.com> <e81be8b2.0304021348.c0f4c9a@posting.google.com>
Message-ID: <3E8B5B2E.2B16689E@engcorp.com>

Dean Goodmanson wrote:
> 
> When sending someone to Google, remember to support page ranking and
> send `em with one of them URL's that are apparrantly so easy to
> generate.

Please clarify.  "Support page ranking"?  This sounds like a
new movement with which I'm not familiar.  Are there rallies
and things too?  :-)

(Shorter version: huh?)

-Peter



From salvatore at wanadoo.fr  Sat Apr  5 05:03:16 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Sat, 05 Apr 2003 12:03:16 +0200
Subject: In memory File
In-Reply-To: <3E8EA197.3AE807EB@alcyone.com>
References: <b6m741$dqo$1@news-reader12.wanadoo.fr> <3E8EA197.3AE807EB@alcyone.com>
Message-ID: <3E8EA9E4.60704@wanadoo.fr>

Erik Max Francis wrote:

> Salvatore wrote:
> 
> 
>>How can I create file-like objects ?
>>I want to work in memory only.
> 
> 
> You want the StringIO module.
> 
Thanks Max,

I wanted to redirect sys.stdout to StringIO object but il doesn't seem
to work




From tomchance at gmx.net  Wed Apr  9 13:20:02 2003
From: tomchance at gmx.net (Tom Chance)
Date: Wed, 09 Apr 2003 18:20:02 +0100
Subject: a size-intelligent progress bar?
Message-ID: <1PYka.1140$RX1.361@newsfep3-gui.server.ntli.net>

Hullo,

I'm making a shell-based program that includes a progress bar (like wget's
one). At the moment, I've got it printing an equals sign every 2% so I get
a progress bar 50 "="s wide, which fits my Konsole perfectly. This isn't
very handy for other shell sizes though :)

Does anyone have, or can anyone point me in the direction of a way of making
a decent progress bar that will automagically detect the width of the shell
open, even resizing it mid-"progress"?

Cheers in advance,
Tom



From aleax at aleax.it  Wed Apr 23 07:39:18 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 11:39:18 GMT
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org>
Message-ID: <GXupa.83382$T34.2079161@news2.tin.it>

<posted & mailed>

Gr?goire Dooms wrote:

> Hi,
> I need a lower and higher integer for initializing lower and higher
> bounds in an algorithm.  I know integers have no limits(maximum and
> minimum) in python and infimum and supremum would do the job too.
> 
> I empirically discovered (Python 2.2.2) that None could be considered an
> infimum and any string could be considered a supremum.
> e.g "a">int(x) is always true and None<int(x) is always false
> 
> But this fact is not stated in the language reference.

Indeed, it's not a property of the language: it's a property of your
specific implementation (and, indeed, the results of comparisons between
different arbitrary types are even allowed to vary between runs).

If you ARE indeed talking about int (not about long, which IS indeed
unbounded), sys.maxint >= any int, and -sys.maxint-1 <= any int.  But
since int() itself IS now allowed to return a long, that may perhaps
be of little use to you.

> Citing http://www.python.org/doc/current/ref/comparisons.html :
> "The objects need not have the same type. If both are numbers, they are
> converted to a common type. Otherwise, objects of different types always
> compare unequal, and are ordered consistently but arbitrarily."
> 
> Could/should that order be hardcoded in the language ?

It surely COULD -- all comparisons could e.g. specialcase type None
to ensure it compares < any different type.  As for SHOULD, I dunno.

For your algorithm, why not just, for example:

class Infimum:
    def __cmp__(self, other): return -1
infimum = Infimum()

class Supremum:
    def __cmp__(self, other): return 1
supremum = Supremum()

as coded these have obvious anomalies, e.g. infimum < infimum -- but
those are easy to remove if they give problems to your algorithm.


Alex




From pan-newsreader at thomas-guettler.de  Sat Apr  5 09:22:32 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Sat, 05 Apr 2003 16:22:32 +0200
Subject: POP3 and selective download/deletion
References: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com>
Message-ID: <pan.2003.04.05.16.22.31.832230.878@thomas-guettler.de>

On Sat, 05 Apr 2003 12:12:26 +0200, Dino Levy wrote:

> Hello!
> I'm making a mail program and I have a little problem: I don't know how
> to achieve the selective downloads/deletion from POP3 servers.

Maybe this recipe from the python cookbook does help you to get started:
  http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/134945

It prints all messages which are biger then N bytes and ask if it should
be deleted.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From hawkeye.parker at autodesk.com  Wed Apr  2 19:09:21 2003
From: hawkeye.parker at autodesk.com (haughki)
Date: 2 Apr 2003 16:09:21 -0800
Subject: windows-cmd shell doesn't reflect _winreg changes
Message-ID: <b6e56f7a.0304021609.521700e6@posting.google.com>

hi

i'm using _winreg to 'hard-change' and environmental variable in
windows.  the change is 'successful': regedit correctly reflects the
changes, and crtl panel > system > properties envronmental vars have
been correctly changed.  but, when i start a cmd shell (dos), the
shell evironment still reflects the old values until i manually open
the env vars cntrl panel and click 'ok'.  anyone?

thanks
hawkeye



From grahamd at dscpl.com.au  Tue Apr 29 03:17:17 2003
From: grahamd at dscpl.com.au (Graham Dumpleton)
Date: 29 Apr 2003 00:17:17 -0700
Subject: socket module's listen backlog problem
References: <ba1e306f.0304281519.1304ce5a@posting.google.com>
Message-ID: <dc6f5c99.0304282317.7012e3fd@posting.google.com>

SOMAXCONN value of 0x7fffffff means use the largest value the OS supports.
Problem is that Micro$oft differentiates their products and low end products
only support  a backlog of 5. You will need Windows 2000 Server in order
to get more than 5 to work. I can't remember what the actual number the
Server version of their product supports though.

janeaustine50 at hotmail.com (Jane Austine) wrote in message news:<ba1e306f.0304281519.1304ce5a at posting.google.com>...
> I'm running Python 2.3b1, which has been downloaded from python.org as
> binary, on my Win XP Pro machine.
> 
> socket.SOMAXCONN is 5. I want it to be higher. However, listen with
> higher backlog number doesn't work.
> 
> I'm using asyncore and removed the following code but it still doesn't
> work with more than 5 backlogs.
> 
> #from asyncore
>         if os.name == 'nt' and num > 5:
>             num = 1
> 
> I suppose the socket module uses winsock2.h and it's SOMAXCONN is
> 0x7fffffff.
> 
> What's wrong? What can I do to increase listen backlog?



From aleax at aleax.it  Tue Apr 29 09:42:16 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 13:42:16 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051571264.31905.python-list@python.org>
Message-ID: <Yivra.19878$3M4.571305@news1.tin.it>

Bjorn Pettersen wrote:

>> From: Alex Martelli [mailto:aleax at aleax.it]
>> 
>> Bjorn Pettersen wrote:
>>    ...
>> > Then define a hierarchy:
>> > 
> [...]
>> >  class CountedPoint(Counted, Point):
>> >      def __init__(self):
>> >          super(CountedPoint.self).__init__(self)
>> 
>> hmmm. I think you want a comma there, not a dot, in the super args.
>> And you shouldn't pass self as an argument here, either.  Just:
> 
> Sorry about the dot, a little premature copying. That __init__ doesn't
> take self is an interesting artifact of the implementation [super(..)
> seems to want to be object substitutable, but only is some of the
> time...]

It returns an instance, and you call methods on that instance -- just
as you would call methods on other instances, as opposed to calling
them on classes.  What's confusing about it?

> 
> [...]
>> > metaclasses aren't related by inheritance. Trying to fix
>> > this by, e.g.
>> 
>> "Related by inheritance" may be a bit too weak.  More specifically:
> 
> Yes, but not in this case, which is why I didn't want to say:
> 
> [..for all B in C.__bases__, C.__metaclass__ must be subclass of
> B.__metaclass__...]
> 
> Not that I didn't know them, I just haven't figured out why they must be
> so restrictive ? I'm guessing it has to do with attribute lookup rules,
> however I haven't been able to find a good description of attribute
> lookup for new-style classes... (references?) E.g. why is the metaclass'
> __getattr__ never called?

Uh, what do you mean with this last question...?

>>> class meta(type):
...   def __getattr__(self, name):
...     if len(name)==1: return name
...     raise AttributeError, name
...
>>> __metaclass__=meta
>>> class X: pass
...
>>> X.h
'h'
>>>

Of COURSE meta.__getattr__ gets called -- when you get an attribute
from an instance of meta, i.e., a class having meta as its metaclass.

X().h will NOT call meta.__getattr__ -- NO operation on an instance
EVER goes "up" to the metaclass of that instance's class, EVER.  NOT
ONE.  Think about it -- would you WANT all operations invoked on
an instance to go up to the metaclass?  So for example EVERY instance
would be callable -- and use the metaclass's __call__ (like calling
the CLASS of the instance) thus create a new instance totally unrelated
to the calling one (except by having the same type)?  I would find
THAT hypothetical feature extremely confusing, personally.

So maybe I don't understand correctly what you're asking...?


>> > making metaIC a subclass of metaMetaBunch (and changing the
>> > __new__ call appropriately), gives you:
>> > 
>> > Traceback (most recent call last):
>> >   File "M:\python\cpoint.py", line 38, in ?
>> >     class CountedPoint(Counted, Point):
>> >   File "M:\python\cpoint.py", line 25, in __new__
>> >     return metaMetaBunch.__new__(cls, name, bases, dict)
>> >   File "M:\python\cpoint.py", line 20, in __new__
>> >     return type.__new__(cls, classname, bases, newdict)
>> > TypeError: multiple bases have instance lay-out conflict
>> > 
>> > which means you're stuck (?)
>> 
>> Well no -- you MIGHT be stuck if you ever needed two different AND
>> incompatible __new__ in the same class, but here metaIC may just as
>> well use __init__ instead (as it ain't messing with e.g. __slots__:-)
>> so everything works if you recode the whole mess as, e.g.:
> 
> Which is fine as long as you have access to the code of both the Point
> and Counted classes and their metaclasses, understand metaclasses in
> general, the ramifications of the particular implementation and how
> other things would break if you started changing things, and even then
> you're stuck if they both need __new__, unless you can code up a
> compromise. Which was basically my only point :-) While the abstract

I don't get this "only point".  How does it differ from the coupling
from any class and its multiple bases?  I.e., aren't you just as stuck
if multiple bases all need __new__, just as if multiple metaclasses do?

"Understanding metaclasses in general" isn't all that different from
"understanding classes in general", methinks -- as long as you recall
that a class is to a metaclass like an instance is to a class, no more,
no less.  OK, we (collectively) aren't very uses to this, yet, but it's
nothing all that terrible.

> base class solution does have its problems, so does the metaclass
> solution (I wouldn't consider mucking with superclasses' implementation
> just because you need to multiply inherit as beeing in the OO spirit
> <wink>) -- the metaclass solution is aesthetically more pleasing though
> :-)

When you need to multiply inherit and the superclasses were never
designed to coexist (e.g. they BOTH have their own special __new__)
you have issues, of course.  With OR without custom metaclasses.

> Well, _I_ just wanted a countable point, someone else forced me into
> hacking "magic code" just because I wanted to use regular inheritance
> <grin>.

If you only want countable points, it may be silly to use ANY kind
of inheritance or custom metaclasses -- just put everything into the
one class that you say you want.  But if you also want points that
need not be countable, countables that need not be points, and so on,
then refactoring commonalities may be worth it.  Custom metaclasses
are one powerful way to refactor out certain kinds of commonalities
for future reuse (and to avoid code duplication), just like, for some
other kinds, inheritance is.  Not ALL that much difference here -- in
no way is one more "magic" or abstruse than the other, except perhaps
transiently by reasons of familiarity.


Alex




From abli at freemail.hu  Wed Apr  2 15:27:18 2003
From: abli at freemail.hu (Abel Daniel)
Date: Wed, 2 Apr 2003 22:27:18 +0200
Subject: Entry widget bind problems
In-Reply-To: <3E8A30BA.7000805@netscape.net>
References: <3E8A30BA.7000805@netscape.net>
Message-ID: <20030402202718.GB751@hooloovoo>

vector wrote:
> Im a little confused with an entry widget bind. I had it working b4 but 
> In an attempt to tidy up the code I have confused the class/callback 
> reference or something.
> As you can see from the text below, there is a MyFrame instance which 
> builds the TK Gui window, in which a frame contains an entry widget. 
> after typing in text into the entry widget a <return> event(enter key) 
> is supposed to run the doEntry, but all i get this error.
> what am i doing wrong?
 [ ... code snipped ... ]

The short asnwer is that you doEntry is wrong. An event callback will
get (at most) one argument when it is called, which will be an Event
instance. Thats what becomes page in your doEntry function.
This stores all sorts of usefull information, of which the
widget attribute will be important for you, as it is the widget where
the event occured.

Changing that function to:

def doEntry(event):
    """Gets text from Entry widget on <enter> hit"""
    page=event.widget
    p=page.get()       #gets all text
    print p              #print it
    page.delete(0,END)        #clear entry widget

should get it working.

Apart from that it looks like you have the arguments to bind() mixed up.
You use the first and second correctly, however the docs write the
following about the third one:
(ADD is the third argument, FUNC is the second one)
    An additional boolean parameter ADD specifies whether FUNC will
    be called additionally to the other bound function or whether
    it will replace the previous function.

So that is simply to allow multiple bindings to the same event, thereby
doing two different things if an event occurs.

An additional remark:
in the code you posted, entry looks like being a method of a class,
however, you don't use it like that. I would change the name of the
first argument from self to something else (like frame) to avoid
misleading other people who read your code.
On the other hand, what you are doing with entry() and doEntry() is
essentially making a costumized widget with some special functionality,
and that can be done much nicer in OO style. Either subclassing from
Tkinter.Entry, or by making an own class. (In fact, I would think putting
the functionality of entry inside the __init__ of MyFrame would be better,
if you don't use it elsewhere.) This would have the advantage
of being able to do:
   self.outEntry.bind('<Return>', self.doEntry)
(notice that in this case doEntry is a method of a class)
and doEntry might look like this:
    def doEntry(self, event):
        p=self.outEntry.get()       #gets all text
        print p              #print it

Notice that in this case, the class instance was automatically passed to
doEntry, and we can avoid reading the contents of the event instance.
But this is (at least partly) a question of aesthetics.

Abel Daniel




From salvatore.didio at wanadoo.fr  Tue Apr  1 17:51:37 2003
From: salvatore.didio at wanadoo.fr (Salvatore)
Date: Wed, 2 Apr 2003 00:51:37 +0200
Subject: Creating instance
References: <b6d191$j7s$1@news-reader12.wanadoo.fr> <u3cl17sfv.fsf@online.no>
Message-ID: <b6d536$pml$1@news-reader12.wanadoo.fr>

"Syver Enstad" <syver-en+usenet at online.no> a ?crit dans le message de news:
u3cl17sfv.fsf at online.no...
> Salvatore <salvatore at wanadoo.fr> writes:
>
> > Hello,
> >
> > I would like to create an instance of class
> > from the string name of the class
> >
> > class Foo:
> > pass
> >
> > classname = 'Foo'
>
> Does this suit you?
>
> instance = eval(classname)()


The beauty is in the simplicity
Thank you very much Syver :o)








From juneaftn at REMOVETHIShanmail.net  Wed Apr 16 13:25:05 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Thu, 17 Apr 2003 02:25:05 +0900
Subject: How many of you are Extreme Programmers?
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com> <mailman.1050504987.840.python-list@python.org> <3E9D8A80.FD76BDE9@engcorp.com>
Message-ID: <b7k3kj$4gj$1@news.hananet.net>

"Peter Hansen" <peter at engcorp.com> wrote in message
news:3E9D8A80.FD76BDE9 at engcorp.com...
> Christopher Blunck wrote:
> >
> > What's discouraging is that it often takes books, buzzwords,
> > and fads to get the attention of higher-ups.
>
> Thankfully, in the case of XP all it took (at first) was
> a Wiki, not a book.  The book came later...
>
> http://c2.com/cgi/wiki?ExtremeProgramming
>
> > Developing software
> > is not all that difficult - it's always so confusing why people make
> > software development more complex than it actually is.
>
> Perhaps one reason some people think it's complex or difficult is
> because so many (90%?) of software projects *fail* in various ways,
> including by being so far over their scheduled release date that
> their customer vanishes.

Check out the Standish Group CHAOS report.

http://www.softwaremag.com/archive/2001feb/CollaborativeMgt.html (refers to
2001 report)
http://www.pm2go.com/sample_research/chaos_1994_1.php

>
> There *is* a reason why agile methodologies are being developed,
> although perhaps some people aren't ready for them yet. ;-)
>
> -Peter




From aleax at aleax.it  Wed Apr 16 11:08:31 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 15:08:31 GMT
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com> <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com> <cnoq9vci702imrunkenpulnefautq0n5ua@4ax.com>
Message-ID: <Plena.22932$LB6.544795@news1.tin.it>

Dale Strickland-Clark wrote:

> Dale Strickland-Clark <dale at riverhall.NOTHANKS.co.uk> wrote:
> 
>>danra at walla.co.il (Danra) wrote:
>>
>>>def f(fileobj):
>>>    if not hasattr(fileobj,'write'):
>>>        raise TypeError, "Given arg must have the 'write' attribute."
>>>    .
>>>    .
>>>    fileobj.write('...')
>>
>>You want assert, you do:
>>
>>def f(fileobj):
>>    assert hasattr(fileobj, 'write'), 'Bang! Sort your code out.'
> 
> Obviously, there's a 'not' missing from the above :-(.

Hmmmm, are you sure...?  This raises when the attribute is
missing, which seems likely to be what one desires.

But I prefer a slightly different approach, exemplified at
http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/52291 --
``Accurate "look before you leap"''.  It mostly comes in handy
when you want to reduce the risk of having done some hard to
reverse "partial state change" by the time you get the exception.

In a case such as this, the idiom would reduce, at a minimum, to:

def f(fileobj):
    write = fileobj.write
    # rest of function follows -- just call write('...') when needed

by extracting the bound-method as the first action of the function,
you ensure the exception happens ASAP -- *AND* you get a nice little
performance boost in many cases too.  You can spend some of that
boost to do a little bit more checking than the above snippets show,
for example:

def f(fileobj):
    write = fileobj.write
    write('')
    # rest of function follows -- just call write('...') when needed

NOW, this ensures you, if you get to the "rest of function" part:
   1. fileobj has a write attribute, AND
   2. that attribute is callable, AND
   3. it is, in particular, callable with a single argument
      that is a plain string

Of course, you need to identify meaningful "no-operation" ways to
call the relevant callables in order to LBYL _very_ accurately.  But
it's often possible and sometimes fruitful to do so.  Of course,
you'll probably want to document this usage accurately.  Incidentally,
the inspiration for "accurate LBYL" came to me from the "Template
Constraints" idiom of C++ (of course, in C++ templates &c are entirely
resolved at compile-time, and so are the "constraints" in the idiom,
but the similarities, in terms of signature-based polymorphism, between
C++ templates and Python functions are still strong).


Alex




From gh at ghaering.de  Wed Apr 30 20:28:17 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Thu, 01 May 2003 02:28:17 +0200
Subject: Python and Java Servlet ISP?
In-Reply-To: <009901c30f5f$6eca3040$c90aa8c0@PDANIELL>
References: <009901c30f5f$6eca3040$c90aa8c0@PDANIELL>
Message-ID: <3EB06A21.50704@ghaering.de>

Ponce Dubuqe wrote:
> Does anyone know if there's an Web hosting Provider that supports both 
> Python and Java Servlets/JSPs?

http://python-hosting.com/

-- Gerhard

PS: Please, text-only postings are preferred here. You can somehow turn 
off the HTML in OE so it send text-only instead of multipart/alternative.




From tug at wilson.co.uk  Tue Apr 22 15:42:29 2003
From: tug at wilson.co.uk (John Wilson)
Date: Tue, 22 Apr 2003 20:42:29 +0100
Subject: xmlrpc introspection
References: <20030422170450.33256.qmail@web13907.mail.yahoo.com>
Message-ID: <020201c30907$4f062650$640a0b0a@handel>

Is this what you're after?

http://scripts.incutio.com/xmlrpc/introspection.html

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From peter at engcorp.com  Wed Apr 16 15:17:50 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 15:17:50 -0400
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com> <v9r6m7864jmic6@news.supernews.com> <3E9DAA66.171CAC9@engcorp.com>
Message-ID: <3E9DAC5E.D4422FA4@engcorp.com>

Peter Hansen wrote:
> 
> John Roth wrote:
> >
> > > So how many of you guys use XP processes?
> >
> > I do. Writing tests before code is something that I've thought would
> > be a good idea for literally decades, but it took both XP and having
> > the test tool delivered with the language to get me to start.
> 
> I know you knows this, but for the benefit of others who aren't
         ^^^
D'oh!  "I know *John* knows this..."    

(Grumble grumble stupid keyboard, mistaking the letters j-o-h-n for y-o-u.)

-Peter



From cben at techunix.technion.ac.il  Mon Apr  7 11:59:00 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Mon, 7 Apr 2003 18:59:00 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2u1dcw22s.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <Pine.GSO.4.51.0304071659170.8306@techunix.technion.ac.il>

Patrick K. O'Brien wrote on 2003-04-06:

> If you were crazy enough to think that you could create a program
> along the lines of Emacs, but written in Python, how would you go
> about it?  How would you design the domain model for files, buffers,
> windows, and frames?  How would you allow the same level of
> customizability?  How would you map functions (or methods or whatever)
> to keys?  Any thoughts?
>
I wrote some implementation ideas in other posts, I'll now write my
wishes for being better than emacs...

My #1 wish is not to have one major mode with one syntax per buffer.
Suppose I'm writing Python so I want python-mode - but I'm also
writing reStructuredText in my doc-strings so I want rest-mode [being
written by somebody now, still incomplete].  And then the python is
generating xml snippets, so I want xml-mode inside specific places.

Emacs started out by assuming there is just one major mode.  That
means there is one syntax table, one paragraph-start regexp, one
evrything.  Some specific things are allowed to be modified by minor
modes and text properties/overlay and the rest can be half-emulated by
mmm-mode.sf.net but I strongly believe this can and should be
addressed from the first day.

When Python mode discovers that something is a string, it should mark
it as foreign territory and let the user decide what modes get to
control its content.  A buffer will thus contain a hierarchy of modes.

This reminds me, there must be better ways to achieve syntax
sensitivity than emacs.  It's not something to add at an after-thougt
in font-lock in an optional way; there should be standard mechanisms
for modes to extract and communicate syntactic information.

As a whole, I want more flexible ways to describe syntaxes, so that
specialized modes could be constructed from pieces reather than
written from scratch.  E.g. I want <tag>...</tag> in xml to be
described as sort of "big parentheses" in a way that show-paren will
understand.

#2 (or is it 4 already?): this is a dangerous issue but perhaps flat
text buffers are not the best way.  They are, for most editing modes;
however when I need a list-of-dicts buffer, a repr-of-list-of-dicts is
not the ideal thing to work with.

Emacs has quite a lot of modes that are not about editing flat text at
all (dired, buffer-list, custom, rmail's inbox, speedbar, etc.).
They could well benefit from an being able to manipulate other
structures.  To be able to use page-up/down and any other bindings
that are not overriden, any such structure will need to provide view
of itself in simpler terms, all down to flat text...

I have many other ideas but I can't put them down clearly yet...

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>
When I grow up I want to write an Emacs.




From the_3_project at yahoo.com  Tue Apr 29 01:41:40 2003
From: the_3_project at yahoo.com (Joe Green)
Date: 28 Apr 2003 22:41:40 -0700
Subject: Arg names in class and function code - Newbie question
References: <64e13ba5.0304281424.52cb6b5d@posting.google.com>
Message-ID: <a5b86892.0304282141.26f2920a@posting.google.com>

newspost at lolmc.com (Lol) wrote in message news:<64e13ba5.0304281424.52cb6b5d at posting.google.com>...
> HI all,
> 	Could anyone let me know - do the names you give as arguments HAVE to be
> actual variable names?
> e.g def readFile(filename,dir):
> 	blah blah
> 	return blah
> or even class FileRead(filename,dir):
> 	def doSomething (self,this):
> 		yadda yadda
> 		return yadda
> 
> In the above examples do filename and dir have to be the name of the
> variables to pass as args to the class/functions OR are they descriptive
> so I can pass the real name of the file and directory?


No, they don't. However, if you are using default values, then they
do.

For example:

def myfunc(arg1=None, arg2=None)

Here, if you are calling 'myfunc' and passing it any arguments, you
must call them by name. e.g.

myfunc(arg1='hi', arg2='hello')

Otherwise the interpreter will have no idea what value goes to what
(the order by which you pass the arguments doesn't matter).

If you only have something like:

def myfunc(arg1, arg2)

then you don't have to explitly call them anything, you can do

s1 = 'hi'
s2 = 'hello'

myfunc(s1,s2)

and that should work fine. Experiment on your python shell.

Hope that helps.



From linux_pt2003 at netcabo.pt  Tue Apr  8 10:54:26 2003
From: linux_pt2003 at netcabo.pt (Carlos Sousa)
Date: Tue, 08 Apr 2003 15:54:26 +0100
Subject: PyQT
Message-ID: <3e92e21e$0$1986$a729d347@news.telepac.pt>

Hi

Can someone tell me who to install eric3 with PyQT ?
I try from the sources but something goes wrong


Thanks



From fusion at thuule.pair.com  Sat Apr 12 21:58:20 2003
From: fusion at thuule.pair.com (Alan McIntyre)
Date: Sat, 12 Apr 2003 21:58:20 -0400
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <3E985DC8.9060907@thuule.pair.com> <mailman.1050178039.28676.python-list@python.org> <bdb5b58f.0304121731.1b1312f4@posting.google.com>
Message-ID: <3E98C43C.8050109@thuule.pair.com>

logistix wrote:
> This is wrong.  You should be doing t.append(time.time()-starttime). 
> That will show how much more time 1k, 2k, 3k etc operations will take
> compared to 1k.
> 
>  Right now you're only timing over and over for groups of 1000.  If
> this was linear [O(n)], you'd get a flat line on the graph with your
> current equation.

All I wanted to find was the average time required to tack a 2 character 
string onto the end of an existing string by using +=.  I know it's not 
entirely accurate to take the average of the time required to append a 
string to an existing string of lengths 0 to 1998, 2000-3998, etc and 
divide by 1000, but I just wanted ballpark numbers.

I wasn't interested in the time to execute the entire 200k append 
operations; as Mr. Martelli pointed out earlier, it's O(n^2), but the += 
append (considered by itself) appears to be O(n).





From MFranklin1 at gatwick.westerngeco.slb.com  Thu Apr 17 06:13:07 2003
From: MFranklin1 at gatwick.westerngeco.slb.com (Martin Franklin)
Date: Thu, 17 Apr 2003 11:13:07 +0100
Subject: scrollbar
In-Reply-To: <b7ltfi$49c$1@atlantis.news.tpi.pl>
References: <b7kh3g$51u$1@nemesis.news.tpi.pl>
 <MAlna.4040$ej5.2777@news01.roc.ny.frontiernet.net>
 <b7ltfi$49c$1@atlantis.news.tpi.pl>
Message-ID: <3E9E7E33.3090801@gatwick.westerngeco.slb.com>

zyziu wrote:
> ----- Original Message -----
> From: "Lee Harr" <missive at frontiernet.net>
> Newsgroups: comp.lang.python
> Sent: Thursday, April 17, 2003 1:22 AM
> Subject: Re: scrollbar
> 
> 
> 
>>In article <b7kh3g$51u$1 at nemesis.news.tpi.pl>, zyziu wrote:
>>
>>>Hello. I made lanChat in Python. I have problem with Text widget. I want
> 
> to
> 
>>>text from Entry goes to Text widget and scrollbar goes down, because I
> 

<snip>

> 
>     def Pisz(self):
>          self.dane = self.tekst.get()
>          self.tekst.delete(0,'end')
>          if (self.dane<>''):
>              self.okno.insert('end',strftime("%H:%M:%S")+'\n')
>              self.okno.insert('end',self.dane+'\n'+'\n')

  A call to yview_pickplace here should do the trick.....

self.okno.yview_pickplace("end")


Regards
Martin





From no at spam.invalid  Tue Apr 29 16:02:36 2003
From: no at spam.invalid (Russell E. Owen)
Date: Tue, 29 Apr 2003 13:02:36 -0700
Subject: Python 2.3b1: RuntimeError using rexec
References: <b8mh4l$2oo0$1@nntp6.u.washington.edu>
Message-ID: <b8mlot$1s6u$1@nntp6.u.washington.edu>

In article <b8mh4l$2oo0$1 at nntp6.u.washington.edu>,
 "Russell E. Owen" <no at spam.invalid> wrote:

>I'm trying out Python 2.3b1 and am having trouble with rexec:

Sorry, stupid question. I reread the what's new more carefully and saw 
that rexec has been killed (though the 2.3b1 docs for rexec say nothing 
of this; I have report it as a documentation bug).

I'll use eval instead to convert string representations of simple 
dictionaries (containing only numbers and strings) to dictionaries. But 
if anyone has a better (e.g. safer) suggestion I'd be interested.

-- Russell



From jerry.seutter at pason.com  Fri Apr 11 15:04:38 2003
From: jerry.seutter at pason.com (Jerry Seutter)
Date: Fri, 11 Apr 2003 13:04:38 -0600
Subject: string.split() documentation incorrect?
In-Reply-To: <qard9v4ndv0g89bffoseign9g3isvjt6og@4ax.com>; from carroll@tjc.com on Fri, Apr 11, 2003 at 09:39:00AM -0700
References: <mailman.1050076398.7023.python-list@python.org> <qard9v4ndv0g89bffoseign9g3isvjt6og@4ax.com>
Message-ID: <20030411130438.B16080@ash.int.ca.pason.com>

Doh, I had never realized that this was there.  I've been maintaining
Python 1.6 code and it wasn't used or wasn't available.

Bad code breeds more bad code.  Thank you.

Jerry Seutter

On Fri, Apr 11, 2003 at 09:39:00AM -0700, carroll at tjc.com wrote:
> However, use of the string module is deprecated in favor of methods of
> the string class, which work as documented.  The string module only
> calls the equivalent class method anyway, e.g., string.strip(s) turns
> into s.strip().
> 
> Save yourself some trouble and use the class method directly, e.g.,
> instead of these:
> 
>  >>> string.strip(s,",")
>  >>> string.strip("id INT NOT NULL,",",")
> 
> use these:
> 
>  >>> s.strip(",")
>  >>> "id INT NOT NULL,".strip(",")
> 




From ianb at colorstudy.com  Mon Apr  7 23:09:27 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 07 Apr 2003 22:09:27 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <b6tcb5$8p1s0$1@ID-169208.news.dfncis.de>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	 <6ee58e07.0304061144.35c5ab92@posting.google.com>
	 <m2k7e78j5s.fsf@localhost.localdomain>
	 <mailman.1049662570.29853.python-list@python.org>
	 <yfsr88fkppr.fsf@black132.ex.ac.uk>
	 <mailman.1049687898.18892.python-list@python.org>
	 <b6tcb5$8p1s0$1@ID-169208.news.dfncis.de>
Message-ID: <1049771367.749.65.camel@lothlorien>

On Mon, 2003-04-07 at 21:41, Greg Ewing (using news.cis.dfn.de) wrote:
> Ian Bicking wrote:
> > Given a novel form of input device, perhaps there's something to what
> > he's trying to do.  I'm skeptical, though.
> 
> Foot pedals. Hold down the Search pedal and type away.
> 
> Maybe one of those car-racing-game accessories could
> be used to prototype this...?

There are foot pedals that map to things like Shift, for use by people
with RSI.  Little expensive ($50?  Maybe $100...), but easy to get
started with otherwise.

  Ian






From missive at frontiernet.net  Thu Apr 10 19:19:19 2003
From: missive at frontiernet.net (Lee Harr)
Date: Thu, 10 Apr 2003 23:19:19 GMT
Subject: class too big for one file! How can I split it?
References: <e956b5f1.0304100512.597654d3@posting.google.com>
Message-ID: <XZmla.304$ZX6.159@news02.roc.ny.frontiernet.net>

In article <e956b5f1.0304100512.597654d3 at posting.google.com>, HHaegele wrote:
> Hello,
> I have the following structure:
> 
>===============
> ot_starter.py :
> ---------------
> 
> import OTLib
> 
> ot = OTLib.Ot(A, E)
> 
> ot.Att()
> 
> ot.Btt()
> 
> ---------------
> 
> 
>===============
> OTLib.py :
> ---------------
> 
> XX = None
> 
> class Ot(a, e):
> 
>     def __init__(self, a = None, e = None):
>         global XX
>         XX = e    
>        
>     def Att(self):
>         global XX
>         print XX
>         ...
>         
> ---------------
> 
> Now, OTLib.py has more then 32000 lines. And that doesn?t work. Python
> 1.5 can?t execute then. So, how can I add the method Btt()? How can I
> make another file, which knows the global XX and who can be handled
> from the ot_starter, like it is in one file? Good would also be, if
> the Btt() can access XX the same way like Att() can. Is that possible?
> Or how must I change the structure?
> 


How about a subclass?

=================
OTLib1.py :
-----------------

import OTLib

class Ot(OTLib.Ot):
    def Btt(self):
        ...





From benoit.besse at club-internet.fr  Tue Apr 15 19:52:02 2003
From: benoit.besse at club-internet.fr (Benoit BESSE)
Date: Wed, 16 Apr 2003 01:52:02 +0200
Subject: Réception et décodage d'une trame UDP
Message-ID: <3e9c9b1e$0$1509$7a628cd7@news.club-internet.fr>

Bonsoir tout le monde,

qui a d?j? ?crit un module qui ?coute, re?oit, d?code et insert dans une BD
une trame UDP.
Je voudrais d?coder la trame UDP ainsi et l'ins?rer dans une table :
    - Port Source
    - Port Destination
    - Longueur
    - Somme de contr?le
    - Donn?es

Merci par avance.

Ben





From skip at pobox.com  Wed Apr 30 11:27:06 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 30 Apr 2003 10:27:06 -0500
Subject: Determining Unicode encoding.
In-Reply-To: <ef7a350b.0304300707.7c43ad4c@posting.google.com>
References: <ef7a350b.0304291242.42c9b024@posting.google.com>
        <ef7a350b.0304300707.7c43ad4c@posting.google.com>
Message-ID: <16047.60234.364399.198789@montanaro.dyndns.org>

    sean> From what I'm hearing here, it sounds like I can just use UTF-8
    sean> and not have to worry about it.

Sounds like it.  As long as you know the data encoded in the csv file is in
utf-8, you're golden when you read it back in, precisely because you know
the encoding.

Skip




From exarkun at intarweb.us  Tue Apr 15 16:30:02 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 16:30:02 -0400
Subject: Why functional Python matters
In-Reply-To: <slrnb9on0q.sho.ramen@lackingtalent.com>
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
Message-ID: <20030415203002.GB5198@meson.dyndns.org>

On Tue, Apr 15, 2003 at 07:22:47PM -0000, Dave Benjamin wrote:
>
> [snip - comparisons to and examples from PHP]
> 
> Guido's recent musings in the newsgroups and on his presentation on "Python
> Regrets" have made me a bit uneasy. In particular, they brought to light the
> fact that he generally prefers the object-oriented style to the functional
> style and would like to "deprecate" certain features:
> 
>  - apply, to be replaced with *-notation
>   - map and filter, to be replaced with list comprehensions
>    - lambda
>    

  While they take getting used to, you can't say that list comprehensions
aren't functional.  They're not traditional-lisp functional, but that isn't
the only kind :)

>    Lambda has had the thorniest history of all of these, I think. Lambda has
> its roots in LISP and the Lambda calculus, and can be used to create
> "anonymous functions" at runtime that can be passed around as needed. This
> is a very powerful and useful feature for functional programming, as well as
> for event-driven programming. It baffles me that anyone would want to remove
> it.

  No worries, there are enough people who would shout loudly enough were
this to be removed that its existence is safe, at least for the near future.

  Just make sure you give a holler anytime someone suggests discarding them.

> [snip - default argument hacks, nested scopes]
> 
> The second "problem" that everyone seems to keep bringing up is that lambda
> limits you to expressions, not statements. This makes me laugh every time,
> because this is one of the things that I *loved* about Python when I was
> first learning it! If you get used to writing functions that only return
> expressions, you are on the road to understanding the functional style! One
> thing that I recommend that anyone try as a learning experience is this: try
> writing a program where every single function starts with "return". Better
> yet, write your whole program with lambdas. <screams from the back row> Of
> course, I don't recommend doing this sort of thing "out in industry", but I
> guarantee that you will learn to write code that is more reusable,
> side-effect free, and finely decomposed once you have gained some practice
> writing code this way.

  Lambdas are an excellent tool for obfuscation, that's for sure.  You're
right not to recommend overusing them in production code.  Generally, I try
to limit myself to the simplest expressions, on the order of "x and y" or
"result[index]", and translate to fully fledged functions for larger
expressions.  It's just easier to maintain.

  Also, note that lambdas *aren't* side-effect free, because Python isn't
side-effect free.  You can write ones that are, but then again, you can
write functions with "def" that are side-effect free too.

> 
> I would not like to see lambda support "statements". I think that would ruin
> a good thing. I would not like it removed from the language. I want it the
> way it is. The only thing between "lambda" and perfection is the name. It
> should be "fun" or something. If I had a lambda key on my keyboard, I would
> press it. <grin>

  Statements are, by definition, things which cannot appear in lambdas ;)
No worries here, I think.

> 
> "apply" is less of a big deal to me, but I'd still like to keep it because
> it allows for greater portability. I feel so more strongly about map and
> filter. Yes, I know about list comprehensions. I think they are wonderful, I
> thank Guido for adding them, I thank Haskell for inspiring them, and I don't
> want them to ever go away. But just because you can do a similar set of
> things with list comprehensions than with map and filter doesn't mean that
> they are completely redundant. In fact, I believe that it is advantageous to
> design functions specifically for use with map and filter. If you design
> this way, you will benefit from cleaner syntax than list comprehensions can
> offer. 

  Luckily, map() and filter() are implementable in terms of list
comprehensions.  So even in Python3k, when they're finally removed from the
language, you can implement them yourself in Python, and gain the
performance benefits of using a JIT on python bytecode ;)

> For example:
> 
>  - map(process_record, records)
>  
> compared with:
>  
>  - [process_record(record) for record in records]
>   
>   I'm aware that adding lambdas, nesting, multiple filters, etc. can make
> map/filter much less readable than list comprehensions, but functional
> decomposition is another solution. If you have to nest that much, perhaps
> your functions are too complicated and need to be broken up. I'm not saying
> that this is always the case, but for the sake of clarity - in certain
> situations - I greatly prefer map/filter to list comprehensions. For
> starters, they eliminate the need to name the temporary variable.

  I use both in my programs, but I find more and more that I only use
map/filter for the runtime speedup they sometimes provide.  Again, list
comprehensions are functional, just not in the same way.

> 
> Rather than remove these important functional built-ins, I would like to see
> two last built-ins be added: unzip (to reverse the behavior of zip), and
> curry (see the example by Alex Martelli, et. al. in the Python Cookbook,
> sec. 15.7). Anything else can be thrown into a "functional" standard module.
> The global namespace doesn't need any more clutter, and I sympathize with
> those who wish to reduce it, but--please!--not at the expense of the
> functions that form the foundation of functional programming!

  The inverse of zip is already available (and with one less character than
unzip! ;).  x == zip(*zip(x)) (ignoring sequence types, of course).

> [snip - flexibility of Python]
> 
> Finally, I'd like to make one comment about Perl. I think that this is
> important and relevant. There is clearly a difference in philosophy between
> Perl culture and Python culture. Perl says There's More Than One Way To Do
> It, and Python begs to differ. Perl is all-inclusive, Python is more
> exclusive. However, let's not get carried away with the exclusivity! There
> are clearly multiple of ways of doing many things in Python, and that's not
> a bad thing. Sometimes you want a list, other times you want an iterator.
> Sometimes you want a list comprehension, other times you want map/filter.
> Closures? Objects! Dispatches? Polymorphism! Inheritance? Composition? Just
> because Perl says you can do anything any-which-way doesn't mean that to
> compete Python should force you to a single method of problem-solving.

  I don't think Python is headed the way of Unlambda, don't worry :)

> 
> Perl is a beautiful thing. It encourages more creativity in the sense of
> language design than any other language I have ever used, including Python.
> There are many things to learn from Perl, and open-mindedness is one of
> them. Java, on the other hand, is extremely rigid about its way of doing
> things--trying to do functional programming in Java requires the use of
> bulky wrapper classes instead of true, lightweight higher-order functions,
> and its lack of parametric polymorphism makes the entire type system a pain
> in the butt.

  What I think is an important distinction to make here is the language vs
the library.  Right now, there are overlapping language features.  I'd much
prefer it if some of the things, even the useful things like map(), were
re-implemented in terms of other Python features and moved out of the
language and into the library.  This gives the best of both worlds - a
language simple to learn and use that also provides an expressive set of
features through its stdlib.

> 
> Python is a nice happy medium between regularity and freedom. I'd like to
> see it stay that way.
> 

  Jp

-- 
Somewhere, something incredible is waiting to be known.
                -- Carl Sagan
-- 
 up 26 days, 17:02, 6 users, load average: 1.06, 1.16, 1.12




From aleax at aleax.it  Fri Apr 18 10:07:07 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 14:07:07 GMT
Subject: create a new instance in C of a Python subclass
References: <36ko9vknj9mm6m1ad18r3dm51q3t02pe03@4ax.com> <he8zegsn.fsf@python.net> <4v00avocpki19ijlj8f8revh9tb032ht0f@4ax.com>
Message-ID: <fETna.50706$T34.1368488@news2.tin.it>

Paul Miller wrote:

>>You do the same as in Python: call the type.
> 
> How would I do that from C?

To call any Python callable X (where X is a PyObject*) from C,
you can use C/API calls such as PyObject_CallObject, 
PyObject_CallFunction, PyObject_CallFunctionObjArgs.  Any of them 
returns a new reference -- they differ by how you pass parameters.

This holds for any callable Python object X, including of course
types as well as functions and many other kinds.


Alex




From cswiger at mac.com  Fri Apr 11 16:10:03 2003
From: cswiger at mac.com (Chuck Swiger)
Date: Fri, 11 Apr 2003 16:10:03 -0400
Subject: popen(), sendmail: Success?
In-Reply-To: <b7703o$ejq$1@news.island.liu.se>
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de> <mailman.1050021261.25010.python-list@python.org> <b7703o$ejq$1@news.island.liu.se>
Message-ID: <3E97211B.4030906@mac.com>

Thomas Bellman wrote:
> Chuck Swiger <cswiger at mac.com> wrote:
>> If you plan on sending any significant amount of mail, it is
>> recommended that your code talk SMTP to port 25 and pipeline all of
>> the messages you need to send in one continuous stream.  If you
>> fork a sendmail subprocess for each message you send, that will
>> create a wastefully huge amount of system load.
> 
> If you write your application to talk SMTP directly, instead of using
> the proper Unix API for sending mail (i.e, calling 
> /usr/lib/sendmail), you have a number of problems.

Which "proper Unix API"?  The closest thing is RFC-2822, since 
/usr/lib/sendmail doesn't exist on many Unix platforms:

18-sec% which -a sendmail
/usr/sbin/sendmail
19-sec% uname -a
FreeBSD sec.codefab.com 4.8-STABLE FreeBSD 4.8-STABLE #0: Wed Apr  9
08:07:21 EDT 2003     root at sec.codefab.com:/usr/obj/usr/src/sys/SEC  i386
20-sec% ls -l /usr/lib/sendmail
ls: /usr/lib/sendmail: No such file or directory

> First, which machine should you talk SMTP with?  Localhost?

A site which intends to handle a non-trivial amount of electronic mail
needs to have one or more designated mail servers, very commonly
identified via MX records in the DNS.  Sites with enough traffic likely
have both inbound and outbound SMTP boxes.

> Not portable.  There are lots of Unix machines which do not have any 
> MTA running listening to port 25.

True, but I don't recall insisting that one need to do SMTP specificly 
to localhost:25.  If you want to use another machine to queue the mail,
such as a dedicated mail server, instead of locahost, that's fine by me.

> Every machine I administrate run Sendmail in queue mode only (except
> for the mail servers, of course).

By this you mean, "you do not configure your clients run sendmail as a 
daemon listening on port 25" and not "I always run sendmail in 'actual 
delivery is deferred until the next queue run' delivery mode (-odq)"...?

> Thus, you need your application to be configurable for
> the mail server to use.  If the system changes, and another mail 
> server should be used, the sysadmin might not remember, or know, that
> that particular application must be reconfigured.

That's why sysadmins need to document how systems are configured and 
note what the dependencies are.  Just as developers need to write 
applications to be configurable enough to select a MTA-- possibly not 
sendmail, and possibly not running on the localhost-- to handle email 
generated on behalf of the system.

> It becomes even worse if it is an application where each user on the 
> machine has their own configuration, like Netscape or Mozilla...

Try specifying "option smtp-server _IP_;" within your DHCP server's 
configuration, or perhaps use port forwarding, IPFW divert sockets, etc 
so that connections made to localhost:25 get forwarded to a machine 
which will handle such mail.

> Second, you better be prepared to queue the mails in case of network
> or mail server problems and retry sending them at a later time.

Certainly.  If your app can hand email off to a local MTA which is 
always available, it doesn't have to queue the mail itself-- let the MTA 
do that.

[ ... ]
> If you talk SMTP directly, you run the risk of skipping any local 
> policy for sending email, like masquerading as a specific domain 
> instead of as the FQDN of the local host, or a number of other 
> settings.

I think you may have misunderstood my suggestion: I don't believe or 
recommend that someone writing a python program should perform MX 
lookups and try to do SMTP to the mail server of the addressee, retry 
undelivered messages, etc.

Talking SMTP to a local mailserver (25/tmp, or via the MSP, 587/tcp) 
lets your local mailserver perform masquerading, mailertable lookups, 
virus scanning, forwarding mail through a firewall which would otherwise 
block outbound SMTP, etc.

> There are many subtleties involved in the game of being, or 
> pretending to be, an MTA, many of which are not handled by the 
> smtplib Python module, and many of them are subject to local policy,
> decided by the system administrator.  Sendmail, Postfix and Qmail are
> written to take care of those; you are unlikely to do a better job
> when writing an application whose primary purpose *isn't* being an
> MTA.

Agreed.  Where we seem to disagree is how to get the email to an MTA:

I suggest that invoking /usr/lib/sendmail -t via popen() is a poor 
approach for sending more than a trivial amount of mail.  By any chance, 
Thomas, have you ever tried delivering more than 1000 messages or so at 
a time by forking sendmail subprocesses?  :-)

> I know many other sysadmins that run their systems similarly to the
> way I run mine.  Any application that talks SMTP on their own instead
> of calling /usr/lib/sendmail to do it, is considered broken by us..

If doing so makes you happy, I guess that's okay.  Certainly, there are 
less reasonable positions that one could insist upon.

> So, when you write any program that sends email, at *least* give the
> option of configuring it to use /usr/lib/sendmail, and preferably to
> use that by default.  It is at least as likely to work out of the box
> as any other setting is.

Unless you write code for Windows, or Unix flavors that use a different 
path to sendmail, or use qmail/smail/postfix/etc _instead_ of sendmail.

[ Or M$-Exchange, DEC's Pathworks for VMS SMTP connector, Cisco's 
wretched SMTP firewall proxy (MailGuard?), or any number of other things 
that handle SMTP the way a "Do Not Fold, Spindle, or Mutilate" stamp on 
a real letter seems to be handled: ie, by redirecting such email to the 
virtual equivalent of a paper shredder.  :-) ]

Portability sometimes means more than just "This works under Red Hat, 
SuSE, and Debian."

-- 
-Chuck





From leomasi at libero.it  Tue Apr 22 13:10:57 2003
From: leomasi at libero.it (QWE)
Date: Tue, 22 Apr 2003 17:10:57 GMT
Subject: Keyboard Commands
References: <Pine.SOL.4.51.0304201825001.6411@herald.cc.purdue.edu> <81l7avg7r77ko9p80mggjlfdmb2bd98b51@4ax.com> <d62f103e.0304211736.3330b672@posting.google.com>
Message-ID: <fjtaavcm6cju3tu1e0gfdfbcbs3qc3411c@4ax.com>

On 21 Apr 2003 18:36:15 -0700, dick at charts.org (Rich) wrote:

>OK that helps some but it still isn't working perfectly.
>is this right:
>
>    self.Game.bind('<Control-h>',self.HighlightKey())
>
>    def HighlightKey(self):
>        self.FileOptionsHelp.Highlight()
>     	if self.Yote.HighlightEnable == 1:
>	    self.FileOptionsHelp.Highset.set(1)
>        else:
>	    self.FileOptionsHelp.Highset.set(0)
>    def Highlight(self):
>        if self.GameScreen.Yote.HighlightEnable == 1:
>            self.GameScreen.Yote.HighlightEnable = 0
>        else:
>            self.GameScreen.Yote.HighlightEnable = 1
>
>
>Shouldn't hitting ctrl-h cause HighlitghtEnable to switch values?

try
self.Game.bind('<Control-h>', self.HighlightKey)

or
self.Game.bind('<Control-h>', lambda e: self.HighlightKey())

if is not work, see chapter 7 of 'An Introduction to Tkinter'



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Tue Apr 29 18:04:11 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Wed, 30 Apr 2003 00:04:11 +0200
Subject: apache modules in python ?
In-Reply-To: <pan.2003.04.29.22.09.59.56666@sdf-eu.org>
References: <pan.2003.04.29.22.09.59.56666@sdf-eu.org>
Message-ID: <3eaef6dc$0$49110$e4fe514c@news.xs4all.nl>

Geirgat Jonas wrote:
 > Is it possible to write modules for the apache http server in python
 > ?
 >

Not quite the same, but I think you might be interested in mod_python:
http://www.modpython.org

"Mod_python is an Apache module that embeds the Python interpreter within the 
server. With mod_python you can write web-based applications in Python that 
will run many times faster than traditional CGI and will have access to 
advanced features such as ability to retain database connections and other 
data between hits and access to Apache internals."

--Irmen de Jong




From tundra at tundraware.com  Mon Apr  7 18:03:53 2003
From: tundra at tundraware.com (Tim Daneliuk)
Date: 07 Apr 2003 22:03:53 GMT
Subject: [Slightly OT]: More on ints and floats
Message-ID: <fjss6b.1021.ln@boundary.tundraware.com>

OK, I don't want to resurrect another interminable discourse in ints,
floats, and how they are/ought/can be handled in the language. But, I
have a sort of theoretical mathematical question which the whole
business brought to mind.

As I understand it, integers and floats are distinct mathematical
entities. A colleague of mine claims, that insfar as we use them in
computing, ints are merely a proper subset of floats. He furthermore
asserts that (again as regards to computing) the distinction between
them was made as a purely as a practical matter because floating point
arithmetic was historically computationally expensive. He argues that
any place one can use an int, these days (with cheap FP hardware), one
could use a float 0 extended to the precision of the machine and get
equivalent computational results. Say the hardware supported 4 digits of
precision. He is arguing that:


    3/4.0000  is equivalent to 3.0000/4.0000

(Never mind the old Python modulo vs. divison debate.)

In effect he is saying that, unless there is a practical
performance/cost issue at hand, there is no real reason to
differentiate between ints and floats in practical programming
problems.

As a matter of 'pure' mathematics, I argued that ints and floats are
very different critters. My argument (which is no doubt formally very
weak) is that the integer 3 and the float 3.0000 are different because
of the precision problem.  For instance, the integer 3 exists at a single
invariant point on the number line, but 3.0000 represents all numbers
from 2.99995 through 3.00004.

Could one of the genius mathematicians here bring some clarity to this
discussion, please?

TIA,
-- 
----------------------------------------------------------------------------
Tim Daneliuk     tundra at tundraware.com
PGP Key:         http://www.tundraware.com/PGP/TundraWare.PGP.Keys.txt




From alloydflanagan at attbi.com  Fri Apr  4 10:56:15 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 4 Apr 2003 07:56:15 -0800
Subject: Roman assignments (Re: Overloadable Assignment PEP)
References: <9383481e.0304012011.78f05548@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <a8b7f07a.0304031432.3c585657@posting.google.com> <b6in8i$600fi$1@ID-169208.news.dfncis.de> <b6j2ks$8gg$1@panix2.panix.com>
Message-ID: <a8b7f07a.0304040756.61276087@posting.google.com>

aahz at pythoncraft.com (Aahz) wrote in message news:<b6j2ks$8gg$1 at panix2.panix.com>...
...
> ...and if we add a
> 
>     from __past__ import FORTRAN
> 
> it only applies when the names are I, J, K, L, M, or N.

This is brilliant!!  But how are the implementors of the sets module
going to feel?  They could have skipped all that work and just done

from __past__ import Pascal



From mykbourassa at cogeco.ca  Tue Apr  8 16:14:15 2003
From: mykbourassa at cogeco.ca (myk)
Date: Tue, 08 Apr 2003 16:14:15 -0400
Subject: Pathname problem
In-Reply-To: <3e9321b5$1@rmcnt03a.rmc.ca>
References: <3e9321b5$1@rmcnt03a.rmc.ca>
Message-ID: <JaGka.1231$h%2.212680@read1.cgocable.net>

Many thanks :)

bourassa wrote:

> Hi there.  I am just trying to load files using PIL.  I get the 
> following:
>
> im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
> Traceback (most recent call last):
>   File "<pyshell#13>", line 1, in ?
>     im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
>   File "C:\Python22\Lib\site-packages\PIL\Image.py", line 1479, in open
>     fp = __builtin__.open(fp, "rb")
> IOError: [Errno 2] No such file or directory: 
> 'C:\\MATLAB6p1\toolbox\\pout.tif'
>
> I have stepped through and if the file in C:\ or C:\MATLAB6p1 it loads 
> with no problem.  The minute I try to get it out of "toolbox" (which 
> exists) I get the above error.  I'm using  PIL-1.1.4a3.win32 on a 
> Win2K PIV.  TIA.
>
> majb
>




From tic at dtek.chalmers.se  Fri Apr 18 15:42:18 2003
From: tic at dtek.chalmers.se (Mikael Jansson)
Date: Fri, 18 Apr 2003 21:42:18 +0200
Subject: ncurses and thread problems
Message-ID: <Pine.GSO.4.10.10304182137100.13710-100000@licia.dtek.chalmers.se>

Hi there,

I have some code where I put my window's loop for retrieving user input
via a getch() loop in a thread, so I can go on doing other things in the
mean time.

Now, the problem is that for /some/ reason, I don't get the ESC-sequence
mangling from getch() if I call it within a thread, which I /do/ get if I
have no other thread than the main application. I.e.: I have to check for
27, [ and A to get KEY_UP, instead of checking against KEY_UP directly.

I tried setting various values for ESCDELAY (500,1000,2000,10000), but
still no go.

Why's that?


I'm using Python-2.2.2 on Debian GNU/Linux-3.0 and BeOS R5/BONE.



Thanks in advance,
-- Mikael J.





From max at alcyone.com  Wed Apr 16 19:20:02 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 16:20:02 -0700
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com> <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com> <a8b7f07a.0304161330.1979b409@posting.google.com>
Message-ID: <3E9DE522.23C3A1C7@alcyone.com>

"A. Lloyd Flanagan" wrote:

> An assert is technically a
> statement that a certain condition can't happen under any
> circumstances.  It's used to verify that the code in the function has
> the expected effect on the state of the program.  But being called
> with an incorrect parameter is NOT a flaw in the code that contains
> the assert -- it's a problem with the caller.  Basically, instead of
> saying, "hey, buddy, that parameter is wrong" you're now saying
> "whoops, there's a bug in my code somewhere".  For a bad parameter you
> should raise ValueError or something similar, not AssertionError.

Some programming houses break asserts into two types:  logic asserts and
data asserts.  Logic asserts are the traditional "can't happen; if it
did something is disastrously wrong" variety, but data asserts are the
"you sent me bad data, dummy" form.  This way you select differing
behavior (ignore them completely, print a warning message in a log file,
or abort immediately) on each of them separately.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Would a harvest from my heart / Find its season
\__/ Sandra St. Victor
    Python chess module / http://www.alcyone.com/pyos/chess/
 A chess game adjudicator in Python.



From gh at ghaering.de  Sat Apr  5 14:21:56 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Sat, 05 Apr 2003 21:21:56 +0200
Subject: List of Numbers
In-Reply-To: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
Message-ID: <3E8F2CD4.6090307@ghaering.de>

Simon Faulkner wrote:
> I have a list of about 5000 numbers in a text file - up to 14 digits
> each.
> 
> I need to check for duplicates.
> 
> What would people suggest as a good method?

Use the numbers as keys for a dictionary. Before inserting, check if 
such an entry already exists. If this key exists already, you have found 
a duplicate.

-- Gerhard





From bbfoto2002 at netscape.net  Tue Apr  1 22:24:40 2003
From: bbfoto2002 at netscape.net (BBFoto)
Date: Wed, 02 Apr 2003 03:24:40 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org>
Message-ID: <3e8a586c.56142743@news-server>

"module not found"

Thanks for trying to help.

BB

On Wed, 2 Apr 2003 00:05:10 +0400, "Raseliarison nirinA"
<nirina at mail.blueline.mg> wrote:

>"BBFoto" <bbfoto2002 at netscape.net> wrote:
>
>> I run the Python executable and type in "PySol" at the prompt on the
>> black screen and get "name not defined."
>
>What happens if you type:
>import PySol
>
>There is a colorful screen at
>C:\Python23\Tools\idle\idle.py
>
>Hope this help
>
>nirinA
>
>




From andersjm at inbound.dk  Wed Apr  2 16:27:45 2003
From: andersjm at inbound.dk (Anders J. Munch)
Date: Wed, 2 Apr 2003 23:27:45 +0200
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <x7of3qjh4g.fsf@guru.mired.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com>
Message-ID: <wBIia.2000$y3.310439@news010.worldonline.dk>

"Dave Benjamin" <ramen at lackingtalent.com> wrote:
> In article <b6ckjr$903$1 at panix2.panix.com>, Roy Smith wrote:
> > The bottom line is exactly what Mike said: in interpreted languages,
> > the more you can push down into the libraries and interpreter core,
> > the better off you are most of the time.  And profiling helps too :-)
>
> This has been my experience as well. It is a very gratifying discovery,
> IMHO, because it encourages a less-is-more style of programming where you
> say what you mean, and little else.

The downside is that sometimes code is contorted to fit whatever has a
fast high-level implementation in the current interpreter, instead of
adapting to the nature of the problem.

For a Python example, see the Decorate-Sort-Undecorate pattern.  In
C++ the straightforward and fast solution is to use a suitable
comparison predicate.  In Python you can do the same, but that's slow,
and you are left with a choice between clarity and speed.

That coding style tends to adapt to performance accidents of the
implementation, happens in all interpreted languages.  It's a
particular problem in TIMTOWTDI languages where chosing a way then
becomes a micro-optimisation instead of a clue to the program's
structure.

- Anders






From andersjm at dancontrol.dk  Fri Apr 11 08:25:54 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Fri, 11 Apr 2003 14:25:54 +0200
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050048561.24768.python-list@python.org>
Message-ID: <3e96b435$0$66259$edfadb0f@dread11.news.tele.dk>

"Christian Tanzer" <tanzer at swing.co.at> wrote:
> Carl Banks <imbosol-1049990205 at aerojockey.com> wrote:
> > I think dynamic scpoing is a wholly bad idea, and simplifying a very
> > misguided programming practice doesn't make it any less so.
> 
> You are of course right. Surprisingly though, it works very well for
> Emacs. Passing arguments would *not* work in the context of Emacs.
> 

What makes you say that?  Why wouldn't lexical scoping supplemented by
a temporary-global-binding construct work just as well or better?

- Anders





From bellman at lysator.liu.se  Tue Apr  1 14:12:49 2003
From: bellman at lysator.liu.se (Thomas Bellman)
Date: Tue, 1 Apr 2003 19:12:49 +0000 (UTC)
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it>  <bd993a2f.0303310333.5a95ecb3@posting.google.com>  <roy-D262FA.09230131032003@reader2.panix.com>  <b6aam1$7o6$1@news.island.liu.se> <uk7ef3w3c.fsf@boost-consulting.com>
Message-ID: <b6cobh$hck$1@news.island.liu.se>

David Abrahams <dave at boost-consulting.com> writes:

> Thomas Bellman <bellman at lysator.liu.se> writes:

>> After a while it was realized that that was untenable, and
>> templates were invented to patch around it.

> That's a very interesting view of the origin of templates!  Did you
> know there's a book by Bjarne Stroustrup called "The Design and
> Evolution of C++" where you can read the real story?

Yes, I actually own that book myself (despite hating C++ :-),
though it has been several years since I read it.  It says that
the motivation was for container classes, but IMHO that's
essentially the same problem: you want to use types/classes with
different inheritance in the same functions or data structures,
as long as the interface is the same.  (For pure container
classes, the needed interface typically is just "you can get a
pointer to it" or something similar.)

And of course, Bjarne doesn't consider templates as a kludge to
work around a basic design deficiency, but that's his problem. :-)


-- 
Thomas Bellman,   Lysator Computer Club,   Link?ping University,  Sweden
"God is real, but Jesus is an integer."      !  bellman @ lysator.liu.se
                                             !  Make Love -- Nicht Wahr!



From http  Sun Apr  6 13:27:41 2003
From: http (Paul Rubin)
Date: 06 Apr 2003 10:27:41 -0700
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <7xn0j3fsdu.fsf@ruckus.brouhaha.com>

pobrien at orbtech.com (Patrick K. O'Brien) writes:
> If you were crazy enough to think that you could create a program
> along the lines of Emacs, but written in Python, how would you go
> about it?  How would you design the domain model for files, buffers,
> windows, and frames?  How would you allow the same level of
> customizability?  How would you map functions (or methods or whatever)
> to keys?  Any thoughts?

I think you're going to need some C extensions to represent buffers
similar to how Emacs does it.  It's hard to move blocks of characters
around in Python with the array module without constantly re-allocating 
the arrays.  Of course it can't be done with strings because Python
strings are immutable.

Aside from that, the general architecture can probably be similar to
Emacs.  You might want to use a multi-threaded interpreter to make
extensions more flexible (e.g. so they can respond to keystrokes while
doing other stuff).  You generally won't need as much C code as Emacs
used, because computers are much faster now than when Emacs was
written.  Python and Emacs Lisp are fairly similar in speed, I think.




From delete.this.part.alex.news at attbi.com  Wed Apr 30 15:50:25 2003
From: delete.this.part.alex.news at attbi.com (Alex)
Date: Wed, 30 Apr 2003 19:50:25 GMT
Subject: question on the compile built in function.  Filename ignored?
References: <7rSra.706428$S_4.744795@rwcrnsc53> <m31xzjhm6a.fsf@mira.informatik.hu-berlin.de>
Message-ID: <5OVra.708775$S_4.745564@rwcrnsc53>

Martin v. L?wis wrote:

> Alex <delete.this.part.alex.news at attbi.com> writes:
> 
>> From the docs I would expect the line
>> 
>>   File "<string>", line 1
>> 
>> to read
>> 
>>   File "abc", line 1
>> 
>> 
>> What am I missing?
> 
> You are right: this is what should have been printed. You have found a
> bug.
> 
> Actually, you are not the first one to find this bug; it will be fixed
> in Python 2.3.
> 
> Regards,
> Martin

Ok.  I will ignore it for now and/or upgrade to python 2.3 beta.

My thanks to everyone.

Alex





From lekishvili at python.qartu.com  Fri Apr  4 12:04:37 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 4 Apr 2003 09:04:37 -0800
Subject: Pythonic way of web-programming
Message-ID: <a3b5b608.0304040904.146dfff5@posting.google.com>

Hi all!

Let me start with the statement that I am not a professional
web-programmer. My tasks are rather linked with the scientific
calculations and algorithm prototyping.

Recently I have started to make a Georgian language based web-site for
Python. It was a natural feeling not to use, say, php or perl for
web-content management. I have gone through different approaches
available in Python; namely:
- CGI (with or without mod-python and / or cpy ) ;
- ZOPE ;
- Jython / J2EE ;
- ASP .NET ;
- Webware PSP.

I didn't examine all of them in depth as my real choice was somehow
restricted with either CGI or (utter maximum), webware appserver,
(which still has to be set up in a due manner by my ISP provider; I am
doing the site as they are willing to promote Python in Georgia).

My opinion, or rather feeling, is that "there is more than one way of
doing this"; something utterly unpythonic


Hence my questions to the Python web-programming gurus:

- Is Python a real web-programming language at all? I mean, has
somebody ever used it for real, commercial tasks? Please, do not
consider my question too trivial. There is no language equally
suitable for all tasks, however, Python seems to me a language
naturally tailored for web-programming; this or that way, I have more
than two years of experience of professional programming in Python.

- If it is an intention of the Python community to keep or to
establish Python as a web-programming language on the level of Java,
C# and Co., what is the standard, or at least, recommended way to
web-program in Python? Which technology?

- Hasn't yet come the time to have the only *standard* way of
web-programming in Python? For example, will anybody ever use Java for
web-programming outside the J2EE framework? In Python, and I still
emphasize, it might be only my feeling, there are lots of ways to do
this job.

- Is there a need to design a genuine python framework for
web-programming? I admit that it should sound weird and fuzzy and that
I myself have no exact idea of how to do so, however, a mere porting
of technologies compels us to program, say, Java, in Python. Why
should we keep doing this? Python itself is one of the most powerful,
clear and emotionally beautiful language.
Or, maybe, ZOPE is the genuine pythonic way? I simply do not know. I
do not want to come with some "new ideas" simply not feeling myself
confident. Therefore, please, give me some hints where to start
web-programming in Python.

Thank you very much for your attention.

Best wishes,
Giorgi



From max at alcyone.com  Wed Apr 23 17:05:24 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 23 Apr 2003 14:05:24 -0700
Subject: True/False
References: <mailman.1051104517.3339.python-list@python.org>
Message-ID: <3EA70014.3E02A0A5@alcyone.com>

Skip Montanaro wrote:

> What does this have to do with conditional assignment to True and
> False?

>From the context, which has been excessively discussed, it should be
obvious.  There has been a great deal of talk recently about doing the
same SyntaxWarning protection for True and False.

> Assigning to True or False today doesn't produce a SyntaxWarning, and
> given
> that it's rather common to define them in programs, it's likely that
> it
> won't for a long time.

It came out that it was decided that any such change (if ever
implemented) would be held off for a long time, but it was not clear at
the time it was discussed and the beginning of this thread, when the
topic was being addressed.

It also came out that if a SyntaxWarning were to be added for True/False
manipulations, they'd be done at parse time before anything gets
executed so the try: True; except NameError: ... trick wouldn't work
anyway.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If I had known, I would have become a watchmaker.
\__/ Albert Einstein
    7 Sisters Productions / http://www.7sisters.com/
 Web design for the future.



From jdominion at gmx.de  Fri Apr 25 07:25:31 2003
From: jdominion at gmx.de (jdominion at gmx.de)
Date: Fri, 25 Apr 2003 13:25:31 +0200 (MEST)
Subject: embedding python2.2 in C++/Qt
Message-ID: <23049.1051269931@www49.gmx.net>

hi!

Who has up-to-date experience in embedding python into C++?  All examples i
found in the internet or in the (not recommendable) book of Himstedt/M?tzel
are for python version 1.5.

I include the Python.h and cannot perform the command Py_Initialize(). what
is the problem?
I know, this is more a C++-question. but  everybody promise that it is "very
easy" to embed/extend python.

jan

-- 
+++ GMX - Mail, Messaging & more  http://www.gmx.net +++
Bitte l?cheln! Fotogalerie online mit GMX ohne eigene Homepage!





From escalation746 at yahoo.com  Wed Apr  2 11:29:15 2003
From: escalation746 at yahoo.com (robin)
Date: 2 Apr 2003 08:29:15 -0800
Subject: distributed computing implementations
Message-ID: <a626bbd7.0304020829.1e5cac2f@posting.google.com>

I am attempting to summarise the distributed computing implementations
available to Python programmers. My conclusions so far are as follows:

If you want the biggest, boldest approach and don't care about
overhead, use CORBA.

If you want a simpler approach but with performance and implementation
difficulties, use SOAP or XML-RPC.

If you want the ultimate in simplicity and are willing to foresake
multi-language support, use Dopy, Pyro, or Twisted Spread.

Comments? Specifically, what are the relative advantages and
disadvantages of the last three mentioned products?

-- robin



From gh at ghaering.de  Wed Apr  9 05:28:23 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 09 Apr 2003 11:28:23 +0200
Subject: Making python module using Borland C++ 5
In-Reply-To: <b70jiq$men$1@news.cc.tut.fi>
References: <b6u7fp$d2c$1@news.cc.tut.fi> <b6ujd4$pe5$1@news.hccnet.nl> <b70jiq$men$1@news.cc.tut.fi>
Message-ID: <3E93E7B7.9010801@ghaering.de>

Rami Saarinen wrote:
> Anton Vredegoor wrote:
>>> ImportError: dynamic module does not define init function (initpyclips)
>>
>> Please have a look at:
>>
>> http://www.cyberus.ca/~g_will/pyExtenDL.shtml
>>
> I finally managed to build the module using bcpp. Thank you Anton and 
> Anders. It was all about including the EXPORT initpyclips=_initpyclips 
> in a def file (see the link above).

Why don't you just use disutils? It will work without any further setup, 
then. All you need to read is the part of the Python documentation about 
using non-MS compilers on Windows.

-- Gerhard






From dickey at saltmine.radix.net  Sun Apr  6 19:26:17 2003
From: dickey at saltmine.radix.net (Thomas Dickey)
Date: 6 Apr 2003 23:26:17 GMT
Subject: curses: xterm vs. linux terminal
References: <3E9065FE.1090108@sympatico.ca>
Message-ID: <b6qd2p$ib9$1@news1.radix.net>

Brian Alexander <brian094 at sympatico.ca> wrote:
> Hello;

> Why is it that the curses special (alt) characters do not display on xterms?

> Ex:

> theScreenObj.box(curses.ACS_VLINE, curses.ACS_HLINE)
> theScreenObj.refresh()

> This code puts a box with graphical characters around the window when 
> the terminal is of type linux. When the terminal is of type xterm, 
> blanks appear where the line characters should be.

> Is there anyway to fix this behaviour? I find it hard to believe an 
> xterm does not support the drawing of these simple graphical characters.

normally it doesn't (linux console doesn't draw lines either).

I modified XFree86 xterm a few years ago to draw lines when the font
does not contain them; you're apparently not using XFree86 (or else
thinking that "rxvt" is "xterm" ;-).

-- 
Thomas E. Dickey <dickey at radix.net> <dickey at herndon4.his.com>
http://dickey.his.com
ftp://dickey.his.com



From jeremy at zope.com  Mon Apr  7 15:27:46 2003
From: jeremy at zope.com (Jeremy Hylton)
Date: Mon, 07 Apr 2003 15:27:46 -0400
Subject: subclassing "object" class in C?
In-Reply-To: <slrnb93cj7.193.mac@mac.dgp.toronto.edu>
Message-ID: <web-1441295@digicool.com>

On Mon, 07 Apr 2003 18:18:04 GMT
 Maciej Kalisiak <mac at die.spammer.die.dgp.toronto.edu>
wrote:
> How does one subclass the "object" class in C?  A lot of
> examples I've seen of
> writing picklable subclasses in Python first subclass
> "object" (e.g. PEP 307).
> I'd like to do this in C, as that's where I'm writing an
> extension class/type.

Actually, you should make your C type a subclass of
PyType_Type.  This is the C equivalent of saying 
"class C(object):".

You might look at the datetime classes, implemented in C,
as examples of how you would write a picklable type in C.
 In my very limited experience, it is harder than I
expected.

One thing you can do is define a helper function that makes
new instances and define a __reduce__ method on your class
that returns this helper function as the first argument.

Jeremy




From staschuk at telusplanet.net  Tue Apr 29 16:00:56 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 14:00:56 -0600
Subject: Where did you learn to unit test?  How did you learn?
In-Reply-To: <1147e466.0304290753.5cbf3c8c@posting.google.com>; from blunck@gst.com on Tue, Apr 29, 2003 at 08:53:03AM -0700
References: <1147e466.0304290638.287687f6@posting.google.com> <b8m3cj$8vu$1@ginger.mathworks.com> <1147e466.0304290753.5cbf3c8c@posting.google.com>
Message-ID: <20030429140056.A1336@tibia.amotlpaa.bogus>

Quoth Christopher Blunck:
  [Josh suggests writing unit tests to tickle bugs]
> We do that already.  It works to a certain extent.  But usually
> requires brow beating of the bug fixer.  I guess we really just need
> some time to pass.  Unit testing is an investment and by definition
> investments don't pay off immediately.

In the case of bug-tickling tests, one of the big selling points
is that they let you make sure bugs don't come back.  This is an
easier sell if you can demonstrate (say, by analysis of historical
bug reports) that in the past bugs have been fixed and come back
again.  Even better if you can determine how much time was spent
troubleshooting reintroduced bugs from scratch.

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From member27604 at dbforums.com  Thu Apr 17 21:43:51 2003
From: member27604 at dbforums.com (dolephi9080)
Date: Fri, 18 Apr 2003 01:43:51 +0000
Subject: [Que]Do any SGML lib good for Python?
References: <2778697.1050595278@dbforums.com> <WWAna.28468$LB6.652269@news1.tin.it>
Message-ID: <2781053.1050630231@dbforums.com>

Originally posted by Alex Martelli 
> dolephi9080 wrote:
>
> > 
> > I search  the SGML Lib for python, but it's hard to find. even
>     the
> > python contain a sgml parser itself. But that's a limite
>     parser.
> > Does some one know it? pls tell me. Thx
>
> Perhaps:
>
> http://www.garshol.priv.no/download/software/pysp/"]http://www.-
> garshol.priv.no/download/software/pysp/[/url]
>
> ???  Doesn't seem to be very actively maintained, but perhaps
> it doesn't need to be... I wouldn't know, it's ages since I last
> had to parse very general SGML input.  Worst case, you could
> re-wrap James Clark's SP, e.g. with Boost Python...
>
>
> Alex 


I download it, and run the setup.py, it seems not work, my python
version is 2.2.

--
Posted via http://dbforums.com



From marcus at infa.abo.fi  Mon Apr 14 09:47:22 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 14 Apr 2003 15:47:22 +0200
Subject: Complexity of standard Python data structures (Re: Speed of string += string)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi> <Oqwma.11524$LB6.319489@news1.tin.it>
Message-ID: <slrnb9lbbq.v4f.marcus@infa.abo.fi>

On Mon, 14 Apr 2003 10:53:34 GMT, Alex Martelli <aleax at aleax.it> wrote:
>Marcus Alanen wrote:
>> Indeed, that was the first thing I wondered about when I started
>> learning Python.
>Interesting!  What other languages (besides C++ for a subset of its
>standard libraries) give O() specifications for their operations?
>In other words, what other languages did you learn, before Python,
>that didn't leave you wondering about such performance issues?

It must've been a combination of issues:

1) Coming from a C/C++ background, I want to use data structures
efficiently.  This belongs to the "premature optimization" group. So
sue me.

2) The project I was doing could with time lead to handling quite big
sets of data, so algorithmic efficiency was a concern.

3) Python's "list" had on insert() routine, but also a __getitem__
routine. Both probably couldn't be O(1), so I got a bit confused.
I didn't want to choose a "slow" method on purpose.

4) And hey, I did want to learn the language properly.

Number 3 was quite irritating, since C++ "list" is a double-linked
list, while Python "list" is what is called a "vector" in C++.

Probably not very good excuses (except number 4), but I'm just telling
what I felt at the time. On the other hand, I don't understand why the
Python documentation simply can't say "speed O(N)", "speed amortized
O(1)" for the various methods.

>> Others answered this very thoroughly, I'd just like to emphasize the
>> linearity of keys() --- most often you can manage with .iterkeys() or
>> .itervalues() just fine, and it's a lot faster.

>As for the performance issues, I wouldn't be so sure.  E.g.:

Sorry, I really didn't explain that very well. In several cases I have
had, I don't need to traverse the whole array of dict(), so iterkeys()
does feel like saving some memory, and possibly speed. Naturally, both
functions in general visit all elements, and are linear in speed.

By the way, I tested a similar program with 10 million elements a >15%
speed gain with iterkeys():

d = {}
for i in xrange(10000000): d[i] = i

import time

for i in xrange(3):
        st = time.time()
        for i in d.iterkeys():
            pass
        et = time.time()
        print et - st

        st = time.time()
        for i in d.keys():
            pass
        et = time.time()
        print et - st

Definitely not a significant performance difference, though.

>That's most definitely NOT a significant performance difference, 
>of course, but calling iteration on iterkeys "a lot faster" than 
>iteration on keys would appear to be substantially misleading, at 
>least without a lot of qualifications and hedging!-)

True. Sorry for the confusion.

-- 
Marcus Alanen
maalanen at abo.fi



From exarkun at intarweb.us  Sun Apr 13 15:16:56 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 13 Apr 2003 15:16:56 -0400
Subject: Protected Methods and Python
In-Reply-To: <221d8dbe.0304131058.1ecd1c3c@posting.google.com>
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com>
Message-ID: <20030413191656.GB12415@meson.dyndns.org>

On Sun, Apr 13, 2003 at 11:58:36AM -0700, srijit at yahoo.com wrote:
> I do not see a  concept of Protected Methods in Python 2.2 like C++ and Ruby.
> Is this a serious limitation for Object Oriented Programing with Python? 

  Not in my experience.

> Or may be it is possible to emulated Protected Methods in Python.

  If you have methods you don't want people calling, simply do one of the
folloowing:

    1) Document them as such

    2) Prefix their names with "_"

    3) Don't document them at all

> I look forward to some discussion by OOP experts.

  It may just be me, but I don't see how this has anything to do with object
oriented design/programming.

  Jp

-- 
Somewhere, something incredible is waiting to be known.
                -- Carl Sagan
-- 
 up 24 days, 15:01, 1 user, load average: 0.99, 0.90, 0.85




From glenfant at NOSPAM.bigfoot.com  Tue Apr  1 05:48:38 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Tue, 1 Apr 2003 12:48:38 +0200
Subject: Stackless Perl 4.0
References: <mailman.1049160602.3367.python-list@python.org>
Message-ID: <b6bq31$1jd5$1@biggoron.nerim.net>

"Christian Tismer" <tismer at tismer.com> a ?crit dans le message de news:
mailman.1049160602.3367.python-list at python.org...
> Dear Python/Perl Community,
>

My newsreader troll finder plugin doesn't beep ! Why ?

Oh we're April 1st !

--Gilles




From bkc at Murkworks.com  Wed Apr  9 08:31:32 2003
From: bkc at Murkworks.com (Brad Clements)
Date: Wed, 9 Apr 2003 08:31:32 -0400
Subject: Sockets on PocketPC
References: <b711rb$15rb$1@news.uit.no>
Message-ID: <3e94135d_1@spamkiller.newsgroups.com>

It should be. I had forgotten to compile it in, but I think telion released
a version with sockets.. maybe I did.. I can't recall.

I haven't seen Telion anywhere on the lists for months.

Did you check the pythonce list?


--
Novell DeveloperNet Sysop #5

_
"Thomas Aanensen" <thoaan at stud.cs.uit.no> wrote in message
news:b711rb$15rb$1 at news.uit.no...
> Hi
>
> Have anybody used sockets on a PocketPC?
>
> I get an "ImportError: No module named _socket"
>
> Isn't sockets supportet in the PPCPython lib?
>
>
> Thomas
>
>





From aahz at pythoncraft.com  Sat Apr 26 21:15:47 2003
From: aahz at pythoncraft.com (Aahz)
Date: 26 Apr 2003 21:15:47 -0400
Subject: please comment on technologies
References: <slrnbalplq.254.luc@trillian.dont-panic.info>
Message-ID: <b8fb03$jmk$1@panix3.panix.com>

In article <slrnbalplq.254.luc at trillian.dont-panic.info>,
luc wastiaux  <luc at nospam.com> wrote:
>
>Hello, I am writing the specifications for a school project (the
>subject is free for us to choose), I was thinking about doing a
>news-server archive database just like google groups, but on a smaller
>scale (it would only archive a limited number of newsgroup, not all of
>USENET). I got this idea because my school uses a news server and a lot
>of valuable information is lost when old messages are erased from the
>server's spool.

One problem you *do* want to think about up-front is how you're going to
query the database.  Full-text searching can be a daunting task.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From abpillai at lycos.com  Sat Apr  5 07:00:51 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 5 Apr 2003 04:00:51 -0800
Subject: FTP Upload through Proxy
References: <cbdb2fee.0304040558.200651b8@posting.google.com>
Message-ID: <bd993a2f.0304050400.40526ea@posting.google.com>

I think you need to use ProxyHandler along with 'FTPHandler'
for authorization.

This is the link to a similar post with some answers. Mighe be
helpful.

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&frame=right&th=1cfe3e41fd22595b&seekm=bd993a2f.0303190641.28ca1072%40posting.google.com#link1

Regards

Anand Pillai


soulayrol at free.fr (David) wrote in message news:<cbdb2fee.0304040558.200651b8 at posting.google.com>...
> Hi.
> 
> I'm new to Python, and for my first app, I try among other things to
> upload a file with FTP.
> 
> I read many messages here about urllib2 and proxies, but it seems that
> urllib2 only authorize URL resolution and read access on content.
> On the other hand, there is ftplib, but there is no proxy related
> methods on it.
> 
> What is the good solution, since I don't know anything of my proxy
> here, and i'd like it to work almost everywhere ?
> 
> Thanks.
> David.



From rmunn at pobox.com  Wed Apr  9 18:08:40 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 09 Apr 2003 22:08:40 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain> <6j8u8v0hjh3n9dm1jjflu3a970ada6qho4@4ax.com> <slrnb933p7.46e.rmunn@localhost.localdomain> <0au59v0n82d016vs3mai634d4088uf9f0k@4ax.com>
Message-ID: <slrnb98hor.8ee.rmunn@localhost.localdomain>

carroll at tjc.com <carroll at tjc.com> wrote:
> On Mon, 07 Apr 2003 20:39:13 GMT, Robin Munn <rmunn at pobox.com> wrote:
> 
>>The two approaches you describe in the paragraph above are really the
>>same approach, in a *slightly* different disguise. When you say "if the
>>caller invoked it with the wrong object type, he'll get an error," I'm
>>guessing that your idea of the "wrong object type" is an object that's
>>not a subclass of "file". But in Python, the "wrong object type" would
>>be... (drumroll)... Any object type that didn't have a method named
>>"write"!
> 
> I don't look at it that way.  The doc for the method specifies that
> the parameter is (if not a string) a file object.  That means that the
> wrong object type is any object type that isn't a file object.

But that is not the Python philosophy. The Python philosophy is "if it
quacks like a duck, then treat it like a duck". :-) In other words, you
can create a class that *behaves* like a file object, and you should be
able to pass it to any function that expects a file object. When I see
the words "a file object" in the documentation, I understand it to mean
any object that behaves like a file object. If this were not the case, I
would complain.

> Now, it happens that the only other method used for this object is the
> write method, but that's not guaranteed.  If someone called the method
> using an object that is not a file object, but that has a write method
> (e.g., a StringIO object), sure, it would work.  For now.  In this
> release.  But if the nntp method were changed in the future to do
> something else with the file object that actually depended on it being
> a file object, it would fail.  

But if it were changed to actually check the type of the object, so that
it refused (for instance) to accept a StringIO object, then I (and
dozens of other Python programmers) would complain bitterly, because we
would feel that the documentation contract had been broken.

Yes: I would feel that the documentation contract had been *broken* if
"this parameter is a file object" did not mean "this parameter behaves
like a file object". If what was meant was "a file object, AND NO OTHER
TYPE OF OBJECT, is required here" then I would expect to see that
wording explicitly stated.

The reason for all this is to support all kinds of nifty polymorphism.
Passing a StringIO object instead of a file object, for example, can be
very handy when you want to profile a piece of code and want to
eliminate the slowness of accessing a filesystem. Or maybe you want to
set up for some unit tests. Or you want to capture the output and do
some processing with it, but you don't want to have to make a temporary
file, write it, close it, read it, and delete it, all to get the data
back into a string. And that's just what I came up with in thirty
seconds of thinking!

> I submit that the failure is because the user is not passing it an
> object of the type that it is documented to require, because it says
> pretty plainly that it requires a file object, not merely any object
> that includes a method named write.  Using a non-file object is just
> something you can get away with.

You are right that the documentation could have been phrased better. At
the least it could have said "a file-like object". See, for instance,
the specifications for the functions dump() and load() in:

    http://www.python.org/doc/2.2/lib/node61.html

>>The key concept to understand is this: in Python, types usually don't
>>matter. What matters is interfaces. If you have a function that expects
>>to be passed a file object, for instance, someone could pass it a
>>StringIO object (http://www.python.org/doc/2.2/lib/module-StringIO.html)
>>instead. 
> 
> FWIW, nntplib will work with StringIO (or anything else that has a
> write method) -- not because it does interface checking on the object
> passed, but because, after it's type-checked whether it's a string
> (and if it's a string, it creates a file object with open() on the
> file named in the string), it assumes that any non-string object it's
> been passed is a file object (if you rely on the docs) or a object
> that supports write() (if you read the code), and goes ahead and uses
> write().  If the object doesn't have a write method, you'll get an
> exception.

Looking at the source for nntplib, I see what you mean. It checks
whether the file parameter is an instance of str (this does have one
failing, that it will fail to catch Unicode objects -- if your
filesystem accepts Unicode filenames, this might cause a problem), and
if it's not a string, then it treats it like a file.

Note that it does *not* explicitly check whether it was passed a file
object. This is how it should be. I couldn't find it off-hand in a
little bit of Googling, but I do remember coming across a tutorial on
writing in Python for those used to other languages. It had some good
advice: Never explicitly check the type of your parameters. Instead,
just treat them as the appropriate type, and allow AttributeError (or
whatever exception) to be raised if an inappropriate type is passed.

> In any event, I'm not planning on re-writing NNTPlib's file support --
> I'm just extending its existing file support to some additional NNTP
> commands.

And now that I look at it, I see you won't have to do any type-checking
code yourself. That's really all I was trying to do was warn you off of
doing explicit type-checking. I hope I've been able to explain properly
why that's not the Python way.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From chrtoeller2 at freenet.de  Mon Apr 14 05:10:26 2003
From: chrtoeller2 at freenet.de (Chris)
Date: 14 Apr 2003 02:10:26 -0700
Subject: conversion of INTEGER into string
Message-ID: <a01b8ad6.0304140110.2b055605@posting.google.com>

Hello everybody,

seems that this problem is very easy to solve, but I have absolutely
no idea...

I want to convert an integer, e.g. 2000 into a string (i.e. "2000")
As far as I know, the chr() funktion only works with integers  in the
range of -256 to + 256

So, what is the right method ??


cheers


Chris
Rostock, Germany



From tjreedy at udel.edu  Tue Apr  8 21:29:56 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 8 Apr 2003 21:29:56 -0400
Subject: STRING TO LIST
References: <mailman.1049843053.19630.python-list@python.org>
Message-ID: <sNicneNGn5EZ6g6jXTWcpQ@comcast.com>

<jubafre at brturbo.com> wrote in message
news:mailman.1049843053.19630.python-list at python.org...
> Hello,
>
> I need to change a variable from string to list.
> > >>> x=str("[[12,13],[14,15]]")
> > must be a list like this:
> > [[12, 13], [14, 15]]

>>> eval('[[12,13],[14,15]]' )
[[12, 13], [14, 15]]

Terry J. Reedy





From vze4rx4y at verizon.net  Sat Apr 12 03:59:58 2003
From: vze4rx4y at verizon.net (Raymond Hettinger)
Date: Sat, 12 Apr 2003 07:59:58 GMT
Subject: Unit testing "templates"?
References: <v9fcrcmbaa0eca@corp.supernews.com>
Message-ID: <2IPla.1532$883.720@nwrdny02.gnilink.net>

"Francis Avila"
> I'm now writing tests for these methods which instantiate the class with
> known data and then call a method of that instance to get a known result.
> But it's getting repetitive, and I'm wondering if there is a better way.>

Whenever you're doing something repetitive, there's *always* a
better way.  For a good example, take a look at test_sets.py in Py2.3.
It takes advantage of inheritance to factor out common code.  In your
example, it would look like this (untested code ahead):

class DatatypeUnpack(unittest.TestCase):
    """Verify xdmcplib.Unpacker conversions."""
    def testUnpack(self):
        # unpack should return known result from known valid input
        for value, data in self.vdpairs:
            obj = xdmcplib.Unpacker(data)
            result = getattr(obj, self.method)()
            self.assertEqual(value, result)

Class DatatypeUnpackCard32(DatatypeUnpack):
    vdpairs = ( (0,'\x00\x00\x00\x00'), (1,'\x00\x00\x00\x01'),
                              (2,'\x00\x00\x00\x02'),
                              (65535,'\x00\x00\xff\xff'),
                             (4294967295L,'\xff\xff\xff\xff') )
    method = "unpack_card32"

Class DatatypeUnpackArray8(DatatypeUnpack):
    vdpairs = ( ('','\x00\x00'), ('Test','\x00\x04Test') )
    method = "unpack_array8"


Raymond Hettinger






From hat at se-46.wpa.wtb.tue.nl  Fri Apr 25 06:56:29 2003
From: hat at se-46.wpa.wtb.tue.nl (Albert Hofkamp)
Date: Fri, 25 Apr 2003 10:56:29 +0000 (UTC)
Subject: What means exactly "Memory error"?
References: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net> <mailman.1051198958.16669.python-list@python.org>
Message-ID: <slrnbai52u.8cv.hat@se-46.wpa.wtb.tue.nl>

On Thu, 24 Apr 2003 18:43:04 +0200, Bo M. Maryniuck <b.maryniuk at forbis.lt> wrote:
> On Thursday 24 April 2003 17:26, Tim Peters wrote:
>> [Bo M. Maryniuck. gets a traceback ending with
>>     File "/usr/lib/python2.2/site-packages/_xmlplus/dom/Event.py",
>>         line 93, in initEvent
>>     self.bubbles = canBubbleArg
>>     MemoryError
>> ]
>> I can't help with your app, but I can tell you what MemoryError means. 
> 
> [skipp]
> 
>> The most likely cause is that you've
>> simply run out of memory (#1 above) -- but that's more a symptom than an
>> explanation.  Don't expect it to be easy to resolve.
> 
> Yes, you're right: I'm just out of memory, though server has 4G RAM... :/ Have 
> you at least any ideas how to decrease/pack collected data to avoid fat 
> growing?

Do you need all data DOMmed at once?
You may be able to have one DOM tree at a time, dropping and reloading
everytime you switch file.
(or a bit more luxurious, cache the last n files).

Albert
-- 
Unlike popular belief, the .doc format is not an open publically available format.



From jzgoda at gazeta.usun.pl  Wed Apr  9 17:58:07 2003
From: jzgoda at gazeta.usun.pl (Jarek Zgoda)
Date: Wed, 9 Apr 2003 21:58:07 +0000 (UTC)
Subject: Nice quote
References: <just-60BB53.18302728032003@news1.news.xs4all.nl> <pan.2003.03.28.20.18.26.483935@jerf.org>
Message-ID: <b7251f$1ma$2@atlantis.news.tpi.pl>

Jeremy Bowers <jerf at jerf.org> pisze:

>>   """The Python people also piped to say "everything's just
>>      fine here" but then they always do, I really must learn that
>>      language."""
> 
> Would somebody amplify on this please? What XML library do we have that's
> easier to use then DOM or SAX for real world tasks that doesn't require
> reading the whole file into memory to work?

Is it a problem to read whole file to parse the XML? Really? I'm happy
using ReportLab's pyRXP and effbot's ElementTree (loosely based on
J.Clark's expat parser...) - they both use similar approach to build 
document structure. And both are fast, with pyRXP being faster but with
less intuitive interface.

-- 
Jarek Zgoda
http://www.zgoda.biz/|JID:zgoda at jabber.atman.pl|http://zgoda.jogger.pl/



From roy at unice.fr  Thu Apr 17 03:55:46 2003
From: roy at unice.fr (Jean-Paul Roy)
Date: Thu, 17 Apr 2003 09:55:46 +0200
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr>
Message-ID: <roy-8E37D1.09554617042003@malibu.unice.fr>

Hi, snakes,

I'm teaching Scheme in our Univ. to 2nd year students. The cursus is now:
- Java in 1st year
- Scheme in 2nd year
- Java, C++... in 3rd year and above (+ some Scheme to build an 
interpreter).

Now, is it a good move to swap from Java to Python in 1st year ? Some of 
us find Java a bit heavy for algorithmics. The probable discussion will 
be about:

- Python is bad : no static types (it's ok for me)
- Pyhton is bad : recursion limit (I am astonished at that one)
- Python is bad : tail recursion is not iterative (also astonished at 
that, I can understand with an interpreter, but compiler ?)
- Python is bad : we are obliged to use message passing paradigm to add 
an element to a list (it's ok for me but some of us get angry with 
objects for beginners)
- Python is bad : where are classical arrays ? Does:
     from array import array
  actually imports contiguous arrays ? How can they stay contiguous if
  they support adding an element ? Is the access only O(1) amortized ?
  Comes from : wanting to play with big arrays of bitmap images.

Aside from that points, a personal question as a Schemer who reads the 
Python primer:
- how do you ask if a list is empty ? Is the following ok ?
      if L == [] :
- how do you ask if the list has only one element, with time 0(1) ? The
  Scheme/Lisp analog is (null? (cdr L))
- more generally, is the length of a list a builtin ? Is the L[k] access
  time O(1) with a list ? Probably not. And with a tuple ?

Sorry for so many questions and your time.

But thanks a lot, I guess that Python is a good guy :-)

    -jpr



From aussie2010 at yahoo.com  Tue Apr  1 03:14:38 2003
From: aussie2010 at yahoo.com (Chris)
Date: 1 Apr 2003 00:14:38 -0800
Subject: Comments
Message-ID: <435d40e0.0304010014.503f9f8f@posting.google.com>

I would like to request that Python support multi-line comments (e.g. /* */ in C).



From aahz at pythoncraft.com  Tue Apr 15 08:04:04 2003
From: aahz at pythoncraft.com (Aahz)
Date: 15 Apr 2003 08:04:04 -0400
Subject: Python FAQ (Was: Re: Confusion: assignments create object references.)
References: <mailman.1050080058.23175.python-list@python.org> <3E970241.483698C7@engcorp.com> <slrnb9kveb.1q1.rmunn@localhost.localdomain>
Message-ID: <b7gsfk$98n$1@panix1.panix.com>

In article <slrnb9kveb.1q1.rmunn at localhost.localdomain>,
Robin Munn  <rmunn at pobox.com> wrote:
>
>In fact, the Python FAQ needs some trimming, IMHO. With the "Programming
>in Python" section containing over 100 questions, few are going to have
>the patience to sit down and read the entire list to see if their
>question is on the list. Thus we get the same questions over and over in
>the newsgroup. The FAQ is not serving its purpose.
>
>I am not suggesting we get rid of the information currently contained
>in the FAQ, but we should probably refactor the FAQ into two levels: one
>for the truly frequently-asked questions, and one for "Here are more
>questions that crop up from time to time; look through this for useful
>nuggets of information." I think a shorter FAQ would be read a lot more
>often.

"What 'we', white man?"

The python.org webmasters would love for someone to do this work.  Any
volunteers?....
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From max at alcyone.com  Wed Apr 23 03:27:56 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 23 Apr 2003 00:27:56 -0700
Subject: unused args in GUI callbacks: pychecker
References: <20030422145218.6e9e6549.gry@ll.mit.edu> <3EA5B0C4.7ACB349@alcyone.com> <slrnbaccg8.du4.simo_salminen@pikkukorppi.cs.tut.fi>
Message-ID: <3EA6407C.337EAE99@alcyone.com>

Simo Salminen wrote:

> I think he ment this:
> int foo(int){}  <- no formal parameter name

That occurred to me as well, but from the context I got the impression
he was talking about preexisting code.  In retrospect, you're probably
right.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Scars are like memories.  We do not have them removed.
\__/ Chmeee
    Lsystem / http://www.alcyone.com/pyos/lsystem/
 A Lindenmayer systems explorer in Python.



From hupp at cs.wisc.edu  Tue Apr 29 17:54:48 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Tue, 29 Apr 2003 16:54:48 -0500
Subject: os.spawn
In-Reply-To: <Xns936CF02836B77jwolffenertelnl@194.109.133.29>
References: <Xns936C9D9E91B1Bthewolffman@194.109.133.29> <mailman.1051624647.29525.python-list@python.org> <Xns936CF02836B77jwolffenertelnl@194.109.133.29>
Message-ID: <20030429215448.GA2679@upl.cs.wisc.edu>

On Tue, Apr 29, 2003 at 09:36:07PM +0000, Jeroen Wolff wrote:

> Thx, but i need the .txt file because i have to parse it several time...
> with os.system( ) is the script waiting until the command is finished?
> I have to wait until i can open the generated .txt file

Yes, os.system will block until the command is completed.  If you need
to run through the data multiple times you can just store a list of
the lines though:

lines = os.popen("command").readlines()

-Adam




From peter at engcorp.com  Tue Apr  1 06:15:49 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 06:15:49 -0500
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com>
Message-ID: <3E8974E5.88B861D7@engcorp.com>

Chris wrote:
> 
> I would like to request that Python support multi-line 
> comments (e.g. /* */ in C).

There are at least two reasons why someone would want to do
that, and of course it's been suggested before, but what's
your rationale behind this?  There are generally simple 
answers that address each possible use case, but perhaps you've
actually got an original idea about this...

-Peter



From yann.morin.1998 at anciens.enib.fr  Sun Apr  6 10:16:40 2003
From: yann.morin.1998 at anciens.enib.fr (Yann E. MORIN)
Date: Sun, 6 Apr 2003 16:16:40 +0200
Subject: Install bug : incorrect permissions
Message-ID: <20030406142002.ABD3D40E30@kraid.nerim.net>

Hi all!

(sorry, I'm not subscribed)

I've found that python 2.2.2 installs its lib-dynload files with what I think
are incorrect permissions : 0750. That prevent any user, safe root and the
group, to use those files, like the expat module.

Changing manualy to 0755 solves this.

I've built on LFS-4.0 with BLFS cvs from 20030405. lib-dynload is
/usr/lib/python2.2/lib-dynload

Regards,
Yann.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| --==< ?_? >==-- ?---.----------------:  X  AGAINST      |  \e/  There is no  |
| web: ymorin.free.fr | SETI at home 1187 | / \ HTML MAIL    |   v   conspiracy.  |
?---------------------?----------------?------------------?--------------------?
 \__ np: 22 - P... de toi ____________________________________________________/




From DennisR at dair.com  Tue Apr 22 00:02:48 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Tue, 22 Apr 2003 04:02:48 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com> <mailman.1050978337.8048.python-list@python.org>
Message-ID: <I93pa.1054$JB2.90303532@newssvr14.news.prodigy.com>

>   Is the app that generates these available (commercially or for free)?

(1)  http://www.dair.com/autorun/arpyexe.cgi
Use of the web page is free.

(2) A serious app (SpamAI sig link below)
I packaged the Python tree manually.  I started with the main DLL and then
included modules as Python found them not present in the tree at the time.
Packaging = Python = Free.  The trick here is to assign system.path and
prevent Python from searching its own install, using only the tree being
constructed.

I used TinySetup (www.tinysetup.com) for setup tasks.  Free after
registration (tinysetup is my site, btw).  I used WinRAR (www.rarlabs.com)
for self-extracting compression.  Commercial = $29
--

Dennis Reinhardt

DennisR at dair.com   http://www.spamai.com?ng_sig





From rob at zoism.org  Sat Apr  5 11:48:33 2003
From: rob at zoism.org (Rob Brown-Bayliss)
Date: 06 Apr 2003 04:48:33 +1200
Subject: segfault with imp.load_module()
In-Reply-To: <pan.2003.04.05.13.23.43.694594.12686@metaslash.com>
References: <mailman.1049534708.5967.python-list@python.org>
	 <pan.2003.04.05.13.23.43.694594.12686@metaslash.com>
Message-ID: <1049561326.22972.14.camel@musicbox>

On Sun, 2003-04-06 at 01:23, Neal Norwitz wrote:

> What version of python, what platform, what compiler, what options?
> Assuming item = 'oda' this code works for me.

Redhat 8
[rob at musicbox rob]$ python
Python 2.2.1 (#1, Aug 30 2002, 12:15:30)
[GCC 3.2 20020822 (Red Hat Linux Rawhide 3.2-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.


If item = "abc" then I get (as expected) ImportError: No module named
abc

So any way, I did some more playing and my module had a line:

from bhh import verbose

But the bhh module is not in the path, it's in the parent of the
./items/ directory.  If I remove that line then there is no segmentation
fault.






-- 

*  Rob Brown-Bayliss
*  =================	
*      zoism.org






From andrew-pythonlist at puzzling.org  Sat Apr 26 22:39:25 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 12:39:25 +1000
Subject: Class problem: loosing my mind
In-Reply-To: <v0fmavkm9iknef7hit5n17d4m0qi6tpnia@4ax.com>
References: <v0fmavkm9iknef7hit5n17d4m0qi6tpnia@4ax.com>
Message-ID: <20030427023925.GG26530@frobozz.local>

On Sun, Apr 27, 2003 at 02:21:32AM +0000, <stuffduff> wrote:
> I'm 'doing the class pickle thing.'
> 
> I understand that both the module doing the pickling and the module
> doing the unpickling need to reference the same class definition.
> 
> So I go and make a directory under site-packages named gV and in the
> directory i add a file gV.py class gO and a completely empty
> __init__.py
> 
[..]
> 
> At the python prompt I enter import gV, no error.
> 
> Next I try and create an instance:
> 
> x=gV.gO()
> 
> and I get AttributeError: 'gV' module has no attribute 'gO'
> 
> What am I missing here?

Try printing gV.  You'll see that it's the package, i.e. the directory with
the __init__, not the module in it.

You want to do "from gV import gV", then gV.gO() should work.

-Andrew.





From nanthan79 at lycos.com  Mon Apr  7 08:28:40 2003
From: nanthan79 at lycos.com (Pathmenanthan Ramakrishna)
Date: Mon, 07 Apr 2003 12:28:40  0000
Subject: MAKE ERROR Scintilla!!Need Help
Message-ID: <HCPNNINNKLOCJCAA@mailcity.com>

hi,
i couldnt install Scite.
when i build Scintilla using make command

i get this error:
PlatGTK.cxx:11:18:glib.h:No such file or directory
PlatGTK.cxx:12:21:gdk/gdk.h:No such file or directory
PlatGTK.cxx:14:28:gtk/gtk.h:No such file or directory
PlatGTK.cxx:11:18:glib.h:No such file or directory
PlatGTK.cxx:14:28:gdk/gdkkeysyms.h:No such file or directory
make:[PlatGTK.o error]

i've installed the gtk and the dependencies ,but what is cause of this? i need help as i cant continue the installation of Scite.

hope anyone could help

thanks

nanthan.r






____________________________________________________________
Get 25MB, POP3, Spam Filtering with LYCOS MAIL PLUS for $19.95/year.
http://login.mail.lycos.com/brandPage.shtml?pageId=plus&ref=lmtplus




From aleax at aleax.it  Mon Apr 14 17:57:08 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 21:57:08 GMT
Subject: encapsulation checking features
References: <3e9b1298@cs.colorado.edu>
Message-ID: <U8Gma.14362$LB6.388431@news1.tin.it>

<posted & mailed>

Eric Scharff wrote:

> I love python's semantics for rapid prototyping, but I was wondering if
> there was a way to turn on warnings so that for other kinds of code, I
> can avoid certain common uglinesses. :)

PyChecker is the popular "lint-like" tool for Python.  However, I
very much doubt it can possibly warn about the specific usages that
you worry about, as they're quite common and idiomatic in Python.

Perhaps you can build some kludges to help, depending on exactly
what you're after.

> Specifically, is there a "lint-like" tool to:
> 
>  * warn if instance variables are created outside of a constructor?
> 
>   class K:
>     def __init__(self):
>       self.a = 6  #okay
>     def example(self):
>       self.b = 66 # not as good
> 
>   myVar = K()
>   myVar.c = 666  # really bad

To catch assignments you need a __setattr__.  In turn, it has to
know when __init__ has finished executing, so it can then warn
for assignments it wouldn't warn about otherwise.  To avoid having
to ensure your __init__ does explicit marking of its termination,
you're probably better off designing a custom metaclass that
overrides type.__call__ and does the "marking" that "initialization
is now finished" (and it may as well insert the __setattr__ too).


>  * warn if instance variables are used outside of instance methods
>    (encouraging the use of getters and setters)

That's even harder, because knowing if you ARE "inside an instance
method" or outside it is really tough.

But I really don't see the problem here.  If you want myVar.x to
call an accessor, you make x into a property of myVar's class,
period.  Absolutely no need to design trivial do-nothing accessors
either -- you can perfectly well have all client code access all
instance variables you want, and if and then such "variables"
need to be "computed on the fly" or have special effects when
set, then you make them into properties, hey presto.

> These are great features to have, but it'd be nice to have warnings to
> discourage their use at times.

I can see that for the "ensure all instance variables exist at
the end of __init__".  I can't see that, given the existence of
properties, for "no access to instance attributes from outside
of methods".  Perhaps you can clarify what advantage the latter
would provide, that properties just can't...?


Alex





From kp at kyborg.dk  Mon Apr  7 08:57:09 2003
From: kp at kyborg.dk (Kim Petersen)
Date: Mon, 07 Apr 2003 14:57:09 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com>
Message-ID: <3E9175A5.1060607@kyborg.dk>

Lothar Scholz wrote:
>>"comparable in power."  The reason is that there is a ton of elisp
>>code out there, including most of Emacs itself, that could be
>>translated to work with PyAlaMode.  My goal isn't to be able to
>>convert in a completely automated fashion, but close enough to be
>>attractive to both an elisp programmer and a Python programmer.
> 
> 
> Sorry, have you ever written a elisp program, something that is larger
> then
> 100 lines. I can't think about a way to provide this in python without
> writting a lisp like interpreter. Remember that the scope of elisp
> variables for example are not lexical. This is something very hard to
> emulate in python.

Why would the variable scheme that elisp uses be tougher to emulate in 
python than a lexical version ? From my pov it should make it a heck of 
a lot easier (think of emacsen as having 1 global dictionary and 1 
buffer specific dictionary for global/buffer local variables) then every 
time you call a function in elisp - do a shallow copy of local variable 
space (dictionary) and send it as the local vars - you can safely throw 
away the dict after function returns).

eg. something like this:

    def do_elisp_fun(local_name_space,fun,funargs,args):
       local_name_space=local_name_space.copy()
       # apply the args to the funargs (simple ignore :optional etc.)
       map(lambda n,v,d=local_name_space:
              local_name_space[n]=v,zip(funargs,args))
       retval=lisp_eval(funbody(fun),local_name_space)
       return retval

Excuses made for misremembering how lisp actually does optional args and 
rest ;-)

> I would really suggest to look at the structure of keyhandling,
> keymaps and buffers. And then forget about lisp and design a good
> python editor.





From oren-py-l at hishome.net  Wed Apr  9 04:50:20 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Wed, 9 Apr 2003 04:50:20 -0400
Subject: How to split string in equally long parts
In-Reply-To: <ad052e5c.0304082351.6cc33feb@posting.google.com>
References: <b6vd0e$i1b$01$1@news.t-online.com> <mailman.1049836334.25959.python-list@python.org> <ad052e5c.0304082351.6cc33feb@posting.google.com>
Message-ID: <20030409085020.GA8102@hishome.net>

On Wed, Apr 09, 2003 at 12:51:04AM -0700, Dan Bishop wrote:
> "Mike Rovner" <mike at bindkey.com> wrote in message news:<mailman.1049836334.25959.python-list at python.org>...
> > def split(S,N):
> >   """S - source string
> >      N - number of parts
> >   """
> >   lens=len(s)
> >   l=lens/N
> 
> Don't you mean lens//N ?

To get integer division that works consistently on all Python versions 
I would suggest int(lens/N). A bit longer but readable and unambigous.

    Oren





From gtalvola at nameconnector.com  Thu Apr 10 14:37:01 2003
From: gtalvola at nameconnector.com (Geoffrey Talvola)
Date: Thu, 10 Apr 2003 14:37:01 -0400
Subject: ssl error "PRNG not seeded"
Message-ID: <61957B071FF421419E567A28A45C7FE59AF3AF@mailbox.nameconnector.com>

Jeremy Hylton [mailto:jeremy at zope.com] wrote:
> On Thu, 2003-04-10 at 11:13, Geoffrey Talvola wrote:
> > sslerror: (1, 'error:24064064:random number
generator:SSLEAY_RAND_BYTES:PRNG
> > not seeded')
> 
> > Is this likely to be a problem in the Python SSL implementation or an
> > OpenSSL problem?
> 
> The error quoted about is generated in OpenSSL.  The Python SSL code
> just wraps the error message in a Python exception and propagates it.
> 

Right, I've figured out that much.  What I'm stumped on is why this would
only fail sporadically.  I'd expect it to either always complain about the
unseeded PRNG, or never.

I'm looking in the Python SSL code to see if it explicitly seeds the PRNG,
and I can't find any code that does.  So I guess OpenSSL automatically seeds
itself somehow, but some condition caused that to fail.  Anyhow, it does
clearly look like the problem is in OpenSSL, as you suggested.

- Geoff




From andrew-pythonlist at puzzling.org  Sat Apr 26 06:38:19 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sat, 26 Apr 2003 20:38:19 +1000
Subject: MD5 to long?
In-Reply-To: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de>
References: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de>
Message-ID: <20030426103819.GD26530@frobozz.local>

On Sat, Apr 26, 2003 at 12:20:34PM +0200, Axel Grune wrote:
> Hi, is there any function which can convert a MD5 or SHA1 sum
> (e.g. fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8) to a long integer?

Yep:

    >>> long('fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8', 16)
    1446090383108487004598230180906600288613882766568L

-Andrew.





From carsten at gehling.dk  Tue Apr 29 07:43:32 2003
From: carsten at gehling.dk (Carsten Gehling)
Date: Tue, 29 Apr 2003 13:43:32 +0200
Subject: SV: Function to convert fetchone() result from a list to a dictionary
In-Reply-To: <7jrra.18677$3M4.547894@news1.tin.it>
Message-ID: <BHEFKIOIMEDGILGDACKDIEBEDFAA.carsten@gehling.dk>

> -----Oprindelig meddelelse-----
> Fra: python-list-admin at python.org
> [mailto:python-list-admin at python.org]Pa vegne af Alex Martelli
> Sendt: 29. april 2003 11:09

> Yes.  I think what you REALLY want (even though you have no way to
> know you want it:-) is Greg Stein's wonderful dtuple.py, see:
>
> http://www.lyra.org/greg/python/dtuple.py
>
> The best explanation and set of examples I know for this is in Steve
> Holden's "Python Web Programming" (New Riders) -- I highly recommend

I already have this book, but unfortunately (for me) the chapters about
database manipulation is about the only section of the book, that I haven't
shuffled through. :-)

Looking at dtuple.py I think it may be shooting a bit over target. I could
be wrong - after all, I'm very new to Python :)

> if you can't purchase it I think you can at least get its examples
> free off the web (start looking at Steve's site, www.holdenweb.com).

Will do this!

> Pretty good (except that you're using tabs, which don't show up
> when I reply with KDE 3's KNode:-)  You can do a bit better (for

Heh - I like using tabs. It makes it easier for me to control indentation.

> So here's a step-by-step version (using SPACES, not TABS:-)...:
>
>
> def cfd(cursor):
>     values = cursor,fetchone()
>     if values is None:
>         return None
>     keys = [d[0] for d in cursor.description]
>     result_dict = dict(zip(keys, values))
>     return result_dict

Hmm... (looking up zip() in the documentation...). Okay I understand
everyting except this contruct:

keys = [d[0] for d in cursor.description]

Let me guess: it loops through cursor.description (obvious), for each of the
items (which are tuples) it "outputs" element 0. The output is then
automatically formatted in a way, that conforms to the list creation syntax.

Or something like that. Kind of neat - haven't seen that in any other
language, that I've used.

>     return dict(zip([d[0] for d in cursor.description],values))
>
> if you're into one-liners!-)

I'm not - I'd like to be able to read my own code 6 months from now... ;-)

- Carsten

BTW: Wish me welcome - I've just subscribed. And the way I continue to be
marveled by Python, I'm not likely to unsibscribe again. :-)





From aleax at aleax.it  Thu Apr 24 03:02:47 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 07:02:47 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <3ea434fd$0$180$a1866201@newsreader.visi.com> <25a672ee.0304211504.4f3cd067@posting.google.com> <mailman.1050968618.16005.python-list@python.org> <va956b925cj1df@corp.supernews.com> <25a672ee.0304231444.1ce0dd1b@posting.google.com>
Message-ID: <r_Lpa.89137$T34.2223764@news2.tin.it>

Denis Agar wrote:
   ...
> Ok, try to ignore the thickness of my skull, but you're saying that it
> is possible to create progams using mfc42.dll, a GUI that most people
> have already? I have the win32 extensions and am in the process of

Yes, with Python and a simple add-on you can choose to reuse any DLL
you may have lying around.  win32all (PythonWin in particular) does
that for MFC specifically.  Other add-ons such as the oldie (but
goldie) calldll, and the newer (and more powerful) ctypes, let you
do that for ANY dll (with more Python-side work to adapt it).  Good
luck -- it IS, of course, lower-level programming than typical Python
work, since you want to use DLLs designed for lower-level languages.

> trying to understand them. (PS: Anyone got some (s)ample code they
> could throw my way?)
> But, as I understand it, this would be hard.

Of course it would.

> This isn't a request nor a demand nor a manifesto of any kind, but ---
> would it be possible to make a simple PythonCardish GUI shaping-type
> languagey thing (I'm not yet up on the lingo) that interfaces with the
> win32 extensions and dumbs it down to about my level? I'm just
> wondering about it is all. . .

The anygui project, anygui.sf.net, aims to design a uniform wrapper
that can use most any GUI toolkit as a back-end -- this would include
(at least potentially) MFC, or the bare Win32 API's, etc.  It's not
moving very fast, but there's nothing specific holding it back save
for big ambitions and scarce time being spent on it;-).  Once anygui
is stable, and there exists a working back-end for your chosen GUI
toolkit[s], it would surely be possible to write a screen-painter
front-end on top of it -- there is no conceptual block against that.

The pragmatical likelihood of this actually happening, like for any
other open-source software project, depends on how likely it is that
some active developers find that investing their time and energy into
it "scratches their itch".  anygui's slow progress so far seems to
suggest that, while intellectually interesting and stimulating, it
may not speak to an actual NEED of its developers -- so, it becomes a
"back-burner", "in our copious spare time" project, and those do not
tend to move forwards all that rapidly.


Alex




From kirk at strauser.com  Sun Apr  6 12:47:48 2003
From: kirk at strauser.com (Kirk Strauser)
Date: Sun, 06 Apr 2003 16:47:48 GMT
Subject: Strauser family release #3
Message-ID: <rs16mxfer.ln2@news.conpoint.com>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

########################################

#!/usr/bin/python

# This program forks().  That should be plenty for a few years' entertainment
# Copyright (C) 2003 Kirk & Jennifer Strauser

# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.

# However, the output of this particular instance shall remain
# exclusively licensed to the authors for a period of up to eighteen
# years.

# $Id: baby.py,v 1.3 2003/04/06 06:12:04 kirk Exp $

import time
import os

eta = 1049547660
name = "Jennifer Strauser"

if eta > time.time():
    # Not done yet
    time.sleep(eta - time.time())

pid = os.fork()
if pid:
    print "%s: At %s, I got my own address space back!" % (name, time.ctime(eta))
else:
    length = 20.5   # inches
    weight = 5.4    # pounds.ounces
    name = "Ariana Nicole Strauser"
    print "%s: Hello, world!" % name

########################################

- -- 
Kirk Strauser

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+kFhu5sRg+Y0CpvERAu+1AJ4/yzpXOtYtyLoZb/so+lraylLHPwCfd5uQ
4HRtwzPqMmm7gcIfsv+9hVo=
=lIgp
-----END PGP SIGNATURE-----



From hupp at cs.wisc.edu  Tue Apr  8 14:14:22 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Tue, 8 Apr 2003 13:14:22 -0500
Subject: STRING TO LIST
In-Reply-To: <316753154.1049824568087.JavaMail.nobody@webmail1>
References: <316753154.1049824568087.JavaMail.nobody@webmail1>
Message-ID: <20030408181422.GA29642@upl.cs.wisc.edu>

On Tue, Apr 08, 2003 at 02:56:08PM -0300, jubafre at brturbo.com wrote:
> Hello, 
> 
> I need to change a variable from string to list. 
> > 
> > >>> x=str("[[12,13],[14,15]]") 
> > >>> x 
> > '[[12,13],[14,15]]' 

>>> b = eval("[[12,13],[14,15]]")
>>> type(b)
<type 'list'>
>>> b
[[12, 13], [14, 15]]

 
-Adam




From mwh at python.net  Thu Apr  3 14:34:34 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 3 Apr 2003 19:34:34 GMT
Subject: newbie os.stat question
References: <AV_ia.7382$4P1.570905@newsread2.prod.itd.earthlink.net> <slrnb8p1ut.174.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <7h3vfxvjrxy.fsf@pc150.maths.bris.ac.uk>

Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> writes:

> In article <AV_ia.7382$4P1.570905 at newsread2.prod.itd.earthlink.net>,
> Tipton Bandy wrote:
> > 
> > I've got my first python utility working pretty good, but
> > I just need help on one point.
> > 
> > I need to test a filename and return true or false based on 
> > whether the file is readable and writable by the user running
> > my utility. Oh, yes, this is on Linux:)

[...]

> os.access

is almost never what you want, because it uses the real user id, not
the effective user id.

> (But this is fairly useless, because the situation could change by
> the time you actually try to use the file.  You should just attempt
> to use the file and catch OSError exceptions:
> 
>     try:
>         f = open(filename)
>     except OSError, e:
>         if e.errno == errno.EPERM:
>            print 'Operation is not permitted'
> 
> )

This is still true, though.

Cheers,
M.

-- 
  The above comment may be extremely inflamatory. For your
  protection, it has been rot13'd twice.
                           -- the signature of "JWhitlock" on slashdot



From apolloner at bap.at  Wed Apr  9 01:05:49 2003
From: apolloner at bap.at (apolloner)
Date: Wed, 9 Apr 2003 07:05:49 +0200
Subject: embedding on aix fails
Message-ID: <6mf07b.s8q.ln@news.bap.at>

AIX 4.3.3
VisualAge C++ Professional / C for AIX Compiler, Version 6
Python 2.2.2 

I have an executable named t and it works fine .
then I do a link 
$ ln t x ( or: $ ln t z )
and try to run this x ... and I get a Segmentation fault(coredump) !
$ ln t y
and run y and it works.

The program is using the python library and if the segfault happens,
it is because the adress of the Py_InitModule4 function is 
wrong in initstrop after loading the library strop.so dynamically.
Executing t or y this adress is ok ! ( verified with testprints ) 

Python itself works great , loading all these libraries without any problem !

the program :

#include "Python.h"
int main(int argc, char ** argv)
  {
  Py_Initialize();
  PyRun_SimpleString("from string import *\n");
  }

cc_r -Wl,-bE:Modules/python.exp -I. -I./Include -o t t.c \
    libpython2.2.a -lld  -ldl -lm

ptyhon is compiled in the following way ( output of make ) 

cc_r  -Wl,-bE:Modules/python.exp -lld -o python  Modules/python.o  \
    libpython2.2.a -ldl    -lm

what could be wrong with my linkage ??

regards
udo



From staschuk at telusplanet.net  Wed Apr 16 22:19:57 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 20:19:57 -0600
Subject: compatibility problem
In-Reply-To: <1f4c702e.0304161116.5b0f6565@posting.google.com>; from chad.scherrer@pnl.gov on Wed, Apr 16, 2003 at 12:16:55PM -0700
References: <1f4c702e.0304161116.5b0f6565@posting.google.com>
Message-ID: <20030416201957.A6846@tibia.amotlpaa.bogus>

Quoth Chad Scherrer:
  [...]
> Python 2.1 gives the correct solution, but somehow 2.2.2 is only
> giving me a subset. Not only that, but the order is different! The
> ordering is arbitrary anyway, but I would still expect it to be
> consistent between the two.

This doesn't answer your question, but:  If your code has a decent
suite of unit tests, surely they'll pinpoint the problem for you
in a hurry.  If you have no unit tests, maybe this is an
opportunity to write some.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From mac at die.spammer.die.dgp.toronto.edu  Fri Apr  4 20:37:14 2003
From: mac at die.spammer.die.dgp.toronto.edu (Maciej Kalisiak)
Date: Sat, 05 Apr 2003 01:37:14 GMT
Subject: pickling new-style extension types: examples?
Message-ID: <slrnb8scn8.d6f.mac@mac.dgp.toronto.edu>

I've implemented a new extension type in C, and am now attempting to make it
picklable, with little success.  Can someone point out any online examples of
doing this, or even post some code snippets?  Ideally I would want to implement
the pickling functionality in C, but will settle for a Python class wrapper
around the extension type, if necessary.  Either the __reduce__ protocol, or
using the "copy_reg" module will do.



From carroll at tjc.com  Sat Apr  5 13:57:07 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Sat, 05 Apr 2003 10:57:07 -0800
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain>
Message-ID: <6j8u8v0hjh3n9dm1jjflu3a970ada6qho4@4ax.com>

On Sat, 05 Apr 2003 18:11:50 GMT, Robin Munn <rmunn at pobox.com> wrote:

>First, from what you've written above, it sounds like you're doing
>type-checking to determine whether the parameter is a string or a file.

Actually, I'm just accepting the additional parameter, representing a
string (interpreted as a filename) or an open file object, and passing
that along to an internal method longcmd(), that accepts either, which
in turn passes it along to the internal getlongresp().  getlongresp()
is the part that acts differently depending on the presence of the
parameter.

getlongresp() does do type checking.  In the version I have, it's

  if isinstance(file, types.StringType):

and the latest CVS is:

  if isinstance(file, str):

But that's exiting code, nowhere near anyplace I'm touching.  My
inclination is to leave it alone.

Up until now, the aspect of writing to a file instead of the returned
list is only used for the body() method; I'm just extending it to
xover() and probably a few others.


>You may not be doing type-checking, in which case this advice doesn't
>apply to you, but type-checking is the wrong way to go about that. Do
>interface checking instead, e.g.:
>
>    # Is it a file-like object?
>    try:
>        output_file.write
>        print "output_file is a file-like object"

Just curious; why is this preferred?  Serious question; I'm relatively
new to Python, and still learning a lot.

I'm thinking it might be two different philosophies.  The one in the
present nntplib will try to invoke write() in whatever object was
passed to it, and if the caller invoked it with the wrong object type,
he'll get an error.  The one above seems to change the spec to support
any object that has a method named "write."

>Second, do *not* use the name "file" as your parameter name; you will
>shadow the file builtin.

That's already in nntplib.  I'm just making xover() consistent with
body().

>Third, I'm not sure I like the functionality of passing a string
>filename. It sounds convenient, but all it really does is *add*
>ambiguity to your function which could be avoided.

I'd probably do it that way if I were doing it from scratch, but I
don't want to reinvent the wheel, and I think maintaining consistency
with the body() call is a Good Thing.

>Other than that, nice job.

Thanks, and thanks for the feedback.

By the way, I should really point out that my changes are relatively
trivial.  It's Travers Naran, who modified nntplib to support a file
parameter on body(), and did it in a particularly flexible way to make
it easy to extend to the other user-called methods, that deserves the
bulk of the credit.




From cswiger at mac.com  Wed Apr 23 10:40:40 2003
From: cswiger at mac.com (Chuck Swiger)
Date: Wed, 23 Apr 2003 10:40:40 -0400
Subject: dbm concurrency
In-Reply-To: <6c492956.0304230623.74b92d7f@posting.google.com>
References: <6c492956.0304230623.74b92d7f@posting.google.com>
Message-ID: <3EA6A5E8.1040903@mac.com>

Clemens Hermann wrote:
> I want to use a standard unix dbm database to store user->regexp
> combinations for squid ACLs. The dbm is written from a python cgi and
> it is assured that only one write-access at a time occurs.
> On the other side the same dbm is accessed read-only from a
> C-programm. There may be many concurrent read-only accessess and also
> requests for read-access while python writes to the dbm.
> 
> Might this be a problem when the C-program requests read-only access
> while python is writing to the dbm or vice versa? If so, how can this
> be solved?

If it's a problem that the C program may cache old user/regexp 
combinations after they've been updated (and it does do), yes, you'd 
have an obvious concurrency/race-condition issue.

What you should first consider is making changes to the database atomic 
to the C program.  By this I mean, have python update a copy of the DBM 
database, not the actual one that the C code is looking at.  Once python 
has written all of the changes it wants to make, use "mv" [or 
unlink()/rename()] to swap the new database in.

The next thing is to either restart the C programs to make them re-read 
the new DB (via 'squid -k reconfigure'?), or else have them fstat the 
database file and refetch any cached results if the DB file's 
last-modified timestamp changes.  You could also look into something 
like the BerkeleyDB 4 from sleepycat.com, and what sendmail does for the 
hash or btree map types like /etc/mail/aliases or mailertable or access.

-Chuck






From mhammond at skippinet.com.au  Sun Apr  6 18:53:07 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Sun, 06 Apr 2003 22:53:07 GMT
Subject: PyWin: Redraw during long loops?
In-Reply-To: <3E8EA676.90008@web.de>
References: <3E8EA676.90008@web.de>
Message-ID: <nd2ka.2425$Tb6.22760@news-server.bigpond.net.au>

Ingo Linkweiler wrote:
> Hello,
> 
> I am using PythonWin.
> When my programs are in a long loop, the redraw of Pywin is blocked.
> PyWin only redraws its window and handles events, when a print (with 
> newline) or input is done.
> 
> Is there a way to avoid this?
> 
> Ingo

Calling win32ui.PumpWaitingMessages() generally avoids this.

Mark.




From qvx3000 at yahoo.com  Fri Apr 18 04:59:14 2003
From: qvx3000 at yahoo.com (Qvx)
Date: 18 Apr 2003 01:59:14 -0700
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <8b2d5523.0304180059.722de25a@posting.google.com>

blunck at gst.com (Christopher Blunck) wrote in message news:<1147e466.0304160630.5b510724 at posting.google.com>...
> So how many of you guys use XP processes?

I see that most of you, when talking about XP, are referring to TDD,
and on some occasions about desing/coding issues. So let me speak more
about TDD.

Testing as a part of development process seems so natural in Python.
When developing in Python I strive to write modules/classes which
satisfy the low coupling "criteria" so that I can remove them easily
(to some extent) from the project. This enables me to write the
required tests. All of this can be achieved because the Python is
dynamically typed language and it doesn't stand in my way. It doesn't
need the rest of the system in order to test one of modules.

Must I say that all of this is happening in my home projects and some
minor utility projects at work and I would like to use more of TDD at
work. But there is a catch. As hard as it is to admit this I must
admit it: I don't know how to to use it. All work is done in Oracle
database and Forms and they don't fit nicely with TDD (at least to
me). How do you test a form? How do you test a lengthy database job
whose input depends on dozens of tables and which modifies/creates
records across dozens of tables?

I might be in a position to influence some things here at my company
so any help would be appreciated. Any thoughts on this? Books? Links?

--
Sokol



From shagshag13 at yahoo.fr  Tue Apr  1 03:22:25 2003
From: shagshag13 at yahoo.fr (shagshag13)
Date: Tue, 1 Apr 2003 10:22:25 +0200
Subject: Mimic a web surfer ?
Message-ID: <3e894b7d$0$29158$626a54ce@news.free.fr>

hello,

i would like to develop a script which will mimic a web surfer, that is :

- follow links as a human and thus always come from previous page (i think
there is something to do with http 'referer' and cookies)
- be able to log if necessary

is there anything already existing ? i'm looking for any tips which will
help me to start...

note : this could be used to have "automatic" access to free webmails not
providing pop protocol.

thanks in advance.





From nirina at mail.blueline.mg  Tue Apr  8 16:54:10 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Wed, 9 Apr 2003 00:54:10 +0400
Subject: selecting a class in execution time
References: <AE59EE7E-691B-11D7-8CA9-000393B10A78@mac.com>
Message-ID: <019401c2fe2c$c77bf140$f7a2383e@raseliar>

Maybe you want something like this:

>>> class TheClass:
             def Name(self):
                  print 'a name from a class'

 
>>> TheClass
<class __main__.TheClass at 0x00AD7270>
>>> object = TheClass()
>>> object
<__main__.TheClass instance at 0x00AE35F8>
>>> object.Name
<bound method TheClass.Name of <__main__.TheClass instance at 0x00AE35F8>>
>>> object.Name()
a name from a class


If this is saved as 'MyPackage.py', you can select TheClass with:


>>> from MyPackage import TheClass


or create the object like


>>> import MyPackage
>>> object = MyPackage.TheClass()
>>> object
<MyPackage.TheClass instance at 0x00AF7378>
>>> object.Name
<bound method TheClass.Name of <MyPackage.TheClass instance at 0x00AF7378>>
>>> object.Name()
a name from a class


HTH
nirinA


----- Original Message ----- 
From: <cybersamurai at mac.com>
To: <python-list at python.org>
Sent: Monday, April 07, 2003 9:09 PM
Subject: selecting a class in execution time


> have some way to do this on Python?
> 
> Object x = Class.forName( "mypackage.TheClass" ).newInstace();
> (this is a java code)






From mwh at python.net  Wed Apr 30 05:25:21 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 30 Apr 2003 09:25:21 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304251030.30fd1f94@posting.google.com> <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk> <mailman.1051635832.26364.python-list@python.org> <7h3el3kgy4j.fsf@pc150.maths.bris.ac.uk> <k7dc1gxv.fsf@python.net>
Message-ID: <7h3ade8gwww.fsf@pc150.maths.bris.ac.uk>

Thomas Heller <theller at python.net> writes:

> See also http://www.python.org/sf/660098

Bollocks, I submitted a new bug.  I'll close it again.

Cheers,
M.

-- 
  About the use of language: it is impossible to sharpen a
  pencil with a blunt axe.  It is equally vain to try to do
  it with ten blunt axes instead.
      -- E.W.Dijkstra, 18th June 1975. Perl did not exist at the time.



From cybersamurai at mac.com  Mon Apr  7 15:56:56 2003
From: cybersamurai at mac.com (cybersamurai at mac.com)
Date: Mon, 7 Apr 2003 16:56:56 -0300
Subject: selecting a class in execution time
In-Reply-To: <003401c2fd2a$75970340$21795418@dell1700>
Message-ID: <15FAE31B-6933-11D7-8CA9-000393B10A78@mac.com>

Thanks about your help

I am developing a way to describe a application in XML language. De 
idea is define only the data and your type, using a engine in execution 
time to create a functionality of some application.

take a look in this diagram to get the idea.
in this idea is only needed create a green file to make a application
http://homepage.mac.com/cybersamurai/public/model2.jpg
http://homepage.mac.com/cybersamurai/public/datamodel.jpg
http://homepage.mac.com/cybersamurai/public/data_definition.html

On segunda-feira, 7 abr, 2003, at 14:23 America/Sao_Paulo, Brian 
Quinlan wrote:

> I'd be nice to know the why of this to suggest the best idiom.
>
>> Object x = Class.forName( "mypackage.TheClass" ).newInstace();
>
> x = getattr(__import__("mypackage"), "TheClass")()
>
> The last set of parens is the argument list for TheClass' constructor.
>
> If you know the name of the package at compile time then:
>
> import mypackage
> x = getattr(mypackage, "TheClass")()
>
> Cheers,
> Brian
>





From andrew at acooke.org  Tue Apr 29 19:24:53 2003
From: andrew at acooke.org (andrew cooke)
Date: Tue, 29 Apr 2003 19:24:53 -0400 (CLT)
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 29)
Message-ID: <4320.10.2.0.2.1051658693.squirrel@127.0.0.1>


Jack Diederich said:
>
> 2) What happened on the 27th to cause a blind spot in google?
>    Best conspiracy theory wins a cookie.

http://lcamtuf.coredump.cx/evilfinder/ef.cgi?said=python+27th+google

andrew

-- 
http://www.acooke.org/andrew



-- 
http://www.acooke.org/andrew




From max at alcyone.com  Tue Apr 22 17:14:44 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 14:14:44 -0700
Subject: unused args in GUI callbacks: pychecker
References: <20030422145218.6e9e6549.gry@ll.mit.edu>
Message-ID: <3EA5B0C4.7ACB349@alcyone.com>

george young wrote:

> So, is there a way to define a python(member) function and tell
> python to ignore particular arguments? [so that pychecker could
> ignor them too...]

A quick check of the pychecker help (--help) indicates that the
-E/--unusednames option is what you want.  By default it will already
ignore arguments named _, empty, unused, and dummy.

> I recall that C++ had a way to tell the compiler this.

C++ is just the language, so it was just a particular compiler that you
are referring to with that feature.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From m.faassen at vet.uu.nl  Tue Apr  1 17:52:36 2003
From: m.faassen at vet.uu.nl (Martijn Faassen)
Date: 1 Apr 2003 22:52:36 GMT
Subject: I need a lot of help!
References: <31ef0f43.0304011336.1d2777ac@posting.google.com>
Message-ID: <b6d57k$lvo$1@newshost.accu.uu.nl>

Jewlz408 <Jewlz408 at aol.com> wrote:
> I need a lot of help
> What do you know about Arrays

I know about 50% of what there is to know about Arrays. They're usually
green but occasionally blue. They generally have big eyes.

> How many storage Locations will be allocated for each of the given arrays?
> 1) Dim A(1 to 10)

17.5

> 2) Dim B(-3 to 5)

The number pi.

> 3) Dim C$( 12)

Trick question! This will be one Perl.
 
> 4) Dim D( 1 to 4, -2 to 2)

Frequently.

> 5) Dim E(4,6)

I don't know the answer to this one. I suspect it has to do with advanced
quantum numerology.

> And What does the following programs do with its output?

> Dim Table( 1 to 5, 1 to 3)
> For I = 1 to 5
> ????? For J = 1 to 3
> ????? Let Table(I,J) = 1
> ????? Next J
> Next I
> End
> 
> What does the program do?

It is a dishwasher.
 
> And what is the output?

Dirty water, and clean dishes. You can also put in cookies however,
in which case you will get chocolate chips.

> Thankyou so much for your time 
> Please help me!

You're welcome!

Regards,

Martijn
-- 
History of the 20th Century: WW1, WW2, WW3?
No, WWW -- Could we be going in the right direction?



From roy at panix.com  Fri Apr 11 23:28:23 2003
From: roy at panix.com (Roy Smith)
Date: Fri, 11 Apr 2003 23:28:23 -0400
Subject: Singletons
Message-ID: <roy-20E369.23282211042003@reader1.panix.com>

Hunting around for ideas on implementing the singleton pattern in 
Python, I found

http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/52558

It seems like a pretty reasonable implementation, but I don't understand 
one thing about it.  What is the point of:

# Store instance reference as the only member in the handle 
self.__dict__[ '_Singleton__instance' ] = Singleton.__instance

I don't see that it does anything useful.  Is there something subtle 
that I'm missing?



From mnations at airmail.net  Thu Apr  3 19:02:54 2003
From: mnations at airmail.net (Marc)
Date: 3 Apr 2003 16:02:54 -0800
Subject: <KeyPress>, focus, and Pmw widgtes
Message-ID: <4378fa6f.0304031602.33c3bcee@posting.google.com>

Hi,

I was trying to use a keypress to trigger an event in a Gui. Reading
through past postings I discovered that the widget has to have focus
to actually register the event. However, even setting the focus on the
widget component I am unable to trigger the event. Obviously doing
something wrong, I look to the masses for help. Here is my code:

    def widgets(self):
	self.fps = Pmw.EntryField(self.master, labelpos=W, label_text='FPS:',
                                   entry_width=10, validate=None,
value = '0')
	self.fps.pack(side=LEFT, padx = 5)
        self.fps.bind( '<Leave>', self.getTBPS)
        self.fps.bind( '<KeyPress-Up>', self.getTBPS)
    
	entry = self.fps.component('entry')
        entry.focus()


    def getTBPS(self, event):
        print "INTO TBPS"
        fps = int( self.fps.getvalue() )
        length = int( self.size.getvalue() )
        
        #fps = rate * pow(10,6) / ( ( ( length + 4) * 8) + 64)
        tbps = fps /  ( pow(10,6) / ( ( ( length + 4) * 8) + 64) )

        self.tbps.clear()
        self.tbps.setvalue(tbps)

So when I start the application, I assume that the entry widget has
focus. But when I enter data, the keypresses still aren't registered.
But I know that the callback works because when I leave the space, my
<Leave> trigger is registered. What am I missing?

Thanks,
Marc



From tim_one at email.msn.com  Sun Apr 20 12:28:19 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sun, 20 Apr 2003 12:28:19 -0400
Subject: Sorting without transitivity
In-Reply-To: <3EA24A7A.10900@rogers.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCMEIEEHAB.tim_one@email.msn.com>

[Frank Niessink]
> [It's been while since I've dealt with transitivity and stuff, so
> I hope I got the terminology right]
>
> I have objects that are not transitivily ordered, e.g.:
>
> A > B
> A == C
> B == C
>
> Now I have a list [A, B, C] and I want to sort it.

What does sorting mean to you?  What it means to most people <wink> is that
a list x is sorted if and only if

    x[i] <= x[i+1]

for each i in 0, 1, ..., len(x)-2.  But the list [A, C, B] is sorted by that
definition, given your odd meanings for "==" and "<".  You don't accept [A,
C, B] as being sorted, though, so you need to explain what sorting means to
you.  Perhaps you consider list x to be sorted if and only if

    x[i] <= x[j]

for all i in 0, 1, ..., len(x)-2 and j in i, i+1, ..., len(x)-1.  When you
make up meanings for < and ==, it requires proof first that such an ordering
always exists.

> The resulting list should have B before A, I don't care where C is.

This prompted people to suggest a topological sort, which may be what you
want (it's hard to say given just one example).  What they haven't suggested
is an efficient way to prepare input for a topsort algorithm -- those
usually work with an explicit sequence of (predecessor, successor) pairs,
but you don't have such a sequence.  You could create one with a quadratic
algorithm comparing every Point r to every other Point s, materializing a
pair (r, s) only when r < s.

It may be possible to do better than that:

"""
class Point:
    def __init__(self, x, y):
        self.x = x
        self.y = y

    def __repr__(self):
        return 'Point(%d, %d)'%(self.x, self.y)

    def __cmp__(self, other):
        if self.x == other.x and self.y == other.y:
            return 0
        elif self.x <= other.x and self.y <= other.y:
            return -1
        elif self.x >= other.x and self.y >= other.y:
            return 1
        else:
            return 0
"""

This __cmp__ makes the intent of code like

        self.failUnless(pointList.index(a) < pointList.index(b))

pretty mysterious; e.g.,

    [Point(20, 3)].index(Point(3, 20))

is 0:  Point.__cmp__ says Point(3, 20) is equal to Point(20, 3), and index()
uses __cmp__.

Glossing over that (I'll just suggest it's unlikely that .index() does what
you want either), note that in Python's lexicographic tuple comparison,

    (a, b) < (c, d)

iff

    a < c or (a == c and b < d)

(a, b) < (c, d) does not imply Point(a, b) < Point(c, d); for example,

    (3, 20) < (20, 3)
but
    Point(3, 20) == Point(20, 3)

However, I think the other direction works:  Point(a, b) < Point(c, d)
implies (a, b) < (c, d).  So whenever Point.__cmp__ thinks p1 < p2, native
tuple comparison would also think (p1.x, p1.y) < (p2.x, p2.y).  So if you
let Python sort naturally on (self.x, self.y) tuples, it should respect your
notion of "<" too (in addition to calling some pairs "<" that Point.__cmp__
doesn't call "<"):

def point_sort(s):  # s is a list of Points
    temp = [(p.x, p.y, p) for p in s]
    temp.sort()
    return [t[-1] for t in temp]

s = [Point(i, j) for i in range(5) for j in range(5)]
import random
random.shuffle(s)
print s
print point_sort(s)

displays (for example -- random.shuffle() may produce any permutation):

[Point(0, 1), Point(1, 0), Point(1, 3), Point(2, 0), Point(1, 4),
 Point(3, 2), Point(4, 0), Point(0, 2), Point(0, 4), Point(1, 1),
 Point(3, 3), Point(2, 3), Point(4, 1), Point(2, 4), Point(2, 1),
 Point(0, 3), Point(3, 4), Point(0, 0), Point(4, 3), Point(2, 2),
 Point(3, 1), Point(4, 4), Point(1, 2), Point(3, 0), Point(4, 2)]

[Point(0, 0), Point(0, 1), Point(0, 2), Point(0, 3), Point(0, 4),
 Point(1, 0), Point(1, 1), Point(1, 2), Point(1, 3), Point(1, 4),
 Point(2, 0), Point(2, 1), Point(2, 2), Point(2, 3), Point(2, 4),
 Point(3, 0), Point(3, 1), Point(3, 2), Point(3, 3), Point(3, 4),
 Point(4, 0), Point(4, 1), Point(4, 2), Point(4, 3), Point(4, 4)]

Using big words <wink>, a partial ordering can be extended to a total
ordering (usually in many different ways), and a sort wrt such a total
ordering necessarily respects the partial ordering it extends.  In this
case, native tuple comparison appears to be a suitable extension of your
partial ordering.





From martin at v.loewis.de  Wed Apr 23 15:16:16 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 23 Apr 2003 21:16:16 +0200
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <3EA5B173.9420324C@alcyone.com> <mailman.1051056037.8077.python-list@python.org> <3EA5DE1E.22704437@alcyone.com> <m3r87tdbo3.fsf@mira.informatik.hu-berlin.de> <mailman.1051081416.21801.python-list@python.org>
Message-ID: <m34r4pgh33.fsf@mira.informatik.hu-berlin.de>

Ian Bicking <ianb at colorstudy.com> writes:

> Note that the idiom I've been using (and which I've seen elsewhere) is:
> 
> True, False = 1==1, 0==1
> 
> Which retains proper boolean objects.

For Python 2.x, this does not really matter much. Yes, you do get
references to Py_True/Py_False, but they are still just integer
objects with the value 1/0.

Regards,
Martin




From tim_one at email.msn.com  Sat Apr 12 21:22:13 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sat, 12 Apr 2003 21:22:13 -0400
Subject: Speed of string += string
In-Reply-To: <roy-06B012.20145112042003@reader1.panix.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCCENGEGAB.tim_one@email.msn.com>

[Roy Smith]
> One of the very few things I like about C++ over Python is that the STL
> is very careful to document the complexity of the various operations.
> AFAICT, no such documentation exists for Python.

That's so.  It will be repaired as soon as someone volunteers to write the
docs <wink>.

> True, it's documented that strings are immutable, and from that you
> should be able to infer that string addition requires lots of creating
> and destroying of objects, which sounds like it should work out to
> being quadradic, but that's not quite the same as having the
> algorithmic complexity spelled out for you.

Not the same at all.  For example, string[i:j] in Python takes time
proportional to j-i, but that doesn't *follow* from immutability (there are
more complicated implementation schemes than Python uses that would allow
faster slicing -- in return for losing elsewhere).

> As another example, I can guess that Python dictionaries work like STL
> maps on the inside, but that would be just a guess.  Is inserting an
> element logN?

In all versions of CPython released to date, expected time O(1), worst O(N),
and no amoritized bound guaranteed better than O(N).  If you supply a
particularly poor __hash__ implementation, or have spectacularly unlucky key
distributions, it can be expected-case O(N).  Note that these don't count
the time for hashing the key.  If, e.g., the key is a string s, and s isn't
interned, and s hasn't been hashed before, then hash(s) takes O(len(s)) time
by itself; but if s has been hashed before, hash(s) takes constant time (its
hash code is cached in the string object).  So the full story can be quite
involved.

> Is keys() linear?

Yes, and that's both best and worst case.

> Is del on an element logN?  Is has_key() logN?

Both the same as insert.

> Those would be my guesses.  But they're just guesses.

Now you know.  The *language* doesn't guarantee any of this, though, and
probably never will.

Guido worked on an earlier language (ABC) that used implementation
techniques with good worst-case bounds, chiefly balanced binary trees.  But
the overhead of those schemes was a major drag on performance for most apps
most of the time.  Python's implementation is generally more pragmatic,
aiming for simpler schemes with excellent normal-case performance, not
worrying much about worst-case performance.

For example, until Python 2.3, the worst-case performance of list.sort() was
quadratic (although in 1.5.2 that was extremely difficult to provoke even
with a deep understanding of the 1.5.2 samplesort's implementation; it was
easy to provoke in pre-1.5.2 quicksort variants).

Python 2.3 has a new worst-case N log N sort, not because it has better
worst-case performance, but because the 2.3 sort proved much faster than the
2.2 sort in many common cases where the list already has significant
pre-existing order, and proved no slower than the 2.2 sort when the input
data is randomly ordered.  Worst-case analysis played no significant role in
deciding to switch to the new sort implementation, and I expect we would
have switched even if the new sort's worst-case performance were cubic (but
exceedingly rare).





From skip at pobox.com  Mon Apr 14 12:02:06 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 11:02:06 -0500
Subject: Displaying Unicode on the console (Windows)
In-Reply-To: <182bcf76.0304140744.1b3d43ee@posting.google.com>
References: <182bcf76.0304140744.1b3d43ee@posting.google.com>
Message-ID: <16026.56190.239395.339962@montanaro.dyndns.org>

    Paul> To use a concrete example, I'd like to print the Euro symbol. A
    Paul> Unicode string for this is u'\20a0'

    >>>> unicodedata.name(u'\u20a0')
    Paul> 'EURO-CURRENCY SIGN'

    ...
    >>>> print u'\u20a0'.encode("latin-15")
    ...
    >>>> print u'\u20a0'.encode("latin-1")
    ...
    >>>> print u'\u20a0'.encode("iso8859_15")
    ...
    >>>> print u'\u20a0'.encode("cp1258")
    ...
    >>>> print u'\u20a0'.encode("mbcs")
    ...
    Paul> [BTW, in case it's relevant - I know it probably is - the output
    Paul> from "chcp" at the console prompt is "Active code page: 1252"]

Seems that cp1252 is the most likely encoding you didn't try.  What happens
if you execute

    print u'\u20a0'.encode("cp1252")

?

Skip




From eric.brunel at pragmadev.com  Fri Apr 25 05:24:58 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Fri, 25 Apr 2003 11:24:58 +0200
Subject: Tkinter - tkFileDialog - need help
References: <oprn5mygrymu3s0q@news.cis.dfn.de> <b8aofk$iqf$1@news-reader14.wanadoo.fr> <oprn6gl1klmu3s0q@news.cis.dfn.de>
Message-ID: <b8auhi$n7l$1@news-reader14.wanadoo.fr>

Klaus Meyer wrote:
> Hello Eric,
> 
>> Tk uses the regular "open file" dialog from Windows , so the only 
>> known way (at least to me) is to configure Windows to display the 
>> hidden files.
> 
> 
> thanks, but this is not the problem, because i always have enabled this 
> option. And i can see the hidden directorys in win-explorer, but not in 
> tkFileDialog. So it seems that tkFileDialog has its own "show/don't show 
> hidden dirs" flag???
> But how to enable this?

Sorry: I didn't correctly read the OP. The dialog to select a directory is 
indeed provided with tcl/tk, as you can tell by its "old-fashioned" look. There 
is apparently no native directory selection dialog in Windows, or if there is, 
maybe we should tell the tcl/tk guys... ;-)

> If this is not possible, than i have a problem... :-(

Then apparently, you have one: according to the tcl/tk documentation for the 
underlying function (tk_chooseDirectory), such an option does not exist. See:

http://dev.scriptics.com/man/tcl8.4/TkCmd/chooseDirectory.htm

> Does Tkinter have a other function to open a directory select box?

Nope. I couldn't find any in the common Tkinter extensions, so I guess you'll 
have to make one yourself. This should be quite easy with the os module and a 
few list boxes.

HTH
-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From roy at panix.com  Sun Apr 13 16:58:49 2003
From: roy at panix.com (Roy Smith)
Date: Sun, 13 Apr 2003 16:58:49 -0400
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <%djma.11763$T34.381354@news2.tin.it>
Message-ID: <roy-031327.16584913042003@reader1.panix.com>

Alex Martelli <aleax at aleax.it> wrote:
> #define private public
> #define protected public
> at the top of your .cpp file, before any #include, in most C++
> implementations

I'm confused.  You started out to show that C++'s "protected" keyword is 
evil, but what you've really shown is that C++'s preprocessor is evil.  
Can they both be true? :-)



From max at alcyone.com  Tue Apr  8 21:46:40 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 08 Apr 2003 18:46:40 -0700
Subject: usage of assert and -O
References: <mailman.1049844497.9845.python-list@python.org>
Message-ID: <3E937B80.62262184@alcyone.com>

> Matthew Russell wrote:

> I understand that when you invoke the interpreter with the -O flag
> assert statments are not executed - is this 100% correct? (as i belive
> it to be)

Yes, it is correct.

> and if so, shouldn't the use of assert (if not assert itself) be
> depricated?

No, why should it be?  Traditionally, asserts are used in debug mode
where you're checking invariants, things that are never supposed to
happen, and so on.  You're including them for extra layers of immediate,
in-your-face testing -- if something goes wrong and you're running in
debug mode (in Python, no -O option), you want to know about it
immediately.  The asserts themselves add no additional functionality, so
when you're not running in debug mode you want them totally skipped
over.  (That's why, in C or C++, assert is actually a macro that expands
to nothing in non-debug mode.)

> I note that in unittest.py assert is not use,
> we have failIf or assert_ instead - thus if you avoided using assert
> in your own tests you can saftly run all your unittests optimised....

But that's because unit tests are something different, though obviously
both are related to testing.  Unit tests are standalone tests of a
component to make sure it behaves the way that's expected; asserts are
internal mechanisms used to make sure that invariants stay invariant or
that things that shouldn't happen, don't.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The most exhausting thing in life is being insincere.
\__/ Anne Morrow Lindbergh
    PyUID / http://www.alcyone.com/pyos/uid/
 A module for generating "unique" IDs in Python.



From sophy.cheeran at nettasking.com  Fri Apr  4 01:53:18 2003
From: sophy.cheeran at nettasking.com (Sophy)
Date: 3 Apr 2003 22:53:18 -0800
Subject: Python ODBC crash on failure of connection to MS SQL
Message-ID: <48e1ebc2.0304032253.5d8dd7b6@posting.google.com>

Hi,

I would appreciate ur advice on this problem that i am facing. 
I am running a python application and using the MSSQL database through
a python ODBC interface. I find that the python ODBC interface crashes
whenever there is a connection failure to the MSSQL. I get an "unable
to read memory" message coming up and a print message "disconnected"
which i later found in the ODBC.cpp source file in the method
odbcPrintError() which handles the dbi errors.
I use the same with MySQL and dont seem to have any problems with
connection failure handling.

Can somebody advice me as to how i can resolve this issue.

Thanks
Sophy



From grante at visi.com  Thu Apr 17 11:18:31 2003
From: grante at visi.com (Grant Edwards)
Date: 17 Apr 2003 15:18:31 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com> <3e9d9bb9$0$151$a1866201@newsreader.visi.com> <a8b7f07a.0304161314.361a80fa@posting.google.com> <3e9dca74$0$170$a1866201@newsreader.visi.com> <dSxna.114901$It5.77194@news2.central.cox.net>
Message-ID: <3e9ec5c7$0$178$a1866201@newsreader.visi.com>

In article <dSxna.114901$It5.77194 at news2.central.cox.net>, Steve Holden wrote:
> "Grant Edwards" <grante at visi.com> wrote in message
> news:3e9dca74$0$170$a1866201 at newsreader.visi.com...
>> In article <a8b7f07a.0304161314.361a80fa at posting.google.com>, A. Lloyd
> Flanagan wrote:
>>
>> >> > python << DONE
>> >> > print 2
>> >> > for i in (1,4):
>> >> >      print i
>> >> > DONE
>> >>
>> >> But how do you do it in a Makefile?
>> >
>> > Err.  Put it in a script file, and call the script file from make?
>>
>> That's one solution, but now you've got another file to keep
>> track of.
>>
>> > I tried sticking it in a Makefile, can't get it to work.  So far.
>>
>> How 'bout this:
>>
>> all:
>>         python2 -c $$'for i in range(5):\n print i\n\n'
>>
>>
>> I think that the $'<string>' notation is a bash-ism.  If that's
>> a problem, then
>>
>>         echo -e "for i in range(5):\n print i\n\n" | python2
>>
>> should also work.  Adjust the echo options appropriately to get
>> interpretation of \n to work.
>>
>> There may be other ways to get Make to interpret \n in a
>> string...
>>
> Now try putting an import before the "for" statement ...

OK...

  echo -e 'import sys\nfor i in range(5):\n sys.stdout.write(str(i)+"\\n")\n\n' | python2

Works for me.  Was it supposed to fail?

-- 
Grant Edwards                   grante             Yow!  While I'm in
                                  at               LEVITTOWN I thought I'd
                               visi.com            like to see the NUCLEAR
                                                   FAMILY!!



From dgemic at nospam.net  Sun Apr 27 17:51:46 2003
From: dgemic at nospam.net (Drazen Gemic)
Date: Sun, 27 Apr 2003 23:51:46 +0200
Subject: python web server modules
Message-ID: <b8hicc$qa4$1@brown.net4u.hr>

Is there any more detailed and explained example of Python HTTP server 
modules ? A Python documentation ist to brief, at least for me.

                                            TIA

						DG




From lekishvili at python.qartu.com  Fri Apr  4 13:34:32 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 4 Apr 2003 10:34:32 -0800
Subject: mp3 or wave  ---> midi , in python , how?
References: <b6isqq$23hs$1@mail.cn99.com>
Message-ID: <a3b5b608.0304041034.71b4a364@posting.google.com>

"Frank Zheng" <hust_zxq524 at 263.net> wrote in message news:<b6isqq$23hs$1 at mail.cn99.com>...
> Hi, All:
> 
>    Who can tell me how to do this in python ,are there any modules existing
> can help me?
> 
> 
> Thanks in advance
> 
> Frank

As far as I know, PyGame is a good option...



From kristofer at hotpop.com  Fri Apr 25 01:04:57 2003
From: kristofer at hotpop.com (Kristofer Wouk)
Date: Fri, 25 Apr 2003 05:04:57 GMT
Subject: variable scoping problem
In-Reply-To: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com>
References: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com>
Message-ID: <Zl3qa.1979$001.1638721@monger.newsread.com>

Tung Wai Yip wrote:
> Consider this
> 
> 
>>>>x=1
>>>>def foo():
> 
> ...   print x
> ...
> 
>>>>def bar():
> 
> ...   #print x   <- referenced before assignment error
> ...   x=2
> ...   print x
> ...
> 
>>>>foo()
> 
> 1
> 
>>>>bar()
> 
> 2
> 
>>>>print x
> 
> 1
> 
> 
> What rules in Python explain the scoping of x? Is there a name for the
> scope of the first x? 'module' scope? Why is bar() not able to access
> x but foo() can?
> 
> Wai Yip Tung
> 
Sorry, missed the bit about the error. I messed up...




From anton at vredegoor.doge.nl  Mon Apr 28 09:25:17 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 28 Apr 2003 15:25:17 +0200
Subject: Python-2.3b1.exe install error on W98se
References: <b8itdt$53h$1@news.hccnet.nl> <3EAD1ADF.27E9E4A5@engcorp.com>
Message-ID: <b8ja52$k77$1@news.hccnet.nl>

Peter Hansen <peter at engcorp.com> wrote:

>Hmm.... clearly the download was unsuccessful.  In fact, Python 2.3b1
>*has* grown to over 1.8G, and you didn't wait long enough to download
>the whole thing.  (Are you on dialup or something?)  The reason for
>the new, much larger installation package, in addition to all the new
>features which were crammed in at the last minute, is that the PSU

Last minute? That must be the sum builtin then. This finally proves
that using reduce results in shorter code :-) Ok, having a little fun
is nice, but please lets not forget that this really happening on
three W98se machines here. No one else has confirmed this result yet,
so I can't be sure whether it's a local issue or not.

Anton



From heiland at ncsa.uiuc.edu  Fri Apr 25 09:40:34 2003
From: heiland at ncsa.uiuc.edu (Randy Heiland)
Date: Fri, 25 Apr 2003 08:40:34 -0500
Subject: PyXML - setContentHandler??
References: <3EA91EA8.2050505@ncsa.uiuc.edu> <b8b87m$rs4$01$1@news.t-online.com>
Message-ID: <3EA93AD2.30208@ncsa.uiuc.edu>

Martin v. L?wis wrote:
> Randy Heiland wrote:
> 
>> I'm using a chunk of code that does:
>>
>>  >>> parser = xml.sax.make_parser()
>> <define a handler>
>>  >>> parser.setContentHandler(handler)
>>
>> however, using PyXML, I get:
>> AttributeError: SAX_expat instance has no attribute 'setContentHandler'
> 
> 
> Where do these come from? I.e. what is xml.sax.__file__, and
> parser.__class__.__module__?
> 
> Regards,
> Martin

from here:

 >>> xml.sax.__file__
'/usr/local/lib/python2.3/site-packages/_xmlplus/sax/__init__.pyc'
 >>>
 >>> parser.__class__.__module__
'xml.sax.drivers.drv_pyexpat'





From SBrunning at trisystems.co.uk  Thu Apr 17 09:21:36 2003
From: SBrunning at trisystems.co.uk (Simon Brunning)
Date: Thu, 17 Apr 2003 14:21:36 +0100
Subject: Python for VB6 programmers (was Re: Complexity of standard Py
	thon data structures)
Message-ID: <31575A892FF6D1118F5800600846864D0120082F@intrepid>

> From:	rzed [SMTP:Dick.Zantow at lexisnexis.com]
> A little off the thread's topic, but ... I always liked RPG, though it
> is far more pleasant when you stay reasonably close to its main
> purpose -- to generate reports. I'm not sure what the "400 kicks at
> the can" means, but I'm quite sure this is not version 400, but the
> version of RPG used by the AS/400. The code looks essentially the same
> as RPGIII, which was what I used first.
 
RPG/400 *is* RPGIII, or rather, it's what you order from IBM if you want it.

Or it *was*. These days *everything* is called WebSphere.

RPG is horrid, but it's usually the only option on a '400, and it gets the
job done.

Cheers,
Simon Brunning
TriSystems Ltd.
sbrunning at trisystems.co.uk




-----------------------------------------------------------------------
The information in this email is confidential and may be legally privileged.
It is intended solely for the addressee. Access to this email by anyone else
is unauthorised. If you are not the intended recipient, any disclosure,
copying, distribution, or any action taken or omitted to be taken in
reliance on it, is prohibited and may be unlawful. TriSystems Ltd. cannot
accept liability for statements made which are clearly the senders own.




From tweedgeezer at hotmail.com  Wed Apr 23 15:41:39 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 23 Apr 2003 12:41:39 -0700
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <NKxpa.85054$T34.2112670@news2.tin.it>
Message-ID: <698f09f8.0304231141.5d458eb0@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<NKxpa.85054$T34.2112670 at news2.tin.it>...
> However, such new functionality doesn't get back-ported to previous
> releases of Python, such as 2.2.*.  In all Python releases from 1,6
> included to 2.3 excluded, "needle.find(haystack) > 0" is the idiom.
> Module string is needed if you want to use antediluvian versions of
> Python, such as 1.5.2, where strings didn't have methods.

Of course, you  meant "haystack.find(needle) >= 0", right?

I always use (and have seen used) "haystack.find(needle) != -1", myself.

Jeremy



From mwm at mired.org  Thu Apr  3 15:13:43 2003
From: mwm at mired.org (Mike Meyer)
Date: Thu, 03 Apr 2003 20:13:43 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <x7of3qjh4g.fsf@guru.mired.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com> <wBIia.2000$y3.310439@news010.worldonline.dk>
Message-ID: <x7adf7ibk6.fsf@guru.mired.org>

"Anders J. Munch" <andersjm at inbound.dk> writes:
> "Dave Benjamin" <ramen at lackingtalent.com> wrote:
> > In article <b6ckjr$903$1 at panix2.panix.com>, Roy Smith wrote:
> > > The bottom line is exactly what Mike said: in interpreted languages,
> > > the more you can push down into the libraries and interpreter core,
> > > the better off you are most of the time.  And profiling helps too :-)
> > This has been my experience as well. It is a very gratifying discovery,
> > IMHO, because it encourages a less-is-more style of programming where you
> > say what you mean, and little else.
> The downside is that sometimes code is contorted to fit whatever has a
> fast high-level implementation in the current interpreter, instead of
> adapting to the nature of the problem.
> For a Python example, see the Decorate-Sort-Undecorate pattern.  In
> C++ the straightforward and fast solution is to use a suitable
> comparison predicate.  In Python you can do the same, but that's slow,
> and you are left with a choice between clarity and speed.

The statement about the preferred method in C is simply false. Anytime
the "suitable comparison predicate" is sufficiently complex - for
instance, what the unix sort(1) command allows - a DSU that allows
simple comparisons will be faster than a sort with complex
comparisons.

In fact, any time you can simplify the comparison with DSU, the DSU
sort will always perform faster on n objects for sufficiently large n.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From pinard at iro.umontreal.ca  Tue Apr  1 20:39:56 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 01 Apr 2003 20:39:56 -0500
Subject: converting from shell script to python
In-Reply-To: <kfn0jaoig1.fsf@shaoz.activesky.com.au>
References: <kfn0jaoig1.fsf@shaoz.activesky.com.au>
Message-ID: <oqsmt17k43.fsf@alcyon.progiciels-bpi.ca>

[Shao Zhang]

> I need to convert a very simple shell script to python in order get it
> run on both windows and unix.

Another possibility is to install a ported `bash' on Windows.  I do not use
Windows often, but the last time I had to, a while ago, `bash' was working
quite well.  I used the one from DJGPP -- Cygwin probably has one too.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From afriere at yahoo.co.uk  Wed Apr 30 03:21:01 2003
From: afriere at yahoo.co.uk (Asun Friere)
Date: 30 Apr 2003 00:21:01 -0700
Subject: Raising objects
References: <453d10b2.0304290750.222fb453@posting.google.com> <tpltav0bbeehseajdo3osampi7flsvhk6g@4ax.com>
Message-ID: <38ec68a6.0304292321.2d7b5647@posting.google.com>

Jan Olderdissen <jolderdissen at ixiacom.com> wrote in message news:<tpltav0bbeehseajdo3osampi7flsvhk6g at 4ax.com>...

> That said, I question the point of deriving from object as classes
> (and in fact all objects as far as I can tell) are derived from object
> as it is. Since object isn't a class, deriving from it probably is
> bogus to begin with. 

Surely deriving from 'object' would have to be the most commonly
method of ensuring that your class is a NewStyleClass, no?



From sholden at holdenweb.com  Wed Apr 16 15:13:30 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 19:13:30 GMT
Subject: Python and Schools
References: <mailman.1050499405.10336.python-list@python.org> <eGdna.110601$It5.106431@news2.central.cox.net> <mailman.1050503666.19709.python-list@python.org>
Message-ID: <uXhna.110755$It5.95103@news2.central.cox.net>

"Andrew Bennetts" <andrew-pythonlist at puzzling.org> wrote ...
> On Wed, Apr 16, 2003 at 02:22:02PM +0000, Steve Holden wrote:
> [...]
> >
> > For a real example, consider the contortions I have to go through when
> > writing VBScript ASP code for a client's project if I want a list
(array) of
> > strings. Where in Python I can write:
> >
> >     sarr = ['the', 'quick', 'brown', 'fox', 'jumps' 'over', 'the',
'lazy',
> > 'dog]
> >
> > in VBScript I must resort to one of [nontested code follows]:
> >
[...]
>
> Or, if my memory isn't failing me:
>
>     sarr = Array("the", "quick", ... "dog")
>

Indeed, though that's new to me (never used VB, just the scripting version).
Now let's create a dictionary-of-objects equivalent using the
DictionaryOfObjects() function ;-)

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From abonet at yahoo.com  Wed Apr 23 20:51:59 2003
From: abonet at yahoo.com (Angelo B)
Date: 23 Apr 2003 17:51:59 -0700
Subject: Tkinter itemconfigure and memory leaks
Message-ID: <8e9d8e42.0304231651.43d7ef9e@posting.google.com>

The itemconfigure() method of Listbox seems to create a memory leak in
the following use case.  Comment out the call to itemconfigure() and
the memory leak disappears.  I'm running Python 2.2 (I think you need
at least 2.2 for the itemconfigure method to even be present on
Listbox).  I've seen the leak on both Compaq Tru64 5.1A and SunOS 5.6.

Is this a genuine bug?  Or is there some technique that I don't know
about to prevent this (i.e. a different way of coloring items in a
Lisbox)?

Any help or insight is appreciated.


import Tkinter as Tk
Lb = None

def update_lb ():
    global Lb    
    Lb.delete(0, Tk.END)

    for ii in range(100):
	Lb.insert(Tk.END, 'Item %d' % ii)
	Lb.itemconfigure(ii, bg='red')

    Lb.after(10, update_lb)


root = Tk.Tk()
Lb = Tk.Listbox(root)
Lb.pack()
Lb.after(1000, update_lb)
root.mainloop()



From skip at pobox.com  Fri Apr 11 09:58:39 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 11 Apr 2003 08:58:39 -0500
Subject: collecting data from file
In-Reply-To: <3E9671A3.2050004@sympatico.ca>
References: <3E9671A3.2050004@sympatico.ca>
Message-ID: <16022.51727.12214.751076@montanaro.dyndns.org>

    Mustafa> I want to scroll thru a file;
               [ and do stuff with it ]

    Mustafa> Any tips?

As others have mentioned, your problem was underspecified.  If your data is
tabular with a unique separator between fields, the new csv package
(available from Python CVS) should fill the bill.  Curren development
version of the docs are here:

    http://www.python.org/dev/doc/devel/lib/module-csv.html

Skip




From puccio_13 at yahoo.it  Fri Apr 25 05:36:01 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Fri, 25 Apr 2003 11:36:01 +0200
Subject: GUI libs
Message-ID: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>

Hi, I'd like to make a GUI for a program written in python, meant to run on
mac, win and linux: which is the preferable solution? I saw around tkinter,
wxpython and pyqt....

-- 
bye
Alessio Pace



From aahz at pythoncraft.com  Sat Apr 26 15:37:45 2003
From: aahz at pythoncraft.com (Aahz)
Date: 26 Apr 2003 15:37:45 -0400
Subject: In which versions is list.sort stable?
References: <slrnbalhft.bc1.mlh@furu.idi.ntnu.no> <yyAqa.8330$3M4.231032@news1.tin.it>
Message-ID: <b8en69$263$1@panix2.panix.com>

In article <yyAqa.8330$3M4.231032 at news1.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>Magnus Lie Hetland wrote:
>>
>> Has list.sort always been stable? Or, if not, in which Python versions
>> is it not stable?
>
>list.sort has not been stable in 1.5.2, 2.0.*, 2.1.*, 2.2.*.  I do not
>know about previous versions.  Note that list.sort may once again cease
>being stable (in 2.4.* or further versions) if the timbot dreams up one
>that's even better on _other_ criteria -- stability's not going to
>constrain his algorithm choices.

OTOH, Uncle Timmy won't gratuitously remove stability, either.  But I
still wouldn't bet on Python's sort staying stable past 2.3.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From no at mail.com  Tue Apr 29 17:36:07 2003
From: no at mail.com (Jeroen Wolff)
Date: 29 Apr 2003 21:36:07 GMT
Subject: os.spawn
References: <Xns936C9D9E91B1Bthewolffman@194.109.133.29> <mailman.1051624647.29525.python-list@python.org>
Message-ID: <Xns936CF02836B77jwolffenertelnl@194.109.133.29>

> You want to use os.system or os.popen.  With popen you can read the
> results directly without using the intermediate file. 
> 
> os.system('flow-cat <dirname> | flow-stat <param> > outputfile.txt')
> 
> or
> 
> flow = os.popen('flow-cat <dirname> | flow-stat <param>')
> 
> for i in flow.readlines():
Thx, but i need the .txt file because i have to parse it several time...
with os.system( ) is the script waiting until the command is finished?
I have to wait until i can open the generated .txt file

Jeroen



From staschuk at telusplanet.net  Sun Apr 20 12:37:26 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 20 Apr 2003 10:37:26 -0600
Subject: Sorting without transitivity
In-Reply-To: <7h3y925o596.fsf@pc150.maths.bris.ac.uk>; from mwh@python.net on Sun, Apr 20, 2003 at 10:03:19AM +0000
References: <20030419213405.GA588@obstler> <mailman.1050821403.13707.python-list@python.org> <7h3y925o596.fsf@pc150.maths.bris.ac.uk>
Message-ID: <20030420103726.A255@tibia.amotlpaa.bogus>

Quoth Michael Hudson:
  [...]
> Do you (or anyone else) know why it's called a *topological* sort?

Not really.  Knuth says: "The problem of topological sorting is to
embed the partial order in a linear order."  [TAoCP, 2.2.3.]

> Don't see no topology here, off hand.
>
> It could be that the open subsets of a topological space are a poset
> under inclusion, I guess.

Hm... poset == partially ordered set?

Certainly inclusion is a partial order of subsets, and a
topological sort finds an inclusion-preserving map from any space
S to a space T in which inclusion satisfies the trichotomy
condition.  Is that what topologists mean by "embed S in T"?

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From max at alcyone.com  Tue Apr  1 22:53:45 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 01 Apr 2003 19:53:45 -0800
Subject: recursion confusion
References: <2621b014.0304011909.d56a12a@posting.google.com>
Message-ID: <3E8A5EC9.854BF60B@alcyone.com>

Jon Schull wrote:

> Sometimes recursion befuddles me.  This is one of those times.
> 
> I want to convert a graph structure into a nested string
> For example I want this...
> 
> g={0:[], 1:[11], 2:[21,  22], 22:[221,222]}
> 
> ...to become this
>         "0,     1[11], 2 [21,  22[221, 222]] "
> 
> (with 221 and 222 nested under 22 which is also nested under 2)
> 
> Of course, I want this to work for arbitrary nesting, and for strings
> as well as numbers.

Seems relatively straightforward.  First, get a list of the graph nodes
that are top-level (i.e., those which are not the children of any node).
Then, write a recursive function which displays a node by its ID, and
its children, if it has any, in brackets afterward.  Then take that list
of top-level nodes and run the function on each one, separating the
result with commas.  Add spaces and/or formatting to taste.

The function would be something like this:

	def render(node):
	    result = str(node)
	    children = getNodeChildren(node) # write this function
	    if children:
	        result += '[%s]' % ', '.join(map(render, children))
	    return result

>From looking at your routine, it looks like getNodeChildren would simply
be written as

	def getNodeChildren(node):
	    return G.get(node, []) # calling your global dictionary G

Obviously, this might need improvement depending on your graph.  If it
has cycle, then this will recurse indefinitely.

I leave it as an exercise for you to figure out how to find the
top-level nodes.  (Hint:  What property makes them top-level?  How would
you test for that?  If you use a recursive algorith, how would you
communicate that information down through the function calls?)

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The people are to be taken in very small doses.
\__/ Ralph Waldo Emerson
    Erik Max Francis' bookmarks / http://www.alcyone.com/max/links/
 A highly categorized list of Web links.



From BPettersen at NAREX.com  Thu Apr 24 15:35:02 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Thu, 24 Apr 2003 13:35:02 -0600
Subject: Question about accessing class-attributes.
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE101@admin56.narex.com>

> From: Alex Martelli [mailto:aleax at aleax.it] 
> 
[...]
> >
> >   class AbstractBaseCounter(object):
> >       __InstCount = {}
> >       def InstCount(cls):
> >           return AbstractBaseCounter.__InstCount[cls]
> >       InstCount = classmethod(InstCount)
> >
> >       def __init__(self):
> >           cls = self.__class__
> >           if cls == AbstractBaseCounter:
> >               raise TypeError("Attempt to instantiate abstract base
> > class")
> >           d = AbstractBaseCounter.__InstCount
> >           if cls in d:
> >               d[cls] += 1
> >           else:
> >               d[cls] = 1
> >
> >   class A(AbstractBaseCounter):
> >       def __init__(self):
> >           AbstractBaseCounter.__init__(self)
> >
> >   class B(A): pass
> >
> >   a1=A()
> >   a2=A()
> >   b1=B()
> >   b2=B()
> >   print B.InstCount()
> 
> comparing the complication of this one, and the simplicity of 
> the custom metaclass I posted and you snipped, PLUS the minor 
> risks such as "B overriding __init__ and forgetting to call 
> its superclass's"(:-), I think the custom metaclass is 
> vindicated;-).

Oh, no argument it's easier. And although the above has problems,
metaclasses have their own set of problems, e.g.:

For reference:
>>> class metaIC(type):
...     def __new__(cls, name, bases, dict):
...         dict['InstCount'] = 0
...         return type.__new__(cls, name, bases, dict)
...     def __call__(cls, *args, **kwds):
...         cls.InstCount += 1
...         return type.__call__(cls, *args, **kwds)

and the (now canonical) metaMetaBunch (munched version [my apologies],
real version here:
http://mail.python.org/pipermail/python-list/2002-July/112007.html):

 class metaMetaBunch(type):
    def __new__(cls, classname, bases, classdict):
        def __init__(self, **kw):
            for k in self.__dflts__: setattr(self, k, self.__dflts__[k])
            for k in kw: setattr(self, k, kw[k])

        newdict = {'__slots__':[], '__dflts__':{}, '__init__':__init__}

        for k in classdict:
            if k.startswith('__'):
                newdict[k] = classdict[k]
            else:
                newdict['__slots__'].append(k)
                newdict['__dflts__'][k] = classdict[k]

        return type.__new__(cls, classname, bases, newdict)

Then define a hierarchy:

 class Counted(object):
	__metaclass__ = metaIC
	
 class Point(object):
	__metaclass__ = metaMetaBunch
	x = 0
	y = 0
	
 class CountedPoint(Counted, Point):
     def __init__(self):
         super(CountedPoint.self).__init__(self)
		
 cp = CountedPoint()
 print cp	

we get:

 Traceback (most recent call last):
  File "M:\python\cpoint.py", line 36, in ?
    class CountedPoint(Counted, Point):
  File "M:\python\cpoint.py", line 6, in __new__
    return type.__new__(cls, name, bases, dict)
 TypeError: metatype conflict among bases

not exactly clear _what_ the conflict is, i.e. it's not that the meta
classes must have a common base (both isinstance(x, type), it's not that
type.__new__ should perhaps be super(x, type).__new__, it _is_ that the
metaclasses aren't related by inheritance. Trying to fix this by, e.g.
making metaIC a subclass of metaMetaBunch (and changing the __new__ call
appropriately), gives you:

Traceback (most recent call last):
  File "M:\python\cpoint.py", line 38, in ?
    class CountedPoint(Counted, Point):
  File "M:\python\cpoint.py", line 25, in __new__
    return metaMetaBunch.__new__(cls, name, bases, dict)
  File "M:\python\cpoint.py", line 20, in __new__
    return type.__new__(cls, classname, bases, newdict)
TypeError: multiple bases have instance lay-out conflict

which means you're stuck (?)

> > which happens to work...? I've verified that
> > AbstractBaseCounter.__init__ gets called for B(). Is this
> > correct/documented/guaranteed/works with MI/multiple args/old-style
> > classes, etc., etc.?
> 
> well, you COULD get dual calls to that __init__ in presence of
> "inheritance diamonds" -- using super may be safer.

True, although for general use, super might have more issues than it's
worth...

-- bjorn
 




From Mike.Prager at noaa.gov  Tue Apr 15 12:15:18 2003
From: Mike.Prager at noaa.gov (Michael Prager)
Date: Tue, 15 Apr 2003 12:15:18 -0400
Subject: book on Python for Fortran programmers
References: <3064b51d.0304150458.595afc30@posting.google.com>
Message-ID: <vhbo9vgvsqjg8l7s0aluijrran56cg8mlm@4ax.com>

beliavsky at aol.com wrote:

>Python has been suggested by several posters to comp.lang.fortran as a
>scripting language worth learning. What are some good book(s) and web
>site(s) to get an experienced Fortran 95 programmer started in Python?

Of the books that I've looked at, I prefer the "Python 2.1
Bible" by Brueck and Tanner.  It's rather complete and doesn't
start by explaining what a bit is.

An . . . essential reference . . . is "Python Essential
Reference, Second Edition" by David Beazley.  It's a concise
version of everything (reminiscent in its brevity of Metcalf and
Reid) whose main drawback is the very small type size used.

Whatever you buy, make sure it's current at least to Python 2.1.
(The current version is 2.2.2.)  AVOID any reference that is
1.5.x or 1.6.x -- some worthwhile features have been added..

-- 
Mike Prager
NOAA, Beaufort, NC
* Opinions expressed are personal and not represented otherwise.
* Any use of tradenames does not constitute a NOAA endorsement.



From cct at tavaresstudios.com  Sat Apr 12 21:08:58 2003
From: cct at tavaresstudios.com (Chris Tavares)
Date: Sun, 13 Apr 2003 01:08:58 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com>
Message-ID: <KM2ma.22486$4P1.1970973@newsread2.prod.itd.earthlink.net>

"Roy Smith" <roy at panix.com> wrote in message
news:roy-06B012.20145112042003 at reader1.panix.com...
[ ... snip ... ]
>
> As another example, I can guess that Python dictionaries work like STL
> maps on the inside, but that would be just a guess.  Is inserting an
> element logN?  Is keys() linear?  Is del on an element logN?  Is
> has_key() logN?  Those would be my guesses.  But they're just guesses.
>

STL maps are implemented using binary trees (red-black trees usually), so
that's why they have their logN performance numbers.

Python dicts are implemented using a hash table. As a result, their typical
behavior is O(1) - constant time to get any key. Unless there are hash
collisions of course.

The big difference besides performance is that STL maps store the keys in
sorted order, where Python dict keys are "randomly" ordered.

-Chris





From news03 at zurgy.org  Tue Apr 22 14:36:50 2003
From: news03 at zurgy.org (Ben Lamb)
Date: Tue, 22 Apr 2003 19:36:50 +0100
Subject: Compiling a program the embeds python
References: <3ea51711$0$4862$cc9e4d1f@news.dial.pipex.com> <slrnbaabum.uvq.rmunn@localhost.localdomain>
Message-ID: <3ea58bbe$0$11385$cc9e4d1f@news.dial.pipex.com>

Thanks very much for your reply which solved the problem. My logic and gcc's
are clearly at odds. As the for man page, I'm being shortchanged, mine only
has 2135 lines. I'll get the full docs.

Thanks again,

Ben.

Robin Munn wrote:

> Ben Lamb <news03 at zurgy.org> wrote:
>> 
>> I'm trying to compile an example program written in C that embeds Python.
>> The porgram is:
>> 
> [snip]
>> 
>> I use the following to compile it:
>> 
>> ben at localhost:~/prog/pyembed$ gcc -I/usr/local/include/python2.2
>> -L/usr/local/lib/python2.2/config -Xlinker -export-dynamic -lm
>> -lpython2.2 -ldl -lpthread -lutil pyembed.c
>> 
> [snip]
>> 
>> Please can someone point out what I am doing wrong.
> 
> The -l options, which specify which libraries to link against, need to
> come *after* the file(s) that need those libraries. Yes, this breaks the
> Unix rule-of-thumb that options always come first, and parameters
> second. Your command line should be:
> 
>     ben at localhost:~/prog/pyembed$ gcc -I/usr/local/include/python2.2
>     -L/usr/local/lib/python2.2/config -Xlinker -export-dynamic pyembed.c
>     -lm -lpython2.2 -ldl -lpthread -lutil
> 
> Check the gcc manpage, and read the section that describes the -l
> command. On my system, it starts at line 3630 of the manpage. You'll
> find the following paragraph:
> 
>     It makes a difference where in the command you write this option;
>     the linker searches and processes libraries and object files in the
>     order they are specified.  Thus, foo.o -lz bar.o searches library z
>     after file foo.o but before bar.o.  If bar.o refers to functions in
>     z, those functions may not be loaded.
> 
> Hope this helps.
> 

-- 
news.dsl.pipex.com



From dale at riverhall.NOTHANKS.co.uk  Wed Apr 16 12:01:13 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 16 Apr 2003 17:01:13 +0100
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com> <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com> <cnoq9vci702imrunkenpulnefautq0n5ua@4ax.com> <Plena.22932$LB6.544795@news1.tin.it>
Message-ID: <3cvq9vc9smcp7fbott8ivb1ujv0bgsk855@4ax.com>

Alex Martelli <aleax at aleax.it> wrote:

>Dale Strickland-Clark wrote:
>
>> 
>> Obviously, there's a 'not' missing from the above :-(.
>
>Hmmmm, are you sure...?  This raises when the attribute is
>missing, which seems likely to be what one desires.

No. I'm being a dick. :-(

>In a case such as this, the idiom would reduce, at a minimum, to:
>
>def f(fileobj):
>    write = fileobj.write
>    # rest of function follows -- just call write('...') when needed
>
>by extracting the bound-method as the first action of the function,
>you ensure the exception happens ASAP -- *AND* you get a nice little
>performance boost in many cases too.  You can spend some of that
>boost to do a little bit more checking than the above snippets show,
>for example:
>
>def f(fileobj):
>    write = fileobj.write
>    write('')
>    # rest of function follows -- just call write('...') when needed

Way too fussy and obscure for my liking. Also, the assert is debugging
code and nicely goes away in production. 

-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From sismex01 at hebmex.com  Wed Apr 16 15:22:03 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Wed, 16 Apr 2003 14:22:03 -0500
Subject: How many of you are Extreme Programmers?
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F0623304615031@mail.hebmex.com>

> From: Steven Taschuk [mailto:staschuk at telusplanet.net]
> Sent: Wednesday, April 16, 2003 1:24 PM
> 
> Quoth sismex01 at hebmex.com:
>   [...]
> > Many of us have used XP techniques without actually knowing
> > that our methodology was labelled as "XP", it comes natural
> > in the scope of our tools and goals.
> > 
> > Tool:   Python.
> > Goal:   Write specs-conforming programs and modules.
> > Method: Create tests which verify the code.
> > 
> > I don't know; "OOP", "XP", "Structured Programming", etc.. are
> > all monickers, buzzwords, fads, which come and go; [...]
> 
> To be fair, there's a great deal more to XP than just writing
> tests (as two minutes on the web will show you).  And verifying
> conformance to specs is only a small part of the function of tests
> in XP, or in TDD generally.
>

Of course; "XP" is merely the "standarization" of said
practices.  What I was trying to say was that these practices
didn't appear with XP, as many authors would like to claim,
but only that "XP" is a distillation of said practices.

> 
> The hype may well be annoying -- it annoys me, certainly -- but
> that doesn't mean the methodologies are in fact substanceless,
> or not worth learning about.
>

The tragic part about the hype is that, many people (such
as myself) who like a certain tendency, or technology, or
process, etc (what have you), automatically are turned off
by excessive marketing, detecting a large possibility of it
being vapour, which many times, it is. :-/

-gus




From hal_wine at yahoo.com  Wed Apr  2 11:21:24 2003
From: hal_wine at yahoo.com (Hal Wine)
Date: Wed, 02 Apr 2003 08:21:24 -0800
Subject: Tkinter under cygwin (was Re: converting from shell script to python)
In-Reply-To: <mailman.1049248805.1309.python-list@python.org>
References: <mailman.1049248805.1309.python-list@python.org>
Message-ID: <3E8B0E04.7000203@yahoo.com>

Jason Tiller wrote:
> I'm using Cygwin's Python 2.2.2 + Tkinter + Pmw to go through the
> examples in John Gray's "Python and Tkinter Programming," and the
> environment is *very* good.

I, too, couldn't live on win32 w/o cygwin. Could you give a few 
more details of how you have Tkinter integrated with cygwin?

I.e. the official cygwin packages don't have Tkinter included. 
What all did you have to compile, and from where did you fetch it?

Thanks,
--Hal




From wweexxsseessssaa at telusplanet.net  Wed Apr  2 10:13:33 2003
From: wweexxsseessssaa at telusplanet.net (John Hall)
Date: Wed, 02 Apr 2003 15:13:33 GMT
Subject: A question about help() and docstrings
References: <20030402092404.7a466b0c.rjkuhns@geetel.net> <mailman.1049294705.429.python-list@python.org>
Message-ID: <k8vl8v42ms027dunfsunsf6forcq34uhp4@4ax.com>

On Wed, 2 Apr 2003 06:43:10 -0800, Sean 'Shaleh' Perry
<shalehperry at attbi.com> wrote:

>look at site.py in your python directory.  help() is defined there as a 
>function wrapping an instance of a class.  The class comes from pydoc.py in 
>the same directory.  Give it a peak.


I think you mean "peek" rather than "peak" here.
And probably not "pique" either.

I haven't done a programming "peek" or "poke" since the days of Atari
Basic in the early 80's. I love those graphic names.

-- 
John W Hall <wweexxsseessssaa at telusplanet.net>
Calgary, Alberta, Canada.
"Helping People Prosper in the Information Age"



From paul at prescod.net  Mon Apr  7 16:33:57 2003
From: paul at prescod.net (Paul Prescod)
Date: Mon, 07 Apr 2003 13:33:57 -0700
Subject: Problem setting COM property using win32com
In-Reply-To: <E1B4F80EB080194BBAD3092392F511B317F2B7@XCH-NW-13.nw.nos.boeing.com>
References: <E1B4F80EB080194BBAD3092392F511B317F2B7@XCH-NW-13.nw.nos.boeing.com>
Message-ID: <3E91E0B5.9010904@prescod.net>

I don't know why that's happening, but I would suggest you take more 
control of the instantiation by doing something like this:

from win32com.client.gencache import EnsureModule
MyModule = EnsureModule('{3BBEE982-742B-11D2-BD0A-00A024C36143}', 0, 1, 0)

reclist = MyModule.SmRecList()

In my recent COM project, I used the "dispatch" mechanism and pythoncom 
was smart enough to pick up my gen_py'd module. But the mechanism above 
is fine too. In fact, it is probably better by virtue of being more 
explicit...

Green, Gregory P wrote:
> I don't think the gen_py class is being used:
> 
> print obj: <COMObject SmRecList.SmRecordList>
> 
> print obj.__class__: win32com.client.CDispatch
> 
> print dir(obj): ['_ApplyTypes_', '_LazyAddAttr_', '_NewEnum', '_Release_', '__AttrToID__', '__LazyMap__', '__call__', '__cmp__', '__doc__', '__getattr__', '__getitem__', '__init__', '__int__', '__len__', '__module__', '__nonzero__', '__repr__', '__setattr__', '__setitem__', '__str__', '_builtMethods_', '_enum_', '_find_dispatch_type_', '_get_good_object_', '_get_good_single_object_', '_lazydata_', '_make_method_', '_mapCachedItems_', '_oleobj_', '_olerepr_', '_print_details_', '_proc_', '_unicode_to_string_', '_username_', '_wrap_dispatch_']






From exarkun at intarweb.us  Sun Apr  6 01:24:57 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 6 Apr 2003 01:24:57 -0500
Subject: properties and get/set methods
In-Reply-To: <3E8FB1A9.20507@rogers.com>
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <3E8FB1A9.20507@rogers.com>
Message-ID: <20030406062457.GB25203@meson.dyndns.org>

On Sat, Apr 05, 2003 at 11:48:41PM -0500, Mike C. Fletcher wrote:
> Dan Bishop wrote:
> 
> >Timothy Grant <tjg at craigelachie.org> wrote in message 
> >news:<mailman.1049581810.25706.python-list at python.org>...
> > 
> >
> ...
> 
> >>Is it possible to determine which attribute was used to as the trigger 
> >>for the call to setvar() or getvar()? so that set_appropriate_attribute() 
> >>or appropriate_attribute() can set the correct attribute or get the 
> >>correct attribute?
> >>   
> >>
> >
> >Why not just use two different getvar and setvar functions?
> >
> 
> When you're doing generic "framework" level programming, it's just not 
> acceptable to have to define 2 (really 3) new functions for every 
> property.  Compare:

  Indeed not.  So move things around a little:

    def makeSetter(name):
        def set(self, value):
            doGeneric(self, name, value)
            self.__dict__[name] = value
        return set
    def makeGetter(name):
        def get(self):
            doGeneric(self, name)
            return self.__dict__[name]
        return get
    def makeProperty(name):
        return property(makeGetter(name), makeSetter(name))

    class Foo:
        bar = makeProperty('bar')
        baz = makeProperty('baz')

  Logically, the two layers still exist (good thing), but the first is
resolved entirely at class-creation time, nicely negating the efficiency
concerns (good thing).

> [snip]

  Jp

-- 
In the days when Sussman was a novice Minsky once came to him as he sat
hacking at the PDP-6. "What are you doing?" asked Minsky. "I am training a
randomly wired neural net to play Tic-Tac-Toe." "Why is the net wired
randomly?" asked Minsky. "I do not want it to have any preconceptions of how
to play." Minsky shut his eyes. "Why do you close your eyes?" Sussman asked
his teacher. "So the room will be empty." At that moment, Sussman was
enlightened.
 -- 
 up 17 days, 2:00, 5 users, load average: 1.13, 1.17, 1.09




From trentm at ActiveState.com  Mon Apr 14 15:28:34 2003
From: trentm at ActiveState.com (Trent Mick)
Date: Mon, 14 Apr 2003 12:28:34 -0700
Subject: Case-insensitive string comparison
In-Reply-To: <3e9b09bc@news.012.net.il>; from avner@skilldesign.com on Mon, Apr 14, 2003 at 09:19:20PM +0200
References: <3e9b09bc@news.012.net.il>
Message-ID: <20030414122834.B30958@ActiveState.com>

[Avner Ben wrote]
> Question:
> 
>     Does standard Python have case-insensitive string comparison and search?


You can do it by just lowercasing (or uppercasing) your strings before
searching:

    >>> s = "Blah FOO spam"
    >>> q = "Foo"
    >>> s.lower().find(q.lower())
    5


Or, if you are using regular expressions, use the IGNORECASE flag:

    >>> import re
    >>> match = re.search(q, s, re.IGNORECASE)
    >>> match.start()
    5
    >>> match.end()
    8


Cheers,
Trent

-- 
Trent Mick
TrentM at ActiveState.com




From mwh at python.net  Mon Apr 14 10:29:48 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 14 Apr 2003 14:29:48 GMT
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it> <yfshe92dmpt.fsf@black132.ex.ac.uk> <a8vma.11117$LB6.312431@news1.tin.it>
Message-ID: <7h365phtan7.fsf@pc150.maths.bris.ac.uk>

Alex Martelli <aleax at aleax.it> writes:

> If AMK could be fooled, so could be anyone -- I'm not interested in
> defending descrintro.html as particularly lucid, well-written, and
> detailed in its coverage of __slots__.

It was written by Guido!

Cheers,
M.

-- 
  Remember - if all you have is an axe, every problem looks 
  like hours of fun.                                        -- Frossie
               -- http://home.xnet.com/~raven/Sysadmin/ASR.Quotes.html



From dan_griffith at sprynet.com  Tue Apr 15 10:38:57 2003
From: dan_griffith at sprynet.com (Dang Griffith)
Date: Tue, 15 Apr 2003 14:38:57 GMT
Subject: any benchmark results for python?
References: <b7h40n$28s$1@skates.gsfc.nasa.gov>
Message-ID: <5QUma.25905$4P1.2327934@newsread2.prod.itd.earthlink.net>

Check out technical report #2, and the refereed journal papers #6 and #7
at

http://www.ipd.uka.de/~prechelt/Biblio/#pspeTR

   --dang


J Li wrote:

> Hi,
> 
> I am looking for benchmark test results for python, in comparison to other
> languages like Perl or Java or C++
> 
> thanks
> 
> jylimd(at)yahoo(dot)com
> 
> 




From pobrien at orbtech.com  Sun Apr  6 17:29:39 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 16:29:39 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org>
Message-ID: <m2k7e771rw.fsf@localhost.localdomain>

Ian Bicking <ianb at colorstudy.com> writes:

> On Sun, 2003-04-06 at 15:28, Patrick K. O'Brien wrote:
> > Actually, we don't have to clone everything about the interface.  I'm
> > using wxPython, so the interface is pretty and more easily configured.
> > But some people actually like the minibuffer, so we'll need to support
> > something like it and something less ugly.  Heck, I'm actually running
> > three different versions of the interface now, and I only wrote the
> > code this week (minus the almost two year's worth of code in PyCrust).
> 
> No, don't change the interface!  It's one of the best parts about
> Emacs... it doesn't use all the stupid crap that most GUIs put in.  The
> minibuffer is vastly superior to popup dialogs.  The keyboard focus is
> simple and actually works, unlike most GUIs.  M-x actually works, where
> menus for that many commands wouldn't work well at all.

I agree.  Plus dired, plus the cvs integration.  No need for a mouse
at all.  My main point was that it should be easy for us to experiment
with variations, or allow the user to change the interface, since
we're using a real GUI toolkit.  For example, some users might like
the minibuffer to be at the top of the frame, rather than the bottom.
Maybe there's a way to configure that in Emacs, but I've never tried
or seen mention of that one.  And some users will rather have a dialog
box and click on things with a mouse.  But my goal is to make it as
close to Emacs in terms of keyboard friendliness.

> I came to appreciate Emacs more when I looked at THE
> http://humane.sf.net/the/ -- an editor made by Jef Raskin, former Mac
> interface designer.  The interface he describes, without realizing it,
> is Emacs, and THE is just a poor implementation of Emacs.  I point out
> the similarities in:
> https://sourceforge.net/forum/forum.php?thread_id=796533&forum_id=193382

Interesting read, thanks for the link.

> Now, I'm not saying Emacs couldn't be improved.  But it's a far, far
> better starting point that "modern" text editors.  I'm thoroughly
> convinced that modality is the superior interface for most situations,
> and I think GUIs are leaning more in that direction as time goes on.  I
> think this is why vi remains popular -- even though I personally don't
> like it -- because it's modal.  At least anything that's keyboard
> friendly should be modal, and a text editor that isn't keyboard friendly
> is clearly stupid.

I like Emacs plenty, and use it (and Gnus) for just about everything:
text, email, newsgroups, and IRC.  That's why I want to use it as a
model.

> But I'd still like a Python Emacs to make experimentation with UI
> possible.  It would also be neat if it could be an environment for
> application development, like Emacs is, like THE wants to be... Oberon
> was like this too, and while I didn't like many parts of Oberon, the
> overall environment was really neat.  Squeak keeps trying to move this
> way too.  I'd love to see an environment where there was less boundary
> between module, one-off program, and end-user application.  But ignore
> such high ideas, because they will distract and overwhelm :)  The editor
> is a great starting point for any number of things, but if you get to
> thinking about all those things it'll keep you from starting the
> foundation ;)

Okay, thanks.  Back to the grindstone...

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From max at alcyone.com  Tue Apr  1 17:24:09 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 01 Apr 2003 14:24:09 -0800
Subject: Questions on linked lists
References: <kl8ia.27315$OV.127379@rwcrnsc54> <5vonl-vm3.ln1@beastie.ix.netcom.com> <slrnb8ja7a.421.ramen@lackingtalent.com>
Message-ID: <3E8A1189.3F806B4B@alcyone.com>

Dave Benjamin wrote:

> That said, what would be an example of a legitimate need for a custom
> linked
> list data structure in Python? I can't say I've come across one yet.
> Do
> linked lists outperform native Python lists at certain operations?

Yes.  Python lists are (computer science) vectors, which means they're
dynamically resizeable arrays.  It's inexpensive to index into the list
[O(1)] and add to the list [amortized O(1)], but it's expensive to
insert or delete from the list [O(n)].  A linked list, although it's
expensive to index [O(n)], is very inexpensive to insert or delete
objects at _any_ location [O(1) for either].

If you find you need a data structure where you are doing a lot more
inserting and deleting than indexing, then a linked list might well make
a great deal more sense.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ A good indignation brings out all one's powers.
\__/ Ralph Waldo Emerson
    Esperanto reference / http://www.alcyone.com/max/lang/esperanto/
 An Esperanto reference for English speakers.



From tim_one at email.msn.com  Sat Apr 12 22:24:50 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sat, 12 Apr 2003 22:24:50 -0400
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <Q0Wla.4427$LB6.127243@news1.tin.it>
Message-ID: <LNBBLJKPBEHFEDALKOLCCENJEGAB.tim_one@email.msn.com>

[Alex Martelli]
> IMHO the most maddening use of __repr__ is that done by the
> implementation of __str__ for all built-in container types:
>
> >>> class X:
> ...   def __str__(self): return 'str'
> ...   def __repr__(self): return 'rep'
> ...
> >>> print [X()]
> [rep]
> >>>
>
> There's supposed to be a good reason for this to print [rep],
> rather than [str] as common sense would suggest, but I keep
> forgetting it (so it's probably nor very compelling;-).

It's the problem of what to do with contained strings.  If str(container)
applied str() to the containees, then, e.g.,

    print [2, "2", "[3, 4]"]

would display

    [2, 2, [3, 4]]

instead of today's

    [2, '2', '[3, 4]']

Many would find the former unbearably misleading, including Guido.  It would
likely be confusing whenever a string contained commas, colons, square
brackets, curly braces, or parentheses (the delimiter characters used by the
convert-to-string methods of the basic builtin containers).  OTOH, if
str(string) were changed to act like repr(string), then WYSIWYG would be
violated when printing strings directly.

So strings really seem to want three display gimmicks:  pump out the bytes
as-is to the display device (current str.__str__); produce a portable 7-bit
printable-ASCII encoding, with escape sequences as needed (current
str.__repr__); and pump out the bytes as-is, except with enough decoration
so that it can still be identified as a string when part of the display of a
container.  That last one is an ongoing sore point for people who can't use
7-bit ASCII to spell their name:  when their name appears in a container
(say, as a dict key), displaying the container converts the "funny
characters" in their name to backslash hex escapes.

OTOH, no type other than a string type seems to feel a pressing need for
more than two convert-to-string gimmicks, and the basic container types
don't even distinguish repr from str (str(list) and repr(list) act the same
today because they are the same today).

In any case, unless containers are to special-case the snot out of string
containees, their convert-to-string methods will apply str() or repr()
uniformly to all containees.  Because of the example above, they choose to
apply repr().  That problem with strings is really all there is to this
decision.





From sholden at holdenweb.com  Sat Apr 19 09:50:21 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sat, 19 Apr 2003 13:50:21 GMT
Subject: variable in Python
References: <b7lvbh$an3$1@news-reader13.wanadoo.fr> <mailman.1050586367.28668.python-list@python.org> <DD4oa.113758$0X.26571104@twister.columbus.rr.com> <mailman.1050730020.31828.python-list@python.org>
Message-ID: <xucoa.134489$It5.69547@news2.central.cox.net>

"Jp Calderone" <exarkun at intarweb.us> wrote in message
news:mailman.1050730020.31828.python-list at python.org...
> On Sat, Apr 19, 2003 at 04:53:55AM +0000, berg_ wrote:
> >
> > > object it refers to; additionally, there is no concept of a "null"
> > > reference.
> >
> > Python does have the concept of null, it is represented by None.
> >
> > cursor = None
> >
>
>   This is merely semantic quibbling, of course - but None is an object.
Is
> it often used like NULL?  Sure, but then, [] and () and '' are also often
> used like NULL (They all have a false truth value, after all).  The
concept
> of a variable that doesn't point to anything, though, doesn't exist in
> Python.  "cursor = None" binds cursor to the immutable singleton object
> "None".
>

Of course, there's always

    del cursor

but even that will fail if the name isn't currently bound.

and-now-back-to-the-angles-on-the-head-of-a-pin-debate-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From lucio at movilogic.com  Mon Apr  7 14:56:50 2003
From: lucio at movilogic.com (lucio)
Date: Mon, 07 Apr 2003 15:56:50 -0300
Subject: Debugging embedded python
In-Reply-To: <u4r5bgyfx.fsf@online.no>
References: <u4r5bgyfx.fsf@online.no>
Message-ID: <3E91C9F2.5090006@movilogic.com>

Syver Enstad wrote:

>I have the python interpreter embedded in a daemon/service (this is
>win2k so it's a service). I'd really like to be able to debug the
>python code that is running. I was thinking of popping up a console
>window to be able to run pdb, how does one go about it to accomplish
>this? Are there any code doing similar things or articles.
>
>  
>

Have you tried HAP?

http://hapdebugger.sourceforge.net/

Remote debbuger with extra cheese.

Lucio.





From aleax at aleax.it  Tue Apr 29 18:23:47 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 22:23:47 GMT
Subject: Another itertool function?
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <45pra.20060$xw4.19534@nwrdny01.gnilink.net> <C0ura.19455$3M4.563854@news1.tin.it> <mailman.1051650464.17483.python-list@python.org>
Message-ID: <TXCra.34117$K35.1091440@news2.tin.it>

Steven Taschuk wrote:

> Quoth Alex Martelli:
>> Raymond Hettinger wrote:
>>    ...
>> > Here's another idea that was submitted:
>> > 
>> > def windowed(iterable, group=2):
>> >     "returns (i[0], i[1], ... i[group-1]), (i[1], i[2], ... i[group]),
>> 
>> I think this would be perfect if the overlap between successively
>> returned windows was also a parameter, rather than being fixed to 1.
> 
> Alexander Schmolck proposed such a function back in February [1];
> his version included a "step" argument doing what you suggest.
> 
> In the discussion of whether such a function was useful enough to
> be included in itertools, I provided several use cases [2], but
> was unable to find a convincing one which needed a step other than
> one.  Do you have one in mind?

Simplest example: I have a flat sequence k1 v1 k2 v3 k3 v3 ... and I
want a mapping k1->v1, k2->v2 etc.  dict(windowed(x, 2, 2)) would do it.

Second example: if I have a sequence of daily earnings and I want a "moving 
daily average" -- for each day, the average of the seven days centered on
it -- imap(sum, windowed(earnings, 7)) would suffice [ok, sum is not the
same as avg, but dividing each number by 7 ain't hard if needed].  But say
I want a "moving weekly average" instead -- for each WEEK, the average
(or sum) of the three weeks centered on it.  Why not:
imap(sum, windowed(earnings, 21, 7)) -- it seems simplest. Or maybe faster:
imap(sum, windowed(imap(sum, windowed(earnings, 7, 7)), 3)).

In general, 1 and N are going to be the most popular 'steps' for a
window of length N, but other values may help for moving averages &c.


Alex





From abpillai at lycos.com  Thu Apr 17 10:14:05 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 17 Apr 2003 07:14:05 -0700
Subject: "pythonic"
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com> <mfuna.26072$LB6.616225@news1.tin.it>
Message-ID: <bd993a2f.0304170614.48c92a76@posting.google.com>

'Guido's thesaurus' gives 4 different 
meanings for this.

pythonic

  It is a term used by programmers who have found
solace in the ultimate. It also means the following
generally used terms, though not accepted with grudge
by followers of other programming faiths.

not perl
__easy__ and __elegant__
weakref.ref("C")
pickle.dump("C++")
pickle.dump("Java")

Anand Pillai
 
Alex Martelli <aleax at aleax.it> wrote in message news:<mfuna.26072$LB6.616225 at news1.tin.it>...
> Afanasiy wrote:
> 
> > Too many times now I have seen the term "pythonic" used to vaguely
> > describe something which is never explicitly defined, as if it cannot be.
> > 
> > Now it seems I can just replace "pythonic" with "my idea of utopia".
> > 
> > I do not know if this ignorance, laziness, or a meme, but I would like
> > to know if there a real definition of "pythonic". Anyone?
> 
> If you want a thesaurus:
> 
> http://www.bartleby.com/62/49/P1214900.html
> 
> """
> pythonic
> 
> ADJECTIVE:
> Of extraordinary size and power: behemoth, Brobdingnagian, Bunyanesque, 
> colossal, cyclopean, elephantine, enormous, gargantuan, giant, gigantesque, 
> gigantic, herculean, heroic, huge, immense, jumbo, mammoth, massive, massy, 
> mastodonic, mighty, monster, monstrous, monumental, mountainous, 
> prodigious, stupendous, titanic, tremendous, vast. Informal : walloping. 
> Slang : whopping. See BIG. 
> 
> Roget?s II: The New Thesaurus, Third Edition. Copyright ? 1995 by Houghton 
> Mifflin Company. Published by the Houghton Mifflin Company. All rights 
> reserved.
> """
> 
> 
> If you prefer a dictionary:
> 
> http://www.bartleby.com/61/25/P0692500.html
> 
> """
> ADJECTIVE:
> 1. Of, relating to, or resembling a python. 2. Of or resembling an oracle; 
> prophetic. 3. Of extraordinary size and power. 
> 
>    
> The American Heritage? Dictionary of the English Language, Fourth Edition. 
> Copyright ? 2000 by Houghton Mifflin Company. Published by the Houghton 
> Mifflin Company. All rights reserved.
> """
> 
> 
> Of course, around here we tend to use a nuance of "Of, relating to,
> or resembling Python" rather than "a python", and while "extraordinary
> power" is surely sometimes part of the connotations, I don't think
> the specific connotations towards BIG SIZE, which the thesaurus points
> out so well, are normally implied in our discourse.
> 
> 
> Alex



From grayrest at grayrest.com  Thu Apr 17 12:19:10 2003
From: grayrest at grayrest.com (grayrest)
Date: Thu, 17 Apr 2003 12:19:10 -0400
Subject: Random web page?
In-Reply-To: <mailman.1050587310.4664.python-list@python.org>
References: <b7m4ql$ril$1@news.hccnet.nl> <mailman.1050587310.4664.python-list@python.org>
Message-ID: <b7mkad$aur$1@news-int.gatech.edu>

Bryan L. Fordham wrote:
> So why not just use google instead?  Write something to generate a 
> random word or phrase, then submit that to google using the "I'm feeling 
> lucky" feature.

Because then the most linked to pages will show up more frequently than 
the less linked pages, which is the whole point of google and not what 
he asked.

-- 
grayrest
http://grayrest.com/moz




From amk at amk.ca  Wed Apr 16 09:25:37 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Wed, 16 Apr 2003 08:25:37 -0500
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <roy-92A049.08504216042003@reader1.panix.com>
Message-ID: <dGGdnXL6VMlMxACjXTWcoQ@speakeasy.net>

On Wed, 16 Apr 2003 08:50:42 -0400, 
	Roy Smith <roy at panix.com> wrote:
  <excellent discussion deleted...>
> algorithm tuning.  Imagine you've got a program which runs in O(N^2) 
> time.  If you could find a way to reduce it to O(N), for an input set of 
> just 10 items, you would have speeded it up by a factor of 10!  Compared

... except for one nit; the above inference isn't valid.  O() notation
usually ignores differing multipliers because they don't matter much as N
grows larger, but they still have an effect on the running time.

Consider two algorithms: A is O(N) where each loop takes one second, and B
is O(N**2) where each loop takes .001 second.  For 10 items, A takes 10
seconds and B takes .1 sec to run, so switching from A to B would actually
run slower.  The situation changes for N=1000, where A takes 1000 seconds
and so does B.  For N=2000, A takes 2000sec and B takes 4000sec, twice as
long.

Algorithms often achieve better O() performance at the cost of making the
loop more complicated or requiring some precomputing step, so for small
problem sizes the straightforward algorithm might be faster.  Better O()
performance buys you something at larger problem sizes, of course.

(BTW, you should turn your post into a web page and put it somewhere; it
might be useful for educators.)

> complexity and algorithmic complexity.  This, of course, is what makes 
> high level languages so useful in the first place, but I fear we may be 
> training a new generation of programmers for whom algorithmic complexity 
> is not as familiar a concept as it used to be.

I don't see that as being a problem; students can first learn the ideas of
big-O notation by counting the number of Python operations, ignoring the
wall-time performance of those operations. 

--amk                                                             (www.amk.ca)
Now I'll never know if I was right.
      -- Adric's last words, in "Earthshock"



From staschuk at telusplanet.net  Wed Apr 30 10:01:58 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 30 Apr 2003 08:01:58 -0600
Subject: Self modifying systems question
In-Reply-To: <DCB542A4-7AD6-11D7-93CD-003065DC4D52@as4-2-8.an.g.bonet.se>; from skvamme@as4-2-8.an.g.bonet.se on Wed, Apr 30, 2003 at 08:42:07AM +0200
References: <DCB542A4-7AD6-11D7-93CD-003065DC4D52@as4-2-8.an.g.bonet.se>
Message-ID: <20030430080158.C334@tibia.amotlpaa.bogus>

Quoth Sten Kvamme:
> Self modifying code is a well known technology. Has anyone done this in
> Python? What I'm seeking for is examples of software that adds to
> itself (methods and variables) in runtime. I'd like an instantiated 
> object to
> be able to add a method to itself.

Others have already shown examples of how Python's intrinsic
dynamism lets you add attributes (including methods) to objects at
runtime.

For a more alarming example, in which a class, at runtime,
rewrites its own source and gets recompiled, see Michele
Simionato's post
    <http://groups.google.com/groups?th=e6cf42cbc54a7420>
from 17 March.

-- 
Steven Taschuk                staschuk at telusplanet.net
"I tried to be pleasant and accommodating, but my head
 began to hurt from his banality."   -- _Seven_ (1996)




From a.schmolck at gmx.net  Tue Apr 15 18:20:31 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 15 Apr 2003 23:20:31 +0100
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <m2fzokfgni.fsf@mycroft.actrix.gen.nz> <slrnb9oajn.493.rmunn@localhost.localdomain>
Message-ID: <yfs3ckjcsi8.fsf@black132.ex.ac.uk>

Robin Munn <rmunn at pobox.com> writes:

> Paul Foley <see at below.invalid> wrote:
> > On Mon, 14 Apr 2003 16:50:40 GMT, Robin Munn wrote:
> > 
> >> What I really want to know is why you would consider sample #1 "broken
> >> and fragile code". I can't think of any circumstances where sample #1
> >> would fail to achieve its desired effect. Can you?
> > 
> > What if another thread (B) does the same thing, and changes sys.stdout
> > while this thread (A) is still inside the "try" block?  And if thread
> > A runs its "finally" clause and resets sys.stdout while thread B is
> > still in its "try" block, they screw each other.
> 
> This is the classic problem that arises when two threads need to use any
> shared resource, isn't it? I wasn't considering threads at all when I
> wrote those two code samples, so of course I didn't use locks to ensure
> good behavior.
> 
> Actually, now that I start thinking about it, any kind of
> save-and-restore mechanism would start to get pretty hairy around
> threads, wouldn't it? 

Yes and that's why dynamic variables *do* make a real difference: different
threads don't have to share the same global variable.

'as



From eugene_druker at yahoo.com  Fri Apr 25 16:16:33 2003
From: eugene_druker at yahoo.com (Eugene Druker)
Date: 25 Apr 2003 13:16:33 -0700
Subject: Hierarchy of types ?
Message-ID: <37cbc1de.0304251216.6068146f@posting.google.com>

Hi All,

I'd like to understand how to use types properly.
I want to find extremums of non-sequences. 

>>> min(0,[1,2,3],4)
0
>>> max(0,[1,2,3],4)
[1, 2, 3]

It does not look very naturally...
So, may be first compared are types:

>>> type(0) < type([1,2,3])
1

It's better - I can find extremums type-by-type. However:

>>> type(int) < type(list)
0

Any explanations ?
(Python 2.2.1 (#34, Apr  9 2002, 19:34:33) [MSC 32 bit (Intel)] on win32)
Thanks,
Eugene



From Patrick.Carabin at sciencesnaturelles.be  Tue Apr 15 10:14:27 2003
From: Patrick.Carabin at sciencesnaturelles.be (Patrick Carabin)
Date: Tue, 15 Apr 2003 16:14:27 +0200
Subject: files: direct access - NOT sequential
Message-ID: <03041516142702.01079@pc20_118>

Terry,
""" open in 'w' mode or (probably better for Windows, ?'wb') mode and 
use tell() and seek() methods. """

Thank You.

i found tell() and seek() as properties/mrthods of the ( opened ) 
file object...

But i still have a problem : open in 'w' mode  ( or 'wb' : what can 
be the difference? ) does ERASE the contents of the file, so 
everything is lost ...

1 more question :
how do i have to manage it to keep the contents that are already in 
the file ? Do You have some example?

TIA

Patrick Carabin.

Institut Royal des Sciences Naturelles de Belgique 
http://www.SciencesNaturelles.be/
Koninklijk Belgisch Instituut voor Natuurwetenschappen 
http://www.NatuurWetenschappen.be/ 

?Plant in de tuin van je hart slechts de roos der liefde?
?Dans le jardin de ton coeur, ne plante que la rose d' amour?
						Baha'u'llah.




From tim at bladerman.com  Fri Apr  4 09:02:12 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Fri, 04 Apr 2003 14:02:12 GMT
Subject: Interactive loop with InteractiveInterpreter
References: <Fffja.3972$Du.12917@newsc.telia.net>
Message-ID: <Efgja.3979$Du.12985@newsc.telia.net>

"Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote

> Is it possible to step through a loop at all? With interactive
interpreter?
> I can't seem to find an easy solution, I guess InteractiveConsole won't do
> it either. Is it maybe better to do it in a completley different way?

Hmm replying to myself here but I just figgured out a way that maybe is
good... or bad... ad an extra layer before I call

    InteractiveInterpreter.runcode()

If I am about to send a for statement or a while statement to the
interpreter I skip that line and set up my own loop with the parameters from
the the loopline and then feed each line to the interpreter several times,
handling the looping myself.

It seems like an awkward way but I guess it is possible.

Or am I maybe better of using the debuggermodule, I just remembered it and
havn't had a look at it yet, is this maybe what it is intended for?

Tim





From h_schneider at marketmix.com  Sun Apr 20 06:45:05 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Sun, 20 Apr 2003 12:45:05 +0200
Subject: SOLVED: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net> <b7mbnb$pp2$06$1@news.t-online.com> <8yu9gpnb.fsf@python.net> <b7mf9u$2jb$03$1@news.t-online.com> <b7mjlc$6tm$03$1@news.t-online.com> <b7p1mq$i2t$02$1@news.t-online.com> <7k9rh748.fsf@python.net>
Message-ID: <b7ttn9$ltf$05$1@news.t-online.com>

OOps .. your're right. Thank you!

--Harald


"Thomas Heller" <theller at python.net> schrieb im Newsbeitrag
news:7k9rh748.fsf at python.net...
> "Harald Schneider" <h_schneider at marketmix.com> writes:
>
> > Strange ... I changed the order of the import statements from
> >
> > import win32serviceutil,win32service,win32event,pywintypes
> > import os,string,rpcserver,xmlrpclib,confighandler
> > from agent import *
> > from setuptool import RegistryTool
> >
> > to
> >
> > from agent import *
> > from setuptool import RegistryTool
> > import win32serviceutil,win32service,win32event,pywintypes
> > import os,string,rpcserver,xmlrpclib,confighandler
> >
> > and not it works!
> >
> > The registry access is done in a class contained in the agent module.
Under
> > Python 2.1 the order played no role, under 2.2 it does!
> >
>
> I think I've seen this before, but I'm not sure.  IIRC, using 'from
> agent import this, that, whatever' instead of the 'from agent import *'
> may also fix the problem.
>
> Usually it is bad style anyway ;-).
>
> Thomas





From roy at panix.com  Tue Apr 15 08:51:50 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 15 Apr 2003 08:51:50 -0400
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org> <mailman.1050406764.904.python-list@python.org>
Message-ID: <roy-9568B2.08515015042003@reader1.panix.com>

Beni Cherniavsky <cben at techunix.technion.ac.il> wrote:
> Don't forget that strings are immutable and you need the old object to
> still represent the old string.  I see two approaches:
> 
> 1. Each string object points to a string buffer and has an end index
>    and multiple strings can share the same buffer.

I think this is a cool idea.  The problem I see is that it adds about 8 
bytes (2 32-bit indicies) to the size of a string object.  For 
applications that use lots of small strings, this becomes significant.  

> Ironically, this
>    works well for substrings but not for concatenation because of::
> 
>      s = "some long string"
>      print s + "a"
>      print s + "b"
> 
>    (you can only grow the initial string for one of them).  This might
>    not be a problem because this usage is rare.

I think you get around that completely by defining a new string method 
for this.  Instead of s1 += s2, you do s1.append(s2).  The + and += 
operators give you the existing behavior.

I wonder if the right solution to that is to have two different kinds of 
strings, perhaps MutableStringType being a subset of StringType.  You 
get a mutable string by doing s = "\mThis is a mutable string"?  Mutable 
strings would inherit all of the methods of regular strings, plus they 
would have append() and reserve(), and maybe a few others.  A mutable 
string's __str__() method would return an immutable copy.



From mcclaine at comcast.net  Mon Apr 14 15:29:45 2003
From: mcclaine at comcast.net (Em)
Date: Mon, 14 Apr 2003 15:29:45 -0400
Subject: inserting Python commands into HTML documents
In-Reply-To: <zpqdne0Kztcb0AejXTWcow@comcast.com>
References: <zpqdne0Kztcb0AejXTWcow@comcast.com>
Message-ID: <6-CdnUrJoLKFkQajXTWcqA@comcast.com>

These options all seem to require me to install 3rd party software onto 
the server of this company, and I believe that to be outside of the 
scope of the test project.

The only issue is that I want it to be invisible that I'm pulling from a 
CGI script (I don't want the URL bar to say 
"www.mywebsite.com/cgi-bin/myProj.py?page=projects", I want it to simply 
say "www.mywebsite.com/projects.html" or "www.mywebsite.com/projects.php").

Is there a way for me to send my output to /index.html, /projects.html, 
etc.?

Em wrote:
> Hello,
> 
> I am an extremely new member of the Python family. I had a job interview 
>  where they charged me with the task of creating a (fake) corporate 
> website using Python in ten days--with two left now.
> 
> The problem is that what I have read seems to say Python is a CGI 
> language... Am I remiss in believing that my website is going to be 
> completely run from inside a couple .PY files in the CGI-BIN?
> 
> Is there a way for it to act more like PHP, as in run Python commands 
> from inside the (index|contacts|about|etc).html or at least simulate 
> that behavior? (This is on a linux server and I am not to use PHP, FYI.)
> 
> 
> Thank you all for any assistance!
> 




From max at alcyone.com  Thu Apr 10 20:18:43 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 10 Apr 2003 17:18:43 -0700
Subject: Can I define "static" (like in C++) methods in Python?
References: <169a7283.0304101430.2bec922b@posting.google.com>
Message-ID: <3E9609E3.D20746BC@alcyone.com>

sdieselil wrote:

> Can I define methods which are attributes of class and not attributes
> of
> instance?

>>> class C:
...  X = 1 # X is a class attribute
...  def __init__(self, y):
...   self.y = y # y is an instance attribute
... 
>>> C.X
1
>>> c = C(10)
>>> c.X
1
>>> c.y
10
>>> C.X = 2
>>> c.X
2
>>> c.y 
10

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ We're here to preserve democracy, not to practice it.
\__/ Capt. Frank Rasmey
    REALpolitik / http://www.realpolitik.com/
 Get your own customized newsfeed online in realtime ... for free!



From the_3_project at yahoo.com  Wed Apr 30 02:06:23 2003
From: the_3_project at yahoo.com (Joe Green)
Date: 29 Apr 2003 23:06:23 -0700
Subject: simplest question on win32com.client
References: <a5b86892.0304290534.31528b9e@posting.google.com> <CtDra.7016$lD4.74951@news-server.bigpond.net.au>
Message-ID: <a5b86892.0304292206.615d0d1@posting.google.com>

Mark Hammond <mhammond at skippinet.com.au> wrote in message news:<CtDra.7016$lD4.74951 at news-server.bigpond.net.au>...
> 
> I've seen this happen when an old state excel process is hanging around. 
>   Check using the Task Manager that there is no Excel currently running.
> 

Nope. No Excel running. This is on a Win2kpro machine with
ms-office2k.

In other unrelated questions, would it be possible to starta COM
object without it being visible, write, save to a file, close it?

Thanks



From f.steinhaeusler at amatic.com  Tue Apr  1 01:02:31 2003
From: f.steinhaeusler at amatic.com (Franz Steinhäusler)
Date: Tue, 01 Apr 2003 08:02:31 +0200
Subject: Text Editor with wxPython
References: <3e7c561a.8718311@news.utanet.at> <af8b8v0lol2vqvqr09ag5aalbiid70huja@4ax.com> <bd993a2f.0303310456.2f201c43@posting.google.com>
Message-ID: <9oai8vshpth0hvll13kbv85q0kefrpti0s@4ax.com>

On 31 Mar 2003 04:56:56 -0800, abpillai at lycos.com (Anand B Pillai)
wrote:

>I am thinking of doing something similar using wxWindows
>(C++ port). I have not started doing anything on it, but
>you can view my post on wxWindows mailing list.
>[...]

Hello Anand,

thank you for your answer.

It is only an idea, I'm still a "Python newbie",
but I will see.
It seems, Scintilla and using wxStyledTextCtrl will be the best
approach.

best regards,


-- 
Franz Steinhaeusler



From theller at python.net  Wed Apr  2 04:23:57 2003
From: theller at python.net (Thomas Heller)
Date: 02 Apr 2003 11:23:57 +0200
Subject: PyObject_CallFunction with keywords?
References: <mailman.1049251384.31007.python-list@python.org> <7876a8ea.0304012330.63f405b9@posting.google.com>
Message-ID: <65pxntg2.fsf@python.net>

woodsplitter at rocketmail.com (David Rushby) writes:

> Richard Cook <rcook at llnl.gov> wrote in message news:<mailman.1049251384.31007.python-list at python.org>...
> > Long question:
> > I am using an embedded Python interpreter in our C program to execute 
> > a method from a Python module (pexpect, to be exact).  At a certain 
> > point, I wish to do the equivalent of the following Python code in my 
> > C code:
> > 
> > result =  obj.expect("hello", timeout=50);
> > 
> > So I have a PyObject *f which points to obj.expect, and a PyObject *s 
> > which is a list containing the string "hello".  I figure I have to do:
> > PyObject_CallObject(f, s, k)
> > where k is the timeout parameter, but this does not match the param 
> > list for PyObject_CallObject().
> > 
> > My question is, how can I pass keyword parameters such as the timeout 
> > parameter in my example to a function which accepts them?
> 
> With Python 2.2 and later, there's PyObject_Call(callable, args,
> kwargs), but strangely, it seems to have been omitted from section
> "6.1 - Object Protocol" of the standard docs.

Even when programming for Python 2.2, you should at least sometimes peak
in the devel version of the docs:

http://starship.python.net/crew/theller/pyhelp.cgi?keyword=PyObject_Call&version=devel

Thomas



From mwh at python.net  Mon Apr 28 11:22:30 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 28 Apr 2003 15:22:30 GMT
Subject: Forgetting "()" when calling methods
References: <mailman.1051540303.31725.python-list@python.org>
Message-ID: <7h3bryqk5pf.fsf@pc150.maths.bris.ac.uk>

Tim Peters <tim.one at comcast.net> writes:

> [comparison stuff]
>
> No, the current state makes no sense.  In very early Pythons, it wasn't
> *possible* (for technical reasons) for a comparison to raise an exception
> (well, it could, but it got silently ignored).  When it became possible for
> comparisons to raise exceptions, new types started doing so, but
> pre-existing types generally didn't.

There was also the issue that before rich comparisons there was no way
for a type to know whether it was supposed to be testing equality or
order... having 2 != "42" raise an exception *would* cause anguish.
The time machine broke badly on this one.

Cheers,
M.

-- 
93. When someone says "I want a programming language in which I
    need only say what I wish done," give him a lollipop.
  -- Alan Perlis, http://www.cs.yale.edu/homes/perlis-alan/quotes.html



From rlratzel at siliconmetrics.com  Wed Apr  9 03:19:55 2003
From: rlratzel at siliconmetrics.com (Rick Ratzel)
Date: Wed, 09 Apr 2003 07:19:55 GMT
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <3E93CA4C.6070409@siliconmetrics.com>

Jim Shady wrote:

> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string. For
> example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?

    I'm sure i'll be one of many to help, but FWIW:

#
# returns the longest substring in inString, as delimited by "/"
#
def longest_str( inString ) :

     currentLen = 0
     maxLen = 0
     biggestString = ""

     #
     # check each substring, save the biggest one
     #
     for subString in inString.split( "/" ) :

         currentLen = len( subString )

         if( currentLen > maxLen ) :
             biggestString = subString
             maxLen = currentLen

     return biggestString


#
# test
#
for ts in ["abcd/df/a/iiwk/abcdefghijkl/b/c",
             "",
             "//",
             "abcd/df4444444444444444//////iiwk/abcdefghijkl/b/c"] :

     print "BIGGEST SUBSTRING IN: %s IS: %s" % (ts, longest_str( ts ))




From syver-en+usenet at online.no  Tue Apr  8 10:12:04 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 08 Apr 2003 16:12:04 +0200
Subject: Pythonwin COM and GDI
References: <3e926fb4$1@news>
Message-ID: <u7ka5gjt7.fsf@online.no>

Tim Hoffman <timh at zute.net> writes:

> Hi
> 
> I have gone through the activestate docs, I must be thick,
> cause I can't quite fathom the difference. 

gdi32.dll is not a COM dll so you can't use makepy on it. See below
for what you can do.

> Also it seems some of the
> GDI
> 
> stuff is in the win32ui module but unfortunately not the things I need
> 
> like CreateEnhWMF.

If functions are not present in win32gui or win32ui (win32ui is MFC
based), install ctypes and use this module to get what you want, or
you could try using pyrex.
 

-- 

Vennlig hilsen 

Syver Enstad



From exarkun at intarweb.us  Mon Apr 14 13:47:16 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 14 Apr 2003 13:47:16 -0400
Subject: an ugly file-reading pattern
In-Reply-To: <3e9ae78f$1@news.012.net.il>
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050201322.13220.python-list@python.org> <slrnb9h7u4.it3.rmunn@localhost.localdomain> <3e9a857b$1@news.012.net.il> <slrnb9ksaa.1q1.rmunn@localhost.localdomain> <3e9ae78f$1@news.012.net.il>
Message-ID: <20030414174716.GA25237@meson.dyndns.org>

On Mon, Apr 14, 2003 at 06:53:32PM +0200, Avner Ben wrote:
> 
> "Robin Munn" <rmunn at pobox.com> wrote in message
>
> > You're right, I didn't close the file explicitly. But explicit is better
> > than implicit, so IMHO it's better to give a name to the file object so
> > that you can close it explicitly:
> >
> >     input_file = file("somefile")
> >     for line in input_file:
> >         process(line)
> >     input_file.close()
> 
> 
>     The C# language has a nice construct: using (variable = whatever) { /*
> code */ }. The variable in the using header is guaranteed to be
> garbage-collected at the end of the block (or so I think). Could be useful
> here.
> 

  In CPython, the reference count to the file object in:

    for line in file(...):
        ...

  will drop to 0 immediately after the loop, and the file will be closed at
that point.

  Jp

-- 
In the days when Sussman was a novice Minsky once came to him as he sat
hacking at the PDP-6. "What are you doing?" asked Minsky. "I am training a
randomly wired neural net to play Tic-Tac-Toe." "Why is the net wired
randomly?" asked Minsky. "I do not want it to have any preconceptions of how
to play." Minsky shut his eyes. "Why do you close your eyes?" Sussman asked
his teacher. "So the room will be empty." At that moment, Sussman was
enlightened.
 -- 
 up 25 days, 13:02, 2 users, load average: 2.04, 2.02, 2.00




From JoeyTaj at netzero.com  Wed Apr 30 09:46:24 2003
From: JoeyTaj at netzero.com (Paradox)
Date: 30 Apr 2003 06:46:24 -0700
Subject: Calling dotNet components from Python.
References: <924a9f9c.0304282138.15aa7635@posting.google.com> <LqDra.7014$lD4.74951@news-server.bigpond.net.au>
Message-ID: <924a9f9c.0304300546.2cb5617d@posting.google.com>

Thank you all for your replies. I was hoping for something better than
just putting a COM wrapper on my components. I would like to be able
to script my notNet components without having to take an extra step
every time I change something. I will give it a try though and see how
tedious it is. Also this KOBRA is not exactly what I was talking about
but it does look useful does anyone know how I could get a copy? All
the links I've found lead to the same broken website.

                                         Thanks Joey.



From lacuidadanadelmundo at yahoo.com  Mon Apr 28 14:31:24 2003
From: lacuidadanadelmundo at yahoo.com (La Uruguaya)
Date: 28 Apr 2003 11:31:24 -0700
Subject: Fix for datetime conversion to integer and back to string; was 1 hour off
Message-ID: <4f705b08.0304281031.ddd8f75@posting.google.com>

I searched this group for a solution to a problem I had converting
dates from a string format to integer (in seconds since 1970) and back
to strings.  During some times of the year, the conversion didn't do
a perfect round trip and I ended up with a datetime that differed from
the original datetime by one hour.  I found 2 postings with variations
on the problem, but not a solution.  I implemented the following
solution and wanted to share it in case somebody needs it later.


import time

DATE_FORMAT = "%Y/%m/%d %H:%M:%S"

def date2int(strdate):
    result = 0
    try:
       tuple = time.strptime(strdate, DATE_FORMAT)
       result = time.mktime(time.strptime(strdate, DATE_FORMAT))
    except:
       pass
    return (result, tuple)

def correctForDaytime(tup):
    lst = list(tup)
    lst[3] -= tup[-1] # if this is zero, then this is a no op
    lst[-1] = 0
    return tuple(lst)

def int2date(intdate):
    result = ''
    if intdate > 0:
       try:
          tuple = time.localtime(intdate)
          tuple = correctForDaytime(tuple) # commenting this line out
will show you the bug
          result = time.strftime(DATE_FORMAT, tuple)
       except:
          pass
    return (result, tuple)

def f(t1str):
    (t1int, tup1) = date2int(t1str)
    (t1str2, tup2) = int2date(t1int)
    print t1str, "->", t1int
    print t1int, "->", t1str2
    if tup1 == tup2:
       print "SAME", tup1, tup2
    else:
       print "DIFF", tup1, tup2
    if t1str != t1str2:
       print "FAILED", t1str, t1str2 
    else:
       print "OK"

f("2003/01/15 11:15:00")
f("2003/02/15 11:15:00")
f("2003/03/15 11:15:00")
f("2003/04/15 11:15:00")
f("2003/05/15 11:15:00")
f("2003/06/15 11:15:00")
f("2003/07/15 11:15:00")
f("2003/08/15 11:15:00")
f("2003/09/15 11:15:00")
f("2003/10/15 11:15:00")
f("2003/11/15 11:15:00")
f("2003/12/15 11:15:00")

f("2003/01/03 11:15:00")
f("2003/02/03 11:15:00")
f("2003/03/03 11:15:00")
f("2003/04/03 11:15:00")
f("2003/05/03 11:15:00")
f("2003/06/03 11:15:00")
f("2003/07/03 11:15:00")
f("2003/08/03 11:15:00")
f("2003/09/03 11:15:00")
f("2003/10/03 11:15:00")
f("2003/11/03 11:15:00")
f("2003/12/03 11:15:00")

f("2003/01/27 11:15:00")
f("2003/02/27 11:15:00")
f("2003/03/27 11:15:00")
f("2003/04/27 11:15:00")
f("2003/05/27 11:15:00")
f("2003/06/27 11:15:00")
f("2003/07/27 11:15:00")
f("2003/08/27 11:15:00")
f("2003/09/27 11:15:00")
f("2003/10/27 11:15:00")
f("2003/11/27 11:15:00")
f("2003/12/27 11:15:00")



From max at alcyone.com  Sun Apr 27 00:14:26 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sat, 26 Apr 2003 21:14:26 -0700
Subject: Class problem: loosing my mind
References: <v0fmavkm9iknef7hit5n17d4m0qi6tpnia@4ax.com> <mailman.1051411241.22289.python-list@python.org> <diimavsd60audlnjm80m6dcan0e8b67hae@4ax.com>
Message-ID: <3EAB5922.A4F8DDA5@alcyone.com>

"" wrote:

> Seems pretty wierd that:
> 
> import gV
> 
> fails while:
> 
> from gV import gV
> 
> works.
	...
> P.S. Any of you 'language lawyers' feel free to enlighten me as to WHY
> this works and the other doesn't, I'm all ears!

It doesn't _fail_, it just didn't do what you thought it did.  The
confusion here is because you're not seeing the difference between
packages (a directory containing a __init__.py file, and some modules
and perhaps even other packages) and modules (things which contain
Python things in them).  You want the module, but you're asking for the
package.  The confusion arises is because you've named the module and
the package containing the one module the same thing.  There's no need
for them to be the same thing.  If ... is in PYTHONPATH and you have a
directory structure like this:

	.../package/
	.../package/__init__.py
	.../package/module.py

and .../package/module.py contains an object x, then I need to access x
by importing the _module_, not the package.  So I need to write:

	import package.module; package.module.x
	from package import module; module.x
	from package.module import x; x

or something along those lines.  Yours is just a generate case where x
is in a module named blah, which is itself in a package named blah, so 

	import blah

imports the package named blah, not the module, which is what you really
wanted.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ But tell me, who _are_ they, these wanderers ... ?
\__/ Rainer Maria Rilke
    ZOE / http://www.alcyone.com/pyos/zoe/
 A simple Python OpenGL rendering engine.



From gh at ghaering.de  Wed Apr 30 09:51:05 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 30 Apr 2003 15:51:05 +0200
Subject: Can urllib be used with nonblocking sockets?
In-Reply-To: <877k9c2nq0.fsf@r2d2.rfc1149.org>
References: <877k9c2nq0.fsf@r2d2.rfc1149.org>
Message-ID: <3EAFD4C9.9020401@ghaering.de>

Arne Schwabe wrote:
> Hi,
> 
> I have written a little python program that uses non-blocking sockets
> with select and I now want to urrlib in my program and I cannot find a
> method in urllib that would allow me to tell urllib that it should use
> non blocking sockets.
> 
> Is there a easy solution to this? One solution would to do implement
> something like a non blocking urllib but I don't want to do that.

<insert standard Twisted response here>

I haven't used these myself yet, but

http://www.twistedmatrix.com/documents/TwistedDocs/current/api/public/twisted.web.client.HTTPDownloader.html

seems to do what you want for HTTP and

http://www.twistedmatrix.com/documents/TwistedDocs/current/api/public/twisted.protocols.http.HTTPClient.html

is a more low-level interface (like httplib in the std. lib. is).

-- Gerhard




From jjl at pobox.com  Tue Apr 22 15:18:53 2003
From: jjl at pobox.com (John J. Lee)
Date: 22 Apr 2003 20:18:53 +0100
Subject: Event about a selection
References: <3EA57520.3F39B8D5@emi.u-bordeaux.fr>
Message-ID: <87ptne49ya.fsf@pobox.com>

thomas sabathe <sabathe at emi.u-bordeaux.fr> writes:

> In a text, a selection appears when a "double-clic" is done.
> But I don't know how to get this selection back. How can I do ?
> 
> For example:
> 
> def modif(event):
>     index = myText.tag_ranges(SEL)
>     myText.tag_add('valid', index[0], index[1])
> 
> myText.bind('<ButtonRelease-1>',modif)
[...]

Give us a clue, Thomas: are you using a GUI toolkit here?  If so, say
which one you're using in the subject line.


John



From eichin at metacarta.com  Mon Apr 28 22:21:59 2003
From: eichin at metacarta.com (eichin at metacarta.com)
Date: 28 Apr 2003 22:21:59 -0400
Subject: POD vs. reST for standalone writing?
Message-ID: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>

So, after a recent bad experience with LyX and version skew, I was
looking at alternatives for software end-user and administrator
documentation.  The constraints are pretty simple:

* produce good html and PDF
  ** via linux tools, and highly preferably free ones
* be editable on multiple platforms
  ** linux first, but also, osx, windows
* support "literal computer output" and screen shots

DocBook was under consideration (and was used for some man pages and
standalone docs) but has a high learning curve, and none of the
windows dtd-aware or docbook-aware editing tools actually seemed to
*work* well enough to hand them to non-techies (the techies, on linux,
are already happy with emacs and vi :)  Also, the subject of the
documentation changes rapidly enough that having document specialists
just isn't acceptable, everyone involved needs to be able to update
things.

Anyway, POD came up as a documentation format in a related context (in
particular, "I'm going on vacation, here's a walk through of how that
code works, don't call me" :) and it was very effective for output
that was dominated by verbatim code snippets interspersed with text;
in fact *easier* to write in than raw HTML would have been, though
probably no more expressive.  PDF output was not unreasonable (via
html or more directly via latex, so I haven't had to dig further.)

Now I start seeing mention of reStructuredText (especially in comments
about pyCon.)  reST files look very "ascii art" on the input side, but
after a little experiment, that seems to make them *harder* to write,
if somewhat (and I think only a little) easier to read. 

Am I missing something?  Is there actually a reST clear short hand,
and a tool that expands it to full-blown ascii-art?  Or is at least
fuzzy in the counting of dashes and other punctuation for marking
different levels?

(POD is perhaps a little too simple, but a first cut POD Parser class
in python, with a derived pod2html was the work of a recent
cross-country plane flight, so that's not all bad... and I'm not sure
that a pod2reST would be that hard either...)




From staschuk at telusplanet.net  Wed Apr 30 12:27:34 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 30 Apr 2003 10:27:34 -0600
Subject: Raising objects
In-Reply-To: <OFLra.36816$K35.1143840@news2.tin.it>; from aleax@aleax.it on Wed, Apr 30, 2003 at 08:18:54AM +0000
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it> <mailman.1051678851.21661.python-list@python.org> <OFLra.36816$K35.1143840@news2.tin.it>
Message-ID: <20030430102734.A1201@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
> Steven Taschuk wrote:
  [inability to raise new-style objects]
> > I don't have a real use case either, but it does seem rather
> > arbitrary.  Is there an implementation reason for this restriction?
> 
> In a sense, yes: there is no way to check if something "is a new-style
> class" as opposed to "any type whatsoever" -- so one could end up
> raising lists, tuples, strings (shades of the past...), whatever.

How about this?

    >>> def newstyle(obj):
    ...     """Whether obj is an instance of a new-style class."""
    ...     return hasattr(obj, '__class__') \
    ...            and issubclass(obj.__class__, object)
    ...
    >>> class OldStyle: pass
    ... 
    >>> class NewStyle(object): pass
    ... 
    >>> newstyle(OldStyle())
    0
    >>> newstyle(NewStyle())
    1
    >>> newstyle(Exception())
    0

This function can even be used directly on the classes:

    >>> newstyle(OldStyle)
    0
    >>> newstyle(NewStyle)
    1
    >>> newstyle(Exception)
    0

(Since the type 'type' is new-style but the type 'class' is not.)
It would produce counterintuitive results in such use if it were
possible for a new-style class object to be an instance of an
old-style class...  is that possible?

This function does report that strings and lists are new-style.
Aren't they?

> What I hope to see happen eventually is a spec that raise can use:
> 
>   -- any subclass of Exception (by that time a new-style class), OR
>   -- for backwards compatibility any old-style class, OR
>   -- for VERY backwards compatibility, any string
> 
> [and also INSTANCES as well as classes for the first two cases].

That seems reasonable.

But is there a problem with being able to raise anything?  Perhaps
some class of easy-to-produce bugs that I'm not seeing?

What I'm imagining at the moment is this: suppose you can raise
anything, and except clauses identify the exceptions to be caught
by their class.  The only problem I can see with this is that it
clashes with the present (backwards-compatible) treatment of
raised strings:
    try:
        raise 'foo'
    except 'foo':
        print 'wah'
wouldn't print 'wah', since 'foo' is not an instance of 'foo'.
You'd have to do something like
    try:
        raise 'foo'
    except str, e:
        if e == 'foo':
            print 'wah'
        else:
            raise
which is ugly.

So, suppose that strings were special-cased for backwards
compatibility:

    def exceptmatch(exception, classtomatch):
        """Whether an except clause specifying classtomatch should
        catch exception.
        """
        if isinstance(classtomatch, str):
            return exception == classtomatch
        else:
            return isinstance(exception, classtomatch)

(I ignore for simplicity except clauses which match more than one
class.)

Is there a problem with this scheme?

> Exactly: even in my hoped-for new spec, you'd have to wrap this
> in a subclass of Exception.  Facilitating the [ab]use of the
> exception mechanism for generalized control flow is, I suspect,
> NOT on the list of priorities in the BDFL's mind;-).  Feel free
> to open a PEP to settle that beyond any doubt, though...;-)

Heh.  I imagine you're right.  I'd argue in favour of a scheme
such as above chiefly on the basis of regularity (rather than for
making my silly example work).

I just don't see a good reason not to be able to raise new-style
objects, so to me it's an arbitrary restriction and as such makes
the language slightly uglier.  (It is, of course, a very minor
issue; I hadn't even noticed that Exception is old-style until
this thread.)

-- 
Steven Taschuk                               staschuk at telusplanet.net
"[T]rue greatness is when your name is like ampere, watt, and fourier
 -- when it's spelled with a lower case letter."      -- R.W. Hamming




From rganesan at myrealbox.com  Tue Apr 15 00:29:53 2003
From: rganesan at myrealbox.com (Ganesan R)
Date: 15 Apr 2003 09:59:53 +0530
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it>
Message-ID: <ouadesfkn2.fsf@andlx-anamika.cisco.com>

>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:

> Note that any Python function call always passes VALUES -- i.e.
> there is no difference between your call to sys.getrefcount in
> the above snippet and one that directly does sys.getrefcount(1).
> In either case, it's the VALUE (1) that you're asking info about;
> it makes no difference HOW you obtain the value to pass (I guess
> this is probably already clear to you, but, just in case...).

Then, I find this confusing

===========================================================================
Python 2.2.1 (#1, Jul 29 2002, 23:15:49) 
[GCC 2.95.4 20011002 (Debian prerelease)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import gc
>>> a = "hello world"
>>> gc.get_referrers(a)
[{'__builtins__': <module '__builtin__' (built-in)>, '__name__': '__main__', 'gc': <module 'gc' (built-in)>, '__doc__': None, 'a': 'hello world'}]
>>> b = "hello world"
>>> gc.get_referrers(a)
[{'a': 'hello world', 'b': 'hello world', 'gc': <module 'gc' (built-in)>, '__builtins__': <module '__builtin__' (built-in)>, '__name__': '__main__', '__doc__': None}]
>>> gc.get_referrers("hello world")
[('hello world', None)]
>>> 
===========================================================================

Why am I getting a different list for the last call? 

Ganesan

-- 
Ganesan R




From gh at ghaering.de  Thu Apr 17 15:59:54 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Thu, 17 Apr 2003 21:59:54 +0200
Subject: Downloading TV listings with urllib
In-Reply-To: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4>
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4>
Message-ID: <3E9F07BA.1070800@ghaering.de>

Josh,

there seems to be a problem with your email account that you might want 
to investigate:

Reporting-MTA: dns; mail.ghaering.de
Arrival-Date: Thu, 17 Apr 2003 21:47:39 +0200 (CEST)

Final-Recipient: rfc822; jslm at earthlink.net
Action: failed
Status: 5.0.0
Diagnostic-Code: X-Postfix; host mx03.earthlink.net[207.217.120.78] 
said: 550
     jslm at earthlink.net...User unknown


-- Gerhard




From theller at python.net  Wed Apr 16 15:18:38 2003
From: theller at python.net (Thomas Heller)
Date: 16 Apr 2003 21:18:38 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <v9qot54nmk87cb@news.supernews.com> <u1cyqvfi.fsf@python.net> <9Wfna.34058$T34.892831@news2.tin.it>
Message-ID: <k7dup7xt.fsf@python.net>

Alex Martelli <aleax at aleax.it> writes:

> Thomas Heller wrote:
>    ...
> > Hm, the latest version of the patch which I posted typically prints this:
> > 
> > ======================================================================
> > FAIL: test_failUnlessEqual (__main__.FailingTests)
> > ----------------------------------------------------------------------
> > TestFailed: 0 != 1
> >   File "xunit.py", line 12, in test_failUnlessEqual
> >     self.failUnlessEqual(self.a, self.b)
> > 
> > Is this poorly formatted? I won't argue about this.
> 
> It's sure WAY better than what it does out of the box!!!
> 
> > Anyway, this is IMO the information I want to have:
> > filename, line number, source code line, and the actual#
> > values which made the test fail.
> 
> Seconded.
> 
> > Concerning the 'more is better': If the test crashes (raises an
> > unexpected exception), the full traceback is printed.
> > If the test fails because a condition is not fulfilled, I'm *not*
> > interested to see how unittest raises this error.
> 
> Absolutely!!!  I think unittest would be MUCH better with
> your mods.

Given your and the other positive responses, and ignoring the negative
responses ;-), I have posted it to the Python patch manager:

http://www.python.org/sf/722638

Thomas



From adalke at mindspring.com  Wed Apr 23 22:40:49 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 23 Apr 2003 20:40:49 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
References: <mailman.1051023275.29338.python-list@python.org> <b8420j$629$1@slb3.atl.mindspring.net> <mailman.1051043856.10118.python-list@python.org> <b86gim$60r$1@slb5.atl.mindspring.net> <mailman.1051127503.25664.python-list@python.org>
Message-ID: <b87il2$3rf$1@slb6.atl.mindspring.net>

Steven Taschuk:
> I do not (and did not) endorse this notion for __and__.  I simply
> do not agree with the previous claim that there's no way to
> implement overloading and short-circuiting for the same operator.
>
> No *good* way given how Python works now, perhaps.

Ahhh, a nit-picker, eh?  :)

Let's see.  I said:
] There's no way to implement that behaviour in Python, so
] it cannot be overridden.

I submit that a language which worked in your suggested manner
would not be Python.

The nit is now in your court!

 :)

                    Andrew
                    dalke at dalkescientific.com





From staschuk at telusplanet.net  Tue Apr 15 12:46:11 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 15 Apr 2003 10:46:11 -0600
Subject: hex -> 16bit signed int (newbie)
In-Reply-To: <TTRma.1502$KH1.321574@news20.bellglobal.com>; from zf_acs@DELETETHIShotmail.com on Tue, Apr 15, 2003 at 07:18:36AM -0400
References: <TTRma.1502$KH1.321574@news20.bellglobal.com>
Message-ID: <20030415104611.A1863@tibia.amotlpaa.bogus>

Quoth zif:
> How can I get 16bit signed int from a hex string?
> Eval('0x...') gives me 16bit unsigned,

... which you could easily convert to signed.

    def hex2uint16(s):
        # assumed s is up to four hex digits
        i = eval('0x' + s)
        if i >= 2**15:
            i -= 2**16
        return i

> struct.pack('i', '\x..\x..') works fine, but
> Python doesn't allow me to concatenate strings
> with '\' (str(92) gives me '\\', which doesn't
> work as well).

To be clear: concatenating strings containing backslashes is fine.
What you want, however, is to construct on the fly a string
containing backslash-escapes, and have those escapes interpreted
the same way they would be in source.  That calls for an eval.

    >>> hexstr = 'AB0E'
    >>> s = ''
    >>> for i in range(0, len(hexstr), 2):
    ...     s += '\\x%s' % hexstr[i:i+2]
    ... 
    >>> s
    '\\xAB\\x0E'
    >>> eval('"' + s + '"')
    '\xab\x0e'

But there's an easier way:

    >>> hexstr = 'AB0E'
    >>> import binascii
    >>> binascii.unhexlify(hexstr)
    '\xab\x0e'

Whereupon you can do your struct.unpack trick if you want.
(Though you should be aware that the meaning of 'i' is
machine-dependent; on my machine it's 32 bits, for example.  If
you want 16-bit unsigned on all machines, you want '>h' or some
such.  See the struct module's documentation.)

-- 
Steven Taschuk                                7\ 7'Z {&~         .
staschuk at telusplanet.net                        Y r          --/hG-
                                            (__/ )_             1^1`




From gmuller at worldonline.nl  Tue Apr 29 15:15:30 2003
From: gmuller at worldonline.nl (GerritM)
Date: Tue, 29 Apr 2003 21:15:30 +0200
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org>      <1bf4033f.0304290627.40f7d831@posting.google.com>      <uptn56xxi.fsf@fitlinxx.com> <mailman.1051638646.15981.python-list@python.org> <3EAEC78E.9D51D668@easystreet.com>
Message-ID: <b8mj06$h2l$1@reader1.tiscali.nl>

<achrist at easystreet.com> schreef in bericht
news:3EAEC78E.9D51D668 at easystreet.com...
> When attempting to introduce a new programming language into a
> corporate shop, you've got to look at it from the manager's
> side.
>
Yes, very true. I have written a how-to for architects:
"How to present architecture issues to higher management"
http://www.extra.research.philips.com/natlab/sysarch/ArchitectManagementInte
ractionPaper.pdf

> 1. There is a tremendous potential downside to having any language
> in your shop if you have fewer than 3 fluent developers.
>
> 2. If you introduce a new language, you hope that it will succeed and
> be used in your shop for 5-10 years.
>
> 3. Multiply (1) times (2) and you see that the decision to introduce
> a new language is well into a seven-figure $US investment.
>

This is of course only true from the ill-informed fearful manager.

<...snip..>
>
> 6. The way to counter all of this in favor of Python is to make the
> case that fewer lines of code result in much lower maintenance cost,
> so that those who look beyond the short-term see benefits and not
> costs.  One thing that might be interesting to try would be to take
> a VB application from 1995 or so (16-bit Win 3.1) and see how easy it
> is to get it running under the current version of VB, ie VB.Net.  Then
> take a python version 1.4 app and bring it up to verson 2.3.  I believe
> that python will look pretty good in this comparison.
>
Yes, working by example, supported by hard facts (numbers!) preferably
transformable in MegaDollars is an effective way.

> Al

regards Gerrit

--
www.extra.research.philips.com/natlab/sysarch/






From dino_levy at mindlessSPAM.com  Sat Apr  5 05:12:26 2003
From: dino_levy at mindlessSPAM.com (Dino Levy)
Date: Sat, 05 Apr 2003 12:12:26 +0200
Subject: POP3 and selective download/deletion
Message-ID: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com>

Hello!
I'm making a mail program and I have a little problem: I don't know how
to achieve the selective downloads/deletion from POP3 servers. For example,
I want to check whether there is an e-mail with subject:HELLO on my POP3 
server, and all I want for my program to do is to connect to my POP3 server,
check whether that specific mail is on my server, if it is then download
it, and if it isn't then break the connection. Is is possible? With this
kind of information, I could also make my program spam aware ( connect to
POP3, check whether spam mail with known subject or known sender is on
the server, and then delete it ). I really need your help...
Thanks in advance.



From staschuk at telusplanet.net  Sat Apr  5 16:18:16 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 5 Apr 2003 14:18:16 -0700
Subject: List of Numbers
In-Reply-To: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>; from news@titanic.co.uk on Sat, Apr 05, 2003 at 08:13:45PM +0100
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
Message-ID: <20030405141816.A1100@tibia.amotlpaa.bogus>

Quoth Simon Faulkner:
> I have a list of about 5000 numbers in a text file - up to 14 digits
> each.
> 
> I need to check for duplicates.
> 
> What would people suggest as a good method?

On Unixy systems,
	sort file |uniq -d
is probably easiest.  The approach is to sort the file and check
for consecutive duplicates.  This would be easy to implement in
Python too.

(This approach is in general more memory-intensive than a solution
using a hash table instead of sorting, the difference being more
or less pronounced as there are many or few duplicates.  But five
thousand 14-digit numbers in ASCII representation is 75000 bytes,
which is negligible.)

-- 
Steven Taschuk                                        staschuk at telusplanet.net
"Study this book; read a word then ponder on it.  If you interpret the meaning
 loosely you will mistake the Way."         -- Musashi, _A Book of Five Rings_




From stern at itgssi.com  Mon Apr 14 15:31:05 2003
From: stern at itgssi.com (James M. Stern)
Date: 14 Apr 2003 19:31:05 GMT
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <3E99B586.5040107@skynet.be> <mailman.1050344555.31977.python-list@python.org>
Message-ID: <slrnb9m33m.khv.stern@itgssi.com>

On Mon, 14 Apr 2003 20:19:49 +0200, Dirk Gerrits <dirk at gerrits.homeip.net>
wrote:
> Jan Dries wrote:
>> Don't worry too much about protected methods. Bjarne Stroustrup, creator 
>> of the C++ language, considers them to be a design mistake, and now 
>> regrets they were ever added to C++. See "The Design and evolution of 
>> C++" by Bjarne Stroustrup, page 302.
>> Java and I assume Ruby as well simply copied the concept from C++.
> 
> I don't have the book with me, but are you sure it wasn't protected DATA 
> members that Bjarne regretted?

It was.  From _The Design and Evolution of C++_, Section 13.9 (Protected
Members):

    "Fortunately, you don't have to use protected data in C++; 'private'
    is the default in classes and is usually the better choice.  Note that
    none of these objections are significant for protected member
    *functions*.  I still consider 'protected' a fine way of specifying
    operations for use in derived classes.

That paragraph is near the end of the section, and "these objections"
refers to arguments just made against protected data.

Admittedly, the next paragraph says that 'protected' "overcame my better
judgment [...]" but I think he means (data) members there.

... As an aside, I view every Python method I write as intended either
for the world, for this class and derived classes only, or for this
class only.  In short, as public, protected, or private.

Maybe it's my C++ heritage.

-- 
Opinions expressed above are not necessarily my employer's.
James M. Stern    stern at itgssi.com (213) 270-7955
ITG Software Solutions, Inc.
Culver City, CA 90230



From rmunn at pobox.com  Wed Apr  2 14:38:18 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 19:38:18 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org> <3e8afeb9.1125578@news-server> <3E8B0060.D9C449A3@engcorp.com> <3e8b1b0f.8380222@news-server>
Message-ID: <slrnb8lq1h.d7q.rmunn@localhost.localdomain>

BBFoto <bbfoto2002 at netscape.net> wrote:
> I apologize for top posting.  I clicked on "Followup Usenet Message".
> My mistake was unintentional.  I will try to find out more about it as
> was suggested.  And as to why I tried to execute a .pyc file:  If I
> knew what I was doing, then I would do it right.  I do appreciate all
> attempts to help.
> 
> I did a search on the PySol directory.  Nowhere is there a file called
> "PySol.py".  There is a series of files in sequence named
> "PySol_15.pyc"  "PySol_16.pyc"  "PySol_20..pyc"  "PySol_21.pyc" and
> "PySol_22.pyc".
> 
> 
> I tried to execute this line,
>  
> execfile('c:/program files/pysol/pysol-4.80/data/PySol_22.pyc') 
> 
> The path is correct.  After it failed I copied this error message from
> the Dos screen:
> 
>   File "c:/program files/pysol/pysol-4.80/data/PySol_22.pyc", line 1
>     -_
>      ^
> SyntaxError: invalid syntax

OK, try this:

C:\Python22\Python.exe "c:/program files/pysol/pysol-4.80/data/PySol_22.pyc"

(Substitute your path to Python as appropriate for the Python.exe).

If you get an error message about "global name 'win32_gethomedir' is not
defined", that's because you somehow got the Linux version of PySol
instead of the Windows version. If that's the case, go here:

    http://avitous.net/software/pysol-windows/index.shtml

and download the PySol 4.81 for Windows installer. I just tried that one
five minutes ago on my Windows box and it worked just fine; that should
get you up and running.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From tjreedy at udel.edu  Sun Apr 27 21:59:20 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Sun, 27 Apr 2003 21:59:20 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com> <7ULqa.9804$3M4.264757@news1.tin.it>
Message-ID: <g3qdncyq1Zx2FzGjXTWcrg@comcast.com>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:7ULqa.9804$3M4.264757 at news1.tin.it...
> Terry Reedy wrote:
> > I consider it a feature that all objects have truth value (except
for
> > perversely written user classes), so that conditional expressions
and
> > statements always work without exception.
>
> It's a feature, just like it WAS a feature that you USED TO be able
> to always test "if a>b:" without exception.

And I am not convinced that removing that invariant was entirely a
good idea.  I noticed on PyDev a couple of months ago that even Guide
has some sympathy for that view and the idea of adding a before()
method or something that would restore it.

Having some things like this that are/were true without exception (in
both the English and Python meanings of the word) are part of what
attracted me to Python.  They certainly made learning the language
easier for me (and other, I would think -- and I thought you also
cared about that ;-).

At one time, it was easy for anyone to get 100% on a quiz like 'how
many of the following lists can Python sort?".  Now, even some
experienced users, might miss the correct answer for the following
pair (0, 1, or 2?):

[1, 1j].sort()
['1', 1j].sort()

(For anyone who is not sure, make a guess and then cut&paste to fine
out.)

One bad turn does not deserve another.

Terry J. Reedy





From amk at amk.ca  Tue Apr 22 07:48:43 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Tue, 22 Apr 2003 06:48:43 -0500
Subject: Python implementation for a web spider \ site mirroring
References: <662cede5.0304220114.54e4dd10@posting.google.com>
Message-ID: <S8WdnR2SKryGsTijXTWcrg@speakeasy.net>

On 22 Apr 2003 02:14:34 -0700, 
	Tzahi <tzahij at hotmail.com> wrote:
> I need to copy complete web sites form the web.

Try GNU wget (http://www.gnu.org/software/wget/wget.html); not written in
Python, but quite featureful.

--amk



From randall at tnr.cc  Wed Apr 16 22:57:28 2003
From: randall at tnr.cc (Randall Smith)
Date: Thu, 17 Apr 2003 02:57:28 GMT
Subject: PyQT and Eric
In-Reply-To: <b7is7g$i1d$1@rex.ip-plus.net>
References: <b7is7g$i1d$1@rex.ip-plus.net>
Message-ID: <sKona.82884$Q27.4096742@twister.austin.rr.com>

I had the same problem and this fixed it.

 From the website:

eric3 needs PyQt snapshot 20021217 or newer and the respective 
QScintilla version

Hope this does it.

Randall

Daniele Trovato wrote:
> Hi
> 
> I tried to install Eric, but it won't work and I don't no why.
> I'm using SuSE 8.2 and I installed Python, PyQT and PyQT Develop,
> QScintilla, SIP ..... packages from the SuSE distribution.
> My problem is, when I'm trying to install Eric with the following command
> "python install.py", the shell returns me an error: "Sorry, Please install
> PyQT"  but I have PyQT already installed.
> Does anybody now where the problem is or how i can solve the problem ?
> 
> Thanks for any help
> 
> Daniele
> 
> 
> 




From donnal at donnal.net  Fri Apr 18 06:58:25 2003
From: donnal at donnal.net (Donnal Walter)
Date: Fri, 18 Apr 2003 05:58:25 -0500
Subject: Teaching : Python, Scheme, Java...
In-Reply-To: <20030417155900.A10011@tibia.amotlpaa.bogus>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr> <20030417155900.A10011@tibia.amotlpaa.bogus>
Message-ID: <b7ollh$phs$1@main.gmane.org>

Steven Taschuk wrote:
> Quoth Jean-Paul Roy:
>>- Python is bad : tail recursion is not iterative (also astonished at 
>>that, I can understand with an interpreter, but compiler ?)
> 
> 
> For teaching purposes, imho tail recursion is a Bad Thing.
> 
> Writing an iterative process using tail recursion has always
> struck me as a clever but obfuscated technique.  Write what you
> mean: if you want an iterative process, write a loop; if you want
> a recursive process, make recursive calls.

What is the difference between "recursive calls" and "tail recursion"? 
Sometimes I have written methods that call the same method on a parent 
or child object, with the depth of recursion being limited by a method 
of the same name NOT calling itself. In the past I have thought of this 
as tail recursion. Is this incorrect?

Thanks,

Donnal Walter
Arkansas Children's Hospital






From david.bear at asu.edu  Tue Apr  1 19:28:26 2003
From: david.bear at asu.edu (David Bear)
Date: Tue, 01 Apr 2003 17:28:26 -0700
Subject: stop exeuction and give my a prompt
Message-ID: <b6dara$5mv$1@news.asu.edu>

I would like to have the ability to insert a statement in my code that would 
stop program execution at that point and give me a python prompt.  I don't 
know what to call this. 

What I would like to do is to stop program execution and inspect the state 
of certain dictionaries --- I don't want to insert a bunch of print 
statements and watch output fly of the screen.  I'd also like to be able to 
fiddle with these dictionaries when my program stops.  

Any pointers?



From mads at troest.NEVERMORE.dk  Sun Apr 13 06:08:42 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Sun, 13 Apr 2003 12:08:42 +0200
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <bJ2ma.5447$883.4511@nwrdny02.gnilink.net>
Message-ID: <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net>

On Sun, 13 Apr 2003 01:05:11 GMT, Carl Banks wrote:

> Constant.
> Constant.
> Constant.

But is that /guaranteed/ (as it is with STL), or does it just happen to be 
that way with the current implementation (however unlikely it is to 
change)?

Regards,
   /\/\\ads Orbesen Troest



From schulz.online at gmx.de  Tue Apr 22 15:21:29 2003
From: schulz.online at gmx.de (Birgit Lachner)
Date: Tue, 22 Apr 2003 21:21:29 +0200
Subject: Problem with PyQt on SuSE 8.2
References: <b842hl$dsa$1@online.de>
Message-ID: <b844nl$fst$1@online.de>

Bernd Lachner wrote:

> Hello,
> 
> we try ro run several Python programs which are using PyQt on SuSE 8.2.
> Every time the same error message occurs:
> 
> Traceback (most recent call last):
>   File "./kdevpy.py", line 16, in ?
>     from qt import *
>   File "/usr/lib/python2.2/site-packages/qt.py", line 39, in ?
>     import libqtc
> ImportError: /usr/lib/python2.2/site-packages/libqtcmodule.so: undefined
> symbol: _ZNK16QAssistantClient9classNameEv
> 
> We tried the original PyQt packages which comes with SuSE 8.2. We also
> have compiled PyQt from the source RPMs, but the same error occurs.
> 
> Is ther e anybody who can help us? Thanks in advance.
> 
> Best regards
> 
> Bernd & Birgit Lachner
Okay ... as we compiled the original riverbank-sources the programms are
running now.

Bye ... Birgit & Bernd Lachner




From do-not-spam-ben.hutchings at businesswebsoftware.com  Tue Apr 15 06:03:31 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 15 Apr 2003 10:03:31 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <mailman.1050173000.12088.python-list@python.org>
Message-ID: <slrnb9nm7j.1bc.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <mailman.1050173000.12088.python-list at python.org>,
Alan McIntyre wrote:
> Just off the cuff here, it seems to me that the time required to do 
> "string += string" is going to be proportional to the size of the 
> original string whether you do it in Python or with the STL's 
> std::string or with C strings (so long as you don't know beforehand how 
> long the final string will be).
<snip>

Mutable string implementations may allocate more memory than their
current contents require, in order to avoid future reallocation and
copying.  A string implementation that grows its buffer exponentially
(e.g. doubling it every time it needs to grow) will have amortised
linear time complexity for additions, just like the STL vector or
Python list.



From armchairmillions at hotmail.com  Tue Apr 22 00:04:48 2003
From: armchairmillions at hotmail.com (Arm)
Date: 21 Apr 2003 21:04:48 -0700
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com> <mailman.1050087439.11600.python-list@python.org> <435d40e0.0304130212.5659bd4d@posting.google.com> <7Bgma.8749$LB6.237526@news1.tin.it> <435d40e0.0304152211.5a9511d4@posting.google.com> <ZZ7na.27561$T34.822409@news2.tin.it>
Message-ID: <14848047.0304212004.5aff060d@posting.google.com>

> > I had some small exposure to "iterators" in C++, and I remember them
> > being used in loops - iterating over the items in a container. The
> > example you have given appears to create a list from an "iterator"
> > with no loop. Something called an "iterator" creating a list with no
> > loop is very perplexing to me. But maybe my problem is I don't
> > recognize the syntax [ mo.group(0) for ... ]. It looks like a method
> > call on an object, followed by an incomplete for loop.
> 
> the syntax:
> 
>     [ <expression> for <target> in <iterable> ]
> 
> is known in Python as a "List Comprehension".  There is no issue of
> "with no loop" -- the "for" keyword inside the list comprehension
> is indicating exactly the fact that the loop is taking place.  (List
> Comprehensions may optionally have other clauses too -- for clauses
> to indicate nested loops, if clauses to use only SOME of the items
> in an iterable -- but I'm using the simplest form in the above).
> 
> If you're familiar with Python 1.5.2 but none of the Python changes
> in the last 4 years or so, then one way to explain list comprehensions
> is to say that
> 
>     return [ <expression> for <target> in <iterable> ]
> 
> is just the same as:
> 
>     templist = []
>     for <target> in <iterable> :
>         templist.append( <expression> )
>     return templist
> 
> The list comprehension notation is a Pythonization of Haskell's, which
> uses punctuation rather than the for and in keywords -- it would say:
>     [ <expression> | <target> <- <iterable> ]
> which reads just fine in Haskell but definitely not in Python; in turn,
> Haskell's list comprehension notation is a simple adaptation of the
> widespread "set comprehension" notation used in maths, e.g.
>     { x*x | x <- S }
> to mean "the set of the squares of the elements of set S" (where
> the <- set-membership indicato is often indicated by some glyph
> that's more reminiscent of the Greek letter epsilon).
> 
> 
> Python iterators are rather different beasts from C++'s, though
> the analogy between a Python iterator and a C++ "input iterator"
> is reasonably good... where in C++ you might have *it++ to
> indicate "fetch the current value of the iterator then advance
> the iterator", in Python you'd have it.next() for just the same
> purpose [the for loop makes that call intrinsically on your
> behalf!]; where in C++ you need to find out whether an iterator
> is done by comparing it with some "end marker", in Python an
> iterator lets you know it's done by raising StopIteration when
> you call its .next method [again, the for loop handles that for
> you intrinsically, catching StopIteration and just taking it as
> the indicator for normal termination of the iteration/loop].
> 
> 
> So, back to our muttons: are.finditer(instring) returns an iterable
> (which is actually already an iterator, but that's not very
> important here) whose items are the match-objects for each of
> the nonoverlapping matches of compiled regular expression object
> are inside string instring.  So, a loop of the form:
>     for mo in are.finditer(instring):
>         ...
> (whether written out like this, or inside a list comprehension,
> makes no difference) makes mo assume, one after the other, the
> values of RE matchobjects for each of those non-overlapping
> matches, left to right inside instring.  So for example a call
> to mo.group(0) gives the substring of instring that corresponds
> to the specific nonoverlapping match we're currently at, inside
> a loop such as the above.
> 
> 
Alex,
Thanks for the great explanation. It seems like both List
Comprehension and Iterators could use more documentation on
python.org. I think it would have been a lot more straightforward and
easier to comprehend (at least for beginners), if the object of a
"list comprehension" had been limited to a "sequence" or an
"iterable". (I would even prefer a different keyword when using
iterables). It doesn't make sense to be iterating through an iterator.
What is the gain from allowing the iterator to be placed in the
position a container should be - given that iterable
containers/objects must return the iterator anyway? And it seems like
in order to implement the "iterating through an iterator" syntax, an
equally confusing call was added to iterators - "return self".



From mstenner at phy.duke.edu  Wed Apr  9 13:09:42 2003
From: mstenner at phy.duke.edu (Michael Stenner)
Date: Wed, 9 Apr 2003 13:09:42 -0400
Subject: signals and system calls
Message-ID: <20030409130941.A30482@phy.duke.edu>

So, searching the archives, I see that this issue has been touched on
several times over the last 8 or so years.  I'm wondering if anyone
has a solution to my little problem:

I have a long-running python program.  I want it to terminate nicely
(ie not lose data) upon certain common signals.  The problem comes
with system calls.  If a system call is interruped by a signal, it
ends up raising an exception.  Now, the way you address this in C is
to set SA_RESTART via sigaction, but that doesn't appear the be
available in python.  [ I trust there is a good reason, btw, and am
curious about it, but am trying to keep this email goal-oriented :) ]

I've seen it suggested that you simply catch the exception and redo
the system call, but that gets REALLY ugly.  To do it reliably, you
need to wrap ALL system calls, including all the writes, reads, sends,
recvs, selects, etc. in ALL OF THE MODULES YOU USE.

Just to be clear, I'm not trying to "use" signals here.  They exist,
they WILL be sent (when a system shuts down, for example), and I'm
just trying to tolerate them in a way that doesn't lose data.

So, the bottom line is:  what do people recommend?  

				Thanks,
				 Michael
-- 
  Michael Stenner                       Office Phone: 919-660-2513
  Duke University, Dept. of Physics       mstenner at phy.duke.edu
  Box 90305, Durham N.C. 27708-0305




From rmunn at pobox.com  Tue Apr 22 13:58:37 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 22 Apr 2003 17:58:37 GMT
Subject: Compiling a program the embeds python
References: <3ea51711$0$4862$cc9e4d1f@news.dial.pipex.com>
Message-ID: <slrnbaabum.uvq.rmunn@localhost.localdomain>

Ben Lamb <news03 at zurgy.org> wrote:
> 
> I'm trying to compile an example program written in C that embeds Python.
> The porgram is:
> 
[snip]
> 
> I use the following to compile it:
> 
> ben at localhost:~/prog/pyembed$ gcc -I/usr/local/include/python2.2
> -L/usr/local/lib/python2.2/config -Xlinker -export-dynamic -lm -lpython2.2
> -ldl -lpthread -lutil pyembed.c
> 
[snip]
> 
> Please can someone point out what I am doing wrong.

The -l options, which specify which libraries to link against, need to
come *after* the file(s) that need those libraries. Yes, this breaks the
Unix rule-of-thumb that options always come first, and parameters
second. Your command line should be:

    ben at localhost:~/prog/pyembed$ gcc -I/usr/local/include/python2.2
    -L/usr/local/lib/python2.2/config -Xlinker -export-dynamic pyembed.c
    -lm -lpython2.2 -ldl -lpthread -lutil

Check the gcc manpage, and read the section that describes the -l
command. On my system, it starts at line 3630 of the manpage. You'll
find the following paragraph:

    It makes a difference where in the command you write this option;
    the linker searches and processes libraries and object files in the
    order they are specified.  Thus, foo.o -lz bar.o searches library z
    after file foo.o but before bar.o.  If bar.o refers to functions in
    z, those functions may not be loaded.

Hope this helps.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From shalehperry at attbi.com  Wed Apr  2 09:43:10 2003
From: shalehperry at attbi.com (Sean 'Shaleh' Perry)
Date: Wed, 2 Apr 2003 06:43:10 -0800
Subject: A question about help() and docstrings
In-Reply-To: <20030402092404.7a466b0c.rjkuhns@geetel.net>
References: <20030402092404.7a466b0c.rjkuhns@geetel.net>
Message-ID: <200304020643.10799.shalehperry@attbi.com>

On Wednesday 02 April 2003 06:24, Richard Kuhns wrote:
> Given a file dummy.py as follows:
> ====
> def dummy_function(one, two):
>     """This is a rather stupid function.
>     """
>     print "1 + 1 = 3"
> ====
>
> The doc string shows up just fine.  However, if I use 'from dummy import
> *' help() no longer knows anything about it.
>

in honor of teaching people to fish ....

look at site.py in your python directory.  help() is defined there as a 
function wrapping an instance of a class.  The class comes from pydoc.py in 
the same directory.  Give it a peak.




From afd00 at aub.edu.lb  Mon Apr 28 17:19:58 2003
From: afd00 at aub.edu.lb (Ali Dada)
Date: Mon, 28 Apr 2003 21:19:58 GMT
Subject: file transfer question
Message-ID: <E19AG2W-0001d2-00@fast.aub.edu.lb>

hi all:

i want to read a file from a network pc and use it in a local python script. i
thought of an easy way: i log on from within the python script to the host pc
using ssh then copy the file to my working directory using scp. below is the code:

import os
os.popen('ssh username at hostname')
os.popen('scp username at hostname:~/requiredFile ./copiedFile')

but what's happening is that when i execute the code, i get prompted for the
password by ssh, so i enter it, and then the program freezes and i can do
nothing.  how do you think i should alter the code? thanks in advance..


----------
Ali Dada




	









From hupp at cs.wisc.edu  Wed Apr 23 03:53:16 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Wed, 23 Apr 2003 02:53:16 -0500
Subject: why not say just what wants an integer?
In-Reply-To: <87d6jeeyo3.fsf@jidanni.org>
References: <87d6jeeyo3.fsf@jidanni.org>
Message-ID: <20030423075316.GA13945@upl.cs.wisc.edu>

On Wed, Apr 23, 2003 at 10:27:08AM +0800, Dan Jacobson wrote:
> Say, does lack of ability to figure out how to make this work 
> $ python -c 'print range(map(lambda x:x+1,(11,131,11)))'
> Traceback (most recent call last):
>   File "<string>", line 1, in ?
> TypeError: an integer is required
> 
> indicate 1. newbie, 2. lazy newbie, 3. can't really blame me?

How about I don't blame you for being a newbie :)

I assume you want to call range with those 3 arguments.  There are 2
ways to do this:

args = map(lambda x:x+1, (11,131,11))

range(*args)
apply(range, args)


As an aside, a nicer way of building the arguments is by using a list
comprehension:

args = [i + i  for i in (11,131,11)]

It's shorter and more readable, a rare combination.


> BTW, couldn't python be improved by saying just what wants an
  integer?

It's generally bad form to smash everything on one line like that.  A
message like "range requires an integer" would be somewhat more useful
though.


> Yes, it happened after I added range(), so range wants the integer, I
> can tell that.  So list() won't help etc.
> 
> P.S. "pydoc print" gives nothing. Should it give some clue to the user
> wanting to know about "print", or is that some violation of principles?

Sort of a violation.  Print is a keyword just like 'if' and 'def'.  so
pydoc would have to special-case all of those.  Not a bad idea but I
doubt it comes up very often.  

-Adam





From robin at jessikat.fsnet.co.uk  Sat Apr 12 05:18:57 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Sat, 12 Apr 2003 10:18:57 +0100
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
 <d116fbae.0304100704.68ef4833@posting.google.com>
 <c48$3e960655$516049d2$3498@nf1.news-service.com>
 <3e96efac$0$49107$e4fe514c@news.xs4all.nl>
 <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk>
 <3e96f481$0$49099$e4fe514c@news.xs4all.nl>
 <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>
 <3e970abc$0$49101$e4fe514c@news.xs4all.nl>
 <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>
 <3e972e43$0$49109$e4fe514c@news.xs4all.nl> <b77hpo$59a$0@216.39.172.122>
Message-ID: <6ZY$RKABo9l+Ew+N@jessikat.fsnet.co.uk>

In article <b77hpo$59a$0 at 216.39.172.122>, Bengt Richter <bokr at oz.net>
writes
.....
>>I really can't explain these numbers either? Localhost is 10 times slower
>>than over the network???
>
>IIRC, there was some discussion some time ago about this being a particular
>FreeBSD pessimality problem that wasn't true of Linux, but someone was fixing 
>it?
>But FreeBSD has good networking reputation I thought, so this needs double 
>check,
>for sure. Maybe it was for an old version?
>
....this is version 4.5. I guess things would be easier if one could be
sure that it's not a dns type thing, but I see from the dev list that
there's already some discussion about how/where dns lookups can be
avoided/speeded up with sockets which have a supposedly known address.
Naively I thought it would be relatively easy to just get a 32 / 64 bit
int back from some socket module function and then shove that into the
address part of the socket address.

Given the above perhaps it's not dns at all, but some other things that
get in the way. Do OSes impose limits on number of ports or
connections/second etc? 
>Regards,
>Bengt Richter

-- 
Robin Becker



From mhammond at skippinet.com.au  Thu Apr 10 20:41:56 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Fri, 11 Apr 2003 00:41:56 GMT
Subject: Problem setting COM property using win32com
In-Reply-To: <mailman.1049750232.29020.python-list@python.org>
References: <mailman.1049750232.29020.python-list@python.org>
Message-ID: <obola.8429$1h6.64545@news-server.bigpond.net.au>

The problem will be that the COM object you are using does not report 
any type information at runtime (it should).  Thus, we have no way of 
associating the object with the makepy data.


Green, Gregory P wrote:
> I missed the last line in your example code:
> 
> so if I do (using the EnsureModule line):
> 
> prmidattrs = MyModule.ISmRecordList()
> 
> I get:
> Traceback (most recent call last):   
> 	File "example.py", line 43, in ?     
> 		prmidattrs = MyModule.ISmRecordList()   
> 	File "C:\PYTHON22\lib\site-packages\win32com\client\__init__.py", line 310, in __init__     
> 		oobj = pythoncom.new(self.CLSID) pywintypes.com_error: (-2147221164, 'Class not registered', None, None) 
> pywintypes.com_error: (-2147221164, 'Class not registered', None, None) 
> 
> If I use 
> prmidattrs = MyModule.SmRecordList()

What will generally work is to make sure that the class name you are 
using is an *interface* (rather than a CoClass), and you can pass an 
existing COM object - eg:

ob = win32com.client.Dispatch("whatever")
# Should be early-bound, but it isn't - convert
ob = MyModule._SmRecordList() # the interface, not CoClass

You can use this technique regardless of how you were handed the object. 
  If you have access to the source of the object or the developers, ask 
them to implement the typeinfo classes for their objects (they already 
have the type info there, it just isn't exposed via their runtime objects)

Mark.




From cben at techunix.technion.ac.il  Tue Apr 15 07:38:47 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Tue, 15 Apr 2003 14:38:47 +0300 (IDT)
Subject: Speeding up: s += "string"
In-Reply-To: <mailman.1050393203.4704.python-list@python.org>
References: <mailman.1050393203.4704.python-list@python.org>
Message-ID: <Pine.GSO.4.51.0304151423510.2075@techunix.technion.ac.il>

Lulu of the Lotus-Eaters wrote on 2003-04-15:

> The topic of the O() problems of the "s += 'string'" operation has been
> discussed in several recent threads.  The aspect that seems not to have
> really come up is "why not FIX it?"
>
I'm +1 if it is feasible ;-).

> It seems to me that the behavior CAN be improved.  If strings were to
> use a preallocation strategy--similar to what lists do--MANY string
> appends could avoid a memory copy on the original string.  The intial
> value could stay where it was, and the extra characters could live in
> the already-available contiguous space.  Obviously, the utilized length
> would need to be stored somewhere.  Once in a while, of course, the
> preallocation would fill up; in that case you'd need to do a costly
> strcopy()... but this could be avoided MOST of the time, ISTM.
>
Don't forget that strings are immutable and you need the old object to
still represent the old string.  I see two approaches:

1. Each string object points to a string buffer and has an end index
   and multiple strings can share the same buffer.  Then it makes
   perfect sense to also have a start index.  The real gain is from
   allowing substrings (slicing) to be O(1)!  The hard part is memory
   management, so that when you retain only a small part of a big
   string, the unneed memory could be reclaimed.  Ironically, this
   works well for substrings but not for concatenation because of::

     s = "some long string"
     print s + "a"
     print s + "b"

   (you can only grow the initial string for one of them).  This might
   not be a problem because this usage is rare.

2. Lazy concatenation: string + string returns a proxy object that
   simply points to both; its materialized to a real string only when
   needed (initially on any other string operation but it's easy to
   implement things like iteration without requiring it).  Similar
   approaches could work for lists and dictionaries, BTW.  For the
   later I mean the non-commutative union of two dicts where one has
   priority and the other is the fallback; this would be handy for
   cases where consing to alists is used in LISP...

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From jkraska at san.rr.com  Tue Apr 15 22:42:29 2003
From: jkraska at san.rr.com (Courageous)
Date: Wed, 16 Apr 2003 02:42:29 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <mailman.1050438685.19213.python-list@python.org>
Message-ID: <sggp9v8lpk0045klmquc8gec1ju7q4kmpe@4ax.com>

>  Lambdas are an excellent tool for obfuscation, that's for sure.  You're
>right not to recommend overusing them in production code.  Generally, I try
>to limit myself to the simplest expressions, on the order of "x and y" or
>"result[index]", and translate to fully fledged functions for larger
>expressions.  It's just easier to maintain.

Anonymous functions are occasionally useful, as in Java's anonymous
class expressions, really used as a poor man's lambda of sorts. I've
seen this overused, though, that's for sure.

C//




From ramen at lackingtalent.com  Wed Apr 16 02:13:41 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Wed, 16 Apr 2003 06:13:41 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <mailman.1050453385.17091.python-list@python.org> <b7iigt$1d3jt$1@ID-169208.news.dfncis.de>
Message-ID: <slrnb9pt59.iin.ramen@lackingtalent.com>

In article <b7iigt$1d3jt$1 at ID-169208.news.dfncis.de>,
Greg Ewing (using news.cis.dfn.de) wrote:
> There isn't any widely understood name for what Python used
> to do, because no other language has ever been quirky
> enough to do it that way. It was lexical, but with only
> three levels visible at any one time.

Three? I thought it was two! Locals, globals, and...?

Dave



From duncan at NOSPAMrcp.co.uk  Thu Apr 17 06:13:51 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 17 Apr 2003 10:13:51 +0000 (UTC)
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <d6jmqr2q.fsf@python.net>
Message-ID: <Xns936072232AA84duncanrcpcouk@127.0.0.1>

Thomas Heller <theller at python.net> wrote in news:d6jmqr2q.fsf at python.net:
> "Terry Reedy" <tjreedy at udel.edu> writes:
>> Hmmm.  What I would really like is a structure comparison function
>> that spit out the first difference found  so I did not have to print
>> each and compare by eye.  That is not directly in the scope of your
>> problem, but it would be the most useful succinct display.
> 
> As you say, it's a different problem.
> The patches simply rearrange the output.
> 

Anyone fancy a patch that gives the output below? (I do hope it doesn't word wrap).

FFFF
======================================================================
FAIL: test_failUnlessEqual (__main__.Test)
----------------------------------------------------------------------
TestFailed: failUnlessEqual
'xxxxxxxxxxxxxxxxxx...xxxxyzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz...
                   ...    ^
'xxxxxxxxxxxxxxxxxx...xxxxazzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz...
                   ...    ^
  File "J:\temp\test.py", line 7, in test_failUnlessEqual
    self.failUnlessEqual(s, t)

======================================================================
FAIL: test_failunlessEqual2 (__main__.Test)
----------------------------------------------------------------------
TestFailed: <object object at 0x007E63B8> != <object object at 0x007E63C0>
  File "J:\temp\test.py", line 10, in test_failunlessEqual2
    self.failUnlessEqual(object(), object())

======================================================================
FAIL: test_failunlessEqual3 (__main__.Test)
----------------------------------------------------------------------
TestFailed: failUnlessEqual
{'A': 65, 'C': 67, ...0, 'S': 83, 'R': 82, 'U': 85, 'T': 84, 'W': 87, 'V': 8...
                   ...    ^^^^^^
{'A': 65, 'C': 67, ...0, 'P ': 0, 'R': 82, 'U': 85, 'T': 84, 'W': 87, 'V': 8...
                   ...    ^^^^^^                                            ...
  File "J:\temp\test.py", line 18, in test_failunlessEqual3
    self.failUnlessEqual(d1, d2)

======================================================================
FAIL: test_failunlessEqual4 (__main__.Test)
----------------------------------------------------------------------
TestFailed: failUnlessEqual
{'A': 65, 'C': 67, ...0, 'P ': 0, 'R': 82, 'U': 85, 'T': 84, 'W': 87, 'V': 8...
                   ...    ^^^^^^                                            ...
{'A': 65, 'C': 67, ...0, 'S': 83, 'R': 82, 'U': 85, 'T': 84, 'W': 87, 'V': 8...
                   ...    ^^^^^^
  File "J:\temp\test.py", line 26, in test_failunlessEqual4
    self.failUnlessEqual(d2, d1)

----------------------------------------------------------------------
Ran 4 tests in 0.187s

FAILED (failures=4)

The patch follows (it assumes you already applied Thomas Heller's patch):

----- unittest.diff -----
*** unittest.py.orig	Thu Apr 17 10:08:03 2003
--- unittest.py	Thu Apr 17 11:07:19 2003
***************
*** 146,152 ****
  
  class TestFailed(Exception):
      pass
!   
  class TestCase:
      """A class whose instances are single test cases.
  
--- 146,199 ----
  
  class TestFailed(Exception):
      pass
! 
! def shortdiff(x,y):
!     '''shortdiff(x,y)
! 
!     Compare strings x and y and display differences.
!     If the strings are too long, shorten them to fit
!     in one line, while still keeping at least some difference.
!     '''
!     import difflib
!     LINELEN = 79
!     def limit(s):
!         if len(s) > LINELEN:
!             return s[:LINELEN-3] + '...'
!         return s
! 
!     d = difflib.Differ()
!     diffs = list(d._fancy_replace([x], 0, 1, [y], 0, 1))
!     if len(diffs) == 3:
!         if diffs[1].startswith('+ '):
!             diffs.insert(1, '')
!         else:
!             diffs.append('')
!     if len(diffs) != 4:
!         # String were the same!
!         #return '\n'.join(diffs)
!         return limit(diffs[0][2:])
! 
!     # Remove '- ', '+ ', '? ' markers from start of lines,
!     # and newlines from end.
!     diffs = [ s[2:].rstrip() for s in diffs]
!     if max([len(s) for s in diffs]) < LINELEN:
!         return '\n'.join(diffs)
! 
!     # Need to shorten the string, but make sure
!     # first difference is visible.
!     diff1, diff2 = diffs[1], diffs[3]
!     a = len(diff1) - len(diff1.lstrip())
!     b = len(diff2) - len(diff2.lstrip())
!     if a == 0 or (b != 0 and b < a):
!         a = b
! 
!     if a > LINELEN/2:
!         diffs = [ s[:LINELEN/4] + '...' + s[a - 4:]
!                   for s in diffs ]
! 
!     diffs = [ limit(s) for s in diffs ]
!     return '\n'.join(diffs)
! 
  class TestCase:
      """A class whose instances are single test cases.
  
***************
*** 317,324 ****
             operator.
          """
          if first != second:
              raise self.failureException, \
!                   (msg or '%s != %s' % (`first`, `second`))
  
      def failIfEqual(self, first, second, msg=None):
          """Fail if the two objects are equal as determined by the '=='
--- 364,374 ----
             operator.
          """
          if first != second:
+             reprfirst, reprsecond = repr(first), repr(second)
+             if not msg and len(reprfirst) + len(reprsecond) > 60:
+                 msg = "failUnlessEqual\n" + shortdiff(reprfirst, reprsecond)
              raise self.failureException, \
!                   (msg or '%s != %s' % (reprfirst, reprsecond))
  
      def failIfEqual(self, first, second, msg=None):
          """Fail if the two objects are equal as determined by the '=='
-------------------------

For completeness test.py with the failing tests
---- test.py ----
import unittest

class Test(unittest.TestCase):
    def test_failUnlessEqual(self):
        s = 5000*'x'+'y'+5000*'z'
        t = 5000*'x'+'a'+5000*'z'
        self.failUnlessEqual(s, t)

    def test_failunlessEqual2(self):
        self.failUnlessEqual(object(), object())

    def test_failunlessEqual3(self):
        d1 = {}
        for i in range(65, 96):
            d1[chr(i)] = i
        d2 = {'P ': 0}
        d2.update(d1)
        self.failUnlessEqual(d1, d2)

    def test_failunlessEqual4(self):
        d1 = {}
        for i in range(65, 96):
            d1[chr(i)] = i
        d2 = {'P ': 0}
        d2.update(d1)
        self.failUnlessEqual(d2, d1)
        
if __name__=='__main__':
    try:
        unittest.main()
    except SystemExit:
        pass

-----------------

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From bokr at oz.net  Mon Apr 21 19:02:54 2003
From: bokr at oz.net (Bengt Richter)
Date: 21 Apr 2003 23:02:54 GMT
Subject: negative lookahead question
References: <mailman.1050942341.2204.python-list@python.org>
Message-ID: <b81tau$b9e$0@216.39.172.122>

On Mon, 21 Apr 2003 11:24:34 -0500, Skip Montanaro <skip at pobox.com> wrote:

>This re.sub call lives in Lib/stmplib.py as a way to make line endings
>canonical: 
>
>    re.sub(r'(?:\r\n|\n|\r(?!\n))', CRLF, data)
>
>This certainly seems to do what's desired, however, it looks overly complex
What is actually desired? Something that will work on a binary image of a text
file from any of Unix, Windows and Mac? How about the mixed uglies that happen
if you capture windows text in binary with \r\n CRLFs, and send it to Unix as
binary and cook it coming back to windows? This error doesn't even need a round
trip to Unix, if file modes on windows are mixed for some reason. The result is
typically /r/n -> /r/n -> /r/r/n. Should that be canonicalized as a single CRLF?

>to me.  First, the non-grouping parens are unnecessary.  Second, I don't
>think the negative lookahead assertion is required.  This simpler function
>call seems to do the trick:
>
>    re.sub(r'\r\n|\n|\r', CRLF, data)
Neither of the above do \r\r\n -> CRLF, whereas

     re.sub(r'\r+\n|\n|\r', CRLF, data)

would, I think:

 >>> import re
 >>> CRLF = '<CRLF>'
 >>> pats = [r'\r\n|\n|\r(?!\n)', r'\r\n|\n|\r', r'\r+\n|\n|\r']
 >>> data = 'line 0\r\nline 1\nline 2\rline 3\r\r\nline 4\n'
 >>> for pat in pats: print re.sub(pat,CRLF,data)
 ...
 line 0<CRLF>line 1<CRLF>line 2<CRLF>line 3<CRLF><CRLF>line 4<CRLF>
 line 0<CRLF>line 1<CRLF>line 2<CRLF>line 3<CRLF><CRLF>line 4<CRLF>
 line 0<CRLF>line 1<CRLF>line 2<CRLF>line 3<CRLF>line 4<CRLF>

I don't know what ugliness can happen ftp-ing inappropriately between
Mac and Windows, but I suppose you could get \r\n -> \r\n -> \r\n\n.
Maybe the canonicalizing  pattern should be r'\r+\n+|\n|\r' ;-)

>
>A simple test case containing a combination of different line endings seems
>to yield identical results:
>
>    >>> data = 'line 0\r\nline 1\nline 2\rline 3\r\r\nline 4\n'
>    >>> re.sub(r'\r\n|\n|\r(?!\n)',CRLF,data) == re.sub(r'\r\n|\n|\r',CRLF,data)
>    True
>
>Is there a case where the negative lookahead assertion will produce correct
>results but the simpler regular expression won't?
>
Don't see one, but that's no guarantee ;-)

Regards,
Bengt Richter



From Bill.Scherer at verizonwireless.com  Thu Apr 24 07:26:28 2003
From: Bill.Scherer at verizonwireless.com (Scherer, Bill)
Date: Thu, 24 Apr 2003 07:26:28 -0400 (EDT)
Subject: Problems at python.org??
In-Reply-To: <3EA7C922.5050708@ghaering.de>
Message-ID: <Pine.LNX.4.44.0304240725390.12990-100000@bagend.odc.vzwcorp.com>

On Thu, 24 Apr 2003, Gerhard Haering wrote:

> Scherer, Bill wrote:
> > Pointing my browser at http://www.python.org, I receive:
> > 
> > """
> > No more connections can be made to this remote computer at this 
> > time because there are already as many connections as the 
> > computer can accept.
> > """
> > 
> > Has there been a sudden Python Popularity Explosion?
> 
> That's strange, because a Google search suggests that you're really 
> getting this message from a Windows 2000 Server, not from the python.org 
> webserver.

Yep. Sorry for the alarm. It's a local issue. 

Now back to our regular programming...



> 
> -- Gerhard
> 
> 
> 
> 

-- 
Bill.Scherer at Verizon Wireless
RHCE 807101044903581





From cjw at sympatico.ca  Mon Apr 14 20:49:34 2003
From: cjw at sympatico.ca (Colin J. Williams)
Date: Mon, 14 Apr 2003 20:49:34 -0400
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net> <3E9B4D4A.3C1F53F@sympatico.ca>
Message-ID: <3E9B571E.50879BA4@sympatico.ca>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030414/77372afd/attachment.html>

From anton at vredegoor.doge.nl  Thu Apr 17 07:55:08 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Thu, 17 Apr 2003 13:55:08 +0200
Subject: Random web page?
Message-ID: <b7m4ql$ril$1@news.hccnet.nl>

>From the universe of all currently available webpages (where
"currently available" means available on the world wide web at the
runtime of the script) I want to select a page in such a way that each
page has an equally likely chance of being selected. Any idea about
how to accomplish this is equally likely to be welcome ;-)

Anton.



From mustafa.celik at sympatico.ca  Fri Apr 11 11:19:46 2003
From: mustafa.celik at sympatico.ca (Mustafa Celik)
Date: Fri, 11 Apr 2003 11:19:46 -0400
Subject: collecting data from file
References: <3E9671A3.2050004@sympatico.ca> <3E96B6BC.E9DBE74F@engcorp.com> <3E96DA23.6070500@sympatico.ca>
Message-ID: <3E96DD12.10903@sympatico.ca>

Correction on what columns really mean:

2nd column : if someone has arrived/left (HELLO/BYE)
3rd column : who he is (0x... ; hex)
4th column : how much $ they brought/took



Mustafa Celik wrote:
> The file that I'm analyzing guarantees to have >4 columns, and lines 
> contain string, integer, float..., they are seperated by spaces.
> 
> An example line is as below:
> 
>  1850099.32 HELLO 0xfce 6 OTTAWA stree_number_200 exit_Metcalfe 
> take_hw_417 arrive_at_Ottawa_Airport
> 
> 2nd column will tell me if someone has arrive, 3rd column will tell me 
> how much $ they'll bring with them, and 4th column will tell who he is 
> (a hex guy).
> 
> I will find out;
>   * how many people arrived and did not leave,
>   * how many have arrived and left
>   * how much we have in Ottawa at the end
>   * how much transient money flow have occured
> 
> My file may contain some lines that only leaves Ottawa, so I should 
> ignore them (they probably arrived earlier, and should not decrement my 
> dollar gain)  - This condition is not very important, if it'll 
> complicate things, I have other ways to prevent this.
> 
> I found a good awk document  last night (can't remember the web site), I 
> was planning to dig in it today, but you guys have supported me with my 
> initial plan with Python. I'm not a beginner in Python, this is probably 
> not the best project to start with, but I need to have this done by Awk 
> or Python, ...
> 
> Thanks,
>   Mustafa
> 
> 
> 
> 
> 
> Peter Hansen wrote:
> 
>> Mustafa Celik wrote:
>>
>>> I want to scroll thru a file;
>>>   * find lines that match a string (e.g. HELLO) on the 2nd column
>>>   * add up the 4th column (an integer) on each matching line, say the
>>> variable is TOTAL
>>>   * subtract the 4th column from TOTAL if another is string (e.g. BYE)
>>> is hit on 2nd column of a line
>>>
>>> Any tips?
>>
>>
>>
>> Yes, post an example showing what you intend.  The instructions are
>> decent, but implementable only by making a few assumptions about what
>> you mean.  Also, are there any error conditions to be handled, or do
>> you guarantee the input is always perfect (e.g., always at least four
>> columns, and so on)?
>>
>> -Peter
> 
> 




From aleax at aleax.it  Mon Apr 14 07:42:49 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 11:42:49 GMT
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk>
Message-ID: <Z8xma.11688$LB6.322565@news1.tin.it>

Matt Bergin wrote:

> Alex Martelli wrote:
>> Matt Bergin wrote:
>>    ...
>>>I'm just worried that if an exception occurs in a thread then the cursor
>>>will never be returned, using the 'no black magic' method which appeals
>>>to me the most. Could I prevent this by careful use of try: finally?
> 
>> Yes.  And you can *in addition* keep the cursors you've handed out
>> in a list and ALSO periodically do the scan checking refcounts just
>> to help you debug user threads that aren't returning cursors properly
>> (or use other tricks as above but for debugging-help rather than
>> functional purposes).
> 
> Just out of interest, when I run a very simple script, i.e.
> 
> import sys
> a=1
> print sys.getrefcount(a)
> 
> Why is the refcount of a so high (50 from the command line)? Where are
> these references lurking? Should I check be checking for changes in the
> refcount of an object rather than its absolute value?

Python can cache any immutable objects for performance reasons.
In practice, Python does such caching for small integers, and 1
is clearly a popular value -- any number of system-maintained
data structures may have attributes and items that happen to
be worth 1 (or 0, or other small, popular numbers).

import gc
gc.get_referrers(1)

will at any time give you a list of all containers which have
a reference to the value 1, if you're curious.  Of course, having
such a container in hand will not necessarily make it easy to
understand WHAT the container in fact IS -- but many will be
dicts with string keys, which lets you do some detective work;-).


Note that any Python function call always passes VALUES -- i.e.
there is no difference between your call to sys.getrefcount in
the above snippet and one that directly does sys.getrefcount(1).
In either case, it's the VALUE (1) that you're asking info about;
it makes no difference HOW you obtain the value to pass (I guess
this is probably already clear to you, but, just in case...).


In your case, you were talking about tracking objects of a specific
different type -- not ones that Python will implicitly cache and
reuse, such as small integers.  Thus, you do not necessarily need
to work with _changes_ in refcounts rather than with refcounts
directly -- I think using the refcounts directly may be simpler.


Alex




From tjreedy at udel.edu  Tue Apr 15 17:54:55 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 15 Apr 2003 17:54:55 -0400
Subject: Printing Columns
References: <mailman.1050426466.11198.python-list@python.org> <3E9C5035.F52E63BB@engcorp.com>
Message-ID: <9yCdnXfO2aMi4AGjXTWcpQ@comcast.com>

"Peter Hansen" <peter at engcorp.com> wrote in message
news:3E9C5035.F52E63BB at engcorp.com...
> c = 3
> print '\n'.join([' '.join([myList[i] for i in range(s, len(myList),
c)])
>     for s in range(len(myList)/c+1)])
...
> Note: relies on current integer division, otherwise you need to add
an
> int() call in the last bit.

You can guarantee future truncating behavior by using // instead of /
now.

TJR





From michaels at rd.bbc.co.uk  Fri Apr  4 04:30:28 2003
From: michaels at rd.bbc.co.uk (Michael Sparks)
Date: Fri, 04 Apr 2003 10:30:28 +0100
Subject: Best practice for object attributes?
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk> <3E8AEC44.2090909@removethis.free.fr> <b6eusn$avn$1@nntp0.reith.bbc.co.uk> <2uGia.77$nG6.43@nwrdny01.gnilink.net>
Message-ID: <b6jjkc$b30$1@nntp0.reith.bbc.co.uk>

Carl, Jp Calderone, laotseu:

Thanks for the comments between the 3 of you I now understand the
behaviour of the following ....

> python2.3
Python 2.3a2 (#1, Feb 20 2003, 09:54:48)
[GCC 3.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>1> class A: time='today'
...
>2> B=A()
>3> B_CB=A()
>4> B.time='yesterday'
>5> A.time='tomorrow'
>6> C=A()
>7> C.time
'tomorrow'          #1#
>8> B.time
'yesterday'         #2#
>9> B_CB.time
'tomorrow'          #3#

... to be this ....

1 Creates a class with class attribute "time", value 'today'.
2 Creates an object "B" with an empty __dict__, no local attributes
3 Creates an object "B_CB" with an empty __dict__, no local attributes
4 Inserts into "B" a local attribute "time" with value 'tomorrow'
5 Changes the class attribute "time" to 'tomorrow'
6 Creates an object "C" with an empty __dict__, no local attributes

7 The system looks in C for a local attribute time, and does not find
  one. It then looks in the class to see if it has attribute time, finds
  one and treats *that* as the value of C.time. 
  Hence the system gets the value of A.time - specifically 'tomorrow'
  hence the output on line #1#

8 The system looks in B for a local attribute time - it finds one with
  the value 'yesterday' in B's dict, hence the output on line #2#

9 The system looks in B_CB for a local attribute time, and does not find
  one. It then looks in the class to see if it has attribute time, finds
  one and treats *that* as the value of B_CB.time. 
  Hence the system gets the value of A.time - specifically 'tomorrow'
  hence the output on line #3#

This is based on these two explanations:

JP Calderone> When an attribute lookup on an instance fails, the lookup
JP Calderone> proceeds to the instance's class object (if it fails
JP Calderone> there, it proceeds to the class object's bases, and so
JP Calderone> on). 

Carl Banks> Before you modified the attribute, there was no
Carl Banks> such attribute in the instance dict

Carl Banks wrote:
> Even if you fully understood the dynamics (and it seems from this post
> you don't), another person working on the code might not.  

I didn't before I posted - I'd noticed the behaviour but wasn't aware
what was happening - value copying, copy on write, something different.
Turns out it's the something different - which *could* be termed
localisation on write, but that'd be almost as misleading as copying.
(since I presume you have all the shallow vs deep issues to consider)

For the system I'm writing this behaviour strikes me as a useful
optimisation of the system rather than necessary. The system will have
large numbers of objects all with a number of identical attributes -
which is why I considered placing them in the class rather than
instances. I accidentally discovered the syntax self.attribute to get
at the class attribute and was then rather surpised to be able to
modify the value on a per instance basis.

As a result if I hit speed or memory bottlenecks using this behaviour
would be useful, otherwise using the form of shrubbery2 - __init__
initialisation rather than class attributes is the better, less
surprising to others route to go down.

ie only use this shallow "localisation on write" if I hit speed/memory
problems, and if that happens document it's usage well.

> Generally, you should only use class variables when you want *all*
> (not most, all) instances to have the same attribute, and you want
> *all* instances to reflect any changes to it.

This is the case for the situation I was considering using the class
variables. Specifically I create large amounts of components with the
same externally defined interface defined in two attributes inboxes &
outboxes (which are then used by the superclass constructor to set up
more interesting instance attributes). 

Also each of these has access to a local debugger, again globally
defined. The thing I found suprising which prompted this question what
the fact that I was able to both globally change debugging levels -
which I wanted - and locally to an instance change debugging levels -
which was an unexpected bonus. However, I didn't understand where I
gained that behaviour and as a result didn't want to rely on it without
understanding it, especially if it's an unintentional side effect.

Upshot - I'll probably keep the components with the class attributes
they have - they really are the same for all instances of the class,
and have a long think about the debugger.

laotseu> oops ! I should have shut my mouth :(

Absolutely NOT - it made me explain myself better! Thanks :)

Thanks to all for your comments/advice - much appreciated!


Michael
-- 
Michael.Sparks at rd.bbc.co.uk    
British Broadcasting Corporation, Research and Development
Kingswood Warren, Surrey KT20 6NP

This message (and any attachments) may contain personal views
which are not the views of the BBC unless specifically stated.





From titanrebel at comcast.net  Sat Apr 26 00:32:13 2003
From: titanrebel at comcast.net (kkennedy)
Date: 25 Apr 2003 21:32:13 -0700
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com>
Message-ID: <14bd1399.0304252032.71752cb5@posting.google.com>

achrist at easystreet.com wrote in message news:<3EA97D61.3CC9FE74 at easystreet.com>...
> Any suggestions of the shortest path from a Python program to RTF
> output?  RTF seems to be the best (reasonable combination of 
> features and compatability) common denominator for exporting
> data so that (windows) users can bring it into their own word
> processing program, whatever that may be.  (Any problem with
> that?)
> 
> Requirements are about like this:
> 
> 1. Don't need an edit control, just want to write a file, but if
> I have to open an edit window, I can.
> 
> 2. Don't want to do too much fancy.  Text, headings, bulleted
> lists,  maybe some variations in font size.
> 
> 3. Development platform is WinNT, target platform is any reasonably
> recent Windows machine (eg either anything newer than Win95 release 1
> or Win98 or better). 
> 
> TIA,
> 
> 
> Al


Be sure to code to the least common RTF spec.  From my experience, the
RTF format (MS open format???) is a moving target.  RTF files created
by MS Word 97 will look odd or worse in MS Word 95, and be completely
unreadable in WordPad. Either find out what version of RTF WordPad
uses and use that, or require your users to have a specific version of
MS Word (or the free download - MS Word viewer).

Just a thought...HTML would be a far easier solution for simple
formating.  You said MS Word ruined the formatting of your HTML file,
but why do your users have to open the document in a word processor if
it isn't going to be edited?  Why not just open to document in a
browser.  It would be far simpler to code IE 3 compatable HTML code
than WordPad x.x version of RTF.



From staschuk at telusplanet.net  Tue Apr  1 00:05:20 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 31 Mar 2003 22:05:20 -0700
Subject: Is Python overhyped (just like Java)?
In-Reply-To: <b6aam1$7o6$1@news.island.liu.se>; from bellman@lysator.liu.se on Mon, Mar 31, 2003 at 09:07:13PM +0000
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <bd993a2f.0303310333.5a95ecb3@posting.google.com> <roy-D262FA.09230131032003@reader2.panix.com> <b6aam1$7o6$1@news.island.liu.se>
Message-ID: <20030331220520.B8627@tibia.amotlpaa.bogus>

Quoth Thomas Bellman:
> Roy Smith <roy at panix.com> writes:
> 
> > I'm all for C++ bashing, but in what way do you think C++ is not "a true 
> > object oriented language"?
> 
> A common complaint is that it confuses implementation (inheritance)
> with interface.  If a function expects a Frobnitz instance, then it
> better get an instance of a class that derives its implementation
> from the Frobnitz class, or the compiler will reject your program.
> A class that implements exactly the same interface, but uses its
> own implementation without inheriting from Frobnitz, won't work.

It's been a long time since I've done any C++ programming, but
don't pure virtual classes do the work of Java interfaces?

> After a while it was realized that that was untenable, and
> templates were invented to patch around it.

What do templates have to do with interfaces?

  [...]

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr  4 12:00:16 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 4 Apr 2003 17:00:16 GMT
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <mailman.1049388384.29614.python-list@python.org> <9383481e.0304031917.1d0100bf@posting.google.com> <b6j711$5m5dh$2@ID-169208.news.dfncis.de>
Message-ID: <slrnb8regv.rs.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <b6j711$5m5dh$2 at ID-169208.news.dfncis.de>,
Greg Ewing (using news.cis.dfn.de) wrote:
> Drew Moore wrote:
>> # modify the voltages, in a very simple, readable syntax.
>> vdduc = -3.8 # try a reduced supply
>> vreset = -3.2 # and increased reset drive...
>> vbias = 2.3 # and decreas the bias a little..
> 
> Maybe our brains work differently, but I'd find it
> FAR less confusing if the dac were an object and
> one did
> 
>    dac.vdduc = -3.8
>    dac.vreset = -3.2
>    dac.vbias = 2.3

I implemented a change to Python that would let you access
attributes of a 'default' object without naming it.  The default
object would be a 'for' loop's iterator (if a simple variable)
or the function's first argument.  This would let you (ab)use a
'for' statement as something like VB's 'with' statement:

    for dac in [dac]:
        .vdduc = -3.8
        .vreset = -3.2
        .vbias = 2.3
        
It would also let you omit many uses of 'self' in a method.

Whether that's desirable or whether I could release that, I do
not know.



From skip at pobox.com  Mon Apr  7 11:30:19 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 7 Apr 2003 10:30:19 -0500
Subject: PYTHONPATH question
In-Reply-To: <5ai29vogfthpqvbmlic09pbllltun61d2l@4ax.com>
References: <E1927CG-0007k5-00@fast.aub.edu.lb>
        <mailman.1049632149.9888.python-list@python.org>
        <5ai29vogfthpqvbmlic09pbllltun61d2l@4ax.com>
Message-ID: <16017.39307.610301.219881@montanaro.dyndns.org>

    nomad> For future safety, it might be an idea to make it:

    nomad>     PYTHONPATH=$PYTHONPATH:<colon-separated list of directories>
    nomad>     export PYTHONPATH

What's the interpretation of an empty element in $PYTHONPATH?

Skip




From theller at python.net  Sun Apr  6 11:42:03 2003
From: theller at python.net (Thomas Heller)
Date: 06 Apr 2003 17:42:03 +0200
Subject: pb py32exe and PIL
References: <20030406074208.01628.00000159@mb-cb.aol.com>
Message-ID: <brzj4oqc.fsf@python.net>

ryomanga at aol.com (Ryomanga) writes:

> As in the title I got a pb with py2exe and the PIL library.
> When I use the exe i've got the following error message :
> 
>   File "Image.pyc", line 960, in open
> IOError: cannot identify image file
> 
> It seems it can't read my pictures. I've try to change their format and
> directory but nothing changes. Is there something special to do with PIL like
> with PMW??
> 
You have to import the plugins you need in your script so that
py2exe knows they are used. You should be able to google for this
questions and find some answers.

Thomas



From peter at engcorp.com  Sat Apr  5 21:08:16 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sat, 05 Apr 2003 21:08:16 -0500
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
Message-ID: <3E8F8C10.A7F921CF@engcorp.com>

Simon Faulkner wrote:
> 
> I have a list of about 5000 numbers in a text file - up to 14 digits
> each.
> 
> I need to check for duplicates.
> 
> What would people suggest as a good method?

And after all those other replies, I still have to wonder *what*
you plan to do once you've found duplicates?  Remove them, 
quiet-like, so your list has just non-duplicated entries?
Report on all the duplicates?  Something else?

The solution can be different depending on what you plan to do...

(E.g. the dictionary approach is excellent for ending up with
a list with no duplicates, provided you don't care about
the ordering of the resulting list.  It's also the simplest
by far in Python, if you don't even care to know whether there
*were* duplicates, nor which they were.)

-Peter



From aleax at aleax.it  Wed Apr 16 13:05:50 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 17:05:50 GMT
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com> <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com> <cnoq9vci702imrunkenpulnefautq0n5ua@4ax.com> <Plena.22932$LB6.544795@news1.tin.it> <3cvq9vc9smcp7fbott8ivb1ujv0bgsk855@4ax.com>
Message-ID: <O3gna.23369$LB6.556843@news1.tin.it>

Dale Strickland-Clark wrote:
   ...
>>def f(fileobj):
>>    write = fileobj.write
>>    write('')
>>    # rest of function follows -- just call write('...') when needed
> 
> Way too fussy and obscure for my liking. Also, the assert is debugging
> code and nicely goes away in production.

...while the bound-methods stay in local variables and accelerate operations
even "in production" -- oh woe!  No accounting for tastes, I guess;-)


Alex




From svbrk at start.no  Thu Apr  3 08:32:29 2003
From: svbrk at start.no (Svein Brekke)
Date: 3 Apr 2003 05:32:29 -0800
Subject: IDLEfork: How to open all files in a single window
Message-ID: <21298535.0304030532.2c632e1f@posting.google.com>

I am using IDLEfork on Windows 2K and XP and on Mandrake Linux 9.0. I
have fruitlessly fiddled around in IDLEfork code to figure out how I
can modify it to my preferred window handling:
 - file->open opens a new file in the same window (I realize this
   can be done with the "open in same window" hook, but how?)
 - "idlefork foo.py" on the command line opens foo.py in 
   the existing idle window (if it does exist), i.e. same behavour
   as file->open (How do i send file name from the new process to
   the existing?)

Maybe it's in the code already. If not: I thought to implement this
myself, but would be greatful if any "housewarm" person could give a
hint of where to start.

Svein Brekke



From drew at astro.pas.rochester.edu  Tue Apr  1 23:11:36 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 1 Apr 2003 20:11:36 -0800
Subject: Overloadable Assignment PEP
Message-ID: <9383481e.0304012011.78f05548@posting.google.com>

Howdy! 

I submitted a pre-PEP on overloadable assignment. 

The idea is:

In situations where the assignment token "=" appears, and an augmented
assignment symbol (such as +=, -=, *=, etc..) would also be
syntactically correct, the assigned-to object would be checked for an
__assign__(self, other) method. If this __assign__ method existed, it
would be called, and the reference that it returned would be used for
the assignment.. otherwise, assignment would proceed as it currently
does now.

Thus, a few simple cases:

# do-nothing assign method...
  def __assign__(self,other) :
      return other    #  same as current implementation. 

# "un-writeable" object.. kinda 'const-ish'
  def __assign__(self,other) :
      return self     #  ignore attempt at assignment.. 

and a slightly more complex one:

  def __assign__(self,other) :
      self.value = int(other)
      self.notify() # do the other things needed when value changes
      return self  # and of course, don't lose identity.

David Goodger suggested I search on "__assign__", as this topic has
been discussed before. Well, I did the search, but I did not find a
thread that had quite the same slant on this topic as I have. In any
event, I want to champion this PEP, either to bring it to fruition, or
definitively explain why it won't work.

Most people who proposed this before had the "__assign__" operator as
a method of the "assigned-from" object, not the "assigned-to" object.
I agree.. that does not make sense.. but in the context of augmented
assignment, it seems sensible. (one person did briefly propose
something akin to this approach, but nobody seemed to take notice..)

The situations I can think of where an = assignment is syntactically
correct, but the augmented assignments are not:

1:
multiple assignment..
a = b = 4  # a = b += 4 does not work...

2:
tuple unpacking
a,b,c = 1,2,3  # a,b,c += 1,2,3 does not either.

Why the augmented assingment operators are syntactically incorrect in
these cases is another issue that I'd rather defer discussion on.. I
think probably  it is simpler to consider allowing overloading the
assignment operator in the same situations that the augmented
assignment operators allow overloading.. Just more consistent..

Anyway, That's the gist of it.
(Any insights into how to hack the "ceval" source to implement this
would be appreciated. Insights into why the idea won't work are
equally welcome; I figured I gain that insight by trying to implement
it, too..)

Drew



From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Mon Apr 14 13:57:22 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Mon, 14 Apr 2003 19:57:22 +0200
Subject: A python server
In-Reply-To: <15f445fb.0304140751.170d410e@posting.google.com>
References: <15f445fb.0304140751.170d410e@posting.google.com>
Message-ID: <3e9af68e$0$49112$e4fe514c@news.xs4all.nl>

dansan wrote:
> Hello all!
> 
> I have read and re-read the PY Sockets HOWTO, and I guess I'm still in
> the dark about how to do some stuff.
> 
> I'm trying to write a little server that will allow connections via
> telnet-like clients to play a board game.  The server would wait for a
> wile for connections, and at some point, it would read a command from
> the CLI on the server that says <startTournament>.  The server would
> then generate a tournament schedule and start the games.  Since these
> are turn-based games, the server would have to give each player its
> <go> command (and start a timer) in the proper sequence.
> 
[...]

Is it mandatory to support telnet clients?
If not, may I suggest you take a look at Pyro (http://pyro.sourceforge.net).

If you're able to write your game system and clients as a regular
python application, you can make it a distributed application with
a game server hub using Pyro in no time.

Pyro comes with a lot of examples (including a chat box),
perhaps you find something you need to get you going.

Bottom line: if you're already having trouble with the low-leve guts
of the system (the communication/socket code) I strongly suggest
that you drop that and use a higher level library or framework first.

--Irmen




From imbosol-1049311994 at aerojockey.com  Wed Apr  2 15:46:28 2003
From: imbosol-1049311994 at aerojockey.com (Carl Banks)
Date: Wed, 02 Apr 2003 20:46:28 GMT
Subject: PIL altrenatives?
References: <qWFia.668$n95.847183@twister.nyc.rr.com>
Message-ID: <E_Hia.101$nG6.62@nwrdny01.gnilink.net>

Max Khesin wrote:
> I am looking for a python func to do edge detection. A good function would
> be a version of im.point(func), where func would be passed the pixel's
> neighbourhood, so that different edges can be produced. As far as I can tell
> PIL does not support this directly. Of course, this could probably be done
> with PIL the long way, by getting individual pixels. But are there
> alternatives? Also, is there a good source for PIL code samples?


You can use Numeric, along with clever slicing, to do a whole edge
detect in one swoop.  I explained how to do something similar, a
cellular automaton, in another thread:

http://tinyurl.com/8ori

Basically, what you do is convert the image to a Numeric array
(this example assumes you have a grayscale image):

    width,height = image.size
    array = Numeric.fromstring(image.tostring(),Numeric.UnsignedInt8)
    array = Numeric.reshape(array,(height,width))

It sucks, but Numeric arrays are arraged in reverse order from Image
arrays, so you have to reverse width and height.  It would be helpful
for you to change the type of the array into Float:

    array = array.astype(Numeric.Float)

Then do the edge detect operation.  Maybe something like this
(my expample is very primitive, and probably wrong, of course):

    edge = Numeric.zeros((height-2,width-2),Numeric.Float)
    edge += array[1:-1,1:-1]
    edge -= 0.25*array[2:,1:-1]
    edge -= 0.25*array[1:-1,2:]
    edge -= 0.25*array[:-2,1:-1]
    edge -= 0.25*array[1:-1,:-2]

Then it back to an image:

    edge = edge.astype(Numeric.UnsignedInt8)
    edgeimage = Image.fromstring('L',(width-2,height-2),edge.tostring())

I hope you get the idea.


-- 
CARL BANKS




From marcel_achim at videotron.ca  Fri Apr 11 23:11:31 2003
From: marcel_achim at videotron.ca (Marcel Achim)
Date: Fri, 11 Apr 2003 23:11:31 -0400
Subject: mirroring text window to file
Message-ID: <3E9783E3.6010108@videotron.ca>

Hi,

I was wondering if there's an existing solution other than writing
to both a file and an output text window (Tk text widget), as in :

   text_window.insert('end', output)
   log_file.write(output)

thanks
Marcel




From avner at skilldesign.com  Tue Apr 15 14:07:44 2003
From: avner at skilldesign.com (Avner Ben)
Date: Tue, 15 Apr 2003 20:07:44 +0200
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il> <Xns935E66BE3E498duncanrcpcouk@127.0.0.1> <3e9c1413$1@news.012.net.il> <jJUma.23762$T34.707989@news2.tin.it>
Message-ID: <3e9c4a74$1@news.012.net.il>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:jJUma.23762$T34.707989 at news2.tin.it...

> You'll have to measure the performance for yourself, but you may
> find it quite good -- do, and let us know!
>

    Thank you all for the help. Actually, the problem is a little more
complicated and involves topological sort of the names to make sure that
names within names are not processed twice, but never mind. I got some
interesting hints. give me a few days and I'll let you know about the
performance.

        Avner.





From donn at drizzle.com  Tue Apr 29 02:15:46 2003
From: donn at drizzle.com (Donn Cave)
Date: Tue, 29 Apr 2003 06:15:46 -0000
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <20030417205405.A10672@tibia.amotlpaa.bogus> <mailman.1051310385.1948.python-list@python.org> <b8cgst$20ho$1@nntp6.u.washington.edu> <mailman.1051490038.11587.python-list@python.org> <NTjra.23897$WQ4.21608@news2.central.cox.net>
Message-ID: <1051596944.630307@yasure>

Quoth "Steve Holden" <sholden at holdenweb.com>:
| "Steven Taschuk" <staschuk at telusplanet.net> wrote in message
| news:mailman.1051490038.11587.python-list at python.org...
...
|> Hm... "conversion to type string".  So under this concept str()
|> converts an object to a string, while repr() represents the object
|> in a string, right?  Is this what you're saying when you
|> distinguish "the data as a string" from "the object as a string"?
|> If so, I quite like it.
|>
| </lurk>
| Finally, the murk lifts and a dim light becomes perceptible. "str() is what
| the users will recognize. repr() is what the programmers will thank you
| for."

Ouch, next someone will observe that str() should be friendly.  For me,
the murk starts rolling in at this point.  I'm sure _when_ str() output
is intended to be presented to someone, its success can be expressed in
terms of how friendly or recognizable it is, but that seems rather
generally true of anything and hard way to get specific about str().
Taken on its own, it has the potential to drag in an infinite number
of considerations (what education does this user have, etc.), or be
sort of irrelevant (where it isn't clear when or how the data will be
presented to anyone.)

One interesting thing about this discussion is that I suspect we all
would implement essentially similar __str__ and __repr__ functions,
even if we can't understand the concept in similar terms.

| The print __str__()/__repr__() dichotomy is probably the best compromise for
| the average Python user.

I didn't understand this sentence.  How does "print" figure in there?
Is the average Python user actually a computer programmer, or are you
talking about a user of software that happens to be written in Python?
The former, I guess?  What is the best compromise - the various ways
repr() and str() have actually been used lo these many years, or their
currently documented normative use?  Is this a lot of questions to ask
about such a short sentence?

	Donn Cave, donn at drizzle.com



From rob  Wed Apr 16 10:37:48 2003
From: rob (Arion)
Date: Wed, 16 Apr 2003 14:37:48 GMT
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net> <30uo9vo46k53v4to6d65koh72nqf0p8nvn@4ax.com>
Message-ID: <Xns935F6288EB740arionSPAMMENOTuseles@24.28.95.158>

> [1] My major objection to Programming Python is that 1) part of it's
> too advanced to be a good tutorial; 2) it's not properly organized to
> be a good reference; and 3) too many sections of it depend on having
> read all the prior sections, which is somewhat acceptable for a
> tutorial, but really falls short for a book where you're going to skip
> around -- the examples aren't very self-contained.

I have found Programming Python 2e to be an outstanding work, partly because 
I didn't have prerequisites for the book not to live up to, and partly 
because I was someone whou could benefit from an advanced tutorial guiding 
me gently through a lot of Pythonic concepts. I feel it greatly enhanced my 
modest ability with the language by starting at the beginning and working 
through.

For someone who has plumbed the depths of a few basic tutorials and 
appreciates a certain flavor of organized tutorial, it's a fine book. I 
think it comes down largely to what one wants/expects.

-Rob Andrews



From mwm at mired.org  Tue Apr  1 11:51:27 2003
From: mwm at mired.org (Mike Meyer)
Date: Tue, 01 Apr 2003 16:51:27 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com>
Message-ID: <x7of3qjh4g.fsf@guru.mired.org>

"Max Khesin" <max at cNOvSisiPonAtecMh.com> writes:

> > Yes, but humans aren't good optimizing compilers.
> So who writes optimizing compilers? :). Anyway, there are lots of
> super-optimized C++ libs that take advantage of C++ access to the machine
> and compile-time decision-making

And many of them are part of the runtime support of languages like
Python, or otherwise available to Python programmers.

The trick to getting good performance out of interpreted languages is
doing things to take advantage of those libraries. Sometimes, you may
have to wrap the libraries yourself. But after you've done that,
you've got an interactive environment to experiment with the libraries
in. That may be the most valuable thing to come out of doing the
wrapping.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From alloydflanagan at attbi.com  Wed Apr 16 17:36:29 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 16 Apr 2003 14:36:29 -0700
Subject: Using property in classic class may not work
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it> <a8b7f07a.0304150527.5b0b1187@posting.google.com> <hRUma.23848$T34.709378@news2.tin.it> <3E9C99E9.9000201@eb2.net.au> <3E9C9D1B.2000303@alldunn.com>
Message-ID: <a8b7f07a.0304161336.6a3bbec2@posting.google.com>

Robin Dunn <robin at alldunn.com> wrote in message news:<3E9C9D1B.2000303 at alldunn.com>...
> I have it on my ToDo list for figuring out how what exactly the problem 
> is (I have a general idea) and fixing it.  Not sure when I'll get to it 
> though, hopefully soon...
> 
> --Robin
> 

Hey, thanks for the response.  It's certainly a "nice to have", but
it's probably not critical.  Until, of course, classic classes go
away...

And while I'm at it, let me express my appreciation for a really nice
GUI library.  I've used several, wxPython is much easier to use than
most.



From staschuk at telusplanet.net  Tue Apr 22 16:41:58 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 22 Apr 2003 14:41:58 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
In-Reply-To: <b8420j$629$1@slb3.atl.mindspring.net>; from adalke@mindspring.com on Tue, Apr 22, 2003 at 12:38:30PM -0600
References: <mailman.1051023275.29338.python-list@python.org> <b8420j$629$1@slb3.atl.mindspring.net>
Message-ID: <20030422144158.A1009@tibia.amotlpaa.bogus>

Quoth Andrew Dalke:
> Bjorn Pettersen:
  [...]
> > Does anyone know the reason for not allowing an overload of the and
> > operator?
> 
> The 'and' and 'or' operators work on the boolean-ness of the objects,
> and apply short-circuit behaviour.  This is different than other binary
> operators, which don't short-circuit.
  [...]
> There's no way to implement that behaviour in Python, so
> it cannot be overridden.  Eg, suppose there is an "__and__".
> You might think it could look like
> 
>   def __and__(self, other):
>       if bool(self): return self
>       if bool(other): return other
>       return False

__and__ could be passed other as a callable which computes and
returns the second operand.  Then the standard __and__ could be
    def __and__(self, other):
        if bool(self):
            return self
        return other()

(Must return the second object when it is false, thus:
    >>> 1 and []
    []
rather than False.)

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From staschuk at telusplanet.net  Fri Apr 25 11:12:14 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 09:12:14 -0600
Subject: variable scoping problem
In-Reply-To: <_j3qa.1978$001.1638401@monger.newsread.com>; from kristofer@hotpop.com on Fri, Apr 25, 2003 at 05:02:50AM +0000
References: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com> <_j3qa.1978$001.1638401@monger.newsread.com>
Message-ID: <20030425091214.E261@tibia.amotlpaa.bogus>

Quoth Kristofer Wouk:
> Tung Wai Yip wrote:
  [...]
> >>>>x=1
  [...]
> >>>>def bar():
> > 
> > ...   #print x   <- referenced before assignment error
> > ...   x=2
> > ...   print x
  [...]
> OK, I'm a newbie but let me give this a try. I think since foo() has no 
> internal variable x it is forced to use the global(?, or whatever) x. 
> However, since bar() actually assigns something to x, it just creates a 
> new variable in bar()'s scope. Am I correct? I'd like to know myself...

Almost perfect.

The only thing you didn't mention is that assigning to x anywhere
in bar makes *all* references to x in bar refer to the local
variable, even those which precede the assignment.  This is why
referring to x before assigning to it raises an error (rather
than, say, using the value of the global variable).

(This is one of relatively few places where Python does static
analysis.  Another is generators, where the presence of yield
anywhere in the function body (except in nested functions) makes
the whole function a generator.)

-- 
Steven Taschuk                                                   w_w
staschuk at telusplanet.net                                      ,-= U
                                                               1 1




From sde at recombinant.demon.co.uk  Wed Apr  9 19:56:00 2003
From: sde at recombinant.demon.co.uk (Stephen D Evans)
Date: Thu, 10 Apr 2003 00:56:00 +0100
Subject: convert dictionary to string
References: <b72a7l$ifj$1@lust.ihug.co.nz>
Message-ID: <b72buh$9uo$1$8300dec7@news.demon.co.uk>

d={'A':45.10, 'F':56.50,}

# assuming that you would like 2 decimal places...

s1 = ''.join(['%s%.2f' % item for item in d.iteritems()])

# or without the list comprehension...

s2 = ''
for (key, value) in d.items():
    s2 = '%s%s%.2f' % (s2, key, value)


Stephen D Evans


"matthew" <matthew at newsgroups.com> wrote in message
news:b72a7l$ifj$1 at lust.ihug.co.nz...
> hi,
>
> can anyone tell what is the fastest way to convert a dictionary to a
string?
>
> eg: d={'A':45.10, 'F':56.50,...}
> to 'A45.10F56.50...'
>
> Thanks very much. matthew
>





From aleax at aleax.it  Sat Apr 12 18:41:42 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 22:41:42 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <1c8pot8sof9m3$.1a39huke99f0m.dlg@40tude.net>
Message-ID: <GC0ma.5919$LB6.169092@news1.tin.it>

Mads Orbesen Troest wrote:

> On Sat, 12 Apr 2003 18:45:26 GMT, Alex Martelli wrote:
> 
>> Building up lists of strings with mylist.append(piece), then joining
>> them together at the end with ''.join(mylist), is the canonical solution;
>> others, which I've seen already pointed out to you, include using
>> cStringIO.
> 
> Alex, thanks for your response. I had a hunch the += approach would be a
> performance killer. I'll try the list.append approach.

You're welcome, and my compliments for your accurate hunch -- to most
people, this particular trap comes as a surprise (it did to me the
first time I met it [which wasn't in Python, but in a C++ library
that didn't meet std::vector::push_back performance guarantees...]), 
though I think it becomes quite clear as soon as it's explained.


Alex




From adalke at mindspring.com  Tue Apr 29 14:13:15 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Tue, 29 Apr 2003 12:13:15 -0600
Subject: how to parse numeric data files
References: <20030429132519.4e92271e.gry@ll.mit.edu>
Message-ID: <b8mf41$jik$1@slb9.atl.mindspring.net>

george young:
> Currently we have a nasty mess of awk/shell/C/fortran programs that
> extract and process some data from these files.  I have a dream of
> a suite of simple, clear, maintainable python programs to do these tasks.

> Ideally, for each file format(there may be a dozen or so), a clear,
> concise descriptor file (yacc-like language definition? whatever)
> would drive the parsing process for reading that kind of data file,
> providing a nice clear API to the data.

Bioinformatics is another field which has this problem of many
"simple" file formats.  By simple I mean they are easy to parse
with regular expressions, and don't need a context-free grammar
used in yacc and other parser generators.  But they are complex
because they are very stateful - it's hard to partition the language
up into "tokens" which apply to a "grammar".

The solution most bioinformatics groups have done is to hand-write
a parser for each format.  Some build a data structure directly,
while others generate events so that a handler can build the
appropriate data structure.  The event model is nice because if
the event types are standardize, then it's sometimes possible to
reuse one handler for several file formats.

I've extended that one further and developed a parser generator
called Martel (part of the Biopython project) which takes a possibly
very large regular expression as the language definition, and creates
SAX2 events, so everything can be processed as if it's in XML.  See
  http://www.dalkescientific.com/Martel/
for more information about it.

Martel builds on mxTextTools, which is a low-level tool which
takes a while to understand, but is quite powerful.  Other relevent
tools, like SimpleParse, also build on that engine.

> [Below is a sample of one of the worst formats, shortened from a 40MB
file!]
> -- George

Martel parses in-memory and uses about 2x - 4x the memory
of the basic file.  If most of the file is a large number of blocks in
the same format (that is, optional header + repeats of a common format +
optional footer) then Martel includes some workarounds.

> PH_lot_id: YES_2_1.25V_1.0V
> Device_id: 4Mb_dc1
> Operator_id: EEA

A Martel grammer would look something like

from Martel import *
def Field(name):
  return Str(name + ": ") + ToEol(name)

header = (Field("PH_lot_id") + Field("Device_id") +
          Field("Operator_id") + ... )

> PFD: YES_2_1.25V_1.0V 4Mb_dc1 4M_digit_capture1 VDD=1.25V \
   VLEAKLOW=0V V_TCSETP=1.0V

or if you wanted more detail

PFD =   Str("PFD: ") + Word("PFD_YES") + Str(" ") + Word("PDF_SPAM") +
  Str(" ") + Word("PFD_SOURCE") + Str(" VDD=") + Float("PDF_VDD") +
  Str("V ") + Str("VLEAKLOW=") + Float("PFD_VLEAKLOW")  + ...

The parser is like XML

expression = header + PFD + ....

parser = expression.make_parser()
parser.setContentHandler( ... )
parser.parse(open("filename"))

                    Andrew
                    dalke at dalkescientific.com





From staschuk at telusplanet.net  Wed Apr 16 22:20:40 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 20:20:40 -0600
Subject: Standard way to express a date in a real number?
In-Reply-To: <v9rrq9ercgto01@news.supernews.com>; from johnroth@ameritech.net on Wed, Apr 16, 2003 at 08:07:49PM -0400
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net> <420ced91.0304152121.2af84a33@posting.google.com> <9t8na.27736$T34.825154@news2.tin.it> <v9rrq9ercgto01@news.supernews.com>
Message-ID: <20030416202040.B6846@tibia.amotlpaa.bogus>

Quoth John Roth:
> "Alex Martelli" <aleax at aleax.it> wrote in message
> news:9t8na.27736$T34.825154 at news2.tin.it...
  [...]
> > In Python 2.3, module datetime may serve your needs:
> >
> > >>> datetime.date(2303,4,16).toordinal()
> > 840893
> > >>> datetime.date(1303,4,16).toordinal()
> > 475651
> 
> Unfortunately, this appears to be based off of 1AD or
> some such. I usually want 4712BC as the base - in other
> words, the astronomical Julian date.

So fix it:

    >>> _epoch = datetime.date(-4711, 1, 1)
    >>> def date2ordinal(date):
    ...     return date.toordinal() - _epoch.toordinal()

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From andersjm at dancontrol.dk  Tue Apr  8 12:15:58 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Tue, 8 Apr 2003 18:15:58 +0200
Subject: ANN: Capon 0.1 - a software build framework
Message-ID: <mailman.1049901964.20840.clpa-moderators@python.org>

Capon 0.1
---------
Capon is a software build framework.
Or you might say, a Make replacement.
Or you might say, an alternate notation for Python.
Or you might say, a delegation-based object-oriented programming
language.

Capon is free software.

Why Capon?
----------
Capon blurs the line between a conventional make utility and a
general-purpose programming language to provide the best of both
worlds.

Requirements
------------
Capon needs Python 2.0 or higher.  (But has only been tested with
2.2.2.)

Where to get it
---------------
http://www.dancontrol.net/share/capon/index.html

Anders Munch
Haderslev, Denmark, 2003-04-07






From nas-usenet at arctrix.com  Mon Apr 14 11:15:32 2003
From: nas-usenet at arctrix.com (Neil Schemenauer)
Date: Mon, 14 Apr 2003 15:15:32 GMT
Subject: Parsing strings?
References: <698f09f8.0304132224.6a28c76f@posting.google.com> <7h3znmtpo0r.fsf@pc150.maths.bris.ac.uk>
Message-ID: <ogAma.15965$U34.9541@nwrddc02.gnilink.net>

Michael Hudson <mwh at python.net> wrote:
> tweedgeezer at hotmail.com (Jeremy Fincher) writes:
> 
>> Is there any way to turn a string (in Python format, escapes and all)
>> into an actual Python string without using eval()?
>> 
>> I.e., I'd like to turn the four-character string "\\x00" into the
>> one-character Python string "\x00" without using eval().  Is there any
>> way to do that?

In Python 2.3 courtesy of MvL:

    >>> '\\x00'.decode('string-escape')
    '\x00'

> Don't think so -- and if there is, Lib/compiler/transformer.py would
> like to know...

I guess someone should tell it.

  Neil



From eric.brunel at pragmadev.com  Tue Apr  8 08:48:00 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Tue, 08 Apr 2003 14:48:00 +0200
Subject: [help]Is this a bug or my code's fault
References: <2741546.1049794193@dbforums.com>
Message-ID: <b6ug1b$sn2$1@news-reader12.wanadoo.fr>

dolephi9080 wrote:
> import Tkinter
> import sys
> import os
> import os.path
> 
> 
> 
> if __name__ == "__main__":
>     rt = Tkinter.Tk()
>     btn = Tkinter.Button(rt,text='one')
>     btn.setvar('v','one')  #<=here I set a var to the btn object
>     with v='one'
>     btn.pack()
>     btn1=Tkinter.Button(rt,text="aa")
>     btn1.setvar('v','aa')   #<=here I set a vat to the btn1 object
>     v='aa'
>     btn1.pack()
>     print btn1.getvar('v') #<= print the btn1 v
>     print btn.getvar('v') #<= print the btn v
> 
>     rt.mainloop()
> 
> 
> 
> 
> please look at the comment, I set the value to two buttons, but when I
> print it at last, two buttons got the same data. I don't know why when I
> set the btn1's value, it will rewrite the btn's value, is this a bug? by
> the way I'm using activestate python 2.2.2

I don't quite understand what you're trying to do; maybe you should start by 
explaining your exact purpose with this code.

I particularly can't understand why you use the setvar method here. AFAIK, this 
method is used to manipulate variables in the tcl/tk layer behind Tkinter. Since 
Python has its own variables, I can't see any reason why you'd want to do this.

And the explanation to your problem is simple: tk has nothing looking like 
attributes on objects, just plain variables. So you don't do what you think you 
do: when doing btn.setvar('v', 'one'), you're setting a tcl/tk variable named v 
to the value "one". This has nothing to do with btn at all. So when you're doing 
btn1.setvar('v', 'aa'), you're modifying the *same* tcl/tk variable v to the 
value "aa".

But please explain what you're trying to do here, so that we can provide more 
useful advices.

HTH however...
-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From achrist at easystreet.com  Sat Apr 12 22:59:13 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Sat, 12 Apr 2003 19:59:13 -0700
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it>
Message-ID: <3E98D281.8C0CE87D@easystreet.com>

We went around on this one about 4 years ago. Is this a url:

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&threadm=37054490.E79ADCC9%40easystreet.com&rnum=1&prev=/groups%3Fnum%3D100%26hl%3Den%26lr%3D%26ie%3DISO-8859-1%26as_drrb%3Db%26q%3Dgroup%253Acomp.lang.python%2Binsubject%253Astring%2Bauthor%253Aachrist%2540easystreet.com%26btnG%3DGoogle%2BSearch%26as_mind%3D12%26as_minm%3D5%26as_miny%3D1993%26as_maxd%3D12%26as_maxm%3D4%26as_maxy%3D1999

About 6 messages into the thread, Tim Peters came up with a way that
was about 5 times faster than string.join().

Now, which is preferred:

	if astring == "":

or 

	if len(astring) == 0:

????


Al

Alex Martelli wrote:
> 
> Mads Orbesen Troest wrote:
> 
> > Given that strings are immutable, is it exceedingly slow (and memory
> > spiking) to do a lot of "string += string" operations in one's code? IOW,
> 
> Yes -- it's THE one major performance trap in naively coded Python,
> since, when "a lot" is a sufficiently large N, a lot of += on short
> strings ends up being O(N squared) where O(N) approaches would be
> available.  And turning something potentially O(N) into something
> O(N squared) is the kind of performance hit that profiling won't
> necessarily catch, because it shows up (as unboundedly bad) only
> for N sufficiently large...
> 
> So, it's basically the one performance trap I find it most necessary
> to teach even to total beginners!
> 
> > does each and every += evaluation lead to a new concatenated copy of the
> > previous string (now freed for garbage collection) and the new string, so
> 
> Exactly (it's not necessarily the case that the previous string is freed
> for garbage collection, but whether it is, or not, a new concatenated copy
> is indeed unavoidable).
> 
> > that the longer the string to which stuff is appended is, the longer times
> > each += operation takes?
> 
> Again, exactly -- s += l takes roughly O(len(s)+len(l)) time (as it
> necessarily has to copy the len(s) previous chars followed by the len(l)
> new ones).
> 
> Building up lists of strings with mylist.append(piece), then joining
> them together at the end with ''.join(mylist), is the canonical solution;
> others, which I've seen already pointed out to you, include using cStringIO.
> 
> Alex



From gh at ghaering.de  Sat Apr 19 09:02:04 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Sat, 19 Apr 2003 15:02:04 +0200
Subject: getting an item's list index
In-Reply-To: <1050757065.735.2.camel@carpathia>
References: <1050757065.735.2.camel@carpathia>
Message-ID: <3EA148CC.4030708@ghaering.de>

Michael Anckaert wrote:
> Hello all,
> 
> I have a list called contents. If one of the elements has the value
> "item1" for example, how can I get that elements index? [...]

contents.index("item1")

:-)

-- Gerhard




From LogiplexSoftware at earthlink.net  Thu Apr 10 12:56:42 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 10 Apr 2003 09:56:42 -0700
Subject: PEP305 csv package: from csv import csv?
In-Reply-To: <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>
References: <mailman.1049814434.25284.python-list@python.org>
	 <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>
Message-ID: <1049993801.13006.47.camel@software1.logiplex.internal>

On Wed, 2003-04-09 at 17:03, Magnus Lie Hetland wrote:

> csv.py, right? (I consistently make this mistake myself...)

It gets worse when you've been typing "csv" for an hour then try to
commit the work to cvs ;)

> As for reasons -- the checkin message for the __init__.py (made by
> Cliff Wells) is:
> 
>   "Make cvs [sic] into a package so we can have a util package"

See?

> So it seems that prior to this decision, csv was simply a module, and

Yep.

> the names were directly available. I'd think this to be preferrable,
> and more consistent with existing library modules. We may have modules
> containing classes with the same name, but this convention of placing
> a module inside a package with the same name does not occur at all in
> the libs. (Yes, I checked ;)

We needed to have a sub-package (util) and that required having csv
itself as a package (unless someone knows how to get around this - my
knowledge of the import system is fairly limited).

> Of my currently installed packages, I see that numarray uses the
> convention, for example, in its RandomArray2 module/package. However,
> it also uses the above suggested convention of placing
> 
>   from RandomArray2 import *
> 
> in RandomArray/__init__.py.
> 
> So ... Are any of the PEP authors reading this? (Just checking -- as
> this probably belongs on the discussion list... Sorry about that ;)

Skip has this on his list.

Regards,

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From paul at prescod.net  Mon Apr  7 18:41:55 2003
From: paul at prescod.net (Paul Prescod)
Date: Mon, 07 Apr 2003 15:41:55 -0700
Subject: Problem setting COM property using win32com
In-Reply-To: <E1B4F80EB080194BBAD3092392F511B3DADEF1@XCH-NW-13.nw.nos.boeing.com>
References: <E1B4F80EB080194BBAD3092392F511B3DADEF1@XCH-NW-13.nw.nos.boeing.com>
Message-ID: <3E91FEB3.6070209@prescod.net>

Bare in mind that I'm not a COM expert. We should probably be talking 
about this on the python-win32 list. But I see something strange in your 
makepy code...

Green, Gregory P wrote:
> 
> # This CoClass is known by the name 'SmRecList.SmRecordList'
> class SmRecordList(CoClassBaseClass): # A CoClass
> 	# An object used for storing data. A collection of headers and records, each composed of a collection of nodes.  In turn, each node can contain information of various types.
> 	CLSID = pythoncom.MakeIID("{C85E7014-8B4F-11D1-8E20-00A02498EA3C}")
> 	coclass_sources = [
> 	]
> 	coclass_interfaces = [
> 		'{C85E7013-8B4F-11D1-8E20-00A02498EA3C}',
> 		'{C85E7015-8B4F-11D1-8E20-00A02498EA3C}',
> 	]
> 	default_interface = '{C85E7013-8B4F-11D1-8E20-00A02498EA3C}'

In my makepy'd classes, default_interface is never a string.

class ResourceManager(CoClassBaseClass): # A CoClass
	# ResourceManager Class
...	default_interface = _ResourceManager

# This CoClass is known by the name 'XMetaL.Global'
class Global(CoClassBaseClass): # A CoClass
....
	default_interface = _Global

That's probably why you're getting the "cannot call string" problem. 
Look at the __init__ that's causing a problem. It probably looks like this:

class CoClassBaseClass:
	def __init__(self, oobj=None):
		if oobj is None: oobj = pythoncom.new(self.CLSID)
		self.__dict__["_dispobj_"]=self.default_interface(oobj)

I don't know how your default_interface got to be a string (perhaps a 
genpy bug, or hand-tweaking, or something else I don't understand), but 
I'd suggest you replace it with something like:

default_interface = _SmRecordList

or maybe

default_interface = ISmRecordList

  Paul Prescod





From frithiof.jensen at removethis.ted.ericsson.dk  Fri Apr 11 06:42:10 2003
From: frithiof.jensen at removethis.ted.ericsson.dk (Frithiof Andreas Jensen)
Date: Fri, 11 Apr 2003 12:42:10 +0200
Subject: Python 2.2.2 & Tix on win32 - What must I do?
Message-ID: <b7666e$oto$1@newstree.wise.edt.ericsson.se>

On spotting the Tix documentation in the module docs, I gleefully set about
giving it a spin.

This is what I got however:
" Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
  Type "copyright", "credits" or "license" for more information.
  IDLE 0.8 -- press F1 for help
  >>> import Tix
  >>> root = Tix.Tk()
  Traceback (most recent call last):
    File "<pyshell#1>", line 1, in ?
      root = Tix.Tk()
    File "C:\Python22\lib\lib-tk\Tix.py", line 213, in __init__
      self.tk.eval('package require Tix')
  TclError: can't find package Tix
  >>>
"
What must I *do* exactly to fix this?





From oren-py-l at hishome.net  Tue Apr 15 09:59:50 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Tue, 15 Apr 2003 09:59:50 -0400
Subject: Speeding up: s += "string"
In-Reply-To: <Pine.GSO.4.51.0304151423510.2075@techunix.technion.ac.il>
References: <mailman.1050393203.4704.python-list@python.org> <Pine.GSO.4.51.0304151423510.2075@techunix.technion.ac.il>
Message-ID: <20030415135950.GA8115@hishome.net>

On Tue, Apr 15, 2003 at 02:38:47PM +0300, Beni Cherniavsky wrote:
> > It seems to me that the behavior CAN be improved.  If strings were to
> > use a preallocation strategy--similar to what lists do--MANY string
> > appends could avoid a memory copy on the original string.  The intial
> > value could stay where it was, and the extra characters could live in
> > the already-available contiguous space.  Obviously, the utilized length
> > would need to be stored somewhere.  Once in a while, of course, the
> > preallocation would fill up; in that case you'd need to do a costly
> > strcopy()... but this could be avoided MOST of the time, ISTM.
> >
> Don't forget that strings are immutable and you need the old object to
> still represent the old string.  I see two approaches:
> 
> 1. Each string object points to a string buffer and has an end index
>    and multiple strings can share the same buffer.  Then it makes
>    perfect sense to also have a start index.  The real gain is from
>    allowing substrings (slicing) to be O(1)!  The hard part is memory
>    management, so that when you retain only a small part of a big
>    string, the unneed memory could be reclaimed.  Ironically, this
>    works well for substrings but not for concatenation because of::
> 
>      s = "some long string"
>      print s + "a"
>      print s + "b"
> 
>    (you can only grow the initial string for one of them).  This might
>    not be a problem because this usage is rare.
> 
> 2. Lazy concatenation: string + string returns a proxy object that
>    simply points to both; its materialized to a real string only when
>    needed (initially on any other string operation but it's easy to
>    implement things like iteration without requiring it).  Similar
>    approaches could work for lists and dictionaries, BTW.  For the
>    later I mean the non-commutative union of two dicts where one has
>    priority and the other is the fallback; this would be handy for
>    cases where consing to alists is used in LISP...

Guido is reluctant to add major new code to the code unless it gets some
serious real-world testing as an extension module. But in the case of
strings there is a problem: you can't create an object that will be a
drop-in replacement for a string in Python unless it's actually derived
from Py_StringObject and has the same internal structure. It can be an
object that behaves much like a string and it can have an __str__ method
to convert it into a string - but it will not be a compatible with a
string in many situations. For example, it will not be accepted as an
argument to string methods. It's not enough to make an object that can
be converted into a string (almost any Python object can). It needs to
*be* a string without sharing the same internal structure.

Such objects could be derived from the basestring class without being a
subclass of either str or unicode. The only method they must implement
is __str__. All other string methods and operations will be emulated by
calling __str__ and passing the result to the original string method.
The object may optionally override other methods to provide a more
efficient implementation. With this kind of infrastructure in place it
would be possible to experiment with mutable strings, lazy concatenation, 
shared-buffer strings, etc as extension modules. The best solution could
be considered for eventual integration into the core.

It would require quite a lot of changes (ParseTuple, almost all the
places that call PyString_Check, etc) but I think the Python code might
actually benefit from it and become smaller and cleaner.  Currently 
there are a lot of places with hard-wired check for string,/unicode and
objects exposing the buffer interface. These could be handled in a more
generic way.

   Oren




From aleax at aleax.it  Thu Apr 24 17:21:57 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 21:21:57 GMT
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com> <Xns9367885737E62duncanrcpcouk@127.0.0.1> <EARpa.120$3M4.2728@news1.tin.it> <Xns93679D1BF4A2Cduncanrcpcouk@127.0.0.1>
Message-ID: <VzYpa.1990$3M4.48787@news1.tin.it>

Duncan Booth wrote:

> Alex Martelli <aleax at aleax.it> wrote in
> news:EARpa.120$3M4.2728 at news1.tin.it:
> 
>> ..isn't this smoother and handier?  Or does it just look that way
>> to me, due to overexposure to custom metaclasses?-)
>> 
> Yes, I agree that it is better, but I have one niggling worry about this:
> 
> Your metaclass works fine for the one requirement given here, and other
> metaclasses you have proposed work fine in their place, but what happens
> if you want to combine these and get e.g. a Bunch which counts its
> instances?

You make a new metaclass multiply inheriting from the ones you have
(and solve conflicts as usual for multiple inheritance).


> I haven't tried it to see what happens, maybe you just need to indirect
> type.__new__ calls through super, and make a new submetaclass with the old
> metaclasses as bases, but I'd like to know if there are issues here or if
> I'm just worrying needlessly.

Both;-).  Yes, there are issues (custom metaclasses haven't been used
widely enough yet to shake out all the early problems -- I keep thinking
about them but I don't use them in production code yet, for example,
partly because I generally don't *need* the class-based functionality,
such as automatically counting instances, which is SO often requested
in this newsgroup!-).  However, I think that in general your worries may
be needless -- or at least, no more needful than any other worry that
relates to multiple inheritance, a construct that, of course, _does_
give you headaches SOME of the time.


Alex




From aleax at aleax.it  Tue Apr 29 08:28:27 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 12:28:27 GMT
Subject: Calling dotNet components from Python.
References: <924a9f9c.0304282138.15aa7635@posting.google.com>
Message-ID: <Ldura.19534$3M4.565359@news1.tin.it>

Paradox wrote:

> I was wondering if it is possible to call a dotNet dll written in C#
> from within a Python script. If anyone knows of any documentation or
> example code that accomplishes something like this. It seems that you
> can call a com component but all my searching on dotNet has not found
> anything about using dotNet components in Python. Thanks

There's a project by Chetan Gadgil called Kobra for this purpose, but 
right now I cannot access its page,
http://home.attbi.com/~chetangadgil/DotNetWrapperForPython.htm -- I
hope it's just some transient glitch.


Alex





From roger_hancock at hotmail.com  Thu Apr 17 01:00:01 2003
From: roger_hancock at hotmail.com (Roger Hancock)
Date: 16 Apr 2003 22:00:01 -0700
Subject: Memory leak while looping
Message-ID: <1020cbf8.0304162100.6fc49e6f@posting.google.com>

I'm just starting to play with writing C extensions for Python.  In
the process of doing so, I called one of my extension functions
thousands of times only to discover that a leak was eating up all the
memory in my system.  Upon further investigation, I found I could
produce the same result with the following python code with my
extension completely out of the picture:

#!/usr/bin/env python2

for i in range(1,10000000):
    item = i


I would have thought that the memory for "item" would be freed as new
"item"s were created.  However, this doesn't seem to be the case.  I
would appreciate any clarification on this issue.


Thanks,

Roger



From rmunn at pobox.com  Thu Apr  3 04:55:48 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 03 Apr 2003 09:55:48 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org> <3e8afeb9.1125578@news-server> <3E8B0060.D9C449A3@engcorp.com> <3e8b1b0f.8380222@news-server> <slrnb8lq1h.d7q.rmunn@localhost.localdomain> <3e8b5589.10546880@news-server> <3E8B6716.11FDDC7F@engcorp.com>
Message-ID: <slrnb8nc9b.eaq.rmunn@localhost.localdomain>

Peter Hansen <peter at engcorp.com> wrote:
> BBFoto wrote:
>> 
>> I copied your line and pasted it.
>> 
>> File "<stdin>", line 1
>>     C:\Python22\Python.exe "c:/program
>> files/pysol/pysol-4.80/data/PySol_22.pyc"
>> 
>>      ^
>> SyntaxError: invalid syntax
> 
> This is a message from the Python interpreter.  You should
> not be *in* the interactive interpreter when you type that.
> It is an MS-DOS command, to execute the bytecode in the .pyc
> file using the python.exe executable.
> 
> May I suggest you read some parts of the FAQ at www.python.org, 
> and also go through the Python tutorial available there, before
> trying to do much more?  You are missing a substantial amount of
> background knowledge necessary to understand the advice you are
> being given, both in Python and Windows/DOS.

Oh my, yes. I was assuming that you would immediately recognize that
line as a DOS command. It won't work in the Python interpreter.

Let me repeat my suggestion that you instead try to download and run the
Windows version of PySol found here:

    http://avitous.net/software/pysol-windows/index.shtml

It has a simple installer program that you can double-click on, which
will put icons in your Start menu and set everything up for you.

Now, once you've got the game up and running via that Windows installer,
you may be tempted to give up on understanding why your previous
attempts didn't work. *DON'T* give in to that temptation! Continue to
experiment, and try to understand the results. When you do finally "get"
why some of your attempts failed, and why others succeeded, the "aha!"
experience will be very rewarding, *and* it will provide you with a much
deeper level of understanding of how computers work.

Yes, it's hard work trying to reach that "aha!" moment; but it's worth
it, I promise you. Keep trying.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From peter at engcorp.com  Thu Apr 17 13:51:46 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 17 Apr 2003 13:51:46 -0400
Subject: More
References: <teBna.9109$pK2.11130@news.indigo.ie>
Message-ID: <3E9EE9B2.E0FDADF2@engcorp.com>

kpop wrote:
> 
> I usualy print strings the normal way
> print x, but if the string is too big it will
> go by the screen too fast and i wont be
> able to see anything. I notice the python
> internel help system hasa thing like the
> unix prgram more, where you press enter
> to go down. How can i use that in my own
> programs? Is there a function or a module.
> I am using windows so i dont have more.

Although this one is non-obvious, you can learn a lot by trying
to check the source.  I found how it works by doing this:

>>> type(help)
<type 'instance'>
>>> help.__class__
<class site._Helper at 0x00794990>
>>> help.__module__
'site'

Then I looked at site.py in the standard library, and found
that it uses pydoc for the help feature.

Looking at the source in pydoc.py, I could see a variety of 
"pagers" that are related to this function.

After a false start trying to do this:

 pydoc.ttypager(longstring)

which did nothing effective, I noticed getpager() just above it.

>>> pydoc.getpager()(longstring) 
long strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong strinxlon
g strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong s
trinxlong strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong stri
nxlong strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong strinxl
ong strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong
 strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong strinxlong st
-- More --


Hmm... seems to work okay.  I'm sure you can learn more by looking
at the internals of pydoc a little more... or maybe this stuff is
even documented if you check the docs!  

-Peter



From djc at object-craft.com.au  Tue Apr 29 08:50:30 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 29 Apr 2003 22:50:30 +1000
Subject: Web programming and a different "type" problem
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com> <slrnbarhfb.67t.jon+usenet@snowy.squish.net> <b8lj2q$l14$1@newshost.accu.uu.nl>
Message-ID: <m3fzo1v561.fsf@ferret.object-craft.com.au>

>>>>> "Martijn" == Martijn Faassen <m.faassen at vet.uu.nl> writes:

Martijn> Doesn't jonpy allow the Evil User to manipulate the type of
Martijn> your Python variables too? The Evil User could get rid of the
Martijn> '*' in the form and submit, right? Though I grant it looks
Martijn> like the code will always fail in that case, which is
Martijn> good. It's a neat idea.

Martijn> Formulator tries quite hard to avoid the Evil User
Martijn> manipulating HTML forms problem as well, by the
Martijn> way. Validation information is not sent to the client but
Martijn> kept on the server, and requests are interpreted by it.

Albatross solves this yet another way.  As input tags are processed by
the template interpreter the form tag accumulates a summary of all the
input fields.  When the form is closed the summary is then pickled,
MD5 signed with a secret, compressed, and base64 encoded in a hidden
field.

When the browser submits a request from the form the data in the
hidden field is restored and used to interpret the rest of the
request.  This is good enough for casual use.

It can still be defeated by manually crafting a request which does not
include the hidden field.  There are simple enhancements that can be
made to the request handling which simply rejects requests which do
not contain the hidden field.

- Dave

-- 
http://www.object-craft.com.au



From bokr at oz.net  Tue Apr  1 01:21:16 2003
From: bokr at oz.net (Bengt Richter)
Date: 1 Apr 2003 06:21:16 GMT
Subject: regular expression help (was GUI re builder)
References: <3E88647D.6010008@motorola.com> <b69rg8$p2k$1@slb2.atl.mindspring.net> <3E887F3D.1030107@motorola.com>
Message-ID: <b6bb4s$fkr$0@216.39.172.122>

On Mon, 31 Mar 2003 11:47:41 -0600, Stephen Boulet <stephen.boulet at motorola.com> wrote:

>Matt Gerrans wrote:
>> There should be an redemo.py in the tools/scripts directory, under Python22.
>
>Thanks Matt.
>
>I have my date-matching re (for incidences of 3/31/03 or 12/3/2003, for example):
>
>s = '3/31/03 some text 12/3/2003 some more text.'
>p = re.compile('\d+/\d+/\d+')
>s = p.sub('\n\n', s)
>
>Now, I want to precede each match with two newlines. How do I do this, instead 
>of replacing each match like above?
>
Do you mean that your
    s = '3/31/03 some text 12/3/2003 some more text.'
should become just have \n\n inserted, to become equal to
    s2 = """

3/31/03 some text 

12/3/2003 some more text."""
?

If so, perhaps this would serve:

 >>> import re
 >>> rx = re.compile(r'(\d+/\d+/\d+)')
 >>> s = '3/31/03 some text 12/3/2003 some more text.'
 >>> rx.sub('\n\n\\1',s)
 '\n\n3/31/03 some text \n\n12/3/2003 some more text.'
 >>> print rx.sub('\n\n\\1',s)


 3/31/03 some text

 12/3/2003 some more text.

Regards,
Bengt Richter



From bdesth at removethis.free.fr  Wed Apr 30 19:14:35 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 30 Apr 2003 23:14:35 +0000
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com> <mailman.1051630724.14928.python-list@python.org>
Message-ID: <3EB058DB.8070502@removethis.free.fr>

Sean 'Shaleh' Perry wrote:
> On Tuesday 29 April 2003 07:38, Christopher Blunck wrote:
> 
>>Hi all-
>>
>>This is my second generic question posted to c.l.p (the first was 'How
>>many of you are Extreme Programmers?') but it seems like a somewhat
>>appropriate discussion for the group.
>>
>>I've noticed in my development that Python programmers (in general)
>>write more unit tests than their java counterparts.  If I download a
>>project off of jakarta.apache.org, it very rarely has unit tests.
>>However, when I download a python module, it frequently has unit
>>tests.
>>
> 
> 
> I believe this is due in part to Python being an interpreted language.  With 
> Java you have a compiler thereto catch many problems but with Python most
> bugs, misfeatures, etc. can only be found by exercising the code. 

Java has the same mixture of compilation/interpretation than Python. The 
only differences are that :
- the python runtime (== VM) byte-compile the source code by itself if 
necessary,
- Python has dynamic typing where Java has static typing

So with Java you have to manually call the compiler, which may refuse to 
compile because an object is not *declared* to implement a given 
interface even if it does. And that's the only 'bug' (may be one, may be 
not) the Java compiler may find for you.

In both cases, you *need* unit test to catch *real* bugs.

> Python 
> coding is all about handling the runtime which makes for rapid writing of 
> code but almost requires stronger testing.

Ok for one point : a given object does not implement the interface it 
should. The Java compilers ensure the interface is *formally* 
implemented (which, IMHO, doesn't buy you that much).

(snip the rest)

Laotseu




From roy at panix.com  Mon Apr 28 15:08:42 2003
From: roy at panix.com (Roy Smith)
Date: 28 Apr 2003 15:08:42 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <s4uqavg0fjuucqve2vl4g03atpu25pdn97@4ax.com>
Message-ID: <b8ju7q$ij3$1@panix2.panix.com>

In article <s4uqavg0fjuucqve2vl4g03atpu25pdn97 at 4ax.com>,
Dale Strickland-Clark  <dale at riverhall.NOTHANKS.co.uk> wrote:
>Frantisek Fuka <fuka at fuxoft.cz> wrote:
>
>>When I try to call methods, I sometimes forget to include the 
>>parentheses. Instead of:
>>
>>if object.isGreen():
>>	do something...
>>
>>i sometimes write:
>>
>>if object.isGreen:
>>	do something...
>>
>
>It's the curse of Visual Basic.

I thought it was a perl-ism.  You can use ()'s, or leave them out.
You can use & or leave it out too.  Sometimes.  Maybe.  Depends if
it's a built-in or not, I think.  Ugh.

>I suggest you take your frustration out on the person who persuaded
>you to learn VB in the first place.

Same comment goes for perl.

>Recovery may be found in hours of coding in just about any grown-up
>language.

Ditto, but the recovery time may be longer than just a few hours.



From martin at v.loewis.de  Fri Apr 25 17:52:18 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 23:52:18 +0200
Subject: GUI libs
In-Reply-To: <3ea9ac4c$0$179$a1866201@newsreader.visi.com>
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com> <3ea947e1$0$198$a1866201@newsreader.visi.com> <mailman.1051301620.25438.python-list@python.org> <3ea9ac4c$0$179$a1866201@newsreader.visi.com>
Message-ID: <b8camh$g65$05$1@news.t-online.com>

Grant Edwards wrote:
> But those C runtime library routines aren't redundant. they're
> providing functionality that Python doesn't implement. 

They are *also* providing functionality that Python
doesn't implement. There is quite some redundancy, too. E.g.
a Posix C library supports regular expressions, which Python
also does, and Python doesn't use the C implementation.

 > OTOH, having both Tcl and Python interpreters seems a bit redundant.

"A bit" redundant would be fine, I think.

> I have used Scheme with Tk bindings, and that didn't use Tcl.
> The Scheme interpreter was bound directly to the Tk widgets.
> That results in several advantages to the application
> programmer, but disadvantages to the binding maintainer. 

What are those advantages? Aren't they also obtained by exposing
the Tcl object interface?

Regards,
Martin




From Henrik =?iso-8859-1?b?SORya/ZuZW4i?= <radix at kortis.to>  Wed Apr 23 10:28:04 2003
From: Henrik =?iso-8859-1?b?SORya/ZuZW4i?= <radix at kortis.to> (Henrik =?iso-8859-1?b?SORya/ZuZW4i?= <radix at kortis.to>)
Date: Wed, 23 Apr 2003 17:28:04 +0300
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz>
Message-ID: <pan.2003.04.23.14.28.04.640479@kortis.to>

On Wed, 23 Apr 2003 15:44:57 +0200, Frantisek Fuka wrote:

> I want a simple test to see if str contains any occurences of substr. I 
> am currently using this:
> 
> import string
> if string.find(str,substr) + 1:
>      ...do something...
> 
> But it somehow looks "ugly" to me. Is there more "standard" way to test 
> this?

Well, there are number of ways I guess, but after cleaning up a bit it 
could look like this:

if str.count(substr) > 0:
    do_stuff()

and where the 'str' is the string variable wanted to be checked of
occurances of 'substr', using string-methods. So you don't have to
import string module, but of course, if using old interpreter, you must do
it the old way.


Henrik



From sjmachin at lexicon.net  Wed Apr 30 17:03:27 2003
From: sjmachin at lexicon.net (John Machin)
Date: 30 Apr 2003 14:03:27 -0700
Subject: Removing last line while appending to file
References: <20030430152608.3c51bc77.twegener@fastmail.fm> <20030430164323.256ee71a.twegener@fastmail.fm> <3EAFB754.41CE2258@engcorp.com>
Message-ID: <c76ff6fc.0304301303.969955a@posting.google.com>

Peter Hansen <peter at engcorp.com> wrote in message news:<3EAFB754.41CE2258 at engcorp.com>...
> Tim Wegener wrote:
> > 
> > Someone emailed me directly with the solution: use open(filename, "r+")
> > This solved my problem.
> > (I was using the wrong mode for open())
> 
> Maybe mode "rb+" would be a little better for you, if these needs
> to be used on Windows or to be portable to Windows.

Generally seeking is very dodgy in text mode, but it *seems* to work
in this case on Windows (2000). Using binary mode to create files that
need to be read as text files by other software is a programming
nuisance -- see "outsep" in the following short script which attempts
to be portable and demonstrates using text or binary mode. Works OK on
Windows 2000. Not tested on any other OS.

HTH,
John
8<---
import os, sys
tag = "</log>"
taglen = len(tag)
def add_lines(fname, bin=1):
   f = file(fname, ["r+", "rb+"][bin])
   outsep = ["\n", os.linesep][bin]
   f.seek(-taglen-len(os.linesep), 2)
   guff = f.read(taglen)
   print repr(guff)
   f.seek(-taglen, 1)
   for k in range(3):
      f.write("added %d%s" % (k, outsep))
   f.write(tag + outsep)
   f.close()
if __name__ == "__main__":
   add_lines(sys.argv[1], int(sys.argv[2]))
8<---



From me at privacy.net  Sun Apr 20 02:14:16 2003
From: me at privacy.net (Kyoskyu)
Date: Sat, 19 Apr 2003 23:14:16 -0700
Subject: debugging fileinput
Message-ID: <pan.2003.04.20.06.14.16.600197@privacy.net>

I'm using fileinput to go through the lines in a file using something
like:

import fileinput, string
for line in fileinput.input():
	process(strip(line))

However, for some reason, this snippet of code gets to the end of the file
and starts again at the beginning.  I'm at a loss as to how to debug this
problem.  If someone has any ideas or knows the reason, I'd appreciate the
help.  Thanks.



From a.schmolck at gmx.net  Wed Apr 16 17:24:39 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 16 Apr 2003 22:24:39 +0100
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr>
Message-ID: <yfsu1cy9luw.fsf@black132.ex.ac.uk>

laotseu <bdesth at removethis.free.fr> writes:

> Paul Foley wrote:
> > On Tue, 15 Apr 2003 23:49:37 +0000, laotseu  wrote:
> >
> >>Even for OO programmers, functionnal features in Python are IMHO a
> >>great plus, and BTW functionnal and OO paradigm does not have to
> >>conflict (that would be functionnal vs imperative and object vs
> >>procedural). CLOS is one of the great system objects out here, and
> >>it's been implemented on top of a functionnal language,
                                      ^^^^^^^^^^^^^^^^^^^^

You go and say that in comp.lang.lisp :)

'as



From achrist at easystreet.com  Tue Apr 29 19:44:29 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Tue, 29 Apr 2003 16:44:29 -0700
Subject: Installer Problem WinNT to Win 98
References: <3EADD5C2.ABC48D5E@easystreet.com> <Xns936C27AC56317JanC@213.224.134.81> <3EADE94B.379C63A0@easystreet.com> <TpDra.7013$lD4.74951@news-server.bigpond.net.au>
Message-ID: <3EAF0E5D.2F3449A6@easystreet.com>

Mark Hammond wrote:
> 
> 
> You really need a way to see the exception.  This could be as simple > as hacking your app to open a file and set sys.stderr and sys.stdout > to it.
>
> Or simply install Python on that machine and run from source code, 
> just to determine the error.
> 

Yes, I know I'm a victim of dumb optimism.

The machine in question is 550 miles from here, and I doubt that I
will return to it any time before I want to deliver this app.  I
was just traveling to see a friend over the weekend and wanted to show
off my work. 

If I had just left the app the way that it was running here, with
the console on, I'm sure I would have found out what the problem
was.  Unfortunately, I turned it off to make it look slicker. 
Eventually we'll find out.

Based on the point where the app failed, I'm about 75% sure that it
had something to do with either the wxWindows feature of being able
to plant widgets in the middle of the wxHtmlCtrl or with accessing
the registry, as it failed right about where it should have been
filling a combo box in the middle of a wxHtmlCtrl with values from
the registry accessed via the win32api that comes with python. 

I'm not sure whether the registry access is done through advapi.dll or
oleaut32.dll, but windows does have several versions of these.  When
looking through google for info on these, I have seen that some 
software vendors do mention that the original version of one of these
that came with Win98 is not sufficient for their app -- it must be
updated with a download from MS.  I wonder if there is any way that it
might likewise need to be updated to run a python app(?).


Al



From rs55862 at NOSPAMyahoo.com  Tue Apr  8 08:08:11 2003
From: rs55862 at NOSPAMyahoo.com (Rami Saarinen)
Date: Tue, 08 Apr 2003 15:08:11 +0300
Subject: Making python module using Borland C++ 5
In-Reply-To: <3e92b527$0$32016$edfadb0f@dread12.news.tele.dk>
References: <b6u7fp$d2c$1@news.cc.tut.fi> <3e92b527$0$32016$edfadb0f@dread12.news.tele.dk>
Message-ID: <b6ue3b$hfb$1@news.cc.tut.fi>

Anders J. Munch wrote:

>>I am trying to build a simple module for python (2.1.1) using the 
>>Borland C++ Builder 5.
> 
> 
> Piece of cake with distutils, assuming you're just writing an
> ordinary extension module, and not trying to embed Python in your
> application.
> 
> Just write a setup.py and run it with --compiler=bcpp.

Ah, the bcpp option is good to know. Thanks. I did build the module with 
VC++ using the "python setup.py  build". However the problem is that 
I'll be using modified Python that includes custom Python21.lib and the 
include-files. At this point I am just trying to get it work with normal 
Python installation so that I know I am doing it right before trying to 
do it on the custom system. Thus I need to learn how to do it by hand.

I have compiled the standard CLIPS 6.2 as a static library which I 
include along with the python21.lib.

--
Rami Saarinen




From ark at research.att.com  Mon Apr 28 00:36:24 2003
From: ark at research.att.com (Andrew Koenig)
Date: Mon, 28 Apr 2003 04:36:24 GMT
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <3EAC2962.2050609@soraia.com>
Message-ID: <yu99he8j1bnb.fsf@europa.research.att.com>

Joe> The Deitels have a textbook called _Python How to Program_ that
Joe> is aimed at beginning programmers.

Aimed at beginning programmers?  I certainly didn't get that impression.
You can look at the table of contents here:

        http://www.deitel.com/books/pythonHTP1/pythonHTP1_toc.pdf

and make up your own mind.

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From garys at ihug.com.au  Mon Apr  7 18:24:23 2003
From: garys at ihug.com.au (Gary Stephenson)
Date: Tue, 8 Apr 2003 08:24:23 +1000
Subject: 'Catastrophic failure's trying to use OCX
Message-ID: <b6stun$mcf$1@lust.ihug.co.nz>

I am trying to interface to an .OCX file.  After successfully running makepy
over it, this is what happens:

ActivePython 2.2.2 Build 224 (ActiveState Corp.) based on
Python 2.2.2 (#37, Nov 26 2002, 10:24:37) [MSC 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import win32com.client
>>> x1 = win32com.client.Dispatch('PosEft.PosEftCtrl.1')
>>> `x1`
'<win32com.gen_py.PosEft ActiveX Control module._DPosEftMethods>'
>>> x1.ApiVersion
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "C:\activePython22\Lib\site-packages\win32com\client\__init__.py",
line 3
47, in __getattr__
  File "C:\activePython22\Lib\site-packages\win32com\client\__init__.py",
line 3
41, in _ApplyTypes_
pywintypes.com_error: (-2147418113, 'Catastrophic failure', None, None)
>>>

Exactly the same thing occurs if I use late-binding instead (via
win32.com.client.Dispatch.dynamic('....').

What am I doing wrong?

many tias,

gary






From nagylzs at freemail.hu  Thu Apr  3 13:37:24 2003
From: nagylzs at freemail.hu (=?ISO-8859-1?Q?Nagy_L=E1szl=F3_Zsolt?=)
Date: Thu, 03 Apr 2003 20:37:24 +0200
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org>        <b6fj57$626$1@news-reader11.wanadoo.fr>        <3E8B53BA.DBB0C84C@engcorp.com>        <mailman.1049319846.7316.python-list@python.org>        <3E8B67AD.7AD74C0C@engcorp.com> <32609.143.166.224.133.1049392052.squirrel@iamlearning.com>
Message-ID: <3E8C7F64.7020208@freemail.hu>

>
>
>>danny at iamlearning.com wrote:
>>    
>>
>>>I want my users run *.pyo or *.pyc files instead of *.py.
>>>      
>>>
>>To expand on Jp's response a little: please explain why you think
>>you want this, so we can better understand the need that's driving it.
>>Maybe you are trying to do something you shouldn't be trying
>>to do, or maybe you're just going about it the wrong way.  What
>>makes you think doing what you are asking is a good thing?
>>
>>-Peter
>>    
>>
>
>I am working in test environment, and I didn't like testers alter my
>script (this happened to me). Basicly, I want my scripts are under control
>with my knowledge. If the problem come up, I know where to start
>troubleshooting.
>  
>
You should sing your code in a way. E.g. with M2Crypto or PyCrypto. You 
can keep your private key's
password. You can re-sign your files when you distribute them to testers 
and you can check anytime if they
changed something. Note: you should include a last modification date in 
your file to be totally secure
because they are able to replace a file/signature pair with an older 
pair of the same file.

  Laci 1.0

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030403/f122006d/attachment.html>

From carroll at tjc.com  Tue Apr 29 16:33:45 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Tue, 29 Apr 2003 13:33:45 -0700
Subject: Determining the volume ID for a file/drive?
References: <uv7savkms6lsgchcv6jh8ahtccu9lon3oi@4ax.com> <8ytt4utc.fsf@python.net>
Message-ID: <94otavg1a3s8tlsgd2pfsjkfe7gtcd31dg@4ax.com>

On 29 Apr 2003 09:38:39 +0200, Thomas Heller <theller at python.net>
wrote:

>win32api.GetVolumeInformation("c:\\")

Excellent!  Thanks.

Funny, I grepped through my Python22/Docs directory looking for likely
suspects (including the string "volume"), and it didn't find anything;
now I see that PythonWin32Extensions.chm documents
win32api.GetVolumeInformation().

Thanks to all other who replied as well.




From ckea25d02 at sneakemail.com  Tue Apr  8 00:07:52 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 16:07:52 +1200
Subject: Calling python from a C function that is embedded in python?
In-Reply-To: <mailman.1049722812.17109.python-list@python.org>
References: <mailman.1049722812.17109.python-list@python.org>
Message-ID: <b6thcv$8d30k$1@ID-169208.news.dfncis.de>

VanL wrote:
> In short:  I want python to be able to call wrapped functions like they 
> are python functions, and a wrapped C library to call wrapped python 
> functions like they were in C.
> 
> Is there any way to do this?

Yes, there are ways of doing this. I have a demo showing
how to do it with Pyrex, but I don't have it here right
now. I'll try to remember to let you know when I've
posted it somewhere.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From max at alcyone.com  Fri Apr 25 00:39:01 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 24 Apr 2003 21:39:01 -0700
Subject: variable scoping problem
References: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com>
Message-ID: <3EA8BBE5.433AFD4C@alcyone.com>

Tung Wai Yip wrote:

> What rules in Python explain the scoping of x? Is there a name for the
> scope of the first x? 'module' scope? Why is bar() not able to access
> x but foo() can?

Python has two scopes, local and global (really, module-level).  The
Python interpreter is pretty smart about which one you mean, but since
Python is fully dynamic, it can't always be sure.  When it's getting
confused, use the global statement:

	x = 1
	def bar():
	    global x # "I mean the global x, not a local one"
	    x = 2
	    print x

One strictly only needs the global statement in cases where Python would
have otherwise guessed the wrong thing or would generate an error.  My
personal style is to always use the global declaration whenever I'm
referencing a global variable in a scope, which is something I do rarely
enough anyway.  That way when code changes, I don't have to worry about
adding global statements or get exposed to any unpleasant surprises.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Love, the itch, and a cough cannot be hid.
\__/ Thomas Fuller, M.D.
    Bosskey.net / http://www.bosskey.net/
 A personal guide to online multiplayer first person shooters.



From staschuk at telusplanet.net  Mon Apr  7 23:09:03 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 7 Apr 2003 21:09:03 -0600
Subject: [Slightly OT]: More on ints and floats
In-Reply-To: <fjss6b.1021.ln@boundary.tundraware.com>; from tundra@tundraware.com on Mon, Apr 07, 2003 at 10:03:53PM +0000
References: <fjss6b.1021.ln@boundary.tundraware.com>
Message-ID: <20030407210903.B4539@tibia.amotlpaa.bogus>

Quoth Tim Daneliuk:
  [...]
> As I understand it, integers and floats are distinct mathematical
> entities. A colleague of mine claims, that insfar as we use them in
> computing, ints are merely a proper subset of floats. [...]

Not if both use the same amount of storage.

For example, with four decimal digits to work with, the usual
unsigned int representation can represent 9999.  But with a
floating point representation in which one digit is dedicated to
the exponent, 9999 cannot be represented; it is between the two
representable numbers 9.99e3 and 1.00e4.

> [...] He furthermore
> asserts that (again as regards to computing) the distinction between
> them was made as a purely as a practical matter because floating point
> arithmetic was historically computationally expensive. He argues that
> any place one can use an int, these days (with cheap FP hardware), one
> could use a float 0 extended to the precision of the machine and get
> equivalent computational results. [...]

This may be true if you assume that overflow and underflow never
occur, and that error never accumulates to significant proportions.

But it is easy to construct counterexamples.  Working again with
my notional four-decimal-digit computer, unsigned ints can
correctly compute
	9980 + 0017 = 9997
while the floating point representation has rounding error for the
same computation:
	9.98e3 + 1.70e1 = 9.99e3
(or 1.00e4, depending on rounding mode).

Perhaps your friend's notion of a "practical programming problem"
is one in which the precision needed is much less than the
precision available.  (Lots of problems do indeed have this
property; but lots do not.)

  [...]
> As a matter of 'pure' mathematics, I argued that ints and floats are
> very different critters. My argument (which is no doubt formally very
> weak) is that the integer 3 and the float 3.0000 are different because
> of the precision problem.  For instance, the integer 3 exists at a single
> invariant point on the number line, but 3.0000 represents all numbers
> from 2.99995 through 3.00004.
> 
> Could one of the genius mathematicians here bring some clarity to this
> discussion, please?

I'm hardly a genius mathematician, but here's a few comments.

First, I disagree that the int 3 represents exactly the integer 3.
If int 3 is the result of wrapping int addition, it can be said to
represent some integer congruent to 3 modulo 2^w (where w is the
word size in bits).  If the int 3 is the result of a floor
division (that is, round-to-negative-infinity division of ints)
then it can be said to represent some number in [3,4).  (This is
exactly analogous to rounding error with floats.)

There are at least two ways to approach these situations formally.

One way is to describe computer addition, etc., as operators
directly on real numbers.  For example, you could define the int
addition operator (+) by
	x (+) y := (floor(x) + floor(y)) mod 10^4
In this way of thinking, the int 3 and the float 3.0 are not
distinct; the difference is actually between the operators 'int
addition' and 'float addition'.

Another way is to describe the computer values as entities
distinct from real numbers, and to define computer addition, etc.,
as operators on these entities, with a separate function which
maps from these entities to real numbers.  In this way of
thinking, ints and floats may be different kinds of entities.
(The operators are also distinct in this view.)

In any case, the important practical and theoretical question is
this: if you do a computer calculation and get a value x, what can
you infer about the number you would have gotten if you'd done
real arithmetic?  For such questions, whether you use fixed-point
or floating-point representations, what precision the
representations offer, and so forth, are vitally important points.

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From theller at python.net  Fri Apr 11 08:44:50 2003
From: theller at python.net (Thomas Heller)
Date: 11 Apr 2003 14:44:50 +0200
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com> <b766fi$p3b$1@newstree.wise.edt.ericsson.se> <16411664.REEpLcBOk5@lothar.cengija.com>
Message-ID: <he95yzi5.fsf@python.net>

Davor Cengija <dcengija_remove_ at inet.hr> writes:

> I wrote a simple python program with gui for a friend of mine and his 
> projects at local university. His co-workers didn't want to use it since 
> 'it's too small, obviously it doesn't do what we need'. It's about 45k of 
> python source. Then I created self-standing exe with mcmillan's installer. 
> It's now about 7mb in size. Now it looks 'respectable' and 'worth of trying 
> but not for serious work since it doesn't have setup.exe with EULA and 
> other things -- so it isn't a real program'. Now I'm looking for a free 
> install-shield like software.
> 
NSIS or Inno Setup.

> So, it isn't the quality of a software what only counts. Actually, I'm not 
> sure that the quality is the most important thing these days.

That reminds me of the first installer programs I used - they had an
option to slow down the installation: to impress the user.

Thomas



From gry at ll.mit.edu  Tue Apr 22 14:52:18 2003
From: gry at ll.mit.edu (george young)
Date: Tue, 22 Apr 2003 14:52:18 -0400
Subject: unused args in GUI callbacks: pychecker
Message-ID: <20030422145218.6e9e6549.gry@ll.mit.edu>

[python-2.3a1, pychecker-0.8.12]

I've started using pychecker a lot.  It has guided me to substantially
improve my code.  BUT: the "Parameter (event) not used" messages
for GUI(pygtk) callbacks are driving me crazy.  I could turn off 
*all* "parameter unused" messages, but some might be useful. 

So, is there a way to define a python(member) function and tell
python to ignore particular arguments? [so that pychecker could
ignor them too...]  

I recall that C++ had a way to tell the compiler this.

-- George

-- 
 I cannot think why the whole bed of the ocean is
 not one solid mass of oysters, so prolific they seem. Ah,
 I am wandering! Strange how the brain controls the brain!
	-- Sherlock Holmes in "The Dying Detective"



From cben at techunix.technion.ac.il  Tue Apr 15 10:17:59 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Tue, 15 Apr 2003 17:17:59 +0300 (IDT)
Subject: Speeding up: s += "string"
In-Reply-To: <20030415135950.GA8115@hishome.net>
References: <mailman.1050393203.4704.python-list@python.org>
 <Pine.GSO.4.51.0304151423510.2075@techunix.technion.ac.il>
 <20030415135950.GA8115@hishome.net>
Message-ID: <Pine.GSO.4.51.0304151706360.27142@techunix.technion.ac.il>

Oren Tirosh wrote on 2003-04-15:

> Guido is reluctant to add major new code to the code unless it gets some
> serious real-world testing as an extension module. But in the case of
> strings there is a problem: you can't create an object that will be a
> drop-in replacement for a string in Python unless it's actually derived
> from Py_StringObject and has the same internal structure. It can be an
> object that behaves much like a string and it can have an __str__ method
> to convert it into a string - but it will not be a compatible with a
> string in many situations. For example, it will not be accepted as an
> argument to string methods. It's not enough to make an object that can
> be converted into a string (almost any Python object can). It needs to
> *be* a string without sharing the same internal structure.
>
I bumped into this once while trying to create really transparent
proxies logging all accesses.  Many primitives written in C not only
require arguments to be isinstance() of some primitive class (string,
list, etc.) but also dirtily access its internal by-passing
Python-land customizations.  This means subclassing built-in types
doesn't really work.  At best I missed logging of some accesses, at
worst I got outright bugs.

> Such objects could be derived from the basestring class without being a
> subclass of either str or unicode. The only method they must implement
> is __str__. All other string methods and operations will be emulated by
> calling __str__ and passing the result to the original string method.
> The object may optionally override other methods to provide a more
> efficient implementation.

Seems a good approach to me.

> With this kind of infrastructure in place it
> would be possible to experiment with mutable strings, lazy concatenation,
> shared-buffer strings, etc as extension modules. The best solution could
> be considered for eventual integration into the core.
>
> It would require quite a lot of changes (ParseTuple, almost all the
> places that call PyString_Check, etc) but I think the Python code might
> actually benefit from it and become smaller and cleaner.  Currently
> there are a lot of places with hard-wired check for string,/unicode and
> objects exposing the buffer interface. These could be handled in a more
> generic way.
>
I'm +1 on anything that will extend the Pythonic way of informal
interfaces and "type checking by behavior" to interactions with
primitives.  I don't know the C code enough to comment on how to do it
:-).

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From gtalvola at nameconnector.com  Thu Apr 10 15:34:10 2003
From: gtalvola at nameconnector.com (Geoffrey Talvola)
Date: Thu, 10 Apr 2003 15:34:10 -0400
Subject: ssl error "PRNG not seeded"
Message-ID: <61957B071FF421419E567A28A45C7FE59AF3B0@mailbox.nameconnector.com>

Geoffrey Talvola wrote:
> ...
> When my 
> service started, it
> couldn't make any SSL requests.  My logfile reveals that 
> every time it tried
> to call socket.ssl(sock) the following exception was raised:
> 
> sslerror: (1, 'error:24064064:random number 
> generator:SSLEAY_RAND_BYTES:PRNG
> not seeded')
...

I'm closer to figuring this out.  Here's what I've determined:

1) OpenSSL's RAND_poll() gets called the first time it needs some random
bytes.

2) On Windows, RAND_poll() is implemented in crypto/rand/rand_win.c and
basically gets a bunch of unpredictable stuff from various Win32 API calls
such as the CryptoAPI, various Performance counters, timer data, processes,
threads, modules, heap, etc. to build up randomness.

3) For some reason, on the occasion when I experienced the exception,
RAND_poll() failed to generate enough randomness.  OpenSSL doesn't let you
proceed with an inadequately seeded PRNG so all SSL calls fail.

4) I'm not sure if this was a chance failure, or if it could have been
exacerbated by the fact the the service was starting due to a reboot.
Perhaps some of the calls that RAND_poll() was making to try to gather
randomness failed because the services that provide those calls weren't
started yet.  Hard to tell.

I think I'll report this as a bug in OpenSSL (although I don't know how to
fix it).

A workaround would be for me to call socket.RAND_add() with some chunk of
pseudo-random bytes, to fool OpenSSL into thinking it has enough randomness.

- Geoff




From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr  4 12:24:17 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 4 Apr 2003 17:24:17 GMT
Subject: What are the differences between _weak_ references and "normal" references?
References: <169a7283.0304040901.4bc3531b@posting.google.com>
Message-ID: <slrnb8rfu0.rs.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <169a7283.0304040901.4bc3531b at posting.google.com>,
sdieselil wrote:
> Hi
> 
> I noticed a module weakref in Python standard library. What is weak
> reference and in what way it differs from "normal" reference? By the
> way, how can I implement usual reference in Python (I mean C++
> reference)?

All Python variables are references, but unlike C++ references they can
be rebound.  In Python, copying objects is unusual.

In a garbage-collected environment like Python, a weak reference is one
that does not prevent its target from being destroyed, but is instead
nullified (in Python, set to None) when the target is destroyed.



From jerry.seutter at pason.com  Fri Apr 11 11:52:32 2003
From: jerry.seutter at pason.com (Jerry Seutter)
Date: Fri, 11 Apr 2003 09:52:32 -0600
Subject: string.split() documentation incorrect?
Message-ID: <20030411095232.A8283@ash.int.ca.pason.com>

I'm trying to use string.strip to remove the comma off the end of 
a string, eg:

string.strip("id INT NOT NULL,", ",")

The current documentation says I can do this.  Python 2.2.2 gives
the following error:

>>> strip('  hello  ', ',')
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: strip() takes exactly 1 argument (2 given)

So, I think either the documentation is incorrect or there is 
a bug.  Thoughts?

Jerry Seutter





From v451v at yahoo.com  Mon Apr 21 10:33:00 2003
From: v451v at yahoo.com (Jorl Shefner)
Date: Mon, 21 Apr 2003 10:33:00 -0400
Subject: _socket problem
Message-ID: <b80ves$52ova$1@ID-181029.news.dfncis.de>

I am unable to use the urllib module because of the error:
ImportError: No module named _socket

I have seen this issue reported in this newsgroup previously.  Some 
postings have indicated this is related to a compatibility issue with 
openSSL, though most of the details are beyond my ability to follow. 

My question is mainly focused on determining whether the problem likely 
resides in the python installation or some other program.

The important points:

   The system with the problem is:
   Unix version SGI IRIX 6.5
   It is running- OpenSSL 0.9.7a Feb 19 2003
   Python 2.2.2 (#2, Apr  9 2003, 11:54:07)

   I was recently running (on this machine) an earlier version of python 
(2.2, I believe) installed locally in my account and had no _socket 
error.  When I had the administrator install python system-wide, the 
problems began.

   I do much of my work on Mac OsX with very similar versions python and 
openSSL without any such problems.
   
   OS 10.2.5 
   OpenSSL 0.9.6i Feb 19 2003
   Python 2.2.2 (#1, 01/12/03, 07:51:34)

   Is there any known workaround for this problem, or does it require a 
re-installation of python.

   Thanks,

   J.S.



From tim.one at comcast.net  Mon Apr 28 10:29:34 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 28 Apr 2003 10:29:34 -0400
Subject: Forgetting "()" when calling methods
In-Reply-To: <cHadnX5vTdzoWDGjXTWcpw@comcast.com>
Message-ID: <BIEJKCLHCIOIHAGOKOLHCEKIFHAA.tim.one@comcast.net>

[Tim]
> I'm pretty sure that if Python had never supported meaningless
> (although consistent within a single run) comparisons, a compelling case
> for adding them couldn't have been made.

[Terry Reedy]
> Nor can I imagine a compelling case for exactly the current rules if
> they were not whay they currently are.  Allowing 1j to compare to '1'
> but not 1 is pretty exoteric.  The argument seem to be this: cmp(1j,
> '1') is so meaningless that it is ok to give an answer while cmp(1j,1)
> might be seen as meaningful (and lexicographically, it is) so therefor
> it must be forbidden.

No, the current state makes no sense.  In very early Pythons, it wasn't
*possible* (for technical reasons) for a comparison to raise an exception
(well, it could, but it got silently ignored).  When it became possible for
comparisons to raise exceptions, new types started doing so, but
pre-existing types generally didn't.  complex is a quirky exception.  It's
obvious that, in Python 3, Guido doesn't want

    2 < "42"

to pass silently any more than he wants

    2 + "42"

to pass silently.  The current state is an inconsistent mish-mash moving
toward that.





From frandebo at latt.if.usp.br  Mon Apr 14 13:22:11 2003
From: frandebo at latt.if.usp.br (Francisco Borges)
Date: Mon, 14 Apr 2003 14:22:11 -0300
Subject: parser stack size limit
Message-ID: <20030414172211.GA20921@latt.if.usp.br>

I'm getting (both in Python2.2 and Python2.3): 

"""
s_push: parser stack overflow
MemoryError
"""

When trying to run the following line:
a = eval("[('top_start_xp','start'),[[('max_xp(root)','max'),[[('non_wh_topicalization(np)','root'),[[('het_noun','np'),'Het'],[('optpunct(e)','optpunct'),[]],[('imp','sv1'),[[('verb(hebben,sg3,alsof_sbar_subj)','v'),'lijkt'],[('v2_vp_vproj','v2_vp'),[[('vpx_vproj','vpx'),[[('vp_v_komma_arg(sbar)','vproj'),[[('vproj_vc','vproj'),[[('vc_vb','vc'),[[('vb_v','vb'),[[('vgap','v'),[]]]]]]]],[('optpunct(e)','optpunct'),[]],[('sbar(vp)','sbar'),[[('alsof_complementizer','comp'),'alsof'],[('vp_vpx','vp'),[[('vpx_vproj','vpx'),[[('vp_arg_v(np)','vproj'),[[('np_pron_weak','np'),[[('wkpro_pronoun(nwh,thi,both,de,both,def)','pron'),'ze']]],[('vp_arg_v(np)','vproj'),[[('np_pron_weak','np'),[[('wkpro_pronoun(nwh,je,sg,de,both,def)','pron'),'je']]],[('vp_vp_colon','vproj'),[[('vproj_vc','vproj'),[[('v_v_v','vc'),[[('verb(hebben,pl,aux(inf))','v'),'willen'],[('vc_vb','vc'),[[('vb_v','vb'),[[('verb(hebben,inf,np_sbar)','v'),'zeggen']]]]]]]]],[('punct(dubb_punt)','punct'),':']]]]]]]]]]]]]]]]]]]]]]]]],[('optpunct(e)','optpunct'),[]]]]")

(the brackets *are* balanced...)

There are some bogus lists inside it I know, but as soon as I start
using this for more complicated sentences (linguistics being done here)
the number of nested lists will increase.

Any ideas on how to avoid this? Could it be as simple as a file to
modify and then recompile python? 

-- 
  Francisco.
             __o
            `\<,
     _____(*)/(*)_____




From bdesth at removethis.free.fr  Thu Apr 17 09:56:35 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Thu, 17 Apr 2003 13:56:35 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <7h3y92ap1uk.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3E9EB293.4070100@removethis.free.fr>

Michael Hudson wrote:
> laotseu <bdesth at removethis.free.fr> writes:
> 
> 
>>Paul Foley wrote:
>>
>>>On Tue, 15 Apr 2003 23:49:37 +0000, laotseu  wrote:
>>>
>>>
>>>>Even for OO programmers, functionnal features in Python are IMHO a
>>>>great plus, and BTW functionnal and OO paradigm does not have to
>>>>conflict (that would be functionnal vs imperative and object vs
>>>>procedural). CLOS is one of the great system objects out here, and
>>>>it's been implemented on top of a functionnal language,
>>>
>>>Oh yes?  Which functional language would that be?
>>>
>>
>>CLOS means Common Lisp Object System.
> 
> 
> Paul knows that :-)

He does. Some may not...

> I'd wager the point he was making is that Common
> Lisp isn't "really" a functional programming language...

Ho yes ? Why ? Because it doesn't enforce a strict functionnal 
programming style ? In this case, Python is not "really" an object 
oriented language !-)

AFAIK, Common Lisp support all the features that make the definition of 
a functionnal language, so it *is* a functionnal language.

> Ocaml would be a better example, I'd have thought.

Let's say another example...

Laotseu




From bitbucket at safe-mail.net  Fri Apr 25 07:35:22 2003
From: bitbucket at safe-mail.net (poiboy)
Date: 25 Apr 2003 04:35:22 -0700
Subject: xmlrpc introspection
References: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com> <af6b5bdd.0304240434.fd7e6ad@posting.google.com> <UR0qa.77049$gK.181217@rwcrnsc52.ops.asp.att.net>
Message-ID: <af6b5bdd.0304250335.17ce7cd1@posting.google.com>

> thanks!!!!  the xml-rpc debugger is awsome.

Happy to help. Quick note - the debugger link was courtesy of Graham
Dumpleton, posted earlier.

> i see a scalability problem with listMethods. we have a
> server that dynamically loads many providers.
> it would be nice if it was standard that there was a
> listServices request and listMethods took an optional
> service parameter. then you can scale this up.

Re: scalability - I'd vote for such a solution, where your server's
listMethods response includes listServices and methods particular to
your server (listServiceMethods, listMethodDirectory, etc.) rather
than aiming for listMethods transparency - if only to draw attention
to the varying degrees of energy required to fulfill a request. Of
course this is entirely dependent on client relationships.

Dictate. Become the standard.

A hui hou.



From michael at stroeder.com  Sat Apr 19 10:56:44 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Sat, 19 Apr 2003 16:56:44 +0200
Subject: ANN: python-ldap-2.0.0pre09
Message-ID: <mailman.1050926644.15847.clpa-moderators@python.org>

HI!

Find a new pre-release of python-ldap:

   http://python-ldap.sourceforge.net/

python-ldap provides an object-oriented API to access LDAP directory
servers from Python programs. It mainly wraps the OpenLDAP 2.x libs for
that purpose. Additionally it contains modules for other LDAP-related
stuff (e.g. processing LDIF, LDAPURLs and LDAPv3 schema).

Download link:
http://prdownloads.sourceforge.net/python-ldap/python-ldap-2.0.0pre09.tar.gz?download

Changes since 2.0.0pre08:

Modified setup.py to support Cyrus-SASL 2.x.

ldap.ldapobject:
* apply() is not used anymore since it seems deprecated
* Fixed __setstate__() and __getstate__() of ReconnectLDAPObject

ldap.schema:
* Completed classes for nameForms, dITStructureRules and
   dITContentRules






From revanna at mn.rr.com  Tue Apr 15 17:26:16 2003
From: revanna at mn.rr.com (Anna)
Date: Tue, 15 Apr 2003 21:26:16 GMT
Subject: Post-OOP: not just an "unfortunate acronym"?
Message-ID: <pan.2003.04.15.21.27.33.796814@mn.rr.com>

Interesting article.


Every generation has to reinvent the practice of computer programming. In
the 1950s the key innovations were programming languages such as Fortran
and Lisp. The 1960s and '70s saw a crusade to root out "spaghetti code"
and replace it with "structured programming." Since the 1980s software
development has been dominated by a methodology known as object-oriented
programming, or OOP. Now there are signs that OOP may be running out of
oomph, and discontented programmers are once again casting about for the
next big idea. It's time to look at what might await us in the post-OOP
era (apart from an unfortunate acronym).


http://www.americanscientist.org/Issues/Comsci03/03-03Hayes.html#fig3



From deja-google at wotevah.com  Tue Apr  8 02:28:37 2003
From: deja-google at wotevah.com (L.C.)
Date: 7 Apr 2003 23:28:37 -0700
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com>
Message-ID: <e0bf36b5.0304072228.4b47a93a@posting.google.com>

I would ask a different question. What exactly would be the BENEFIT 
of doing everything in FP ?

I believe programming usually deals with practical, performance/cost
issues, so discarding that to make a point does make the issue sort 
of academical anyway. 

Integers are more efficient in storage space as they do not have to
store a mantissa. They are also precise, as you mentioned. Floats 
have the bad habit of not fitting properly in our binary formats and
thus propagating errors even within their precision range (try 
representing 0.3 or 0.1).

After all, they make hammers in different sizes for a reason.

--
L.C. (Laurentiu C Badea)

Tim Daneliuk <tundra at tundraware.com> wrote in message news:<fjss6b.1021.ln at boundary.tundraware.com>...
> OK, I don't want to resurrect another interminable discourse in ints,
> floats, and how they are/ought/can be handled in the language. But, I
> have a sort of theoretical mathematical question which the whole
> business brought to mind.
> 
> As I understand it, integers and floats are distinct mathematical
> entities. A colleague of mine claims, that insfar as we use them in
> computing, ints are merely a proper subset of floats. He furthermore
> asserts that (again as regards to computing) the distinction between
> them was made as a purely as a practical matter because floating point
> arithmetic was historically computationally expensive. He argues that
> any place one can use an int, these days (with cheap FP hardware), one
> could use a float 0 extended to the precision of the machine and get
> equivalent computational results. Say the hardware supported 4 digits of
> precision. He is arguing that:
> 
> 
>     3/4.0000  is equivalent to 3.0000/4.0000
> 
> (Never mind the old Python modulo vs. divison debate.)
> 
> In effect he is saying that, unless there is a practical
> performance/cost issue at hand, there is no real reason to
> differentiate between ints and floats in practical programming
> problems.
> 
> As a matter of 'pure' mathematics, I argued that ints and floats are
> very different critters. My argument (which is no doubt formally very
> weak) is that the integer 3 and the float 3.0000 are different because
> of the precision problem.  For instance, the integer 3 exists at a single
> invariant point on the number line, but 3.0000 represents all numbers
> from 2.99995 through 3.00004.
> 
> Could one of the genius mathematicians here bring some clarity to this
> discussion, please?
> 
> TIA,



From nav at adams.patriot.net  Fri Apr 25 13:32:47 2003
From: nav at adams.patriot.net (Nick Vargish)
Date: 25 Apr 2003 13:32:47 -0400
Subject: Divided by a common language (was: OO + Python Developers Wanted)
References: <mailman.1051183297.29110.python-list@python.org> <yyyznmezekm.fsf@adams.patriot.net> <vais01o44vtc0d@corp.supernews.com>
Message-ID: <yyyvfx2zdmo.fsf@adams.patriot.net>

claird at lairds.com (Cameron Laird) writes:

> >Which "City" would that be?
> <URL: http://www.cityoflondon.gov.uk/ >

Oh, then he should have specified, "The City", and I would have known
he meant Mother London. Here in Amerika, we have a number of urban
locales that consider themselves "the City"...

Nick

-- 
#include<stdio.h> /* SigMask 0.3 (sig.c) 19990429 PUBLIC DOMAIN "Compile Me" */
int main(c,v)char *v;{return !c?putchar(*v-1)&&main(0,v+ /* Tweaks welcomed. */
1):main(0,"Ojdl!Wbshjti!=obwAqbusjpu/ofu?\v\1");}  /* build: cc -o sig sig.c */



From aleax at aleax.it  Wed Apr 23 18:02:13 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 22:02:13 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com> <e9570f37.0304220625.11312687@posting.google.com> <K2qdnXJ_ZeZMVTujXTWcqg@speakeasy.net>
Message-ID: <F3Epa.51023$LB6.1152908@news1.tin.it>

A.M. Kuchling wrote:
        ...
> Money does not make interested developers materialize,


Surely you're joking, Mr Kuchling?


Alex




From sholden at holdenweb.com  Sat Apr 19 10:00:28 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sat, 19 Apr 2003 14:00:28 GMT
Subject: wxPython 1.52 under Win95
References: <7d22d8d6.0304190231.4f6da5b1@posting.google.com>
Message-ID: <0Ecoa.134541$It5.18605@news2.central.cox.net>

[posted and mailed]


"ekranawetter_steyr" <ekranawetter at bmw.co.at> wrote in message
news:7d22d8d6.0304190231.4f6da5b1 at posting.google.com...
> Hi all,
>
> I tried to run wxPython 1.52 under Win95 (because the application I
> use is hard to port to version 2) and Python crashes on the command:
>
> from wxPython.wx import *
>
> This is my Python installation:
>
> - py152.exe
> - win32all-132.exe
> - a lot of files in \Python\wxPython (wx22_2.dll from 21st Oct 00)
> - win32com
> - win32
>
> How can I fix this problem? Please send an email also to
> ekranawetter at utanet.at
>
It looks very much like you've loaded wxPython for Python 2.2, which has no
chance of working with Python 1.5.2.

A search in www.wxpython.org and https://sourceforge.net/projects/wxpython/
didn't reveal any versions capable of working with 1.5.2. That's not to say
they don't exist. Take a look at those sites, and maybe consider mailing the
developers for assistance. If they ever did support 1.5.2 they may have an
installer lying round.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/






From lost at soul.net  Sun Apr 13 15:41:45 2003
From: lost at soul.net (lost)
Date: Sun, 13 Apr 2003 20:41:45 +0100
Subject: Help with CGI
Message-ID: <pan.2003.04.13.19.41.42.420658@soul.net>

Hi there, 
I'm trying to write a program which prints the html page, and when form
data has been submitted, it prints the html again, but with the keys and
values from the form.
here's the code i have so far:(bottom of post)

I can't get the "for key in form.keys():" part to run....
I'd also like to ONLY print the html if there are no values in the cgi, so
I'd need something like:
if form.keys == [""]: #if there are no keys submitted
	do this...

any help for what I need to do? 
cheers!


#!/usr/bin/python
print "Content-type: text/html\n\n"                                                                                                                                  
#get cgi data into a list...
import cgi
form = cgi.FieldStorage
                                                                                                                                   
print """
<html>
<head><title>Hello</title></head>
<body>
<form name=\"login_form\" method=\"post\" action=\"http://pertelotte/cgi-bin/test.py\">
<table width=\"85%\" align=\"right\" border=\"1\">
        <tr>
                <td width=\"40%\" align=\"right\"><font size=\"-2\">Enter your details:</font></td>
                <td align=\"right\"><font size=\"-2\">Username: </font><input type=\"text\" name=\"uname\" size=\"15\"></td>
                                                                                                                                   
                <td align=\"right\"><font size=\"-2\">Password: </font><input type=\"password\" name=\"pword\" size=\"15\"></td>
                <td colspan=\"2\" align=\"right\">
                <input type=\"submit\" name=\"submit\" value=\"Login\">
                </td>
        </tr>
</table>
</form>
<br><br>
<hr width=\"100%\">
"""
for key in form.keys():
        print key
                                                                                                                                   
print """
</body>
</html>
"""




From ekranawetter at bmw.co.at  Sat Apr 19 07:55:41 2003
From: ekranawetter at bmw.co.at (ekranawetter_steyr)
Date: 19 Apr 2003 04:55:41 -0700
Subject: Simple wxPython Application
Message-ID: <7d22d8d6.0304190355.6098714c@posting.google.com>

Hi all,

1.Problem
---------

I was not able to run this simple example from the tutorial:

from wxPython.wx import wxPySimpleApp, wxFrame
  app = wxPySimpleApp()
  frame = wxFrame(None, -1, "Hello World")
  frame.Show(1)
  app.MainLoop()

I got a syntax error in line 2


2. Problem
----------

Running wxPython on Win95 (previously posted) doesn't even run when I
add the script "sitecustomize.py" in the python installation directory
with following commands:

import sys
sys.path.append(wxPythonpath)


Please help and send an email also to ekranawetter at utanet.at

Thanks, Erwin



From joshua at goodish.org  Sat Apr  5 09:39:51 2003
From: joshua at goodish.org (Joshua Goodlett)
Date: 5 Apr 2003 06:39:51 -0800
Subject: Calling C libraries from a Python script
Message-ID: <e81e0f3d.0304050639.37b59655@posting.google.com>

I'm trying to find examples of calling C library functions from within
a Python script.  Any advice or pointers on where to look for examples
would be appreciated.

Cheers,
Joshua Goodlett



From mross1 at nycap.rr.com  Thu Apr 24 21:50:37 2003
From: mross1 at nycap.rr.com (M Ross)
Date: 24 Apr 2003 18:50:37 -0700
Subject: Errors during xmldiff install on xp
Message-ID: <b4002bf3.0304241750.7b2a840c@posting.google.com>

Getting the following error msgs when trying to install the xmldiff
module:

c:\Python22\include\pyconfig.h(37) : fatal error C1083: Cannot open
include file: 'io.h': No such file or directory

running 
> python setup.py install
from the xmldiff directory.

io.h is on my machine-I have VS.NET installed.  It originally was
unable to find mspdb70.dll.  That was solved by adding the dll
location to my PATH.   However, it doesn't work for io.h.

How do I configure my setup so the xmldiff install will find all
required files?

I am a python newbie...

Thanks for your time.
Mike



From glenfant at NOSPAM.bigfoot.com  Tue Apr 22 17:15:53 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Tue, 22 Apr 2003 23:15:53 +0200
Subject: Installing MySQL-python-0.9.2 on Solaris 2.9
References: <mailman.1051034858.6946.python-list@python.org>
Message-ID: <b84ald$k8k$1@biggoron.nerim.net>

"Eduardo Castro" <ecastro at calarts.edu> a ?crit dans le message de news:
mailman.1051034858.6946.python-list at python.org...
>
> Hi, has anyone had problems installing MySQL-python-0.9.2 on Solaris 2.9?
>
> I am getting the following error when I run 'setup.py build'
>
> ld: fatal: library -lmysqlclient: not found
>
>
> This is part of the error messages:
>
> building '_mysql' extension
> skipping _mysql.c (build/temp.solaris-2.9-sun4u-2.2/_mysql.o up-to-date)
> gcc -shared build/temp.solaris-2.9-sun4u-2.2/_mysql.o -L/usr/lib/mysql
> -L/usr/local/lib/mysql -L/usr/loca/mysql/lib -L/usr/local/mysql/lib/mysql
> -L/usr/local/mysql/include -L/usr/bin
> -Wl,-R/usr/local/lib:/usr/openwin/lib:/usr/dt/lib -lmysqlclient -lz -o
> build/lib.solaris-2.9-sun4u-2.2/_mysql.so
> ld: fatal: library -lmysqlclient: not found
> ld: fatal: File processing errors. No output written to
> build/lib.solaris-2.9-sun4u-2.2/_mysql.so
> collect2: ld returned 1 exit status
> error: command 'gcc' failed with exit status 1
>
>
> Thanks
>
> Eduardo Castro
> California Institute of the Arts.
>

Fell on this on a FreeBSD box

$ locate libmysqlclient

should find something like...

/usr/local/lib/mysql/libmysqlclient.so.10

but perhaps somewhere else for you.

Just cd to a place of your ld path (/usr/local/lib/mysql may be a good
place)

$ ln -s /usr/local/lib/mysql/libmysqlclient.so.10 libmysqlclient.so

If "locate" doesn't find anything, you must install MySQL client package
before building MySQL-python.

HTH

--Gilles




From lale at fotonation.com  Tue Apr 29 12:09:42 2003
From: lale at fotonation.com (Ilariu Raducan)
Date: Tue, 29 Apr 2003 17:09:42 +0100
Subject: win32 display jpg file
Message-ID: <irxra.10972$pK2.14342@news.indigo.ie>

Hi All,

Does anybody knows how to display a JPG file in pythonwin and draw lines 
over it?

Thank you,
Lale




From sholden at holdenweb.com  Sun Apr  6 13:13:58 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sun, 06 Apr 2003 17:13:58 GMT
Subject: negative values in profile output
References: <mailman.1049479748.24160.python-list@python.org> <7h38yunk17w.fsf@pc150.maths.bris.ac.uk>
Message-ID: <qfZja.18781$It5.10774@news2.central.cox.net>

"Michael Hudson" <mwh at python.net> wrote in message
news:7h38yunk17w.fsf at pc150.maths.bris.ac.uk...
> "Matthew M. Boedicker" <mboedick at mboedick.org> writes:
>
> > When using the profile module to profile a long-running program,
> > I am getting some negative values:
> >
> > 6753628 -4036.027   -0.001 -4036.027   -0.001
gobject_sql.py:36(<lambda>)
> >
> > Am I correct to assume these don't have special meaning and are the
> > result of something wrapping?
>
> They may be a result of mis-calibration of the profiler -- see the
> docs.
>
> Last time i used the profiler, I more-or-less ignored the values
> output and just assumed I was doing the right thing when they got
> smaller...
>

But when a number comes up negative, does "smaller" mean more negative or
closer to zero?

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From tim_one at email.msn.com  Mon Apr 14 00:46:54 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Mon, 14 Apr 2003 00:46:54 -0400
Subject: How to subclass datetime?
In-Reply-To: <ba1e306f.0304130850.1db6bf69@posting.google.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCMEACEHAB.tim_one@email.msn.com>

[Jane Austine]
> ...
> For example,
> 
> >>> class mydate(datetime.date):
> 	def abc(self):
> 		print "abc"
> >>> z=mydate(2003,4,14)
> >>> z.abc()
> Traceback (most recent call last):
>   File "<pyshell#6>", line 1, in -toplevel-
>     z.abc()
> AttributeError: 'datetime.date' object has no attribute 'abc'
> >>> type(z)
> <type 'datetime.date'>
>
> ...
>
> I reckon it is not subclassing the datetime.date class in
> its proper definition of subclassing.

Indeed it's not!  That wasn't intended.  I opened a bug report:

    http://www.python.org/sf/720908




From peter at engcorp.com  Tue Apr  1 14:04:30 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 14:04:30 -0500
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <x7of3qjh4g.fsf@guru.mired.org> <b6ckjr$903$1@panix2.panix.com>
Message-ID: <3E89E2BE.608144F9@engcorp.com>

Roy Smith wrote:
> 
> For example, let's say you're building up a list and at the end, you
> want to know how many elements you added to the list.  The obvious
> thing to do would be to keep track of how many elements you add as you
> build the list.  [...]
> 
> I just tried it in Python, and (not surprisingly), found that it's not
> true in Python either.  [...]

What you're saying is that it's not true that redundant code in Python 
runs faster than code that skips the redundant part...  I don't find
that surprising either.

Put another way, and as you found out after your experiment, one
should expect a language like Python, or really any decent collection 
object, to keep track of how many items it contains without needing to
do an expensive scan of the entire collection each time the length is
asked for...

-Peter



From exarkun at intarweb.us  Wed Apr  2 01:23:44 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 01:23:44 -0500
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304012011.78f05548@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com>
Message-ID: <20030402062344.GA4169@meson.dyndns.org>

On Tue, Apr 01, 2003 at 08:11:36PM -0800, Drew Moore wrote:
> Howdy! 
> 
> I submitted a pre-PEP on overloadable assignment. 

  -1e100

  Why?

  Assignment is not an object operator.  It's a name operator.

  "x = y" doesn't operate on the object "x" references or the object "y"
references, it operates on x and y.

  Jp

-- 
"The problem is, of course, that not only is economics bankrupt but it has
 always been nothing more than politics in disguise ... economics is a form
 of brain damage."  -- Hazel Henderson
-- 
 up 13 days, 2:00, 3 users, load average: 1.00, 1.00, 1.03




From abpillai at lycos.com  Fri Apr 25 08:08:12 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 25 Apr 2003 05:08:12 -0700
Subject: problem starting python interpreter in emacs (win98)
References: <V65pa.4289$AB5.573029@news02.tsnz.net> <b831pv$2d65$1@serv.vrn.ru> <Kq8pa.4362$AB5.585892@news02.tsnz.net>
Message-ID: <bd993a2f.0304250408.10f67426@posting.google.com>

I think you got it wrong there. Emacs uses 'load-path'
for looking up lisp packages only. For setting the path
to executables like python.exe, you need to set the 'exec-path'
variable.

;; in your .emacs file
(add-to-list 'exec-path "C:\\Python22")

If you do this, you need not modify the PATH
variable also.

ABP

selwyn <selwyn at home.net.nz> wrote in message news:<Kq8pa.4362$AB5.585892 at news02.tsnz.net>...
> Lexy Zhitenev wrote:
> > "selwyn" <selwyn at home.net.nz> wrote in message:
> > news:V65pa.4289$AB5.573029 at news02.tsnz.net...
> > 
> >>system: win98
> >>python version 2.2.2
> >>emacs  21.2.1
> >>..........
> >>hi all,
> > 
> > 
> > <snip>
> > 
> >>My .emacs file is as follows:
> >>
> >>(setq load-path (cons "C:\Python22" load-path))
> > 
> > 
> > 
> > Not sure about ELisp, but try to use escaped backslash:
> > 
> > (setq load-path (cons "C:\\Python22" load-path))
> > 
> > 
> > Regards, Lexy
> > 
> > 
> 
> cheers for the advice.
> however it doesn't seem to help my problem- Emacs seems to be able to 
> find the python-mode.elc ok, as evidenced by the syntax highlighting 
> working...
> 
> cheers again,
> Selwyn



From bdesth at removethis.free.fr  Sun Apr  6 10:59:27 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Sun, 06 Apr 2003 16:59:27 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <3E9040CF.2040606@removethis.free.fr>

Patrick K. O'Brien wrote:
> If you were crazy enough to think that you could create a program
> along the lines of Emacs, but written in Python, how would you go
> about it?  How would you design the domain model for files, buffers,
> windows, and frames?  How would you allow the same level of
> customizability?  How would you map functions (or methods or whatever)
> to keys?  Any thoughts?
> 
> Here is the background.  I'm working on a file editor (called
> PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> PyCrust code base.  I've got a basic version working, and it supports
> editing multiple files and all the usual, basic features.  But I want
> it to be as customizable and extensible as Emacs.  In fact, I want to
> model it after Emacs, so that elisp code could be rewritten in Python
> and work with PyAlaMode.  To do that, I've got to support the same
> primitives as Emacs; expose files, buffers, regions, and windows the
> same way Emacs does; handle the Emacs keybindings and ability to remap
> keys, the same way Emacs does, etc.
> 
> Sound like fun?
> 

Sounds like Real Fun !

Is there a way to get the current code ? (I was recently thinking of 
extending pycrust to make it a full programming environnement, something 
like a mix of Emacs and Smalltalk's programming environnement... Not 
sure if I'm able to do it, but...)

Laotseu




From jkraska at san.rr.com  Mon Apr 21 12:04:11 2003
From: jkraska at san.rr.com (Courageous)
Date: Mon, 21 Apr 2003 16:04:11 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <erk5avkh423eh8bc13ctvtbjqt9f6kt4kv@4ax.com> <84smscm5df.fsf@boost-consulting.com>
Message-ID: <pf58avcm8u8r84183tia6alfo8gctc4fdu@4ax.com>

>> actually the meat of what it is. In unix, theoretically one could
>> set up the heavy weight equivalent of a continuation by the use of
>> setjmp and
>
>Really?  I thought longjmp could only go back up the call stack,...

One *can* on some systems, although this isn't portable, and not
supported on many, and guaranteed on none. Hence in the part you
deleted: "although certain patterns of setjmp/longjmp aren't
always safe." :)

>       1. execution context is preserved
>       2. You can transfer control up, down, or between call stacks.

That's correct.

>Also, "stacklessness" is really an implementation detail AFAICT - it
>just says that all execution context is stored in
>dynamically-allocated structures instead of using local variables on
>the 'C' stack (via a recursive-calling implementation).
>
>Have I misunderstood this whole thing?

Nope. You've gotten it. "Stacklessness" as part of an implementation
detail essentially advertises certain probable features of the continuation
environment, but nothing more than whatever you decide is implied. :)

C//




From tim.one at comcast.net  Mon Apr 28 11:08:41 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 28 Apr 2003 11:08:41 -0400
Subject: Forgetting "()" when calling methods
In-Reply-To: <ge5ra.13872$3M4.378491@news1.tin.it>
Message-ID: <BIEJKCLHCIOIHAGOKOLHOEKLFHAA.tim.one@comcast.net>

[Tim]
>> OTOH, how many would have scored 100% if also asked what the result of
>> sorting would be?  I'm not sure anyone could have.  For years, ints came
>> out "less than" lists, which in turn came out "less than" longs, so that
>>
>>     [10, [9], 7L, 8]
>>
>> was considered to be "in sorted order" as-is.  It was something like 7
>> years before someone finally noticed how nuts that was.

[Alex Martelli]
> *Blink* -- I don't get it.  Even if int < list < long, how CAN the
> two int's here be ``considered in sorted order''???

They weren't, it was the list itself that was considered to be in sorted
order, because checking for sorted order doesn't compare every pair of
elements (if it did, sorting would be best-case quadratic time), it only
compares adjacent pairs, and

    10 < [9] < 7L < 8

Comparison wasn't always transitive.  Python eventually wormed around that,
by internally pretending that all "numeric types" have a type name
consisting of an empty string for purposes of comparing an object of numeric
type to one of non-numeric type.  That hack forces objects of numeric types
next to each other, and then their inter-numeric-type coercions can come
into play.  That's as much of a hack as it sounds, and it only works
reliably for builtin numeric types (where Python knows for sure to pretend
the type name is an empty string).

For the rest, I'm out of time.  Briefly:  If pretty printing a dict is the
killer use case, then that could surely be addressed by writing a dict
pretty-printing routine that didn't rely on meaningless comparison results.

> ...
> What's the ADVANTAGE gained, worth breaking prettyprinting for?

A language in which

    100 < "2"

is true has got at least one deep problem.  If it goes on to claim

    math.sin < xrange(10)

too, its view of comparisons is simply bizarre, and *on the face of it*.  I
believe this is a case where absurd behavior is defended simply because some
highly experienced Python programmers "got used to it".  Newbies who point
out that if Python doesn't let them add a string to an int, it shouldn't
allow comparing them either, make more sense to me.

> It seems to me it's basically one of "purity"

More that errors shoudn't pass silently.  1.5.2 took a wholly pure view.





From staschuk at telusplanet.net  Thu Apr 17 18:09:35 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 16:09:35 -0600
Subject: variable in Python
In-Reply-To: <ad052e5c.0304171017.45b8884f@posting.google.com>; from danb_83@yahoo.com on Thu, Apr 17, 2003 at 11:17:24AM -0700
References: <b7lvbh$an3$1@news-reader13.wanadoo.fr> <ad052e5c.0304171017.45b8884f@posting.google.com>
Message-ID: <20030417160935.B10011@tibia.amotlpaa.bogus>

Quoth Dan Bishop:
  [...]
> The latter: All variables in Python are pointers in disguise.
  [...example...]
> Note that I used a list instead of an int.  With ints, the distinction
> between value and reference semantics doesn't matter as much because
> ints are immutable.

And this is an extremely important point: C programmers who are
familiar with pointer semantics would expect that after
    a = 3
    b = a
    a = 4
we would find b == 4.  This is why, as Aahz pointed out, thinking
of Python variables as pointers can lead one astray, and why you
often see Python variables described as simply "names for
objects", and = as not a value-copying operation (as it is in C)
but a name-binding operation.

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From remi at remove-me.python-hosting.com  Thu Apr  3 03:29:31 2003
From: remi at remove-me.python-hosting.com (Remi Delon)
Date: Thu, 3 Apr 2003 10:29:31 +0200
Subject: web server in python
References: <b6gnau$d7q$1@news.freedom2surf.net>
Message-ID: <3e8bf0e9$0$26365$626a54ce@news.free.fr>

> i am curerntly trying to make simple dynamic site in python which i intend
> to expand on. i can make the html using python but i have a few questions.
> using the BaseHTTPServer class, is it then possible to write all the code
> for a site inside one python program, which then serves based on requests
> from a website?
> ive been having trouble finding information about this and i would be
> greatfull if someone could give me an example of how to start.

Check out CherryPy (http://www.cherrypy.org), it does exactly what you want
(build a single file containing everything to run the website).
The code for the HTTP server is a heavily customized version of
BaseHTTPServer so you can look at it to see how we did it ...


Remi.     remi at remove-me.python-hosting.com
----------------------
Specialized python hosting: http://www.python-hosting.com
Python application server: http://www.cherrypy.org
Free CherryPy hosting: http://www.freecherrypy.org
----------------------






From alloydflanagan at attbi.com  Fri Apr 18 17:41:14 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 18 Apr 2003 14:41:14 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com> <RoacnSAeGeS6CACjXTWcoQ@comcast.com>
Message-ID: <a8b7f07a.0304181341.5e4b69f3@posting.google.com>

"Terry Reedy" <tjreedy at udel.edu> wrote in message news:<RoacnSAeGeS6CACjXTWcoQ at comcast.com>...
> "Tj" <tj_scarlet at yahoo.com> wrote in message
> > fold-left) that explain nothing.  Lambda could also stand renaming;
> > Pythonistas probably think it's an arrogant name.  (Or exotic?  I
> > can't tell.)
> 
> For people without knowledge of Lisp, and maybe lacking complete
> knowledge of the Greek 'alpha-beta' and math people's habit of using
> such letters, try 'obscure'.  To me (who had such knowledge) it is
> just a bit jarring.  All other Python keywords are English words (or
> abbreviations thereof) with Python meanings similar to their everyday
> meanings.  'Lambda' sticks out like a sore thumb.  'Func' would have
> been better.
> 
> Terry J. Reedy

I'll second that.  Maybe we could get func added as a synonym for lambda, at least.



From mis6 at pitt.edu  Thu Apr 24 18:14:26 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 24 Apr 2003 15:14:26 -0700
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com> <Xns9367885737E62duncanrcpcouk@127.0.0.1> <EARpa.120$3M4.2728@news1.tin.it> <Xns93679D1BF4A2Cduncanrcpcouk@127.0.0.1>
Message-ID: <2259b0e2.0304241414.675b1e14@posting.google.com>

Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in message news:<Xns93679D1BF4A2Cduncanrcpcouk at 127.0.0.1>...
> Alex Martelli <aleax at aleax.it> wrote in 
> news:EARpa.120$3M4.2728 at news1.tin.it:
> 
> > ..isn't this smoother and handier?  Or does it just look that way
> > to me, due to overexposure to custom metaclasses?-)
> > 
> Yes, I agree that it is better, but I have one niggling worry about this:
> 
> Your metaclass works fine for the one requirement given here, and other 
> metaclasses you have proposed work fine in their place, but what happens if 
> you want to combine these and get e.g. a Bunch which counts its instances?
> 
> I haven't tried it to see what happens, maybe you just need to indirect 
> type.__new__ calls through super, and make a new submetaclass with the old 
> metaclasses as bases, but I'd like to know if there are issues here or if 
> I'm just worrying needlessly.

I did quite a lot of experiments on multiple inheritance of metaclasses,
and as far as I can tell it works quite well. You should not worry, except
in the case in which you want to multiple inherits from a metaclass and
a normal class by redefining __new__. Then I have found a nasty subtility
that made me loose an afternoon. You can find the solution if you Google
on the newsgroup under "another super wart" (read only my last post, the
first was wrong). This only happens in very complicate hierarchies, anyway.

My motivation for multiple inheritance from a class and a metaclass was
the counting of classes, since I had a Counter class:

class Counter(object): 
    counter=0  
    def __new__(cls,*args,**kw):
        cls.counter+=1 
        return super(Counter,cls).__new__(cls,*args,**kw)  

and I promoted it to a metaclass as follows:

class MetaCounter(Counter,type):
      pass

The subtle problem does not appears here but in complications of this
pattern. I would advice against this pattern, also because type.__new__
is not cooperative with Counter.__new__  and 

class MetaCounter(type,Counter):
      pass

would not work.

HTH,

                            Michele



From peter at engcorp.com  Tue Apr 15 12:23:07 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 15 Apr 2003 12:23:07 -0400
Subject: any benchmark results for python?
References: <b7h40n$28s$1@skates.gsfc.nasa.gov>
Message-ID: <3E9C31EB.21E97689@engcorp.com>

J Li wrote:
> 
> I am looking for benchmark test results for python, in comparison to other
> languages like Perl or Java or C++

Questions like this are what Google is for:

http://www.google.ca/search?q=benchmark+python+perl+java+C%2B%2B

Note: leads directly to http://www.python.org/doc/Comparisons.html
which provides links to pages answering your question in various
ways. 

Learn to use Google; it's your friend.

-Peter



From gh at ghaering.de  Fri Apr 25 19:44:45 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Sat, 26 Apr 2003 01:44:45 +0200
Subject: Dealing with files...
In-Reply-To: <pan.2003.04.25.23.37.08.286778@soul.net>
References: <pan.2003.04.25.22.39.14.432753@soul.net>
	<AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.37.08.286778@soul.net>
Message-ID: <3EA9C86D.5020004@ghaering.de>

"lost" wrote:
> On Fri, 25 Apr 2003 22:51:12 +0000, Alex Martelli wrote: 
>>for each_file in os.dir('/var/path/to/files/'):
> 
> actually, not quite what i was looking for :)
> At this line, I get:
> AttributeError: 'module' object has no attribute 'dir'

The long answer is to read the docs for the os module.

The short answer is that Alex meant 'listdir'.

-- Gerhard




From mcfletch at rogers.com  Sun Apr  6 14:31:30 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 06 Apr 2003 14:31:30 -0400
Subject: properties and get/set methods
In-Reply-To: <5kl09vses40u5v7jqrecgu5co8epaqbidh@4ax.com>
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org> <5kl09vses40u5v7jqrecgu5co8epaqbidh@4ax.com>
Message-ID: <3E907282.4020900@rogers.com>

Greg Chapman wrote:

>On Sat, 05 Apr 2003 23:48:41 -0500, "Mike C. Fletcher" <mcfletch at rogers.com>
>wrote:
>  
>
>>As for how the problem ((a deficiency in the property API IMO) of not 
>>knowing the name being set/get within the get/set methods) gets worked 
>>around.  I (as seen above) do some comparatively minor (but still 
>>    
>>
...

>You could always use somthing like this:
>  
>
<snip evil ;) hack>

Sure, but then the council of bishops and dervishes gets after you for 
being an apostate and an unbeliever ;) .  It's tempting, but I'm not 
sure it's so tempting that I'd actually use it in a general library.  
However, I saw that pattern used at PyCon by someone, so the chances 
that you can start a decent little cult are probably pretty good :) .

I just don't trust the whole "reach out and change a parent namespace" 
approach.  Maybe I'll revise that opinion, but for now I consider it 
something that should be addressed by Guido & co, rather than something 
I want to hack around myself.  If it never does get addressed I may wind 
up softening.

Temptation to either side, I walk the tightrope,
and wonder for a moment if even the fall itself might bring joy :) .

Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From Bill.Scherer at verizonwireless.com  Thu Apr 24 07:15:48 2003
From: Bill.Scherer at verizonwireless.com (Scherer, Bill)
Date: Thu, 24 Apr 2003 07:15:48 -0400 (EDT)
Subject: Problems at python.org??
Message-ID: <Pine.LNX.4.44.0304240713090.12990-100000@bagend.odc.vzwcorp.com>

Pointing my browser at http://www.python.org, I receive:

"""
No more connections can be made to this remote computer at this 
time because there are already as many connections as the 
computer can accept.
"""

Has there been a sudden Python Popularity Explosion?

-- 
Bill.Scherer at Verizon Wireless






From peter at engcorp.com  Fri Apr 25 18:55:57 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 18:55:57 -0400
Subject: Reading a non-standard floating point format
References: <mailman.1051286382.24604.python-list@python.org> <vaip4ts74ajqd0@corp.supernews.com> <Xns9327331C4783Bfidtzxyzabc@195.8.68.209>
Message-ID: <3EA9BCFD.53D41309@engcorp.com>

Fidtz wrote:
> 
> The closest thing I have to a definition from the AlphaBasic programming
> manual: The reason for this is that floating point numbers occupy six
> bytes of storage. Of the 48 bits in use for each 6-byte variable, the
> high order bit is the sign of the mantissa. The next 8 bits represent
> the signed exponent in excess-128 notation, giving a range of
> approximately 2.9*10^-39 through 1.7*10^38. The remaining 39 bits
> contain the mantissa, which is normalized with an implied high-order bit
> of one. This gives an effective 40-bit mantissa which results in an
> accuracy of 11 significant digits.
> 
> I had a go this afternoon and managed to read the sign and something
> approximating the exponent, though I will have to test the edge
> conditions carefully :)

Note that this is a *perfect* case for some test-driven development,
or at least some good unit tests with hand-checked data covering all
the edge cases you can find, and some other cases.

If you didn't know, Python has a good unit test framework as a
standard module, and it's very easy to build and run tests like this.
We'd be happy to help guide you too. :-)

-Peter



From peter at engcorp.com  Wed Apr  2 15:28:06 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 15:28:06 -0500
Subject: Some newbie questions the FAQs didn't answer!
References: <6ffd97db.0304021132.62f86c7b@posting.google.com>
Message-ID: <3E8B47D6.65EBC88E@engcorp.com>

"Jason C." wrote:
> 
> I was thinking about taking Python up this summer.  I was initially
> drawn towards it because I had heard how easy it was to program GUIs
> using the various packages.  But then I read that in order to run a
> python program, the machine had to have the python VM installed, and
> the user had to go to a command line and type "python <program
> name>.py".
> 
> What I had planned to do was write Win32 GUIs, but I would guess that
> most common users don't have Python installed.  Is there some way to
> write the program in C++ and then have it bring up the GUI written in
> Python without Python being installed?  Wouldn't that require the code
> being compiled into machine code?

The usual approach is different.  You use a special utility (there
are at least several, called py2exe and the McMillan Installer)
to package the program into an executable which *bundles* the
Python runtime (i.e. the interpreter) much like other languages
do it.  Then your users just install that on their machine as they
do with any other Windows software and they're happy and unaware
that the program is written in Python (except for being happily
surprised with how robust the program is, and how quickly you respond
to new feature requests, for example ;-).

-Peter



From dale at riverhall.NOTHANKS.co.uk  Mon Apr 28 15:29:33 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Mon, 28 Apr 2003 20:29:33 +0100
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com>
Message-ID: <gkuqavo4cgrfuljhokpli5q41pdlvf8v76@4ax.com>

"Anna" <revanna at mn.rr.com> wrote:

>My daughter just asked me about this "python stuph". So I pulled up
>livewires.org.uk and showed her some basics.
>
>She just coded her first while loop and for loop... she thought it was
>kewl! In fact, we had to kick her off the computer and send her to bed cuz
>she's got so much to do tomorrow...
>
>FWIW, she's 12...
>
>Anna

Get her on the turtle graphics which are included in Python 2.2.

>>> import tutle
>>> turtle.demo()

My daughter was coding simple shapes and patterns months ago and she
is 9.

Watch out for Livewires, though. It's blatant religious evangelistic
twaddle wrapped in a veneer of respectable education. 

Too much of that stuff and your intellect takes a dive. Telltale signs
include your spelling of simple words like 'stuff', 'because' and
'cool' goes all to Hell.
-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From bokr at oz.net  Wed Apr 23 16:13:08 2003
From: bokr at oz.net (Bengt Richter)
Date: 23 Apr 2003 20:13:08 GMT
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <NKxpa.85054$T34.2112670@news2.tin.it>
Message-ID: <b86s4k$as7$0@216.39.172.122>

On Wed, 23 Apr 2003 14:50:21 GMT, Alex Martelli <aleax at aleax.it> wrote:
[...]
>However, such new functionality doesn't get back-ported to previous
>releases of Python, such as 2.2.*.  In all Python releases from 1,6
>included to 2.3 excluded, "needle.find(haystack) > 0" is the idiom.

If needle is a substring to be found in haystack, that seems reversed. I.e.,

 >>> haystack = 'needle-in-haystack with second needle fthoi'
 >>> needle = 'needle'
 >>> haystack.count(needle)
 2
 >>> haystack.find(needle)
 0

And a nit:
 >>> haystack.find(needle) > 0
 0
 >>> haystack.find(needle) >= 0
 1

Were you recently using re to find needles? E.g.,

 >>> import re
 >>> needle = re.compile('needle')
 >>> needle.findall(haystack)
 ['needle', 'needle']

Regards,
Bengt Richter



From dirk at gerrits.homeip.net  Wed Apr 23 16:59:26 2003
From: dirk at gerrits.homeip.net (Dirk Gerrits)
Date: Wed, 23 Apr 2003 22:59:26 +0200
Subject: Tkinter text entry questions: set_text, onchange.
Message-ID: <b86ups$hec$1@main.gmane.org>

I've been creating a little GUI application with Python 2.3a2 and 
Tkinter and I've been running into some issues with Entry widgets.

I have some of them with state='readonly', and I want to update their 
text values from Python functions. For reading their text values there 
is the simple method .get(), but there doesn't seem to be a 
corresponding .set() method? So I've tried using delete followed by 
insert, but this doesn't seem to work for state='readonly'. Finally, I 
made the following helper function:

def set_text(entry, text):
     old_state = entry["state"]
     entry["state"] = "normal"
     entry.delete(0, END)
     entry.insert(0, str(text))
     entry["state"] = old_state

This seems to do the trick, but it seems to me that it's a lot more 
trouble then necessary. Isn't there a more obvious, more 'pythonic' 
solution?

Also, I have some state='normal' Entry widgets for which I want to call 
a function whenever their value changes. Currently, I use the following 
helper function, but just as with set_text, I feel it is a non-obvious hack:

def bind_onchange(entry, command):
     for sequence in ["<Key>", "<KeyPress>", "<KeyRelease>"]:
         entry.bind(sequence, command)

Is there a better solution for this (too)?

Thanks in advance,

Dirk Gerrits






From lamont at security.hp.com  Sat Apr 12 12:06:46 2003
From: lamont at security.hp.com (LaMont Jones)
Date: Sat, 12 Apr 2003 16:06:46 GMT
Subject: DNS lookup
References: <eTFaa.3697$Fc.114939@tornado.fastwebnet.it> <b4fb0l$u0q$03$1@news.t-online.com>
Message-ID: <qQWla.732$5n6.120@news.cpqcorp.net>

"Martin v. L?wis" <martin at v.loewis.de> wrote:
> I recommend to use socket.gethostbyname (or socket.getaddrinfo if you 
> need support for IPv6 as well).

And if I need asynchronous NS/SOA lookups?  (What I really want is
socket.res_query(), but that's beyond my current level of python-hacking
ability...)

lamont



From alloydflanagan at attbi.com  Sat Apr 12 07:52:58 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 12 Apr 2003 04:52:58 -0700
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk>
Message-ID: <a8b7f07a.0304120352.75b53dd8@posting.google.com>

Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfsistmh4ec.fsf at black132.ex.ac.uk>...
 
> __slots__ is the one addition in python2.2 that really annoys me. It seems an
> ill-conceived jumble of conceptually unrelated things (bug avoidance,
> optimization, safety, encapsulation) that screws up reflection big time for
> little apparent gain. As Knuth said ...
> 
> 'as

I agree.  The more I looked at it, the more it became clear that
__slots__ turns a class into something very non-pythonic, with
significantly different behavior.  A lot of utilities, debuggers, and
advanced programming tricks will be coming with this disclaimer:
"does not work with any class which defines __slots__"

Unless you absolutely have to have it, don't use it.



From oblivious at web.de  Sat Apr 12 13:50:33 2003
From: oblivious at web.de (Dominic)
Date: Sat, 12 Apr 2003 19:50:33 +0200
Subject: Macros in Python? (PEP 310)
In-Reply-To: <wJRla.3135$LB6.88227@news1.tin.it>
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org> <b770t8$2b8$00$1@news.t-online.com> <b771jb$ac6$03$1@news.t-online.com> <a8b7f07a.0304111546.3e5a5636@posting.google.com> <wJRla.3135$LB6.88227@news1.tin.it>
Message-ID: <b79jii$njr$06$1@news.t-online.com>

> 
> def wrapfunc(func, pre=mkprint("hi"), post=mkprint("bye")):
>     def wrapped(*args, **kwds):
>         pre()
>         try: return func(*args, **kwds)
>         finally: post()
>     return wrapped

Ups, I am surprised that wrapped still has access
to func I thought that this would only work
with Python's lambda. My fault.
The pre/post calls are only executed once
when the file is loaded so this should be
handled with care...
However I do not need to pass pre/post as
parameters, so it does not matter very much.

The only drawback is that the code needs
to be separated into a fuction unless the
generator "hack" is used. So wrapfunc seems
to be the optimum so far.

Ciao,
  Dominic





From ramen at lackingtalent.com  Thu Apr 24 03:44:08 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Thu, 24 Apr 2003 07:44:08 -0000
Subject: Why functional Python matters
References: <mailman.1051154381.32098.python-list@python.org> <slrnbaf35b.ebi.ramen@lackingtalent.com>
Message-ID: <slrnbaf5fa.gb9.ramen@lackingtalent.com>

In article <slrnbaf35b.ebi.ramen at lackingtalent.com>, Dave Benjamin wrote:
> Perhaps this is an over-generalization, but I think that nestability
> encourages readability. The nice thing about lambdas in Python are that they
> enforce this property.

Before I get completely blasted for this, I meant to say "reusability", not
"readability". I am aware of the effect that nesting can have on readability.
It's not necessarily a good thing. ;)

Ducking,
Dave



From mcfletch at rogers.com  Mon Apr 28 10:52:26 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Mon, 28 Apr 2003 10:52:26 -0400
Subject: Recursive Algorithm getting stuck (because of Python Limits?)
In-Reply-To: <236da4ed.0304280616.349e8bdd@posting.google.com>
References: <236da4ed.0304280616.349e8bdd@posting.google.com>
Message-ID: <3EAD402A.6030705@rogers.com>

You're creating close to a million objects in that tree for depth=8, 
that takes >233MB on my machine and 98s.  I'd guess your machine is just 
going into swapping fits or the like when it starts running out of 
memory.  By the way, doing this can get you down to 81MB and 46s for the 
same depth under Python 2.2.2:

class Node(object):
    __slots__ = ('childNodes',)

Good luck,
Mike

Marc Floessel wrote:

>I am trying to execute a simple Tree Building algorithm.
>
>"
>nTreeBreadth = 7
>nTreeDepth = 7
>
>
>class Node:
>    def __init__(self):
>        self.childNodes = []
>    
># Create a recursive Sub-Tree with <nChildren> until we reach the
>final <nDepth>
>def CreateSubTree (oRoot, nChildren, nDepth):
>    for i in range(nChildren):
>        oNewNode = Node()
>        oRoot.childNodes.append(oNewNode)
>        if (nDepth > 1):
>            CreateSubTree (oNewNode, nChildren, (nDepth-1))
>
>oRoot = Node() # Create the Root Node
>CreateSubTree (oRoot, nTreeBreadth, nTreeDepth-1)
>"
>
>This works. However increasing the Depth or Breadth to 8 seems to
>break Python. Python rises to about 100 MB memory consumption (as
>indicated by the Windows Task Manager) and then seems to drop and rise
>randomly. The algorithm never finishes. Am I breaking the stack? I
>tried executing the code (with a Depth of 8) in Stackless Python, but
>it didn't finish either, so it doesn't appear to be the stack. Are
>there any other hard limits?
>  
>

-- 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From aleax at aleax.it  Mon Apr 28 04:02:20 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 08:02:20 GMT
Subject: Forgetting "()" when calling methods
References: <mailman.1051497342.21537.python-list@python.org>
Message-ID: <ge5ra.13872$3M4.378491@news1.tin.it>

<posted & mailed>

Tim Peters wrote:

> [Terry Reedy]
>> ...
>> At one time, it was easy for anyone to get 100% on a quiz like 'how
>> many of the following lists can Python sort?".
> 
> OTOH, how many would have scored 100% if also asked what the result of
> sorting would be?  I'm not sure anyone could have.  For years, ints came
> out "less than" lists, which in turn came out "less than" longs, so that
> 
>     [10, [9], 7L, 8]
> 
> was considered to be "in sorted order" as-is.  It was something like 7
> years before someone finally noticed how nuts that was.

*Blink* -- I don't get it.  Even if int < list < long, how CAN the
two int's here be ``considered in sorted order''???

> two within a single program run).  When that's so, sorting brings objects
> of
> the same type next to each other, and that's about it.  But if that's "the
> use case" (if it isn't, what is the use case?!), it's much more efficient
> (O(N) vs O(N log N)) to segregate objects via a dict mapping type to a
> list of objects with that type.

Here is a typical use case: prettyprinting a dictionary.

>>> import pprint
>>> adic = dict.fromkeys([ i*1j for i in range(4) ])
>>> pprint.pprint(adic)
{0j: None, 3j: None, 1j: None, 2j: None}

So far so good -- we CAN prettyprint a dictionary.  I'd rather have
the keys show (within a prettyprint operation only) in an order that
seems less arbitrary, but, oh well, can't have everything, I guess.

So, let's prettyprint a slightly bigger dict...:

>>> adic = dict.fromkeys([ i*1j for i in range(8) ])
>>> pprint.pprint(adic)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/local/lib/python2.3/pprint.py", line 54, in pprint
    printer.pprint(object)
  File "/usr/local/lib/python2.3/pprint.py", line 105, in pprint
    self._stream.write(self.pformat(object) + "\n")
  File "/usr/local/lib/python2.3/pprint.py", line 109, in pformat
    self._format(object, sio, 0, 0, {}, 0)
  File "/usr/local/lib/python2.3/pprint.py", line 142, in _format
    items.sort()
TypeError: cannot compare complex numbers using <, <=, >, >=
>>>

Oops!  We can't prettyprint it any more -- kaBOOM.

It's perfectly natural to want to sort a dict's keys in order to
prettyprint it -- as you see, the standard library's own pprint
does it (but only if the dict is large-ish).  What one would LIKE
to happen is to have the keys come out bunched by types and in
SOME vaguely sensible order within each type (but probably with
number-types mixed together, and string-types too) -- but "pretty" 
may be in the eye of the beholder, so we can quibble about that
separately, I guess.

However, it seems totally arbitrary to me that the standard
library's pprint.pprint should raise an exception just because 
the user's trying to prettyprint a dictionary with some complex
numbers as keys.  I guess one could class this as a bug in
pprint, but from my POV what it does is highlight a typical
case in which the inability to ask "just sort this stuff so
it displays in an order that doesn't appear totally random"
*hurts*, and hurts pretty badly too.

What's the ADVANTAGE gained, worth breaking prettyprinting for?

It seems to me it's basically one of "purity" -- that 1.5.2
was far more pragmatic in this regard.  I know I'm not gonna get
the pragmatism back, sigh, but is it SO hard to understand why
I would want it?!  I need to show beginners what's going to
result from various operations -- and the fact that when the
results are dicts they display in random-ish order is a small
but unpleasant distraction from whatever point I'm trying to
teach -- so, I might want to (e.g.) hook displayhook to give
"pretty" results rather than raw repr's.  pprint.pprint might
be a good starting point... BUT it's broken by this purity-
based "enhancement"...


Alex




From carroll at tjc.com  Fri Apr 18 01:45:47 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Thu, 17 Apr 2003 22:45:47 -0700
Subject: Downloading TV listings with urllib
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4>
Message-ID: <d04v9vkls8vdittb63u01mvotndfh1b6gv@4ax.com>

On 17 Apr 2003 17:23:34 GMT, Josh <mlsj at earthlink.net> wrote:

>Can someone give me general pointers on how it could be done? I am trying 
>to go to tvguide.com and download the TV listings for my area. The problem 
>is how to specify my location and cable provider. 

Josh --

I don't have specific pointers on your question.  However, you should
be aware that there's already a project to do this fetching, although
in Perl, not Python: XMLTV, at
<http://sourceforge.net/projects/xmltv>.

It brings down the listings and stores them in an XML file.  You might
find it most convenient to use XMLTV to download your listings, and
then use Python's XML facilities to process them (I've had just this
sort of project on the back burner for some time).  

In any event, you might find looking at the XMLTV code instructive,
even though it's Perl.

Good luck.



From andrew-pythonlist at puzzling.org  Sun Apr 27 07:12:45 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 21:12:45 +1000
Subject: Forgetting "()" when calling methods
In-Reply-To: <bAOqa.14719$K35.482952@news2.tin.it>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <mailman.1051423062.28410.python-list@python.org> <bOLqa.13827$K35.455183@news2.tin.it> <mailman.1051430877.5518.python-list@python.org> <bAOqa.14719$K35.482952@news2.tin.it>
Message-ID: <20030427111245.GR26530@frobozz.local>

On Sun, Apr 27, 2003 at 10:48:39AM +0000, Alex Martelli wrote:
> Andrew Bennetts wrote:
> 
> > On Sun, Apr 27, 2003 at 07:38:47AM +0000, Alex Martelli wrote:
>    ...
> >> See?  PyChecker CAN and DOES warn about "this case"!  But OF COURSE
> >> it can't help with the following erro1.py:
> > 
> > Sure, but I said: "removing __nonzero__ from functions still doesn't help
> > this case".  You can catch that today, with PyChecker, as you've
> 
> I.e., it doesn't help with a case that doesn't really need help -- is THAT
> the "point" you thought it was worth a post to make (and DIDN'T mention!)?

It just seemed like you ignored my point, and then proceeded to show that
pychecker is sometimes helpful.  I was trying to say that I was aware of
that, and it didn't affect my opinion.

> There are a zillion cases where no "help" is particularly needed nor
> warranted.  The fact that any proposed change "doesn't help" with any
> subset of the cases that don't particularly need help is so clearly
> irrelevant to any assessment of that proposed change that I _still_
> don't see why ever one would want to start enumerating such cases.

Ok, the meta-arguments are getting so convoluted I propose we just give them
up and stick to the actual underlying argument :)

> > It's still not a justification for breaking __nonzero__ on functions. 
> 
> Would it damage you if that __nonzero__ raised a warning that is
> suppressed by default, but can easily be (selectively) enabled?

Yes -- if it's a PendingDeprecationWarning or similar, that suggests what I
consider a perfectly good idea will at some point be removed from the
language.  If it's a new kind of warning that is merely a "Hey!  If you're a
newbie and want handholding, you might want to double-check this", then
that's ok with me.  If possible, I think that sort of thing belongs in
PyChecker, along with various other warnings with high false positive rate.

> > There are plenty of cases where "if func:" is perfectly valid, and no
> > amount of "SINERGY" is going to convince me that making that fail (either
> > in PyChecker or at runtime) is a good idea :)
> 
> I apologize for the misspelling, but it doesn't affect my argument.

Sorry; that was a low blow on my part.  It was the combination of the
misspelling, the ALL-CAPS, and the word "synergy" (which I have trouble
taking seriously at the best of times ;) that somehow affected my judgement.

> You mentioned that you have no extensive experience teaching newbies,
> particularly ones that come from the many languages where trailing
> empty parentheses are not necessary to call an argument-less function,
> such as Pascal and its descendants, as well as Visual Basic.  Yet it
> is exactly to help such newbies that I'm considering making __nonzero__
> raise an optional warning for function and method objects -- help them,
> together with PyChecker for the diagnosis of other cases, get info
> ASAP about the fact that once again they've forgotten a trailing '()',
> rather than having to deduce it from subtle errors arising later and
> quite possibly in quite separate points of their code.

I do have some relevant experience here, if only a little: I used to do alot
of VB before I started doing alot of Python, and for a fair while in the
middle I did alot of both.  So I *did* occasionally make this very error,
because I was switching between languages -- but never did it cause me more
than a couple of seconds' worth of anguish.  That's the basis upon which I
consider this proposed change to be unjustified, especially given that I 
consider it more useful the way it is.

[Oh -- and I *do* have some experience teaching Python to a class of
newbies, I just remembered, but just one class.]

-Andrew.





From a.schmolck at gmx.net  Thu Apr 24 21:15:57 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 25 Apr 2003 02:15:57 +0100
Subject: variable scoping problem
References: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com>
Message-ID: <yfs65p3wf5e.fsf@black132.ex.ac.uk>

Tung Wai Yip <tungwaiyip at yahoo.com> writes:

> Consider this
> 
> >>> x=1
> >>> def foo():
> ...   print x
> ...
> >>> def bar():
> ...   #print x   <- referenced before assignment error
> ...   x=2
> ...   print x
> ...
> >>> foo()
> 1
> >>> bar()
> 2
> >>> print x
> 1
> >>>
> 
> What rules in Python explain the scoping of x? Is there a name for the
> scope of the first x? 'module' scope? Why is bar() not able to access
> x but foo() can?

Q1: global scope

Q2: Because when python sees that x is assigned to in a function it is treated
as a local variable *anywhere* in that function (even *before* the
assignment); *unless* you've declared it as global in that function. As long
as you don't assign to it, you're fine.

This makes sense -- otherwise you could easily introduce devious bugs by
changing global variables without any intent of doing so. OTOH, you don't want
to have to declare a module, function or a constant global just so that you
can use it from within the function.

Consider (all untested):

def aTest():
    print "HI"

def bTest():
    def aTest(): print "Just a local definition"
    aTest()

then consider:

>>> aTest
HI
>>> bTest()
Just a local definition
>>> aTest()
HI

def aTest():
    print "HI"

def bTest():
    aTest() 
    def aTest(): print "NEVER"
    return aTest(3)
>>> bTest()


and finally:

def aTest():
    print "HI"
def bTest():
    global aTest
    aTest()
    def aTest(): print "REDEFINED GLOBALLY"
    aTest()
>>> aTest()
Hi
>>> bTest()
HI
REDEFINED GLOBALLY
>>> aTest()
REDEFINED GLOBALLY


'as



From exarkun at intarweb.us  Tue Apr 15 13:19:05 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 13:19:05 -0400
Subject: an ugly file-reading pattern
In-Reply-To: <slrnb9nkhp.493.rmunn@localhost.localdomain>
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050201322.13220.python-list@python.org> <slrnb9h7u4.it3.rmunn@localhost.localdomain> <3e9a857b$1@news.012.net.il> <slrnb9ksaa.1q1.rmunn@localhost.localdomain> <3e9ae78f$1@news.012.net.il> <mailman.1050342503.3825.python-list@python.org> <slrnb9nkhp.493.rmunn@localhost.localdomain>
Message-ID: <20030415171905.GA32515@meson.dyndns.org>

On Tue, Apr 15, 2003 at 03:28:31PM +0000, Robin Munn wrote:
> Jp Calderone <exarkun at intarweb.us> wrote:
> > On Mon, Apr 14, 2003 at 06:53:32PM +0200, Avner Ben wrote:
> >> 
> >> "Robin Munn" <rmunn at pobox.com> wrote in message
> >>
> >> > You're right, I didn't close the file explicitly. But explicit is better
> >> > than implicit, so IMHO it's better to give a name to the file object so
> >> > that you can close it explicitly:
> >> >
> >> >     input_file = file("somefile")
> >> >     for line in input_file:
> >> >         process(line)
> >> >     input_file.close()
> >> 
> >> 
> >>     The C# language has a nice construct: using (variable = whatever) { /*
> >> code */ }. The variable in the using header is guaranteed to be
> >> garbage-collected at the end of the block (or so I think). Could be useful
> >> here.
> >> 
> > 
> >   In CPython, the reference count to the file object in:
> > 
> >     for line in file(...):
> >         ...
> > 
> >   will drop to 0 immediately after the loop, and the file will be closed at
> > that point.
> 
> But the Python language spec doesn't guarantee that objects will be
> garbage-collected immediately when their refcount drops to 0; 

  Python language spec? >:)  I thought CPython was the language spec, for
all practical purposes.

> and, in fact, Jython does *not* garbage-collect zero-refcount objects
> immediately. Relying on garbage-collection is not safe. For most
> operations, it's safe enough (if a file stays open a bit longer than you
> expected, it's probably not a big deal) but there are times when it is
> important to know exactly when the file is closed; I prefer to simply make
> a habit of it.

  I agree.  What I would *really* like to see is a PEP from the "with ...:"
construct debate that occurred several weeks ago on python-dev.

  Jp

-- 
No, `Eureka' is Greek for `This bath is too hot.'
                -- Dr. Who
-- 
 up 26 days, 13:02, 2 users, load average: 0.00, 0.02, 0.04




From sholden at holdenweb.com  Mon Apr 14 17:39:42 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Mon, 14 Apr 2003 21:39:42 GMT
Subject: encapsulation checking features
References: <3e9b1298@cs.colorado.edu>
Message-ID: <yUFma.88352$It5.54564@news2.central.cox.net>

"Eric Scharff" <edsZZZ_2 at yahoo.com> wrote in message
news:3e9b1298 at cs.colorado.edu...
> I love python's semantics for rapid prototyping, but I was wondering if
> there was a way to turn on warnings so that for other kinds of code, I
> can avoid certain common uglinesses. :)
>
> Specifically, is there a "lint-like" tool to:
>
>  * warn if instance variables are created outside of a constructor?
>
>   class K:
>     def __init__(self):
>       self.a = 6  #okay
>     def example(self):
>       self.b = 66 # not as good
>
>   myVar = K()
>   myVar.c = 666  # really bad
>
>  * warn if instance variables are used outside of instance methods
>    (encouraging the use of getters and setters)
>
> These are great features to have, but it'd be nice to have warnings to
> discourage their use at times.
>

There's pychecker, but I don't know whether it specifically checks for thos
conditions. See

    http://pychecker.sf.net/

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/



>  (genuine email replies, remove ZZZ from email address)





From brian094 at sympatico.ca  Sun Apr  6 13:38:06 2003
From: brian094 at sympatico.ca (Brian Alexander)
Date: Sun, 06 Apr 2003 13:38:06 -0400
Subject: curses: xterm vs. linux terminal
Message-ID: <3E9065FE.1090108@sympatico.ca>

Hello;

Why is it that the curses special (alt) characters do not display on xterms?

Ex:

theScreenObj.box(curses.ACS_VLINE, curses.ACS_HLINE)
theScreenObj.refresh()

This code puts a box with graphical characters around the window when 
the terminal is of type linux. When the terminal is of type xterm, 
blanks appear where the line characters should be.

Is there anyway to fix this behaviour? I find it hard to believe an 
xterm does not support the drawing of these simple graphical characters.

Many thanks,

Brian.




From rs55862 at NOSPAMyahoo.com  Tue Apr  8 09:12:55 2003
From: rs55862 at NOSPAMyahoo.com (Rami Saarinen)
Date: Tue, 08 Apr 2003 16:12:55 +0300
Subject: Making python module using Borland C++ 5
In-Reply-To: <3e92b527$0$32016$edfadb0f@dread12.news.tele.dk>
References: <b6u7fp$d2c$1@news.cc.tut.fi> <3e92b527$0$32016$edfadb0f@dread12.news.tele.dk>
Message-ID: <b6uhsn$kj9$1@news.cc.tut.fi>

Anders J. Munch wrote:

>>I am trying to build a simple module for python (2.1.1) using the 
>>Borland C++ Builder 5.
> 
> 
> Piece of cake with distutils, assuming you're just writing an
> ordinary extension module, and not trying to embed Python in your
> application.
> 
> Just write a setup.py and run it with --compiler=bcpp.

(Sorry if this comes duplicated, but I sent a message over an hour ago 
and it still has not shown up..)

Yes, I did use the "python setup.py build" thingy when using vc++. I did 
not know about the compiler option which is good to know. Thanks Anders.

However, the thing is that eventually I'd be using a custom python 2.1 
that has its' own python21.lib and includes. I am pretty sure that there 
  is no distutils for that installation so I need to learn to do it by 
hand. I'd like to do it using normal python installation just to be on 
the safe side. Currently I have made the CLIPS 6.2 as static library and 
use it along with python21.lib.

The module (in C) is very simple (adapted form the embedding.. docs), 
providing only three functions to python. I bet my pants off, that there 
is some really simple solution for this. 8|

So to summarize:

* I'm using 2 static libs python21.lib and clips.lib
* one dead-simple c-file - the module - to work as glue between clips 
and python.
* The compiler is Borland C++ builder 5 on WinXP.

--
Rami Saarinen




From nils.hensel at web.de  Sat Apr 26 08:56:14 2003
From: nils.hensel at web.de (Nils Hensel)
Date: Sat, 26 Apr 2003 14:56:14 +0200
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com> <S5tqa.9420$K35.308092@news2.tin.it> <b8do9j$91iv5$1@ID-126101.news.dfncis.de>
Message-ID: <b8dv9n$6de$06$1@news.t-online.com>

LMAO!

A very good one there!
May god beware the PERL part!

Nils


"Axel Grune" <NineOfSix at gmx.de> schrieb im Newsbeitrag
news:b8do9j$91iv5$1 at ID-126101.news.dfncis.de...
> Alex Martelli wrote:
> > we all know how teenagers tend to rebel against parental authority), the
>
> ...and then they start to do drugs and program Perl! Jesus, save them! ;-)
>
> Axel
>





From sdhyok at yahoo.com  Thu Apr 17 22:09:51 2003
From: sdhyok at yahoo.com (sdhyok)
Date: 17 Apr 2003 19:09:51 -0700
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com>
Message-ID: <420ced91.0304171809.26fe1221@posting.google.com>

sdhyok at yahoo.com (sdhyok) wrote in message news:<420ced91.0304141131.73a6206e at posting.google.com>...
> I am looking for a way to express a date in a real number
> in Python.
> 
> For instance, in MatLab,
> n = datenum(2001,12,19,18,0,0) returns n = 731204.75.
> 
> This kind of expression turned out to be very helpful
> when handling or storing dates.
> 
> Is there any funtion for the purpose in Python (e.g. in datetime module)?
> 
> Daehyok Shin

Another solution I found in mx.DateTime is:

from mx.DateTime import *
d = DateTime(1999,9,9,3,10,5)
d.absdays



From donn at drizzle.com  Thu Apr 17 03:18:02 2003
From: donn at drizzle.com (Donn Cave)
Date: Thu, 17 Apr 2003 07:18:02 -0000
Subject: Using __repr__ or __str__ for own printable class?
References: <mailman.1050525028.7854.python-list@python.org> <b7km3s$238m$1@nntp6.u.washington.edu> <mailman.1050537782.19056.python-list@python.org> <3E9DF377.E94BFCB8@engcorp.com>
Message-ID: <1050563880.979772@yasure>

Quoth Peter Hansen <peter at engcorp.com>:
| Steven Taschuk wrote:
...
|> But information-rich and unambiguous representations, however
|> useful for (e.g.) debugging, are not always easy to read or
|> understand; for the humans then, we'd like an "informal",
|> easier-to-read representation, whence __str__.
|
| I've always considered __str__ to be for use in the application,
| such as when you have a class which is sometimes asked for a 
| string representation for display purposes to a real user, or in
| an application log, or something.

Yes!  An object implements the __str__ interface to function as a
string in some computation that requires a string.

| I've always considered __repr__ to be for use in *debugging* the
| application, where a *debugging* log needs to display information,
| or in a traceback.  Since it's next to impossible for most interesting
| objects to be reproduced easily from the output of __repr__ if it's
| also desirable that the output be meaningful and useful to a human 
| reader, I don't buy into the 'eval(repr(x)) == x' theory very much,
| other than to note that it's the most obvious thing to do for 
| primitive data types.
|
| I've never had any confusion in my own mind about when it's 
| appropriate to use one or the other.  Perhaps this perspective 
| should be used in the documentation, rather than the focus on 
| reproducing objects from their __repr__.  

Works for me.  Of course the documentation mentions repr's use
in debugging etc., but it's distinctly subordinated.  __repr__
must conform to the expression requirement if possible, and
in the __str__ documentation even that concession (``if possible'')
is omitted.

	Donn Cave, donn at drizzle.com



From max at alcyone.com  Thu Apr 17 04:09:09 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 17 Apr 2003 01:09:09 -0700
Subject: "pythonic"
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
Message-ID: <3E9E6125.C928E910@alcyone.com>

Afanasiy wrote:

> Too many times now I have seen the term "pythonic" used to vaguely
> describe something which is never explicitly defined, as if it cannot
> be.
> 
> Now it seems I can just replace "pythonic" with "my idea of utopia".
> 
> I do not know if this ignorance, laziness, or a meme, but I would like
> to know if there a real definition of "pythonic". Anyone?

Your characterization is not entirely inappropriate.  People use the
term to mean "that which seems to follow good style laid about by the
language Python."  There's no yes or no here; it's a judgement call.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Can I walk with you / 'Till the day that my heart stops beating
\__/ India Arie
    Polly Wanna Cracka? / http://www.pollywannacracka.com/
 The Internet resource for interracial relationships.



From alloydflanagan at attbi.com  Wed Apr 16 13:46:54 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 16 Apr 2003 10:46:54 -0700
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org>
Message-ID: <a8b7f07a.0304160946.73760d3a@posting.google.com>

Dan Jacobson <jidanni at dman.ddts.net> wrote in message news:<87he90b8rp.fsf at jidanni.org>...
> I need to use one liners every day, especially in Makefiles.  I thus
> need to know how to make any kind of python program into a big long one
> liner. However, right out of the starting gates I run into
> $ python -c 'print 2;for i in (1,4):print i'
>   File "<string>", line 1
>     print 2;for i in (1,4):print i
>               ^
> SyntaxError: invalid syntax
>
... 
> 
> This makes no output and returns no error value to the shell:
> $ python -c 'import math' -c 'for i in range(200,500,50): print
>   i,360/2/math.pi*math.asin(i/1238.4)'
...

You don't have to do these as one liners.

You can use a shell trick called a 'here document'.  It lets you put a
set of lines directly in the shell script, and pass them as input to a
program.  Here's an example:

#!/bin/ksh
python << EOF
y = 0
for i in range(10):
        y += i
        print i, y
EOF

The '<< EOF' tells the script to send everything to python until the
line that contains just EOF  (and you can use any label, not just
EOF).  So for your example:

python << DONE
print 2
for i in (1,4):
     print i
DONE

python << END_MATH
import math
for i in range(200,500,50):
     print i,360/2/math.pi*math.asin(i/1238.4)
END_MATH

These all worked fine on HP/UX 10, using korn shell, with python 2.1
(no snide remarks, please, I didn't choose the setup :).



From tim at bladerman.com  Tue Apr  8 13:24:15 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Tue, 08 Apr 2003 17:24:15 GMT
Subject: Loading a class the name of which is only known runtime
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net>
Message-ID: <3BDka.4759$Du.16741@newsc.telia.net>

"Mads Orbesen Troest" <mads at troest.NEVERMORE.dk
> Is it possible to load (import and instantiate) a class dynamically in
> Python? I have a feeling the answer is Yes, but as I am unfortunately not
> quite the python expert (yet :) forgive me if I ask something obvious.

If this is what ou are asking for it is certanly possible.

if test:
    import modul1
    a=class1()
else:
    import module2
    b=class2()

Good luck

Tim






From ckea25d02 at sneakemail.com  Thu Apr 17 01:43:54 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 17 Apr 2003 17:43:54 +1200
Subject: wxPython: Intercepting keyboard events from multiple widgets?
Message-ID: <b7le9j$266le$1@ID-169208.news.dfncis.de>

I have a wxPython problem:

I have a wxNotebook, each wxPanel of which contains a
number of controls.

I want to catch certain keystrokes and use them for
things pertaining to the whole collection of controls,
such as switching from one tab of the notebook to another.

Is there any way I can do this without having to attach
event handlers individually to every control?

In the wxWindows docs under "Pluggable Event Handlers"
it hints at it being possible to do things like
intercepting all the events directed at a dialog box,
but it doesn't go into any details, and it doesn't
seem to be obvious how you would do this. Attaching
an event handler to the dialog itself doesn't seem
like it would work, because keyboard events seem to
originate in whatever control has the keyboard focus
and propagate upwards.

Can anyone point me to an example which shows how
to do this, if it's really possible at all?

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From anti-spam.dbroadwell at mindspring.com  Mon Apr 28 12:54:38 2003
From: anti-spam.dbroadwell at mindspring.com (David Broadwell)
Date: Mon, 28 Apr 2003 16:54:38 GMT
Subject: Newbie: ( if good in self: print "Foobar" )
References: <uN1ra.15662$kc1.13753@nwrddc02.gnilink.net> <%r4ra.19022$K35.649255@news2.tin.it>
Message-ID: <i1dra.3534$TR3.612@nwrddc04.gnilink.net>

BTW: Thank you, You do some good responces to the group. I especially liked
the 'forgot ()' thread solution even though i'm sure you got flames for it.
I'll have to read that 'primer' on metaclasses ... loooks usefull and
potentially code saving.

<posted & mailed>

>> class foo:

>> bar = ['B','A','R']

>>

>> def baz():

>> for item in bar: print item

>> boggle()

>> for item in bar: print item

> This baz function just won't work -- its unqualified references to 'bar'

> and 'boggle' can't be solved, and the "signature" (pattern of arguments,

> namely none) is only compatible with a staticmethod or a free function,

> not with an ordinary method as this appears to be.

Ok. (this didn't make sence on the first read)

>> Question 1: Do I read it right that; a classes method called with self
can

>> change the 'global' value without having to type the bloddy global for

>> every one?!? (ie: the above code not the below code)

> No, it's NOT a *global* attribute -- it's an attribute of the CLASS. You

> can access an attribute of a class through any instance of that class, as

> long as the instance hasn't "overridden" the attribute (i.e., bound an

> attribute of the same name as a per-instance attribute on itself).

Oh. That on the other hand maked perfect sence. All i was ever trying to do
was access a variable one level up and was seemingly forced me to a global
to do the assignment. While a list[index] dosen't seem to have that problem,
strings and integers did. (problem with programmer, not language) That is
untill i go all classy. And that is about exactly what i might do.

>>>>> def boggle(self):

>> self.bar[1] = 'a'

>> self.bar[2] = 'r'

> However, this function needs an argument. You're not passing that

> argument when you call it, thus, boom.

Except within an instance of a class when self is defined? You see i did try
to have boggle called with self and got that little complaint in the
interpreter.

> Question 3: So what exactly does my code have to look like to work?

> Question 4: Is this notation only valid inside a class?

> You can use the "notation" wherever you want, but you cannot call a

> free function without arguments if it wants an argument.

But you can call a class method from an instancce of a class without
arguments and 'self' is required to be received?

> To make the above-exemplified free-standing boogle function work,

> you would have to pass it an object having an attribute named bar.

> In this case, the object in question is the module in which the

> functions baz and boggle live. You can get at the module object

> in the caller, though it's a rather fishy usage IMHO:

>>>> def baz():

> for item in bar: print item

> import sys

> boggle(sys.modules[__name__])

> for item in bar: print item

Oh, a non class instance answer ... fish for the name. But in some cases
will save me a global vaiable declaration, which while simple, i'm NOT fond
of for some reason or other. And in a few cases it will save a STACK of
global variable declarations, which offend my sence of namespace smoothness
... any only requires a bit of fish sauce.





From thoaan at stud.cs.uit.no  Thu Apr 10 12:58:13 2003
From: thoaan at stud.cs.uit.no (Thomas Aanensen)
Date: Thu, 10 Apr 2003 18:58:13 +0200
Subject: Is it possible to run more than one shell in PPCPython?
Message-ID: <b747pl$239g$1@news.uit.no>

Yep, that's the question...


Thomas





From jjl at pobox.com  Tue Apr 29 18:05:01 2003
From: jjl at pobox.com (John J. Lee)
Date: 29 Apr 2003 23:05:01 +0100
Subject: apache modules in python ?
References: <pan.2003.04.29.22.09.59.56666@sdf-eu.org>
Message-ID: <863ck1c642.fsf@pobox.com>

"Geirgat Jonas" <kemu at sdf-eu.org> writes:

> Is it possible to write modules for the apache http server in python ?

Google for mod_python and mod_snake.  There are others too, IIRC.


John



From skip at pobox.com  Mon Apr 14 09:57:02 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 08:57:02 -0500
Subject: an ugly file-reading pattern
In-Reply-To: <3e9a857b$1@news.012.net.il>
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
        <mailman.1050201322.13220.python-list@python.org>
        <slrnb9h7u4.it3.rmunn@localhost.localdomain>
        <3e9a857b$1@news.012.net.il>
Message-ID: <16026.48686.550571.403229@montanaro.dyndns.org>

    Avner> "Robin Munn" <rmunn at pobox.com> wrote in message
    Avner> news:slrnb9h7u4.it3.rmunn at localhost.localdomain...
    >> Skip Montanaro <skip at pobox.com> wrote:
    >> > Recent versions of Python (2.2 and later) allow this:
    >> >
    >> >     file = file("somefile")
    >> >     for line in file:
    >> >         process(line)
    >> 
    >> Ewww! Don't *do* that! You just shadowed the "file" builtin. Use a
    >> different name for your file object -- I prefer "input_file":

Yes, as I replied to Christian's post, I had one foot in the boat and one on
the dock.  "file" shouldn't have been a variable name.  This would have
worked fine:

    for line in file("somefile"):
        process(line)

presuming you want to slurp the entire file in that one loop (otherwise you
should assign the file object to a variable, just not to a variable named
"file" ;-).

Skip




From dmgriffith at tasc.com  Wed Apr 23 11:53:48 2003
From: dmgriffith at tasc.com (Dang Griffith)
Date: Wed, 23 Apr 2003 11:53:48 -0400
Subject: another python style question -- copying a data file
References: <zTepa.127177$Vzu.30273@news02.bloor.is.net.cable.rogers.com>
Message-ID: <b86cve$ol8$1@gateway.northgrum.com>

Guy Middleton wrote:

> Ok, here's another style question.
> 
> I want to copy a data file, but if the file doesn't already exist, I don't
> care.  I do care about any other errors, such as wrong permissions.
> 
> I have the following code, is this a typical Python way to do this?
> 
>     try:
>         os.stat(DATAFILE)
>         os.system("cp %s %s.new" % (DATAFILE, DATAFILE))
>     except OSError, arg:
>         # ENOENT is ok, we didn't do the copy
>         # anything else is a problem
>         if arg.errno != errno.ENOENT:
>             raise
> 

Look into the shutil module.  It has functions for copying files.
   --dang



From logistix at cathoderaymission.net  Fri Apr 18 17:27:20 2003
From: logistix at cathoderaymission.net (logistix)
Date: Fri, 18 Apr 2003 17:27:20 -0400
Subject: FInd nic speed/duplex (Win32)
In-Reply-To: <3E9EDE85.88D5C0BD@engcorp.com>
Message-ID: <000101c305f1$4b7c4b40$20bba8c0@XP>


> -----Original Message-----
> From: python-list-admin at python.org 
> [mailto:python-list-admin at python.org] On Behalf Of Peter Hansen
> Sent: Thursday, April 17, 2003 1:04 PM
> To: python-list at python.org
> Subject: Re: FInd nic speed/duplex (Win32)
> 
> 
> j2 wrote:
> > 
> > Is there any sample code to be able to find NIC settings (speed and 
> > duplex) for a NIC in a Win98 system?
> 
> You could look in the registry, or there might be some Win32 
> call. Either way, the bulk of the answer is Windows-specific 
> and could be found by searching with Google, probably.  After 
> you have the answer in "Windows-speak", post the APIs 
> required, or the names of the registry keys required, and 
> someone can post example code.
> 

Actually, the exact registry entry will be vendor specific, so don't try
to do that unless you know all the machines you're testing have the
exact same hardware.  In general, NIC cards are too low of a layer on
the OSI model for Windows itself to care about how it's working.  That's
left as an "implementation detail" for the vendor.

However, you might be able to make a good guess by scanning the text
returned by "os.popen('ipconfig /all')"





From davec999 at yahoo.co.uk  Thu Apr 10 07:45:06 2003
From: davec999 at yahoo.co.uk (David Currie)
Date: 10 Apr 2003 04:45:06 -0700
Subject: Zope large PDF question
References: <e09988ac.0304070214.4212e2e9@posting.google.com> <oqh8m-ldq.ln1@news.lairds.org>
Message-ID: <e09988ac.0304100345.122b8ab4@posting.google.com>

Thanks to everyone for the replies.  I fired off the question and then
promptly got sidetracked on other stuff, which is why I haven't
replied before. :)

Kyler Laird <Kyler at news.Lairds.org> wrote in message news:<oqh8m-ldq.ln1 at news.lairds.org>...
> Zope has problems dealing with byteranges the way that
> Acrobat Reader specifies them for "large" files.  I 
> reported this a long time ago, but haven't followed it
> lately.  (I just use my own patch to disable it.)  
> 
> As I recall, Reader sends a request for pieces of a
> large file
> 	1-500, 1001-1500, 501-1000
> and Zope decides to be efficient and join all of the
> contiguous regions.  Unfortunately, that defeats the
> purpose of requesting the ranges *and* breaks Reader.

I see.  Is there any chance of explaining how to disable this
behaviour, or even better, of sharing the patch? ;)

Cheers.

Dave



From exarkun at intarweb.us  Fri Apr  4 12:27:51 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 4 Apr 2003 12:27:51 -0500
Subject: What are the differences between _weak_ references and "normal" references?
In-Reply-To: <169a7283.0304040901.4bc3531b@posting.google.com>
References: <169a7283.0304040901.4bc3531b@posting.google.com>
Message-ID: <20030404172751.GB10506@meson.dyndns.org>

On Fri, Apr 04, 2003 at 09:01:44AM -0800, sdieselil wrote:
> Hi
> 
> I noticed a module weakref in Python standard library. What is weak
> reference and in what way it differs from "normal" reference? 

  Weak references to objects don't increment the refcount for that object.

> By the way, how can I implement usual reference in Python (I mean C++
> reference)?

  You can't, since this requires changing the semantics of assignment.  What
you can do is use wrappers of some form (from the simple, hackish
list-of-length-one to much more complete, OO-style solutions) to move from
name rebinding to attribute modification, giving the same behavior as
C++-style references.

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 15 days, 14:00, 8 users, load average: 0.24, 0.22, 0.09




From aleax at aleax.it  Sat Apr 12 16:08:53 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 20:08:53 GMT
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org> <kauu6b.f2d1.ln@boundary.tundraware.com> <v960rni1esvo94@corp.supernews.com>
Message-ID: <pn_la.7721$T34.214175@news2.tin.it>

Cameron Laird wrote:
   ...
> I'm a (former) mathematician.  I've always taken "1.0"
> to be a local, linguistic act, requiring me to infer
> the speaker's context.  "1.0" means different things in
> the mouths of different people.

I'm an engineer, and I fully concur with you (once you
generalize "different people" to include the SAME people
speaking ``with different hats on'', so to speak).  

When I speak about a given software program, and I assert 
that "1.0 is much better than 1.1, 'cause they added far 
more bugs than features in the upgrade", I expect it to be
quite clear that I'm using "release numbers" -- strange
beasts that can well have more than one decimal point
(e.g., the interlocutor might well object "That's unfair,
you're really compating 1.0.3 with 1.1.0!"...;-).  

In another context, though, I might well use "1.0" as being
distinct from "1.00" because the latter implies I've measured 
"it" (whatever "it" we're talking about) "to two decimal
places" (i.e., in such a context saying "the length is
1.00" DOES mean I claim it's between 0.995 and 1.005,
while if I said "the length is 1.0" I'd be making a far
less precise claim).  That's perhaps the sense in which
some people say that a floating-point number "stands
for an interval" (?).

In another context yet, I might well be using "1.0" as
quite equivalent to "float(1)" (after all, "1.0" IS the
form in which both str and repr render float(1)...:-).
I'd be likely to use that in just about any programming
language because I never remember what languages let me
skip the trailing zero AND "1." is less readable anyway;-).


Alex




From aahz at pythoncraft.com  Fri Apr 18 01:52:55 2003
From: aahz at pythoncraft.com (Aahz)
Date: 18 Apr 2003 01:52:55 -0400
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org>
Message-ID: <b7o3rn$3hc$1@panix1.panix.com>

In article <mailman.1050587068.2733.python-list at python.org>,
Jp Calderone  <exarkun at intarweb.us> wrote:
>
>-- 
>It is practically impossible to teach good programming style to
>students that have had prior exposure to BASIC: as potential
>programmers they are mentally mutilated beyond hope of
>regeneration.        -- Dijkstra

I like to believe that I'm a counter-example to this generalization.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From gh at ghaering.de  Sat Apr 26 01:19:12 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Sat, 26 Apr 2003 07:19:12 +0200
Subject: PoPy (postgresql DA for ZOPE)
In-Reply-To: <3eaa13e8@duster.adelaide.on.net>
References: <3eaa13e8@duster.adelaide.on.net>
Message-ID: <3EAA16D0.9080904@ghaering.de>

Luke wrote:
> Hi, 
> 
> I hope this is not too off-topic,
> 
> I have installed ZOPE 2.6.1 and want to get it working with Postgresql 7.3 
> on my RedHat 8.0 box.
> I undertook to install ZPoPyDA-2.0.8 but it needs the PoPy module.
> [build error] 
> I want to get ZOPE working with Postgresql, 
> any help greatly appreciated,

I might be wrong, but the whole PoPy stuff seems to be little supported. 
For accessing PostgreSQL from ZOPE, I'd recommend you use psycopgDA instead.

-- Gerhard




From stephen at theboulets.net  Tue Apr  1 00:13:19 2003
From: stephen at theboulets.net (Stephen Boulet)
Date: Mon, 31 Mar 2003 23:13:19 -0600
Subject: Regular expression help
Message-ID: <2FGdnaUSmIhkghSjXTWc-g@speakeasy.net>

How do I use the sub function to replace my match with my match 
prepended by a newline?

Thanks.

-- Stephen




From hammer at maxwell.com  Sun Apr 13 16:13:21 2003
From: hammer at maxwell.com (Maxwell Hammer)
Date: Sun, 13 Apr 2003 16:13:21 -0400
Subject: Question regarding indirect function calls
Message-ID: <pan.2003.04.13.20.13.21.504295@maxwell.com>

I am just starting to learn python but I love it already..
Ok the question..
In the program I'm trying to write I have a module called "plugins"
containing a number of functions. I only know at runtime the name of a
function to call from "plugins", i.e. need to call a function indirectly.
The problem is that assigning the name of one of plugin's functions to a
variable, I cannot then call the function. IOW:
 func=function1
 plugins.func() ..........will fail
 plugins.function1() .......will pass

Perhaps the following code snippet explains it better...

 <example>
 import plugins # <--------module containing functions I want to call
                # indirectly, say function "function1"

 def test(job)
 # determine whether the job binary is either an internal plugin
 # or an exernal executable
 job_bin = string.strip(job[2])  <---the name of the actual job is here
 
 if os.access(job_bin, os.X_OK):
   print "external job found"  # job binary is an external program
 else:
   # find out which names our "plugin" module defines
   # and try to find our job in that list of names
   for element in dir(plugins):
    if job_bin == element:
      print "found match for job_bin: ",element
      if callable(plugins.function1):
        print "explicit call: job is callable" <-----------explicit call works
	try:
          plugins.job_bin <---------------------implicit call fails
          print "implicit call: job is callable"
        except:
          print "implicit call: job is not callable"
      else:
        print "explicit call: job is NOT callable"

To show what works and doesn't here's the sample output:
>> 
  found match for function1
  explicit call: job is callable
  implicit call: job is not callable
>>

The above is for illustration only to show that I can only call 
plugins.function1 directly. I have read python.org docs and am also using
book "The Complete Python Reference", but no where can I find the reason
why my code fails.

Thanks in advance for your help




From aleax at aleax.it  Tue Apr 15 13:25:57 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 19:25:57 +0200
Subject: pread/pwrite (was Re: files: direct access - NOT sequential)
In-Reply-To: <20030415161936.GA30739@unpythonic.net>
References: <mailman.1050398560.8881.python-list@python.org> <S_Sma.22945$T34.690529@news2.tin.it> <20030415161936.GA30739@unpythonic.net>
Message-ID: <200304151925.57736.aleax@aleax.it>

On Tuesday 15 April 2003 06:19 pm, Jeff Epler wrote:
> On Tue, Apr 15, 2003 at 12:33:54PM +0000, Alex Martelli wrote:
> > Patrick Carabin wrote:
> > > i search a means to get " direct access " inside files, eg mixing
> > > read & write operaions for the same file
> > > example:
> > > given a file with contents : "qwerty" , positionning on the "w", and
> > > writing a "N" , so the file becomes "qNerty"
> > > How can i achieve this in Python ?
> >
> > thefile = open('thefile', 'r+')    # must be already-existing
> > thefile.seek(1)
> > thefile.write('N')
> > thefile.close()
> >
> > Here, in fact, we're NOT mixing read and write operations -- the
> > only _operation_ is a write -- but, if we wanted to, we might (that
> > is what mode 'r+' means).
>
> Speaking of which, (if a patch were supplied,) is it remotely likely
> that the os module would get pread/pwrite functions?  These make
> lseek+read or lseek+write atomic, which is useful if two threads share
> the same file descriptor and want to use random-access to it in a
> threadsafe way.

I know Guido is reconsidering the whole structure of fileobjects, so
it is quite possible he might entertain the idea (for Python 2.4 -- I
think it's too late for such major surgery in Python 2.3!) -- I think a
clear suggestion to python-dev, with examples and use cases, might
be more useful in this case than a patch, given the prospect of
pensioning off the whole current implementation of file objects.


> (nicest would be if file.read and file.write took an optional pos=
> argument, but pread/pwrite are for file-descriptors, not FILE*s so stdio
> buffering would louse things up)

I think the positioning argument[s] should be the same as for .seek --
ONE argument (unless it's a tuple specifying origin and offset) would
not suffice.  Consider a frequent case: "I want to write these bytes
at the END of the file" - no matter how long it's now and where it's
currently positioned; or, ditto but "at the START of the file"; I don't
think the Pythonic convention of -1 to meand end would work here,
i.e. to write _after the end_ -- I think that, like for seek, offset and
indicator of origin (start, end, current) would in general be needed.

I don't know if adding optional argumens to write and read is the
right way, or rather new methods should be used -- I do think that
the key issue is defining the syntax, semantics and use cases, as
opposed to implementing via a patch (buffering might not be a
problem, if such operations implied previous flushing of any
pending buffers or semantics equivalent thereto, for example).


Alex





From andrew-pythonlist at puzzling.org  Wed Apr 30 19:14:54 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Thu, 1 May 2003 09:14:54 +1000
Subject: looping through a file
In-Reply-To: <iuTra.42265$K35.1222818@news2.tin.it>
References: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au> <uWOra.24347$3M4.683692@news1.tin.it> <mailman.1051711067.28906.python-list@python.org> <iuTra.42265$K35.1222818@news2.tin.it>
Message-ID: <20030430231454.GA21890@frobozz.local>

On Wed, Apr 30, 2003 at 05:12:46PM +0000, Alex Martelli wrote:
> Andrew Bennetts wrote:
>     ...
> > 
> > I thought there might be something in 2.3's itertools module to help with
> > this -- I was hoping you'd be able to write something like:
> > 
> >     InFile = open('users.dat', 'r')
> >     
> >     for username, password in multistep(InFile, 2):
> 
> Careful!!!  You'd still need to strip them before you compare them,
> else the trailing \n or any other whitespace will make the comparisons
> fail.  imap(str.strip, multistep(InFile, 2)) might work here (if
> multistep existed).

Oops!  Yes, I forgot about that.

> > Where "multistep" would probably have a better name, and be defined like:
> > 
> >     def multistep(iterable, count):
> >         # XXX: What should happen when len(iterable) % count != 0 ?
> 
> I think the last trailing part X with 0<=len(X)<count should be dropped.
> Otherwise, unpacking assigments such as the above would fail -- and,
> "padding with None's" isn't in fashion any more;-).

That's my inclination too, because it's the easiest thing to do :)

> >         iterator = iter(iterable)
> >         while 1:
> >             values = []
> >             for i in range(count):
> >                 values.append(iterator.next())
> >             yield values
> 
> I think I'd prefer expressing this loop (naming issues apart) as:
> 
>    while True: return [iterator.next() for i in range(count)]

I think you meant to use "yield", not "return".

But yes, that occurred to me after I posted :)   It's much nicer that way.

> > 
> > I couldn't see anything in itertools to save me writing that out.  But I
> > just figured out how to do it with itertools -- izip can do it:
> > 
> >     def multistep(iterable, count):
> >         return itertools.izip(*((iter(iterable),)*count))
> > 
> > I guess you could write this directly:
> > 
> >     iterator = iter(InFile)
> 
> Unneeded, btw: iter(inFile) is inFile (I also exploit that in my
> less-elegant snippet above where I call inFile.next(), secure that
> this will consume a line).

Hmm.  I couldn't remember if that was safe or not; I tested this stuff
interactively with xrange, and 

    x = xrange(10)
    print list(izip(x, x))

didn't do what I expected until I remembered to change the first line to

    x = iter(xrange(10))

So I played it safe.

> >     for username, password in izip(iterator, iterator):
> >         ...
> > 
> > So it's probably not worth defining a seperate function...
> 
> I think it is (if anything can still be added to new module itertools
> in 2.3 -- I'm not sure ANY functionality can be added after beta-1), as
> it's frequently needed.  But, if it isn't, the "trick" with the
> iterator explicitly repeated is good and IMHO worth keeping in mind.
> 
> But one does still need the .strip call...;-)

Yes :)

-Andrew.





From jordan at krushen.com  Fri Apr  4 01:07:39 2003
From: jordan at krushen.com (Jordan Krushen)
Date: Fri, 04 Apr 2003 06:07:39 GMT
Subject: Is there a python realaudio player? For windows?
References: <b6i6uq$rcs$0@216.39.172.122>
Message-ID: <Li9ja.28908$vs.1947268@news3.calgary.shaw.ca>

Bengt Richter wrote:
> I recently wanted to play a .ram file without having to accept
> all the baggage around the minimal player capability. Is there
> such a thing? Even in C/C++, in case I got motivatged to write
> a wrapper? TIA.

Check out https://www.helixcommunity.org/ for most things related to Real
Networks and Open Source.  Specifically,
https://www.helixcommunity.org/2002/intro/client.

J.





From grante at visi.com  Fri Apr 25 10:36:17 2003
From: grante at visi.com (Grant Edwards)
Date: 25 Apr 2003 14:36:17 GMT
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com>
Message-ID: <3ea947e1$0$198$a1866201@newsreader.visi.com>

In article <vai9f3ug54144 at corp.supernews.com>, Cameron Laird wrote:

>>That's what I had in mind, but after reading some tutorials on
>>Tkinter I saw that wxpython and pyqt have more features still
>>remaining portable...mmm...

[...]

> I think it's reasonable for you to start with Tkinter.

Tkinter is indeed a pretty good place to start -- just try not
to think about the fact that you're pulling in a Tcl
interpreter also.  ;)

-- 
Grant Edwards                   grante             Yow!  Why don't you
                                  at               ever enter and CONTESTS,
                               visi.com            Marvin?? Don't you know
                                                   your own ZIPCODE?



From BPettersen at NAREX.com  Mon Apr 28 19:05:06 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Mon, 28 Apr 2003 17:05:06 -0600
Subject: Question about accessing class-attributes.
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE14C@admin56.narex.com>

> From: Alex Martelli [mailto:aleax at aleax.it] 
> 
> Bjorn Pettersen wrote:
>    ...
> > Then define a hierarchy:
> > 
[...]
> >  class CountedPoint(Counted, Point):
> >      def __init__(self):
> >          super(CountedPoint.self).__init__(self)
> 
> hmmm. I think you want a comma there, not a dot, in the super args.
> And you shouldn't pass self as an argument here, either.  Just:

Sorry about the dot, a little premature copying. That __init__ doesn't
take self is an interesting artifact of the implementation [super(..)
seems to want to be object substitutable, but only is some of the
time...]

[...]
> > metaclasses aren't related by inheritance. Trying to fix 
> > this by, e.g.
> 
> "Related by inheritance" may be a bit too weak.  More specifically:

Yes, but not in this case, which is why I didn't want to say:

[..for all B in C.__bases__, C.__metaclass__ must be subclass of
B.__metaclass__...]

Not that I didn't know them, I just haven't figured out why they must be
so restrictive ? I'm guessing it has to do with attribute lookup rules,
however I haven't been able to find a good description of attribute
lookup for new-style classes... (references?) E.g. why is the metaclass'
__getattr__ never called?

[..]
> > making metaIC a subclass of metaMetaBunch (and changing the 
> > __new__ call appropriately), gives you:
> > 
> > Traceback (most recent call last):
> >   File "M:\python\cpoint.py", line 38, in ?
> >     class CountedPoint(Counted, Point):
> >   File "M:\python\cpoint.py", line 25, in __new__
> >     return metaMetaBunch.__new__(cls, name, bases, dict)
> >   File "M:\python\cpoint.py", line 20, in __new__
> >     return type.__new__(cls, classname, bases, newdict)
> > TypeError: multiple bases have instance lay-out conflict
> > 
> > which means you're stuck (?)
> 
> Well no -- you MIGHT be stuck if you ever needed two different AND
> incompatible __new__ in the same class, but here metaIC may just as
> well use __init__ instead (as it ain't messing with e.g. __slots__:-)
> so everything works if you recode the whole mess as, e.g.:

Which is fine as long as you have access to the code of both the Point
and Counted classes and their metaclasses, understand metaclasses in
general, the ramifications of the particular implementation and how
other things would break if you started changing things, and even then
you're stuck if they both need __new__, unless you can code up a
compromise. Which was basically my only point :-) While the abstract
base class solution does have its problems, so does the metaclass
solution (I wouldn't consider mucking with superclasses' implementation
just because you need to multiply inherit as beeing in the OO spirit
<wink>) -- the metaclass solution is aesthetically more pleasing though
:-)

[..nice self consistent reimplementation..]

> But given the dependencies I don't think this particular mixin
> mesh makes much sense.  I'd rather merge two metaclasses into a
> third one (solving any __new__-related conflicts of course;-)
> rather than arbitrarily deriving one from the other (and it IS
> arbitrary -- now you're forcing ANYTHING that's counted to be
> SOME kind of metabunch...).

Well, _I_ just wanted a countable point, someone else forced me into
hacking "magic code" just because I wanted to use regular inheritance
<grin>.

-- bjorn




From mis6 at pitt.edu  Mon Apr 28 08:46:26 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 05:46:26 -0700
Subject: dynamic function names
References: <20030428102400.21c5074a.bram@nospam.sara.nl> <iP5ra.14084$3M4.382689@news1.tin.it>
Message-ID: <2259b0e2.0304280446.5e38f9af@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<iP5ra.14084$3M4.382689 at news1.tin.it>...
> <posted & mailed>
> 
> Bram Stolk wrote:
> 
> > Hello,
> > 
> > Is it possible in Python to define functions with a dynamic name?
> 
> Yes, but not with the def statement.
> 
> 
> > something like:
> > 
> > def make_func(funcname) :
> >   def funcname :
> 
> Syntax error -- you're missing parentheses here before the colon.
> 
> >     print "This func is named", funcname
> > 
> > make_func(foo)
> > make_func(bar)
> > 
> > The code above fails due to foo being unknown,
> > however, when calling as make_func("foo"), the "foo" string
> > cannot be used as a funcname.
> 
> Actually, if you fix the syntax error it will work, but not do
> what you want:
> 
> >>> def makefunc(funcname):
> ...   def funcname(): print "this function is named", funcname
> ...   return funcname
> ...
> >>> f = makefunc('foo')
> >>> f()
>  this function is named <function funcname at 0x402c09cc>
> >>>
> 
> See?  "def funcname():" *RE-BINDS* identifier funcname to
> refer to the function object.
> 
> So, what you need is to call function 'function' in module
> 'new'.  You can first build a temporary function with a def
> to give you the right code, globals, defaults and closure,
> and then pass each of these to new.function together with
> the funcname you want to use:
> 
> >>> import new
> >>> def makefunc(funcname):
> ...   def temp(): print "this function is named", funcname
> ...   f = new.function(temp.func_code, globals(), funcname,
> ...           temp.func_defaults, temp.func_closure)
> ...   return f
> ...
> >>> f = makefunc('foo')
> >>> f()
>  this function is named foo
> >>> f.func_name
>  'foo'
> >>>
> 
> 
> Alex

In Python 2.3 most types are callable and one can often avoid the 'new'
module. In this example 

new.function(temp.func_code, globals(), funcname,
             temp.func_defaults, temp.func_closure)

could be replaced with

types.FunctionType(temp.func_code, globals(), funcname,
                   temp.func_defaults, temp.func_closure)


In addition to that, I have a question I always wanted to ask ;) 
why I can change the name of a class (es. C.__name__='newC' works)
and I cannot change the name of a function (f.__name__='newf' does not
work) ?



From imbosol-1050339953 at aerojockey.com  Mon Apr 14 14:09:25 2003
From: imbosol-1050339953 at aerojockey.com (Carl Banks)
Date: Mon, 14 Apr 2003 18:09:25 GMT
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050307807.19679.python-list@python.org>
Message-ID: <pPCma.12520$883.10698@nwrdny02.gnilink.net>

Christian Tanzer wrote:
> All these functions (except the one you're writing) might have been
> implemented years before the text properties and
> `inhibit-point-motion-hooks` were introduced.

See, that's really the problem.  It's obvious why history made Emacs
what it is, but it doesn't change the fact that, as it stands now,
Emacs is a terrible software design.

Now, Emacs has a variable inhibit-point-motion-hooks.  If Emacs were
currently designed well, this variable would not be needed, because
functions would be able to operation without touching the point.

Now, Emacs has a current bufferm, and a current point in a buffer,
which all text operations take place it.  Right now, you have to use
macros such as save-excursion and with-current-buffer, because right
now, text manipulation functions implicitly use the current buffer and
current point.  If Emacs were currently designed well, functions would
accept a buffer to work on as an argument, and there would be no need
to change the buffer temporarily.

And so on.

So, basically, what I'm saying is, most of the time when you need
these implicit parameters is because Emacs isn't a good design.

A redesign in Python does not need to make the same mistakes.

(As for things like buffer-local variables, I don't think it's a big
deal to make them attributes of a buffer class.  Frame local
variables, frame objects.  I think it's quite short-sighted to try to
force buffer-local variables to be regular variables, and not what
they obviously should be, attributes of the buffer, just because Emacs
does it that way.)



-- 
CARL BANKS



From tjreedy at udel.edu  Tue Apr  1 19:51:30 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 1 Apr 2003 19:51:30 -0500
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <3E8974E5.88B861D7@engcorp.com> <435d40e0.0304011457.56e32bd0@posting.google.com>
Message-ID: <sKSdnfwAUb0mrxejXTWcog@comcast.com>

> What are the reasons for not having multi-line comments?

Because we more or less have them already -- the compiler can dump '''
''' strings that are not assigned to anything or otherwise used.

> I just remembered that IDLE will color-code """ as a string and not
a
> comment - even when it is intended only as a comment.

Unlike the compiler, IDLE does not know the differrence.

Once could propose that ###, for instance, delimit multiline comments,
but I suspect Guido would say, "Why bother?".

TJR





From aleax at aleax.it  Mon Apr 21 12:02:27 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 21 Apr 2003 16:02:27 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net>
Message-ID: <nCUoa.41676$LB6.941312@news1.tin.it>

Steve Holden wrote:
   ...
>> Interesting... I have my doubts on whether Guido would
>> accept such a patch just a week before a beta, but maybe
>> I should try submitting it anyway...
> 
> Only one way to resolve those doubts.

Yup -- proof of the pudding, and all that.  So, I submitted it and
Guido accepted it for 2.3, EXCEPT...:

> The use of sum() to concatenate a list (please, not a tuple!) of strings
> in such a deterministic way is a masterstroke. It would surely deserve to
> become immediately idiomatic.

...that sum(manystrings) was rejected -- now, when called with any
sequence of strings, sum will raise a TypeError, and remind the user
to use ''.join instead.  Not sure why you wanted sum to differentiate
between its argument being a list or tuple when all similar built-ins 
these days accept any iterable, but anyway the problem that led to
this came when the argument is an *iterator* -- sum had to jump through
weird hoops to guarantee correct behavior on an iterator (particularly
one freely mixing unicode and non-unicode strings), leading to a factor
of two slowdown with respect to ''.join, and still couldn't ensure the
same semantics as a loop with += in sufficiently weird cases such as:

class Gotcha:
    def __add__(self, *whatever): return 23

weirdlist = [ 'xx', Gotcha() ]
print sum(weirdlist)


This _should_ have printed 23, just like reduce(operator.add,weirdlist)
does -- however, due to the string special-case, it ended up delegating
to ''join and thereby raising a TypeError instead.  All in all, it did
make much more sense to forbid the first item from being a string (so
as to avoid the performance trap of strings-concatenation done by +).

Interestingly, and characteristically, Tim Peters had expressed strong
instinctive dislike to sum being applicable to (and indeed special-case
optimized for) sequences of strings, even before we had unearthed these
weird examples making the prohibition well-nigh inevitable.  timbot
argued that "sum is about numbers" -- and soon he was proved right.


As a bonus -- sum([]) now returns 0, while in my original design it
raised a ValueError, just like e.g. max([]) does.  I thought raising an
error was best in the name of "errors shouldn't pass silently" and "in
the face of ambiguity, refuse the temptation to guess".  But -- once
we decree that "sum is about numbers", there's no real ambiguity, so we
can have the little extra handiness of having sum([]) not be an error;-).


Alex




From mcfletch at rogers.com  Tue Apr 15 16:32:34 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 15 Apr 2003 16:32:34 -0400
Subject: Big-O notation (was: Re: Python and Schools)
In-Reply-To: <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
Message-ID: <3E9C6C62.8010808@rogers.com>

Greg Brondo wrote:
...

>
> I understand why the code is bad (copying the string each time in
> memory as it grows).  However, I've never understood the notation's
> O(N) and O(N2)?

...

Here's a crack at explaining it:

    Characterisation of an algorithm's (as distinct from an
    implementation's) performance in relation to the size of the input
    set.  Normally of interest in determining the
    scalability/suitability of a particular algorithm.  The static (e.g.
    setup) and proportional (e.g. implementation) factors in the
    algorithm are factored out under the assumption that they are
    swamped by the relative effects of the size of the input set as that
    goes to infinity.

Common examples of big-O performance types:

    * O(1) -- time is linear, that is, it takes the same amount of time
      to do something regardless of how large the input set is.
    * O(log N) -- time grows as the log of N, (for those who don't
      remember high-school math, this means that it grows very slowly
      compared to N).
    * O(N) -- time grows linearly as the size of the set grows (e.g.
      process every item in the set in turn)
    * O(N log N) -- see O(N) and O(log N), basically, you do an amount
      of work tied to the size of the whole set for every stage in an O(
      log N ) operation (this is comparatively rarely discussed in my
      experience).
    * O(N**2) O(N**3) etceteras -- time grows as an exponent (e.g.
      2,3,3.4,10) of the size of the set, so you get geometric explosion
      of running time as set size increases.  These algorithms tend to
      fail on even small N values.

You can have all sorts of fun learning how to combine and re-factor the 
equations to get a final big-O notation, but for the most part, 
practical coders can get away with recognising and avoiding O(N**X) 
where X > 1 situations.  Recognising the underlying patterns in O(log N) 
(binary searches and the like), and O(1)-ish situations is useful as 
well, as they provide very good general scalability.  The calculus of 
big-O notations is just a convenient shorthand for people doing rigorous 
analysis AFAICS.

Keep in mind that the setup and constant factors of the algorithm are 
factored out of the equations.  Something which is O(1) may be slower 
than O(N**3) for all practical values of N if the setup time for the 
O(1) algorithm approaches the age of the universe.  Similarly, two 
algorithms which are O(N), but with vastly different constant factors 
(implementations) may have dramatically different running times.

HTH,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From aleax at aleax.it  Mon Apr 28 04:33:54 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 08:33:54 GMT
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <3EAC2962.2050609@soraia.com> <3EAC31B0.43C50CDC@sympatico.ca>
Message-ID: <SH5ra.19700$K35.664227@news2.tin.it>

Colin J. Williams wrote:
   ...
>> > other hand, is ideal.  If I can find a textbook that is even ok, I may
   ...
>> The Deitels have a textbook called _Python How to Program_ that is aimed
>> at beginning programmers.  I'm not sure how good/bad it is.  Three out
>> of four reviewers on Amazon seem to like it (the fourth guy gave it one
>> star based on its lack of Perl coverage).
> 
> The book is very expensive and heavy reading.  The Nutshell is excellent,

Why, thanks!

> but would need more examples for your purpose.

...AND it would need EXERCISES, that it completely lacks... it's *not*
meant as a textbook.  Besides, while I've been quite pleased to hear that
programmers already experienced in other languages have found the Nutshell
to be a good introduction to Python (I aimed for that, as a second order
issue -- I'm glad I managed to achieve it), I think it's out of the question
that the Nutshell can be usable as an introduction by *beginners* -- people
who do NOT already have substantial programming experience (albeit in other
languages).  There are many excellent books I would recommend as beginners'
introductions ("Practical Python" and "Learning Python" are two that have
already been mentioned in this context, and I would second these -- but, the
book I chose to give as a birthday gift to a bright beginner who was 
starting to learn Python was Holden's "Python Web Programming" instead!-).
But, none of these is really a *textbook*, either...


Alex




From andrew-pythonlist at puzzling.org  Sun Apr 27 04:30:31 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 18:30:31 +1000
Subject: Forgetting "()" when calling methods
In-Reply-To: <9%Lqa.13853$K35.457009@news2.tin.it>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <mailman.1051422522.15507.python-list@python.org> <9%Lqa.13853$K35.457009@news2.tin.it>
Message-ID: <20030427083031.GO26530@frobozz.local>

On Sun, Apr 27, 2003 at 07:52:37AM +0000, Alex Martelli wrote:
> Andrew Bennetts wrote:
>    ...
> > lack of a better word.  It's usually not any clearer to say
> >     
> >     if len(myList) == 0:
> >         print "I have a list of stuff"
> 
> The "== 0" part here is making it to the reverse of what you probably mean.

Oops, yes.  Thanks.  Another reason not to spell it that way ;)

> > than
> > 
> >     if myList:
> >         print "I have a list of stuff"
> 
> Absolute agreement here.

It's good to know we have at least *some* common ground :)

> > I think the same applies to testing to see if a variable holds a callable,
> 
> Not at all.  "if callable(thevariable):" is the one obvious way to check
> THAT.  "if thevariable:" is going to have you try to call the number 23,
> the string 'loolapalooza', and non-empty dicts and lists.

I disagree.  Attempting to call any of those things will fail immediately,
and with an obvious error message.  I see no need for LBYL here.

> > just as much as it applies to checking the length of something.  Removing
> 
> The two cases are very different.  "if len(myList):" is equivalent to
> "if myList:" *for a known given type of myList* -- a container.  In
> the callable case, you _don't know_ the type -- and are checking the
> VALUE to try and infer the TYPE.  Checking callable(f) is more obvious
> and direct -- it clearly and unambiguously shows what you're checking.

I don't think they're different at all.  If 

    if x:
         val = x[0]

is ok, why not

    if x:
         val = x()

as well?  If x isn't a container, and it is expected to be, then x[0] will
break -- and it can also break in the case where x is a dict, and len(x)
wouldn't necessarily help there.

Similarly, callable(x) doesn't guarantee that you can successfully call x
with some arbitrary arguments.  The only way is to try.  I don't see any
difference here at all.

In my opinion, "callable(f)" also clearly and unambiguously clutters the
code.  Maybe I'm overly sensitive, but I find "if func:" much more pleasant
to the eyes and faster to read than "if callable(func):".  The extra
punctuation and syllables get in the way of the meaning.

In both English and programming, I prefer to use the simplest, most direct
way of expressing something that I can.  I'm sure that most of the time I
could do better, but I try, and I think both my prose and my code is better
for it.  Requiring "if callable(func):" rather than "if func:" goes against
this instinct, with the only justification being to prevent an error that
has struck me only very rarely, and never proved confusing or difficult to
debug.  I don't like the tradeoff.

> > this feature will lead to unnecessarily longer code for minimal benefit.
> 
> I dispute the "unnecessarily" and the "minimal".  Tests for "is None"
> add clarity (when they're applicable), and tests for callable(f) even
> more so.  And giving newbies a little help by diagnosing a very common
> newbie error more promptly would help (or at least it might, were it
> not for some newbies' insistence that the diagnosis should be performed
> by static analysis instead -- it just can't, and thus said insistence
> convinces me that this is not worth putting to a PEP -- however, I find
> the arguments against it quite unconvincing, except for the "we can't
> break this idiom because even though it's inferior it's widely used").

Personally, most of the time I use "if x:" rather than "if x is not None:",
even if I know the only expected false value will be None.  Why?  Because,
to me, adding the three extra words slows down comprehension -- particularly
the "not" -- and obscures the essence of what I'm trying to say.  So I
disagree that tests for "is not None" add clarity.  I reserve testing
against None to those times when I *really* need it, so that when I see it
in my code, I immediately recognise there probably an intentional
distinction between None, and 0 or '' or [].

I don't teach newbies, so I can't really comment on how much of an
impediment this is to them.  But recalling my own experiences, and the few
times I have helped newbies, I can't say that I remember it ever being an
issue.

So I disagree with you that "if x:" is necessarily inferior to "if
callable(x):" or "if x is not None:".

So I stick by my assesment: it is an unnecessary change will at best minimal
benefit.  I've never been seriously inconvenienced by callables having
__nonzero__.  I feel I would be inconvenienced regularly if they suddenly
became the only Python type that didn't.

-Andrew.





From ckea25d02 at sneakemail.com  Sun Apr  6 22:12:12 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 14:12:12 +1200
Subject: Roman assignments (Re: Overloadable Assignment PEP)
In-Reply-To: <a8b7f07a.0304040756.61276087@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <a8b7f07a.0304031432.3c585657@posting.google.com> <b6in8i$600fi$1@ID-169208.news.dfncis.de> <b6j2ks$8gg$1@panix2.panix.com> <a8b7f07a.0304040756.61276087@posting.google.com>
Message-ID: <b6qm8g$7u3sb$1@ID-169208.news.dfncis.de>

A. Lloyd Flanagan wrote:
> This is brilliant!!  But how are the implementors of the sets module
> going to feel?  They could have skipped all that work and just done
> 
> from __past__ import Pascal

No, we're not allowed to import from __past__, so
it would have to be

   from __languages_named_after_famous_french_mathematicians__ \
     import Pascal

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From berkowit at silcom.com  Tue Apr 15 01:31:43 2003
From: berkowit at silcom.com (Paul Berkowitz)
Date: Mon, 14 Apr 2003 22:31:43 -0700
Subject: Built-ins vs. modules
Message-ID: <BAC0E74F.35089%berkowit@silcom.com>

On 4/14/03 9:02 PM, in article
mailman.1050379428.20298.python-list at python.org, "Tim Peters"
<tim_one at email.msn.com> wrote:

> [Paul Berkowitz]
>> I see from the Python Library Reference (for 2.2) that a lot of
>> things which I thought required a module to be imported can also be
>> accessed without doing so:
>> 
>>>>> "boo".capitalize()   #built-in
>>     'Boo'
>>>>> import string
>>>>> string.capitalize("boo")
>>     'Boo'
>> 
>> Why import the string module if you can do the same things anyway?
> 
> Nostalgia; inertia; laziness; perverseness; whimsy; Luddism.
> 
>> The page of built in String Methods bears an uncanny resemblance to the
>> page on the string module. Is there a historical reason?
> 
> Yes:  string methods didn't exist before Python 1.6.  The string module has
> always been there.
> 
>> Are the built in type methods quite recent?
> 
> Python 1.6 was released in 2000.  Whether that's quite recent depends on
> whether you reckon in, say, geological time or computer time <wink>.
> 
> If you like investigating this kind of thing, see Misc/HISTORY and Misc/NEWS
> in your Python distribution (the latter gets prepended to the former as the
> centuries fly by).
> 
> 

Thank you - it's a privilege and pleasure to hear it from - I guess I'm not
supposed to say it. ;-) I've just been learning Python from "Learning
Python" which was written as v1.5-1.5.2 were prevalent, so naturally had
nothing about the built-in string methods.

I can see how much more there is to learn from the online Documentation,
which seems very thorough and clearly written. It's amazing how much has
been built up in such a short time.  There's a lot to learn. Thanks for the
History tips.

-- 
Paul Berkowitz






From heiko at hhenkelmann.de  Wed Apr  9 15:27:12 2003
From: heiko at hhenkelmann.de (Heiko Henkelmann)
Date: Wed, 9 Apr 2003 21:27:12 +0200
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com> <3E935B75.64968B9@engcorp.com>
Message-ID: <b71s3u$lf3$01$1@news.t-online.com>

to be more precise: The fastest way...

Thanx for all of the answers


"Peter Hansen" <peter at engcorp.com> schrieb im Newsbeitrag
news:3E935B75.64968B9 at engcorp.com...
> Heiko Henkelmann wrote:
> >
> > I need to split a string into an array of equally long substrings, like
> > e.g.:
> >
> > '1234567890'   -> ['12', '34','56','78','90']
> >
> > What is the most efficient way to do this?
>
> Define "efficient" ...
>
> And see also http://groups.google.ca/groups?th=7848e47f9918f3f7
>
> -Peter





From password.april03.20.netman at spamgourmet.com  Mon Apr 21 22:17:05 2003
From: password.april03.20.netman at spamgourmet.com (Chris)
Date: Tue, 22 Apr 2003 02:17:05 GMT
Subject: When not to use an RE -- an example
References: <c76ff6fc.0304191459.2adf73a@posting.google.com>
Message-ID: <BC1pa.23530$5f4.9990@twister.nyroc.rr.com>

What's wrong with:

    def findrepeat(s):  return len(s) > 1 and not filter(lambda x: x !=
s[0], s)

Chris



"John Machin" <sjmachin at lexicon.net> wrote in message
news:c76ff6fc.0304191459.2adf73a at posting.google.com...
> I needed a check for strings consisting of repeated characters -- like
> when users type "ZZZZZZZ" instead of "UNKNOWN" into a database field.
> After implementing the obvious overlapping-substring comparison, I got
> to thinking how this could be done with REs. The following resulted:
>
> import re
> repeats1 = re.compile(r"^(?:(.)(?=\1))+\1\Z", re.DOTALL).match
> def repeats2(s):
>    return len(s) > 1 and s[1:] == s[:-1]
> for testvalue, expected in zip(
>    ['','x','xx','xxx','xxxxxx','xy','xxy','xyy','\n\n\n','aaa\n'],
>    [0,  0,  1,   1,    1,       0,   0,    0,    1,       0     ]):
>    print repr(testvalue), not not repeats1(testvalue),
> repeats2(testvalue), expected
>
> Pergly/phugly, eh? Note the effort required to ensure the newline
> cases worked.





From DennisR at dair.com  Thu Apr 24 16:51:30 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Thu, 24 Apr 2003 20:51:30 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com> <mailman.1050978337.8048.python-list@python.org> <87wuhm2rsu.fsf@pobox.com> <Bkipa.1154$Hm6.104194064@newssvr14.news.prodigy.com> <87adefkbs9.fsf@pobox.com>
Message-ID: <m7Ypa.156$Jr1.14453940@newssvr15.news.prodigy.com>

> Yes, I read your post.  You said you did some kind of manual hack to
> package your app,

I have posted a 390 KB starter Python Hello World that runs on a Windows
computer without Python at: http://www.spamai.com/hello/hello.zip.
Decompress to a directory and click on runmain.bat.  The runmain.bat file
contains:

    bin\python.exe bin\main.py
    exit

The main.py program is not too complicated either:

import sys
sys.path = [".", ".\\bin", ".\\bin\\libstd", ".\\bin\\libxtra"]

import os

def write_file_basic (filepath, filecontents):
    emfile_open_flags = os.O_WRONLY+os.O_CREAT+os.O_TRUNC+os.O_BINARY
    emf = os.open(filepath, emfile_open_flags)
    os.write(emf, filecontents)
    os.close(emf)

html= """<html>
<head><title>Hello from Python</title></head>
<body bgcolor=#e8fbfe>
<h1>Sss...</h1>(how Pythons say "hello")<p>
current directory: %s
</body></html>
""" % os.getcwd()

write_file_basic("hello.html", html)
os.startfile("hello.html")
sys.exit(0)

The path bin/libstd contains the library files needed to import os.  If you
call more functions, add to path bin/libstd or bin/libxtra.

Writing an html file and then shelling to it is just to get started.  For my
own app, I wrote a socket-level http server and call that.

Maintaining this "manual hack" is far less effort than you would guess.  I
have not had to revise the library files I use for months now.  It is very
clear when one is missing and what needs to be added.

> but clearly you have some system for your web page
> packager (neat, BTW :-), which you didn't specify.
Errr..., that's right, I didn't specify how I did it.  I have abandoned this
particular approach myself and describing what is going on is tough   There
is lots of non-portable stuff in there.

Instead, I have posted code for the approach I am actually using above.

The general approach for the web page is that I coded a stub in my own
compiler language (www.pld2.com) which embedded python and library code.  At
run time, this is extracted from the exe and run.  The exe file has internal
pointers which the online Perl code can use to patch the stub code to
produce a custom exe.  Cross-program dependencies abound.

> So, does your thing work better on space than the alternatives?  Maybe
> it's worth making it available if so...
Yes 390 KB is smaller than any other solution I am aware of.  I just have
made it available.  Thanks for you interest.

Oh, btw, I made a zip file available above because lots of people can access
this format.  The same file compressed with WinRar in rar format is about 80
KB smaller. WinRar provides a self-extracting format also.
--

Dennis Reinhardt

DennisR at dair.com   http://www.spamai.com?ng_clp





From vlindberg at verio.net  Wed Apr  9 19:26:17 2003
From: vlindberg at verio.net (VanL)
Date: Wed, 09 Apr 2003 23:26:17 GMT
Subject: limitation  in urllib.urlretrieve(url, fn) ??
In-Reply-To: <10caf2e2.0304091421.2ec9bcbc@posting.google.com>
References: <10caf2e2.0304091421.2ec9bcbc@posting.google.com>
Message-ID: <3E94ABF4.5080504@verio.net>

There isn't a general python limitation;  I have used urllib.urlretrieve 
to automate the downloading of files as large as 50 MB with 
python2.2.2/freebsd 4.3.  Can't comment about Solaris, tho.

VanL

HP wrote:
> For example,
> url = 'http://www.londonstockexchange.com/techlib/word/ser/Live8102.doc';
> fn  = '/local/disk/file_name';
> 
> urllib:urlretrieve(url, fn) hangs in the air !!
> 
> This happens for many of the files in the above 
> web site.  Those files have one thing in common:
> their file size is larger than 200K.
> 
> Is there a limit on the file size in the function:
> urllib:urlretrieve(url, fn) ??
> 
> (I am running python.2.2.2 on Sun Solaris 5.8).
> 
> HP




From drew at astro.pas.rochester.edu  Thu Apr  3 22:39:14 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 3 Apr 2003 19:39:14 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <mailman.1049398267.20730.python-list@python.org>
Message-ID: <9383481e.0304031939.320c661b@posting.google.com>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> wrote in message news:<mailman.1049398267.20730.python-list at python.org>...
> drew at astro.pas.rochester.edu (Drew Moore) wrote previously:
> |voltage += 5  # (raise the voltage by 5 volts)
> |voltage *= 2  # (double the current voltage)
> |voltage = 3  # (set the voltage to 3 volts)
> |my voltage object is clobbered with an integer object.
> 
> The case makes sense... but IMO, not nearly enough to change the
> behavior of assignment.
> 
> For this particular case, you could simply coopt the meaning of some
> other augmented operator to set voltage.  For example:
> 
>     voltage |= 3
> 
> The need to bitwise-and on voltage seems unlikely, so you should not
> step on anything important.  Of course, experienced Python programmers
> might do a little double take on the meaning, but documentation can
> explain it.
> 
> Then again, I wouldn't mind adding the new operator ":=" which was
> "generic augmentation"... the operator could do -absolutely nothing- for
> standard types, but could be available for custom meaning in custom
> types.
> 
> Yours, Lulu...

Use a different operator? not a bad idea..

I been there.. kinda. I used 

voltage << 3
which was a little more intuitive to me syntactically..

However I still found myself accidentally typing 
voltage = 3 every once in a while, as that makes the most "sense."


Drew



From skip at pobox.com  Thu Apr 10 12:14:36 2003
From: skip at pobox.com (Skip Montanaro)
Date: Thu, 10 Apr 2003 11:14:36 -0500
Subject: python 2.3a2 Win32 expected failures in regression tests?
In-Reply-To: <e2d925a8.0304100803.46e63240@posting.google.com>
References: <mailman.1049910795.4707.python-list@python.org>
        <e2d925a8.0304100803.46e63240@posting.google.com>
Message-ID: <16021.39020.38164.886501@montanaro.dyndns.org>

    Harri> I poked around Python CVS, but I did not find any Changes.log, or
    Harri> equivalent, to say what has changed since 2.3a2, or from 2.2.2 to
    Harri> that. What would be the best way to find out?

Misc/NEWS is Python's ChangeLog, available online here:

    http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/*checkout*/python/python/dist/src/Misc/NEWS?rev=HEAD&content-type=text/plain

Skip




From duncan at NOSPAMrcp.co.uk  Thu Apr 17 11:06:49 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 17 Apr 2003 15:06:49 +0000 (UTC)
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <d6jmqr2q.fsf@python.net> <Xns936072232AA84duncanrcpcouk@127.0.0.1> <he8xgr3g.fsf@python.net>
Message-ID: <Xns9360A249561CBduncanrcpcouk@127.0.0.1>

[posted and mailed]

Thomas Heller <theller at python.net> wrote in news:he8xgr3g.fsf at python.net:

> This looks great, but has problems. It is slow. For some input, at
> least.
> 
> I didn't have enough patience to wait for the output of your script,
> instead I changed the 10000 character strings to 100 characters and then
> it only took some seconds.

I'm confused.
On my machine, the test.py with 10000 character strings has a total running 
time of about 1/2 second. Are you saying it takes significantly longer on 
your machine?
...
a little later
...
Ok, now I see the problem. When I tried applying your patch to unittest.py 
in Python 2.2.2, it complained, so I applied it to Python 2.3. Some change 
between the two means that the comparison code is blindingly fast on 2.3 
and depressingly slow on 2.2.
 
> 
> Maybe you should create a new method (failUnlessEqual_ShowDiff or so)
> which will do the fancy formatting.
> 
> Maybe the diff formatting should be put into the TestReport, so that
> at least the
>     FFFF
> line is printed more quickly, and the formatting doen afterwards.
> 
> Maybe a time limit should be introduced.
> 
The first one might be good. A better move though is just to replace the 
2.2.2 difflib with the one from 2.3 (and enable generators and add a python 
definition of enumerate). This brings the running time down to something a 
bit more manageable (although I haven't checked that it hasn't broken 
difflib for other purposes).

The difference code could actually be made much faster quite easily. If I 
added a search for the first difference, then I could do the ellipsis and 
truncating first and only pass what is left to difflib.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From regnivon at netscape.net  Tue Apr 22 09:41:05 2003
From: regnivon at netscape.net (scn)
Date: 22 Apr 2003 06:41:05 -0700
Subject: coverting list back to string
References: <3cf72d52.0304211132.799d8afb@posting.google.com> <878yu33a16.fsf@vercingetorix.vercingetorix>
Message-ID: <3cf72d52.0304220541.144f537b@posting.google.com>

Thank you for your help.  This worked.  I incorporated your suggestion
of 'remove = remove[1:]' instead of my original code. I'd like to
figure out how to get each string to output line-by-line, instead of
one continuous string. If you have any suggestion, I'd appreciate it.
Thanks again!

Dialtone <dialtone at despammed.com> wrote in message news:<878yu33a16.fsf at vercingetorix.vercingetorix>...
> regnivon at netscape.net (scn) writes:
> 
> > my quest:  i know that output.write(remove) is not correct.  you
> > cannot use the 'write' method on a list.  is there a better way to
> > convert the list to a string after the first element of the list has
> > been removed and then output the strings to a file?  i'm a newbie
> > programmer and need guru help. thank you.
> 
> Sure:
> 
> input = file('c:/python22/programs/linetest.txt', 'r')
> output = file('c:/python22/programs/output.txt', 'w')
> 
> for line in input:
>     remove = line.split()
>     remove = remove[1:]
>     output.write(''.join(remove))
> 
> input.close()
> output.close()



From python at dylanreinhardt.com  Tue Apr 15 22:10:31 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 15 Apr 2003 19:10:31 -0700
Subject: find non-empty value in a dictionary
In-Reply-To: <b7ib8b$3rk$1@lust.ihug.co.nz>
References: <b7ib8b$3rk$1@lust.ihug.co.nz>
Message-ID: <1050459030.31581.395.camel@ida.dylanreinhardt.com>

What about:

if x:
   <stuff>
elif y:
   <stuff>

Etc, etc.  

Seems to work fine with a float value of 0.  There may be some edge case
where this doesn't work, but if there is, I haven't found it.

If you're just looking for the result (not which variable had the
non-zero value), you could do something like (untested):

result = max([d[var] / d['P'] for var in ['X','Y','Z']])

HTH,

Dylan



On Tue, 2003-04-15 at 18:26, matthew wrote:
> hi,
> 
> say you have a dictionay d =
> {'X': 0.0, 'Y'': 240.2, 'Z': 0.0, 'P': 13.3445}
> 
> where p divides pretty much evenly into Y; in this case approx 18
> 
> P will always be present and only one of X,Y,Z will have a non-zero value.
> 
> How can I easily find which X,Y, or Z is non-zero; esp. given that 
> comparing floats is 'inaccurate' as in:
> 
> if X != 0:
> 	<stuff>
> elif Y != 0:
> 	<stuff>
> elif Z != 0:
> 	<stuff>
> else:
> 	<stuff-up>
> 
> Also, once the correct axis is found,  exactly the same maths will be 
> applied which means that in the above, the same statement will be 
> replicated excepting the axis variable. any ideas on the best pythonic 
> way to tackle this?
> 
> any hints on best practice with comparing floats.
> 
> Thanks. matthew.
> 
> 





From ajsiegel at optonline.net  Fri Apr 11 17:15:32 2003
From: ajsiegel at optonline.net (Arthur)
Date: Fri, 11 Apr 2003 17:15:32 -0400
Subject: instance creatiion as an "event"
Message-ID: <3E973074.80806@optonline.com>

I am trying to re-engineer, as lightly as possible, a Python app I 
wrote, to be make it more functional when used at the interactive prompt.

The only thing I seem to be missing is some kind of "hook" to trigger a 
general update cycle of all existing class instances upon the creation 
of a new class instance at the interactive prompt

The ideas I come up with to explore do not seem very straight forward - 
though some variation of one or another of them will probably work. But 
am generally weak on introspection issues and am wondering, before 
embarking done a crooked road:

Am I missing a something straight forward.

Ideas appreciated.

Art







From salvatore at wanadoo.fr  Sat Apr 12 11:07:16 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Sat, 12 Apr 2003 17:07:16 +0200
Subject: distutils module
In-Reply-To: <sPVla.6373$T34.168242@news2.tin.it>
References: <b79863$dil$1@news-reader10.wanadoo.fr> <sPVla.6373$T34.168242@news2.tin.it>
Message-ID: <b79a5t$gue$1@news-reader12.wanadoo.fr>

Alex Martelli wrote:
> Salvatore wrote:
> 
> 
>>Hello,
>>
>>I've just installed Debian Knoppix,and i was surprised
>>to see that distutils module could not be found.
>>
>>Does anybody experienced the same problem ?
> 
> 
> Not the specific one you mention, but I've often seen various
> Linux distros package up Python in several and not necessarily
> well-conceived sub-packages.  Personally, I always install Python
> from sources -- I believe this reduces the amount of problems.
> 
> 
> Alex
> 
  Thanks Alex, I will do so





From tim.one at comcast.net  Thu Apr 17 21:45:08 2003
From: tim.one at comcast.net (Tim Peters)
Date: Thu, 17 Apr 2003 21:45:08 -0400
Subject: Python and Schools
In-Reply-To: <eppstein-B6AB90.12424517042003@news.service.uci.edu>
Message-ID: <LNBBLJKPBEHFEDALKOLCAEHEEDAB.tim.one@comcast.net>

[David Eppstein]
> Sorry, bug in untested code.  Instead of A[A[x]]==x the test should be 
> seq[A[x]]==x.

So the algorithm is

def anysame(seq):
    allocate array A, of dimension n**2, WITHOUT INITIALIZING IT
    for i, x in enumerate(seq):
        if A[x] is an integer and 0 <= A[x] < i and seq[A[x]] == x:
            return True
        A[x] = i
    return False

Cute!  I like it.

> Anyway, this is not code you would want to actually use, just an 
> illustration of how random access can sometimes make it useful to have 
> much more memory than you actually touch during an algorithm.

Yup, understood.  I still like it <wink>.




From bitbucket at safe-mail.net  Fri Apr 18 14:01:48 2003
From: bitbucket at safe-mail.net (poiboy)
Date: 18 Apr 2003 11:01:48 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com> <a3b5b608.0304080302.c26ad68@posting.google.com>
Message-ID: <af6b5bdd.0304181001.558da6f@posting.google.com>

> > Of course you can use template systems. But this is not always a good
> > economic decision.
>
> OK, but can't one use XML/XLST as a sole solution for the view?

I use XSLT [partially] for that reason; it is a general templating solution. 
The main reason I use XSLT however is because templating is the best (most 
relevant) excuse I have to keep another X specification in my repetoire.

The primary drawback to XSLT is having to fudge an XML document out of your
template variables before being able to transform them. If you're retrieving
data from a native XML source, great, otherwise you need to build an imaginary
tree and make sure it matches up with the paths in the template.  
 
Next, variable representation in the XSL template is likely to take up a whole
lot more room than a typical {{template_variable}}. Instead of standing out
nicely against a backdrop of HTML tags, the template variables in XSL have to
be searched for in small pools of xsl:for-each|value-of|copy-of|... tags and
select="/path/to/var" attributes (and if you've ever had trouble matching up
variable names in templates, you'll have a lot of fun with variable paths).

In the amount of time spent figuring out how to ensure unwanted text would
not appear without running my variable tree through a schema
(<xsl:template match="text()"/>), I could have implemented five other methods
to render my page (just guessing).

Neglecting the inherent joy of an education, I believe the cost of using XSLT
exceeds what benefits it may have (portability, and..) and is likely more
expensive than successive rewrites with other template systems.
 
Aloha, the poiboy



From brian094 at sympatico.ca  Tue Apr  8 10:01:15 2003
From: brian094 at sympatico.ca (Brian Alexander)
Date: Tue, 08 Apr 2003 10:01:15 -0400
Subject: curses: xterm vs. linux terminal
References: <3E9065FE.1090108@sympatico.ca> <b6qd2p$ib9$1@news1.radix.net> <3E91E40D.8050409@sympatico.ca> <b6udtr$das$1@news1.radix.net>
Message-ID: <3E92D62B.9070809@sympatico.ca>

Thomas;

Thanks for the link (and for writing it). It is very helpful.

Brian.




From bokr at oz.net  Sat Apr 26 12:07:34 2003
From: bokr at oz.net (Bengt Richter)
Date: 26 Apr 2003 16:07:34 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <vuiqa.3522$%_3.2810696@newssrv26.news.prodigy.com> <3EA9B7A4.59F9CE04@easystreet.com>
Message-ID: <b8eas6$c14$0@216.39.172.122>

On Fri, 25 Apr 2003 15:33:08 -0700, achrist at easystreet.com wrote:

>dsavitsk wrote:
>> 
>> Your requirements suggest to me that html is the best common denominator,
>> and it's easier to write.
>> 
>
>Yes, it is easier to write.  But I wrote a dead-plain-simple HTML file
>and opened it in MS Word (the version that came free on my computer in
>late 1996), and it looked just dreadful, with all kinds of visible
>markup that made no sense and generally might make you think that
>the  program that wrote it was written by someone who was not very
>smart.  So, I try to open the same file with Abiword (a relatively
>recent version), and it tells me that the file is bogus and refuses to
>open it at all.  This is simple HTML that opens fine in Netscape and in
>Amaya -- those have no trouble with my HTML (just HTML, BODY, UL, LI,
>TABLE, TR, TD, and B tags, IIRC).  So I figure that HTML is not a
>preferred path into word-processing programs when it is not known what
>version of which word-processing program the user has installed.  I
>intend to offer HTML output for those who want it, and I do have that
>nut cracked already, but it's not a complete solution.  
>
>So, again, wrt the larger nut, how to write RTF from Python? Any ideas?
>
You appear to be on NT, so there should be a RICHTX32.OCX somewhere
(which is a rich text text box) that can read and write rtf files and
also has methods and properties that allow you to mess with text.

Someone has probably interfaced to it via python and one of the win32
python wrappers, I would think.

Alternatively, if you just want a few features, use wordpad (start it by typing
write in a dos box, or use menu start>programs>accessories>wordpad) to make yourself
a template file and save it as .rtf. You can use %(placename)s in the text of the
template, so you can plug in substitution text using a dict, e.g.,
    template % {'placename':'replacement text', 'anotherplace':'etc'}

Here is an example python toy using an .rtf template prepared with
wordpad (aka write.exe) and writing the modified file and starting wordpad to show it:

====< rtfexample.py >========================
import sys, os
if len(sys.argv)<3:
    print 'Usage: rtfexample.py arg1 arg2'
    raise SystemExit
template = file('rtfexample.rtf').read().rstrip('\x00')
places = {'redtext':sys.argv[1], 'bluetext':sys.argv[2]}
file('rtfout.rtf','w').write(template % places)
os.system('start write rtfout.rtf')
=============================================
====< rtfexample.rtf >=======================
{\rtf1\ansi\ansicpg1252\deff0\deftab720{\fonttbl{\f0\fswiss MS Sans Serif;}{\f1\froman\fcharset2 Symbol;}{\f2\fmodern\fprq1 Courier New;}{\f3\froman Times New Roman;}}
{\colortbl\red0\green0\blue0;\red0\green0\blue255;\red255\green0\blue0;}
\deflang1033\horzdoc{\*\fchars }{\*\lchars }\pard\plain\f3\fs20 Plain text in default Times New Roman 10-point font.
\par \plain\f2\fs20 This line is courier new, and the following words
\par describe themselves:
\par \plain\f2\fs20\b bold
\par \plain\f2\fs20\i italic
\par \plain\f2\fs20\ul underlined
\par \plain\f2\fs20\b\i bold-and-italic\plain\f2\fs20 .
\par 
\par Above was a blank line.
\par 
\par Colored words on separate bullet lines:
\par \pard\li720\fi-720{\*\pn\pnlvlblt\pnf1\pnindent720{\pntxtb\'b7}}\plain\f2\fs20 {\pntext\f1\'b7\tab}Red  argv[1]: \plain\f2\fs20\cf2 %(redtext)s\plain\f3\fs20 
\par \plain\f2\fs20\cf0 {\pntext\f1\'b7\tab}Blue argv[2]: \plain\f2\fs20\cf1 %(bluetext)s\plain\f3\fs20 
\par \pard\plain\f2\fs20 This is the last line.
\par \pard\plain\f3\fs20 
\par }
=============================================

Without rtf features, the template text was
___________________________________________________

Plain text in default Times New Roman 10-point font.
This line is courier new, and the following words
describe themselves:
bold
italic
underlined
bold-and-italic.

Above was a blank line.

Colored words on separate bullet lines:
Red  argv[1]: %(redtext)s
Blue argv[2]: %(bluetext)s
This is the last line.
___________________________________________________

You don't have to have a monolithic template, obviously.
You could just snip pieces of boilerplate and commands
from an example containing features you want, and write
them out with write statements piece by piece or whatever.

HTH

Regards,
Bengt Richter



From juneaftn at REMOVETHIShanmail.net  Wed Apr 16 10:58:22 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Wed, 16 Apr 2003 23:58:22 +0900
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com> <3E9D6CFB.829F5217@engcorp.com>
Message-ID: <b7jr1h$r5k$1@news.hananet.net>

"Peter Hansen" <peter at engcorp.com> wrote in message
news:3E9D6CFB.829F5217 at engcorp.com...
> Christopher Blunck wrote:
[snip]
> As the man says, "Python is an agile language!" :-)
>
> (See http://www.agilealliance.org for more on that term.)
>
> -Peter

Ward Cunningham, one of the greatest programmers I know of, and Kevin Altis
came up with the idea of calling Python an agile language[1]. I believe who
fall between the intersection of python and XP must have felt that the two
match so well that they are designed for each other.

[1] http://altis.pycs.net/2003/04/04.html
    http://altis.pycs.net/2003/04/14.html#a59




From santacruz at southern.edu  Tue Apr 15 13:11:18 2003
From: santacruz at southern.edu (dansan)
Date: 15 Apr 2003 10:11:18 -0700
Subject: A python server
References: <15f445fb.0304140751.170d410e@posting.google.com> <3E9ADED2.280A671F@engcorp.com>
Message-ID: <15f445fb.0304150911.385c7202@posting.google.com>

> I believe the best advice for you at this stage might be the
> now-ubiquitous "Twisted does that".  Although at your level 
> figuring out how to work with Twisted (http://www.twistedmatrix.com)
> might be non-trivial

True that.  I've been trying to figure out how make twisted work. 
There are SO many things, that I have no clue what exactly to use.

> Five or ten lines of code on top of Twisted (and there are lots
> of good examples and How-Tos at that web site) is enough to get
> you past the socket issues, at which point the job is entirely in
> your hands to program the tournament scheduling parts, the
> "run games" part, etc.... in other words, it will let you focus
> on the application logic itself, and not on the infrastructure.

That is what I would like.  If twisted would allow me to concentrate
on the game-logic, and tournament logic, that would be great... but I
have no clue where to start.

There are 2 main different type of logics.  One is the tournament
logic (which keeps track of the games being played, the schedules, and
the results), and the other is the actuall game logic (rules,
turn-change, time-outs).  It would be nice if there was somewhat 1
instance of the tournament, and multiple ongoing games.

Anyway, if someone can give me a push-start, it'd be nice :)  Oh, and
I'm not scared to read doco, if the doco is actually going to help me
get the job accomplished (which so far hasn't).

Cheerz, Daniel.



From aleax at aleax.it  Tue Apr 29 09:45:03 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 13:45:03 GMT
Subject: Determining the volume ID for a file/drive?
References: <uv7savkms6lsgchcv6jh8ahtccu9lon3oi@4ax.com> <JZtra.19431$3M4.563286@news1.tin.it>
Message-ID: <zlvra.19900$3M4.571480@news1.tin.it>

Alex Martelli wrote:
   ...
> Quick and dirty...:
> 
> import os
> x = os.popen('C:\\')

OOPS -- TOO quick -- sorry, I meant os.popen('DIR C:\\')!!!


Alex




From jeremy at jdyallop.freeserve.co.uk  Fri Apr 11 05:36:58 2003
From: jeremy at jdyallop.freeserve.co.uk (Jeremy Yallop)
Date: 11 Apr 2003 09:36:58 GMT
Subject: collecting data from file
References: <3E9671A3.2050004@sympatico.ca>
Message-ID: <slrnb9d2sh.1oq.jeremy@saturn.cps.co.uk>

Mustafa Celik wrote:
> I want to scroll thru a file;
>    * find lines that match a string (e.g. HELLO) on the 2nd column
>    * add up the 4th column (an integer) on each matching line, say the 
> variable is TOTAL
>    * subtract the 4th column from TOTAL if another is string (e.g. BYE) 
> is hit on 2nd column of a line

I'd use awk for this.

  { 
    if ($2 == "HELLO")    { total += $4 }
    else if ($2 == "BYE") { total -= $4 }
  }
  END { print total }

Jeremy.



From john.burton at jbmail.com  Tue Apr 15 03:53:19 2003
From: john.burton at jbmail.com (John Burton)
Date: 15 Apr 2003 00:53:19 -0700
Subject: DB API for ADO/DAO?
Message-ID: <2402252b.0304142353.2b4711ae@posting.google.com>

I want to be use an .MDB database from a python application on
windows.

The problem with odbc is that you need to manually set up data sources
using the control panel which isn't something I want the users to have
to do.

So I'm using DAO via com to access the database and it's working fine
except that I have to use the DAO com api instead of the standard
python api. So I was wondering if there is a suitable DB API wrapper
for DAO or ADO I can use? Or a better solution to my problem?



From tim_one at email.msn.com  Tue Apr  1 23:24:35 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Tue, 1 Apr 2003 23:24:35 -0500
Subject: A bug I found in Python
In-Reply-To: <ccc7084.0303301315.6864ed5a@posting.google.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCIEEAEEAB.tim_one@email.msn.com>

[Tj]
> I feel a bit chagrined because I'm the last guy who wrote, "OH MY GOD
> THERE'S A BUG!!!"
>
> Hopefully I made sure it sounded a little more self-doubtful than
> that.

Not to worry:  every message on Usenet is in fact about you!  People have
been known to waste valuable years fighting this self-evident truth.  Better
to accept it early, and get on with enjoying the adulation.

me-me-glorious-me-ly y'rs  - tim





From Scott.Daniels at Acm.Org  Tue Apr  8 20:28:54 2003
From: Scott.Daniels at Acm.Org (Scott David Daniels)
Date: Tue, 08 Apr 2003 17:28:54 -0700
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <mailman.1049839153.2128.python-list@python.org>
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org> <BAB8B22D.2022%dogen@mindspring.com> <mailman.1049839153.2128.python-list@python.org>
Message-ID: <3e936947$1@nntp0.pdx.net>


Dave Brueck wrote:
> ...
> def IsPal(s):
>   r = list(s)
>   r.reverse()
>   r = ''.join(r)
>   ...

This bit should go significantly faster with array:

     from array import array

     def IsPal(s):
        t = array('c', s)
        t.reverse()
        r = t.tostring()

-Scott David Daniels
Scott.Daniels at Acm.Org





From imbosol-1050418529 at aerojockey.com  Tue Apr 15 12:01:59 2003
From: imbosol-1050418529 at aerojockey.com (Carl Banks)
Date: Tue, 15 Apr 2003 16:01:59 GMT
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050394215.20671.python-list@python.org>
Message-ID: <X1Wma.17517$Gb1.9689@nwrdny01.gnilink.net>

Christian Tanzer wrote:
> 
> Carl Banks <imbosol-1050339953 at aerojockey.com> wrote:
> 
>> Christian Tanzer wrote:
>> > All these functions (except the one you're writing) might have been
>> > implemented years before the text properties and
>> > `inhibit-point-motion-hooks` were introduced.
>>
>> See, that's really the problem.  It's obvious why history made Emacs
>> what it is, but it doesn't change the fact that, as it stands now,
>> Emacs is a terrible software design.
> 
> This kind of hubris is normally associated with Perl programmers, I
> thought. Calling Emacs a terrible software design is similar to saying
> that tea clippers had a terrible design because they didn't use an
> engine...
> 
> Emacs is certainly a very successful application. Complex as it is, it
> is very easy to program even for somebody like me who hates lots of
> silly parentheses.

That it's a successful application is irrelevant to whether it's a
good design.  (Exhibit A: Any Microsoft program.)

That it's easy for you to program is something that need not disappear
with a good design.  However, 




>> Now, Emacs has a variable inhibit-point-motion-hooks.  If Emacs were
>> currently designed well, this variable would not be needed, because
>> functions would be able to operation without touching the point.
> 
> Not true. This has nothing to do with touching point but how cursor
> movements behave. For example, searching doesn't work properly in some
> cases unless you let `inhibit-point-motion-hooks` to a true value.

The cursor *is* the point, chief.

Things like searching move the point around, because Emacs is a flawed
design and only works in the context of the point.  That's why you
need inhibit-point-motion-hooks: in a search, the point moves all over
the place unbeknownst to the user, and it would really confuse things
and cause problems if hooks the user installed whenever the point
(i.e., the cursor) moves get called during a search operation.

A good design wouldn't abuse the point like this.  The point would be
the place where user input happens, nothing more.  Functions that
wanted to do something in the buffer away from point wouldn't *have*
to move it.  And we wouldn't need a hack like
inhibit-point-motion-hooks.


>> (As for things like buffer-local variables, I don't think it's a big
>> deal to make them attributes of a buffer class.  Frame local
>> variables, frame objects.  I think it's quite short-sighted to try to
>> force buffer-local variables to be regular variables, and not what
>> they obviously should be, attributes of the buffer, just because Emacs
>> does it that way.)
> 
> Only partly true. Sure, a object-oriented design would make for a
> better architecture and a better API. A pythonic solution certainly
> would have buffer/frame/... objects with their respective attributes.
> But there isn't that much difference between an attribute of a buffer
> object (in the Python sense) and a buffer-locale variable (in the
> Emacs sense).

I don't even understand the point you're trying to make.

In any case, the two are very different: buffer-local variables are
variables, and so they need a namespace, and an odd one.  Buffer
attributes have none of the complicated namespace issues, since
they're attributes.


> You might be able to save some macros, but you would probably still
> have lots of context which needs temporary changes for some functions
> to work.

I think you're very wrong.


-- 
CARL BANKS



From rmunn at pobox.com  Thu Apr  3 05:10:59 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 03 Apr 2003 10:10:59 GMT
Subject: Python commands
References: <3e8bb9c5$1_1@news.iprimus.com.au>
Message-ID: <slrnb8nd5p.eaq.rmunn@localhost.localdomain>

Paul Dexter <dextech at terrigal.net.au> wrote:
>     Hi all,
>                 I have just discovered Python and I was wondering if anyone
> could help me find an equivalent command in Python for the "OUT" and "INP"
> commands in Qbasic. I wish to convert a Qbasic program to Python.
>     The following snippet is an example of the task I wish to perform;
> 
> base0 = &H378:
> 
> interface:
> OUT base0, &HA0
> OUT base0, &H80
> OUT base0 + 2, &H8
> 
> V = (INP(base0 + 1) AND &H10) * &H8

Ewww, direct printer port interfacing. Nasty.

>     The base address is 378 (lpt) and the "interface" subroutine outputs
> data to the port (378) and "V" is the input data statement I have used. I
> also use the "GETKEY", "LOCATE" and "DRAW" commands in Qbasic
> 
>     Does anyone know how to perform the same tasks/commands in Python or are
> they even available. I have come across other Basic computer language
> variations that do not support the commands I require.

Personally, I'd want to stay far away from direct printer interfacing;
let the drivers do the work for you. But if you really have to, then the
winioport module might be what you need:

    http://www.geocities.com/dinceraydin/python/indexeng.html

I found it by doing a Google search for "python parallel port".

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From fperez528 at yahoo.com  Wed Apr  2 19:51:06 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 02 Apr 2003 17:51:06 -0700
Subject: A few errors in "Python in a Nutshell" - Numeric related
References: <b6frd5$uq$1@peabody.colorado.edu> <pan.2003.04.03.00.14.30.217865@mn.rr.com>
Message-ID: <b6g0hr$332$1@peabody.colorado.edu>

Anna wrote:

> Alex Martelli is at Python-UK.
> 
> You might wish to submit this to the online errata form. Here's the url
> for doing so.
> 
> http://www.oreilly.com/cgi-bin/errata.form/pythonian

Done.

Thanks for the info.

Cheers,

f.



From fidtz at clara.co.uk  Wed Apr 16 12:25:29 2003
From: fidtz at clara.co.uk (Fidtz)
Date: 16 Apr 2003 09:25:29 -0700
Subject: Inserting Python commands into HTML documents
References: <zpqdne0Kztcb0AejXTWcow@comcast.com>
Message-ID: <24944f78.0304160825.336057a5@posting.google.com>

Em <mcclaine at comcast.net> wrote in message news:<zpqdne0Kztcb0AejXTWcow at comcast.com>...
> Hello,
> 
> I am an extremely new member of the Python family. I had a job interview 
>   where they charged me with the task of creating a (fake) corporate 
> website using Python in ten days--with two left now.
> 
> The problem is that what I have read seems to say Python is a CGI 
> language... Am I remiss in believing that my website is going to be 
> completely run from inside a couple .PY files in the CGI-BIN?
> 
> Is there a way for it to act more like PHP, as in run Python commands 
> from inside the (index|contacts|about|etc).html or at least simulate 
> that behavior? (This is on a linux server and I am not to use PHP, FYI.)
> 
> 
> Thank you all for any assistance!

If you only have 2 days and are lacking install options (from your
other posts), I would suggest cubictemp , a tiny templating system
from nullcube. It consists of one .py file that can live in your
PYTHONPATH.

http://www.nullcube.com/software/cubictemp.html

It is very simple so you will be able to understand it in a few
hours,if that, and get on with the web design bits. You'll have to
design within it's limitations (don't go hog-wild with conditional
formatting basically). In two days you probably have to get the DB
connections sorted , the graphics done etc. etc. A simple but clean
looking design would be the best bet under the time pressure anyway
IMHO. I use it all the time as the simplicity suits me :)

Dom



From ckea25d02 at sneakemail.com  Tue Apr 15 23:21:06 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 15:21:06 +1200
Subject: [ANN] Pyrex 0.7.1
Message-ID: <mailman.1050463083.12917.clpa-moderators@python.org>

Pyrex 0.7.1 is now available:

    http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/

This release fixes several bugs. See the web page and CHANGES file for
more information.

There is also one new demo included, showing how to pass a
Python function to a C function as a callback.

What is Pyrex?
--------------

Pyrex is a new language for writing Python extension modules.
It lets you freely mix operations on Python and C data, with
all Python reference counting and error checking handled
automatically.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg





From cjw at sympatico.ca  Sun Apr 27 15:38:24 2003
From: cjw at sympatico.ca (Colin J. Williams)
Date: Sun, 27 Apr 2003 15:38:24 -0400
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <3EAC2962.2050609@soraia.com>
Message-ID: <3EAC31B0.43C50CDC@sympatico.ca>

Joe Francia wrote:

> Steven Hodgen wrote:
>
> > Hello,
> >
> > I'm currently teaching a high-school introduction to programming
> > course.  I'm currently teaching in Java, but am seriously considering
> > changing to Python.  The problem is that I cannot seem to find any good
> > textbooks, much less ones for students with no prior programming
> > experience.  This is the one area in which Java excels.  I'm using an
> > awesome programming environment called "BlueJ", and there are lots of
> > textbooks for students at all levels.
> >
> > However, Java is not a great language for beginners.  Python, on the
> > other hand, is ideal.  If I can find a textbook that is even ok, I may
> > consider changing, since the language is so much better for students at
> > this level.  Doesn anyone know of a book like this?  Ideally, it would
> > have examples and excercises, and perhaps even a web-site with other
> > instructor and student resources.
> >
> > Thanks!
> >
> > --Steven
> > steven at twitch.net
>
> The Deitels have a textbook called _Python How to Program_ that is aimed
> at beginning programmers.  I'm not sure how good/bad it is.  Three out
> of four reviewers on Amazon seem to like it (the fourth guy gave it one
> star based on its lack of Perl coverage).

The book is very expensive and heavy reading.  The Nutshell is excellent,
but would need more examples for your purpose.

Colin W.






From tim.one at comcast.net  Fri Apr 11 16:17:01 2003
From: tim.one at comcast.net (Tim Peters)
Date: Fri, 11 Apr 2003 16:17:01 -0400
Subject: Loss of precision / data corruption with Tkinter.DoubleVar ?!
In-Reply-To: <f70e3538.0304111007.b0a72c3@posting.google.com>
Message-ID: <BIEJKCLHCIOIHAGOKOLHCECJFFAA.tim.one@comcast.net>

[Lonnie Princehouse]
> I stumbled across this curiosity yesterday:
>
> >>> x = 0.33333333333333098
> >>> x
> 0.33333333333333098
> >>> import Tkinter
> >>> y = Tkinter.DoubleVar(Tkinter.Tk())
> >>> y.set(x)
> >>> y.get()
> 0.33333333333300003

Please open a bug report on this.  The result is the same as you'd get via

    float(str(x))

and Python's str() doesn't produce enough significant digits to reproduce
the input double.  A bit of digging revealed that Python's _tkinter.c is
using PyObject_Str() to produce a string to pass on to Tk, and that's the C
spelling of Python's str().  That explains the symptom, but I don't know how
to fix it.  If you open a bug report, someone who does know how to fix it
can be found.





From oblivious at web.de  Tue Apr  8 04:47:46 2003
From: oblivious at web.de (Dominic)
Date: Tue, 08 Apr 2003 10:47:46 +0200
Subject: Comms Protocol design dilemma
In-Reply-To: <mailman.1049738831.30630.python-list@python.org>
References: <mailman.1049738831.30630.python-list@python.org>
Message-ID: <3E928CB2.2020604@web.de>

> number generation involved, I can pretty much rule out asynchronous
> sockets. Is there any better strategy to tackle this problem?

Well asynchronous sockets are fine. I've written a load-test
client in Python which handles 1000 tcp connections simultaneously.
(1000 sockets is the per process limit for non root users)
I've used the python asyncore module. If you use asynchronous
sockets you need to keep your session state explicitly since
you do not have threads. You could use a finite-state-machine
or Python's generators. (See my previous mail on comp.lang.python,
Programming Erlang style in Python. You should certainly
check out the Erlang programming language homepage (for the concepts)
from Erricson since that language has been designed for servers which 
you intend to write.)

Ciao,
  Dominic




From a.schmolck at gmx.net  Wed Apr  9 13:09:57 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 09 Apr 2003 18:09:57 +0100
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com>
Message-ID: <yfs7ka3iom2.fsf@black132.ex.ac.uk>

s.thuriez at laposte.net (sebastien) writes:

> Hi,
> 
> Is there any PCA analysis tools for python ?

What is the analysis tool supposed to do?

Maybe this will do what you want, once you (downloaded and installed Numeric):

# Warning: hackish and not properly tested ripped out bit of code ahead
# so no guarantees whatsoever
# Anyway, it should at lesat sort of give you the idea
# try pca(X); if that doesn't do what you want try pca(t(X))

from Numeric import take, dot, shape, argsort, where, sqrt, transpose as t
from LinearAlgebra import eigenvectors

def pca(M):
    "Perform PCA on M, return eigenvectors and eigenvalues, sorted."
    T, N = shape(M)
    # if there are less rows T than columns N, use
    # snapshot method
    if T < N:
        C = dot(M, t(M))
        evals, evecsC = eigenvectors(C)
        # HACK: make sure evals are all positive
        evals = where(evals < 0, 0, evals)
        evecs = 1./sqrt(evals) * dot(t(M), t(evecsC))
    else:
        # calculate covariance matrix
        K = 1./T * dot(t(M), M)
        evals, evecs = eigenvectors(K)
    # sort the eigenvalues and eigenvectors, decending order
    order = (argsort(evals)[::-1])
    evecs = take(evecs, order, 1)
    evals = take(evals, order)
    return evals, t(evecs)


You can download Numeric and use it to compute the eigenvalues and
eigenvectors of an array.


> If it does, do you have any idea on how well it would scale ?

It should scale fine. If you experience speed problems, configure Numeric 23
with ATLAS support (you have to install ATLAS and LAPACK first, of course).
For large matrices, this should be *much* faster than handwritten C code that
doesn't use ATLAS.

> 
> I have already seen PyClimate (but it is not available for Windows
> which will be one of the target). Is there some LAPACK like packages ?

Yes, Numeric and scipy. (www.numpy.org, www.scipy.org, I should think)

'as



From hancock at anansispaceworks.com  Thu Apr 17 13:34:32 2003
From: hancock at anansispaceworks.com (Terry Hancock)
Date: Thu, 17 Apr 2003 10:34:32 -0700
Subject: Zope Sequences and Python Scripting
References: <pan.2003.04.17.12.34.15.777976@fraley.de>
Message-ID: <GN6dnaRxi5queAOjXTWcpw@august.net>

Patrick W. Fraley wrote:
> I have a python script which returns a [] containing {}.  Now I can go
> thru the returned array (I know this is perl lingo, sorry that is my
> background) by using the <dtml-in> method, but I can not
> access the values of the dictionary contained within the sequence item.
> 
> And here is what I am trying to do with the returned array:
> 
> <dtml-in filter_row>
> <dtml-var filter_id><br>
> <dtml-var bez_ref><br>
> <dtml-var col_span><br>
> <hr>
> </dtml-in>

<dtml-in filter_row prefix="sq">
   <dtml-var expr="sq_item['filter_id']"><br>
   <dtml-var expr="sq_item['bez_ref']"><br>
   <dtml-var expr="sq_item['col_span']"><br>
   <hr>
</dtml-in>

or

<dtml-in filter_row>
  <dtml-with seq-item>
       <dtml-var filter_id><br>
       <dtml-var bez_ref><br>
       <dtml-var col_span><br>
       <hr>
  </dtml-with>
</dtml-in>

(sorry, they're untested -- you may have to fix)

The point is, you will loop through the *sequence* of *dictionaries*. Then 
you act on each dictionary -- referencing its elements.  You can use python 
syntax to retrieve dictionary elements from the seq-item (prefix is a DTML 
kludge to make the syntax more python compatible), or "dtml-with" to make 
the dictionary the current namespace.  There are probably other ways to do 
it as well.

Cheers,
Terry

-- 
Terry Hancock
Anansi Spaceworks http://www.AnansiSpaceworks.com/




From jordanm at uts.cc.utexas.edu  Tue Apr 15 15:56:29 2003
From: jordanm at uts.cc.utexas.edu (Jordan McCoy)
Date: Tue, 15 Apr 2003 14:56:29 -0500
Subject: Python and Schools
In-Reply-To: <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
Message-ID: <000b01c30389$1c7cce10$2301a8c0@telos>

The notation O(n) and O(n^2) refers to order notation, a means of
representing the execution time of programs, and in particular,
algorithms that operate over a set of data. The more complete definition
is thus:

The order of an algorithm is O(f n), if a function f and a constant k
exists such that k * f(n) is not more then the execution time of the
algorithm, assuming n is a suitable problem set.

The goal of algorithm development is to achieve O(f n) [or O(n)
abbreviated] as this represents constant execution time; each addition
to the problem set adds a constant amount of time to the execution time.
So, an algorithm that takes, for example, 1 second for n=1 would take
100 seconds for n=100, and so forth.

However, most algorithms can't achieve this: larger problem sets
increase algorithmic complexity and therefore increase execution time at
a greater then constant rate. Two examples:

1. The common but inefficient bubble sort has an order of O(n^2). Since
the bubble sort ends up comparing each set element multiple times to
multiple elements, a sort where n=1000 will require nearly 500,000
comparsions. While the math is somewhat more complex and boring, this
means that each addition to the problem set will result in an
exponential time increase.

2. The more efficient quick sort has an order of O(n * log2 n). Worse
then constant time, but much better then exponential time. Generic
sorting algorithms (those that don't rely on knowing pattern information
inherent in the data) rarely improve upon logarithmic time.

You can find more by searching on 'Order Notation.'

Jordan McCoy

> -----Original Message-----
> From: python-list-admin at python.org 
> [mailto:python-list-admin at python.org] On Behalf Of Greg Brondo
> Sent: Tuesday, April 15, 2003 2:07 PM
> To: python-list at python.org
> Subject: Re: Python and Schools
> 
> 
> Aahz wrote:
> > In article <v9e21mhemm8c2 at corp.supernews.com>,
> > Paul Watson <pwatson at redlinec.com> wrote:
> > 
> >>We need to teach students correct design priciples to build 
> something
> >>greater than what already exists.  We will never get very far if we
> >>require everyone to start with a quark or atom.  Yes, of 
> course we need
> >>some people who design silicon and create microcode.  They 
> will learn
> >>the low-level details what they need to know as they need 
> it.  Knowing
> >>great design and organization principles will enable them 
> to make the
> >>most of it.
> > 
> > 
> > While there's some truth to that, try explaining why the 
> following code
> > is a Bad Idea to someone who has no basic understanding of CS:
> > 
> >     s = ''
> >     for i in range(1000000):
> >         s += str(i)
> > 
> > Knowing the difference between O(N) and O(N^2) is critical 
> to writing
> > even the simplest programs.  OTOH, I do agree with you that 
> focusing on
> > programming as a craft is more important to being a programmer than
> > learning CS.
> 
> I understand why the code is bad (copying the string each time in
> memory as it grows).  However, I've never understood the notation's
> O(N) and O(N2)?
> 
> Thanks,
> Greg Brondo
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 





From alloydflanagan at attbi.com  Fri Apr 25 16:40:13 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 25 Apr 2003 13:40:13 -0700
Subject: Exception and finally question
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com> <3EA86EDA.EFD023C0@alcyone.com>
Message-ID: <a8b7f07a.0304251240.7e0ed334@posting.google.com>

Erik Max Francis <max at alcyone.com> wrote in message news:<3EA86EDA.EFD023C0 at alcyone.com>...
> 
> By design.  Originally, as I understand it, Python supported a unified
> try: ... except: ... finally: ... construct, but evidently it caused a
> lot of confusion with newbies (personally, I don't see where the
> confusion would be, but I guess that's because I'm familiar with the
> construct from Java).  So in modern Python the try can either have an
> except (perhaps several) or a finally, but not both.
> 

Really?  I could swear I saw somewhere that it was an implementation
issue, and likely to go away in the future.  Anyway, I agree with you,
I don't think it would be a problem.



From a-steinhoff at web.de  Mon Apr  7 06:50:29 2003
From: a-steinhoff at web.de (Armin Steinhoff)
Date: 7 Apr 2003 03:50:29 -0700
Subject: Calling C libraries from a Python script
References: <e81e0f3d.0304050639.37b59655@posting.google.com> <r88gc37v.fsf@python.net>
Message-ID: <ddc19db7.0304070250.7ecaf65e@posting.google.com>

Thomas Heller <theller at python.net> wrote in message news:<r88gc37v.fsf at python.net>...
> joshua at goodish.org (Joshua Goodlett) writes:
> 
> > I'm trying to find examples of calling C library functions from within
> > a Python script.  Any advice or pointers on where to look for examples
> > would be appreciated.
> 
> If the C library is a dll, you could use ctypes.

Such DLLs must also be conform to the  _cdecl call convention ... isn't it ?

Armin 

> 
> Thomas



From danb_83 at yahoo.com  Thu Apr 17 14:17:24 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 17 Apr 2003 11:17:24 -0700
Subject: variable in Python
References: <b7lvbh$an3$1@news-reader13.wanadoo.fr>
Message-ID: <ad052e5c.0304171017.45b8884f@posting.google.com>

Salvatore <artyprog at wanadoo.fr> wrote in message news:<b7lvbh$an3$1 at news-reader13.wanadoo.fr>...
> Hello,
> 
> I have a little doubt.
> when you declare x = 3

<nitpick>There are no variable declarations in Python.</nitpick>

> does the variable x 'contains' the value 3 (like in Pascal)
> or does it contains an address which points to
> a memory place containing the value 3 ?

The latter: All variables in Python are pointers in disguise.

This is why, after executing

list1 = [0, 1, 2, 3]
list2 = list1
list2[3] = 0

list1 will be [0, 1, 2, 0] instead of [0, 1, 2, 3].  The assignment
"list2 = list1" copies a pointer to the list, not the contents of the
list.

Note that I used a list instead of an int.  With ints, the distinction
between value and reference semantics doesn't matter as much because
ints are immutable.



From pan-newsreader at thomas-guettler.de  Mon Apr  7 16:16:27 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Mon, 07 Apr 2003 22:16:27 +0200
Subject: COM and MS Word
References: <2eb85cfb.0304070115.6aecb644@posting.google.com> <pan.2003.04.07.11.53.57.522248.788@thomas-guettler.de> <2eb85cfb.0304070613.211f24ac@posting.google.com>
Message-ID: <pan.2003.04.07.22.16.26.743151.788@thomas-guettler.de>

On Mon, 07 Apr 2003 16:13:55 +0200, CTony wrote:

>> Our save the word document as XML and parse this.
>
> Thanks, however (you could hear it coming, couldn't you?) I'm trying to
> avoid making explicit changes to the document in Word, and unless wvWare
> is run from Python...?
> 
> I've just found out - there are literally a few hundred of these files
> (ulp!), which makes any real manual work a bit of a pain! *g*

Maybe abiword or openoffice can import theses files. I think they are
scriptable.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From pinard at iro.umontreal.ca  Wed Apr  2 20:39:55 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 02 Apr 2003 20:39:55 -0500
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to Python. Good idea? Bad Idea? PEP?
In-Reply-To: <b6fv3c$5218t$1@ID-169208.news.dfncis.de>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
	<b68jti$2mj66$1@ID-169208.news.dfncis.de>
	<RbCia.6782$9I.555309@news20.bellglobal.com>
	<b6fv3c$5218t$1@ID-169208.news.dfncis.de>
Message-ID: <oqk7ec740k.fsf@alcyon.progiciels-bpi.ca>

[Greg Ewing (using news.cis.dfn.de)]

> WP wrote:
> > Greg Ewing wrote the following Pyrex snippet:
> > >     for i from 0 <= i < n:
> > >       result.append(a[i])

> Rather, the motivation was readability and expressiveness.  The things
> I like about it are that it gives very clear and unambiguous control
> over the inclusion or exclusion of both endpoints and the direction
> of iteration.

Which admittedly, after having experienced it, is useful.  

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From claird at lairds.com  Mon Apr 28 15:07:59 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 28 Apr 2003 19:07:59 -0000
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <yu99sms2k35f.fsf@europa.research.att.com> <mailman.1051548461.3290.python-list@python.org>
Message-ID: <vaqv0fg30kdc85@corp.supernews.com>

In article <mailman.1051548461.3290.python-list at python.org>,
andrew cooke <andrew at acooke.org> wrote:
			.
			.
			.
>[1,4)+[4,7) = [1,7)
>
>isn't there some (fairly old) popular math book by a famousish person
>(knuth?  hofstadter?) that develops maths (at least arithmetic) based on
>defining sets of points from the "number line"?  i wonder what that says
>about this?  (i should admit that i picked up the above from one of
>celko's db books).
			.
			.
			.
Typical behavior of (modern--post-Frege, let's say) mathematicians
is to create "models".  We know how arithmetic's s'posed to work,
but mathematicians delight in demonstrating that, instead of leaving
"number" as an undefinable imponderable, it can be identified with a
certain set construction.  This sort of thing has been done many
times, in many books.  There are correct and incorrect parts in your
memory above.  "[S]ets of points from the 'number line'" will be the
hardest to unmuddle.

So:  there are several relatively popular mathematically-oriented
books that illustrate, for example, Peano arithmetic.  
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From gh at ghaering.de  Thu Apr 24 19:21:22 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 01:21:22 +0200
Subject: string to variable name
In-Reply-To: <b89qi8$2oij$1@agate.berkeley.edu>
References: <b89qi8$2oij$1@agate.berkeley.edu>
Message-ID: <3EA87172.7050403@ghaering.de>

Benjamin Hartshorne wrote:
> <disclaimer -- perl programmer trying to learn python>
> 
> I'm trying to figure out how to get from a string to a varible name.  I just
> picked up Python last week, so I'm still kinda new at this. 
> 
> Suppose I've got a text file:
> 
> """
> blah blerg blarf
> foo bar baz
> spam eggs ham
> """
> 
> I have a class:
> 
> class myobject:
>   def __init__(self):
>     self.blerg = "blarf"
>     self.bar = "baz"
>     self.eggs = "ham"
> 
> I have a global:
> 
> inst = myobject()
> 
> I have a function that I want to change the variables in my class:
> 
> def myassigner():
>   file = open("mytextfile", 'r')
>   for line in file.readline():
>     m = re.match("^(\w+)\s+(\w+)\s+(\w+)$", line)
>     (newval, var, oldval) = m.groups()
>     inst.var = newval
> 
> The equivelent (of what I _want_ to do) in perl: [...]
>     inst.$2 = $1;
>  [...]

Use setattr:

setattr(inst, var, value)

You can easily guess what getattr and hasattr will do.

> [...]
> p.s.  The inspiration behind asking this question: I want to store all the
> values passed in by a CGI script in an instance of a class.
> 
> for key in mycgi.keys():
>   mystorage.key = mycgi[key].value

I'd suggest you only overwrite attributes you have in a particular list, 
otherwise somebody could enerve you by submitting a variable called 
"__init__" or something like that ;-)

So, something like (untested):

class MyClass:
     __cgi_vars = ["foo", "bar", "baz"]
     def __init__(self):
	...


myobject = MyClass()

for key, value in mycgi.items():
     if key in MyClass.__cgi_vars:
         setattr(myobject, key, values)

> Hey! Could I do it by overloading the special function that's called when you
> treat something like a dictionary?  i.e. could I make my class respond to 
>   key = "bar"
>   inst[key] = "baz"
> as though it got
>   inst.bar = "baz"
> 
> How?

Sure, overload the __setattr__ and __getattr__ and for extra fun, the 
__contains__ special methods. See 
http://www.python.org/doc/current/ref/attribute-access.html

__contains__ will affect the "in" operator and is described in
http://www.python.org/doc/current/ref/sequence-methods.html

These and the pages nearby show you how dynamic you can make Python, 
including making custom callables, new numeric types, ...

If you're using Python 2.2, you might also be interested in using 
properties.

-- Gerhard




From peter at engcorp.com  Fri Apr 11 10:31:32 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 11 Apr 2003 10:31:32 -0400
Subject: collecting data from file
References: <3E9671A3.2050004@sympatico.ca> <3E96B6BC.E9DBE74F@engcorp.com> <MM2dneyWRoYWXwujXTWcow@comcast.com>
Message-ID: <3E96D1C4.D60EE85F@engcorp.com>

Terry Reedy wrote:
> 
> >  Also, are there any error conditions to be handled, or do
> > you guarantee the input is always perfect (e.g., always at least four
> > columns, and so on)?
> 
> Missing data is the great bugaboo of data analysis.  Ignoring that
> possibility
> 
> sum = 0
> for line in file('mydata'):
>   fields = line.split() # assume whitespace
>   if field[1] == 'HELLO':
>     sum += field[3]
>   elif field[1] == 'BYE'
>     sum += field[3]

I believe the last line was intended to read

      sum -= field[3]

Terry, I like how you solution *exactly* matches the requirements
as specified, by not even printing the resulting sum.  Very 
efficient of you.  :-)

-Peter



From aleax at aleax.it  Sun Apr 13 04:38:14 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 08:38:14 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com>
Message-ID: <Wl9ma.6864$LB6.194823@news1.tin.it>

Roy Smith wrote:

> Alex Martelli <aleax at aleax.it> wrote:
>> Yes -- it's THE one major performance trap in naively coded Python,
>> since, when "a lot" is a sufficiently large N, a lot of += on short
>> strings ends up being O(N squared) where O(N) approaches would be
>> available.
> 
> One of the very few things I like about C++ over Python is that the STL
> is very careful to document the complexity of the various operations.

Yes, that's pretty unique -- C, Fortran, and all other languages I used,
never came with such useful specs for libraries and primitives.

> AFAICT, no such documentation exists for Python.

Well, there's the chapter on (testing, debugging, and) optimization in the
Nutshell -- Tim Peters kindly double-checked it, so any errors remaning
are HIS fault;-).  But of course there's a key difference -- the Nutshell's
info is *descriptive*, and that's quite another thing from the info in
the C++ Standard, which is *prescriptive* of how standard library
implementations MUST behave.

> As another example, I can guess that Python dictionaries work like STL
> maps on the inside, but that would be just a guess.  Is inserting an
> element logN?  Is keys() linear?  Is del on an element logN?  Is
> has_key() logN?  Those would be my guesses.  But they're just guesses.

If hash(x) is taken to be constant-time (actually it depends on x, of
course), then inserting, deleting and has_key are all roughly O(1) too
(amortized -- of course SOME insertions will cause the dict to grow,
but the idea is that they're rare enough not to matter in an amortized
sense).  keys is indeed O(N).

The crucial difference is that Python dicts are hashtables, while the
std::map in the standard library of C++ can't be, because it's based
on an order relationship, NOT on hashing.  So, std::map has an order,
but its O(...) performance just can't be as good as a dict's.


Alex




From dave at pythonapocrypha.com  Sat Apr 12 15:31:30 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Sat, 12 Apr 2003 12:31:30 -0700 (PDT)
Subject: Speed of string += string
In-Reply-To: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
Message-ID: <Pine.LNX.4.51.0304121229540.8383@ns0.tbye.com>

On Sat, 12 Apr 2003, Mads Orbesen Troest wrote:

> Given that strings are immutable, is it exceedingly slow (and memory
> spiking) to do a lot of "string += string" operations in one's code?
[snip]
> that the longer the string to which stuff is appended is, the longer times
> each += operation takes?

When you tried it out yourself and timed the results, what did the results
indicate?

-Dave




From pyth at devel.trillke.net  Sat Apr 19 15:03:06 2003
From: pyth at devel.trillke.net (holger krekel)
Date: Sat, 19 Apr 2003 21:03:06 +0200
Subject: Testing methods
In-Reply-To: <9d8ac130.0304190938.755d4b4c@posting.google.com>; from poec@yahoo.com on Sat, Apr 19, 2003 at 10:38:40AM -0700
References: <9d8ac130.0304190938.755d4b4c@posting.google.com>
Message-ID: <20030419210306.B22359@prim.han.de>

Ovid wrote:
> Hello,
> 
> Relatively new to Python and I thought I would teach myself the
> language by building a test harness.  After I'm done, I will probably
> be throwing it away in favor of Pyunit, but it seemed like a good way
> to learn some of the fiddly bits.
> 
> Fiddly bits such as "can this object do that?"  In my test code, I
> want to test whether or not a given class *or* instance of that class
> can execute a particular method (i.e., if it has or inherits the
> method).  For example, one way (in Perl), to test if object "$foo" can
> execute an arbitrary method would be:
> 
>   if ( $foo->can($some_method) )

not knowing much about perl internals what exactly does "can" test? 

Just the existence of a method? Then 

    if hasattr(foo, some_method):
        ...

would be the equivalent and is often refered to as 
"look before you leap"-style. 

have fun,

    holger




From nicodemus at globalite.com.br  Sat Apr 26 12:44:13 2003
From: nicodemus at globalite.com.br (Nicodemus)
Date: Sat, 26 Apr 2003 13:44:13 -0300
Subject: Vim Newbie =?windows-1252?Q?=96_Python_Edit/Compile/Run_?=
 =?windows-1252?Q?cycle_and_integrating_py_help_questions?=
In-Reply-To: <b8aad5$o9b$1@main.gmane.org>
References: <a791ab19.0304240532.5a025b81@posting.google.com> <3EA7EBD3.3010605@ghaering.de> <b8aad5$o9b$1@main.gmane.org>
Message-ID: <3EAAB75D.1050607@globalite.com.br>


Hans Nowak wrote:

> A not entirely unrelated followup...  (I searched the vim site in 
> vain, and before I ask on the vim mailing list or elsewhere, I figured 
> I'd try c.l.py first, since I'm more comfortable here. :)
>
> I have been trying to edit Python source code with vim on Mac OS X.  
> Most of the settings in my .vimrc file (copied from Windows) work just 
> fine, except for one.  I use Tab to indent and Shift-Tab to dedent (a 
> line, not a block).  Tab behaves as usual, but Shift-Tab doesn't seem 
> to work; it just behaves like I pressed Tab.
>
> The line I use in my Windows _vimrc to achieve this looks like this:
>
>   imap <S-Tab> ^D


Unfortunately, I have no solution for the S-Tab problem, but I use C-D 
to do dedent in insert mode.

Regards,
Nicodemus.







From david at nospam-freenet.org.nz  Wed Apr 23 08:53:55 2003
From: david at nospam-freenet.org.nz (David McNab)
Date: Thu, 24 Apr 2003 00:53:55 +1200
Subject: how to link python against static sys libs?
Message-ID: <pan.2003.04.23.12.53.47.253065@nospam-freenet.org.nz>

Hi,

I'm having a bit of a nightmare trying to build python to link the sys
libs (eg glibc) statically.

After setting all the ./configure options (eg --with-libc, --with-libs
etc), I've got it to the point where it builds a python executable.

But the built python segfaults when it is itself invoked later in the
Makefile. Also, there's a 'case' statement in the Makefile that make barfs
at.

If anyone has successfully built python 2.2 to use *only* static system
libs, including glibc, pthread, math, dl and util, can you please post
some instructions?

Thanks in advance for your help

Cheers
David




From mcherm at mcherm.com  Tue Apr  1 10:33:57 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Tue,  1 Apr 2003 07:33:57 -0800
Subject: else: w/o except: - why not?
Message-ID: <1049211237.3e89b1657664a@mcherm.com>

Tim Peters writes: 
> ...a bare except is bad practice...

Cameron Laird writes:
> I want to be clear about the judgment that "a bare except is bad
> practice".  Are you talking about any unqualified except?  Is
>   try:
>       f1()
>   except ZeroDivisionError:
>       f2()
>   except:
>       f3()
> an instance of "a bare except" in the sense that you intend here?
> And, for you, does
>   try:
>       f1()
>   except:
>       f2()
>       raise
> become *not* bad practice?

I can't speak for Tim, but my OWN opinion is precisely in alignment
with what you have written.

I think that a "bare except" is useful in only two ways. The first
is like a "finally" clause that's skipped if everything works. In
other words, the following bits of code is (almost[1]) equivalent to 
your second example:

    everythingWorked = False
    try:
        f1()
        everythingWorked = True
    finally:
        if not everythingWorked:
            f2()

This doesn't come up much, but it's occasionally useful, and your
version (using bare except with a raise) is a little prettier.

The other situation is running "untrusted plug-in" code of some 
sort. Something like this:

    try:
        randomCodeSomebodyUploaded()
    except:
        logProblem()

But really, that hardly EVER comes up. I wish that guido had made
it harder to spell "bare except" because these situations seem so
unusual. The next best solution is to just avoid it unless you're
SURE you need it.

-- Michael Chermside

        






From alloydflanagan at attbi.com  Tue Apr  8 10:03:33 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 8 Apr 2003 07:03:33 -0700
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com> <b6soav$7p2$1@panix1.panix.com> <Qdpka.2889$Zu2.2627@news01.roc.ny.frontiernet.net>
Message-ID: <a8b7f07a.0304080603.473a8d90@posting.google.com>

Lee Harr <missive at frontiernet.net> wrote in message news:<Qdpka.2889$Zu2.2627 at news01.roc.ny.frontiernet.net>...
...
> 
> From the super docs ;o)  it says you might do this:
> 
>     class C(B):
>         def meth(self, arg):
>             super(C, self).meth(arg)
> 
> 
> but why not just do:
> 
>     class C(B):
>         def meth(self, arg):
>             B.meth(self, arg)
> 
> 
> 
> Is there any difference?
> 
> Any advantage either way?

If all your classes only have one parent, it doesn't make a
difference.  If your classes a) use multiple inheritance, b) several
of them define a method with the same name, and c) all of them call
super() to execute that method, then python does some really neat
stuff to ensure that all the methods of that name get called, and that
it occurs in the "correct" order.

So super() is preferable if a) you're redefining a method but want to
call the base class method(s), and b) your class is part of a multiple
inheritance graph OR may be used in a multiple inheritance graph.

Disclaimer:  I'm new to this stuff myself, so if somebody's got a
better explanation, please speak up!



From sabathe at emi.u-bordeaux.fr  Tue Apr 22 13:00:16 2003
From: sabathe at emi.u-bordeaux.fr (thomas sabathe)
Date: Tue, 22 Apr 2003 19:00:16 +0200
Subject: Event about a selection
Message-ID: <3EA57520.3F39B8D5@emi.u-bordeaux.fr>

Hi,

In a text, a selection appears when a "double-clic" is done.
But I don't know how to get this selection back. How can I do ?

For example:

def modif(event):
    index = myText.tag_ranges(SEL)
    myText.tag_add('valid', index[0], index[1])

myText.bind('<ButtonRelease-1>',modif)



This works, but I'd like to modify the selection with a "double-clic".
It works only with '<ButtonRelease-1>'.

I've read that this key returns the position of the beginning and the
position of the ending of this selection.
I would like to do the same thing with a "double-clic" only.

I don't know what to do because it doesn't work, even if I disactivate
the "simple-click".


Thank you in advance for helping me.





From news at snakefarm.org  Tue Apr  8 11:53:28 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 8 Apr 2003 15:53:28 GMT
Subject: fdopen() question
Message-ID: <slrnb95s3m.u9b.news@cipw08.physik.uni-bonn.de>

Hi,

I'm doing some socket programming and use poll() to check if there is
data to read. poll() returns a list of (fd, event)-tuples where fd is
(as I understand) a file descriptor. Since the input consists of lines
I call os.fdopen(fd, 'r') to have a file object with a readline() method.
Unfortunately fdopen() fails with:

Traceback (most recent call last):
  File "./listener.py", line 27, in ?
    f = os.fdopen(fd, 'r')
OSError: [Errno 9] Bad file descriptor

What am I missing?

cg.




From luke at chipcity.com.au  Sat Apr 26 01:05:14 2003
From: luke at chipcity.com.au (Luke)
Date: Sat, 26 Apr 2003 14:35:14 +0930
Subject: PoPy (postgresql DA for ZOPE)
Message-ID: <3eaa13e8@duster.adelaide.on.net>

Hi, 

I hope this is not too off-topic,

I have installed ZOPE 2.6.1 and want to get it working with Postgresql 7.3 
on my RedHat 8.0 box.
I undertook to install ZPoPyDA-2.0.8 but it needs the PoPy module.
I undertook to install PoPy-2.0.8 but it complains as follows:

############################
loading cache ./config.cache
checking for python... /usr/bin/python
checking python version... 2.2
checking python installation prefix... /usr
checking python installation exec_prefix... /usr
checking definitions in Python library makefile... sed: can't read 
/usr/lib/python2.2/config/Makefile: No such file or directory
done
checking for install... /usr/bin/install
checking location of python library... $(prefix)/lib/python2.2/site-packages
checking location of python shared modules... 
$(exec_prefix)/lib/python2.2/site-packages
checking how to run the C preprocessor... /lib/cpp
checking for mxDateTime.h... yes
checking location of mxDateTime... /usr/lib/python2.2/site-packages/DateTime
checking for PQconnectStart in -lpq... no
configure: error: can't build without PostgreSQL libraries
#################################

I have used the option '--with-postgresql-libraries' and most paths 
(including 
/usr/lib, /usr/share/installs/pgsql/lib [where I have postgres installed])

however the error remains the same.

I want to get ZOPE working with Postgresql, 
any help greatly appreciated,

kind regards

Luke

-- 
===========================================
"once you've seen your future it means you've got to live it"
===========================================



From fperez528 at yahoo.com  Wed Apr 30 18:28:46 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Wed, 30 Apr 2003 16:28:46 -0600
Subject: Self modifying systems question
References: <mailman.1051685036.19081.python-list@python.org>
Message-ID: <b8pimu$30q$1@peabody.colorado.edu>

Sten Kvamme wrote:

> Hello,
> 
> Self modifying code is a well known technology. Has anyone done this in
> Python? What I'm seeking for is examples of software that adds to
> itself (methods and variables) in runtime. I'd like an instantiated
> object to
> be able to add a method to itself.

I've attached a fairly non-trivial example below.  This is a function which
creates functions on the fly, and adds it to a currently running instance
of a python interpreter for shell access.  It's part of ipython
(http://ipython.scipy.org), in case you want to see how it works in 'real
life'.  The code snippet below is just a piece of a much larger code, so it
won't run by itself.  It's meant to illustrate the kinds of things you can
do.

Best,

f.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: alias.py
URL: <http://mail.python.org/pipermail/python-list/attachments/20030430/a197c45f/attachment.ksh>

From exarkun at intarweb.us  Thu Apr 17 09:31:31 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 17 Apr 2003 09:31:31 -0400
Subject: variable in Python
In-Reply-To: <b7lvbh$an3$1@news-reader13.wanadoo.fr>
References: <b7lvbh$an3$1@news-reader13.wanadoo.fr>
Message-ID: <20030417133130.GA25373@meson.dyndns.org>

On Thu, Apr 17, 2003 at 12:22:34PM +0200, Salvatore wrote:
> Hello,
> 
> I have a little doubt.
> when you declare x = 3
> does the variable x 'contains' the value 3 (like in Pascal)
> or does it contains an address which points to
> a memory place containing the value 3 ?

  Neither.  It is a statement (not a declaration) which binds the name "x"
to the integer object associated with the literal 3.  The variable does not
contain the object, it refers to it.  In this way, it is similar to a
pointer; however, note the difference: there is no explicit way to
"dereference" a Python reference, using it is exactly the same as using the
object it refers to; additionally, there is no concept of a "null"
reference.

  Hope this helps,

  Jp

--
It is practically impossible to teach good programming style to
students that have had prior exposure to BASIC: as potential
programmers they are mentally mutilated beyond hope of
regeneration.        -- Dijkstra
-- 
 up 28 days, 9:02, 5 users, load average: 0.04, 0.03, 0.00




From pete.forman at westerngeco.com  Sat Apr  5 16:06:58 2003
From: pete.forman at westerngeco.com (Pete Forman)
Date: 5 Apr 2003 13:06:58 -0800
Subject: Cygwin Python regression testers wanted
References: <mailman.1049385789.4709.python-list@python.org>
Message-ID: <de538c28.0304051306.5ea3f849@posting.google.com>

Jason Tishler <jason at tishler.net> wrote in message news:<mailman.1049385789.4709.python-list at python.org>...
> I'm in the process of updating the Cygwin section of the official Python
> README.  Since I only have access to Windows NT 4.0 and 2000, I am
> seeking Cygwin Python regression testers for the following Windows
> platforms:

Here is my report.  I did rebaseall as requested.  4 tests were
excluded.  There is only 64M of memory on this PC which might account
for test_mmap failing.

W95 OSR2
cygwin-1.3.22-1
python-2.2.2-7


3 tests were excluded due to invalid page faults:
    test_fork1 test_popen2 test_pty
1 test was excluded due to a hang:
    test_hotshot

163 tests OK.
3 tests failed:
    test_dircache test_mmap test_socket
23 tests skipped:
    test_al test_cd test_cl test_curses test_dbm test_dl test_email
    test_email_codecs test_gl test_imgfile test_largefile
    test_linuxaudiodev test_locale test_minidom test_nis test_pyexpat
    test_sax test_socket_ssl test_socketserver test_sunaudiodev
    test_unicode_file test_winreg test_winsound

test test_dircache failed -- Traceback (most recent call last):
  File "/tmp/python.676/usr/lib/python2.2/test/test_dircache.py", line
55, in test_listdir
    self.assertEquals(entries, ['test1'])
  File "/usr/lib/python2.2/unittest.py", line 286, in failUnlessEqual
    raise self.failureException, \
AssertionError: [] != ['test1']


test test_mmap crashed -- exceptions.EnvironmentError: [Errno 12] Not
enough core


test test_socket crashed -- socket.herror: (4, 'No address associated
with name')


test test_al skipped -- No module named al
test test_cd skipped -- No module named cd
test test_cl skipped -- No module named cl
test test_curses skipped -- Use of the `curses' resource not enabled
test test_dbm skipped -- No module named dbm
test test_dl skipped -- No module named dl
test test_email skipped -- No module named test.test_email
test test_email_codecs skipped -- No module named
test.test_email_codecs
test test_gl skipped -- No module named gl
test test_imgfile skipped -- No module named imgfile
test test_largefile skipped -- filesystem does not have largefile
support
test test_linuxaudiodev skipped -- No module named linuxaudiodev
test test_locale skipped -- test locale en_US not supported
test test_minidom skipped -- dlopen: Win32 error 1157
test test_nis skipped -- No module named nis
test test_pyexpat skipped -- dlopen: Win32 error 1157
test test_sax skipped -- no XML parsers available
test test_socket_ssl skipped -- Use of the `network' resource not
enabled
test test_socketserver skipped -- Use of the `network' resource not
enabled
test test_sunaudiodev skipped -- No module named sunaudiodev
test test_unicode_file skipped -- No Unicode filesystem semantics on
this platform.
test test_winreg skipped -- No module named _winreg
test test_winsound skipped -- No module named winsound

-- 
Pete Forman                -./\.-  Disclaimer: This post is originated
WesternGeco                  -./\.-   by myself and does not represent
pete.forman at westerngeco.com    -./\.-   opinion of Schlumberger, Baker
http://petef.port5.com           -./\.-   Hughes or their divisions.



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Mon Apr 28 15:00:31 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Mon, 28 Apr 2003 21:00:31 +0200
Subject: How to write RTF from Python?
In-Reply-To: <878ytuo3uj.fsf@andreasen.org>
References: <3EA97D61.3CC9FE74@easystreet.com> <878ytuo3uj.fsf@andreasen.org>
Message-ID: <3ead7a4f$0$49103$e4fe514c@news.xs4all.nl>

Sorry if this has already been mentioned or not appropriate
but I'm hopping late into this thread and I missed the
original messages...

What we did quite successfully was typing a letter in Word, with special
placeholders such as "%%%DATE%%%", "%%%AGE%%%" on the locations where
you want your data items to appear.
Save this as RTF.
Load it as a 'template' in your program and replace the
placeholders with the actual data values. Presto.

Irmen




From jmeyer at pdi.com  Tue Apr  8 11:54:00 2003
From: jmeyer at pdi.com (Jim Meyer)
Date: 08 Apr 2003 08:54:00 -0700
Subject: exceptions == errors?
In-Reply-To: <16018.5597.933725.584314@montanaro.dyndns.org>
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
	 <16018.5597.933725.584314@montanaro.dyndns.org>
Message-ID: <1049817240.3158.883.camel@utonium.pdi.com>

Hello!

On Mon, 2003-04-07 at 17:20, Skip Montanaro wrote:
>     Mark> Are exceptions always considered errors, or there the same feeling
>     Mark> as in C++ that exceptions can be used for non-error situations?
> 
> Exceptions are frequently used to signal non-error conditions.  For example,
> they can be used to break out of nested loops:

It's always been my impression that exceptions impose considerable
performance overhead and therefore should be used to handle only the
truly exceptional cases; for any predictable case, one should perform
validation and/or error-handling outside the exception mechanism. 

[BTW, I picked up this habit after reading "Effective C++" by Scott
Meyers, a book I cannot recommend highly enough. =]

Some meager benchmarking of Python's performance when handling
exceptions back around v1.6 seemed to indicate that there was a
performance hit. I just went digging for the data from those runs and
can't find it, so I'll be very vague and say that what I saw didn't
dissuade me from using them but persuaded me to use them judiciously.

The counter argument can be found in the archives of this list about a
year ago, which thread starts at http://tinyurl.com/92lv and ends with a
pointer to http://tinyurl.com/92m1 . The thread could be summarized by
saying that you should always implement, profile, and optimize the slow
bits (very true =); and that the performance penalty for exceptions is
distributed over the whole of Python, whether or not you use them.

I wonder if anyone has done recent benchmarking and can offer results
that indicate one direction or another? If not, I'll probably get around
to it in the next six months or so. ;p

Cheers!

--j
-- 
Jim Meyer, Geek at Large                                  jmeyer at pdi.com





From gmuller at worldonline.nl  Mon Apr  7 14:28:59 2003
From: gmuller at worldonline.nl (GerritM)
Date: Mon, 7 Apr 2003 20:28:59 +0200
Subject: python-dev Summary for 2003-03-16 through 2003-03-31
References: <mailman.1049583243.7529.clpa-moderators@python.org>
Message-ID: <b6sg5s$9o5$1@reader1.tiscali.nl>

Brett C wrote:
<...snip...>
> All summaries are archived at http://www.python.org/dev/summary/ .
>
<...snip...>
> I am trying a new way of formatting the Quickies_ section.  I am
> trying non-inline implicit links instead of inlined ones.  I am hoping
> this will read better in the text version of the summary.  If you have
> an opinion on whether the new or old version is better let me know.
> And remember, the last time I asked for an opinion Michael Chermside
> was the only person to respond and thus ended up making an executive
> decision.

--apologies if this appears twice; I posted this yesterday but I did not yet
see it today--

Brett,

I always enjoy reading these summaries. Thnak you.

The last few times I stopped reading the text version in the newsreader at
the link to the archives. Here I click on the link. At this page I look for
the latest version, click on it and start reading again. I would appreciate
a direct link to the html version at the top of the text version.

As a consequence I don't have any feedback on the text version of the
quickies itself.

regards Gerrit
--
www.extra.research.philips.com/natlab/sysarch/









From robin at jessikat.fsnet.co.uk  Thu Apr 10 14:21:32 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Thu, 10 Apr 2003 19:21:32 +0100
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
 <d116fbae.0304100704.68ef4833@posting.google.com>
Message-ID: <AvzRsgAsYbl+Ewbx@jessikat.demon.co.uk>

In article <d116fbae.0304100704.68ef4833 at posting.google.com>, Uche
Ogbuji <uche at ogbuji.net> writes
>martin at v.loewis.de (Martin v. L?wis) wrote in message news:<m3r88kk24p.fsf at mira.
>informatik.hu-berlin.de>...
>> escalation746 at yahoo.com (robin) writes:
>> 
>> > If you want the biggest, boldest approach and don't care about
>> > overhead, use CORBA.
>> 
>> I can't agree with that analysis: CORBA has, of all your alternatives,
>> the least network bandwidth requirements.
>
>True, and Mike Olson did some pretty thorough analysis on this.  See:
>
>http://www-106.ibm.com/developerworks/library/ws-pyth9/
>
>
>--Uche
>http://uche.ogbuji.net

I'm interested to find out why the python server/client pair is so
abominably slow when going across our lightly loaded 100/10 Mbs
ethernet.


I modified the time-client.py script to allow setting of the server name
using an environ script.

When using the time-client on the same machine I see
C:\Python\tmp\test_servers>time-client.py
Connecting to ('localhost', 8080)
Time to connect: 0.040000

Sending a long string to the server
Time to send a string of 21000 chars, 0.000000

Recieving a long stirng from the server
Time to receive a string of 22000 chars, 0.000000

Sending lots of ints to the server
Time to send 5000 ints, 32.297000 (0.006459 per call)



On a machine on the same local net  I see
R:\Python\tmp\test_servers>time-client.py
Connecting to ('192.168.0.3', 8080)
Time to connect: 0.000000

Sending a long string to the server
Time to send a string of 21000 chars, 0.000000

Recieving a long stirng from the server
Time to receive a string of 1455 chars, 0.000000

Sending lots of ints to the server
Time to send 5000 ints, 1007.937000 (0.201587 per call)


What am I missing that causes such painfully slow connections?
-- 
Robin Becker



From carroll at tjc.com  Fri Apr  4 17:48:13 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Fri, 04 Apr 2003 14:48:13 -0800
Subject: nntplib: how to access NNTP response codes on exceptions?
References: <cirn8vsv4bhtp82n1tfil3e2kkounpgvav@4ax.com> <m3brzn27z3.fsf@telia.com>
Message-ID: <hn2s8vs4p97bl4jcskrn34mgi7eartj4h6@4ax.com>

On Thu, 03 Apr 2003 10:25:36 GMT, Andrew Markebo
<andrew.markebo at telia.com> wrote:

>If you catch it through something like:
>
> except NNTPTemporaryError, e:
>
>The response is put in e.response, so you can find the code in the
>first three characters.. 

Thanks!

>You might have missed that parameters can be thrown with the
>exceptions, and thats what e is.. an instansiation of a
>NNTPTemporaryError class.

No, I knew that, but the inclusion of the response code in
NNTPTemporaryError is not in the docs.





From news at snakefarm.org  Thu Apr 24 04:46:49 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 24 Apr 2003 08:46:49 GMT
Subject: Using the MySQLdb module
References: <3EA6E75E.7070305@cis.ohio-state.edu>
Message-ID: <slrnbaf93p.had.news@cipw08.physik.uni-bonn.de>

In Michael Mayhew wrote:
> _mysql_exceptions.OperationalError: (1054, "Unknown column 'atr' in 
> 'field list'")
> 
> sql = "insert into refFlatHum values ( %s, %s, %d, %d )" % ('atr', 
> 'chr1', 90810, 92190)
> 
> c.execute(sql)

What you need here is documentation on SQL, not on MySQLdb. :-) If you want to
use strings in a query you need to enclose them in single or double quotes but
the query string in your example actually looks like this:

"insert into  refFlatHum values ( atr, chr1, 90810, 92190)"

The correct syntax would be:

"insert into  refFlatHum values ( 'atr', 'chr1', 90810, 92190)"

Note the single quotes around atr and chr1. To achieve this in your code you
will have to say:

sql = "insert into refFlatHum values ( %s, %s, %d, %d )" % \
("'atr'", "'chr1'", 90810, 92190)


cg.



From mathias-usenet at valpo.de  Wed Apr  2 01:38:32 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 2 Apr 2003 06:38:32 GMT
Subject: converting from shell script to python
References: <kfn0jaoig1.fsf@shaoz.activesky.com.au> <b6bccj$3fsej$1@ID-15245.news.dfncis.de> <b6db2p$48k88$1@ID-169208.news.dfncis.de>
Message-ID: <b6e0h8$4agg8$2@ID-15245.news.dfncis.de>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com>
wrote:  
> Alternatively, you can substitute things from a
> dictionary instead of a tuple:
> 
>    vars = {'ROOTDIR': '/usr/lib'}
>    CLASSPATH = \
>      "%{ROOTDIR}s/a/a.jar:%{ROOTDIR}s/b/b.jar:%{ROOTDIR}s/c/c.jar
>      " % vars 

Didn't know that before. Its nice - thanks for the hint. 

Mathias



From cjw at sympatico.ca  Mon Apr 14 20:07:38 2003
From: cjw at sympatico.ca (Colin J. Williams)
Date: Mon, 14 Apr 2003 20:07:38 -0400
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net>
Message-ID: <3E9B4D4A.3C1F53F@sympatico.ca>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030414/4f092857/attachment.html>

From exarkun at intarweb.us  Wed Apr  2 10:52:32 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 10:52:32 -0500
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to Python. Good idea? Bad Idea? PEP?
In-Reply-To: <RbCia.6782$9I.555309@news20.bellglobal.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com>
Message-ID: <20030402155232.GB4884@meson.dyndns.org>

On Wed, Apr 02, 2003 at 09:14:24AM -0500, WP wrote:
> Greg Ewing wrote the following Pyrex snippet:
> >    for i from 0 <= i < n:
> >      result.append(a[i])
> 
> This pyrex snippet reminds me (because of its orthogonality to C) of one
> thing, more than tab indents, that I have always thought was bogus in
> Python, which is the handling of integer loops.

  Why?  Just because the current implementation of it happens to be slower
than native-compiled C for-loops?  You're entitled to your opinion, of
course, but Python is a high level language, so usually design decisions
made about it are not driven primarily by runtime efficiency concerns. ;)

> range(0,n) sucks, xrange(0,n) 

  range(n), xrange(n), btw.  0 is the default starting point.

> is an interesting hack which obscures the
> readability of code, and post 2.x versions of python, with generators, and
> other 'inverted' ways of accomplishing the obvious procedural effect of an
> integer for loop, are all just ways of obfuscating otherwise perfectly
> readable python code.  Okay, so I'm being opinionated. I know that. But
> does anyone else thing that for the extremely common (and therefore
> bytecode optimizeable, syntax customizeable) case where you need to do a
> for loop over a range (possibly large range) of integers, that a language
> level keyword construct (such as the above) would be preferable to
> Python's choices of range, xrange, generators, custom class-based
> iterators, and whatever else you can come up with, that the lack of the
> most common idiom, being expressed in a bytecode optimized fashion would
> be good?

  Yes, in fact.  The current behavior of range() is actually one of
Guido's regrets about Python, so you can be pretty sure that it will change
around the time of Python 3.0.  How exactly, who can say?  I'm sure he'll
come up with something good >:)

  Hopefully something better than that Pyrex-style loop you've got above. 
(How do you do steps with that, anyway?)

> 
> Of course, it all comes I suppose from integers being "first class objects" 
> and thus, not actually integers, they are actually immutable objects which 
> contain an integer, and so we get to the ugly condition where "i++" 
> actually decref's one object, and either looks up or creates another 
> object, and increfs it.  Why does it come from this?

  Furthermore, what language are you talking about, when you talk about
"i++"?  Surely not Python.

> Hardly a fast operation to implement, but certainly one of the most common
> idioms in all of programming.

  Ahh, worrying about efficiency again.

> 
> Going back (dizzyingly) from the bytecode/c implementation, to the syntax, 
> for a moment. Imagine if Python had a pyrex-like first-class iteration 
> case, where i (in the above example) is not an integer, but rather a 
> reference to the for-integer-bytecode emitted by the compiler, an object 
> which knows (in C) how to increment i's value quickly, and jump out of a 
> loop when the limit n is reached.

  I'm not sure you're using "first-class" correctly here, but I think I know
what you mean.

  This kind of optimization could be implemented without any syntax change. 
As long as the Python interpreter knows that range(10) is going to return
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9], it can skip actually creating the list of
integers, and just do a highly-optimized loop over those integers.

  Now, in its current state, it is impossible for the Python interpreter to
know this.  What if range() is rebound?  So this can't be done now.  But, if
some of the talk about forbidding the shadowing of builtins bears fruit, it
might someday be possible for CPython to make this optimization, and all
without any ugly syntax hacks.

> 
> Ever since Python 1.5.2, I have been hacking around on my own to see if I 
> could get huge speed improvements out of Python.  I have settled down to 
> the reality that Python+Pyrex, with syntactical speedbumps in between them, 
> are probably the way things are going to stay. But I can dream, can't I?

  That sounds like a good route.  For myself, it is a rare case where I need
any more speed out of the execution of my software than Python gives me, and
I'm much happier with the quality of software that Python's mostly-clean,
mostly-consistent, mostly-readible syntax gives me.  I wouldn't trade a
speedup, even a significant one, for this kind of syntax-creep.

  Jp

-- 
        "I quite agree with you," said the Duchess; "and the moral of
that is -- Be what you would seem to be' -- or, if you'd like it put
more simply -- Never imagine yourself not to be otherwise than what it
might appear to others that what you were or might have been was not 
otherwise than what you had been would have appeared to them to be
otherwise.'"       -- Lewis Carrol, "Alice in Wonderland"
-- 
 up 13 days, 12:00, 3 users, load average: 0.01, 0.02, 0.00




From Kyler at news.Lairds.org  Mon Apr  7 17:23:06 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Mon, 07 Apr 2003 21:23:06 GMT
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <v938h26mdl9s8f@corp.supernews.com>
Message-ID: <h469m-4et.ln1@news.lairds.org>

claird at lairds.com (Cameron Laird) writes:

>I exaggerate a bit.  There are indeed legitimate roles
>for obfuscating technologies.  Managers think they're
>far more common than is true, though, and programmers
>far, FAR overestimate the objective value of source
>code.  

Hey, I program and I think source code is *extremely*
valuable!  It's rare that I can be convinced to use
software for which I can not get the source code.  I
certainly wouldn't ever want to *depend* on software
that I can't control.

What kind of business *pays* to become dependent on
software they don't control?!

--kyler



From matthew at newsgroups.com  Tue Apr 15 21:26:16 2003
From: matthew at newsgroups.com (matthew)
Date: Wed, 16 Apr 2003 13:26:16 +1200
Subject: find non-empty value in a dictionary
Message-ID: <b7ib8b$3rk$1@lust.ihug.co.nz>

hi,

say you have a dictionay d =
{'X': 0.0, 'Y'': 240.2, 'Z': 0.0, 'P': 13.3445}

where p divides pretty much evenly into Y; in this case approx 18

P will always be present and only one of X,Y,Z will have a non-zero value.

How can I easily find which X,Y, or Z is non-zero; esp. given that 
comparing floats is 'inaccurate' as in:

if X != 0:
	<stuff>
elif Y != 0:
	<stuff>
elif Z != 0:
	<stuff>
else:
	<stuff-up>

Also, once the correct axis is found,  exactly the same maths will be 
applied which means that in the above, the same statement will be 
replicated excepting the axis variable. any ideas on the best pythonic 
way to tackle this?

any hints on best practice with comparing floats.

Thanks. matthew.






From rdsteph at earthlink.net  Sat Apr 19 18:15:41 2003
From: rdsteph at earthlink.net (Ron Stephens)
Date: Sat, 19 Apr 2003 22:15:41 GMT
Subject: In praise of PythonCard
Message-ID: <3EA1CB22.4010002@earthlink.net>

I have been playing around with Python for about two years now. I'm 
still a duffer at best, so take whatever I say with a large grain of salt.

Still, I want to point out what a good time I am having with <a href = 
"http://pythoncard.sourceforge.net/">PythonCard</a> as an easy 
gui-builder for Python scripts.

I believe it is the easiest way to create simple gui-based programs with 
Python (with the possible exception of <a href = 
"http://home.att.net/~stephen_ferg/easygui/">easygui</a>, which is also 
really great, but more limited in functionality).

For any newbies or wanna-be gui programmers, I highly recommend you take 
a look at PythonCard. There is a great online <a href = "
http://pythoncard.sourceforge.net/walkthrough2.html">tutorial</a> 
written by Dan Shafer that is simple and one can walk-thru it in a 
couple of hours and be ready to go.

At first glance, one aspect of PythonCard might unnecesarily intimidate 
newcomers. PythonCard programs consist of two code files instead of just 
one. In addition to the normal .py source code file, there is a resource 
file with a .rsrc extension. While this may be intimidating at first to 
extreme newbies (it was for me :-))), in fact it is a great advantage 
for simplicity and ease of learning and use. The resource file is 
actually a very easy to understand little "gui-painter". That is, one 
opens an existing small PythonCard app and modifies its resource file. 
The resource file contains the layout information for all the widgets. 
Best of all, the PythonCard propertyEditor (it opens automatically when 
you open a resource file) allows you to simply choose which widgets you 
want from a menu, and then you simply drag and drop the widget into 
place and maybe change a few attributes; all automaticallY! The layout 
code is, essentially, writtn for you.

But the propertyEditor and resource file combination is simple and easy, 
not like the big GUI-painters in big IDEs like Visual Studio that take 
longer to learn than the whole Python programming language. 
propertyEditor and respource files are the answer to our prayers for a 
simple, easy way to create gui layouts for Python code.

PythonCard loads so easily on Windows that no one can possibly fail. It 
comes in a self extracting .exe file and takes care of all the work. For 
me, installing PythonCard on a SUSE Linux box and also on a Mac OS X 
machine was trickier; I succeeded, but with some difficulties. I will 
just say, PythonCard works on both systems, but one should follow the 
detailed instructions very carefully and to the letter, and it helps to 
have a clean machine without any previous version of wxPython installed. 
(PythonCard runs on top of, so to speak, the wxPython toolkit.)

By all means give PythonCard a try. Kevin Altis and his team are doing a 
great job. I apologize in advance for any technical mistakes I have made 
in this post, I am only trying to explain why PythonCard appeals so much 
to relative newcomers to Python and programming.

Ron Stephens




From skip at pobox.com  Tue Apr 22 17:56:13 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 22 Apr 2003 16:56:13 -0500
Subject: True/False
In-Reply-To: <eppstein-D065BC.14323022042003@news.service.uci.edu>
References: <mailman.1051036237.30103.python-list@python.org>
        <3EA5B173.9420324C@alcyone.com>
        <eppstein-D065BC.14323022042003@news.service.uci.edu>
Message-ID: <16037.47741.606168.32765@montanaro.dyndns.org>

    Erik> This would be okay (although the untyped version of the True and
    Erik> False builtins were introduced in 2.2, not 2.3).

    David> 2.2.2, I thought?  They are not in the 2.2 command-line python
    David> shipped with OS X, forex.

This is why you should avoid version dependencies in your code wherever
possible. ;-) (Not casting aspersions toward Erik or David.  It's just that
you're bound to make mistakes with this sort of minutiae.)

Skip




From max at alcyone.com  Tue Apr 22 17:20:51 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 14:20:51 -0700
Subject: True/False
References: <1051036194.697.177.camel@lothlorien>
		 <16037.36122.922342.405340@montanaro.dyndns.org> <mailman.1051038744.9300.python-list@python.org>
Message-ID: <3EA5B233.81F95B88@alcyone.com>

Ian Bicking wrote:

> >>> None = 10
> <stdin>:1: SyntaxWarning: assignment to None
> 
> Will this happen to True/False?  Probably not, because that would
> *really* mess up backwards compatibility (since try:except: wouldn't
> even work).

There was some talk of it in the past, I don't know whether it's
definitely planned and/or when that would happen, but it seems part of a
general trend of protecting "constants."  Perhaps an exception would be
made for True/False, but given the changes that are occurring here and
there, I'd program defensively around it.  Best way for that is the
version test (ugly) or the physical test for the existence of the name
(elegant but swallows an exception).

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From rmunn at pobox.com  Fri Apr  4 18:49:03 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 04 Apr 2003 23:49:03 GMT
Subject: Converting text to html
References: <3E8E0681.9005524D@easystreet.com>
Message-ID: <slrnb8rhfi.k4j.rmunn@localhost.localdomain>

achrist at easystreet.com <achrist at easystreet.com> wrote:
> 
> Is there any recommended code available to convert plain text into
> text that can go into an html document without breaking it?  I 
> think that I just have to do these three translations:
> 
> 	< =>  &lt;	> =>  &gt;   & 	=>  &amp;
> 
> But I'm not sure if there are any others.  I've got to do this 
> quite a bit while the user is waiting, so it would be nice to
> have something that did it in a way that was not a lot slower
> than any of the alternatives.  
> 
> Surely someone must have cracked this nut sys.maxint times by now.

You want the cgi module, and specifically the cgi.escape() function.

See http://www.python.org/doc/2.2/lib/module-cgi.html and also the
bottom of http://www.python.org/doc/2.2/lib/module-cgi.html for more.

Hope this helps.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From rmunn at pobox.com  Tue Apr 15 11:28:31 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 15 Apr 2003 15:28:31 GMT
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050201322.13220.python-list@python.org> <slrnb9h7u4.it3.rmunn@localhost.localdomain> <3e9a857b$1@news.012.net.il> <slrnb9ksaa.1q1.rmunn@localhost.localdomain> <3e9ae78f$1@news.012.net.il> <mailman.1050342503.3825.python-list@python.org>
Message-ID: <slrnb9nkhp.493.rmunn@localhost.localdomain>

Jp Calderone <exarkun at intarweb.us> wrote:
> On Mon, Apr 14, 2003 at 06:53:32PM +0200, Avner Ben wrote:
>> 
>> "Robin Munn" <rmunn at pobox.com> wrote in message
>>
>> > You're right, I didn't close the file explicitly. But explicit is better
>> > than implicit, so IMHO it's better to give a name to the file object so
>> > that you can close it explicitly:
>> >
>> >     input_file = file("somefile")
>> >     for line in input_file:
>> >         process(line)
>> >     input_file.close()
>> 
>> 
>>     The C# language has a nice construct: using (variable = whatever) { /*
>> code */ }. The variable in the using header is guaranteed to be
>> garbage-collected at the end of the block (or so I think). Could be useful
>> here.
>> 
> 
>   In CPython, the reference count to the file object in:
> 
>     for line in file(...):
>         ...
> 
>   will drop to 0 immediately after the loop, and the file will be closed at
> that point.

But the Python language spec doesn't guarantee that objects will be
garbage-collected immediately when their refcount drops to 0; and, in
fact, Jython does *not* garbage-collect zero-refcount objects
immediately. Relying on garbage-collection is not safe. For most
operations, it's safe enough (if a file stays open a bit longer than you
expected, it's probably not a big deal) but there are times when it is
important to know exactly when the file is closed; I prefer to simply
make a habit of it.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From tismer at tismer.com  Sat Apr  5 20:46:28 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sun, 06 Apr 2003 03:46:28 +0200
Subject: Stackless python
In-Reply-To: <b6n3ce$pdp$1@news-reader12.wanadoo.fr>
References: <221d8dbe.0304050531.107bc603@posting.google.com> <mailman.1049558408.987.python-list@python.org> <b6n3ce$pdp$1@news-reader12.wanadoo.fr>
Message-ID: <3E8F86F4.5000601@tismer.com>

Hi Salvatore,

> When Stackless will be integrate in official Python ?

When Zope Corp is down and I'm able to pay Guido's payroll. :)

> (I hope you'll have success with EveOnline ;-))

Yes, I do believe they are quite happy with Stackless.
I hope to get them to let me visit them once more
and port them to Stackless 3.0, which is again much more
efficient, and performance is what they really love.

ciao & thanks -- chris
-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From skip at pobox.com  Wed Apr  9 11:36:11 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 9 Apr 2003 10:36:11 -0500
Subject: Adding item in front of a list
In-Reply-To: <BIEJKCLHCIOIHAGOKOLHGEIJFEAA.tim.one@comcast.net>
References: <yu99k7e3wxcf.fsf@europa.research.att.com>
        <BIEJKCLHCIOIHAGOKOLHGEIJFEAA.tim.one@comcast.net>
Message-ID: <16020.15851.131590.501310@montanaro.dyndns.org>

    >> I would have thought that after l.insert(-1, 1), l would be
    >> [2, 3, 1, 4], but it doesn't work that way.

    Tim> Alas, list.insert() existed before sequence indices were
    Tim> generalized to give a "count from the right end" meaning to
    Tim> negative index values.  When the generalization happened, it
    Tim> appears that list.insert() was just overlooked.

    Tim> I'd like to change this.  If I did, how loudly would people scream?

No screams from here.  I don't think I've ever used list.insert() with an
index of anything but 0.

Skip




From wijgerden at hetnet.nl  Tue Apr 15 11:02:35 2003
From: wijgerden at hetnet.nl (vwa)
Date: Tue, 15 Apr 2003 17:02:35 +0200
Subject: any benchmark results for python?
References: <b7h40n$28s$1@skates.gsfc.nasa.gov>
Message-ID: <3E9C1F0B.F1B944FE@hetnet.nl>

> benchmark test results for python, in comparison to other languages

Check (where [1] is probably stopped and [2] is active):
[1] http://www.bagley.org/~doug/shootout/ - The Great Computer Language
Shootout.
[2] http://dada.perl.it/shootout/index.html - The Great Win32 Computer
Language Shootout (Win32 version of above Shootout).

HTH,

Frank



From bdesth at removethis.free.fr  Mon Apr 21 21:07:34 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 22 Apr 2003 01:07:34 +0000
Subject: Regarding Dictionaries in python
References: <mailman.1050957134.26691.python-list@python.org>
Message-ID: <3EA495D6.40006@removethis.free.fr>

Mehta, Anish wrote:
> Hello !
> 
> I am having a small query regarding dictionaries in python. From 
> Orelly's Programming Python
> 
>  >>>D= {'a' : 1, 'b' : 2, 'c' : 3, 'd' : 4}
>  >>>D.items()
> [ ('b', 2), ('c', 3), ('d', 4), ('a', 1)]
> 
> 
> why it is not showing the items in order in which i have inserted. 
> Please tell me why it is happing like this.

Because dictionnaries in python are not ordered. I think (not sure 
but...) this is in the Python doc...


Laotseu




From james at anchor.net.au  Fri Apr 25 03:22:51 2003
From: james at anchor.net.au (James Gregory)
Date: 25 Apr 2003 17:22:51 +1000
Subject: Date validation
Message-ID: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au>

Hi all,

I'm writing a killer web app in python and I want to validate all the
input it gets. Amongst said input are dates represented as text strings
of the format 'YYYY-MM-DD' - how can I check whether or not such strings
represent a valid date? I had thought that

    time.strptime('2003-4-31', '%Y-%m-%d')

would throw an exception (since gnome-something tells me that april only
has 30 days), but it just gives me back

    (2003, 4, 31, 0, 0, 0, 3, 121, 0)

indeed I'm even able to use mktime on that without an exception.

The docs for time.strptime inform me that it's libc's fault, which it
probably is, but I'm still stuck for a method to validate this data.

I mean I *could* write this myself, but I had to ask the computer how
many days in April - someone with a better grasp of the date must have
written this stuff before :)

Thanks,

James.






From tebeka at cs.bgu.ac.il  Wed Apr  9 08:42:59 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 9 Apr 2003 05:42:59 -0700
Subject: PythonWin Open(file) path bug?
References: <3E93B4E8.6020003@netscape.net>
Message-ID: <33803989.0304090442.1963536f@posting.google.com>

Hello Vector,

> the file is in the same folder as the python script
> i get the same error if opened from the interactive window
> import sys
> print sys.path
> indeed shows the script dir is on the path so why cant it find the 
> config file?
> 
> If i call the script direct from explorer (no IDE) it works fine
> if I open lptctrl.cfg within pythonWin and save As it then runs and 
> finds the file(it must be save as, even modifing and just save dosent 
> change the error)
There is a notion of 'working folder', this is the folder where the
application is running. In your case when working from the IDE it's
the IDE's working folder ($PYTHON/Lib/site-packages ...). Try giving
full path to your file and you should be fine. Or do os.chdir to the
right directory to change the current working folder.
sys.path has nothing to do with this problem, it's the path where
Python looks for modules when you import them.


HTH.
Miki



From ramen at lackingtalent.com  Tue Apr 15 21:40:12 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Wed, 16 Apr 2003 01:40:12 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <slrnb9p3vo.vg8.ramen@lackingtalent.com> <mailman.1050454017.965.python-list@python.org> <eppstein-8427F2.18234715042003@news.service.uci.edu>
Message-ID: <slrnb9pd4f.vg8.ramen@lackingtalent.com>

In article <eppstein-8427F2.18234715042003 at news.service.uci.edu>,
David Eppstein wrote:
> On Tue, Apr 15, 2003 at 11:04:06PM -0000, Dave Benjamin wrote:
>> > I think that list comprehensions bridge the gap somewhat between FP and OO
>> > methods, and this is one thing I really like them for. For instance:
>> > 
>> >   [obj.method(x) for x in data]
>> > 
>> > looks nicer to me than:
>> > 
>> >   map(lambda x: obj.method(x), data)
> 
> Sure, but wouldn't it be even better to use
>     map(obj.method, data)

Yes, of course. I didn't think that example out thoroughly. This trick
works, but only if that method happens to be a function of one argument. See
Jp Calderone's last post about the message types - he identified the problem
I was trying to convey here.

Thanks,
Dave



From peter at engcorp.com  Tue Apr  1 21:37:42 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 21:37:42 -0500
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de> <CO6cnfDQxO4jnxejXTWc3Q@speakeasy.net>
Message-ID: <3E8A4CF6.ADCCD6A9@engcorp.com>

"A.M. Kuchling" wrote:
> 
> On Tue, 01 Apr 2003 13:58:42 +0200,
>         Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote:
> > I think about moving from Zope to
> > ZODB and (Quixote or Twisted).
> 
> You're mixing levels here.  ZODB is an object database; Zope is for content
> manager and application development, built on top of that object database.
> Twisted is a framework for writing network services; Quixote is a framework
> for writing HTTP/web applications, and can be hosted on top of Medusa or
> Twisted.  (See the March archive of the quixote-users list for some example
> code to run Quixote applications on top of Twisted.)
> 
> So the choices are more like:
>    ZODB or SQL
>    Apache or AOLserver or Twisted or Medusa
>    Quixote or Zope or Webware or ...
> 
> Zope provides content management features that Quixote and Webware don't.
> You could write a content manager in Quixote or Webware, but no one has done
> so yet, as far as I know.

Just a small adjustment: Twisted (as in twisted.web) definitely covers
much more of Zope than just the Medusa part.  In fact, there seems
to be a remarkably small amout (just that related to actual *content
management*, as far as I can tell) which it does not at least begin
to cover, if not cover better and more elegantly than Zope (IMHO).

That doesn't mean it can (yet?) compete with Zope in the majority
of Zope's areas of application, but then I think many people think
Zope is more readily applied to some simple situations than it really
turns out to be.  Twisted, on the other hand, can handle at least
those areas significantly more easily than Zope can.

Not sure where Quixote fits in yet...



From andy at wild-flower.co.uk  Tue Apr 22 14:08:05 2003
From: andy at wild-flower.co.uk (Andy Jewell)
Date: Tue, 22 Apr 2003 19:08:05 +0100
Subject: Python implementation for a web spider \ site mirroring
In-Reply-To: <S8WdnR2SKryGsTijXTWcrg@speakeasy.net>
References: <662cede5.0304220114.54e4dd10@posting.google.com> <S8WdnR2SKryGsTijXTWcrg@speakeasy.net>
Message-ID: <200304221908.05119.andy@wild-flower.co.uk>

On Tuesday 22 Apr 2003 12:48 pm, A.M. Kuchling wrote:
> On 22 Apr 2003 02:14:34 -0700,
>
> 	Tzahi <tzahij at hotmail.com> wrote:
> > I need to copy complete web sites form the web.
>
> Try GNU wget (http://www.gnu.org/software/wget/wget.html); not written in
> Python, but quite featureful.
>
> --amk


Give PAVUK a try if you want something witha GUI interface that can run under 
Windoze or X...  It's not in Python either, but it is rather good:

http://www.idata.sk/~ondrej/pavuk/

hth
-andyj




From aleax at aleax.it  Sat Apr 12 18:43:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 22:43:46 GMT
Subject: Adding item in front of a list
References: <mailman.1050038237.25551.python-list@python.org> <ZUXla.7029$T34.191442@news2.tin.it> <mailman.1050176305.12881.python-list@python.org>
Message-ID: <CE0ma.5922$LB6.169206@news1.tin.it>

Skip Montanaro wrote:

>     Alex> Why can't mylist.insert use just the same "first argument means
>     Alex> different things" convention as range...?
> 
> Guido's feeling is that the convention used by range was a mistake, and
> intends not to repeat it.  (I believe this line of reasoning was already
> posted by Tim.)

It was (but I hadn't seen that post when I asked this question).  And on
reflection I concur -- I *hate* the few places in the Nutshell in which
I couldn't explain a function's behavior by means of Pythonic "optional
arguments", but rather had to resort to brackets meaning "optional" -- and 
range is a prime example of that.


Alex




From kkto at csis.hku.hk  Thu Apr 17 01:21:26 2003
From: kkto at csis.hku.hk (Isaac To)
Date: 17 Apr 2003 13:21:26 +0800
Subject: Memory leak while looping
References: <1020cbf8.0304162100.6fc49e6f@posting.google.com>
Message-ID: <7iistdk8bt.fsf@enark.csis.hku.hk>

>>>>> "Roger" == Roger Hancock <roger_hancock at hotmail.com> writes:

    Roger> I'm just starting to play with writing C extensions for Python.
    Roger> In the process of doing so, I called one of my extension
    Roger> functions thousands of times only to discover that a leak was
    Roger> eating up all the memory in my system.  Upon further
    Roger> investigation, I found I could produce the same result with the
    Roger> following python code with my extension completely out of the
    Roger> picture:

    Roger> #!/usr/bin/env python2

    Roger> for i in range(1,10000000): item = i


    Roger> I would have thought that the memory for "item" would be freed as
    Roger> new "item"s were created.  However, this doesn't seem to be the
    Roger> case.  I would appreciate any clarification on this issue.

It doesn't seem to me that item is eating up any memory.  When running the
loop (with 10000000 changed to 100000) the memory consumption of python
never goes up.  Of course, the object produced by range(1,100000) would take
up 1M RAM by itself, which is not freed until the loop finishes.  If you
dislike this you might try xrange(1,10000000).

Regards,
Isaac.



From pinard at iro.umontreal.ca  Sun Apr  6 17:31:14 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 06 Apr 2003 17:31:14 -0400
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2fzov8iqm.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	<mailman.1049658250.30172.python-list@python.org>
	<m2fzov8iqm.fsf@localhost.localdomain>
Message-ID: <oqr88fgvod.fsf@alcyon.progiciels-bpi.ca>

[Patrick K. O'Brien]

> Do you think Pymacs would help me understand Emacs better, or should I
> just dive into the Emacs source?

Pymacs will _not_ significantly help you at understanding Emacs, in my
opinion.  Diving in Emacs sources might help you better, but probably not
too soon: you will surely see a lot of details, but have more difficulty at
grabbing the overall picture of the problem.  Studying, understanding and
even translating a few Emacs Lisp files might be more fruitful for this.
A careful reading of the Emacs Lisp Reference Manual is surely a good
pre-requisite as well.

> I think getting the core functionality won't be that hard.  And even
> if I only implement the bare minimum, [...]

If you want being able to run an Emacs Lisp file taken at random on the
net, either directly or by translating it by automatic or semi-automatic
means, you surely need a good and wide deal of that minimum functionality.
It might not be excruciatingly hard: not always easy, but likely tedious.

> > All this being said, I sincerely wish that your project will succeed!

> Me too.  Ignorance is bliss, so I might as well enjoy this moment when
> it all seems possible.

I understood you so well! :-)  Keep happy, and have good courage!

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From tj_scarlet at yahoo.com  Wed Apr 16 14:00:00 2003
From: tj_scarlet at yahoo.com (Tj)
Date: 16 Apr 2003 11:00:00 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
Message-ID: <ccc7084.0304161000.39a8252b@posting.google.com>

Can anyone point me to a discussion of Python's scoping system vs.
Scheme-like lexical scoping?  I'd like to know if people find Python's
scoping more intuitive or what.  I know Guido once mentioned it was
because he thought it might encourage a lot of nesting, but now
"nested scopes" have made its way into the language.



From roy at panix.com  Fri Apr 18 19:55:21 2003
From: roy at panix.com (Roy Smith)
Date: Fri, 18 Apr 2003 19:55:21 -0400
Subject: Are reference cycles bad form?
References: <3o11avop6vkh3ake8dgbtst93t1or771ct@4ax.com>
Message-ID: <roy-DE11D6.19552118042003@reader1.panix.com>

Manuel Garcia <news at manuelmgarcia.com> wrote:
> But as a general pricipal, is a design without reference cycles better
> than a design with them?

I think that's a little like asking if singly-linked lists are better 
than doubly-linked lists.  Does your child ever need to know who its 
parent is?  If so, then it probably makes sense to store the reference.

I don't see that there's any "general principal" here one way or the 
other.



From mathie at wossname.org.uk  Tue Apr  1 10:25:31 2003
From: mathie at wossname.org.uk (Graeme Mathieson)
Date: Tue, 01 Apr 2003 16:25:31 +0100
Subject: Catching webbrowser error in RedHat
References: <v8edjq9cc4o13@corp.supernews.com>
Message-ID: <sa6el4ms0ic.fsf@kerouac.projectcolo.org.uk>

In <v8edjq9cc4o13 at corp.supernews.com>,
on Sun, 30 Mar 2003 10:24:56 -0800, "Lance" <lbrannma at cablespeed.com> writes:

>         try:
>             webbrowser.open(fname)
>         except:
>             print "ERROR: "+fname+" not found."

> 2. If the try block is not successful, the Redhat splash screen shows,
> followed by an ugly system message box saying that the file could not be
> found. I want only the message in my except clause to display, not the
> system message.

Check to see if the file exists before loading it up in the web browser?
It will not help for other errors, but that might be the most common
one, eh?  If it's a local file, something along the lines of:

try:
    import urllib
    # This will raise an IOError if the file cannot be found
    fname_check = urllib.URLopener().open(fname)
    webbrowser.open(fname)
except IOError, data:
    print "ERROR: " + fname + "not found: " + data

might do the trick?  Of course, if fname is a remote file, I think
that's going to retrieve it if it does exist, which is probably not a
good idea if it is a large file...    
-- 
Right now, there are scr1pt k1dd13s plotting to DDoS my network, my co-lo
server is not responding  to pings and  the people that I IRC with may be
involved in both.  I'm  sysadmin Graeme Mathieson and this is the longest
day of my life.                               http://www.wossname.org.uk/



From martin at v.loewis.de  Sat Apr 26 03:42:33 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 26 Apr 2003 09:42:33 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com> <3ea947e1$0$198$a1866201@newsreader.visi.com> <mailman.1051301620.25438.python-list@python.org> <3ea9ac4c$0$179$a1866201@newsreader.visi.com> <b8camh$g65$05$1@news.t-online.com> <3ea9b549$0$175$a1866201@newsreader.visi.com>
Message-ID: <m3llxxvh5i.fsf@mira.informatik.hu-berlin.de>

grante at visi.com (Grant Edwards) writes:

> If Python doesn't use a particular library module, then it
> won't get linked in.  The C run-time library isn't a single
> monolithic object.

On most systems, it is, being a shared library.

Now, you may argue that due to demand-paging, the C library won't be
copied entirely into main memory at run-time, which is true. The same
is true for Tcl, though: Tcl commands that aren't used won't get paged
in, either.

> The big advantage was that Tk "understands" Scheme variables.

I see.

> Another advantage was that parameters passed to and and return
> values from Tk methods all use native Scheme data types.
> 
> > Aren't they also obtained by exposing the Tcl object interface?
> 
> Don't know.

This is what the Tcl object API offers: You get native Python data
types in parameters and results.

Regards,
Martin




From ianb at colorstudy.com  Sun Apr  6 16:05:41 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 06 Apr 2003 15:05:41 -0500
Subject: Is there swap function in Python?
In-Reply-To: <b6q0n6$1c55$1@mail.cn99.com>
References: <b6q0n6$1c55$1@mail.cn99.com>
Message-ID: <1049659541.659.67.camel@lothlorien>

On Sun, 2003-04-06 at 14:56, Xin Liu wrote:
> I know it is a too sample question. I can't find anything in my python
> documents about the swap funtion. Of cause, I can write a swap use boost or
> an other way. But the language has grown for 13 years. Why not its library
> include a so basic funtion?

Do you mean, to swap two variables?  You do it like:

a, b = b, a







From R.Brodie at rl.ac.uk  Wed Apr  9 09:34:25 2003
From: R.Brodie at rl.ac.uk (Richard Brodie)
Date: Wed, 9 Apr 2003 14:34:25 +0100
Subject: REST and directory servicies
Message-ID: <b717h2$t3c@newton.cc.rl.ac.uk>

Now I've had a week since Python-UK to reformulate the question:

Suppose I have two RPC services: fruit and vegetables; and two hosts:
hector and achilles. Now I might the services named in some directory
system, say:

//example.com/services/fruit/
//example.com/services/vegetables/

Now, one could reformulate them for REST as:
http://hector.example.com/services/fruit/ and
http://achilles.example.com/services/vegetables

However, that has nodenames in, which is likely to violate the Coolness rule.
So it ought to get rewritten as, for example:

http://services.example.com/fruit/ and
http://services.example.com/vegetables/

Now, it gets ugly enough hacking mod_rewrite to maintain a site of
any complexity. If I throw web services into the mix, it's likely to have
to get done more often; and also robustly, therefore probably redundantly.

Granted, these are implementation issues but has anyone been down this
way before? It seems to be that what I'm looking for is a lightweight,
robust way of mapping a directory service into the http: namespace.
Or maybe I have some fundamental architectural misunderstanding.
(Did you say 'knives')?

Any views?





From aleax at aleax.it  Wed Apr 16 08:51:27 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 12:51:27 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it> <slrnb9o9ob.493.rmunn@localhost.localdomain> <eX%ma.18570$Gb1.6139@nwrdny01.gnilink.net>
Message-ID: <jlcna.22031$LB6.533808@news1.tin.it>

Carl Banks wrote:

> Robin Munn wrote:
>> I suppose I just illustrated how a try ... finally block might be
>> considered "broken and fragile". I tried to use it, and missed some
>> subtleties. I'm still not convinced that a "with"-like statement is
>> necessary in Python, but I think I just went from -1 to -0 on it.
> 
> I wouldn't be so hasty.
> 
> First, you're basing it on a bad example.  Rebinding sys.stdout is bad
> programming, period.  Wanting to save and restore it is even worse.

I disagree very deeply with this statement, and don't want to leave
it unchallenged, lest impressionable readers be left with any
impression that such a statement represents a widely held opinion
among Pythonistas.  The Python Standard library, written by the
best Pythonistas around, shows several examples of how rebinding
sys.stdout (and also restoring it) can be a perfectly good Pythonic
programming practice, and so do sundry other packages authored by
great Pythonistas, such as Aaron Watters' preppy (part of ReportLab's
product line).  No doubt Mr Banks is and will remain convinced he 
knows what is or isn't bad programming, better than (for example)
Tim Peters (author of doctest.py), Guido van Rossum (author, I
believe, of fileinput.py), etc, etc -- but I trust most readers will
be able to judge for themselves on the basis of these facts.

Of course it's not an issue of "argument from authority" -- rather,
it's one of "practicality beats purity".  Say I have to change
the string 'Carl Banks' to the string 'Tim Peters' in a zillion
files, "in-place".  Right now, I write an absolutely trivial script:

import fileinput
for line in fileinput.input(inplace=1):
    print line.replace('Carl Banks','Tim Peters'),

and I just arrange to run this script with the filepaths as
arguments -- done.  And similarly for uncountable other tasks
dealing with "in-place rewriting" of files.  The reason I can
do it so handily is, of course, that module fileinput is quite
obviously indulging in what Mr Banks calls "bad programming"
(and elsewhere has also called "stupid").  Well, what to HIM
is stupid, and bad programming, to ME (and presumably to Tim
Peters, Guido van Rossum, Aaron Watters, and so on) is an
excellent example of Python's superb practicality -- I just
LOVE module fileinput, and it's one of those I use most often
and most happily.  doctest, which also uses just the same "bad
and stupid programming", is close to fileinput in my personal
ranking of most-loved, most-used modules.  And so on.


> Second, that the "rebinding" statement should come before the "try:"
> is *not* a subtlety.  It's the idiomatic way to do it, as familiar to
> those who use it as the "while 1: break" idiom.

I suspect I may have taught, helped and consulted a few more
Python programmers than Mr Banks has, and in my experience
it's a VERY common error indeed (and subtle enough to cause
substantial headscratching to locate it) to write:
    try:
        <initialize>
        <proceed>
        ...
where one MEANS (and should have written)
   <initialize>
   try:
       <proceed>
       ...
because it somehow appears to the brain that "everything" should
be in the try clause, while in fact a small number of early
"initialization" statements should be RIGHT OUTSIDE it.  I have
noticed a similar pattern of errors in the use of quite similar
constructs in Microsoft's proprietary "structured error handling"
C extensions on NT, and also in Java, but I have less experience
using / teaching / helping with / consulting about those similar
environments, so I cannot state with confidence that the problem
does indeed generalize -- that is just an impression I have.

I conceive of 'a try/finally statement' as a conceptual unit, so
_of course_, instinctively, I put the various components thereof
"inside" the (syntactical) unit -- the statement itself -- BUT...
the _initialization_ part must NOT be syntactically "inside", just
CONCEPTUALLY inside but syntactically *just outside*.  That IS a
bit tricky and thus requires careful attention each time one uses
it -- nothing major, of course, but it just doesn't "flow off the
keyboard" quite as naturally as most Python constructs do, which
makes it somewhat more error-prone.

When the initialization is several statements long one even
ends up having to NEST several try/finally statements to make
sure just the right amount of stuff is undone -- then the level
of bothersomeness becomes DEFINITELY major (one of the very few
places where C++ is "higher-level" than Python, thanks to the
use of automatic-lifetime variables whose dtors will automatically
execute, last-in first-out, iff the corresponding ctor has also
executed correctly).  E.g., interleaving two textfiles into one,
in the Classic Python way (using Python 2.3 for convenience):

    
    outfile = open('file12','w')
    try:
        for lines in itertools.izip(open('file1'),open('file2')):
            outfile.writelines(lines)
    finally:
        outfile.close()

is one thing, but if you need to ensure timely finalization
for the files open for reading as well as the one open for
writing, then (sigh)...:

    outfile = open('file12','w')
    try:
        infile1 = open('file1')
        try:
            infile2 = open('file2')
            try:
                for lines in itertools.izip(infile1, infile2):
                    outfile.writelines(lines)
            finally:
                infile2.close()
        finally:
            infile1.close()
    finally:
        outfile.close()

"flat is better than nested", and this rather ridiculous
level of nesting is something I find DEFINITELY bothersome --
the fact that the finalizations aren't readably "aligned"
with the corresponding initializations, e.g.:

            infile2 = open('file2')
                ...
                infile2.close()

doesn't help either!


In practice the "null object" idiom can help a lot here...:

class NullFile:
    def close(self): pass
outfile = infile1 = infile2 = NullFile()

try:
    outfile = open('file12','w')
    infile1 = open('file1')
    infile2 = open('file2')
    for lines in itertools.izip(infile1, infile2):
        outfile.writelines(lines)
finally:
    infile2.close()
    infile1.close()
    outfile.close()

this does a tiny amount of extra work in certain error
situations, but i find that consideration entirely trivial
compared to the very substantial simplification -- removing
the nesting, and letting everything go INSIDE the "try"
clause... just where "instinct" wants to put it.

At least, this is definitely MY personal preference for
coding try/finally statements that would otherwise need
any kind of complicated structure.  Of course, depending
on what kind of finalization operations you need, things
may not necessarily get to be quite THIS simple (restoring
the state of certain globals may not necessarily be quite
as compactly expressed as "closing a file", for example,
though one can sometimes encapsulate it, with some work).


Alex




From aahz at pythoncraft.com  Wed Apr 30 09:47:38 2003
From: aahz at pythoncraft.com (Aahz)
Date: 30 Apr 2003 09:47:38 -0400
Subject: How to call built in function 'open' from inside object that has 'open' method?
References: <db81c492.0304300526.7b8d3c86@posting.google.com>
Message-ID: <b8ok5q$pfk$1@panix1.panix.com>

In article <db81c492.0304300526.7b8d3c86 at posting.google.com>,
vadim <vf at interna.ru> wrote:
>
>I am new to Python and do not know the answer to very simple how to
>'open' from inside another object that has 'open' method. I know that
>in C++ I should use following syntax ::open.
>
>More details on reasons. I have met this problem trying to write a
>simple script use ActiveScripting inside IE.  I am trying "file =
>open(...) but because context window object has 'open' method that
>just opens window it is called first. Sorry for simple question. I was
>trying to find the answer but either it is not on the surface or too
>simple.

Two options:

* Python 2.2+, use file()

* import __builtin__; __builtin__.open()
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From ckea25d02 at sneakemail.com  Tue Apr  8 00:13:29 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 16:13:29 +1200
Subject: Pyrex, enumeration constants, and wrapping #define's
In-Reply-To: <b6soip$okk$1@bob.news.rcn.net>
References: <b6soip$okk$1@bob.news.rcn.net>
Message-ID: <b6thng$8uhs2$1@ID-169208.news.dfncis.de>

Edward C. Jones wrote:
> Does Pyrex accept the following?
> 
>     enum
>         a = 2
>         b = a + 3

Yes, if you change the "enum" to "cdef enum":

   cdef enum:
     a = 2
     b = a + 3

I just ran that through Pyrex 0.7 and it produced:

   enum  {
     a = 2,
     b = (a + 3),
   };

so, as long as your C compiler can cope with that,
all should be well.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mertz at gnosis.cx  Tue Apr 29 12:41:08 2003
From: mertz at gnosis.cx (David Mertz)
Date: Tue, 29 Apr 2003 12:41:08 -0400
Subject: POD vs. reST for standalone writing?
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
 <mailman.1051584944.30800.python-list@python.org>
 <b8kuet$ftn$1@panix1.panix.com> <20030429043815.GB13813@frobozz.local>
 <3EAE801D.4010005@python.org> <mailman.1051625745.20516.python-list@python.org>
 <zypr+kKkXE2Q092yn@gnosis.cx>
Message-ID: <ksqr+kKkX8eH092yn@gnosis.cx>                       

|Yep.  For Twisted's documentation (now over 200 pages as a PDF), I think
|Lore was the right choice (even if we did have to invent it first).  Other
|users will have to make their own decisions about what's right for
|them.

FWIW, I wrote my whole 500+ page book (http://gnosis.cx/TPiP/) using my
own "smart ASCII" format, which is even lighter than reST.  Actually, I
believe it is quite a lot "lighter" (take a look).  The final
book--which I believe is quite attractive--is generated 100% from the
smart ASCII source... including use of fonts, headers, cross-references,
decorative elements, and so on.  The chain is (smart ASCII -> LaTeX ->
PDF -> Publisher puts it on paper; a somewhat more crude HTML outputer
is included in an appendix of the book itself, FWIW).

Then again, there is a lot of special casing involved in this format.
The conversion applies to *this specific* book, rather than to every
documentation project.  For example, all cross references are introduced
by the phrase "SEE ALSO," at the start of a paragraph, and the stuff
that comes after has to be a comma separated list of linkable targets.
But given how I want to book to be, this is no more and no less than
what is needed.  Something like the following is extremely easy to
write... and to read without any special training in the format:

  SEE ALSO, [float], `int`, `long`, `sys.maxint`, [operator]

It's true that there is something slightly "funny" about the square
brackets and backticks.  You would not necessarily know at a glance that
the brackets surround module names, and backticks surround standard
function names.  But not knowing it does not distract much from reading
(and the same convention is used in all contexts). reST has a similar
property in readability, but with a larger "vocabulary" of special
names (which is both more expressive and more confusing).

Similarly, starting a header with the words "CHAPTER" or "SECTION" or
"TOPIC" each mean something a bit different to the final output.  But
deciding among those is *exactly* what I need to do while writing, it's
not at all a matter of *remembering* an odd syntax variant.  Even
the horizontal lines of dashes I occassionally use in the source are
only there to make the source look nicer in 'less', they don't affect
the processing.  And for a reader of the ASCII format, this is a
transparent, IMO, as anything can possibly be.  But again, the
organization and "markup" is specific to a book of a fairly similar sort
to mine.

As well...  I lied slightly in the above.  I also did the compositing of
my book.  I can generate a *pretty good* book with no annoying special
punctuation/markup.  But for a *final* book, there are a lot of things
to worry about:  whitespacing between words cannot be too loose,
consecutive lines should not begin with the same words, facing pages
must balance in length, successive lines should not both end with
hyphen, and so on.  At the end of the process (and not at the prior
URL), I used a number of really picky and obscure extra punctuation to
adjust for these camera-ready requirements.  Mostly, I chose a bunch of
hi-bit ASCII characters or obscure punctuation sequences to force layout
issues on a page-by-page basis.  If I were to change the page size by
1/4" or change any fonts, these odd markups would need to go in entirely
different places.  But frankly, I don't think that reST, DocBook, Lore,
or anything else, would manage to avoid such manual final touches for
book-quality output.

Yours, David...

--
 mertz@  _/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: \_\_\_\_    n o
gnosis  _/_/             Postmodern Enterprises            \_\_
.cx    _/_/                                                 \_\_  d o
      _/_/_/ IN A WORLD W/O WALLS, THERE WOULD BE NO GATES \_\_\_ z e





From BPettersen at NAREX.com  Wed Apr 30 16:00:47 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 30 Apr 2003 14:00:47 -0600
Subject: Question about accessing class-attributes.
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE192@admin56.narex.com>

> From: Michele Simionato [mailto:mis6 at pitt.edu] 
> 
[..]
> I am sure Alex is going to answer, but since you asked me a related
> question in another thread ...

Sorry, not trying to ignore you, just trying to figure out where the
disconnect was :-)

> Finally, now I see where is the source of the confusion:
> 
> len(S) fails because meta.__len__ is NOT defined in
> you first example. The point is that __getattr__ works on instances of
> meta, NOT on meta itself, therefore S.__len__ is recognized thanks

Well, the whole point of a meta class is that it's "instances" are
regular classes, i.e. the class of a class is called the metaclass...

[meta-meta example...]

> I must say, however, that len(S) is not *literally* 
> type(S).__len__(S):

Yes, that was one of my problems. Where did you find this rule? I can't
find a real reference to it anywhere. For classic classes lookup is
o.__class__.__dict__['attr'].__get__(o.__class__, o) for regular
methods, which sort of looks like the above, but...

[...]
> I did not understood your point in the other thread, I didn't imagine
> you where thinking about meta-metaclasses. As you see, there are 
> situations where they may be useful, i.e. in the customization of 
> metaclasses, but most people don't think at this level yet ;)

I wasn't (or at least wasn't trying to :-). They're facinating beasts
though -- a semantic system with recursion in its axioms ;-)

-- bjorn




From jason at injektilo.org  Tue Apr  1 02:24:12 2003
From: jason at injektilo.org (Jason Diamond)
Date: 31 Mar 2003 23:24:12 -0800
Subject: changing icon to a Tkinter gui program
References: <b6abtb$ldp$1@lacerta.tiscalinet.it>
Message-ID: <6768b874.0303312324.10417d17@posting.google.com>

"Federico" <maschio_77 at hotmail.com> wrote in message news:<b6abtb$ldp$1 at lacerta.tiscalinet.it>...
> How can I change the icon of the windows of a Tkinter Gui program?
> Thanks

If you're using Windows, I posted a solution a few weeks ago that
might work for you.

http://groups.google.com/groups?selm=6768b874.0303172051.36a3d482%40posting.google.com

Jason



From h_schneider at marketmix.com  Thu Apr 17 09:56:00 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Thu, 17 Apr 2003 15:56:00 +0200
Subject: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net>
Message-ID: <b7mbpe$pqt$06$1@news.t-online.com>

What exactly are the internal differences between the 2.1 and 2.2 release of
py2exe ?

--Harald

"Thomas Heller" <theller at python.net> schrieb im Newsbeitrag
news:y929gt00.fsf at python.net...
> "Harald Schneider" <h_schneider at marketmix.com> writes:
>
> > Hi,
> >
> > I created a NT service under with py2exe under Python 2.1
> > The service ran without any problems.
> >
> > Then I upgraded to Python 2.2 and py2exe for 2.2 and now I have a big
> > problem:
> > All functions that access registry keys from within the service fail.
> > When I run my application outside the Service class, everything works
well.
> >
> > Any idea?
>
> Whithout any further information, one can only guess.
> Permission problems? Did you not only move from Python 2.1 to 2.2,
> but also from NT to 2000?
>
> What error do you get?
>
> Thomas





From aahz at pythoncraft.com  Mon Apr 21 12:20:44 2003
From: aahz at pythoncraft.com (Aahz)
Date: 21 Apr 2003 12:20:44 -0400
Subject: looping and deleting
References: <3cf72d52.0304210801.2128eca2@posting.google.com>
Message-ID: <b815os$618$1@panix1.panix.com>

In article <3cf72d52.0304210801.2128eca2 at posting.google.com>,
scn <regnivon at netscape.net> wrote:
>
>what is a good method for looping over a file line-by-line and
>deleting the first 3 characters?

Are you trying to modify the file in-place?  If so, look at the
fileinput module.  If not:

input = <somefile>
output = <some other file>
for line in input:
    output.write(line[3:])
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From http  Thu Apr 10 03:18:09 2003
From: http (Paul Rubin)
Date: 10 Apr 2003 00:18:09 -0700
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com>
Message-ID: <7xwui295xq.fsf@ruckus.brouhaha.com>

Alex Martelli <aleaxit at yahoo.com> writes:
> __slots__ is, essentially, an optimization (in terms of memory
> consumption) intended for classes that are likely to be created 
> in very large numbers: you give up the normal flexibility of
> being able to add arbitrary attributes to an instance, and what
> you gain in return is that you save one dictionary per instance
> (dozens and dozens of bytes per instance -- that will, of course,
> not be all that relevant unless the number of instances that are
> around is at the very least in the tens of thousands).

I generally use __slots__ as a way to "declare" instance variables,
so I get an exception if I accidentally assign to an undeclared one.
I'd hope that was one of __slots__'s intended uses.



From staschuk at telusplanet.net  Sat Apr 19 13:17:03 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 19 Apr 2003 11:17:03 -0600
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>; from guy@obstruction-no-spam.com on Fri, Apr 18, 2003 at 08:21:08PM +0000
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
Message-ID: <20030419111703.B587@tibia.amotlpaa.bogus>

Quoth Guy Middleton:
> What is the most Pythonic way to sum the n-th element of a list of lists (or
> of tuples)?
  [...]
> 	>>> x = (("a", 1), ("b", 2), ("c", 3))
> 	>>> # want to sum 1+2+3 = 6

Here's a variant on the reduce solution:

    reduce(operator.add, zip(*x)[1])

zip(*x) transposes a list of lists, turning columns into rows,
whereupon you may simply index for the column you want.  In this
case this technique is surely less efficient than [y[1] for y in
x], since it creates a list for every column in x.

(I make no claim that this is the most Pythonic approach.  It's
only clear if zip(*x) is a well-known idiom, and I think it might
not be.)

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From mcherm at mcherm.com  Tue Apr 29 17:13:32 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Tue, 29 Apr 2003 14:13:32 -0700
Subject: Endorsement of list comprehensions
Message-ID: <1051650812.3eaeeafc316cc@mcherm.com>

Carsten Gehling writes:
> Okay I understand everyting except this contruct:
> 
> keys = [d[0] for d in cursor.description]
> 
> Let me guess: it loops through cursor.description (obvious), for each of the
> items (which are tuples) it "outputs" element 0. The output is then
> automatically formatted in a way, that conforms to the list creation syntax.
> 
> Or something like that. Kind of neat - haven't seen that in any other
> language, that I've used.

YES! Exactly!

This is a feature called "list comprehensions", and it works just like
you describe. You type out:

    [ func_using_var for var in some_list if condition_using_var ]

and Python makes that into the following loop:

    aList = []
    for var in some_list:
        if condition_using_var:
            aList.append( func_using_var )
    return aList

(except, of course, that there's no "aList" variable... python just
uses the list that would be generated that way).

The original idea comes from "Haskell", another cool programming
language, and the strongest argument I've EVER heard for why this
is a good feature is the little "stream-of-consciousness" description
you just gave which seems to show that as a new-to-python programmer,
it took you only a few moments to figure it out (correctly!) on your
own!

Thanks for endorsing one of my favorite Python features!

-- Michael Chermside





From avner at skilldesign.com  Mon Apr 14 05:55:05 2003
From: avner at skilldesign.com (Avner Ben)
Date: Mon, 14 Apr 2003 11:55:05 +0200
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050201322.13220.python-list@python.org> <slrnb9h7u4.it3.rmunn@localhost.localdomain>
Message-ID: <3e9a857b$1@news.012.net.il>

"Robin Munn" <rmunn at pobox.com> wrote in message
news:slrnb9h7u4.it3.rmunn at localhost.localdomain...
> Skip Montanaro <skip at pobox.com> wrote:
> > Recent versions of Python (2.2 and later) allow this:
> >
> >     file = file("somefile")
> >     for line in file:
> >         process(line)
>
> Ewww! Don't *do* that! You just shadowed the "file" builtin. Use a
> different name for your file object -- I prefer "input_file":
>
>     input_file = file("somefile")
>     for line in input_file:
>         process(line)

I have tried this shortcut and it works! (This gets rid of naming problem
for good.)

for line in file('/dir/filename.ext'):
         print line

The only problem with this solution is that it leaves the timing of closing
the file undefined. But then, neither of you guys have closed the file
explicitly either!

        Avner.





From aleax at aleax.it  Sat Apr 12 11:00:29 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 15:00:29 GMT
Subject: am i missing something ?
References: <3e98226d$0$26858$626a54ce@news.free.fr>
Message-ID: <hSVla.6394$T34.168947@news2.tin.it>

shagshag13 wrote:

> hello,
> 
> i would like to strip only non white space from a string, i have found
> strip in documentation as :
> 
> strip([chars])
> Return a copy of the string with leading and trailing characters removed.
> If chars is omitted or None, whitespace characters are removed. If given
> and not None, chars must be a string; the characters in the string will be
> stripped from the both ends of the string this method is called on.

This is true in Python 2.2.2 (the current stable version), 2.3 (the
current alpha ones), and future ones.  It was not true in 2.2.1 and
previous versions.


> but i can't get it work :
> 
>>>> s = 'this is an example \n'
>>>> s.strip("\n")
> Traceback (most recent call last):
>   File "<pyshell#184>", line 1, in ?
>     s.strip("\n")
> TypeError: strip() takes no arguments (1 given)
> 
> i think i'm missing something...

Yes -- you missed telling us what version of Python you use.  E.g.,
with 2.2.2:

[alex at lancelot Modules]$ python2.2
Python 2.2.2 (#1, Oct 24 2002, 11:43:01)
[GCC 2.96 20000731 (Mandrake Linux 8.2 2.96-0.76mdk)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> s = 'this is an example \n'
>>> s.strip("\n")
'this is an example '
>>>

but older versions would give the behavior you report.  Of course,
the docs refer to a specific version, too -- the online ones are
normally appropriate for the latest stable version (well, there are
several online ones, but as you don't tell us where you took your
docs reference, it's hard to guess...).


Alex




From mb at muenster.de  Tue Apr 22 14:37:11 2003
From: mb at muenster.de (Martin Bless)
Date: Tue, 22 Apr 2003 18:37:11 GMT
Subject: Bug? Pythonwin editor trouble
Message-ID: <3ea58721.1845828@news.muenster.de>

(1) I'm using these versions:
PythonWin 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on
win32,  win32all build 152

(2) Problem:
The editor causes trouble once you start using german "Umlaute".
When you type (each line here representing a keypress and what you get
and a comment):

" -> "     # ok
? -> "?     # ok
" -> "?     # " is missing
" -> "?"    # one " is still missing    
a -> "?""    # type a, and " turns up
a -> "?""a   # one a is missing

Happens both in the interactive windows and in file editing windows.
Tends to really turn down usability.

Previous versions used to work all right. I can't tell which version
brought this behavior along.

(3) Can somebody confirm the error?

(4) What can be done about it?

Hi to all,

Martin




From gh at ghaering.de  Sun Apr 20 19:45:55 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: Mon, 21 Apr 2003 01:45:55 +0200
Subject: Mail and text extraction
In-Reply-To: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com>
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com>
Message-ID: <20030420234554.GA1084@mephisto.ghaering.test>

* Dino Levy <dino_levy at mindlessSPAM.com> [2003-04-21 01:19 +0200]:
> I have the following problem:
> I'm trying to make a script which will connect to POP3 server, print
> number of mails, print their size and subject, and then will ask the user
> which mail should be downloaded...
> So far so good, but I have problem: 
> [...]
> The
> mail looks like this:
> Agnell company
> 
> -Worker		Payment
> Joe Doe		500$
> Mary Smith	550$
> Karen Bay	495$
> 
> -New Employee: Peter Maley
> 
> some text...
> 
> -Raises:
> some text
> 
> And all you want to do is to connect to your POP3 server, and download
> the section called Raises ( for example, two or three lines after the
> title ), so all you want to do is to download those few lines containing your
> text... ( ONLY those lines, not the whole e-mail ) [...]

AFAIK, the POP3 protocol doesn't support downloading parts of emails. To
my knowledge, you can either download the headers (via TOP) or the body
(via RETR).

Of course I'd suppose you could download *up to* the part you're
interested in, then after reading as far as you're interested, close the
socket connection. That's not nice, but it should be possible :-) It'd
also involve doing it low-level style with sockets.

Perhaps you could use IMAP instead? I believe that with IMAP it is
possible to download parts of email messages (MIME parts), but even here
I doubt it's easily possible to cleanly download only a few lines of a
MIME part.

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/




From km-news1 at onlinehome.de  Fri Apr 25 04:59:51 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Fri, 25 Apr 2003 10:59:51 +0200
Subject: Tkinter - tkFileDialog - need help
References: <oprn5mygrymu3s0q@news.cis.dfn.de> <b8aofk$iqf$1@news-reader14.wanadoo.fr>
Message-ID: <oprn6gl1klmu3s0q@news.cis.dfn.de>

Hello Eric,

> Tk uses the regular "open file" dialog from Windows , so the only known 
> way (at least to me) is to configure Windows to display the hidden files.

thanks, but this is not the problem, because i always have enabled this 
option. And i can see the hidden directorys in win-explorer, but not in 
tkFileDialog. So it seems that tkFileDialog has its own "show/don't show 
hidden dirs" flag???
But how to enable this?

If this is not possible, than i have a problem... :-(

Does Tkinter have a other function to open a directory select box?

I just tried tkFileDialog.askopenfile(initialdir="C:/", title="Test").
Surprise! This dialog shows the hidden directorys!!

But i could not use this, because i need a dircectory-name, not a file.

-- 
Gru? - regards Klaus :-)



From nils at bergljot.dyndns.org  Mon Apr  7 10:56:13 2003
From: nils at bergljot.dyndns.org (Nils Grimsmo)
Date: 7 Apr 2003 15:56:13 +0100
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no> <3E915BA0.A4B5286@engcorp.com> <pan.2003.04.07.11.36.03.78556@idi.ntnu.no> <3E917450.CBD5CDD9@engcorp.com>
Message-ID: <3e91837d$1@news.broadpark.no>

In article <3E917450.CBD5CDD9 at engcorp.com>, Peter Hansen wrote:
> Well, this time I actually tested your program, under Windows with
> Python 2.2.2 and wxPython 2.3.3.1 and it works just fine.  I click 
> the button and see foo/bar/foo/bar over four seconds.   
> 
> What platform are you on, and what versions?

linux 2.4.19
xfree86 4.2.1
python 2.2.2
wxPython 2.4.0.3
gtk 1.2 and 2.0

is it gtk's fault?


klem fra nils
-- 
nils grimsmo
nilsgri at idi.ntnu.no



From BPettersen at NAREX.com  Fri Apr 18 10:52:31 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Fri, 18 Apr 2003 08:52:31 -0600
Subject: Teaching : Python, Scheme, Java...
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFDE@admin56.narex.com>

> From: Alex Martelli [mailto:aleax at aleax.it] 
> 
> Donnal Walter wrote:
> 
> > 
> > What is the difference between "recursive calls" and "tail 
> > recursion"? Sometimes I have written methods that call the 
> > same method on a parent or child object, with the depth of 
> > recursion being limited by a method of the same name NOT 
> > calling itself. In the past I have thought of this
> > as tail recursion. Is this incorrect?
> 
> A function-call is "tail" if it's the last thing the caller does,
> i.e., basically, if the caller executes "return somefunction(args)"
> or a semantically equivalent sequence.
> 
[..recursive..]
> 
> A function-call is "tail recursive" if it's tail and recursive.
> Such calls can be optimized by avoiding the allocation of a new
> stack frame -- the same frame object that's already on the
> stack to handle this specific call can be re-used (if you're
> willing to lose traceback information -- Python currently does
> not allow that).

[To digress, elaborate, use my MS for something :-)] It's acutally more
general. Any tail calls (recursive or not) can be optimized, trivially
(assuming the return address is stored in sp, etc.), instead of:

   return f()

being compiled to (A) -- excuse my shorthand:

         push sp, loc1
         jump f
   loc1: pop sp
         reg1 = pop fn_result_stack # 1
         push reg1, fn_result_stack # 2
         jump sp

(1 and 2 would disappear with a peephole optimizer, assuming there is
one) tail call optimization creates (B):

         jump f

with the last two lines of f looking like the last two lines of (A) the
semantics are the same.

[[Now what if you came up with an algorithm that converted any function
to a function with tail call? Your code for return would be really fast,
it would never "return" to a previous state, you wouldn't need a stack
anymore, implementing continuations would be trivial... They did it for
SML, and the various steps, algorithms, etc. are nicely described in:

   Compiling with Continuations, Andrew Appel
   Cambridge University Press 1992, ISBN 0-521-41695-7

which is one of the more interesting books I read in college both for
language and compiler theory (couldn't find an online-ref).]]

-- bjorn




From jxlu at fudan.edu.cn  Wed Apr 30 00:10:49 2003
From: jxlu at fudan.edu.cn (=?gb2312?B?wqy9o9Db?=)
Date: Wed, 30 Apr 2003 12:10:49 +0800
Subject: How to make a repeatedly timer with 5ms precision?
Message-ID: <003101c30ece$7bd1e170$020010ac@lujx>

I am trying to make a repeatedly timer with high precision, the following program 'countdown.py' is an example which shows the way how I tried to make it.
The problem is that it doesn't work properly with tick set to 5ms or smaller.
Is there a better way to implement a more precise repeatedly timer?
C:\>type countdown.py
#Syntax: countdown [tickle [count]]

import sys
import threading
import time

tick=1.0
count=5
argc=len(sys.argv)
if argc==1:
    pass
elif argc==2:
    tick=float(sys.argv[1])
elif argc==3:
    tick=float(sys.argv[1])
    count=int(sys.argv[2])
else:
    print 'Syntax: countdown [tickle [count]]'
    sys.exit()
if tick<=0.0 or count<0:
    print 'Condition: tick<=0.0 or count<0 are not allowed.'
    sys.exit()
def countdown(tick,count):
    if count>0:
        t=threading.Timer(tick,countdown,[tick,count-1])
        t.start()
        print count,time.time()
    elif count==0:
        end=time.time()
        print count,end
        print 'end-start=',end-start
t=threading.Timer(tick,countdown,[tick,count-1])
start=time.time()
print count,start
t.start()

C:\>countdown.py 1.0 10
10 1051675124.74
9 1051675125.75
8 1051675126.75
7 1051675127.75
6 1051675128.75
5 1051675129.76
4 1051675130.76
3 1051675131.76
2 1051675132.76
1 1051675133.76
0 1051675134.76
end-start= 10.0230000019

C:\>countdown.py 1.0 10
10 1051675137.52
9 1051675138.52
8 1051675139.52
7 1051675140.52
6 1051675141.52
5 1051675142.52
4 1051675143.52
3 1051675144.52
2 1051675145.53
1 1051675146.53
0 1051675147.53
end-start= 10.0130000114

C:\>countdown.py 0.1 10
10 1051675155.99
9 1051675156.09
8 1051675156.19
7 1051675156.3
6 1051675156.4
5 1051675156.5
4 1051675156.6
3 1051675156.7
2 1051675156.8
1 1051675156.9
0 1051675157.0
end-start= 1.01100003719

C:\>countdown.py 0.1 10
10 1051675160.88
9 1051675160.99
8 1051675161.09
7 1051675161.19
6 1051675161.3
5 1051675161.41
4 1051675161.51
3 1051675161.61
2 1051675161.71
1 1051675161.81
0 1051675161.91
end-start= 1.03199994564

C:\>countdown.py 0.01 10
10 1051675172.42
9 1051675172.43
8 1051675172.44
7 1051675172.46
6 1051675172.47
5 1051675172.48
4 1051675172.49
3 1051675172.5
2 1051675172.51
1 1051675172.52
0 1051675172.53
end-start= 0.110999941826

C:\>countdown.py 0.01 10
10 1051675176.4
9 1051675176.41
8 1051675176.42
7 1051675176.43
6 1051675176.44
5 1051675176.45
4 1051675176.47
3 1051675176.48
2 1051675176.49
1 1051675176.5
0 1051675176.51
end-start= 0.110000014305

C:\>countdown.py 0.005 10
10 1051675213.6
9 1051675213.61
8 1051675213.62
7 1051675213.63
6 1051675213.64
5 1051675213.65
4 1051675213.66
3 1051675213.67
2 1051675213.68
1 1051675213.69
0 1051675213.7
end-start= 0.100000023842

C:\>countdown.py 0.005 10
10 1051675217.88
9 1051675217.89
8 1051675217.9
7 1051675217.91
6 1051675217.92
5 1051675217.93
4 1051675217.94
3 1051675217.95
2 1051675217.96
1 1051675217.97
0 1051675217.98
end-start= 0.100000023842

C:\>




From tanzer at swing.co.at  Mon Apr 14 04:02:57 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Mon, 14 Apr 2003 10:02:57 +0200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: Your message of "Sat, 12 Apr 2003 15:11:44 GMT."
             <Q0Wla.4427$LB6.127243@news1.tin.it>
Message-ID: <E194yvZ-0002aj-00@tswings.swing.cluster>

Alex Martelli <aleax at aleax.it> wrote:

> IMHO the most maddening use of __repr__ is that done by the
> implementation of __str__ for all built-in container types:
>
> >>> class X:
> ...   def __str__(self): return 'str'
> ...   def __repr__(self): return 'rep'
> ...
> >>> print [X()]
> [rep]
> >>>
>
> There's supposed to be a good reason for this to print [rep],
> rather than [str] as common sense would suggest, but I keep
> forgetting it (so it's probably nor very compelling;-).

You aren't seriously suggest the behavior

>>> print ["abc", "d[e]"]
[abc, d[e]]

, are you?

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From tvassila at siac.com  Thu Apr  3 16:40:28 2003
From: tvassila at siac.com (Thanos Vassilakis)
Date: Thu, 3 Apr 2003 16:40:28 -0500
Subject: Socket timeouts for HTTP connections (httplib)?
Message-ID: <OFBA5EB224.B44B8D44-ON85256CFD.0076E783@nsmtp.siac.com>

Don't hack it.

This is easy and usually stable.

Your only possible complications will  be if you are using threads.

thanos



                                                                                                                            
                    Nick Bower                                                                                              
                    <nick_bower at yah       To:     Thanos Vassilakis <tvassila at siac.com>                                     
                    oo.com.au>            cc:                                                                               
                                          Subject:     Re: Socket timeouts for HTTP connections (httplib)?                  
                    04/03/2003                                                                                              
                    02:39 PM                                                                                                
                                                                                                                            
                                                                                                                            




thanks thanos - i'll look at this too.  turns out 2.3 has socket timeouts -
i may be able to hack httplib in some simple way.


 Thanos Vassilakis <tvassila at siac.com> wrote:

 With the httplib I use the classic signal.SIGALRM/signl.alarm combination.

 see the example on signal
 http://www.python.org/doc/current/lib/node216.html

 We use this successfully on the floor of a very big exchange.


 thanos



 -----------------------------------------
 This message and its attachments may contain privileged and confidential
 information. If you are not the intended recipient(s), you are prohibited
 from printing, forwarding, saving or copying this email. If you have
 received this e-mail in error, please immediately notify the sender and
 delete this e-mail and its attachments from your computer.



Yahoo! Mail - For a better Internet experience








-----------------------------------------
This message and its attachments may contain  privileged and confidential information.  If you are not the intended recipient(s), you are prohibited from printing, forwarding, saving or copying this email.  If you have received this e-mail in error, please immediately notify the sender and delete this e-mail and its attachments from your computer.





From tim.one at comcast.net  Wed Apr  9 11:23:51 2003
From: tim.one at comcast.net (Tim Peters)
Date: Wed, 09 Apr 2003 11:23:51 -0400
Subject: Adding item in front of a list
In-Reply-To: <yu99k7e3wxcf.fsf@europa.research.att.com>
Message-ID: <BIEJKCLHCIOIHAGOKOLHGEIJFEAA.tim.one@comcast.net>

[Andrew Koenig, starting with l=[2, 3, 4]]
> ...
> I would have thought that after l.insert(-1, 1), l would be
> [2, 3, 1, 4], but it doesn't work that way.

Alas, list.insert() existed before sequence indices were generalized to give
a "count from the right end" meaning to negative index values.  When the
generalization happened, it appears that list.insert() was just overlooked.

I'd like to change this.  If I did, how loudly would people scream?

Guido says he also wishes list.insert() had been defined with the arguments
in the opposite order, so that list.insert(object) could have a natural
default index argument of 0.  I'd like to change that too, but it's clearly
too late for that one.





From aahz at pythoncraft.com  Thu Apr 10 08:31:50 2003
From: aahz at pythoncraft.com (Aahz)
Date: 10 Apr 2003 08:31:50 -0400
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com>
Message-ID: <b73o7m$lci$1@panix1.panix.com>

In article <b7340h0r7m at enews1.newsguy.com>,
Alex Martelli  <aleaxit at yahoo.com> wrote:
>
>__slots__ is, essentially, an optimization (in terms of memory
>consumption) intended for classes that are likely to be created 
>in very large numbers: you give up the normal flexibility of
>being able to add arbitrary attributes to an instance, and what
>you gain in return is that you save one dictionary per instance
>(dozens and dozens of bytes per instance -- that will, of course,
>not be all that relevant unless the number of instances that are
>around is at the very least in the tens of thousands).

More than that, because of the way modern computer architectures do
caching, using __slots__ tends to speed up instance attribute access.
This also is an optimization, but it's effective with even hundreds of
instances, if you're rapidly switching between instances.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From blaktyger at hotmail.com  Thu Apr 10 12:24:50 2003
From: blaktyger at hotmail.com (Blaktyger)
Date: 10 Apr 2003 09:24:50 -0700
Subject: Python compared toJava
Message-ID: <c9e6ac92.0304100824.eec0f5a@posting.google.com>

Hi!

I recently read that the Python interpreter is like the Java Virtual Machine.
Is that true?



From staschuk at telusplanet.net  Sat Apr 26 12:21:42 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 26 Apr 2003 10:21:42 -0600
Subject: Dealing with files...
In-Reply-To: <Y%wqa.4238$lf3.720331@cletus.bright.net>; from gk_2345@yahoo.com on Sat, Apr 26, 2003 at 10:46:46AM -0400
References: <pan.2003.04.25.22.39.14.432753@soul.net> <AZiqa.5752$3M4.154723@news1.tin.it> <pan.2003.04.25.23.29.48.204375@soul.net> <RZlqa.4182$lf3.712485@cletus.bright.net> <pan.2003.04.26.14.19.35.262492@soul.net> <Y%wqa.4238$lf3.720331@cletus.bright.net>
Message-ID: <20030426102142.B29986@tibia.amotlpaa.bogus>

Quoth George Kinney:
> "lost" <lost at soul.net> wrote in message
> news:pan.2003.04.26.14.19.35.262492 at soul.net...
  [...]
> > from glob import glob
> > file_list = glob.glob('/var/www/html/') for each_file in file_list:
  [...]
> > AttributeError: 'function' object has no attribute 'glob'
  [...]
> Sorry, I should have pointed out that glob does wild-card expansion (not
> just a dir list).

True, but not related to lost's problem above, as a careful
reading of the exception message will show.

Either of
    import glob
    file_list = glob.glob('whatever')
or
    from glob import glob
    file_list = glob('whatever')
will send the AttributeError packing.

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From rmunn at pobox.com  Mon Apr 14 10:22:25 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 14 Apr 2003 14:22:25 GMT
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050201322.13220.python-list@python.org> <slrnb9h7u4.it3.rmunn@localhost.localdomain> <3e9a857b$1@news.012.net.il>
Message-ID: <slrnb9ksaa.1q1.rmunn@localhost.localdomain>

Avner Ben <avner at skilldesign.com> wrote:
> "Robin Munn" <rmunn at pobox.com> wrote in message
> news:slrnb9h7u4.it3.rmunn at localhost.localdomain...
>> Skip Montanaro <skip at pobox.com> wrote:
>> > Recent versions of Python (2.2 and later) allow this:
>> >
>> >     file = file("somefile")
>> >     for line in file:
>> >         process(line)
>>
>> Ewww! Don't *do* that! You just shadowed the "file" builtin. Use a
>> different name for your file object -- I prefer "input_file":
>>
>>     input_file = file("somefile")
>>     for line in input_file:
>>         process(line)
> 
> I have tried this shortcut and it works! (This gets rid of naming problem
> for good.)

I must have been dealing with too much spam recently: when I saw "I have
tried this and it works," I thought the next thing I was going to see
would be instructions for sending $5 to these five addresses... :-)

> 
> for line in file('/dir/filename.ext'):
>          print line
> 
> The only problem with this solution is that it leaves the timing of closing
> the file undefined. But then, neither of you guys have closed the file
> explicitly either!

You're right, I didn't close the file explicitly. But explicit is better
than implicit, so IMHO it's better to give a name to the file object so
that you can close it explicitly:

    input_file = file("somefile")
    for line in input_file:
        process(line)
    input_file.close()

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From danb_83 at yahoo.com  Sun Apr 27 16:39:44 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 27 Apr 2003 13:39:44 -0700
Subject: In which versions is list.sort stable?
References: <slrnbalhft.bc1.mlh@furu.idi.ntnu.no> <yyAqa.8330$3M4.231032@news1.tin.it> <b8en69$263$1@panix2.panix.com> <mailman.1051467524.10087.python-list@python.org>
Message-ID: <ad052e5c.0304271239.1cd0cd8f@posting.google.com>

Beni Cherniavsky <cben at techunix.technion.ac.il> wrote in message news:<mailman.1051467524.10087.python-list at python.org>...

> Why don't we just grow a `list.stablesort` method that can be an alias
> to `list.sort` currently but is guaranteed to be some stable sort
> forever?  Explicit is better than implicit, so in the rare case that
> you do rely on it, why not spell it out?

Or, as a temporary workaround:

# non-in-place sort
def stablesort(lst):
   # insure stable sorting by adding indices to the list to sort
   lst = [(lst[i], i) for i in xrange(len(lst))]
   lst.sort()
   # remove the indices
   return [t[0] for t in lst]

# in-place sort
def stablesort(lst):
   for i in xrange(len(lst)):
      lst[i] = (lst[i], i)
   lst.sort()
   for i in xrange(len(lst)):
      lst[i] = lst[i][0]



From tismer at tismer.com  Mon Apr 21 16:32:01 2003
From: tismer at tismer.com (Christian Tismer)
Date: Mon, 21 Apr 2003 22:32:01 +0200
Subject: Teaching : Python, Scheme, Java...
In-Reply-To: <PMuna.38511$T34.1097150@news2.tin.it>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr> <PMuna.38511$T34.1097150@news2.tin.it>
Message-ID: <3EA45541.9030405@tismer.com>

Alex Martelli wrote:
...

> Yeah, one spot where implementation and pragmatical issues have
> taken priority over didactical and theoretical ones.  I suspect you
> could use _Stackless_ Python and free yourself from the recursion
> limit, if you could accept the limitation of running on Intel and
> Intel-compatible CPU's only, but I don't know it for a fact.

FYI, Stackless 3.0 is almost ready. It re-invents the old principles
of Stackless 1.0, which means, it can avoid 80 percent of all
recursive calls without any hardware dependant code.
In a couple of hours, I will publish this version.
It will compile with hardware stack switching on all supported
platforms. On other platforms, it will at least try to avoid
recursive interpreter calls as much as it can.
With the latter version, you can do unlimited recursions,
unless you are usin deep recursive calls of special __ methods,
which are not supported by the software-only version, yet.

ciao - chris
-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From Olivier.POYEN at clf-dexia.com  Wed Apr 30 09:04:11 2003
From: Olivier.POYEN at clf-dexia.com (POYEN OP Olivier (DCL))
Date: Wed, 30 Apr 2003 15:04:11 +0200
Subject: Visual Basic - Python compare
Message-ID: <8963D6370B323E4AA3241200F0EAF7318C0DEE@FCEXVEXM002.dcl.int.dexwired.net>


> In terms of Python IDEs, most are code/debug only, no GUI.  In the
> wxPython world, BOA Constructor is getting there but still too much of
> a moving target to my mind, and you have to like the Delphi model.
> PythonCard may be promising but is also early on.  wxDesigner is nice,
> but not integrated into a more general IDE.  For Qt, I was initially
> excited about BlackAdder on the commercial front because of it's
> inclusion of a Python Qt and designer at reasonable cost, but it feels
> like a dead project to me at this point (still beta after how long?).
> For Tk, Pythonworks is interesting, but I just don't prefer the look
> of Tk of under Windows.
> 

I find myself quite happy with the pygtk/glade combo. 
I'm much more efficient with glade than with VB building tool. Absolut placement versus auto-sizer....
Anyway, this combo is 
1) working good
2) free as in free beer
3) portable to any place than can get gtk and python AFAIK. 
4) Has a nice look, on all platfomr. You can even get a theme for creating "fake" window widget.


HTH, 

OPQ
-------------- next part --------------
-----------------------------------------------------------------

Ce message est confidentiel ; son contenu ne represente en aucun

cas un engagement de la part de Dexia Credit Local ou de Dexia

CLF Banque. Toute publication, utilisation ou diffusion, meme

partielle, doit etre autorisee prealablement par l'emetteur. Si

vous n'etes pas destinataire de ce message, merci d'en avertir

immediatement l'expediteur.



This message is confidential ; its contents do not constitute a

commitment by Dexia Credit Local or Dexia CLF Banque. Any

unauthorised disclosure, use or dissemination, either whole or

partial, is prohibited. If you are not the intended recipient of

the message, please notify the sender immediately.

-----------------------------------------------------------------

Consultez notre site internet www.dexia-clf.fr

La cle d'une gestion optimisee du secteur public local

-----------------------------------------------------------------

From peter at engcorp.com  Sat Apr 26 13:23:14 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sat, 26 Apr 2003 13:23:14 -0400
Subject: GUI testing with python  (SendKeys)
References: <b8c6c9$p2m$0@198.102.102.163> <b8ecgs$47i$1@sirius.dur.ac.uk>
Message-ID: <3EAAC082.1DD45058@engcorp.com>

Christopher Saunter wrote:
> 
>   You might want to check out Girder - http://www.girder.nl
> It offers more robust and flexible ways of remote controlling Win32 GUI
> stuff than sendkeys.  There is a very noddy Python client for talking to
> Girder at http://www.dur.ac.uk/christopher.saunter/soft.html

"Noddy"?  Please translate. :-)

-Peter



From aleax at aleax.it  Fri Apr 25 06:59:13 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 10:59:13 GMT
Subject: string to variable name
References: <b89qi8$2oij$1@agate.berkeley.edu> <yfsist3wij7.fsf@black132.ex.ac.uk>
Message-ID: <5y8qa.3561$3M4.84356@news1.tin.it>

Alexander Schmolck wrote:

>>   FILE = open("<mytextfile");
>>   while (<FILE>){
>>     /^(\w+)\s+(\w+)\s+(\w+)$/;
>>     inst.$2 = $1;
>>   }
>> }
> 
> 
>> Is there any way to (easily) achieve what I want to do in python?
> 
> Sure: you can do it even shorter and almost as unreadably with (UNTESTED):
> 
> for line in file("mytextfile"):
>     inst.__dict__.update(dict([line.split()[:2]]))

<shudder>:-).  Not to mention a likely key/value exchange.  Since lists 
can't be general-sliced in Python 2.2 (only in 2.3), in the spirit of this
solution, I propose using, as update's argument [ALSO UNTESTED]:
    dict([(lambda x,y: y,x)(*(line.split()[:-1]))])

Now THAT should warm the cockles of any Perlist's heart!


Now, joking apart...:

> a better way would be to use ``setattr(inst, key, value)``

Absolutely better -- for completeness, the loop should then be:

    for line in file('mytextfile'):
        value, key, junk = line.split()
        setattr(inst, key, value)


Alex




From sismex01 at hebmex.com  Tue Apr 29 12:16:03 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Tue, 29 Apr 2003 11:16:03 -0500
Subject: Problem subclassing tuple
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F062330489D3F8@mail.hebmex.com>

> 
> I'm really quite interested in why this doesn't work. I *thought* I'd
> grasped the concept of the new style classes but this seems 
> to demonstrate
> that I haven't!
> 
> Thanks for the response.
> 
> John Wilson
> The Wilson Partnership
> http://www.wilson.co.uk
> 

Most prob'ly you've already grasped it; the problem with
the current implementation is that many internal classes
don't "allow" subclassing, because the interpreter's
internal mechanics access the C code for those classes
(and subclasses), so your Python code doesn't get called.

Bummer, really. :-/

-gus




From b.maryniuk at forbis.lt  Fri Apr 25 14:31:42 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Fri, 25 Apr 2003 20:31:42 +0200
Subject: Pickling objects...
Message-ID: <200304252031.42673.b.maryniuk@forbis.lt>

Hello.

Well, I would like to anybody help me to serialize an instance 
from 4DOM HtmlLib. Somebody know where (and HOW) to push 
__setstate__ / __getstate__ methods? 8-| I tried a lot of different
code, but _always_ I got the same traceback (below).

Yes, I know, I need __setstate__ stuff, but here I dumb where to
push these... Or it is impossible to serialize it?.. Simple classes serialises
without the problem. But this.......

This pack's OK:
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
#!/usr/bin/python
from xml.dom.ext.reader import HtmlLib as HL
import cPickle

class docContainer:
    def __init__(self, doc):
        self.doc = doc

    def __getstate__(self):
        return self.__dict__.copy()

    def __setstate__(self, dict):
        self.__dict__.update(dict)

reader = HL.Reader()
doc = docContainer(reader.fromString(open('foo.html').read(), None, None))

packIt = open('packed', 'w')
p = cPickle.Pickler(packIt, 1)
p.dump(doc)
packIt.close()
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 

This should read, but raises an exception (below):
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
#!/usr/bin/python
import cPickle

class docContainer:
    pass

packedObj = open('packed.file', 'r')
doc = cPickle.load(packedObj)
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 

...and the traceback:
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
Traceback (most recent call last):
  File "./unpkl.py", line 10, in ?
    doc = cPickle.load(packedObj)
TypeError: ('__init__() takes exactly 1 argument (2 given)', <class xml.dom.html.HTMLDocument.HTMLDocument at 0x81dd47c>, (None,))
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 

Can anybody push me where to find and WHAT to subclass, if I need it...?
Thank you.

-- 
Regards, Bogdan

Only great masters of style can succeed in being obtuse.
                -- Oscar Wilde





From exarkun at intarweb.us  Fri Apr 11 02:30:25 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 11 Apr 2003 02:30:25 -0400
Subject: Macros in Python?
In-Reply-To: <b75m9b$kr$07$1@news.t-online.com>
References: <b71qde$q71$07$1@news.t-online.com> <slrnb9b01m.108.do-not-spam-ben.hutchings@tin.bwsint.com> <b74bem$oi3$02$1@news.t-online.com> <m2znmymbaj.fsf@yahoo.com> <b75m9b$kr$07$1@news.t-online.com>
Message-ID: <20030411063025.GA13517@meson.dyndns.org>

On Fri, Apr 11, 2003 at 08:12:13AM +0200, Dominic wrote:
> [snip]
> 
> def make_adder(value):
>    def add(a,value):
>       return a+value
>    return lambda a: value+x
> 
> add3=make_adder(3)
> print add(3)
> 
> I haven't tried it, but should work.

  Close.

    def make_adder(value):
        return lambda a: value + a

    add3 = make_adder(3)
    print add3(3)

  Jp

-- 
"Pascal is Pascal is Pascal is dog meat."
                -- M. Devine and P. Larson, Computer Science 340
-- 
 up 22 days, 3:01, 4 users, load average: 2.26, 2.22, 2.11




From gh at ghaering.de  Tue Apr 22 16:07:45 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 22 Apr 2003 22:07:45 +0200
Subject: Python 2.2.2 and 2.3a2 compile problems on Mingw32
In-Reply-To: <87he8q4993.fsf@pobox.com>
References: <mailman.1051001914.4729.python-list@python.org>
	<87he8q4993.fsf@pobox.com>
Message-ID: <3EA5A111.6000908@ghaering.de>

John J. Lee wrote:
> "Andreas KLostermann" <bluetigger at web.de> writes:
> 
> 
>>I wanted to compile Python on Mingw32. That's a GNU C++ Compiler for
>>Windows and it can configure...  The "make" process begins
>>promissing, but fails at "posixmodule.c"...
>>
>>This Problem is the same for 2.2.2 and 2.3a2.
>>Can someone help me?
> 
> [...]
> 
> Sorry if this is a stupid question, but are you using the cygwin port?

Apparently Andreas was trying to compile Python using the *native* win32 
GNU toolset, aka mingw (minimalist GNU for Windows). This isn't 
currently possible, though I know of several people who have expressed 
interest and/or are currently working on this (that includes me).

> Or is the cygwin port merged into Python proper, now?

It is. Building Python sources with Cygwin is something that works out 
of the box.

-- Gerhard




From Nisamudeen at BNS-solutions.com  Thu Apr 17 08:12:52 2003
From: Nisamudeen at BNS-solutions.com (Nisamudeen)
Date: Thu, 17 Apr 2003 17:42:52 +0530
Subject: For Controlling Python Threads switch from a C/C++ application 
Message-ID: <88191C7199A2164C91BD21D49E4BB20906BDA8@dc-bnsindia.BNS-Software.de>

Hi ,

i have embeded a python(Multithreaded for executing different script file in
each thread ) in a c++ application,  Its working properly but the problem is
, 
I want to control the thread switching from my thread (main thread ) .. Is
it possible ?? help me .. 

Case : 

/// class for the multiple thread state 
class PythonInterpreter
{
public:
	PythonInterpreter()
		: _threadState(0)
	{
		PyEval_AcquireLock();
		_threadState = Py_NewInterpreter();
		assert(_threadState);
		PyThreadState_Swap(_threadState);

	}
	~PythonInterpreter()
	{
		Py_EndInterpreter(_threadState);
		PyEval_ReleaseLock();
	}
public:
	PyThreadState* _threadState;
}; 

// Thread  //////////////////
DWORD _stdcall CheckKey( void *dummy )   
{
	char * filename = (char*) dummy ; 

	PythonInterpreter opython ;  

	pfobj = (PyFileObject * )PyFile_FromString( filename, "r+") ; 

	PyRun_SimpleFile( pfobj->f_fp   , filename  ) ;
	return 0 ;
}



int main(int argc, char *argv[])
{
	char filename1[50],filename2[50] ; 
	DWORD dwID ; 
	HANDLE hHand ,hHand1 ;  

	Py_Initialize();
	PyEval_InitThreads(); // initalize threading
	PyThreadState* state = PyEval_SaveThread(); // release lock 


	strcpy(filename1 ,"c:\\ONE.py"); 
	strcpy(filename2 ,"c:\\TWO.py"); 
 
	

	hHand1 =  CreateThread( NULL , 0, CheckKey , (void *)filename1 , 0,
&dwID ) ;
	printf("Started One ") ;

	Sleep(1000) ; 
	hHand =  CreateThread( NULL , 0, CheckKey , (void *)filename2 , 0,
&dwID ) ;

	Sleep(10000) ;
	Py_Exit(0); 
	Py_Finalize();
	PyEval_RestoreThread(state);
	return 0;
}


its actully working well , its also switching between thread (python taking
care of switching), but what i want is from the main() thread , i want to
suspend or resume each of the thread based on my requrement , for eg:- i
want to execute #1 thread for 100 mill sec and after that , i want to switch
to #2 for another 100 m sec  ,like wise , is it possible ????



Nisamudeen S





From jolderdissen at ixiacom.com  Tue Apr 29 13:14:58 2003
From: jolderdissen at ixiacom.com (Jan Olderdissen)
Date: Tue, 29 Apr 2003 10:14:58 -0700
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk>
Message-ID: <nhctav4n42nlt8t1h2i0tcqqms1b3r9b6h@4ax.com>

Garrison,

I agree with Peter that you're probably better off taking a closer
look at Python instead of attempting to convert an existing Python
program to C. Python is really easy to learn and remember - especially
if you know C.

That said, it would be an interesting project to implement a just in
time compiler for Python that converts Python byte-code to C and uses
gcc to generate a binary form. There is some moderate performance
enhancement to be had, I suspect.

Jan

On Tue, 29 Apr 2003 02:23:59 +0100, "Garrison Tsang" <hygt2 at cam.ac.uk>
wrote:

>Knowing nothing about Python at the moment, I need to convert a python-code
>into compatible c-code.  I recall that were some program/utility that does
>the transformation for you.




From e_fax_t at hotmail._ZAPME_.com  Tue Apr 29 17:05:21 2003
From: e_fax_t at hotmail._ZAPME_.com (Justin)
Date: Tue, 29 Apr 2003 23:05:21 +0200
Subject: mxODBC puzzle
References: <mailman.1051646405.11551.python-list@python.org>
Message-ID: <3eaee997$0$6533$afc38c87@sisyphus.news.be.easynet.net>

> The DB is Access97 via DSN (which is *not* read only), and query seems
> to return OK (1 record affected), but the record never gets updated.


Could it be that you need to commit your changes?






From aahz at pythoncraft.com  Mon Apr 21 09:28:54 2003
From: aahz at pythoncraft.com (Aahz)
Date: 21 Apr 2003 09:28:54 -0400
Subject: __setitem__ puzzle
References: <b80q7c$cmd$1@news.hccnet.nl>
Message-ID: <b80rmm$2c3$1@panix1.panix.com>

In article <b80q7c$cmd$1 at news.hccnet.nl>,
Anton Vredegoor <anton at vredegoor.doge.nl> wrote:
>
>I have subclassed the builtin list type to accept tuples for indexing.
>I am just experimenting with the code to see what I can do with it, so
>asking for what the code will be used for in order to provide a better
>way to do it will be answered with 'ni' :-)

"If you don't know what your program is supposed to do, you'd better not
start writing it."  --Dijkstra

>There's a problem with __setitem__ however. What's the proper way to
>write this class?

While I understand your point about expirimenting, I can't give you any
useful advice unless I know what your code is intended to do, and I doubt
many people on c.l.py will do much better.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From mal at lemburg.com  Wed Apr 23 17:21:51 2003
From: mal at lemburg.com (M.-A. Lemburg)
Date: Wed, 23 Apr 2003 23:21:51 +0200
Subject: ANN: eGenix mxODBC Zope Database Adapter, Version 1.0.4
Message-ID: <3EA703EF.20900@lemburg.com>

________________________________________________________________________
ANNOUNCEMENT

                              EGENIX.COM

                      mxODBC Zope Database Adapter

                             Version 1.0.4

                    Available for Zope 2.3 - 2.6 on
                       Windows, Linux and Solaris

________________________________________________________________________
INTRODUCTION

The eGenix mxODBC Zope Database Adapter (Zope DA) allows you to easily
connect your Zope installation to just about any database backend on
the market today, giving you the reliability of the commercially
supported eGenix.com product mxODBC and the flexibility of the ODBC
standard as middle-tier architecture.

Unlike Zope's ZODBC Zope DA, the mxODBC Zope DA works on Windows
XP/NT/2000/98/95, Linux and Solaris using the same interface on all
platforms.

The mxODBC Zope DA implements thread-safe connection pooling and
multiple physical connects per logical Zope connection. You can safely
run Z SQL Methods in parallel, achieving a much better performance
than ZODBC Zope DA or similar Zope database adapters under heavy load.

This makes it ideal for deployment in Zope Clusters and Zope hosting
environments where stability and high performance are a top priority.

________________________________________________________________________
FEATURES

* Zope Level 3 Database Adapter: the mxODBC Zope DA is fully
   multi-threaded and can handle multiple connections to multiple
   databases.

* Fully compatible to Z SQL Methods.

* Drop-in compatible to the ZODBC DA: the mxODBC Zope DA provides the
   same interfaces as Zope's ZODBC DA to allow a smooth upgrade path
   from this simplistic adapater to the high performance mxODBC Zope
   DA.

* Fully compatible to the Znolk SQL Wizard Product and other similar
   products relying on the common database schema access methods
   .tables() and .columns().

* Connection Pooling: physical database connections are pooled and
   kept open, to reduce the connection overhead to a minimum.  This is
   especially important for high latency database connections and ones
   like Oracle which take a considerable amount of time to setup

* Parallel Execution of Queries on a single logical connection: the
   mxODBC Zope DA can manage any number of physical connections on a
   single logical connection. This enables running truly parallel Z SQL
   Method queries -- a feature not available in other Zope DAs.

* Robust Mode of Operation: connections which have timed out or go
   away due to network problems are automatically reconnected.

* Cross-platform Connection Objects: The Zope DA will automatically
   choose the right platform specific ODBC manager for you.

* Per Connection Adjustable ODBC Interface: mxODBC comes with many
   different subpackages to choose from on Unix. The Zope DA allows you
   to select these subpackages on a per-connection basis.

* Per Connection Error Handling: you can tell each connection whether
   it should report ODBC warnings or not; furthermore all warnings and
   errors are made available as list .messages on the DatabaseConnection
   object.

* Transaction safe automatic reconnect: when the DA finds that a
   connection has timed out, it automatically tries a reconnect
   and replays the transaction on the connection (unlike other DAs
   which break the transaction scheme by doing a reconnect without
   replay).

* Built-in Schema Cache: this results in improved performance under
   heavy load.

* Database Schema Access: all ODBC catalog methods are made available
   for much better database schema inquiry. The catalog methods allow
   building generic database interrogation or manipulation tools and
   facilitates writing database independent Zope products.

________________________________________________________________________
NEWS

We have had a lot of positive user feedback about the mxODBC Zope
Database Adapter 1.0.0.

* Some users have reported problems with certain ODBC drivers which
   required adding work-arounds and special options to the adapter. The
   version 1.0.4 includes a whole set of connection options which make
   the mxODBC Zope DA even more compatible to ODBC drivers in the
   field.

* We also added a few usability enhancements, e.g. the default for
   creating connections is now to create closed connections. This makes
   it much easier setting up and configuring the connection objects.

* Due to popular demand we put a lot of work into getting the mxODBC
   Zope DA to work with the 0.61 release of the FreeTDS ODBC driver.
   This driver is typically used when connecting from Linux to MS SQL
   Server.

* Connections to non-transactional data sources are now also possible
   using a special connection option. This allows direct read-only
   access to e.g. Access or Excel files on Windows. It also makes
   connections to MySQL 3.x possible.

* Because the mxODBC Zope DA supports multiple result sets, we added
   an option to always fetch the last available result set.  This is
   sometimes needed when moving from ZODBC DA to mxODBC Zope DA if you
   are working with stored procedures.

* Last not least, the mxODBC Zope DA is now also available for the
   Solaris platform.

In short: mxODBC Zope DA is moving to become the number one solution
for integrating relational databases with Zope applications.

________________________________________________________________________
UPGRADING

If you have already bought mxODBC Zope DA licenses, you can use these
license for the updated version as well. There is no need to buy new
licenses. The same is true for evaluation license users.

________________________________________________________________________
MORE INFORMATION

For more information on the mxODBC Zope DA, licensing and download
instructions, please visit our web-site:

     http://www.egenix.com/

________________________________________________________________________

Thank you,
-- 
Marc-Andre Lemburg
eGenix.com

Professional Python Software directly from the Source  (#1, Apr 23 2003)
>>> Python/Zope Products & Consulting ...         http://www.egenix.com/
>>> mxODBC, mxDateTime, mxTextTools ...        http://python.egenix.com/
________________________________________________________________________
EuroPython 2003, Charleroi, Belgium:                        62 days left







From ianb at colorstudy.com  Sun Apr 20 20:28:01 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 20 Apr 2003 19:28:01 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <84n0ikommx.fsf@boost-consulting.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	 <6ee58e07.0304061758.35070e7c@posting.google.com>
	 <a3b5b608.0304080302.c26ad68@posting.google.com>
	 <af6b5bdd.0304181001.558da6f@posting.google.com>
	 <84of317cd6.fsf@boost-consulting.com>
	 <mailman.1050870230.2976.python-list@python.org>
	 <84n0ikommx.fsf@boost-consulting.com>
Message-ID: <1050884881.698.12.camel@lothlorien>

On Sun, 2003-04-20 at 17:00, David Abrahams wrote:
> I don't understand; it already *is* plain Python.  While Donovan
> Preston and I were designing the syntax, his friend Bob was sitting on
> the couch coding it up.

Hmmm... I guess so.  I was thinking of lists following an object, but it
actualy gets parsed as a __getitem__ with a tuple as an argument.

> > The use of []'s isn't really necessary, and actually the expressions
> > could maybe be eval'd as well...
> 
> You've lost me here.  Lots of the "syntax" choices aren't based on
> what's neccessary, but on what's clear and expressive.  As far as
> "eval" is concerned, well, I'm not sure if you realize this but those
> if_ and for_each elements are lazily evaluated.  The expression
> creates a callable object, "template", which generates new XML.

So it's not actually a Python program, it's Python syntax that gets
evaluated by your own evaluator...?  I was trying to see if it could be
an actual Python program, which would construct an object you could use
later to handle the XML.

  Ian






From exarkun at intarweb.us  Mon Apr  7 01:38:51 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 7 Apr 2003 01:38:51 -0400
Subject: Handling slices
In-Reply-To: <yfsistrkofa.fsf@black132.ex.ac.uk>
References: <mailman.1049676910.30971.python-list@python.org> <yfsistrkofa.fsf@black132.ex.ac.uk>
Message-ID: <20030407053851.GA27697@meson.dyndns.org>

On Mon, Apr 07, 2003 at 03:54:17AM +0100, Alexander Schmolck wrote:
> [snip]
> 
> Here is a simple example:
> 
> class SliceMaker:
>     """Convinience class to make slices."""
>     def __getitem__(self, a): return a
> 

  It may be worth noting that this is essentially the slice() builtin.

  Jp

-- 
It is practically impossible to teach good programming style to
students that have had prior exposure to BASIC: as potential
programmers they are mentally mutilated beyond hope of
regeneration.        -- Dijkstra
-- 
 up 18 days, 1:01, 0 users, load average: 0.00, 0.00, 0.00




From mis6 at pitt.edu  Fri Apr 25 08:13:31 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 25 Apr 2003 05:13:31 -0700
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org>
Message-ID: <2259b0e2.0304250413.4be8ee45@posting.google.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1051212998.14005.python-list at python.org>...
> <snip>
> we get:
>
> Traceback (most recent call last):
>  File "M:\python\cpoint.py", line 36, in ?
>    class CountedPoint(Counted, Point):
>  File "M:\python\cpoint.py", line 6, in   new  
>    return type.  new  (cls, name, bases, dict)
> TypeError: metatype conflict among bases

>not exactly clear  what  the conflict is

The meta-type conflict is rather annoying, indeed. According to the
metaclass book, the language should generate the correct metaclass
that solves the conflict automatically, without burden for the
programmer. Python is not that magic (yet) therefore you have to
solve the conflict yourself.

I use the following solution:

  def _generatemetaclass(bases,metas):
      "Internal function called by child"

      if metas==(type,): # trivial metaclass
          metabases=(); metaname="_"
      else: # non-trivial metaclass
          metabases=metas
          metaname="_"+''.join([m.__name__ for m in metas])
      trivial=lambda m: m in metabases or m is type

      for b in bases:
          meta_b=type(b)
          if not trivial(meta_b):
              metabases+=(meta_b,)
              metaname+=meta_b.__name__

      if not metabases: # trivial metabase
          return type 
      elif len(metabases)==1: # single metabase
          return metabases[0]
      else: # multiple metabases
          return type(metaname,metabases,{}) # creates a new metaclass
          #shifting the possible conflict to meta-metaclasses

  def child(*bases,**options):
      """Class factory avoiding metatype conflicts: if the base classes have 
      metaclasses conflicting within themselves or with the given metaclass, 
      it automatically generates a compatible metaclass and instantiate the 
      child class from it. The recognized keywords in the option dictionary
      are name, dic and meta."""
      name=options.get('name',''.join([b.__name__ for b in bases])+'_')
      dic=options.get('dic',{})
      metas=options.get('metas',(type,))
      return _generatemetaclass(bases,metas)(name,bases,dic)

Here there is an example of usage:
    
  >>> class M_A(type): pass
  >>> class M_B(type): pass
  >>> A=M_A('A',(),{})
  >>> B=M_B('B',(),{})

M_A is A's metaclass, M_B is B's metaclass. I cannot do
  
  >>> class C(A,B): pass #error
  Traceback (most recent call last):
    File "<stdin>", line 1, in ?
  TypeError: metatype conflict among base

but I can do

  >>> C=child(A,B,name='C')
  >>> C
  <class 'oopp.C'>
  >>> type(C) # automatically generated metaclass
  <class 'oopp._M_AM_B'>


HTH,,
                                                Michele



From jack at performancedrivers.com  Fri Apr 18 17:19:15 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Fri, 18 Apr 2003 17:19:15 -0400
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <HpZna.64494$BQi.43307@news04.bloor.is.net.cable.rogers.com>; from guy@obstruction-no-spam.com on Fri, Apr 18, 2003 at 08:41:11PM +0000
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <7x4r4vy2gl.fsf@ruckus.brouhaha.com> <HpZna.64494$BQi.43307@news04.bloor.is.net.cable.rogers.com>
Message-ID: <20030418171915.C1553@localhost.localdomain>

> > import operator
> > reduce (operator.add, [y[1] for y in x])
> 
> 
> Aha!  That's what I was looking for.  Thanks...
> 

Actually he said
  reduce(operator.add, [y[1] for y in x], 0)
                                        ^^^^
reduce without the third argument will raise an exception if x is empty,
so it is best to get in the habit of including one.

-jack




From skip at pobox.com  Sat Apr 12 19:00:21 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sat, 12 Apr 2003 18:00:21 -0500
Subject: Speed of string += string
In-Reply-To: <Zw0ma.4690$883.4089@nwrdny02.gnilink.net>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
        <65pjlhdx.fsf@morpheus.demon.co.uk>
        <Zw0ma.4690$883.4089@nwrdny02.gnilink.net>
Message-ID: <16024.39557.236558.700373@montanaro.dyndns.org>

    Carl> Surprisingly, += doesn't appear to be a major time waster for
    Carl> small (meaning about 10000 bytes) strings.  I'm guessing it's
    Carl> because the malloc my system (Debian Linux) is optimized for small
    Carl> memory allocations.

Actually, if you're using 2.2.n or later, it's Python's small object
allocator (see Objects/obmalloc.c), which is providing most of the speedup.
I'm not sure if n is 0, 1, or 2.

Skip





From gk_2345 at yahoo.com  Wed Apr 16 23:14:58 2003
From: gk_2345 at yahoo.com (George Kinney)
Date: Wed, 16 Apr 2003 23:14:58 -0400
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it>
Message-ID: <o1pna.3380$lf3.595806@cletus.bright.net>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:Wrena.33096$T34.876856 at news2.tin.it...
> need to understand dotNET to profit from this book".  I don't know if
> there are substantial slices of programmers still stuck on VB5, VB4, ...,
> and so on backwards, that might be scared off by the "6" in the title.

Believe it or not, I am the not so proud possessor of an application written
in a mish-mash of VB5 and dBase.  Its creator claims that he can't upgrade
to VB6, because VB6 is broken.  Yes, it is 8 years old and the same
half-finished
POS it was in '95, but he just won't budge.

The beautiful thing about it is his own insistance on antiquated tools gave
me the
argument I needed convince management to let me re-implement this particular
package...in python. :)

What's the fastest way to convert a VB drone to python in my opinion?
Make them use python for a project, big or small,  they'll never go back.








From tebeka at cs.bgu.ac.il  Mon Apr  7 07:00:23 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 7 Apr 2003 04:00:23 -0700
Subject: PYTHON EDITOR!!
References: <mailman.1049690233.24009.python-list@python.org>
Message-ID: <33803989.0304070300.661d1d9@posting.google.com>

Hello Nanthan,

> im Nanthan, a postgraduate student from Malaysia.im very new to 
> python language but im already getting use of it.
Have fun.

> Currently,i 
> have been working in windows, but my development actually is on 
> Linux, so can u suggest me the best editor(i mean use friendly 
> and graphical,i dont want command line editor)for python?
http://www.python.org/cgi-bin/moinmoin/PythonEditors

Personally I use Vim (http://vim.sf.net) but it far from being graphical.

HTH.
Miki



From pobrien at orbtech.com  Thu Apr 17 12:17:40 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 17 Apr 2003 11:17:40 -0500
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050394215.20671.python-list@python.org> <b7i26g$17t5m$1@ID-169208.news.dfncis.de> <m24r4yh1l8.fsf@orbtech.com> <b7l1ov$22o4q$1@ID-169208.news.dfncis.de>
Message-ID: <m2brz514kb.fsf@orbtech.com>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> writes:

> Patrick K. O'Brien wrote:
> > I suppose one could bind attributes (current frame, current buffer,
> > point) to a module and then import that module everywhere that needed
> > access to these attributes
> 
> That's no better, because they're still global -- there's
> still only *one* of each of them, whereas there really
> needs to be different ones in different contexts.
> 
> A function which needs to operate on a frame should be
> passed a frame. Some buffer will be displayed in that
> frame: frame.buffer. That buffer will have a point:
> frame.buffer.point. Etcetera.

That's easy enough for functions/methods/commands that are part of the
base package.  What I was thinking about were functions/commands that
were written by users to extend the base package.  I haven't figured
out a good way to determine the context they need and be able to pass
that object to them.  For example, let's say someone wanted to create
a new command that operated on the current buffer.  Assuming I had the
mechanism in place to bind keystrokes to a user-created command, how
would I know what parameters that command expected?

Does anyone know of a Python program of any kind that comes close to
having this ability to be extended by users?

Now that I think about it, my dispatcher module does poke into
handlers/receivers to see what parameters they are expecting (based on
parameter name) and only calls the handler with the parameters it
wants.  So I suppose I could do something similar, using a naming
convention for all the objects that are available as parameters, which
should be a reasonably small number.

Here is the relevant code from dispatcher.py:

def _call(receiver, **kwds):
    """Call receiver with only arguments it can accept."""
    if hasattr(receiver, '__call__') and \
       (hasattr(receiver.__call__, 'im_func') or hasattr(receiver.__call__, 'im_code')):
        # receiver is a class instance; assume it is callable.
        # Reassign receiver to the actual method that will be called.
        receiver = receiver.__call__
    if hasattr(receiver, 'im_func'):
        # receiver is a method. Drop the first argument, usually 'self'.
        fc = receiver.im_func.func_code
        acceptable = fc.co_varnames[1:fc.co_argcount]
    elif hasattr(receiver, 'func_code'):
        # receiver is a function.
        fc = receiver.func_code
        acceptable = fc.co_varnames[0:fc.co_argcount]
    else:
        raise DispatcherError, 'Unknown receiver %s of type %s' % (receiver, type(receiver))
    if not (fc.co_flags & 8):
        # fc does not have a **kwds type parameter, therefore 
        # remove unacceptable arguments.
        for arg in kwds.keys():
            if arg not in acceptable:
                del kwds[arg]
    return receiver(**kwds)

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From michael at stroeder.com  Fri Apr 11 14:56:17 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Fri, 11 Apr 2003 20:56:17 +0200
Subject: popen(), sendmail: Success?
In-Reply-To: <b7703o$ejq$1@news.island.liu.se>
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de> <mailman.1050021261.25010.python-list@python.org> <b7703o$ejq$1@news.island.liu.se>
Message-ID: <hnfjm-n9t.ln1@nb2.stroeder.com>

Thomas Bellman wrote:
> Chuck Swiger <cswiger at mac.com> wrote:
> 
>>If you plan on sending any significant amount of mail, it is recommended 
>>  that your code talk SMTP to port 25 and pipeline all of the messages 
>>you need to send in one continuous stream.  If you fork a sendmail 
>>subprocess for each message you send, that will create a wastefully huge 
>>amount of system load.
> 
> If you write your application to talk SMTP directly, instead of
> using the proper Unix API for sending mail (i.e, calling
> /usr/lib/sendmail), you have a number of problems.

1. I wouldn't consider calling /usr/lib/sendmail being an API...
2. I think Chuck talked about using a central SMTP relay.

> First, which machine should you talk SMTP with?

The central corporate SMTP relay, e.g. the one the desktop clients also use.

> If the system changes, and another mail
> server should be used, the sysadmin might not remember, or know,
> that that particular application must be reconfigured.
 > It
 > becomes even worse if it is an application where each user on the
 > machine has their own configuration, like Netscape or Mozilla...

That's why you have a well-known DNS alias for your SMTP relay.

Additionally an application admin should be responsible for checking the 
error logs of the application.

> I turned the mail server at work off for more than 12 hours
> waiting for a patch from RedHat

Depending on the business processes using e-mail a down-time of 12h might be 
more than unacceptable. That's your local policy.

> Sendmail, Postfix
> and Qmail are written to take care of those; you are unlikely to
> do a better job when writing an application whose primary purpose
> *isn't* being an MTA.

But for several reasons calling popen() is a nightmare.

> Any application that talks SMTP on their own
> instead of calling /usr/lib/sendmail to do it, is considered
> broken by us...

That's your local policy.

Ciao, Michael.




From parz at shaw.SpamBucket.ca  Thu Apr 17 00:27:57 2003
From: parz at shaw.SpamBucket.ca (Parzival Herzog)
Date: Thu, 17 Apr 2003 04:27:57 GMT
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com>
Message-ID: <h3qna.143255$vs.16245277@news3.calgary.shaw.ca>

"Aahz" <aahz at pythoncraft.com> wrote in message news:b7kqr8$jsq$1 at panix2.panix.com...
> In article <Wrena.33096$T34.876856 at news2.tin.it>,
> Alex Martelli  <aleax at aleax.it> wrote:
> >Steve Holden wrote:
> >>
> >> You'd do better to call it "Python for VB programmers", expecially since
> >> the change to VB6 is so radical that migrating to a new and more sensible
> >> language might be an attractive option to some!
> >

Maybe when you're done that, you could do "Python for RPG/400"

     H*
     H* RPG/400 VERSION OF THE BOTTLES PROGRAM *
     H*
     FSCREEN  O   F      80            WORKSTN
     C                     MOVE 100       X       30
     C           X         DOWGE0
     C                     EXCPT
     C                     SUB  1         X
     C                     END
     C                     SETON                     LR
     OSCREEN  E
     O                         X          3
     O                                   26 'BOTTLES OF BEER ON THE'
     O                                   31 'WALL,'
     O                         X         36
     O                                   53 'BOTTLES OF BEER'
     O        E
     O                                   22 'TAKE ONE DOWN AND PASS'
     O                                   32 'IT AROUND'

P.S. Can you believe it, at version 400 kicks at the can, they came up with this?

Parzival





From pdaniell at ign.com  Wed Apr 30 17:28:25 2003
From: pdaniell at ign.com (Ponce Dubuqe)
Date: Wed, 30 Apr 2003 14:28:25 -0700
Subject: Python and Java Servlet ISP?
Message-ID: <009901c30f5f$6eca3040$c90aa8c0@PDANIELL>

Does anyone know if there's an Web hosting Provider that supports both Python and Java Servlets/JSPs?

Ponce
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030430/abc7ec14/attachment.html>

From achrist at easystreet.com  Fri Apr 25 18:33:08 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Fri, 25 Apr 2003 15:33:08 -0700
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <vuiqa.3522$%_3.2810696@newssrv26.news.prodigy.com>
Message-ID: <3EA9B7A4.59F9CE04@easystreet.com>

dsavitsk wrote:
> 
> Your requirements suggest to me that html is the best common denominator,
> and it's easier to write.
> 

Yes, it is easier to write.  But I wrote a dead-plain-simple HTML file
and opened it in MS Word (the version that came free on my computer in
late 1996), and it looked just dreadful, with all kinds of visible
markup that made no sense and generally might make you think that
the  program that wrote it was written by someone who was not very
smart.  So, I try to open the same file with Abiword (a relatively
recent version), and it tells me that the file is bogus and refuses to
open it at all.  This is simple HTML that opens fine in Netscape and in
Amaya -- those have no trouble with my HTML (just HTML, BODY, UL, LI,
TABLE, TR, TD, and B tags, IIRC).  So I figure that HTML is not a
preferred path into word-processing programs when it is not known what
version of which word-processing program the user has installed.  I
intend to offer HTML output for those who want it, and I do have that
nut cracked already, but it's not a complete solution.  

So, again, wrt the larger nut, how to write RTF from Python? Any ideas?


Al



From pobrien at orbtech.com  Sun Apr  6 16:28:47 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 15:28:47 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com>
Message-ID: <m2k7e78j5s.fsf@localhost.localdomain>

llothar at web.de (Lothar Scholz) writes:

> pobrien at orbtech.com (Patrick K. O'Brien) wrote in message
> news:<m2u1dcw22s.fsf at localhost.localdomain>...
> > Here is the background.  I'm working on a file editor (called
> > PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> > PyCrust code base.  I've got a basic version working, and it
> > supports editing multiple files and all the usual, basic features.
> > But I want it to be as customizable and extensible as Emacs.  In
> > fact, I want to model it after Emacs, so that elisp code could be
> > rewritten in Python and work with PyAlaMode.  To do that, I've got
> > to support the same primitives as Emacs; expose files, buffers,
> > regions, and windows the same way Emacs does; handle the Emacs
> > keybindings and ability to remap keys, the same way Emacs does,
> > etc.
> 
> Then you have to present the same old and ugly user interface with a
> mini buffer as input line and so on. I don't see any reason at the
> moment for cloning EMACS. You should ask people why they like emacs
> as an editor.

Actually, we don't have to clone everything about the interface.  I'm
using wxPython, so the interface is pretty and more easily configured.
But some people actually like the minibuffer, so we'll need to support
something like it and something less ugly.  Heck, I'm actually running
three different versions of the interface now, and I only wrote the
code this week (minus the almost two year's worth of code in PyCrust).

> There are only a handful of features that are normally the reason to
> keep this as your primary editor. For me its the window handling C-X
> 1, the way to prefix commands with a repeat command, the macro
> recording and a the autoindentation with TAB.

Except everyone will likely disagree about what the handful of
features are.  ;-) In any case, I'm targeting the bare minimum to
start.

> To make clear, i'm using XEMACS 8 hours a day for years now. But
> it's absolutely losing ground compared to modern editors. And i
> don't think that extensibility in the emacs sense is really
> necessary today or a plus, because people want out of the box
> solutions that are specialized to their domain.

What would you like to see in a modern version?

The main reason I started PyAlaMode is that I wanted the same
autocompletion and calltips that I've got in the PyCrust shell, but
while I'm editing a file.  Those features are working now, and are
pretty sweet, if I do say so myself, since they are based on dynamic
namespaces updated whenever you want, one per buffer, with no
conflicts no matter how many buffers you have open, with automatic
reloading of imported modules.

That means if I have two modules open in PyAlaMode, moduleA.py and
moduleB.py, and moduleA.py contains the following import:

    import moduleB

And then I modify and save moduleB.py and switch back to moduleA.py
and update its namespace and type:

    moduleB.

I see a popup list of moduleB's new contents.  Same thing with
calltips.  Every time I change moduleB.py I just have to switch back
to moduleA and update its namespace (one keystroke combo or menu
selection).  Try doing that with Emacs.  ;-)

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From Patrick.Bussi at space.alcatel.fr  Thu Apr 17 14:06:42 2003
From: Patrick.Bussi at space.alcatel.fr (Patrick.Bussi at space.alcatel.fr)
Date: Thu, 17 Apr 2003 20:06:42 +0200
Subject: =?iso-8859-1?Q?R=E9f._:_Downloading_TV_listings_with_urllib?=
Message-ID: <C1256D0B.00665FC2.00@vzmta01.netfr.alcatel.fr>

I suggest you use urllib2, associated with ClientCookie if you need to support
cookie on the client side. They are much more convenient libraries than dealing
with lower level httplib. They also allow you to set headers as you wish.

urllib2 is included in Python
ClientCookie is available at http://wwwsearch.sourceforge.net/ClientCookie/

Hope this help. Patrick.

---
Patrick Bussi
patrick.bussi at space.alcatel.fr

Any opinions expressed are my own and not necessarily those of my Company.




Josh <mlsj at earthlink.net> on 17/04/2003 19:23:34

Pour :    python-list at python.org
cc :   (ccc : Patrick Bussi/ALCATEL-SPACE)
Objet :   Downloading TV listings with urllib



Can someone give me general pointers on how it could be done? I am trying
to go to tvguide.com and download the TV listings for my area. The problem
is how to specify my location and cable provider. I used Proxomitron to
look at the traffic generated while downloading the listings with my
browser. When I used httplib with:

host = 'www.tvguide.com'
pathn = '/Listings/index.asp'

def tvg():
    try:
        h = httplib.HTTPConnection(host)
        h.putrequest('GET', pathn)
        h.putheader('Accept', 'text/html')
        h.putheader('Accept', 'text/plain')
        h.putheader('User-Agent:', 'Mozilla/5.0 (Windows; U; Windows NT
5.1; en-US; rv:1.3) Gecko/20030312')
        h.putheader('Accept:', 'application/x-shockwave-
flash,text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/p
lain;q=0.8,video/x-mng,image/png,image/jpeg,image/gif;q=0.2,*/*;q=0.1')
        h.putheader('Accept-Language:','en-us,en;q=0.5')
        h.putheader('Accept-Encoding:', 'gzip,deflate,compress;q=0.9')
        h.putheader('Accept-Charset:', 'ISO-8859-1,utf-8;q=0.7,*;q=0.7')
        h.putheader('Keep-Alive:', '300')
        h.putheader('Cookie',
'SITESERVER=ID=fd7d0ca25acdd2439ab3f4b48b0827b4; GBA=18;
TVGID=9CE64992B85F4F9A82B57EBC092D11B2; nat=0; ServiceID=75347; zip=63130;
ptfc=yfki')
        h.endheaders()
        err = h.getresponse()
       return err
    except socket.error, er:
        print 'socket error ', er

when I use err.read() to read the object returned I get everything in Hex.

I tried using urllib, but there I can't figure out a way to pass the
location information etc.

As I said, I just need just general pointers on how to deal with this. Any
help would be greatly appreciated.

Thanks

Josh



--
http://mail.python.org/mailman/listinfo/python-list










From aleax at aleax.it  Sun Apr 20 03:45:49 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 20 Apr 2003 07:45:49 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <mailman.1050821359.13091.python-list@python.org>
Message-ID: <Nesoa.37952$LB6.862279@news1.tin.it>

Steven Taschuk wrote:
   ...
> On reflection, I realize that I have prior experience (from math
> reading and play) with many concrete examples of what reduce
> abstracts: big sigma notation for sums, big pi notation for
> products, similar notations for set unions and intersections,

However, big-Sigma (by far the most used of all of these) is
exactly equivalent to the new, simple 'sum' function I propose,
NOT to the generality of the higher-order-function reduce.

> etc..  Familiarity with these concrete examples makes the
> abstraction of reduce sit very comfortably.  "I see, reduce turns
> + into big sigma."  But perhaps such familiarity is not typical in
> Python's, er, target market.

I think the familiarity with SPECIFIC bulk-functions, big-sigma
first and foremost, is quite good.  A higher order function that
abstracts them all, and more, is quite a different issue IMHO --
it does require a penchant for abstract thinking, like so much
of functional programming does.  We claim (an old slogan but
still good) that "Python fits your brain" -- if we want to be
sincere when saying this to "ordinary programmers off the street",
then we should take care not too place an excessive emphasis on 
the generality of functional programming to the detriment of
clarity and simplicity for ordinary operations such as "sum up
these numbers".


Alex





From michael at stroeder.com  Tue Apr 29 02:17:27 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Tue, 29 Apr 2003 08:17:27 +0200
Subject: Web programming and a different "type" problem
In-Reply-To: <slrnbarhfb.67t.jon+usenet@snowy.squish.net>
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com> <slrnbarhfb.67t.jon+usenet@snowy.squish.net>
Message-ID: <o0i1o-i07.ln1@nb2.stroeder.com>

Jon Ribbens wrote:
> e.g. <input type="checkbox" name="foo*">
                                       ^^^
Is the * guaranteed to work?

http://www.w3.org/TR/html401/interact/forms.html#h-17.4

name        CDATA

http://www.w3.org/TR/html401/types.html#type-cdata

------------------------------ snip ------------------------------
ID and NAME tokens must begin with a letter ([A-Za-z]) and may be followed 
by any number of letters, digits ([0-9]), hyphens ("-"), underscores ("_"), 
colons (":"), and periods (".").
------------------------------ snip ------------------------------

Ciao, Michael.




From drs at ecp.cc  Wed Apr 30 16:59:59 2003
From: drs at ecp.cc (drs)
Date: Wed, 30 Apr 2003 20:59:59 GMT
Subject: set cookie problem (slightly OT)
References: <mailman.1051695171.12347.python-list@python.org>
Message-ID: <jPWra.5850$%_3.4001850@newssrv26.news.prodigy.com>

"Carsten Gehling" <carsten at gehling.dk> wrote in message
news:mailman.1051695171.12347.python-list at python.org...
> > Fra: python-list-admin at python.org
> > [mailto:python-list-admin at python.org]P? vegne af drs
> > Sendt: 30. april 2003 04:13
> > Emne: set cookie problem (slightly OT)
> >
> >
> > I have a web directory (apache2/FreeBSD) with ssi enabled but not cgi.
in
> > order to run simple python scripts i use something like
> >
> > <!--#exec cmd="/usr/local/bin/python /path/to/script.py" -->
> >
> > which works surprisingly well.  I cannot figure a way to set a cookie,
> > however, as i am not able to add to the header. Anyone have any
> > suggestions?
>
> Normally, Apache (and other webservers like IIS) pre-process headers
UNLESS
> you prefix the script name with "nph-". Try to do this:
>
> <!--#exec cmd="/usr/local/bin/python /path/to/nph-script.py" -->
>
> And - of course - rename your script accordingly. I haven't tried it, but
it
> should give you greater control with the headers.

[I might note that this is totally off topic now, but maybe worth
mentioning]

The above solution does not seem to work.  My guess is that once ssi has
started, headers are already written.  Instead, I added the following to a
.htaccess file

<Files index.html>
    Header add Set-Cookie "name=value"
</Files>

-d





From rmunn at pobox.com  Wed Apr  2 11:30:15 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 16:30:15 GMT
Subject: list index question
References: <e62b26d3.0304010613.39f98991@posting.google.com> <pan.2003.04.01.15.38.54.181811@mn.rr.com> <pan.2003.04.01.15.42.18.291186@mn.rr.com> <e62b26d3.0304011217.7f41254f@posting.google.com>
Message-ID: <slrnb8lf0v.ccp.rmunn@localhost.localdomain>

Post re-arranged to put replies *after* the questions...


Matt Singer <mrsinger at usa.net> wrote:
> "Anna" <revanna at mn.rr.com> wrote in message news:<pan.2003.04.01.15.42.18.291186 at mn.rr.com>...
>> On Tue, 01 Apr 2003 09:38:54 +0000, Anna wrote:
>> 
>> > 
>> >>>> list1=[(100,0), (10,5), (100,10), (100,0)] print list1
>> 
>> Have I mentioned that I *hate* it when my mailreader does this to a simple
>> copy/paste from IDLE... 
>> 
>> That print statement should, of course, have been on a separate line...
>> 
>> Ugh.
>
> Can I wildcard? I don't know the values in the second column.  I want
> to find the first entry what has a particular value in the first
> column.

The list.index() function won't help you, then: it just does a strict
equality comparison. You'll have to write a simple loop:

    index = -1
    for item in sql_result_list:
        index += 1
        if item[0] == 100:
            print "Found it"
            break
    if index >= len(sql_result_list):
        print "Not found"

I know you said you wanted to avoid looping, but this may not be
possible. Why did you want to avoid looping, anyway?

Another thing you may not have thought of is to loop through your
results once and build a dict that you could use for lookup:

    index_dict = {}
    for item in sql_result_list:
        key = item[0]
        index_dict[key] = item

Now you can use index_dict to do O(1) lookups based on the first item of
each tuple. Note that as written, though, only *one* item with the same
key will be kept. There are other ways around that, but I'm getting a
bit far off your original question.

BTW, I rearranged your post so the questions and answers would come in a
sensible order. Please don't top-post -- it's annoying and leads to
confusion. See http://www.caliburn.nl/topposting.html for reasons why.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From bfordham at socialistsushi.com  Thu Apr 17 09:45:57 2003
From: bfordham at socialistsushi.com (Bryan L. Fordham)
Date: Thu, 17 Apr 2003 09:45:57 -0400
Subject: Random web page?
In-Reply-To: <b7m4ql$ril$1@news.hccnet.nl>
References: <b7m4ql$ril$1@news.hccnet.nl>
Message-ID: <3E9EB015.30200@socialistsushi.com>

Anton Vredegoor wrote:

>From the universe of all currently available webpages (where
>"currently available" means available on the world wide web at the
>runtime of the script) I want to select a page in such a way that each
>page has an equally likely chance of being selected. Any idea about
>how to accomplish this is equally likely to be welcome ;-)
>  
>
I don't think it's really possible to know *all* the available pages.  I 
don't even know how you could really get such a list without running 
something like google yourself and constantly validating all the pages.

So why not just use google instead?  Write something to generate a 
random word or phrase, then submit that to google using the "I'm feeling 
lucky" feature.

--B






From fusion at thuule.pair.com  Sat Apr 12 16:05:53 2003
From: fusion at thuule.pair.com (Alan McIntyre)
Date: Sat, 12 Apr 2003 16:05:53 -0400
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <3E985DC8.9060907@thuule.pair.com>
Message-ID: <3E9871A1.5050304@thuule.pair.com>

Alan McIntyre wrote:
> Now I'm inclined to do a comparison just for my own curiosity's 
> sake...maybe I'll be back with graphs in just a bit. :)

Here's the times required for adding 200k two-character strings to a 
single string (in steps of 1000 += operations):

http://norfolkgraphics.com/testing/test.png

The green line is for Python, red for C++.  It looks kind of O(n)'ish 
for both of them.

Each point in the graph corresponds to the time required to perform 1000 
+= operations (the Python version of the code is below).  I dumped the 
output of the Python and C++ programs to text files and used SciPy to 
plot them.

The total run time for the Python version was 194.069 seconds, the C++ 
std::string version was 157.386 seconds.   I would have done a plain C 
version but my curiosity ran out and it's too nice outside. ;)

Full disclosure: The C++ program was compiled with Borland C++ Builder 
5, all optimizations turned on; the Python program is just as you see it 
below, run on Python 2.2.2 (binary install release from python.org). 
YMMV.  IANAL.

----------------------------------------------------------
import time

x = []
t = []
s = ''

starttime = time.time()

for i in range(200):
     ts = time.time()
     for j in range(1000):
         s += 'aa'

     x.append(i)
     t.append(time.time()-ts)

runTime = time.time() - starttime

for i in range(len(x)):
     print '%d, %.3f' % (x[i], t[i])

print '\nTotal run time: %.3f' % runTime





From aahz at pythoncraft.com  Wed Apr 23 12:17:51 2003
From: aahz at pythoncraft.com (Aahz)
Date: 23 Apr 2003 12:17:51 -0400
Subject: True/False
References: <mailman.1051110756.23235.python-list@python.org>
Message-ID: <b86ebf$ceb$1@panix1.panix.com>

In article <mailman.1051110756.23235.python-list at python.org>,
Anthony Baxter  <anthony at interlink.com.au> wrote:
>
>Wasn't there also some talk of making None, True and False keywords (for
>possible performance gains)? That would probably cause the reassignment
>of them to fail.

That won't happen until Python 3.0.  However, assignment to them may be
prohibited before then.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From cpl.19.ghum at spamgourmet.com  Wed Apr 16 07:59:08 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Wed, 16 Apr 2003 13:59:08 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net>
Message-ID: <Xns935F8DDA3CBC4cpl19ghumspamgourmet@62.153.159.134>

> the cursor in the source buffer jumps to the real cause of the error: the
> self.failxxx() call.

>What do people think? Is this a useful change, and should I submit a
>patch?

+1

Harald



From kemu at sdf-eu.org  Tue Apr 29 18:09:59 2003
From: kemu at sdf-eu.org (Geirgat Jonas)
Date: Tue, 29 Apr 2003 22:09:59 +0000
Subject: apache modules in python ?
Message-ID: <pan.2003.04.29.22.09.59.56666@sdf-eu.org>

Is it possible to write modules for the apache http server in python ?




From aicolburn at yahoo.com  Mon Apr 28 13:51:13 2003
From: aicolburn at yahoo.com (Alan C.)
Date: 28 Apr 2003 10:51:13 -0700
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <1a3a9ffa.0304280951.5f6f71cf@posting.google.com>

Although I don't know enough to recommend a good textbook, an idea
might be to contact Bruce Embry, who has taught Python to high school
students. His web page for high school Python at:
http://bembry.org/tech/python/index.php is a starting point. Perhaps
you could contact him from there.

Good luck! -- Al C.



From yvan_charpentier at hotmail.com  Tue Apr 15 14:40:28 2003
From: yvan_charpentier at hotmail.com (yvan)
Date: 15 Apr 2003 11:40:28 -0700
Subject: wxlistctrl and Sortitems()
Message-ID: <9ee55987.0304151040.35a04089@posting.google.com>

Does anyone know how i can sort on something else than the first column?
I was expecting the following to work, but it keeps on sorting on the first column.

def OnColClick(self, event):
	self.col = event.m_col
	self.list.SortItems(self.ColumnSorter)

def ColumnSorter(self, key1, key2):
	item1 = self.list.GetItem(key1, self.col).GetText()
	item2 = self.list.GetItem(key2, self.col).GetText()
	if item1 < item2:		return -1
	else:				return 1

Any help appreciated,

-Yvan



From BPettersen at NAREX.com  Sun Apr 20 19:28:02 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sun, 20 Apr 2003 17:28:02 -0600
Subject: python dynamic scoping question
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE011@admin56.narex.com>

> From: Julia Goolia [mailto:juliagoolia301 at hotmail.com] 
> 
> Hi,
> 
> I have a question regarding scoping in python.  I wish to have some
> mechanism which puts variables at a function level scope into the
> global scope, dynamically.  I cannot alter my design.

Good luck, given what you're saying, as opposed to only the fragment of
code, you would need perfect knowledge of all assignments that have
dynamically happened in a program, i.e. you can't find these variables
by static/lexical analysis. I'm almost positive there doesn't exists a
language where you can perform this operation in this direction unless
you consider making _all_ variables global (dynamic scoping does
something similar, but in the other direction :-)

> For example, I want this to work:
> 
> def f():
>   def g():
>     x = 5
>   g()
>   print x
> f()
> 
> I want to be able to see x in f, event though it was assigned in a
> lower scope, g.

Try e.g. creating a a global object called e.g. GlobalState and store
the variables there instead?

-- bjorn




From Kevin.Smith at sas.com  Fri Apr 18 10:06:29 2003
From: Kevin.Smith at sas.com (Kevin Smith)
Date: 18 Apr 2003 14:06:29 GMT
Subject: __getattr__ for globals?
Message-ID: <20030418100631860-0400@braeburn.themorgue.org>

Is there a way to get the same type of effect of the __getattr__ method 
on class instances with global variables?  I want to be able to look up 
the value of global variables from an external source if they don't 
exist in the current set of globals.

-- 
Kevin Smith
Kevin.Smith at sas.com



From radix at twistedmatrix.com  Fri Apr 18 19:44:06 2003
From: radix at twistedmatrix.com (Christopher Armstrong)
Date: Fri, 18 Apr 2003 19:44:06 -0400
Subject: Pair of prescription sunglasses left at PyCon sprint
Message-ID: <20030418234406.GA2993@radii>

I found a pair of prescription sunglasses on the table where the
Twisted gang was sprinting. If anyone can claim them, send me an
address to ship them to. Sorry for not posting about this earlier, I
had forgotten about them and they just turned up.

-- 
 Twisted | Christopher Armstrong: International Man of Twistery
  Radix  |          Release Manager,  Twisted Project
---------+     http://twistedmatrix.com/users/radix.twistd/




From tjreedy at udel.edu  Tue Apr  8 10:04:39 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 8 Apr 2003 10:04:39 -0400
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com>
Message-ID: <OIqdnV1p3aljSw-jXTWcpg@comcast.com>

>   I count half a dozen responses to my post, not one of which
provided
> any detail into protecting source code.

Why would you expect different?  People who think their code is
valuable enough to *actually* pay the price to code up a source code
protection method probably also think their source code protection
code to be valuable enough to also keep secret!  Breaking any
protection method is usually a *lot* easier when you know which
protection method was used.

So freely shared answers to questions like this are most likely to
come from people who are not doing it.  The very question is a
contradiction: "Will someone please freely share with me methods that
will enable me to *not* share freely with others?"  People who want
protection code to get more money from customers should either figure
it out themselves or take the hints given or offer to pay for help.

Consider cryptography in general.  Try asking any security/spy agency
for details on their cryptography methods.  Now notice that
publications in the field mostly (all?) come from academics who do not
actually do cryptography!  The science of keeping secrets secret is
generally keep secret, as much as possible, by its practitioners.

Terry J. Reedy





From jmj8237 at cs.rit.edu  Fri Apr  4 18:06:39 2003
From: jmj8237 at cs.rit.edu (Jason M Jurkowski ;002;icsg6;)
Date: Fri, 4 Apr 2003 18:06:39 -0500
Subject: My Object pool isnt working
In-Reply-To: <mailman.1049487908.13958.python-list@python.org>
References: <mailman.1049487908.13958.python-list@python.org>
Message-ID: <Pine.GSO.4.53.0304041759320.9099@wisconsin.cs.rit.edu>

On Fri, 4 Apr 2003, Michael Chermside wrote:

> > Whew! I always thought that I must use locks to access any object in
> > Python before I modify it from a thread.
> > Is it possible to append a list or a dictionary from a thread without
> > locking it?
>
> Yep. If you write two threads that both run around adding to a list
> then the list will wind up with all items, but their order isn't
> known (partial ordering of each thread's items IS preserved).
>
> > What does thread-safe mean here?
>
> Basically, it means that each individual bytecode of python (not each
> statement) is operated in series, not parallel. Erm... to put it
> better, even if you have multiple processors, your threads act as if
> they were running on just _one_ cpu and timeslicing, never interrupting
> the execution of a single bytecode. Except for threads that are
> blocked waiting on I/O, or which are performing C code that is
> "independent" of python (doesn't need to hold the GIL). This effect
> is achieved because there's a single (global) interpreter lock (GIL)
> which must be held for a thread to [start? complete?] a bytecode.
>
and the default number of bytecodes is 10.  see sys.setcheckinterval(n).
> Somebody please correct me if I've gotten some details wrong in the
> above explanation.
>
> > For example, what would happen if I walk through a list in a for loop
> > while another thread modifies the list?
>
> Don't do that... you might accidentally trigger the time machine and
> risk altering the future of the universe. No, seriously, Python won't
> crash, but your program might not do what you expect... it's not usually
> recomended. Someone with deeper knowledge can probably explain EXACTLY
> what will happen, but I suggest using a queue (see the Queue module)
> instead.
well a simple 'bad' scenario would be the following.

thread 1
for i in len(list):
	#do something with list[i]
thread 2
del list[x:]

if thread 1 has not finished iterating and comes upon the elements between
x and the end of the list a IndexError exception will be raised.
>
> -- Michael Chermside
>
>
>



From missive at frontiernet.net  Fri Apr 11 10:08:10 2003
From: missive at frontiernet.net (Lee Harr)
Date: Fri, 11 Apr 2003 14:08:10 GMT
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com> <b766fi$p3b$1@newstree.wise.edt.ericsson.se> <16411664.REEpLcBOk5@lothar.cengija.com>
Message-ID: <e%zla.369$xm2.150@news02.roc.ny.frontiernet.net>

In article <16411664.REEpLcBOk5 at lothar.cengija.com>, Davor Cengija wrote:
> I wrote a simple python program with gui for a friend of mine and his 
> projects at local university. His co-workers didn't want to use it since 
> 'it's too small, obviously it doesn't do what we need'. It's about 45k of 
> python source. Then I created self-standing exe with mcmillan's installer. 
> It's now about 7mb in size. Now it looks 'respectable' and 'worth of trying 
> but not for serious work since it doesn't have setup.exe with EULA and 
> other things -- so it isn't a real program'. Now I'm looking for a free 
> install-shield like software.
> 
> So, it isn't the quality of a software what only counts. Actually, I'm not 
> sure that the quality is the most important thing these days.


It is definitely a point of resistance for getting people to try Python.

If they click on an "applet" in their web browser and it suddenly starts
downloading all of a java runtime, they are quite happy when their 300 line
java program runs. But ask them to go to the python.org website and install
a programming environment which makes all of windows and java look like
a bad joke, and they cannot make the effort.

It seems that some people want everything to converge inside the browser,
but for me that is just not acceptable. I typically have 6-10 browser windows
open to various pages that I am working on. If some crazy applet freaks out
and crashes my browser, that is a huge hit to my productivity.

Ok, now I am way off topic. How about this as food for thought: What are the
chances that MS would include brilliant and free tools like Python or Perl
with their base OS? Ok. I will go laugh myself silly now  :o)




From andrew-pythonlist at puzzling.org  Fri Apr 25 09:34:05 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Fri, 25 Apr 2003 23:34:05 +1000
Subject: What means exactly "Memory error"?
In-Reply-To: <200304241917.47911.b.maryniuk@forbis.lt>
References: <200304241406.20833.b.maryniuk@forbis.lt> <200304241836.38456.b.maryniuk@forbis.lt> <20030424155413.GA88893@hishome.net> <200304241917.47911.b.maryniuk@forbis.lt>
Message-ID: <20030425133405.GJ15695@frobozz.local>

On Thu, Apr 24, 2003 at 07:17:47PM +0200, Bo M. Maryniuck wrote:
> On Thursday 24 April 2003 17:54, Oren Tirosh wrote:
> > Try to use a streaming XML API like SAX instead of DOM. For the splitting
> > operation you describe there shouldn't be any need to store the entire file
> > in memory. You may not even need and XML parser at all - looking for the
> > HR markers could probably be done by a few regular expressions without full
> > parsing of the XML stream.
> 
> Well, I only split it by <HR> and it surely causes no problem. But for the 
> rest (a lot of changings in the document to adapt it for Zope) I probably 
> need DOM -- regular expressions outside this job.

Why can't you use SAX if all you are doing is splitting by <HR>?  Sounds
like a perfect job for SAX to me -- you just write the XML stream out to a
file as you read it in, and switch to a new file everytime you hit <HR>.
The memory requirement for this technique would be minimal.

-Andrew.





From shalehperry at attbi.com  Tue Apr  8 00:40:15 2003
From: shalehperry at attbi.com (Sean 'Shaleh' Perry)
Date: Mon, 7 Apr 2003 21:40:15 -0700
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <b6teqm$956ai$1@ID-169208.news.dfncis.de>
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049731219.27404.python-list@python.org> <b6teqm$956ai$1@ID-169208.news.dfncis.de>
Message-ID: <200304072140.15819.shalehperry@attbi.com>

>
> Now *that's* an interesting idea! We would have, not a text
> editor, but a general hierarchical structure editor. Editing
> plain text would just be a special case, when we're editing
> an object of type "string".
>
> This would solve a number of problems at once. Syntax
> dependent editing is done by parsing the source into
> a tree of objects that know how to display and edit
> themselves. Embedding of "foreign" editing modes is done
> simply by inserting an object of the appropriate type into
> the tree. Want an reSt comment in your Python code? Just hit
> M-x insert_py_rest_docstring and off you go!
>
> The more I think about this, the more it seems like a
> REALLY REALLY interesting idea...
>

and it is.  However the reason most of the emacs modes, vim syntax 
definitions, etc all don't quite work is they are regex based.  To really get 
where the  structure editor needs to be would require a parser for each item 
type.  Not a trivial task.




From dddkk at sina.com  Sat Apr  5 19:39:28 2003
From: dddkk at sina.com (Chunyu)
Date: Sun, 06 Apr 2003 08:39:28 +0800
Subject: Emacs python-mode problem
Message-ID: <m2of3ktq67.fsf@db.hit.edu.cn>


While debugging a C/C++ program with GDB in emacs (version 21.3.50),
there is an arrow symbol `=>' in the source window to indicate the
current line normally. But After I load the python-mode.el file from
my ~/.emacs.el, the arrow symbol disappear while debugging with GDB in
emacs. (The python-mode.el is from
http://www.python.org/emacs/python-mode, version 4.6.)

I load the python in my ~/.emacs.el by
(require 'python-mode)
(setq auto-mode-alist
      (cons '("\\.py$" . python-mode) auto-mode-alist))
(setq interpreter-mode-alist
      (cons '("python" . python-mode) interpreter-mode-alist))

Now, I have to comment these lines, and load python-mode.el by hand
when I want to use python-mode. How can I solve it? Is there another
better python-mode code? Thanks for your replay.



From gerald at vamphq.com  Wed Apr  9 13:52:15 2003
From: gerald at vamphq.com (Gerald Bauer)
Date: 9 Apr 2003 10:52:15 -0700
Subject: How about Python + Mozilla for large apps?
References: <mailman.1049831897.12963.python-list@python.org>
Message-ID: <c4a49f85.0304090952.78b072d8@posting.google.com>

> Is there any project aimed to embed the Python interpreter inside 
> Mozilla OR to make available the XUL/XBL GUI engine of Mozilla to the Python 
> interpreter OR to create a full-capability bridge between Mozilla and Python 
> (maybe via XPCOM)?

  You might wonna check out Luxor - a free open-source XUL
engine/motor in Java that includes Python for scripting - online @
http://luxor-xul.sourceforge.net

  To get started using XUL and Python check out the
luxor-python-examples package in the download section. There's another
package called luxor-python-db-examples in the Luxor contrib download
section.

  - Gerald



From grisha at NOSPAM.modpython.org  Sat Apr 19 01:07:04 2003
From: grisha at NOSPAM.modpython.org (Gregory (Grisha) Trubetskoy)
Date: Sat, 19 Apr 2003 01:07:04 -0400
Subject: Web Hosting
In-Reply-To: <2rjna.171$hT2.182473@news2.news.adelphia.net>
References: <2rjna.171$hT2.182473@news2.news.adelphia.net>
Message-ID: <20030419005630.O28553-100000@eden.ispol.com>

More and more hosting companies pop up that offer FreeBSD jail(8) based
"Virtual Servers". You can compile and run anything your heart desires
because you get root access to the "server" and it's reasonably priced.

Grisha

On Wed, 16 Apr 2003, Bob X wrote:

> Can you name a good (reasonably priced) Web Hosting company that supports
> Python (either CGI or mod_python)?
>
> Bob
>
>
>




From a.schmolck at gmx.net  Sun Apr 13 19:03:26 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 14 Apr 2003 00:03:26 +0100
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it>
Message-ID: <yfshe92dmpt.fsf@black132.ex.ac.uk>

Alex Martelli <aleax at aleax.it> writes:

> A. Lloyd Flanagan wrote:
> 
> > Alexander Schmolck <a.schmolck at gmx.net> wrote in message
> > news:<yfsistmh4ec.fsf at black132.ex.ac.uk>...
> >  
> >> __slots__ is the one addition in python2.2 that really annoys me. It
> >> seems an ill-conceived jumble of conceptually unrelated things (bug
> >> avoidance, optimization, safety, encapsulation) that screws up reflection
> >> big time for little apparent gain. As Knuth said ...
>
> I think it's an important memory optimization, PERIOD.  The other
> conceptually unrelated things (bug avoidance -- I can't even SEE the
> others!) aren't in __slots__' own design (even though a commentator
> who's usually quite perceptive seems to have seen the "bug avoidance"
> feature that wasn't really there).

You may well have some superior knowledge, but to me personally the actual
design intention(s) were not that obvious from descrinto.html (BTW, as to why
 encapsulation and (more arguably) security:
- because you can hide attributes (and thus implementation)
- because you can enforce the layout (and type) of instances to thwart
  copying, and prevent method or class redefinition
(at least in practice, although possibly not by design)).

If the only reason is optimization, why couldn't it have been done less
disruptively (not a rhetorical question)? Specifically:

1. what's the point of not automatically including a slot for '__weakref__'
   (hardly the unreasonable memory consumption such a slot would impose?)?

2. similarly, why doesn't pickling just work out of the box?

3. why can a __slots__ class no just dynamically create an immutable
   pseudo-dict as read-only attribute whenever .__dict__ is accessed (similar
   to e.g. object.__dict__)? That would mean the python's so-far pretty
   unified introspection interface would still work [1].

4. why can't I assign to __class__? (I don't know exactly when else that
   breaks down, the only other case I'm aware of are assigning new-style
   classes to old-style instances and vice versa)

[Note: I just saw Tim's posting which promises that some of those issues will
go away]


> 
> As Knuth said, *premature* optimization is the root of all evil in
> programming. But __slots__ may allow very substantial optimization that need
> not be premature.

The implication was that including __slots__ in the language was a premature
optimization, not that every use of __slots__ is premature (although I fear
many will be). About any feature whose semantics or performance are not easily
achievable with other language features has its potential uses (even goto) --
but *that doesn't mean* that it (goto) should be included in the language in
the first place.

> 
> > Unless you absolutely have to have it, don't use it.
> 
> "Absolutely have to have it", just like "screws up big time", are
> silly overbids IMHO -- over-reactions that I don't see as justified
> at all!  Why can't we all just keep SOME perspective, please...?!

"screws up [reflection] big time"... Well there are 2 components to
reflection:

- intercession (which seems all but gone for classes with __slots__)
  
- introspection (which clearly is at the very least complicated: I'd think the
  majority of introspection features I normally use involve looking at an
  object's __dict__[1] -- such code could obviously be rewritten to also
  handle __slots__, but even if we assume that __slots__ isn't a lie [2], this
  at the very least involves discomfort and code breakage ).

And maybe a third:

- Its creator is not necessarily the only person who reflects a piece of work
  :)

> 
> Consider a typical use case, a class whose instances have and need
> a single attribute 'x', and of which we need to generate 10,000:
[snipped]
> Do we "*absolutely* have to have" the roughly factor-of-2 speedup given
> by __slots__?  

Do we "*absolutely* have to have" the roughly factor-of-100-to-1000 speedup
that [insert your favorite C, Fortran or assembler feature here] would allow
us for certain cases? Depends on implementation difficulty, ease of use and
error-proneness as well as the overall effect on the language
-- but unless it's a really common operation -- probably not. Especially since
pyrex, swig, boost, f2py etc. etc. make it reasonably easy to replace resource
critical sections with C or Fortran extensions.

> Clearly not -- we COULD live with a program that's roughly twice as slow,
> and here we're measuring only initialization anyway.

Exactly -- this is not a real program (which would hardly be twice as slow,
I'd assume?).

> and tens or hundreds of thousands of instances can well be necessary*?

Sure, and if I often felt an urge to have zillions of instances, I'd agree
even more heartily. But biased as I certainly am, I'm tempted to doubt the
problem occurs anywhere frequently to enough to justify shifting the burden
from those people who do encounter it (and might otherwise have had to write
an extension) to the language community as a whole.

> What is it that you think makes __slots__ so terrible, compared with many
> other possible optimizations?!

Optimizations (in a high level language) should either:

-  not cause *any* pain (i.e. be transparently performed by the language
   itself (e.g. constant expression lifting, loop unrolling, JIT compilation
   etc.) -- this of course is all very hard)

-  or the pain should be well-confined, at the very least when everything is
   done correctly (e.g. plastering the code with hints for the compiler or
   disabling usually automatic error checking for *correct* code).

At the *very* least their dangers should be obvious.

*None* of the above holds currently true for __slots__: basically, a simple,
seemingly innocuous addition of in the form of the single line 

__slots__ = ['foo', 'bar', ...

changes the semantics of a class *considerably*: it simply is no longer
behaves like an ordinary python class at all. Usually well defined operations
will now either stop working altogether (e.g. decoration), behave subtly
different (e.g. 'default' class variables) or send demons flying out your nose
(viz, cause undefined behavior).

But do the *particular* things __slots__ break really matter that much?

I really think they do, because I feel the current __slots__ design and
worsens some existing problems and contributes to the errosion of some of the
great things about python (simplicity and malleability). Seeing that it is
getting late and this post is already overly long, I'd better continue this
theme in another posting.

As I already stand accused of hyperbole I can however not withhold a final
comment: :)

> [I really wish I had a good way to find out how much memory a given
> container -- together with all of its contents -- is using... unfortunately,
> I don't know of any good way to answer that question!!!].

This feels a bit like handing a shotgun to a paranoid with blurry eyesight
*before* procuring a pair of glasses (at considerable peril to his own and his
neighborhood's safety).

(and although creation of zillions of instances might be a no-brainer for
memory bottleneck causation, in general even the best programmers are *really,
really* crap at intuiting bottlenecks)


'as


Footnotes: 
[1] .__dict__, vars, inspect.getmembers, dir ... Hmm, it's already getting
    sort of messy (all of these are *subtly* different but seem to have
    substantial overlap).

[2] i.e. __slots__ can be tampered with after class creation, which currently
    works but I guess is not intended to relied upon -- so most likely doing
    so means "undefined behavior" (otherwise the only way to find out seems to
    be to use `dir` (which's behavior is essentially undefined as it is
    intended as a mere convenience) -- or the funny inspect.getmembers(that
    however also cheats by calling dir :)).



From jepler at unpythonic.net  Mon Apr 28 20:55:43 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Mon, 28 Apr 2003 19:55:43 -0500
Subject: Tkinter widget question.
In-Reply-To: <%Ijra.707$5C2.204@tornadotest1.news.pas.earthlink.net>
References: <%Ijra.707$5C2.204@tornadotest1.news.pas.earthlink.net>
Message-ID: <20030429005543.GB22245@unpythonic.net>

Well, you could put them into a list or dictionary yourself:

def make_and_pack_entries(parent, count, **kw):
	ret = [Entry(parent, **kw) for i in count]
	for widget in ret:
		widget.pack()
	return ret

How else would you ever keep track of multiple objects in Python?  You can
keep track of widgets in all those ways... 

Jeff




From staschuk at telusplanet.net  Thu Apr 17 15:26:42 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 13:26:42 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net>; from tim.one@comcast.net on Wed, Apr 16, 2003 at 08:03:52PM -0400
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net>
Message-ID: <20030417132642.D7968@tibia.amotlpaa.bogus>

Quoth Tim Peters:
> [...] What to do with each
> method is in large part art, and the teachable seem to do OK learning by
> example.  [...]

I wonder, then, whether the docs would benefit from being
augmented with a few examples.

Donn -- would that improve their clarity in your view?

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From eppstein at ics.uci.edu  Fri Apr 18 21:13:37 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Fri, 18 Apr 2003 18:13:37 -0700
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <b7q2n1$24d2$1@nntp6.u.washington.edu>
Message-ID: <eppstein-095FC3.18133718042003@news.service.uci.edu>

In article <b7q2n1$24d2$1 at nntp6.u.washington.edu>,
 "Donn Cave" <donn at u.washington.edu> wrote:

> So you see the functional programming answers as "higher level"
> than the imperative, procedural approach taken above?  I think
> many functional programming enthusiasts would see it that way,
> but it's more interesting to hear it from a Python programmer.

I think it would be a mistake to think of me as being a typical Python 
programmer.  Mostly, I am a theoretical computer scientist.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From staschuk at telusplanet.net  Fri Apr 25 10:40:56 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 25 Apr 2003 08:40:56 -0600
Subject: Exception.__init__ needed in derived class?
In-Reply-To: <20030423140754.1946afea.gry@ll.mit.edu>; from gry@ll.mit.edu on Wed, Apr 23, 2003 at 02:07:54PM -0400
References: <20030423140754.1946afea.gry@ll.mit.edu>
Message-ID: <20030425084056.B261@tibia.amotlpaa.bogus>

Quoth george young:
> [python-2.3a1]
> Does class Exception have a constructor __init__ that needs
> to be called?  I couldn't find this in the docs.  I see many
> uses of inheriting from Exception without invoking
> Exception.__init__(self), but nobody ever *said* that this 
> was safe to leave out...
> 
> Pychecker complains about this, which is what brought it to
> mind.  I can certainly invoke it just to make pychecker 
> happy, but I wondered...

Exception does have an __init__, and probably it should be called.
If you don't call it, e.args will not be initialized:

    >>> e = Exception(1, 2, 3)
    >>> e.args
    (1, 2, 3)
    >>> class MyException(Exception):
    ...     def __init__(self, a, b, c):
    ...         self.a = a
    ...         self.b = b
    ...         self.c = c
    ... 
    >>> e = MyException(1, 2, 3)
    >>> e.args
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    AttributeError: MyException instance has no attribute 'args'

This might or might not be important to you.  Afaik the only place
.args is used is in Exception.__str__; if you're overriding
__str__ as well, that won't matter.  But other code might have a
use for .args; debuggers, for example.  In general it's good form
to call __init__ to ensure that the base class's invariants are
preserved -- unless you have a good reason for breaking them.

Many subclasses of Exception are just
    class HullBreachError(Exception):
        pass
and so they inherit Exception.__init__ along with everything else.

(Another, somewhat silly, way to get named attributes:
    class HullBreachError(Exception):
        a = property(lambda self: self.args[0])
        b = property(lambda self: self.args[1])
        c = property(lambda self: self.args[2])
This is inferior.  For one thing, it would be better to have the
meanings of the __init__ arguments documented in __init__'s
docstring.)

-- 
Steven Taschuk                               staschuk at telusplanet.net
"[T]rue greatness is when your name is like ampere, watt, and fourier
 -- when it's spelled with a lower case letter."      -- R.W. Hamming




From aahz at pythoncraft.com  Wed Apr 23 12:16:47 2003
From: aahz at pythoncraft.com (Aahz)
Date: 23 Apr 2003 12:16:47 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net>
Message-ID: <b86e9f$c15$1@panix1.panix.com>

In article <b86bqd$iu4$1 at news.hananet.net>,
Changjune Kim <juneaftn at REMOVETHIShanmail.net> wrote:
>"Aahz" <aahz at pythoncraft.com> wrote in message
>news:b867d7$f53$1 at panix2.panix.com...
>>
>> While writing it, it occurred to me that I *ought* to be using unit
>> tests to make sure that it was working correctly, but two things stopped
>> me.  Secondly, I was in a hurry.
>
>On the second:
>
>KentBeck : "If you write automated tests, you'll get more done every day."
>Johnny : "But we have too much to do. We don't have time to write tests."
>KentBeck : "No, no. You didn't understand. You'll get _more_ done, not
>less."
>Johnny : "Yes, but we don't have time for testing."

Yes, I know; I didn't realize this would be a "program" rather than a
"script" until I started writing it...

Which brings up the question: do the unit-testers here write tests for
throwaway scripts?
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From mhammond at skippinet.com.au  Mon Apr 28 08:50:10 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Mon, 28 Apr 2003 12:50:10 GMT
Subject: win32api registry methods and exception handling problem
In-Reply-To: <8ef9bea6.0304270835.96e779f@posting.google.com>
References: <8ef9bea6.0304270835.96e779f@posting.google.com>
Message-ID: <6s9ra.5572$lD4.43661@news-server.bigpond.net.au>

I've created a bug at sourceforge:

https://sourceforge.net/tracker/index.php?func=detail&aid=728896&group_id=78018&atid=551954

Thanks,

Mark.

Hung Jung Lu wrote:

> Hi,
> 
> The registry handling methods in win32api seem to have problems with
> the Python exception handling. It took me some time, but I have
> managed to isolate the problem to a simple program. But the problem is
> more severe in more complicated situations, and there it seems to
> happen with most (or all) registry methods (win32api.Reg...()).
> 
> The methods of _winreg module do not exhibit this problem. For now, I
> am using _winreg. However, it would be nice if the problem could be
> tracked down inside the win32api's source code and be fixed.
> 
> regards,
> 
> Hung Jung Lu
> 
> #----------------------------------------------------------------------
> # OS used: Windows XP
> #
> # Python 2.1
> #     program works correctly: correct exception message displayed
> #
> # Python 2.2.2 (win32all build 152)
> #    (a) In Pythonwin intractive window, you get
> #        SystemError: error return without exception set
> #    (b) When running with python.exe, it crashes
> #
> # Python 2.3b1 (win32all build 153)
> #    (a) In Pythonwin intractive window, you get
> #        SystemError: error return without exception set
> #    (b) When running with python.exe, program exits silently,
> #        which is not the correct behavior
> #
> # Comments:
> #     (a) equivalent _winreg module methods do not have problem.
> #     (b) problem seems to happen with most, or all, win32api.Reg...()
> #         methods, not in this particular example, but in more
> #         complicated programs (involving classes)
> #
> 
> import win32con
> import win32api
> 
> def reg_operation():
>     hkey = win32api.RegCreateKey(win32con.HKEY_CURRENT_USER,
> r'Software\Spam')
>     x = 3/0 # or a statement like: raise 'error'
> 
> try: reg_operation()
> except: x = 5/0 # or a statement like: raise 'error'




From aleax at aleax.it  Thu Apr 24 09:24:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 13:24:52 GMT
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com> <Xns9367885737E62duncanrcpcouk@127.0.0.1>
Message-ID: <EARpa.120$3M4.2728@news1.tin.it>

Duncan Booth wrote:

> Nikolai Kirsebom <nikolai.kirsebom.NOJUNK at siemens.no> wrote in
> news:0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062 at 4ax.com:
> 
>> class A(object):
>>      InstCount = 0
>>      def __init__(self):
>>           self.__class__.__dict__['InstCount'] += 1
>> 
>> class B(A):
>>      InstCount = 0
>> 
>> class C(A):
>>      InstCount = 0
>> 
>> Is the syntax used for incrementing the class-attribute for every
>> instanciation 'the way' it should be done ?
> 
> It would be more normal just to write:
> 
> self.__class__.InstCount += 1

Yes, but do note that this may produce strange anomalies if some
subclass erroneously forgets to rebind InstCount in its body...:

>>> class A(object):
...   InstCount = 0
...   def __init__(self):
...     self.__class__.InstCount += 1
...
>>> class B(A): pass
...
>>> a1=A()
>>> a2=A()
>>> b1=B()
>>> b2=B()
>>> B.InstCount
4
>>> A.InstCount
2
>>>

VERY unlikely to be what one wants, AND a rather subtle bug too.
The OP's direct access to __class__.__dict__ would ensure an
exception got raised upon instantiation of such an errant subclass.


> Alternatively you could create a class method to do the increment, but
> that is probably needlessly complex here.

Personally, I'd define a custom metaclass for this purpose.  Maybe
I'm getting TOO familiar with those, but, it doesn't sound all that
complex at all to me...:

>>> class metaIC(type):
...     def __new__(cls, name, bases, dict):
...         dict['InstCount'] = 0
...         return type.__new__(cls, name, bases, dict)
...     def __call__(cls, *args, **kwds):
...         cls.InstCount += 1
...         return type.__call__(cls, *args, **kwds)
...
>>> class A(object):
...   __metaclass__ = metaIC
...
>>> class B(A): pass
...
>>> a1=A(); a2=A(); a3=A()
>>> b1=B()
>>> A.InstCount
3
>>> B.InstCount
1
>>>

...isn't this smoother and handier?  Or does it just look that way
to me, due to overexposure to custom metaclasses?-)

When I think of behavior to be associated with a CLASS rather than
its INSTANCES, such as this case, more and more my thoughts turn to
a custom metaclass to encapsulate that behavior.  And why not?  it
does seem to be the normal way of OO -- just as I'd think of the
class in order to define behavior for its instances...


Alex




From peter at engcorp.com  Fri Apr 11 22:25:44 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 11 Apr 2003 22:25:44 -0400
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk>
Message-ID: <3E977928.B3A6DC7E@engcorp.com>

Alexander Schmolck wrote:
> 
> Erik Max Francis <max at alcyone.com> writes:
> 
> > My point was that if you find yourself wanting to make a truly unique
> > copy of a string (an immutable object), then there is almost certainly
> > something wrong with your design.
> 
> Hmm, do you mean if you want ``x is not y`` but ``x == y`` for interned string
> x and some string y?
> 
> But I'm not sure that that's what the OP's up to. E.g. you could have a
> container mapping symbols to other things -- possibly also symbols or further
> containers. You then might want to make a deepcopy of it. You presumably still
> want the symbols to be the same symbols in the copy (so that you can make use
> of the identity), so asking whether ``deepcopy(symbol) is symbol`` doesn't
> seem that unreasonable to me (is it?).

What is this "symbol" that you keep talking about?  If you
are simply using that as an alias for "interned string", then
I believe you are making up reasons for those which they are
*not* intended to have.  You said "the idea of symbols is their
uniqueness".  Well, I don't know what a "symbol" is to you,
but I believe that "interned strings" are *not* intended to 
guarantee some "uniqueness" property which you may count on 
in your code.  Any such property is a side effect of the
implementation or something.  I believe the sole purpose of
an interned string ("symbol"?) is to *optimize* access to those
strings, i.e. for performance reasons.  

See http://www.python.org/doc/current/lib/built-in-funcs.html#built-in-funcs
under "intern()" for what I believe is the only official purpose
of interning strings: "to gain a little performance on dictionary lookup".

-Peter



From gh at ghaering.de  Fri Apr 25 18:24:00 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Sat, 26 Apr 2003 00:24:00 +0200
Subject: Forgetting "()" when calling methods
In-Reply-To: <b8cadp$1fj9$1@ns.felk.cvut.cz>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <3EA9B580.9080005@ghaering.de>

Frantisek Fuka wrote:
> When I try to call methods, I sometimes forget to include the 
> parentheses. Instead of:
> 
> if object.isGreen():
>     do something...
> 
> i sometimes write:
> 
> if object.isGreen:
>     do something...
> 
> If I understand it correctly, the if statement in this case tests if 
> pointer to hasParent method is non-zero, which is of course always True 
> so no error is reported and "do something..." always gets executed, so 
> the application behaves in quite different way than I expected.

Yup. We prefer to talk about references instead of pointers, though.

> You can say to me "Don't forget to always include the parenteses"

Ok, "Don't forget to always include the parentheses" :-)

> but 
> I'm still curious if this cannot be somehow configured, so that I get 
> error when I try to just access the method pointer (".isGreen") instead 
> of calling the method (".isGreen()").

Ok, you asked for it ;-) Here's something special-cased for not 
forgetting the parentheses if the function is a boolean one and is used 
in an if-statement:

 >>> class DontForgetTheParens:
...   def __init__(self, callable):
...     self.callable = callable
...   def __nonzero__(self):
...     raise SyntaxError, "Don't forget the parentheses!"
...   def __call__(self, *args, **kwargs):
...     return self.callable(*args, **kwargs)
...
 >>> def foo(): return 5
...
 >>> foo = DontForgetTheParens(foo)
 >>> if foo(): print "5"
...
 >>> if foo: print "5"
...
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "<stdin>", line 5, in __nonzero__
SyntaxError: Don't forget the parentheses!
 >>>

Your homework for the weekend is to understand how this works ;-)

I hope it's clear that this falls into the "neat but useless" category.

If you want something more useful, I'd recommend you have a look at 
PyChecker. It should probably warn you about such programming mistakes.

-- Gerhard




From yann.morin.1998 at anciens.enib.fr  Sun Apr  6 16:21:24 2003
From: yann.morin.1998 at anciens.enib.fr (Yann E. MORIN)
Date: Sun, 6 Apr 2003 22:21:24 +0200
Subject: Install bug : incorrect permissions
In-Reply-To: <20030406183558.GA25907@meson.dyndns.org>
References: <20030406142002.ABD3D40E30@kraid.nerim.net> <20030406183558.GA25907@meson.dyndns.org>
Message-ID: <20030406202447.1A20B41004@kraid.nerim.net>

Hi!

Thanks Jp for your answer!

Once upon a time (on Sunday 06 April 2003 20:35), Jp Calderone wisely wrote:
 >   Python 2.2.2 installed all the lib-dynload/ files with permission 755 on
 > my system.  Perhaps you have a umask issue?

I double checked the umask for root, and it is 0022. Re-installing python
(after removing the previous installation, of course!) again installed the
lib-dynload files with perm 0750. And all other files installed have the
correct permissions.

config is done with :
./configure --prefix=/usr --with-cxx=g++ --with-signal-module \
  --with-threads --enable-unicode

CFLAGS and CXXFLAGS are -march=pentium4 -mcpu=pentium4 -mfpmath=sse -mmmx
-msse -msse2 (but that shall have no impact!), automake is 1.6.3, autoconf
is 2.57.

I checked the generated Makefile and nowhere the 0750 premission are shown.
A plain ./configure has the same problem with my system... :-(

Regards,
Yann.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| --==< ?_? >==-- ?---.----------------:  X  AGAINST      |  \e/  There is no  |
| web: ymorin.free.fr | SETI at home 1189 | / \ HTML MAIL    |   v   conspiracy.  |
?---------------------?----------------?------------------?--------------------?
 \__ np: Tri Yann - Les pr?paratifs du d?m?nagement __________________________/




From mis6 at pitt.edu  Mon Apr 28 08:32:08 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 05:32:08 -0700
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <2259b0e2.0304280432.4ddfb616@posting.google.com>

Steven Hodgen <steven at twitch.net> wrote in message news:<oproar1sjtx3zh5c at news.mindspring.com>...
> Hello,
> 
> I'm currently teaching a high-school introduction to programming course.  
> I'm currently teaching in Java, but am seriously considering changing to 
> Python.  The problem is that I cannot seem to find any good textbooks, much 
> less ones for students with no prior programming experience.  This is the 
> one area in which Java excels.  I'm using an awesome programming 
> environment called "BlueJ", and there are lots of textbooks for students at 
> all levels.
> 
> However, Java is not a great language for beginners.  Python, on the other 
> hand, is ideal.  If I can find a textbook that is even ok, I may consider 
> changing, since the language is so much better for students at this level.  
> Doesn anyone know of a book like this?  Ideally, it would have examples and 
> excercises, and perhaps even a web-site with other instructor and student 
> resources.
> 
> Thanks!
> 
> --Steven
> steven at twitch.net

I recommend "How to Think Like a Computer Scientist", which is perfect
for high school (see http://www.ibiblio.org/obp/thinkCSpy/)

  Michele



From staschuk at telusplanet.net  Wed Apr 30 10:13:42 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 30 Apr 2003 08:13:42 -0600
Subject: Question about accessing class-attributes.
In-Reply-To: <2259b0e2.0304300528.34b6f522@posting.google.com>; from mis6@pitt.edu on Wed, Apr 30, 2003 at 06:28:28AM -0700
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304251030.30fd1f94@posting.google.com> <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304300528.34b6f522@posting.google.com>
Message-ID: <20030430081342.F334@tibia.amotlpaa.bogus>

Quoth Michele Simionato:
  [...]
> Really, I don't see where is the problem. As far as I understand (correct me
> if I am wrong) dictionaries are indexed trough the identifier of their keys
> i.e. hash(A) == id(A). [...]

Unless __hash__ is overridden.

> [...] In this example, since A is not B, it is clear to me 
> that d[B] is not defined and cannot work.

But it seems wrong that we can have A == B and yet d[A] != d[B].
This can occur if A == B does not imply that hash(A) == hash(B).

> This issue has nothing to do with metaclasses:

Absolutely.  It has to do with not overriding __hash__ to be
consistent with the overridden __eq__.

-- 
Steven Taschuk             "The world will end if you get this wrong."
staschuk at telusplanet.net     -- "Typesetting Mathematics -- User's Guide",
                                 Brian Kernighan and Lorrinda Cherry




From mcfletch at rogers.com  Tue Apr  8 15:23:46 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 08 Apr 2003 15:23:46 -0400
Subject: Loading a class the name of which is only known runtime
In-Reply-To: <d1Fka.7728$y3.941559@news010.worldonline.dk>
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net> <3BDka.4759$Du.16741@newsc.telia.net> <1ruvtw54py90d.1sob8ay5wclmb.dlg@40tude.net> <d1Fka.7728$y3.941559@news010.worldonline.dk>
Message-ID: <3E9321C2.8040007@rogers.com>

In case Mads wanted something a little more fleshed out ;) .  This is 
from basictypes, pretty simple little function that's part of the 
late-binding machinery.  Note that this requires the class to be defined 
in the top-level of it's module.

def importByName( fullName ):
    """Import a class by name"""
    name = fullName.split(".")
    moduleName = name[:-1]
    className = name[-1]
    module = __import__( ".".join(moduleName), {}, {}, moduleName)
    return getattr( module, className )

Enjoy,
Mike


Bjarke Dahl Ebert wrote:

>"Mads Orbesen Troest" <mads at troest.NEVERMORE.dk> wrote in message
>
>  
>
>>Actually, what I am looking for is more alone the lines of:
>>
>>moduleToImport = "SomeModule" # <- Retrieved as CGI argument
>>classToInstantiate = "SomeClass" # <- Retrieved as CGI argument
>>
>>import moduleToImport
>>someInstance = classToInstantiate()
>>    
>>
...

>There is a slight complication, though:
>If themodulename is of the form "foo.bar", then module foo is returned. Some
>themodulename.split('.'), and additional getattr calls should help.
>  
>
...

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From bdesth at removethis.free.fr  Mon Apr 21 20:16:24 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 22 Apr 2003 00:16:24 +0000
Subject: Why functional Python matters
References: <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <b7m546$2a83$1@news.peterlink.ru> <lyptnlxq1n.fsf@cartan.de> <m24r4xdyd6.fsf@mycroft.actrix.gen.nz> <lyd6jlxkpi.fsf@cartan.de> <usft9v01l8fcdg8erdl2kishg0cnr1t9dp@4ax.com> <lyadepp1ck.fsf@cartan.de> <1lt0avslergps3bdv1lpvr48kef6t03640@4ax.com> <87d6jjjuzu.fsf@darkstar.cartan> <bdh1avcltqircf31cnkfro4thkafe2pamq@4ax.com> <Hpcoa.134462$It5.21768@news2.central.cox.net>
Message-ID: <3EA489D8.9040305@removethis.free.fr>

Steve Holden wrote:
> "Thaddeus L Olczyk" <olczyk at interaccess.com> wrote in message

(snip *Dr* Olczyk prose)

> See what happens when the Lispers invade (or did we invade them? - either
> way, it didn't take long to descend to the stage of ugly name-calling). I
> didn't think an 11-year-old would be able to *get* a Ph.D. If it's just
> something he did in the past, why does it appear at the foot of every email
> he sends out?
> 
> glad-to-stick-to-a-friendlier-environment-ly y'rs  - steve

Lol !-)

And sorry for being somewhat responsible for this lisper's invasion 
here. I had forgotten why I'd stop reading comp.lang.lisp first (bad 
signal/noise ratio).

Laotseu




From exarkun at intarweb.us  Fri Apr 25 19:08:34 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 25 Apr 2003 19:08:34 -0400
Subject: The woes of event handling in wxPython
In-Reply-To: <b8ca5o$8fe@dispatch.concentric.net>
References: <b8ca5o$8fe@dispatch.concentric.net>
Message-ID: <20030425230834.GA3870@meson.dyndns.org>

On Fri, Apr 25, 2003 at 09:43:20PM +0000, sparky wrote:
> [snip]
> 
> Using the evtmgr interface to deal with cleaning up the old dead objects 
> created by sending dialogs to the great bit bucket in the sky doesn't 
> seem to work.
> 
> Here is some example code (sorry I cannot give you a complete class, 
> they asked us not to post any code):
> 
> import pubsub, evtmgr
> class EditMemberDialog(EditDialog):
>     def __init__(self, parent, *args):
> .
> .
> <snip a bunch of stuff that works>
> .
> .
> .
> 	eventManager.Register(self.__saveMemberGoals, EVT_TEXT,
> 			      self.descriptionText)
>         server.subscribe(topic    = CommandPattern.CommandHistory,
>                          listener = self.__notifyObjectOfChange)
> 
> <snip a bunch of stuff that works>
> 
>     def __del__(self):
> 
> 	eventManager.Deregister(self.__saveMemberGoals)
> 	server.unsubscribe(self.__notifyObjectOfChange)
> 	self.parent.__del__()
> 
> 
> 
> So, now when I launch the dialog the first time. It works just fine.
> The second time it gets launched I get the:
> 
> wxPython.wx.wxPyDeadObjectError: The C++ part of the EditMemberDialog 
> object has been deleted, attribute access no longer allowed.
> 
> What am I missing?

  If EditMemberDialog instances have a reference to their parent, and their
parent has a reference to them, then what you have is a cycle.  A cycle
become uncollectable if more than one object in it implements __del__.

  Furthermore, when you pass self.__notifyObjectOfChange to
server.subscribe(), you are giving server (or whever eventually holds on to
the reference given by the listener parameter) a reference to self.  This
will probably preclude cleanup of the object entirely.

  Explicit unsubscription seems like the only course.  There are a few
tricks to doing this that might be of interest.  One involves keeping
WeakRefs to objects and calling the unsubscription function when the WeakRef
is no longer valid (it doesn't seem like this will work, because of the
reference I mentioned in the previous paragraph).  Another approach is to
subscribe before a "try:" block and unsubscribe in the matching "finally:"
block.  I'm not sure if this is feasible for use in a GUI app where you lose
control of execution between the subscription and the unsubscription.

  The last, and least magical, approach I can think of is to explicitly call
the unsubscribe method when you are finished with the object, and then clean
up the references to that object so it will be freed.  This probably won't
be easy, since much of the code is already written, but it may be the only
solution.

  Hope this helps,

  Jp

-- 
http://catandgirl.com/view.cgi?90
-- 
 up 36 days, 19:03, 9 users, load average: 0.48, 0.42, 0.37




From roy at panix.com  Tue Apr 15 18:58:44 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 15 Apr 2003 18:58:44 -0400
Subject: Big-O notation (was: Re: Python and Schools)
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org>
Message-ID: <roy-0FAA35.18584415042003@reader1.panix.com>

"Mike C. Fletcher" <mcfletch at rogers.com> wrote:
> You can have all sorts of fun learning how to combine and re-factor the 
> equations to get a final big-O notation, but for the most part, 
> practical coders can get away with recognising and avoiding O(N**X) 
> where X > 1 situations.

Indeed.  Lest people think this is of academic interest only, let me 
describe a situation we had at my company recently.

We've got something which scans a directory (an O(N) process).  It's 
supposed to happen once every time the program is run.  Somebody goofed 
in coding, and moved it into the loop that happens once every input 
file, not once per run.  Now it's an O(N^2) process.

Nobody noticed it in testing because we only tested it with maybe 100 
files.  Then one of our customers ran it with 25,000 files (a perfectly 
reasonable thing to do).  Whoops :-)



From jkraska at san.rr.com  Sun Apr 20 13:13:41 2003
From: jkraska at san.rr.com (Courageous)
Date: Sun, 20 Apr 2003 17:13:41 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl>
Message-ID: <erk5avkh423eh8bc13ctvtbjqt9f6kt4kv@4ax.com>

>Thanks. It seems Tim is making a heroic effort to explain the new in
>terms of the old. This necessarily must fail, but what other option is
>there instead of rebuilding all terminology out of gotos?

"non local goto" describes it well, but another way of looking at
it is as a "first class, user controlled context switch". That's
actually the meat of what it is. In unix, theoretically one could
set up the heavy weight equivalent of a continuation by the use of
setjmp and longjmp, although certain patterns of setjmp/longjmp
aren't always safe. Continuations are (or at least are supposed to
be).

In conjunction with a "stackless" environment, a continuation is
actually a very lightweight state save coupled with the ability
to restore to the last point the continuation was set.

C//




From tzahij at hotmail.com  Sun Apr 13 13:05:19 2003
From: tzahij at hotmail.com (Tzahi)
Date: 13 Apr 2003 10:05:19 -0700
Subject: Using Jython to process HTML
Message-ID: <662cede5.0304130905.6938703d@posting.google.com>

Hi,
I am working on a BI project that reads input HTML from the web, and
needs to do all kinds of manipulations to that HTML.

The project will be written in Java.

I am very new to Java, and have some experience with Python. I like
its dynamic nature and text processing.

I want to use Jython for the HTML processing power, so that I can
access other project and Java system classes. I hope this way I can
read pointers to documents from standard Queues, created by other
parts of the system, and process them with Python power.

I have never used Jython. Can you please tell me if this design make
any sense?

Do other people use Jython I real-life projects?

Thanks for any help or Idea

Tzahi Matar
www.ness.com



From glc at well.com  Tue Apr 15 09:34:47 2003
From: glc at well.com (Greg Chapman)
Date: Tue, 15 Apr 2003 13:34:47 GMT
Subject: Implementing Metatypes in C
References: <mailman.1050275480.8590.python-list@python.org>
Message-ID: <lk2o9v8bamjhgdmas6ce2g2qvuthril237@4ax.com>

On Mon, 14 Apr 2003 00:10:10 +0100, Phil Thompson <phil at river-bank.demon.co.uk>
wrote:

>1. Is my understanding correct, or is there a more correct way to do what I 
>want that avoids the problem?
>
>2. If my solution is correct, can the structure defined in typeobject.c be 
>published in a header file as the comment says? Please?

I believe your solution may break in 2.3a2 if subtypes declare tp_members, since
the PyMemberDefs for these are kept in the variable part of the type structure
which follows the things declared in etype.  At any rate, check out this patch
(accepted after the release of 2.3a2):

   http://www.python.org/sf/696193

which should allow you to cleanly add extra data slots to metatypes.

---
Greg Chapman




From gustav at morpheus.demon.co.uk  Sat Apr 12 14:06:50 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Sat, 12 Apr 2003 19:06:50 +0100
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
Message-ID: <65pjlhdx.fsf@morpheus.demon.co.uk>

Mads Orbesen Troest <mads at troest.NEVERMORE.dk> writes:

> Given that strings are immutable, is it exceedingly slow (and memory 
> spiking) to do a lot of "string += string" operations in one's code? IOW, 
> does each and every += evaluation lead to a new concatenated copy of the 
> previous string (now freed for garbage collection) and the new string, so 
> that the longer the string to which stuff is appended is, the longer times 
> each += operation takes?

I don't know the underlying details too well (and I'm not sure they
are all that important) but yes, string += string does construct a new
string each time, and hence is relatively expensive.

The normal idiom for building a string in chunks is to create a list
of the chunks, and then join them once at the end. So you do:

    chunks = []
    while whatever():
        chunks.append(next_bit_of_string())
    result = ''.join(chunks)

You could also use StringIO. However, all this does is basically what
I describe above, plus extra complexity to handle all the methods you
don't need. So you're better sticking with the above idiom.

Paul.
-- 
This signature intentionally left blank



From luc at nospam.com  Sat Apr 26 16:06:19 2003
From: luc at nospam.com (luc wastiaux)
Date: Sat, 26 Apr 2003 22:06:19 +0200
Subject: please comment on technologies
Message-ID: <slrnbalplq.254.luc@trillian.dont-panic.info>

Hello, I am writing the specifications for a school project (the subject 
is free for us to choose), I was thinking about doing a news-server 
archive database just like google groups, but on a smaller scale (it would 
only archive a limited number of newsgroup, not all of USENET). I got this 
idea because my school uses a news server and a lot of valuable 
information is lost when old messages are erased from the server's spool.

The project would be developped by a group of 4 to 6 CS students 
(including me), over a three months timespan (but not full time, we have 
to go to classes and other stuff)

I have a small experience with web stuff, but mostly with PHP, and some 
perl. and I'm learning python right now.

I'm (almost) sure we will make use of python as the main language, and use
either Mysql or postgresql for the database (I'm familiar with mysql but
maybe it's too limited for what I want to do?). I just found out about
Spyce today and it looks like I could use it for the web interface.  

Then I have this idea about using XSL for templating, the idea is not only 
to separate the data from the representation, but also allow programmers 
to "leech off" the search engine (it's easier to do this if the output is 
XML).

Please comment on everything.. about the technological choices, and if 
you want to suggest any features that we should add to the project, please 
don't hesitate to tell me about it.

thanks a lot!

-- 
luc wastiaux
 $ finger luc at info.4002.org



From max at alcyone.com  Sat Apr 26 16:28:53 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sat, 26 Apr 2003 13:28:53 -0700
Subject: MD5 to long?
References: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de> <b8dnlo$8sco1$1@ID-126101.news.dfncis.de> <aAtqa.6585$3M4.185018@news1.tin.it>
Message-ID: <3EAAEC05.D9F72B3@alcyone.com>

Alex Martelli wrote:

> A speedup of 15 times as a result of using simpler and more direct
> code seems to me to be worth pursuing.

Not to mention the potential security issues.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Laws are silent in time of war.
\__/ Cicero
    Bosskey.net: Return to Wolfenstein / http://www.bosskey.net/rtcw/
 A personal guide to Return to Castle Wolfenstein.



From dbrown2 at yahoo.com  Thu Apr  3 21:22:28 2003
From: dbrown2 at yahoo.com (dbrown2 at yahoo.com)
Date: 3 Apr 2003 18:22:28 -0800
Subject: Beginner: How to copy a string?
References: <mailman.1048813246.30445.python-list@python.org>
Message-ID: <d524bdb3.0304031822.2ed781be@posting.google.com>

Thanks everyone for setting me straight on some important points about
how strings work.  I think I have an improved understanding of the
Python view of things.

I did not need to make a copy of a string in my original case.  My
blunder was in not understanding that for

>>>a='abc'
>>>b=a
and then
>>>a='def' 

the final line does not change b's assignment.  I had somehow thought
I would loose the object 'abc' because b would then refer to 'def'.

But since everyone asked the same question "Horrors! Why in the world
would you ever need to copy a string?" I took it as a challenge.  I
started thinking about some not so esoteric possibilities,
specifically cases where Python deals with the outside world of real
hardware.  So, just for fun, here's a situation where copy might make
sense:

If I have a library that returns data in a "buffer", the library might
return a fixed-length buffer and Python would just assign the object
to an identifier.  Typical cases might be for talking to serial ports
or hardware registers.  If the library is not written in Python (say
it's compiled C or some other external code) then it seems the library
could return that same pointer to the buffer or register every time,
except it might be filled with different data. Perhaps this is not a
correct assumption.  In any case, if I got a string buffer back from
this kind of library I would want to copy the data to another Python
string so I know it won't change later.

Can this situation come about or am I also not understanding function
return values or how external libraries can work?  It seems there are
cases where Python could get a "string" that is not immutable because
the low-level library passes along a string that does not have the
same restrictions as the Python string.

David


> Before anything else ...
> 
> Why do you want to copy a string? What use do you *actually* have for a 
> copy of a string?
> 
> Now, onto the answer.
> 
> 1. In general, you don't. Strings are immutable, so a copy of a string 
> is indistinguishable from the original except by its ID.
> 
> 2. It is impossible to guarantee that you will get a copy of a string. 
> An implementation is free to do whatever it likes with strings e.g. it 
> could decide that *every* string with the same value will refer to the 
> same object - or it could have every string be a separate object.
> 
> 3. CPython interns all literal strings that look like python 
> identifiers. 'abc' (without the quotes) would be a valid python 
> identifier so it is interned. This is done partially so that identifier 
> lookups in dictionaries are as fast as possible.
> 
> 4. A string copy using copy() or [:] is special-cased by the string 
> object to simply return `self`.
> 
> 5. Given all that, to do what you want, the following will work with 
> current versions of Python (no guarantees of future versions though):
> 
>     s = 'abc'
>     print id(s)
>     t = s[:]
>     print id(t)
>     u = t[:1] + t[1:]
>     print id(u)
>     print repr(s)
>     print repr(t)
>     print repr(u)
> 
> Note however what happens if you try to copy an empty string ...
> 
>     s = ''
>     print id(s)
>     t = s[:]
>     print id(t)
>     u = t[:1] + t[1:]
>     print id(u)
>     print repr(s)
>     print repr(t)
>     print repr(u)
> 
> So, I come back to my original question - why do you want to copy a 
> string?
> 
> Tim Delaney



From skip at pobox.com  Fri Apr 18 16:33:45 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 18 Apr 2003 15:33:45 -0500
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
Message-ID: <16032.24873.189718.291023@montanaro.dyndns.org>

    Guy> What is the most Pythonic way to sum the n-th element of a list of
    Guy> lists (or of tuples)?

    Guy> The obvious (to me) way is to use reduce, a list comprehension, and
    Guy> a lambda:
    ...
    Guy> But I don't really like lambda, is there a better way to do this?

You can use the add function in the operator module:

    >>> import operator
    >>> x = (("a", 1), ("b", 2), ("c", 3))
    >>> reduce(operator.add, [y[1] for y in x], 0)
    6

Skip




From peter at engcorp.com  Fri Apr 18 09:31:28 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 18 Apr 2003 09:31:28 -0400
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl> <3E9EAA36.BFB1018D@engcorp.com> <IQmdnQg6xagDdQOjXTWcpA@august.net> <b7neqo$knk$1@news.hccnet.nl> <N_Nna.50877$t_2.4682@afrodite.telenet-ops.be> <b7og5c$h8b$1@news.hccnet.nl>
Message-ID: <3E9FFE30.DE250F28@engcorp.com>

Anton Vredegoor wrote:
> 
> Roel Schroeven <j4nff5m02 at sneakemail.com> wrote:
> 
> >Because not all websites can be reached that way, due to virtual
> >hosting. Take my website for example (don't look at it, it's really not
> >interesting :) ). http://roelschroeven.net gives a normal webpage, while
> >http://64.191.37.195 (which is the ip-address for that domain) gives a
> >generic page (in this case something that shouldn't really be there IMHO).
> 
> This is quite perplexing. My view of the internet is that it is like
> a phonebook where it is possible to link each name to a number. I
> forgot that sometimes a lot of people are living on the same address!
> Also thanks to Terry for bringing this up, although it took another
> post to make me understand what was meant. I guess I am back to the
> drawing board now :-|

I believe it's since version 1.1 that HTTP has required clients to
supply the domain name in the HTTP request, to allow this kind of
virtual hosting.  Think of it as nothing more (to use your analogy) 
than putting a name on an envelope *along with* the address, to 
get it to the right person there.  Prior to 1.1 (?) HTTP allowed
specifying only the street address...

I believe we may be at the point where, technically, it is 
impossible to do what you originally asked, since host names
could even be generated dynamically (although domain names 
cannot, I would think).

-Peter



From mwm at mired.org  Fri Apr  4 13:55:30 2003
From: mwm at mired.org (Mike Meyer)
Date: Fri, 04 Apr 2003 18:55:30 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <x7of3qjh4g.fsf@guru.mired.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com> <wBIia.2000$y3.310439@news010.worldonline.dk> <x7adf7ibk6.fsf@guru.mired.org> <Kp2ja.2590$y3.426173@news010.worldonline.dk>
Message-ID: <x74r5ehz30.fsf@guru.mired.org>

"Anders J. Munch" <andersjm at inbound.dk> writes:

> "Mike Meyer" <mwm at mired.org> wrote:
> > "Anders J. Munch" <andersjm at inbound.dk> writes:
> > > For a Python example, see the Decorate-Sort-Undecorate pattern.  In
> > > C++ the straightforward and fast solution is to use a suitable
> > > comparison predicate.  In Python you can do the same, but that's slow,
> > > and you are left with a choice between clarity and speed.
> > The statement about the preferred method in C is simply false. Anytime
> > the "suitable comparison predicate" is sufficiently complex - for
> > instance, what the unix sort(1) command allows - a DSU that allows
> > simple comparisons will be faster than a sort with complex
> > comparisons.
> The typical case is a simple projection.  And when it's not you can
> turn it into one by storing or caching the projected value in the
> objects themselves.

Uh - storing or caching the values is a decoration algorithm.

> > In fact, any time you can simplify the comparison with DSU, the DSU
> > sort will always perform faster on n objects for sufficiently large n.
> A comparison predicate in C++ (who said anything about C?) that
> forwards to instance variables simplifies the comparison similarly.

My bad - I tend to merge the two, because I only use either one as
high-level assemblers.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From rjkuhns at geetel.net  Wed Apr  2 21:20:43 2003
From: rjkuhns at geetel.net (Richard Kuhns)
Date: Wed, 2 Apr 2003 21:20:43 -0500
Subject: A question about help() and docstrings
In-Reply-To: <b6g0nc$518n3$2@ID-169208.news.dfncis.de>
References: <mailman.1049293506.17715.python-list@python.org>
	<b6g0nc$518n3$2@ID-169208.news.dfncis.de>
Message-ID: <20030402212043.2cad0a64.rjkuhns@geetel.net>

On Thu, 03 Apr 2003 13:02:32 +1200
"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> wrote:

> Richard Kuhns wrote:
> >>>>from dummy import *
> >>>>help('dummy_function')
> > 
> > no Python documentation found for 'dummy_function'
> 
> Try this instead:
> 
>    help(dummy_function)
> 
> i.e. pass it the function *object* itself, rather than
> the *name* of the function.

Thanks for the suggestion.  I started out with
"""help('dummy.dummy_function')""", which has worked fine when doing
something like """import dummy""", and so was rather bamfoozled when the
second form didn't work.

-- 
Richard Kuhns			rjkuhns at geetel.net




From nospam at somedomain.com  Thu Apr 10 09:07:59 2003
From: nospam at somedomain.com (maxx)
Date: Thu, 10 Apr 2003 15:07:59 +0200
Subject: newbie: strange python problem
References: <pan.2003.04.10.12.27.06.600319@somedomain.com> <3E956A20.E3321A5E@engcorp.com>
Message-ID: <pan.2003.04.10.13.07.59.235430@somedomain.com>

On Thu, 10 Apr 2003 08:57:04 -0400, Peter Hansen wrote:

> Check the output of "ls -la /" manually... the first line does
> *not* have more than two items in it...  if you were to change
> the first line above to this, it would work, though I suspect
> there is another option to "ls" which would also suppress
> that extra output.
> 
>> for line in os.popen('ls -la /').readlines()[1:]:
> 
> -Peter

Right! A simple mistake, but those are the ones you make taking your
first steps in a new programming language.

maxx



From rodrigob at elo.utfsm.cl  Mon Apr 28 14:59:33 2003
From: rodrigob at elo.utfsm.cl (Rodrigo Benenson)
Date: Mon, 28 Apr 2003 18:59:33 +0000
Subject: Pyrex Hard Case
References: <20030428171412.ZMGU12530.mta2@mta2.real.mail.entelchile.net> <mailman.1051552208.15532.python-list@python.org>
Message-ID: <3EAD7A15.6060508@elo.utfsm.cl>

Francois Pinard wrote:
>>>I'm not fully sure I'm replying to your question, but this might work:
>>>
>>>---------------------------------------------------------------------->
>>>cdef extern from *:
>>>    MACRO_NAME(EXPECTED_ARGUMENTS)
>>>----------------------------------------------------------------------<
>>
> 
> [rdrb at 123.cl]
> 
> 
>>I had already done that.
>>extern MACRO_NAME(EXPECTED_ARGUMENTS)
> 
> 
> This is not the same thing.  Your macro declaration, in Pyrex, really
> ought to be within the `cdef extern from *:' and indented within it.
> You do not need to repeat `extern', yet I presume it does not hurt either.
> 

I just replyied shortly, in fact I did tested your suggestion and it 
create the same output. It's obvious it should do, because we are 
defining and External Function Prototype. So Pyrex will define the 
prototype in order to have a good compilation.

In normal C I should just simply avoid the prototype and let's the Macro 
do his job. But here in Python/Pyrex I nedd to tell to Pyrex how to 
manage/translate the objects applyied to and resulting form this macro.

I think that Pyrex should have something like:
"""
cdef macro foo_type foo_function (foo_parameters)
"""
That behave like an cdef extern but do not instanciate the extern prototype.

Be the way maybe Pyrex have already a trick to do this in 0.7.

Thanks anyway.
I will be waiting for more attempt.
RodrigoB.




From Vector180W at netscape.net  Thu Apr  3 04:41:34 2003
From: Vector180W at netscape.net (vector)
Date: Thu, 03 Apr 2003 19:41:34 +1000
Subject: writing to windows memory
Message-ID: <3E8C01CE.6070601@netscape.net>

newbie:
pythonwin 2.2.2
Id like to bind a button.widget that simply writes a one to windows 
address 440h. any ideas on where to start.(the button stuf is trivial) i 
need to know how to access the address space.




From jgardner at jonathangardner.net  Wed Apr 16 22:39:05 2003
From: jgardner at jonathangardner.net (Jonathan Gardner)
Date: 16 Apr 2003 19:39:05 -0700
Subject: PyQt Tutorial at NorthWest Linux Fest
Message-ID: <mailman.1050619205.11009.clpa-moderators@python.org>

I am teaching a beginner's PyQt (Python + Qt) tutorial at the
NorthWest Linux Fest in Bellingham, WA on April 26. More information
about the linux fair can be found at http://www.linuxnorthwest.com/

I will be covering the basics about PyQt. If you attend, you will
learn what it is, where you can find the sources for it, how to
compile it and install it on your system, as well as how to use it.
There won't be enough time to cover everything, but there will be
enough time to show you the ropes and get you started in the
documentation. We'll also be able to whip out an application using
some RAD methods.

I used PyQt to develop a claims management and submission system for a
local optometrist with great results. I think it is ready for
production use in a variety of environments.

If you have any questions, feel free to ask.




From bokr at oz.net  Fri Apr 11 19:06:32 2003
From: bokr at oz.net (Bengt Richter)
Date: 11 Apr 2003 23:06:32 GMT
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <d116fbae.0304100704.68ef4833@posting.google.com> <c48$3e960655$516049d2$3498@nf1.news-service.com> <3e96efac$0$49107$e4fe514c@news.xs4all.nl> <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk> <3e96f481$0$49099$e4fe514c@news.xs4all.nl> <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk> <3e970abc$0$49101$e4fe514c@news.xs4all.nl> <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk> <3e972e43$0$49109$e4fe514c@news.xs4all.nl>
Message-ID: <b77hpo$59a$0@216.39.172.122>

On Fri, 11 Apr 2003 23:04:01 +0200, Irmen de Jong <irmen at -NOSPAM-REMOVETHIS-xs4all.nl> wrote:

>Robin Becker wrote:
>
>> sigh :(
>> for real distributed computing I wouldn't want to hold the sockets open.
>
>Sorry if I missed that requirement. I was just responding to Duncan's
>CORBA timings...
>
>> I just tried the original server.py/time-client.py pair again this time
>> on a freeBSD system
>> 
>> freeBSD-->localhost 500 secs
>> freeBSD-->192.168.0.3 50 secs (a local win32 machine).
>> 
>> 
>> I guess I'm just not doing what I think I'm doing.
>
>I really can't explain these numbers either? Localhost is 10 times slower
>than over the network???

IIRC, there was some discussion some time ago about this being a particular
FreeBSD pessimality problem that wasn't true of Linux, but someone was fixing it?
But FreeBSD has good networking reputation I thought, so this needs double check,
for sure. Maybe it was for an old version?

>
>Unless your client & server both struggle for CPU cycles on the same machine 
>at the same time...
Or the struggle involves unnecessary scheduling waits of some kind?

Regards,
Bengt Richter



From andy at wild-flower.co.uk  Wed Apr 16 16:48:26 2003
From: andy at wild-flower.co.uk (Andy Jewell)
Date: Wed, 16 Apr 2003 21:48:26 +0100
Subject: file read : feature or bug ?
In-Reply-To: <3e9c2eea$0$6528$afc38c87@sisyphus.news.be.easynet.net>
References: <3e9c2eea$0$6528$afc38c87@sisyphus.news.be.easynet.net>
Message-ID: <200304162148.26858.andy@wild-flower.co.uk>

On Tuesday 15 Apr 2003 5:10 pm, Asbuilt Easynet wrote:
> Hello,
>
> I have a problem to use the read function that seems read more than
> requested
> as you can see in this sample demo.
>
> Thanks
>
> Jef
>
> # This is a program to read a file in a reverse order by 3 characters block
> # the demo.log file is 2 characters per line in windows so ended with CRLF
> # 01
> # 02
> # 03
> # 04
> #
> # we can see it like 01__02__03__04 where __ is CRLF
> #
> # PROBLEM :
> # the program normaly should return
> #   _04
> #   03_
> #   2__
> #   __0
> #   11
> #
> # BUT it returns
> #   _04
> #   03_
> #   2_0     <-- From where is coming the 0 ?
> #   _02     <-- From where is coming the 2 ?
> #   11
> #  (this is an extract from rgrep.py)
>
> import string
>
> l_bufsize = 3
> l_file    = open("demo.log")
> l_file.seek(0, 2)
> l_pos      = l_file.tell()
>
> while l_pos > 0:
>   l_size = min(l_pos, l_bufsize)
>   l_pos  = l_pos - l_size
>   l_file.seek(l_pos)
>   l_buffer = l_file.read(l_size)
>   print "l_buffer = " + l_buffer + "[" + str(len(l_buffer)) +"]"
>
>   l_lines = string.split(l_buffer, "x")
>   print "Split = " + str(l_lines)
>   del l_buffer


Is this just an exercise or a contrived example to demonstrate your problem? 

Doing file I/O in python with this style of processing is *incredibly* 
inefficient.  If the file can be read into memory, that's the best place to 
process it. 

Here's a more pythonic way of doing what you're doing above:

infile=file("demo.log","r")
lines=infile.readlines()
lines.reverse()
for line in lines:
    print line

If you were merely experimenting with seek, I guess this wont help much! :-)  
But the same principle still applies - get the file into memory if you can 
and then manipulate it there.  Always use builtin functions where they do 
what you want - they're in C and are therefore much faster than python.  

Python's strengths lie elsewhere than sheer computational muscle; readability, 
manageability and a wealth of readily available libraries are just three of 
them.

Hope that helps

-andyj




From tanzer at swing.co.at  Mon Apr 14 12:45:40 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Mon, 14 Apr 2003 18:45:40 +0200
Subject: an ugly file-reading pattern
In-Reply-To: Your message of "Mon, 14 Apr 2003 08:53:39 CDT."
             <16026.48483.929843.87511@montanaro.dyndns.org>
Message-ID: <E19575Q-0003JQ-00@tswings.swing.cluster>

Skip Montanaro <skip at pobox.com> wrote:

>     Christian> Maybe you should have written it as :
>
>     Christian>     open = file("somefile")
>     Christian>     for line in open:
>     Christian>         process(line)
>
>     Christian> <2.2 wink>
>
> Yeah, stupid mistake.  I had part of the OP's original solution (vrbl named
> file) and part of mine...

I wouldn't call it stupid, it was just not newbie-friendly in a
context like this.

Unfortunately, for a long time `file` was an obvious variable name for
an open file and old habits die hard. Grepping over my source tree
shows 183 occurrences of it.

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From alessandro.bottoni at infinito.it  Sun Apr  6 06:12:14 2003
From: alessandro.bottoni at infinito.it (Alessandro Bottoni)
Date: Sun, 6 Apr 2003 10:12:14 +0000
Subject: "Automation" of OpenOffice apps via Python?
Message-ID: <200304061012.14054.alessandro.bottoni@infinito.it>

I'm in the need of remote-controlling OpenOfficeWriter and others OpenOffice 
applications from a custom program and I wonder if there is any way to acces 
the OpenOffice API from Python. The "at a glance" presentation of OpenOffice 
SDK/API tells that that is possible but I'm unable to find any further 
instruction.

Does anybody know if there is a specific Python binding for the OpenOffice 
API? Or is the API accessible from the "standard"  versions of Python and 
OpenOffice via... via what? CORBA? XPCOM? DCOP? Telnet?! What else?

Many thanks in advance for you attention.
 
---------------------
Alessandro Bottoni






From cpl.19.ghum at spamgourmet.com  Mon Apr  7 17:23:12 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Mon, 7 Apr 2003 23:23:12 +0200
Subject: cx_Oracle and Number(11)
References: <3e918724@news.mt.net.mk>
Message-ID: <Xns9356EDF45F8E9cpl19ghumspamgourmet@62.153.159.134>

> I have a column in the Oracle database that is defined as 
>   ORDER_ID NUMBER ( 11),
> but when I select that table with cx_Oracle 2.5 the values in the column
> show up as floats in the python program.

Isn't that by definition? Number turns to float, INT to INT?


Ashes to ashes
strings to strings?





From DennisR at dair.com  Mon Apr 14 16:27:27 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Mon, 14 Apr 2003 20:27:27 GMT
Subject: Reading/writing files
References: <pan.2003.04.14.17.32.01.209427@soul.net> <pan.2003.04.14.19.07.36.357067@effectif.com>
Message-ID: <PQEma.1416$%o7.634@newssvr19.news.prodigy.com>

> input file. That's because print automatically appends a new line
> character to the end of each line it prints, and the file already contains
> new line characters. To get round that, you could import the sys module
> and use sys.stdout.write(line) inplace of the print statement.
> Alternatively you could use the rstrip() method on a string (see below).

Far simpler is a trailing comma:   "print line," instead of "print line"

--

Dennis Reinhardt

DennisR at dair.com
http://www.spamai.com





From staschuk at telusplanet.net  Mon Apr 21 13:55:43 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 21 Apr 2003 17:55:43 -0000
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 21)
Message-ID: <va8c4vbs10e158@corp.supernews.com>

QOTW:  "This suggests that "nobody" doesn't eval to 0, but at least 3."
                                            -- Christian Tismer

"This is one thing Grothendieck got absolutely right: one should
always, always, always consider the relative situation (i.e. think
about morphisms rather than maps)." --  Michael Hudson


Discussion:
    Much ado about the appropriate uses of __str__ and __repr__ (with
    a digression on orbital mechanics, for flavour).
        <http://groups.google.com/groups?th=24b817d49ec3a59b>

    Roy Smith illustrates the "Big O" notation by an informal analysis
    of everyone's favourite Python performance gotcha -- repeated
    string concatenation.
        <http://groups.google.com/groups?th=274dcb113af97b01#link3>

    David Mertz suggests making repeated string concatenation faster.
        <http://groups.google.com/groups?th=5968aec796657f37>

    Continued from last week: the state of the Python web programming
    world; further development of the theme "gee, there's a lot of
    frameworks"; ruminations on simplifying the situation.
        <http://groups.google.com/groups?th=903e379430bffdfd>

Announcements:
    The deadline for Python 11 - OSCON 2003 early bird registration is
    May 23rd, 2003. The conference will be held July 7-11, 2003 in
    Portland, Oregon, USA.
        <http://conferences.oreillynet.com/os2003/>

    PyKota v1.03: PyKota is a python-written, internationalized,
    complete, centralized and extensible print quota solution for CUPS.
        <http://www.librelogiciel.com/software/PyKota/>

    Pyrex 0.7.2: A language for writing Python extension modules.  It
    lets you freely mix operations on Python and C data, with all
    Python reference counting and error checking handled
    automatically.  This release fixes several bugs.
        <http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/>

    Stackless Python 3.0 alpha 1: Christian Tismer's alternative
    Python implementation, once again much transformed.
        <http://www.stackless.com/>

    Tux Math Scrabble v2.0: A math version of the popular board game
    for ages 4-40.
        <http://www.asymptopia.com/>


========================================================================
Everything you want is probably one or two clicks away in these pages:

    Python.org's Python Language Website is the traditional
    center of Pythonia
        http://www.python.org
    Notice especially the master FAQ
        http://www.python.org/doc/FAQ.html

    PythonWare complements the digest you're reading with the
    daily python url
         http://www.pythonware.com/daily  
    Mygale is a news-gathering webcrawler that specializes in (new)
    World-Wide Web articles related to Python.
         http://www.awaretek.com/nowak/mygale.html 
    While cosmetically similar, Mygale and the Daily Python-URL
    are utterly different in their technologies and generally in
    their results.

    comp.lang.python.announce announces new Python software.  Be
    sure to scan this newly-revitalized newsgroup at least weekly.
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python.announce

    Brett Cannon continues the marvelous tradition established by 
    Andrew Kuchling and Michael Hudson of summarizing action on the
    python-dev mailing list once every other week.
	http://www.python.org/dev/summary/

    The Python Package Index catalogues packages.
        http://www.python.org/pypi/

    The somewhat older Vaults of Parnassus ambitiously collects references
    to all sorts of Python resources.
        http://www.vex.net/~x/parnassus/   

    Much of Python's real work takes place on Special-Interest Group
    mailing lists
        http://www.python.org/sigs/

    The Python Business Forum "further[s] the interests of companies
    that base their business on ... Python."
	http://www.python-in-business.org

    The Python Software Foundation has replaced the Python Consortium
    as an independent nexus of activity
        http://www.python.org/psf/

    Cetus does much of the same
        http://www.cetus-links.org/oo_python.html

    Python FAQTS
        http://python.faqts.com/

    The old Python "To-Do List" now lives principally in a
    SourceForge reincarnation.
        http://sourceforge.net/tracker/?atid=355470&group_id=5470&func=browse
        http://python.sourceforge.net/peps/pep-0042.html
     
    The online Python Journal is posted at pythonjournal.cognizor.com.
    editor at pythonjournal.com and editor at pythonjournal.cognizor.com
    welcome submission of material that helps people's understanding
    of Python use, and offer Web presentation of your work.

    *Py: the Journal of the Python Language*
        http://www.pyzine.com

    Tenth International Python Conference 
        http://www.python10.org            

    Archive probing tricks of the trade:
        http://groups.google.com/groups?oi=djq&as_ugroup=comp.lang.python&num=100
        http://groups.google.com/groups?meta=site%3Dgroups%26group%3Dcomp.lang.python.*

Previous - (U)se the (R)esource, (L)uke! - messages are listed here:
  http://www.ddj.com/topics/pythonurl/         
  http://purl.org/thecliff/python/url.html (dormant)
or
  http://groups.google.com/groups?oi=djq&as_q=+Python-URL!&as_ugroup=comp.lang.python


Suggestions/corrections for next week's posting are always welcome.
E-mail to <Python-URL at phaseit.net> should get through.

To receive a new issue of this posting in e-mail each Monday morning
(approximately), ask <claird at phaseit.net> to subscribe.  Mention
"Python-URL!".


-- The Python-URL! Team--

Dr. Dobb's Journal (http://www.ddj.com) is pleased to participate in and
sponsor the "Python-URL!" project.




From tvassila at siac.com  Mon Apr 14 09:58:56 2003
From: tvassila at siac.com (Thanos Vassilakis)
Date: Mon, 14 Apr 2003 09:58:56 -0400
Subject: Inserting Python commands into HTML documents
Message-ID: <OF496B4788.ED594A0B-ON85256D08.004C507B@nsmtp.siac.com>

Try using Python Service Objects (pso) ,see:




user guide: http://www.scriptfoundry.com/modules/pso/doc/pso-guide.html

and  building a site in 10 minutes:
http://www.scriptfoundry.com/modules/pso/doc/pso-example.html





Hello,

I am an extremely new member of the Python family. I had a job interview
  where they charged me with the task of creating a (fake) corporate
website using Python in ten days--with two left now.

The problem is that what I have read seems to say Python is a CGI
language... Am I remiss in believing that my website is going to be
completely run from inside a couple .PY files in the CGI-BIN?

Is there a way for it to act more like PHP, as in run Python commands
from inside the (index|contacts|about|etc).html or at least simulate
that behavior? (This is on a linux server and I am not to use PHP, FYI.)


Thank you all for any assistance!

--
http://mail.python.org/mailman/listinfo/python-list






-----------------------------------------
This message and its attachments may contain  privileged and confidential information.  If you are not the intended recipient(s), you are prohibited from printing, forwarding, saving or copying this email.  If you have received this e-mail in error, please immediately notify the sender and delete this e-mail and its attachments from your computer.





From tim_one at email.msn.com  Fri Apr 11 01:16:13 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Fri, 11 Apr 2003 01:16:13 -0400
Subject: Adding item in front of a list
In-Reply-To: <pan.2003.04.10.13.07.10.593583.815@thomas-guettler.de>
Message-ID: <LNBBLJKPBEHFEDALKOLCKEJGEGAB.tim_one@email.msn.com>

[Duncan Booth]
>> I don't actually see the problem with this. list.insert() currently
>> takes exactly two arguments, therefore there would be no backwards
>> compatibility issues with allowing it to take one or two arguments
>> where a single argument acts as though you had given two but the first
>> defaulted to 0. Just like 'range' does in fact!

[Thomas Guettler]
> I newer looked at the source of python, but I don't think that it would
> be difficult. Should this become a PEP or can someone with write
> access to the CSV just do it?

See my earlier reply to Duncan:  this particular idea is dead on arrival (a
PEP would be rejected; if someone checked in such a change it would be
reverted).  Saving you from needing to type 0 when you mean 0 doesn't
outweigh the ugliness of making the first argument mean different things
depending on how many arguments there are.





From pan-newsreader at thomas-guettler.de  Wed Apr  2 00:45:41 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 02 Apr 2003 07:45:41 +0200
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de> <b6d5s3$lvo$2@newshost.accu.uu.nl>
Message-ID: <pan.2003.04.02.07.45.41.185943.810@thomas-guettler.de>

On Wed, 02 Apr 2003 01:03:31 +0200, Martijn Faassen wrote:

> Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote:
>> I looked at zope3 but it seems very bloated.
> 
> It isn't done yet, obviously. How does it seem bloated? Too much
> framework for your tastes?

Yes, too much framework. 

I like ZODB, but don't like ZPT, DTML, Acquisistion, ZClasses,
Interfaces and Components.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From sdieselil at yahoo.com  Wed Apr  2 07:32:10 2003
From: sdieselil at yahoo.com (sdieselil)
Date: 2 Apr 2003 04:32:10 -0800
Subject: Can double-linked lists be implemented in python?
Message-ID: <169a7283.0304020432.5a824d1@posting.google.com>

Hi

How can I implement double-linked lists and other complex tree-like
structures in Python?



From bram at nospam.sara.nl  Mon Apr 28 04:24:00 2003
From: bram at nospam.sara.nl (Bram Stolk)
Date: Mon, 28 Apr 2003 10:24:00 +0200
Subject: dynamic function names
Message-ID: <20030428102400.21c5074a.bram@nospam.sara.nl>

Hello,

Is it possible in Python to define functions with a dynamic name?

something like:

def make_func(funcname) :
  def funcname :
    print "This func is named", funcname

make_func(foo)
make_func(bar)

The code above fails due to foo being unknown,
however, when calling as make_func("foo"), the "foo" string
cannot be used as a funcname.

Thanks,
  
   Bram Stolk

-- 
------------------------------------------------------------------------------
 Bram Stolk, VR Specialist.
 SARA Academic Computing Services Amsterdam, PO Box 94613, 1090 GP  AMSTERDAM
 email: bram at sara.nl   Phone +31-20-5923059  Fax +31-20-6683167

 "I heard if you play the NT-4.0-CD backwards, you get a satanic message."
 "Thats nothing, if you play it forward, it installs NT-4.0"
------------------------------------------------------------------------------



From aleax at aleax.it  Thu Apr 17 05:13:54 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 09:13:54 GMT
Subject: "pythonic"
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
Message-ID: <mfuna.26072$LB6.616225@news1.tin.it>

Afanasiy wrote:

> Too many times now I have seen the term "pythonic" used to vaguely
> describe something which is never explicitly defined, as if it cannot be.
> 
> Now it seems I can just replace "pythonic" with "my idea of utopia".
> 
> I do not know if this ignorance, laziness, or a meme, but I would like
> to know if there a real definition of "pythonic". Anyone?

If you want a thesaurus:

http://www.bartleby.com/62/49/P1214900.html

"""
pythonic

ADJECTIVE:
Of extraordinary size and power: behemoth, Brobdingnagian, Bunyanesque, 
colossal, cyclopean, elephantine, enormous, gargantuan, giant, gigantesque, 
gigantic, herculean, heroic, huge, immense, jumbo, mammoth, massive, massy, 
mastodonic, mighty, monster, monstrous, monumental, mountainous, 
prodigious, stupendous, titanic, tremendous, vast. Informal : walloping. 
Slang : whopping. See BIG. 

Roget?s II: The New Thesaurus, Third Edition. Copyright ? 1995 by Houghton 
Mifflin Company. Published by the Houghton Mifflin Company. All rights 
reserved.
"""


If you prefer a dictionary:

http://www.bartleby.com/61/25/P0692500.html

"""
ADJECTIVE:
1. Of, relating to, or resembling a python. 2. Of or resembling an oracle; 
prophetic. 3. Of extraordinary size and power. 

   
The American Heritage? Dictionary of the English Language, Fourth Edition. 
Copyright ? 2000 by Houghton Mifflin Company. Published by the Houghton 
Mifflin Company. All rights reserved.
"""


Of course, around here we tend to use a nuance of "Of, relating to,
or resembling Python" rather than "a python", and while "extraordinary
power" is surely sometimes part of the connotations, I don't think
the specific connotations towards BIG SIZE, which the thesaurus points
out so well, are normally implied in our discourse.


Alex




From bokr at oz.net  Tue Apr  1 00:00:01 2003
From: bokr at oz.net (Bengt Richter)
Date: 1 Apr 2003 05:00:01 GMT
Subject: Breaking out of a while loop with a key press?
References: <3e885ad5@mail.hmgcc.gov.uk>
Message-ID: <b6b6ch$75t$0@216.39.172.122>

On Mon, 31 Mar 2003 16:12:20 +0100, "Richard" <richardd at hmgcc.gov.uk> wrote:

>Hi,
>
>Can anyone suggest a simple way of detecting a key press and existing the
>program? I have a program which at the moment uses a while True: statement.
>However this is not ideal as I have to close down the console window to
>exist the program. Can anyone tell me the proper way of doing this? So that
>a CTRL-C, for example, can quit the program correctly?
>
Some time ago I coded this little experiment in interrupting a loop,
including keyboard input echoing the catching of Ctrl-C and Ctrl-Break
and putting off loop exit until the while condition. Perhaps you can
transform it into something serving your purposes. IIRC, if you comment
out the kbhit line, you will get different behavior. I don't recall testing
it on linux. Too lazy right now ;-) Note that the loop ends on the enter key
only if there was a Ctrl-C or Ctrl-Break during the line input and you have
not backspaced over their echoed representation (which should take one backspace
for one control char, even though the echo is multichar).

====< interruptpy.py >================================================
#!/usr/bin/python
# maybe useful part
import signal
class SigHandler:
    def __init__(self):
        self.signaled = 0
        self.sn=None
    
    reset = __init__
    
    def __call__(self, sn, sf):
        self.sn = sn    # latest
        self.signaled += 1
        
def test():
    import msvcrt

    sh = SigHandler()
    old_SIGINT_Handler = signal.signal(signal.SIGINT,sh)
    old_SIGBREAK_Handler = signal.signal(signal.SIGBREAK,sh)
    signames = {
        signal.SIGINT:'SIGINT',
        signal.SIGBREAK:'SIGBREAK'
    }

    # just for this demo, but main loop might be useful
    def puts(s):    # helper
        for c in s: msvcrt.putch(c)
    CTRL_C_ECHO = '<CC>'    # same length as repr, to line up
    SIGINT_ECHO = '<Ctrl-C signal seen>'
    SIGBREAK_ECHO = '<Ctrl-Break signal seen>'

    # this plays the role of your loop body, see loop below           
    def dostuff():
        print 'Enter a line: ',
        line = c = ''
        sigEchoes=[]
        sh.reset()
        while c != '\r':    # Enter key ends input
            # we could break on signals, but we want to finish line, so
            # just echo and record them FTHOI
            if sh.signaled and sh.sn:
                sigEchoes.append((len(line),sh.sn))    # log where
                if sh.sn==signal.SIGBREAK:
                    puts(SIGBREAK_ECHO)
                elif sh.sn==signal.SIGINT:
                    puts(SIGINT_ECHO)
                sh.sn = 0   # reset latest flag
            if not msvcrt.kbhit(): continue # apparently screens ctrl-c?
            c = msvcrt.getch()      # should get ctrl-c but not ctrl-break?
            if c=='\r': break # Enter
            if c=='\b':
                if sigEchoes and len(line)==sigEchoes[-1][0]:
                    # "erase" break effect"
                    puts('\b \b'*len(
                         (SIGINT_ECHO,SIGBREAK_ECHO)[sigEchoes[-1][1]==signal.SIGBREAK]
                    ))
                    sh.signaled -= 1
                    sigEchoes = sigEchoes[:-1]
                    assert sh.signaled == len(sigEchoes)    # how fast can you type ;-)
                    continue
                elif line[-1:]=='\x03': # inc case it's getting to getch
                    puts('\b \b'*len(CTRL_C_ECHO))
                else:    
                    puts('\b \b'*(len(repr(line[-1:]))-2))
                line = line[:-1]
            elif c=='\x03':
                puts(CTRL_C_ECHO)
                line += c     
            else:
                puts(repr(c)[1:-1])
                line += c     
        print '\nIts repr was %s' % `line`
        return sigEchoes
    
    # main loop
    print 'Type Ctrl-C(s) and/or Ctrl-Break(s) anywhere in input line.'
    while not sh.signaled:  # put this condition in your program loop
        
        result = dostuff()           # your stuff goes here
        
    print """\
    This should appear after dostuff() has done everything
    as if not interrupted by signals."""

    print   'Loop was terminated gracefully by %d signal%s\n' \
            '    %s\n' \
            'setting a polled variable (not via exception).' % (
                sh.signaled, 's'[sh.signaled==1:],
                ', '.join(map(lambda x: signames[x], [e[1] for e in result]))
    )

    # restore old signal handlers
    signal.signal(signal.SIGINT,old_SIGINT_Handler)
    signal.signal(signal.SIGBREAK,old_SIGBREAK_Handler)

if __name__ == '__main__':
    test()
======================================================================
Example of running this on windows with python 2.2.2:

[20:55] C:\pywk>interruptpy.py
Type Ctrl-C(s) and/or Ctrl-Break(s) anywhere in input line.
Enter a line: 123<Ctrl-C signal seen>456<Ctrl-C signal seen>789<Ctrl-Break signal seen>abc.
Its repr was '123456789abc.'
    This should appear after dostuff() has done everything
    as if not interrupted by signals.
Loop was terminated gracefully by 3 signals
    SIGINT, SIGINT, SIGBREAK
setting a polled variable (not via exception).

Regards,
Bengt Richter



From claird at lairds.com  Mon Apr 21 14:59:40 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 21 Apr 2003 18:59:40 -0000
Subject: VB curses (was: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?)
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org>
Message-ID: <va8fsscuiqev68@corp.supernews.com>

In article <mailman.1050946774.25112.python-list at python.org>,
Jp Calderone  <exarkun at intarweb.us> wrote:
			.
			.
			.
>instead of just your application.  When you write VB code, you're only
>distributing your application.
			.
			.
			.
Until it turns out that your customer has a VB run-time
which differs subtly but crucially from yours.

At which point you get involved in distribution of 
*larger* files than the corresponding Python-based ones.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From pan-newsreader at thomas-guettler.de  Wed Apr  2 00:57:02 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 02 Apr 2003 07:57:02 +0200
Subject: intro to OOP using Python, tutorial?:newbie
References: <3E8A5AE4.1080902@netscape.net>
Message-ID: <pan.2003.04.02.07.57.02.105639.810@thomas-guettler.de>

On Wed, 02 Apr 2003 05:37:08 +0200, vector wrote:

> Hi all
> is there an introduction to OOp using python tutorial doc,file,pdf etc
> anywhere on the net. something that explains, the interaction
> between,objects,classes,def etc.
> Id prefer something using python examples

Look at chapter 9 of the tutorial:

 http://python.org/doc/current/tut/node11.html

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From Padraig at Linux.ie  Wed Apr  9 05:49:54 2003
From: Padraig at Linux.ie (Padraig at Linux.ie)
Date: Wed, 09 Apr 2003 10:49:54 +0100
Subject: killing child processes
In-Reply-To: <39ee6fa9.0304090148.745fdc46@posting.google.com>
References: <39ee6fa9.0304090148.745fdc46@posting.google.com>
Message-ID: <3E93ECC2.9080207@Linux.ie>

reinhard wrote:
> I've created a process (some batchfile) which in turn creates some
> long running child processes using the win32api CreateProcess call
> (declaring them as a group).
> 
> Is there a way to terminate (kill) all the child processes in the
> group?

yep. just fork() and setpgrp()
Then all processes will have the
same groupid which you can kill(-groupid)
You can look at the code in FSlint:
http://www.pixelbeat.org/fslint/FSlint-2.01.tar.gz

Padraig.




From max at alcyone.com  Tue Apr 22 18:26:53 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 15:26:53 -0700
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <3EA5B173.9420324C@alcyone.com> <eppstein-D065BC.14323022042003@news.service.uci.edu>
Message-ID: <3EA5C1AD.F6207D36@alcyone.com>

David Eppstein wrote:

> 2.2.2, I thought?  They are not in the 2.2 command-line python shipped
> with OS X, forex.

Yeah, you're right.  I always keep synced up with the latest
intermediate version, so I really should have said 2.2.x.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From aleax at aleax.it  Sat Apr 12 15:01:23 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 19:01:23 GMT
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <Z7nla.45$W4.8475@news.uchicago.edu>
Message-ID: <7oZla.5396$LB6.150919@news1.tin.it>

Suchandra Thapa wrote:

> sebastien <s.thuriez at laposte.net> wrote:
>> Is there any PCA analysis tools for python ?
>> If it does, do you have any idea on how well it would scale ?
>> 
>> I have already seen PyClimate (but it is not available for Windows
>> which will be one of the target). Is there some LAPACK like packages ?
> 
> Have you considered using R (www.r-project.com) and RS/Python

Hmmmm -- this site's most succint homepage's in English, but all
links from it seem to be trying to show me lots and lots of Japanese
text.  What's the right way to get R docs & other downloads these
days for the non-Japanese-speaker...?


Alex




From aahz at pythoncraft.com  Fri Apr 18 22:02:34 2003
From: aahz at pythoncraft.com (Aahz)
Date: 18 Apr 2003 22:02:34 -0400
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9FFCB5.79AEA22E@engcorp.com> <b7q72i$nvh$1@slb6.atl.mindspring.net> <roy-416201.21303118042003@reader1.panix.com>
Message-ID: <b7qanq$791$1@panix3.panix.com>

In article <roy-416201.21303118042003 at reader1.panix.com>,
Roy Smith  <roy at panix.com> wrote:
>"Andrew Dalke" <adalke at mindspring.com> wrote:
>>
>> One thing to note, which Dijkstra didn't predict, is that I stopped
>> using a 'traditional' BASIC back in '86, when I got a copy of
>> QuickBasic for my birthday.  No line numbers, real functions,
>> better control structures, compiled code - while editing in the
>> IDE, so the compile-link-run cycle is really fast.
>
>I agree that all those things are good things, but by the time you've 
>made that many changes to the language, what's left that it can still 
>call itself BASIC?

Note that what Andrew describes is much closer to the original Dartmouth
BASIC (which was essentially a simplified FORTRAN) than the abominations
prevalent on the toy PCs of the 1970s and 1980s.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From anti-spam.dbroadwell at mindspring.com  Mon Apr 28 00:06:50 2003
From: anti-spam.dbroadwell at mindspring.com (David Broadwell)
Date: Mon, 28 Apr 2003 04:06:50 GMT
Subject: Newbie: ( if good in self: print "Foobar" )
Message-ID: <uN1ra.15662$kc1.13753@nwrddc02.gnilink.net>

Disclaimer(s):
    I'm a new programmer.
    I'm, a little confuzed.
    I'm working on a 'cut my teeth' program.
    No, it's not any form of classwork.
    This post is long and verbose.

In the course of reading the modules i'm importing to use in my program i've
seen constructions like this:

<code>

class foo:
    bar = ['B','A','R']

    def baz():
        for item in bar: print item
        boggle()
        for item in bar: print item

    def boggle(self):
        self.bar[1] = 'a'
        self.bar[2] = 'r'

</code>

Question 1: Do I read it right that; a classes method called with self can
change the 'global' value without having to type the bloddy global for every
one?!? (ie: the above code not the below code)

<code>

bar = ['B','A','R']

def baz():
    for item in bar: print item
    boggle()
    for item in bar: print item

def boggle():
    global bar
    bar[1] = 'a'
    bar[2] = 'r'

</code>

Question 2: If that is true, what do we call this notation? (anyone got
links?)

Note: In the methods i've read, i have seen the function defined as; def
function (self,arg1,arg2):
I did try to answer the question myself with experimentation, which usually
works, but defining the variable and runctions as they lay and running the
first gave me a traceback (IDLE session paste follows);

>>> bar = ['B','A','R']
>>> def baz():
        for item in bar: print item
        boggle()
        for item in bar: print item
>>> def boggle(self):
        self.bar[1] = 'a'
        self.bar[2] = 'r'
>>> baz()
B
A
R
Traceback (most recent call last):
  File "<pyshell#4>", line 1, in ?
    baz()
  File "<pyshell#2>", line 3, in baz
    boggle()
TypeError: boggle() takes exactly 1 argument (0 given)
>>>

Question 3: So what exactly does my code have to look like to work?
Question 4: Is this notation only valid inside a class?

Brags: I love the cli interface, it's the most usefull thing since my Atari
800 XL (the one with the chipped basic boot prompt that did my homework).

Thanks in advance ...

--

Long live Zork.
    get candle
    light candle
    bask in glow





From fredrik at pythonware.com  Wed Apr 30 09:58:37 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 30 Apr 2003 15:58:37 +0200
Subject: How to call built in function 'open' from inside object that has 'open' method?
References: <db81c492.0304300526.7b8d3c86@posting.google.com>
Message-ID: <b8oko9$12i$1@main.gmane.org>

"vadim" <vf at interna.ru> wrote:
> I am new to Python and do not know the answer to very
> simple how to 'open' from inside another object that
> has 'open' method. I know that in C++ I should use
> following syntax ::open.
>
> More details on reasons. I have met this problem trying to
> write a simple script use ActiveScripting  inside IE.
> I am trying "file = open(...) but because context window
> object has 'open' method that just opens window it is called
> first.

Unless you're typing "file = self.open(...)", the class you're
in has nothing to do with it.

>From your description, chances are that you're using "from
something import *", where "something" defines an "open"
method.  The solution to that is simple: don't do that.

more here:

    http://www.effbot.org/guides/import-confusion.htm

</F>







From hs2l2ATyahoodotcom at yahoo.com  Fri Apr 18 00:11:35 2003
From: hs2l2ATyahoodotcom at yahoo.com (Haran Shivanan)
Date: 18 Apr 2003 04:11:35 GMT
Subject: inheriting variables
References: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de> <mailman.1050586468.29594.python-list@python.org> <rozna.115488$It5.70451@news2.central.cox.net> <b7mn1i$2icfr$2@ID-148859.news.dfncis.de> <LBBna.43583$T34.1210957@news2.tin.it>
Message-ID: <b7nttn$2p743$3@ID-148859.news.dfncis.de>

module "Alex Martelli" core dumped with the following output:
[snipped informative stuff]
> 
> This prints 11, as apparently required by your specs.  I hope
> this can help you make the custom metaclass you need to
> implement your actual, exact specifications, of course.
Thanks *a lot* for your help.
Do you know of any on-line resources where I can get info
on metaclasses and using them?
I was only dimly aware of their presence until your post.
Thanks.


-- 
India and Pakistan may be the next countries to use nukes in war
but no one should forget who was first! "USA!, USA!, USA!"



From the_3_project at yahoo.com  Tue Apr 29 01:56:31 2003
From: the_3_project at yahoo.com (Joe Green)
Date: 28 Apr 2003 22:56:31 -0700
Subject: _winreg newbie question
Message-ID: <a5b86892.0304282156.208b81b1@posting.google.com>

I'm having trouble understanding why _winreg.OpenRegistry() does not
want to open the registry in a class.

An example is something like


from _winreg import *

s = 'HKEY_LOCAL_MACHINE'

if (s is 'HKEY_LOCAL_MACHINE'):
  reg = ConnectRegistry(None, HKEY_LOCAL_MACHINE)
else:
  print "Somehow that didn't work"

That seems to work fine. But when it's inside a class


class myclass:

  def __init__(self, key, cname=None):
    if (cname is 'HKEY_LOCAL_MACHINE'):
      self.reg = ConnectRegistry(cname, HKEY_LOCAL_MACHINE)
    else:
      print "Not working"


it doesn't work. self.reg never points to anything.

Any ideas?



From student00 at angelfire.com  Wed Apr  2 10:37:40 2003
From: student00 at angelfire.com (student)
Date: 2 Apr 2003 07:37:40 -0800
Subject: simulating keyboard events
References: <abfbd6c5.0303301449.655e8756@posting.google.com> <7h31y0nn03m.fsf@pc150.maths.bris.ac.uk>
Message-ID: <abfbd6c5.0304020737.65946e91@posting.google.com>

sorry .... for now i'm developing under windows 2000, although it will
need to work on linux/unix machines in the future

Michael Hudson <mwh at python.net> wrote in message 
> > is there an easy way to simulate a keyboard event in python..
> 
> I doubt it, but I don't actually know...
> 
> > i need to send a 'Delete' key event to a text widget (at the current
> > cursor position/selection)
> 
> But I do know that this will be *very* platform dependent, so drop us
> a clue...
> 
> Cheers,
> M.



From tim_one at email.msn.com  Fri Apr 18 00:36:19 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Fri, 18 Apr 2003 00:36:19 -0400
Subject: Mem "leak" w/ long-running network apps?
In-Reply-To: <Pine.LNX.4.51.0304171941001.22199@ns0.tbye.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCIECLEHAB.tim_one@email.msn.com>

[Dave Brueck]
> I've spent the last few weeks trying to track down a memory leak in a
> long-running network app (an HTTP proxy for large-ish objects) and am
> trying to find out if anyone else has encountered something similar.

Overall, I've spent months of my life tracking down memory leaks in apps
that run for such a short time I can't really conceive of it <wink>.

> I narrowed the problem down to the receiving side of the proxy, and today
> in socketmodule.c I noticed that when you call sock.recv(n) a string of
> size n is created and then resized after the recv to the actual size of
> the data received. There's nothing wrong with that in and of itself, but
> when I replaced it with the function below to do receives the ever-growing
> memory problem went away (so now instead of calling sock.recv(amnt) I do a
> specialmodule.recv(sock, amnt)). The function merely uses a static buffer
> and then creates a Python string of just the size needed - the resize is
> what got eliminated. I initially encountered the problem on Python 2.1.3
> but then moved to 2.2.2 to use the gc module enhancements.
>
> My questions are:
>
> (1) has anybody else run into a similar problem before?

Oh sure.  It can pop up lots of ways.  The ultimate culprit is usually the
platform's malloc implementation.  It should be better under Python 2.3,
because the new release moves heaven and hell to keep the platform malloc
away from managing memory for small objects, which often delays their "Hey,
I've gone insane!" points beyond visible experience.

> This may not be very common because the app is a little unusual in that
> it's long-running, handles hundreds of concurrent connections, each
> connection is usually for a large (tens to hundreds of megabytes) object,
> the data is all proxied rather than being served off disk or generated by
> the app, and both the upstream and downstream connections are generally
> fast (aggregate throughput for the server is usually in the 100-200 Mbps
> range for a P3 900.
>
> (2) Any ideas on why using the normal socket.recv resulted in ever-growing
> memory use? I've spent ages using the gc module and other tools to track
> down objects that should have been freed, cyclic references, etc., and
> don't see any problems there and I don't think the memory is really being
> leaked (in the C sense). Could it be a heavily fragmented heap or
> something like that?

Almost certainly, yes.  You didn't say what "the typical" value for n is,
but since you made your static buffer 1MB, I assume it can be up to that.
Python mallocs that much, then reallocs back to the actual size.  Returns
the object to your program.  Your program then probably allocates a fair
number of tiny objects.  A less-than-stellar malloc can chop those off "the
high end" of the memory realloc returned, leaving a hole in the address
space that isn't big enough to hold the next call to recv.  So the next
block of n comes after the last block, etc -- iterate to disaster.  The used
part of the address space ends up looking like sparse Swiss cheese.

Python 2.3 won't make your platform malloc smarter, but it grabs large
chunks and carves them up itself for small objects.  While it wasn't the
intent, this has the pleasant side effect of making some platform mallocs
much better behaved across the large allocations Python still asks them to
do directly.  It's really difficult to write a malloc that handles a wild
mix of very small and very large requests gracefully; 2.3 helps by taking
the small requests out of the platform malloc's hair.

> I'd notice that after my tests ran for a long time I'd stop them and
> memory usage would drop down after awhile by a few megabytes, and upon
> starting my tests again (without restarting my app) mem usage would drop
> down some more but not all the way down and then it would gradually grow
> again to a new high-water mark, so that overnight my process had hundreds
> of MB of RAM.  With my recv-replacement I'm holding steady at about
> 30 MB, which is normal.

If you want to know more, scour your platform docs for whatever debug info
your local malloc can deliver.

> (3) Does anyone see any glaring errors in my function below? Seems to work
> well enough. :)
>
> Anyway, I don't think there's a bug in Python, and my function is
> certainly not patch-worthy because it really works just for my use, and
> now that my problem is gone it's mostly out of curiosity that I'm trying
> to better understand why my problem is gone, but I'd appreciate any
> insight or hints.
>
> Thanks,
> Dave
>
> Here's the function. I can get away with using a single large buffer for
> all my receives because while the server is a mixture of threading and
> poll-based, all the I/O happens sequentially in one thread against sockets
> that are known to be ready for I/O.

Heh.  That sounds regrettable, eventually <wink>.

> #define MAX_RECV_SIZE 1048576
> static char RECV_BUFF[MAX_RECV_SIZE];
>
> static PyObject *
> recv_wrapper(PyObject *self, PyObject *args)
> {
>     PyObject *py_sock;
>     PyObject *py_str;
>     int sock;              /* Output socket */
>     int len, n;
>
>     if (!PyArg_ParseTuple(args, "Oi:recv", &py_sock, &len))
>             return NULL;
>
>     if (len < 0)
>     {
>       PyErr_SetString(PyExc_ValueError, "negative buffersize in recv");
>       return NULL;
>     }
>
>     if (len > MAX_RECV_SIZE-1)

The "-1" is mysterious here.  If you've got N bytes allocated, why can't you
receive N bytes?

>     {
>       PyErr_SetString(PyExc_ValueError, "buffersize too large in recv");
>       return NULL;
>     }
>
>     sock = PyObject_AsFileDescriptor(py_sock);

Should follow this by

      if (sock < 0)
          return NULL;

although that assumes legit filedescs can't be negative.  Anally correct
would be

      if (sock == -1 && PyErr_Occurred())
          return NULL;

>     Py_BEGIN_ALLOW_THREADS;
>     n = recv(sock, RECV_BUFF, len, 0);
>     Py_END_ALLOW_THREADS;
>     if (n == -1)
>       return PyErr_SetFromErrno(RecvError);
>
>     py_str = PyString_FromStringAndSize(RECV_BUFF, n);
>     if (py_str == NULL)
>       return NULL;
>
>     return py_str;

The last 5 lines could be replaced by

    return PyString_FromStringAndSize(RECV_BUFF, n);





From spam at magnetic-ink.dk  Tue Apr 22 10:57:40 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Tue, 22 Apr 2003 14:57:40 +0000 (UTC)
Subject: Zope TAL
References: <b83j25$3namt$1@hades.rz.uni-saarland.de>
Message-ID: <eefb3157-8e70-4fee-8494-024efe9c3a4c@news.szn.dk>

Uwe Schmitt wrote:

> are there any good resources concerning TAL ?
> I just found a reference on zope.org but nothing more.

How about <http://www.owlfish.com/software/simpleTAL/>?


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From ckea25d02 at sneakemail.com  Wed Apr  2 20:02:32 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 03 Apr 2003 13:02:32 +1200
Subject: A question about help() and docstrings
In-Reply-To: <mailman.1049293506.17715.python-list@python.org>
References: <mailman.1049293506.17715.python-list@python.org>
Message-ID: <b6g0nc$518n3$2@ID-169208.news.dfncis.de>

Richard Kuhns wrote:
>>>>from dummy import *
>>>>help('dummy_function')
> 
> no Python documentation found for 'dummy_function'

Try this instead:

   help(dummy_function)

i.e. pass it the function *object* itself, rather than
the *name* of the function.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From duncan at NOSPAMrcp.co.uk  Mon Apr 28 05:52:57 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Mon, 28 Apr 2003 09:52:57 +0000 (UTC)
Subject: dynamic function names
References: <20030428102400.21c5074a.bram@nospam.sara.nl> <iP5ra.14084$3M4.382689@news1.tin.it>
Message-ID: <Xns936B6D56043E4duncanrcpcouk@127.0.0.1>

Alex Martelli <aleax at aleax.it> wrote in 
news:iP5ra.14084$3M4.382689 at news1.tin.it:

> See?  "def funcname():" *RE-BINDS* identifier funcname to
> refer to the function object.
> 
> So, what you need is to call function 'function' in module
> 'new'.  You can first build a temporary function with a def
> to give you the right code, globals, defaults and closure,
> and then pass each of these to new.function together with
> the funcname you want to use:
> 
>>>> import new
>>>> def makefunc(funcname):
> ...   def temp(): print "this function is named", funcname
> ...   f = new.function(temp.func_code, globals(), funcname,
> ...           temp.func_defaults, temp.func_closure)
> ...   return f
> ...
>>>> f = makefunc('foo')
>>>> f()
> this function is named foo
>>>> f.func_name
> 'foo'
>>>>
> 

Note that this code only works on Python 2.3 (because the closure argument 
to new.function didn't exist earlier), and even there if the function 'foo' 
throws an exception the traceback will still call it 'temp'.


-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From hgg9140 at cola2.ca.boeing.com  Thu Apr 24 10:09:19 2003
From: hgg9140 at cola2.ca.boeing.com (Harry George)
Date: Thu, 24 Apr 2003 14:09:19 GMT
Subject: ascii txt to LaTeX
References: <3ea7085a$0$1032$5402220f@news.sunrise.ch>
Message-ID: <xqxadeg7zsg.fsf@cola2.ca.boeing.com>

marco <marco.rossini at gmx.ch> writes:

> i wrote a python program for converting an word wrapped(!) ascii text into 
> LaTeX format. it actually doesn't do much, just finds(!) and places 
> paragraphs (\n\n), replaces apostrophes and umlauts. it does routine stuff: 
> why do boring things myself if my computer can do them for me?
> 
> feel free to use it (GPL). also i would appreciate critique on my code.
> 
> probably someone has made such (or: a better) program before, but i don't 
> care, it was easy for me to do and it's all _I_ need.
> 
>         marco
> 
> #!/usr/bin/python
> 
> # Converts a regular word-wrapped ascii text into formated LaTeX
> # detects paragraphs, replaces apostrophes, replaces umlauts
> 
> # Program written by Marco Rossini <marco.rossini at gmx.ch>
> # Copyright: GPL
> 
> from sys import argv
> from sys import exit
> from math import ceil
> from string import find
> from string import join
> from string import strip
> from string import replace
> from string import whitespace
> from string import punctuation
> 
> if len(argv) != 2: exit("txt2latex: Argument error!")
> 
> try: f = file(argv[1],"r")
> except: exit("txt2latex: File not found!")
> 
> 
> # GET THE MAXIMAL NUMBER OF CHARACTERS PER LINE
> array = f.readlines()
> linelength = 0
> for i in range(len(array)):
>     array[i] = strip(array[i])
>     if len(array[i]) > linelength: linelength = len(array[i])
> 
> # GUESS IF IT'S A PARAGRAPH BREAK
> for i in range(len(array)-1):
>     nleft = linelength - len(array[i])
>     nright = find(array[i+1]," ")
>     if nright == -1: nright = len(array[i+1])
>     # if it is, append \n\n to the line, else a space
>     if nright+1 <= nleft:
>         if len(array[i]) > 0: array[i] += '\n\n'
>     else:
>         array[i] += ' '
> 
> # the lines are joined, the text is NOT word wrapped anymore
> text = join(array,"")
> 
> # Replace apostrophes intelligent(ly?)
> i = 0;
> while i < len(text):
>     # ... before a word
>     if text[i] == '\"' and find(whitespace,text[i-1]) >= 0:
>         text = text[:i] + "``" + text[i+1:]
>         i += 1
>     # ... before a word (single)
>     if text[i] == '\'' and find(whitespace,text[i-1]) >= 0:
>         text = text[:i] + "`" + text[i+1:]
>     # ... after a word, no punctuation
>     if text[i] == '\"' and find(whitespace,text[i+1]) >= 0:
>         text = text[:i] + "''" + text[i+1:]
>         i += 1
>     if text[i] == '\"' and find(punctuation,text[i+1])>= 0:
>         text = text[:i] + "''" + text[i+1:]
>         i += 1
>     i += 1
> 
> # Here replacements for umlauts. modify if you like to.
> text = replace(text,"?","\\\"{a}") -A
> text = replace(text,"?","\\\"{A}") -A
> text = replace(text,"?","\\\"{o}") -A
> text = replace(text,"?","\\\"{O}") -A
> text = replace(text,"?","\\\"{u}") -A
> text = replace(text,"?","\\\"{U}") -A
> 
> # handle output
> print text
> 

See also:
SDF http://www.xpenguin.biz/download/sdf/doc/

POD http://www.perldoc.com/perl5.6/pod/perlpod.html

PDX http://www.seanet.com/~hgg9140/comp/pdx-1.6.1/doc/index.html
    (my own shot at this)



-- 
harry.g.george at boeing.com
6-6M31 Knowledge Management
Phone: (425) 294-8757



From staschuk at telusplanet.net  Sun Apr 20 02:31:14 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 20 Apr 2003 00:31:14 -0600
Subject: When not to use an RE -- an example
In-Reply-To: <3ea1e522.18284281@news.lexicon.net>; from sjmachin@lexicon.net on Sun, Apr 20, 2003 at 12:15:19AM +0000
References: <c76ff6fc.0304191459.2adf73a@posting.google.com> <yfs7k9qjbmx.fsf@black132.ex.ac.uk> <3ea1e522.18284281@news.lexicon.net>
Message-ID: <20030420003114.B8441@tibia.amotlpaa.bogus>

Quoth John Machin:
  [...]
> I'll get back into my box now. :-)

<welsh>You have it easy.  When I was your age, we didn't have
*boxes*...</welsh>

-- 
Steven Taschuk                               staschuk at telusplanet.net
"[T]rue greatness is when your name is like ampere, watt, and fourier
 -- when it's spelled with a lower case letter."      -- R.W. Hamming




From rmunn at pobox.com  Sun Apr 13 01:16:08 2003
From: rmunn at pobox.com (Robin Munn)
Date: Sun, 13 Apr 2003 05:16:08 GMT
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050201322.13220.python-list@python.org>
Message-ID: <slrnb9h7u4.it3.rmunn@localhost.localdomain>

Skip Montanaro <skip at pobox.com> wrote:
> Recent versions of Python (2.2 and later) allow this:
> 
>     file = file("somefile")
>     for line in file:
>         process(line)

Ewww! Don't *do* that! You just shadowed the "file" builtin. Use a
different name for your file object -- I prefer "input_file":

    input_file = file("somefile")
    for line in input_file:
        process(line)

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From just at xs4all.nl  Fri Apr  4 06:11:50 2003
From: just at xs4all.nl (Just)
Date: Fri, 04 Apr 2003 13:11:50 +0200
Subject: Severe c-module problem
References: <b6jo6v$715$1@nets3.rz.RWTH-Aachen.DE>
Message-ID: <just-1346A6.13114904042003@news1.news.xs4all.nl>

In article <b6jo6v$715$1 at nets3.rz.RWTH-Aachen.DE>,
 Rolf Wester <wester at ilt.fhg.de> wrote:

> Hi,
> 
> I have a c-Module "pyopt" with a function:
> 
> static PyObject *_wrap_c_nlo_init_Hs(PyObject *self, PyObject *args) {
>      PyObject *resultobj;
>      char * arg0;
> 
>      cout << "in init_Hs" << endl;
> 
>      if(!PyArg_ParseTuple(args,(char *)"ss:c_nlo_init_Hs", &arg0))
>              return NULL;

You're asking for two strings ("ss:...") yet you only give one receiving 
argument (arg0), so this will overwrite some arbitrary memory and crash.

Just



From poec at yahoo.com  Sat Apr 19 13:38:40 2003
From: poec at yahoo.com (Ovid)
Date: 19 Apr 2003 10:38:40 -0700
Subject: Testing methods
Message-ID: <9d8ac130.0304190938.755d4b4c@posting.google.com>

Hello,

Relatively new to Python and I thought I would teach myself the
language by building a test harness.  After I'm done, I will probably
be throwing it away in favor of Pyunit, but it seemed like a good way
to learn some of the fiddly bits.

Fiddly bits such as "can this object do that?"  In my test code, I
want to test whether or not a given class *or* instance of that class
can execute a particular method (i.e., if it has or inherits the
method).  For example, one way (in Perl), to test if object "$foo" can
execute an arbitrary method would be:

  if ( $foo->can($some_method) )

For the life of me, I can't figure out the equivalent in Python.  I
suppose I could do:

  if some_method in Foo.__dict__:

But that doesn't tell me whether or not "some_method" is a method. 
Further, if I have an instance of a class, that doesn't seem to work. 
I've been searching through the docs, but I can't quite figure this
out.

I'm a clueless newbie here, so please be kind :)

Cheers,
Ovid



From pobrien at orbtech.com  Sun Apr  6 16:37:53 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 15:37:53 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049658250.30172.python-list@python.org>
Message-ID: <m2fzov8iqm.fsf@localhost.localdomain>

Francois Pinard <pinard at iro.umontreal.ca> writes:

> [Patrick K. O'Brien]
> 
> > If you were crazy enough to think that you could create a program
> > along the lines of Emacs, but written in Python, how would you go
> > about it?
> 
> Hello, Patrick, and all pythoneers. :-)
> 
> I once was crazy enough to at least ponder doing something like
> this, being conscious of the huge amount of work such a project
> might represent.  I did starve for an Emacs environment scriptable
> in Python, but also was aware of the massive amount of knowledge in
> the Emacs distribution and project, and the rapid pace of Emacs
> development.  This would have been fairly difficult to follow and
> cope with in the long run.  No doubt to me that seriously tackling
> such a project could have swallowed me not only whole, but for a
> good part of my life.  I still felt too young to die! :-)

I expect to have a reasonable proximity within a year.  I expect to
have something I'm willing to put up with as my primary editor within
two to three weeks.  I expect that what I'm willing to put up with
will be seen as something wonderful by the Windows and Apple folks who
would never want to use Emacs and instead use stuff like Notepad (been
there, done that).  Remember, since I'm using wxPython this will run
on Linux, Windows, and Mac OS X.

> Pymacs seemed to me like a much more acceptable compromise, as it
> could be written relatively quickly, giving me a good part of all I
> wanted.  Oh, I'm surely not satisfied with the complexity of the
> mix, Emacs plus Python just cannot be simpler than a pure Python
> solution would have been, at least seen from a reasonable distance
> -- another planet, say? :-).  And moreover, there is a cost to the
> communication protocol between Emacs and Python which is better not
> be abused if you want good response times.

Do you think Pymacs would help me understand Emacs better, or should I
just dive into the Emacs source?

> > Sound like fun?
> 
> In a word, I would love to see a pure Python implementation of
> Emacs, but have the prejudice that it requires an extra-ordinary
> involvement, making that goal unreachable in practice for most
> mortal programmers...  So, even if Pymacs is far from ideal, for me,
> it has been a blessing.

I think getting the core functionality won't be that hard.  And even
if I only implement the bare minimum, the results should be better
than most other editors, aside from Emacs and Vi and whatever editor
anyone reading this likes... (please, let's not start any editor wars)

> All this being said, I sincerely wish that your project will succeed!

Me too.  Ignorance is bliss, so I might as well enjoy this moment when
it all seems possible.  If I knew then what I know now, I might never
have tried to create PyCrust either.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From jordan at krushen.com  Wed Apr 30 23:54:38 2003
From: jordan at krushen.com (Jordan Krushen)
Date: Thu, 01 May 2003 03:54:38 GMT
Subject: syslogd (sorta)
References: <z80sa.47803$ey1.4455255@newsread1.prod.itd.earthlink.net>
Message-ID: <oprog6hoqe5ctagx@shawnews>

On Thu, 01 May 2003 03:03:59 GMT, Cecil H. Whitley <cwhitley at earthlink.net> 
wrote:

> I am trying to write a syslogd server in python.  How can I "end" the 
> main thread and have a daemonic thread continue?

You can't.  From the threading library docs:

The entire Python program exits when no active non-daemon threads are left.

Thus, if you have a main, non-daemon thread, and any number of daemonized 
ones, as soon as the main thread ceases activity, they all die.

> thread = threading.Thread(target=syslogd, name=syslogd, args=[], 
> kwargs={})
> thread.isDaemon(True)
> thread.start()

If you want to daemonize the thread, use thread.setDaemon(), not isDaemon() 
.  See below, however.

> Everything works just fine, up until I hit return of course.  So where
> should I be looking in order to write a "clean" background process?

Assuming that your isDaemon() line there is a typo (you'll get an error 
with the above) and that you're actually calling setDaemon(), just comment 
out that line.  The background thread will be non-daemonized, and will 
consequently keep the program alive.

As an example, the following will run until the waiting() function finishes 
(daemonizing it only lets it print "waiting" the first time, as the main 
thread finishes right after start(), terminating the program:

import threading, time

def waiting():
    for i in range(3):
        print "waiting"
        time.sleep(1)

t = threading.Thread(target=waiting)
t.start()

HTH,

J.



From ulope at gmx.de  Fri Apr 11 17:43:29 2003
From: ulope at gmx.de (Ulrich Petri)
Date: Fri, 11 Apr 2003 23:43:29 +0200
Subject: module for doing unix "find"-like things?
References: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com>
Message-ID: <b77cu2$buftr$1@ID-67890.news.dfncis.de>

"Mark Harrison" <mh at pixar.com> schrieb im Newsbeitrag
news:7LDla.1483$Pq2.135830754 at newssvr14.news.prodigy.com...
> Is there such a thing?  I didn't see such a thing, but
> "find" is not the easist thing to search for.
>
> Specifically, I would like to replace an external call to
>     for i in `find $top -type l -maxdepth 1`; ...
>

You can try using os.listdir in a recursive function...


Ciao Ulrich





From aahz at pythoncraft.com  Tue Apr 29 12:22:14 2003
From: aahz at pythoncraft.com (Aahz)
Date: 29 Apr 2003 12:22:14 -0400
Subject: Problem subclassing tuple
References: <mailman.1051631814.18940.python-list@python.org>
Message-ID: <b8m8rm$sef$1@panix3.panix.com>

In article <mailman.1051631814.18940.python-list at python.org>,
John Wilson <tug at wilson.co.uk> wrote:
>
>class Holder(tuple):
>    def __init__(self, *values):
>        tuple.__init__(self, values)
>
>When I try to create an instance:
>
>Holder(1, 2)
>
>I get:
>TypeError: tuple() takes at most 1 argument (2 given)

Use __new__ instead of __init__.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From peter at engcorp.com  Thu Apr 17 13:04:05 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 17 Apr 2003 13:04:05 -0400
Subject: FInd nic speed/duplex (Win32)
References: <Xns9360BC0A648C4spamfilter2muppnet@195.67.237.51>
Message-ID: <3E9EDE85.88D5C0BD@engcorp.com>

j2 wrote:
> 
> Is there any sample code to be able to find NIC settings (speed and duplex)
> for a NIC in a Win98 system?

You could look in the registry, or there might be some Win32 call.
Either way, the bulk of the answer is Windows-specific and could
be found by searching with Google, probably.  After you have the
answer in "Windows-speak", post the APIs required, or the names
of the registry keys required, and someone can post example code.

Or, if you're really lucky, you won't have to do *any* work yourself
and someone else will just post exactly what you need.  I wouldn't
waste time waiting, though, if I were you....

-Peter



From maney at pobox.com  Sat Apr 12 23:13:28 2003
From: maney at pobox.com (Martin Maney)
Date: Sun, 13 Apr 2003 03:13:28 +0000 (UTC)
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org>
Message-ID: <b7akko$5kr$1@wheel2.two14.net>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> wrote:
> Floats are peculiar creatures.

Best description I've ever seen was this classic: floats are like a
pile of sand; every time you move on you lose a little sand and pickup
a little dirt.



From tim.one at comcast.net  Tue Apr  1 22:53:57 2003
From: tim.one at comcast.net (Tim Peters)
Date: Tue, 01 Apr 2003 22:53:57 -0500
Subject: else: w/o except: - why not?
In-Reply-To: <v8j9pb3v2cnra6@corp.supernews.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCEEFCECAB.tim.one@comcast.net>

[Cameron Laird]
> I'm leaving aside Mr. Hand's goals; I'm still unsure about them.
>
> I want to be clear about the judgment that "a bare except is bad
> practice".  Are you talking about any unqualified except?

It's the semantics:  *most* bare excepts swallow all exceptions blindly, and
that's almost never what the author intended.  When it is, it should be
commented.  A syntactic bare except may have unobjectionable semantics, as
in your second example below:

> Is
>   try:
>       f1()
>   except ZeroDivisionError:
>       f2()
>   except:
>       f3()
> an instance of "a bare except" in the sense that you intend here?

Semantically, yes.  It's almost certain that the author did not intend that
(among others):

1. The failure to *find* the name f1 would call f3(), via "f1" raising
   NameError or UnboundLocalError.  The bare except here is a fine way to
   change a typo into mysterious behavior.

2. Trying to terminate the program via Ctrl-C would lead to Ctrl-C getting
   ignored, calling f3() instead, due to the bare except swallowing the
   KeyboardInterrupt exception.

3. Trying to terminate the program via sys.exit() would get ignored, due to
   the bare except swallowing the SystemExit exception.

> And, for you, does
>   try:
>       f1()
>   except:
>       f2()
>       raise
> become *not* bad practice?

Since it doesn't swallow any exceptions, this example is immune to the
surprises of the first example.  It's still questionable whether the author
*really* intended to catch every possible exception coming out of f1(),
though, and I'd be happier with a comment about why this code is appropriate
in context.  For example, perhaps f2() logs the pending exception state.

bad-practice-gets-worse-with-practice-ly y'rs  - tim





From h_schneider at marketmix.com  Fri Apr 25 13:05:31 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Fri, 25 Apr 2003 19:05:31 +0200
Subject: FreeBSD: Python exits on signal 11
Message-ID: <b8bpst$43q$03$1@news.t-online.com>

I get this error on FreeBSD5 with Python 2.2.2, when I run an application
that does its job without errors on NT and Linux platform. It uses threads
and sockets as well (xqmlrpclib). It just stops after a while, displays this
error and does a core dump.

Any hints ?

All the best,
Harald





From Dialtone.NOSPAM.ME at aruba.it  Tue Apr 15 18:47:11 2003
From: Dialtone.NOSPAM.ME at aruba.it (Dialtone)
Date: Wed, 16 Apr 2003 00:47:11 +0200
Subject: Any good donothing?
Message-ID: <slrnb9p2vf.t3.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix>

Hi all. I'm working on a kind of all-in-one cd2mp3 encoder based on lame
and cdda2wav and pygame.

A little code snippet is the following (it is already indented once
because it is part of a bigger class, not so bigger, named cd2mp3).
My problem is to sync the 2 processes which are launched with the
os.system() call but without having the main program locked.
I initially thought that a wait inside a cicle would have got me an easy
solution but it didn't since it locks the main program anyway... Is
there any other way to make a do_nothing_for_some_time(n) function? Or
otherwise... is there another good way to sync those 2 processes without
having the main program locked (and also without having the encoder to
finish ecoding before the ripper has finished).  thanks a lot.

    def start(self, cdda=1, input_file=None, output_file=None, path="./library/audio/", bitrate=128):
#        mkdir_cmd = "mkdir %s%s" % (path, self.artist)
 #       os.system(mkdir_cmd)
  #      path = path + self.artist + "/"
        
        for i in xrange(self.start_track, self.num_tracks):
            try:
                self.infos = "Ripping: %s" % (self.tracks[i])
                self.infos2 = "Encoding: %s" % (self.tracks[i])
            except:
                self.infos = "Ripping track num %d" % (i)
                self.infos2 = "Encoding track num %d" % (i)
                
            self.surface.blit(self.text_bg, self.text_rect)
            pygame.display.flip()
            self.rendered_text = pygame.font.Font(None, 20).render(self.infos, 1, [255, 255, 255])
            self.text_surf.blit(self.rendered_text, (0,0))
            pygame.display.flip()
           
            # If tracks' names found
            if self.succeded:
                output_file = 1
                
            self.cdda2wav_string, self.wav2mp3_string = self.prefs(i, input_file, output_file, path, bitrate)
            if cdda:
                self.wav = os.system(self.cdda2wav_string)
                self.lose_time(20)
                self.progress_bar.inc(self.perc_incr)
                self.text_surf.blit(self.rendered_text, (0,0))
                pygame.display.flip()
            
            self.surface.blit(self.text_bg, self.text_rect)
            pygame.display.flip()
            self.rendered_text = pygame.font.Font(None, 20).render(self.infos2, 1, [255,255,255])
            self.text_surf.blit(self.rendered_text, (0,0))
            pygame.display.flip()
            
            self.mp3 = os.system(self.wav2mp3_string)
            self.lose_time(40)
            self.progress_bar.inc(self.perc_incr)
            self.clean = os.system('rm -rf *.inf && rm -rf *.wav')
        text = "%d track(s) successfully encoded" % (self.num_tracks-self.start_track)
        self.progress_bar.stop(text)
        
    def prefs(self, cur_track, input_file, output_file, path, bitrate=128, dev='/dev/cdrom'):
        """
        settare le preferenze tipo 
        bitrate della conversione etc..
        """
        if not output_file:
            output_file = "audio_%d.mp3" % (cur_track)
        else: output_file = "%s.mp3" % (self.__prepare(self.tracks[cur_track]))

        if not input_file:
            input_file = "audio.wav"
        
        cdda2wav_string = "cdda2wav -D %s -t %d+%d &" % (dev, cur_track, cur_track)
        wav2mp3_string = "lame -b %d %s %s%s &" % (bitrate, input_file, path, output_file)

        return cdda2wav_string, wav2mp3_string

    def lose_time(self, seconds):
        for i in xrange(seconds):
            pygame.time.wait(1000)


-- 
try: troll.uses(Brain)
except TypeError, data: 
   troll.plonk()
Co-Responsabile Hardware Information & Technology http://hit.edengames.net



From trash at kjn-law.de  Mon Apr 28 09:41:43 2003
From: trash at kjn-law.de (Theodor Rash)
Date: Mon, 28 Apr 2003 15:41:43 +0200
Subject: Problems installing the interpreter
Message-ID: <olnvn-ld5.ln1@newsgate.kjn.lan>

Hi,

I am totally new to Python and ran into troubles already by trying to
install the interpreter. I unpacked the tarball (2.3a-2), configured and
then ran 'make'. Obviously the interpreter itself is build and started
automatically to execute setup.py and then stops with an error:
'Platform-independent libraries are not found'. In addition, traceback
shows that import of "os" failed 'cause it was not found.
It gives me the recommandation to set $PYTHONHOME to <prefix> but that
doesn't make no difference.
What am I doing wrong? (system is SuSE 7.3)
Theo 



From max at alcyone.com  Thu Apr 17 03:13:09 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 17 Apr 2003 00:13:09 -0700
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net>
		 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
		 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
		 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
		 <bDSma.22753$T34.687158@news2.tin.it> <slrnb9o8s8.iea.news@datatailors.com>
		 <mailman.1050427771.31620.python-list@python.org>
		 <3E9E0FBA.6050503@indy.rr.com> <mailman.1050561571.26954.python-list@python.org>
Message-ID: <3E9E5405.F28B8B6D@alcyone.com>

Ian Bicking wrote:

> It's also a question of what you're trying to teach.  Something that
> has
> always impressed me with Logo is that it's more of a teaching
> philosophy, and a language goes with it.  The intention behind Logo
> isn't to teach programming... rather, programming is a medium in which
> to teach all sorts of other great stuff.

And, I might point out, Logo really gets a bad rap by the general
populace as a "toy" language that just involves little turtles running
around like silly things.  In actuality, Logo was originally designed
for abstract manipulation of symbols (_logo_ means "word" in Greek); the
turtle graphics stuff came later.  Logo is in fact a fun little cousin
of the Lisp family.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Can I walk with you / 'Till the day that my heart stops beating
\__/ India Arie
    Polly Wanna Cracka? / http://www.pollywannacracka.com/
 The Internet resource for interracial relationships.



From fredrik at pythonware.com  Wed Apr 30 01:49:30 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 30 Apr 2003 07:49:30 +0200
Subject: win32 display jpg file
References: <irxra.10972$pK2.14342@news.indigo.ie> <RuDra.7018$lD4.74951@news-server.bigpond.net.au>
Message-ID: <b8no38$9g2$2@main.gmane.org>

Mark Hammond wrote:

> > Does anybody knows how to display a JPG file in pythonwin and draw lines
> > over it?
>
> With a fair bit of work :)  You would need to use PIL to decode the jpg
> and convert it to a bitmap.  IIRC, Pythonwin still has a PIL demo in it.

if you cannot find it, here's an outline:

when you load the image:

    from PIL import Image, ImageWin

    try:
        im = Image.open(filename)
    except IOError:
        # deal with the error
    else:
        # check image size; resize if necessary
        # check the image mode
        if im.mode not in ("1", "L", "RGB"):
            im = im.convert(Image.getmodebase(im.mode))
        dib = ImageWin.Dib(im)
        # store the "dib" variable somewhere

in the PAINT handler, just do:

    dib.expose(dc)

where "dc" is a device context for the window, cast to an integer.
according to the PIL docs, "you can use the GetHandleAttrib method
of the CDC class to get a suitable handle."

in 1.1.4, you can pass in ImageWin.HWND(wnd) instead of the dc,
where wnd is a window handle cast to an integer.

in all versions, you can also use

    dib.draw(dc, (x0, y0, x1, y1))

where (x0, y0, x1, y1) is the coordinates for where you want the
image to appear.

(expose is the same thing as draw(dc, (0, 0, width, height))

</F>







From ckea25d02 at sneakemail.com  Wed Apr  2 19:01:27 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 03 Apr 2003 12:01:27 +1200
Subject: Don't Panic! (Re: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...))
In-Reply-To: <3E8B48E0.C786D192@engcorp.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com> <mailman.1049298845.7655.python-list@python.org> <3E8B4094.F61515B3@engcorp.com> <ZpHia.23$Sd.4@fed1read01> <3E8B48E0.C786D192@engcorp.com>
Message-ID: <b6ft4r$4cmat$1@ID-169208.news.dfncis.de>

Peter Hansen wrote:
> Hmm.... that is exactly what is needed to make this effective in 
> unit testing.  One imports the module under test, then fakes the
> open method in that other module:

It seems fairly certain that, if this optimisation becomes
a reality, there will be some way of turning it off, perhaps
a command line option. So there's no need to panic just yet.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From duncan at NOSPAMrcp.co.uk  Thu Apr 24 10:30:03 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 14:30:03 +0000 (UTC)
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com> <Xns9367885737E62duncanrcpcouk@127.0.0.1> <EARpa.120$3M4.2728@news1.tin.it>
Message-ID: <Xns93679D1BF4A2Cduncanrcpcouk@127.0.0.1>

Alex Martelli <aleax at aleax.it> wrote in 
news:EARpa.120$3M4.2728 at news1.tin.it:

> ..isn't this smoother and handier?  Or does it just look that way
> to me, due to overexposure to custom metaclasses?-)
> 
Yes, I agree that it is better, but I have one niggling worry about this:

Your metaclass works fine for the one requirement given here, and other 
metaclasses you have proposed work fine in their place, but what happens if 
you want to combine these and get e.g. a Bunch which counts its instances?

I haven't tried it to see what happens, maybe you just need to indirect 
type.__new__ calls through super, and make a new submetaclass with the old 
metaclasses as bases, but I'd like to know if there are issues here or if 
I'm just worrying needlessly.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From duncan at NOSPAMrcp.co.uk  Thu Apr 10 11:07:25 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 10 Apr 2003 15:07:25 +0000 (UTC)
Subject: BEGINNER: converting integer to short month name?
References: <mailman.1049919678.686.python-list@python.org>
Message-ID: <Xns9359A3EE3BC2Aduncanrcpcouk@127.0.0.1>

MCollins at co.seminole.fl.us wrote in news:mailman.1049919678.686.python-
list at python.org:

> that works fine for the numerical month names.  now i'm trying to convert
> the numerical month into it's short month name equivalent to grab those
> files.
> 
> i could always use a long if , elif, elif to do the job, but i was
> wondering if there was built-in function already capable of that.  i was
> just looking for something a little more elegant.

Don't you just use the %b format string in strftime for this, or am I 
missing some subtlety in the question?

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From skip at pobox.com  Fri Apr  4 16:28:44 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 4 Apr 2003 15:28:44 -0600
Subject: output buffering--am I missing something obvious?
In-Reply-To: <yu99istu9ds9.fsf@europa.research.att.com>
References: <yu99istu9ds9.fsf@europa.research.att.com>
Message-ID: <16013.63756.223973.484125@montanaro.dyndns.org>

    >>> import sys
    >>> sys.stdout.write("foo")
    >>> sys.stdout.flush()
    >>> sys.stdout.write("bar\n")
    bar

    Andrew> What am I missing?

Dunno.  Using either 2.2.2 or CVS on Mac OS X I get this:

>>> import sys
>>> sys.stdout.write("foo")
foo>>> sys.stdout.flush()
>>> sys.stdout.write("bar\n")
bar

I tried using 1.5.2 (!) on a Solaris 8 machine (it's what comes on the
Software Companion CD for Sol8) I get the same results:

>>> import sys
>>> sys.stdout.write("foo")
foo>>> sys.stdout.flush()
>>> sys.stdout.write("bar\n")
bar

Bit rot somewhere?  It certainly shouldn't do that.

Skip




From exarkun at intarweb.us  Tue Apr 15 22:15:56 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 22:15:56 -0400
Subject: FTP
In-Reply-To: <e62b26d3.0304151727.6de2e2e4@posting.google.com>
References: <e62b26d3.0304151727.6de2e2e4@posting.google.com>
Message-ID: <20030416021556.GA8693@meson.dyndns.org>

On Tue, Apr 15, 2003 at 06:27:11PM -0700, Matt Singer wrote:
> I have an application with a PYQT interface that uploads large files. 
> Is there an async method to get how much data has been sent with
> storlines to update a progress bar?  Or is there an alternate ftp
> interface?

  Twisted might help.

    http://twistedmatrix.com/documents/TwistedDocs/current/api/public/twisted.protocols.ftp.html

  and doc/examples/ftpclient.py in the source tarball.

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 26 days, 23:02, 5 users, load average: 2.66, 2.47, 1.61




From anand at calbay.com  Tue Apr  8 19:34:05 2003
From: anand at calbay.com (Anand)
Date: 8 Apr 2003 16:34:05 -0700
Subject: Loading a class the name of which is only known runtime
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net>
Message-ID: <c6303995.0304081534.2f386065@posting.google.com>

there is a command called exec

you can try something like ...

yourstring = "generated at runtime"

exec("import %s as dynclass" %(yourstring))

Anand.
Mads Orbesen Troest <mads at troest.NEVERMORE.dk> wrote in message news:<13n0q1sc6o6oh$.hfa48a72df4n.dlg at 40tude.net>...
> Hi folks;
> 
> "And now for something completely different:"
> 
> I have a situation in which I, from a CGI script written in Python, want to 
> instantiate a class, the name of which is known to me only at runtime 
> (because it depends on a CGI parameter).
> 
> Is it possible to load (import and instantiate) a class dynamically in 
> Python? I have a feeling the answer is Yes, but as I am unfortunately not 
> quite the python expert (yet :) forgive me if I ask something obvious.
> 
> Thanks in advance,
>    /\/\\ads Orbesen Troest



From tundra at tundraware.com  Tue Apr 15 16:00:07 2003
From: tundra at tundraware.com (Tim Daneliuk)
Date: 15 Apr 2003 20:00:07 GMT
Subject: O'Reilly Python Books
In-Reply-To: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <3E9C6457.9030907@tundraware.com>

W S wrote:
> Hello,
> 
> I find the material, as presented, in the book Programming Python
> frustrating and extremely difficult to understand.
> 
> Are the nutshell and cookbook O'Reilly Python books easier to read?
> 
> I'm fairly new to programming but trying to learn. I have read Learning
> Python, Core Python, Python Bible and a bunch of perl books
> 
> Thanks,
> Wilfred.
> 

I highly recommend:

The tutorial that comes with the language

then,

http://www.hetland.org/python/instant-python.php

then,

http://www.diveintopython.org/


-- 
----------------------------------------------------------------------------
Tim Daneliuk     tundra at tundraware.com
PGP Key:         http://www.tundraware.com/PGP/TundraWare.PGP.Keys.txt




From gh at ghaering.de  Wed Apr 16 05:12:47 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 16 Apr 2003 11:12:47 +0200
Subject: Embedding Python Engine into a windows executable
In-Reply-To: <433b8cf0.0304160005.651339b2@posting.google.com>
References: <433b8cf0.0304160005.651339b2@posting.google.com>
Message-ID: <3E9D1E8F.2020703@ghaering.de>

Karam Chand wrote:
> Hello,
> 
> Is is possible to embed Python engine completely into a windows
> executable? 

Sure. But almost certainly not worth the effort.

> Basically, I want to give some kind of python scripting to
> allow users extend my applicaion and I dont want to ship python*.dll
> separately.

That's what installers are for.

-- Gerhard





From cben at techunix.technion.ac.il  Mon Apr  7 09:18:00 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Mon, 7 Apr 2003 16:18:00 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
Message-ID: <Pine.GSO.4.51.0304071359090.8306@techunix.technion.ac.il>

> Patrick K. O'Brien wrote:
> > If you were crazy enough to think that you could create a program
> > along the lines of Emacs, but written in Python, how would you go
> > about it?  How would you design the domain model for files, buffers,
> > windows, and frames?  How would you allow the same level of
> > customizability?  How would you map functions (or methods or whatever)
> > to keys?  Any thoughts?
> >
> > Here is the background.  I'm working on a file editor (called
> > PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> > PyCrust code base.  I've got a basic version working, and it supports
> > editing multiple files and all the usual, basic features.  But I want
> > it to be as customizable and extensible as Emacs.  In fact, I want to
> > model it after Emacs, so that elisp code could be rewritten in Python
> > and work with PyAlaMode.  To do that, I've got to support the same
> > primitives as Emacs; expose files, buffers, regions, and windows the
> > same way Emacs does; handle the Emacs keybindings and ability to remap
> > keys, the same way Emacs does, etc.
> >
> > Sound like fun?

Very.  I wanted to write an emacs since I started learning it when I
was about 14(?).  And I wanted to do it in Python since I learnt
Python (about a year ago) :-).  Don't count on me contributing,
however - I'm alwready overcommited :-(.

Carl Banks wrote on 2003-04-06:

> What I recommend is to read through the Emacs Lisp manual and figure
> out the functions you'll want to support.

Yes, `info elisp` is very good to read for this; `info emacs` contains
nothing important.  Here are some more reading pointers:

- http://www.cliki.net/CL-Emacs
- http://www.emacswiki.org/cgi-bin/wiki.pl?WishList
- http://xemacs.org/Architecting-XEmacs/index.html
- http://www.gnu.org/software/emacs/emacs-paper.html
- etc/OTHER.EMACSES in the distro (join the lines):
  http://savannah.gnu.org/cgi-bin/viewcvs/*checkout*/emacs/
  emacs/etc/OTHER.EMACSES?rev=HEAD&content-type=text/plain

> Obviously, you'll want to make each type of object (buffer, windows,
> frames, positions, markers, keymaps, etc.) a Python class.  When you
> read through the manual, take the lisp functions and assign them as
> method of the appropriate class. (For now, don't *write* the methods
> yet; just prototype them.)  This builds a set of interfaces that the
> classes should follow.  The point of this is to give you an idea of
> exactly what the different objects in Emacs do, and how they
> interact.
>
> Don't be too slavish in trying to match the exact behavior of the lisp
> functions, and be careful to assign functions to the class they really
> belong to.  Don't assign a function to the keymap class just because
> it's in the keymap section.
>
Good advices.

> As an example, take the section on buffers.  First function is
> "current-buffer".  What class should it belong to?  Certainly not
> Buffer.  The current buffer is a global property, so "current_buffer"
> should be either a global function, or a method of an application
> context class.
>
Technically I believe it's the current buffer of the current window of
the current frame.  Convenience functions are needed of course.

> Next three functions are "save-current-buffer", "with-current-buffer",
> and "with-temp-buffer".  These are actually not functions, but special
> forms and macros.  Python has no analogue to these (barring some
> obscure metaclass tricks).  Depending on the form, it might be useful
> to have a corresponding method in Python, or it might be better to
> think about a more Pythonic way to do it.  I haven't thought about it
> too much, but off hand, I'd say I'd choose the latter in this case.
> In fact, these seem to be mostly a convenience, so I wouldn't bother
> to include them.
>
This is not a tiny issue trivial to give up.  This kind of macros is
the heart of elisp programming.  The normal way to e.g. find another
place in the buffer is to use functions that actually *move the point*
and record the new position - all this inside `save-excursions'.
Ditto for other tasks.  Elisp has a very imperative, plumbing style.
It's very common for functions to actually change things.

The primitives for building other functions are in part the promitives
the user itself uses when editing - or thin layers upon them.  This
also fits very well with the idea of "interactive functions", which I
think is a great idea.  When you need to combine them in other ways
something else you have all the save- and with- macros that allow you
to plumb these destructive primitives together.

> Next function is "buffer-name".  This obviously belongs in the Buffer
> class.  I'd just call it "name".  Perhaps you want to retain the exact
> name, though.  That's up to you.
>
> Next function is "rename-buffer".  This should belong in the Buffer
> class--however, in Emacs, it only operates on the current buffer.  I
> think that's a design flaw; I would fix it.  I would make "rename" a
> method of Buffer.
>
> You might want to also include global "buffer_name" and
> "rename_buffer" functions that operate on the current buffer.
>
> Next function is "get-buffer".  This should be global, because the
> list of buffers is global.
>
> And so on.
>
These look good.

> Understand what I'm saying?  Eventually, you'll have a nice list of
> classes waiting to be fleshed out.  It should be straightforward to do
> it.
>
> The next thing you should do it refactor your original PyAlaMode code
> to use these classes.  Don't add functionality.  Just implement what
> you need to get back to the original behavior of PyAlaMode, but using
> the new classes.
>
> You said the keymaps aren't editable?  Add one keymap.  Don't worry
> about adding functions to edit keymaps yet.  Say you currently use a
> string to hold the buffer.  Change it to use a Buffer object.
>
> Once you've done all this, then it should be straightforward to fill
> out the interfaces and add features.
>
Sound advices.

> There is one potentially disasterous pitfall you should be aware of.
> Emacs does all kinds of strange things with namespaces.  It uses
> dynamic binding, unlike Python, which uses lexical binding.

This in itself is a minor language issue.  It's true that sometimes
temporarily changing some settings to affect called functions is nice
but it should be easy to do it otherwise.

> Emacs has one monolithic global namespace, as opposed to Python,
> which has an organized infrastructure of "global" namespaces, that
> is, modules.

Again, minor issue.  Emacs libraries are well-disciplined to use
consistent prefixes.  Python should win here a little.

> Also, Emacs has weird namespaces like buffer local variables.  One
> thing you'll have to do is think about how to sort all these
> namespace issues out.
>
This is a tremendously important point.  Think of this before you
think anything else.  Emacs also has a some frame-local and I think
even terminal-local variables nowdays.  It didn't always have them.
Initially there were only buffer-local variables.  The point I'm
driving at is that these were intorduced *without* most code knowing
about them.  Any variable can be become buffer-local.  If you make
thme first global and then rethink them as attributes of a buffer
object, you will need to change all the code - and still one day you
will need frame-local...

If you want ot just mirror emacs, have a special proxie object
containing everything that would be a variable in emacs as attributes.
Then you can invent new kinds of localities without changing code.

The lesson is that emacs is full of quick-and-dirty tricks but there
are good reason they are there.  It's carefully designed to allow
*unforseen* customizations.  That's why there are hooks all over it,
something rare in applications that are thought out from before.
Don't fall into the pitfall of assuming that you can predict the
future customization directions completely.  When choosing good
modular structure, don't do it to minimize the code you already have
but to maximize the flexibility for code nobody thought of yet.

[Excuse my lecture tone - I didn't intend it...]

> It seems to me that a lot of these issues won't matter much if you
> have a good and consistent strategy for organizing stuff and passing
> parameters around.
>
They matter a lot; that's why you must find a good strategy <wink>.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>

When I grow up I want to write an Emacs ;-)




From aahz at pythoncraft.com  Sun Apr  6 20:44:22 2003
From: aahz at pythoncraft.com (Aahz)
Date: 6 Apr 2003 20:44:22 -0400
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <m27ka7aae0.fsf@localhost.localdomain> <b6pri6$drf$1@panix2.panix.com> <mailman.1049656989.5896.python-list@python.org>
Message-ID: <b6qhl6$sht$1@panix2.panix.com>

In article <mailman.1049656989.5896.python-list at python.org>,
Skip Montanaro  <skip at pobox.com> wrote:
>
>    aahz> Bare emacs is unusable.
>
>I'm not sure what you're referring to.  No Emacs installation I'm aware
>of is "bare".  (I presume you're not referring to the predump "temacs"
>executable.)  I find "bare" vi pretty darn inscrutable as well.  I know
>how to 'dd', 'x', 'i', 'a' and ':wq[!]', and can navigate with the
>arrow keys.  Just enough to tweak the occasional config file.  Anything
>more substantial requires Emacs in some fashion (even if it's just X11
>copy/paste from an Emacs buffer running on another machine).  It all
>boils down to what you're productive with.  I began using Emacs more
>than 20 years ago, so at this point any editor in which typing 7-bit
>ASCII doesn't result in those characters winding up in the file I'm
>editing seems weird.

"Bare" == no .emacs

Using emacs with its default keybindings hurts my fingers.  If I were to
seriously use emacs, I'd need a hefty .emacs, which would make me unable
to be productive with it in an emergency.  As for using emacs in vi
mode, why not just stick with vi?
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From artyprog at wanadoo.fr  Thu Apr 17 07:31:30 2003
From: artyprog at wanadoo.fr (Salvatore)
Date: Thu, 17 Apr 2003 13:31:30 +0200
Subject: variable in Python
In-Reply-To: <b7m1g4$qqb$2@news-reader11.wanadoo.fr>
References: <b7m1g4$qqb$2@news-reader11.wanadoo.fr>
Message-ID: <b7m3cq$r4k$1@news-reader12.wanadoo.fr>

Salvatore wrote:

> Hello,
> 
> I have a little doubt concerning variables.
> When one declare x = 3
> does x contains the value 3 or
> does x acts like pointer in C ?
> 
> Regards
> 
> Salvatore
> 
> 
I have the answer thanks:
x is an implicit pointer





From usenet at soraia.com  Sun Apr 27 15:02:55 2003
From: usenet at soraia.com (Joe Francia)
Date: Sun, 27 Apr 2003 19:02:55 GMT
Subject: Textbooks?
In-Reply-To: <oproar1sjtx3zh5c@news.mindspring.com>
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <3EAC2962.2050609@soraia.com>

Steven Hodgen wrote:

> Hello,
> 
> I'm currently teaching a high-school introduction to programming 
> course.  I'm currently teaching in Java, but am seriously considering 
> changing to Python.  The problem is that I cannot seem to find any good 
> textbooks, much less ones for students with no prior programming 
> experience.  This is the one area in which Java excels.  I'm using an 
> awesome programming environment called "BlueJ", and there are lots of 
> textbooks for students at all levels.
> 
> However, Java is not a great language for beginners.  Python, on the 
> other hand, is ideal.  If I can find a textbook that is even ok, I may 
> consider changing, since the language is so much better for students at 
> this level.  Doesn anyone know of a book like this?  Ideally, it would 
> have examples and excercises, and perhaps even a web-site with other 
> instructor and student resources.
> 
> Thanks!
> 
> --Steven
> steven at twitch.net

The Deitels have a textbook called _Python How to Program_ that is aimed 
at beginning programmers.  I'm not sure how good/bad it is.  Three out 
of four reviewers on Amazon seem to like it (the fourth guy gave it one 
star based on its lack of Perl coverage).




From exarkun at intarweb.us  Mon Apr  7 15:07:13 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 7 Apr 2003 15:07:13 -0400
Subject: Comms Protocol design dilemma
In-Reply-To: <000001c2fd30$8c021740$7f88013d@JeethuRao>
References: <000001c2fd30$8c021740$7f88013d@JeethuRao>
Message-ID: <20030407190713.GA29780@meson.dyndns.org>

On Mon, Apr 07, 2003 at 11:37:23PM +0530, Jeethu Rao wrote:
> Hi,
> I've been working on a encrypted instant messenger server for
> quite some time now. I've had to dump the code about 3 times
> (I changed the protocols). And quite frankly I don't feel that
> bad while throwing out code written in python, since its very
> easy to rewrite the same functionality (Throwing away *hard written*
> C code really breaks my heart.)

  Do you mean you've dumped the entire server, not just the protocol
handling code?  If so, it sounds like you could be doing a little better job
keeping these separate areas of your code actually separate.

> 
> Now all the past three implementations are based on a packet
> Oriented scheme <TCP/IP, Blowfish Encryption,16 byte packet hdr>.
> My friends in here are suggesting me to use a line oriented protocol
> (Like Microsoft Messenger), some have even suggested using XML
> (Which I vehemently hate) like in Jabber. I'm of the opinion that
> XML tends to waste a lot of bandwidth.

  I agree.  Don't use XML.  Additionally, since you're not sending
plain-text, and so you're not sending lines, I don't think you should use a
line-oriented protocol.

> 
> Since most of the data is binary, If I go in for line oriented protocol,
> I'll have to base64 (or something similar) all the encrypted data,
> Which will be an added overhead.

  Correct.

> 
> The primary dilemma is, What protocol is better suited, 'text based' or
> 'packet based' ?

  It seems pretty obvious that a text based protocol isn't better suited. 
Maybe if you were actually sending textual data...

> 
> The other problem is in the implementation.
> The old implementations used a simple scheme, every incoming
> Connection would be handled by a thread from a thread pool.

  This sounds like a bad idea.  How can it scale to many thousands (a small
number, given the application, right?) of users this way?

> But now, the client is going to be connected to the server
> All the time its online. Thus, a lot of clients means a lot
> Of threads running in parallel, doing nothing. I'm not quite sure
> If the Python threads waiting on sockets would yield as soon
> as it gets its share of the VM (They're all serialized, right ?).
> And since there's a lot of encryption/ decryption , hashing, random
> number generation involved, I can pretty much rule out asynchronous
> sockets. Is there any better strategy to tackle this problem?
> 

  Ahh, I see, in the previous scheme, "connections" (maybe you were even
using UDP) were short lived, so one thread might handle many of them over a
short period of time.

  I don't think you need to rule out the asynchronous solution.  Check out
Twisted.  You can have all your network code run in a single thread, and all
the encryption/decryption/hashing/random number generation run in any number
of worker threads.  Here's how easy it might be (untested code):

    from twisted.internet.protocol import Protocol
    from twisted.internet.threads import deferToThread

    from Crypto.PublicKey import RSA

    class MessageProtocol(Protocol):
        buffer = ''
        def dataReceived(self, data):
            self.buffer += data
            if len(self.buffer) >= PACKET_SIZE:
                buf = self.buffer[:PACKET_SIZE]
                self.buffer = self.buffer[PACKET_SIZE:]
                d = deferToThread(rsaInstance.decrypt, buf)
                d.addCallback(self.decryptedDataReceived)
                d.addErrback(self.decryptionFailed)

        def decryptedDataReceived(self, data):
            """Override me"""

        def decryptionFailed(self, failure):
            """Override me"""

        def sendData(self, data):
            d = deferToThread(rsaInstance.encrypt, data)
            d.addCallback(self.sendEncrypted)
            return d

        def sendEncrypted(self, data):
            self.transport.write(data)

  Twisted manages the thread pool for you, takes care of getting things
called in the right thread, and lets you use just one thread for network IO.

  Check it out - http://www.twistedmatrix.com/

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 18 days, 15:01, 12 users, load average: 0.12, 0.09, 0.06




From sholden at holdenweb.com  Wed Apr 16 15:03:54 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 19:03:54 GMT
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <mailman.1050502229.7932.python-list@python.org>
Message-ID: <uOhna.110721$It5.15606@news2.central.cox.net>

"Gerhard Haering" <gh at ghaering.de> wrote in message
news:mailman.1050502229.7932.python-list at python.org...
> Steve Holden wrote:
> > "Aahz" <aahz at pythoncraft.com> wrote:.
> >>Right now, I'm trying to get some momentum moving on a book titled
> >>_Python for VB6 Programmers_.  Anyone interested should subscribe to
> >>marketing-python.
> >>
> >>http://wingide.com/mailman/listinfo/marketing-python
> >
> > You'd do better to call it "Python for VB programmers", expecially since
the
> > change to VB6 is so radical that migrating to a new and more sensible
> > language might be an attractive option to some!
>
> You're confusing VB6 with VB.NET (VB7).
>

So I am. Thank you.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From blunck at gst.com  Tue Apr 29 11:53:03 2003
From: blunck at gst.com (Christopher Blunck)
Date: 29 Apr 2003 08:53:03 -0700
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com> <b8m3cj$8vu$1@ginger.mathworks.com>
Message-ID: <1147e466.0304290753.5cbf3c8c@posting.google.com>

Josh <joshway_without_spam at myway.com> wrote in message news:<b8m3cj$8vu$1 at ginger.mathworks.com>...
> Christopher Blunck <blunck at gst.com> wrote:
> ...
> 
> > What advice can you offer on spreading the testing bug around?  
> 
> If you have an established codebase, it might be difficult to convince
> people to go back and add unit tests for all functionality.  One good
> way to ease people into it is to get them to write a unit test for
> every bug-fix they make.


We do that already.  It works to a certain extent.  But usually
requires brow beating of the bug fixer.  I guess we really just need
some time to pass.  Unit testing is an investment and by definition
investments don't pay off immediately.


-c



From news at snakefarm.org  Fri Apr 11 11:03:35 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 11 Apr 2003 15:03:35 GMT
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de> <v29im-ia2.ln1@nb2.stroeder.com>
Message-ID: <slrnb9dma7.1vh.news@cipw08.physik.uni-bonn.de>

In article <v29im-ia2.ln1 at nb2.stroeder.com>, Michael Str?der wrote:
> 
> For security reasons I'd be scared of printing an arbitrary exception 
> message to stdout in a web application since it is sent to the user's browser!
> 

Of course I wouldn't do that in a real world situation. It was just a simple
example of how to handle an exception.

> stderr *is* captured by the web server and most times written to an error 
> log. That's where those error messages belong.

Right. But if you are a really big ISP (and the original poster's ISP *is*
big) you do not want to let your web servers write error logs.

> If you don't have access to 
> the error log redirect stderr to a local web application log file.

ACK.

cg.




From gmuller at worldonline.nl  Tue Apr  8 15:37:31 2003
From: gmuller at worldonline.nl (GerritM)
Date: Tue, 8 Apr 2003 21:37:31 +0200
Subject: Tiobe index algorithm now published
Message-ID: <b6v8id$qli$1@reader1.tiscali.nl>

We have had discussions about the "TIOBE Programming Community Index" at
http://www.tiobe.com/tpci.htm
One of our consistent critics was that the algorithm was not published nor
verifiable. Tiobe has now added the (simple) algorithm on this page. They
also explain which "programming languages" are monitored.

For all clarity I have no relationship with Tiobe, although both Paul Jansen
and myself have worked at different periods at Philips Research.

kind regards Gerrit
--
www.extra.research.philips.com/natlab/sysarch/





From boud at valdyas.org  Wed Apr 16 15:26:03 2003
From: boud at valdyas.org (Boudewijn Rempt)
Date: Wed, 16 Apr 2003 21:26:03 +0200
Subject: How many of you are Extreme Programmers?
References: <mailman.1050513152.32206.python-list@python.org> <3E9DABA4.B2B53153@engcorp.com>
Message-ID: <3e9daec3$0$133$e4fe514c@dreader9.news.xs4all.nl>

Peter Hansen wrote:

> Although the social aspects are not easy, I reject the idea
> that almost all developers have territorial egos.  Territoriality
> was by far the *least* significant social issue we encountered
> in becoming an XP team.  In fact, I'd go so far as to say that
> in most cases (in my experience) people welcome not having to
> feel sole responsibility for a given piece of code.  Not always,
> but very often.

Pair programming was the most fun and most productive way to create code I'd
ever experienced. With the right partner, that is -- and unfortunately the
right partner left the company. So no more pair programming for me.

With the wrong partner, it can be hell. Curiously enough, wrong didn't have
anything to with territoriality -- at my current employer's I've never met
a territorial coder -- but experience. And coding mentality. When your
partner refuses to learn even the language you're supposedly programming in
the fun ends pretty quickly.
-- 
Boudewijn Rempt | http://www.valdyas.org



From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Wed Apr 30 17:02:14 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Wed, 30 Apr 2003 23:02:14 +0200
Subject: network programming problem??
In-Reply-To: <14bd1399.0304301236.7fe19be5@posting.google.com>
References: <14bd1399.0304301236.7fe19be5@posting.google.com>
Message-ID: <3eb03abd$0$49109$e4fe514c@news.xs4all.nl>

kkennedy wrote:
> Here are the results of running the script:
> 
> -------------------- LINUX --------------------------
> Run 0 . 16.69
[...]

> -------------------- WIN XP -------------------------
> Run 0 . 2.69
[...]

> Does anyone have a clue why these results are soooo different.  The
> SuSE system doesn't "feel" sluggish.  My first clue was that a Linux
> user complained about my program being slow, but none of my Windows
> users complained.  My second clue was that I thought I could tell a
> pretty big difference while browsing the web on Win XP vs. SuSE.  I
> just thought it was my imagination, but I guess not!  This is quite a
> difference.

I suspect you have a problem with your DNS server configuration on the
Linux machine. What happens when you type:

ping www.python.org

at a shell prompt? It should start pinging nearly instantly.
I suspect on your system it only starts pinging after about 15 seconds.
And when you type:

ping 194.109.137.226       (www.python.org's IP number)

it will work ok, because no DNS lookup has to be done. Right?

If I'm right, try looking at your DNS server config and see what the
differences are between the Linux and the Windows machine. I suspect
that the first (primary) DNS server on your Linux machine is wrong.

--Irmen




From graham__fawcett at hotmail.com  Thu Apr 24 09:56:51 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 24 Apr 2003 06:56:51 -0700
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com> <e9570f37.0304220625.11312687@posting.google.com> <K2qdnXJ_ZeZMVTujXTWcqg@speakeasy.net> <F3Epa.51023$LB6.1152908@news1.tin.it>
Message-ID: <e9570f37.0304240556.50465c66@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<F3Epa.51023$LB6.1152908 at news1.tin.it>...
> A.M. Kuchling wrote:
>         ...
> > Money does not make interested developers materialize,
> 
> 
> Surely you're joking, Mr Kuchling?


Alex, are you demonstrating the materialization of an interested
developer? ;-)

I think there are a lot of things that could motivate a developer to
contribute time:

money,
pride in workmanship,
sense of contribution,
sense of community and belonging,
unabashed praise from colleagues,
too much free time on hand,
invitations to Python conferences,
win-a-lunch-date-with-the-BDFL opportunities,
t-shirts ("import me", "I hacked Python and made the world a better
place for thousands of developers, and all I got was this lousy
t-shirt"),
plaques and certificates for the den,
thinkgeek toys,
prepaid slashdot subscriptions,
O'Reilly gift certificates,
prepaid Internet access or server hosting (okay, maybe that's too
much),
...
and of course,
the way that girls go crazy for Python contributors.

Of course we have no infrastructure for this (at least for the ones
that cost money). But I'm sure we could create one if the need arose.
If it would help to move our goals along, I would gladly pay an annual
Python tax, and/or participate in some kind of token economy. It would
be a small price to pay in return for the *tremendous* advantages (and
joys!) that the Python community provides to me.

In short, I think we should all send $20 a year to Guido, in return
for which he will dress in a toga once a month and do a
deus-ex-machina to bless and thank some unsuspecting Python
contributor on our common behalf.

-- Graham



From charts at purdue.edu  Sun Apr 20 20:42:19 2003
From: charts at purdue.edu (richard.a.charts.1)
Date: Sun, 20 Apr 2003 19:42:19 -0500 (EST)
Subject: Keyboard Commands
In-Reply-To: <mailman.1050882513.4409.python-list@python.org>
References: <Pine.SOL.4.51.0304201825001.6411@herald.cc.purdue.edu>
 <mailman.1050882513.4409.python-list@python.org>
Message-ID: <Pine.SOL.4.51.0304201942040.3687@herald.cc.purdue.edu>

Sorry, forgot to mention that.  I am using tkinter.


On Mon, 21 Apr 2003, Gerhard [iso-8859-1] H?ring wrote:

> * richard.a.charts.1 <charts at purdue.edu> [2003-04-20 18:27 -0500]:
> > Hey,
> > I'm working on a program and I want to use keyboard shortcuts for some
> > of my menu items.  I can't seem to find the way to implement this.
> > Can anyone point me in the right direction?  ex.  Ctrl-x, ctrl-F1,
> > etc.  [...]
>
> Which toolkit are you using? wxPython, PyQt, Tkinter, curses, ..., ?
>
> Gerhard
> --
> mail:   gh at ghaering.de
> web:    http://ghaering.de/
>
>




From michael at stroeder.com  Tue Apr 15 08:16:00 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Tue, 15 Apr 2003 14:16:00 +0200
Subject: Pythonic way of web-programming
In-Reply-To: <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <mailman.1049653811.11491.python-list@python.org> <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net>
Message-ID: <0p9tm-es4.ln1@nb2.stroeder.com>

A.M. Kuchling wrote:
> The most to hope for is that some of the systems will die out,
> leaving two or three more major ones in control of the space.

Yes, that's my hope too. I'd like to substitute my own little framework with 
another more comprehensive web toolkit. But it's hard to determine at the 
moment which one will survive and will still be maintained in one or two 
years => I stay with mine in the mean-time.

> The best we
> can do now is to make it easier for users to choose which toolkit to use by
> documenting them and comparing their differences.

I strongly agree.

Ciao, Michael.




From bjarke-news2003 at trebe.dk  Tue Apr  8 15:35:27 2003
From: bjarke-news2003 at trebe.dk (Bjarke Dahl Ebert)
Date: Tue, 8 Apr 2003 21:35:27 +0200
Subject: Pathname problem
References: <3e9321b5$1@rmcnt03a.rmc.ca>
Message-ID: <zwFka.7791$y3.946342@news010.worldonline.dk>

"bourassa" <bourassa-m at rmc.ca> wrote in message
news:3e9321b5$1 at rmcnt03a.rmc.ca...
> Hi there.  I am just trying to load files using PIL.  I get the following:
>
> im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
[...]
> IOError: [Errno 2] No such file or directory:
> 'C:\\MATLAB6p1\toolbox\\pout.tif'

Hint: \t is the TAB character ;-)

Either, "C:\\MATLAB6p1\\toolbox\\pout.tif"
or r"C:\MATLAB6p1\toolbox\pout.tif"
will help. The r"..." syntax means that backslashes are interpreted
litterally.


Bjarke







From maschio_77 at hotmail.com  Mon Apr 28 12:43:27 2003
From: maschio_77 at hotmail.com (Federico)
Date: Mon, 28 Apr 2003 18:43:27 +0200
Subject: Tkinter: Position of windows
Message-ID: <b8jllt$ei3$1@lacerta.tiscalinet.it>

How can I control where a window will appear? I'd like to set the initial
position of a window...
Thanks





From ianb at colorstudy.com  Tue Apr 29 05:57:06 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 29 Apr 2003 04:57:06 -0500
Subject: POD vs. reST for standalone writing?
In-Reply-To: <3EADFCCE.7080504@python.org>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
	 <3EADFCCE.7080504@python.org>
Message-ID: <1051610226.675.2656.camel@lothlorien>

On Mon, 2003-04-28 at 23:17, David Goodger wrote:
> eichin at metacarta.com wrote:
> > Now I start seeing mention of reStructuredText (especially in comments
> > about pyCon.)  reST files look very "ascii art" on the input side, but
> > after a little experiment, that seems to make them *harder* to write,
> > if somewhat (and I think only a little) easier to read. 
> 
> That's significant: "somewhat (and I think only a little) easier to 
> read".  I designed reStructuredText to be a *lot* easier to read. 
> Opinions vary of course, I think mostly determined by what you're used 
> to.  (Some people think in XML, others dream in TeX.  Not me.)
> 
> As for being harder to write, that may be true, especially if the author 
> isn't used to the markup yet.  In reStructuredText, readability is 
> deemed far more important than writability.  But with increasing 
> familiarity comes increasing ease of writing.

I would say (emphatically!) that reST is good for writing, and that's
why I like it, more than for reading.  Sure, I know (X)HTML just fine,
but using it in documentation does not appeal to me one bit.  Honestly,
I can't think of a worse format for writing... it's good for
extensibility, and it's well known, but it's not easy on the fingers (or
on the brain, while I'm trying to write about code, not think about what
I'm typing).

As far as learning to write... it takes a little bit.  Linking still
confuses me sometimes.  But you can use it easily enough without knowing
anything about it, and it pretty much works... so I don't think it's
that big of a problem.  Perhaps it would help if a few of the more
obscure kinds of linking and annotation could be turned into a clearer
set of orthogonal features.

  Ian






From dave at boost-consulting.com  Mon Apr 21 07:02:05 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Mon, 21 Apr 2003 07:02:05 -0400
Subject: Pythonic way of web-programming
In-Reply-To: <1050884881.698.12.camel@lothlorien> (Ian Bicking's message of
 "20 Apr 2003 19:28:01 -0500")
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
	<6ee58e07.0304061758.35070e7c@posting.google.com>
	<a3b5b608.0304080302.c26ad68@posting.google.com>
	<af6b5bdd.0304181001.558da6f@posting.google.com>
	<84of317cd6.fsf@boost-consulting.com>
	<mailman.1050870230.2976.python-list@python.org>
	<84n0ikommx.fsf@boost-consulting.com> <1050884881.698.12.camel@lothlorien>
Message-ID: <84he8snmfm.fsf@boost-consulting.com>

Ian Bicking <ianb at colorstudy.com> writes:

> On Sun, 2003-04-20 at 17:00, David Abrahams wrote:
>> I don't understand; it already *is* plain Python.  While Donovan
>> Preston and I were designing the syntax, his friend Bob was sitting on
>> the couch coding it up.
>
> Hmmm... I guess so.  I was thinking of lists following an object, but it
> actualy gets parsed as a __getitem__ with a tuple as an argument.

Yes.

>> > The use of []'s isn't really necessary, and actually the expressions
>> > could maybe be eval'd as well...
>> 
>> You've lost me here.  Lots of the "syntax" choices aren't based on
>> what's neccessary, but on what's clear and expressive.  As far as
>> "eval" is concerned, well, I'm not sure if you realize this but those
>> if_ and for_each elements are lazily evaluated.  The expression
>> creates a callable object, "template", which generates new XML.
>
> So it's not actually a Python program, it's Python syntax 

And that's different from a Python program how? <.02 wink>

> that gets evaluated by your own evaluator...?  

Basically, yes, though the evaluator might be built into the object
that gets constructed.

> I was trying to see if it could be an actual Python program, which
> would construct an object you could use later to handle the XML.

I think if you go in that direction it will end up looking much as it
does... the problem, as you say, is that code blocks are not
first-class objects in Python.  Of course, that goes even further in
the direction of lambda-expressions than we've gone already, which is
the opposite of the direction GvR has been headed.  On the other hand,
he recently started talking about adding support in the future for the
design of domain-specific languages (which this XML example was), and
I don't really see how you can do that job well without first-class
code blocks.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com





From Sorry at ButNo.com  Fri Apr 25 23:11:07 2003
From: Sorry at ButNo.com (Hyperion)
Date: Sat, 26 Apr 2003 04:11:07 +0100
Subject: ActivePython Scripts help with ?
References: <b89bu8$5bp$1@newsg2.svr.pol.co.uk> <slrnbai3cd.84b.rmunn@localhost.localdomain>
Message-ID: <b8ct9o$t3k$1@news7.svr.pol.co.uk>

Excellant thanks for all your help ;)
 I did find out in the end how to do my ( #! what goes here problem),
but after all that I decided to just reinstall the original python back and
forget the ActivePython, the PATH was  to long and I only really wanted
Active for idle, as idle wasnt installed on my system so I could use it.
 I have however decided to force myself to use another editor, as the
indenting was was my biggest worry, is something i really nead to get used
to doing for myself.
 I would of liked to have had idle though, but it doesnt matter ;0
 The   #!/usr/bin/env python  is very helpfull thanks mate.
HYperion




From godoy at metalab.unc.edu  Tue Apr  8 15:34:36 2003
From: godoy at metalab.unc.edu (Jorge Godoy)
Date: Tue, 08 Apr 2003 16:34:36 -0300
Subject: Best way to make a weighted sum
Message-ID: <867ka46awj.fsf@ieee.org>

Hi!


I have the following snippet of code:

----------------------------------------------------------------------
class IEModulo11:

    def __init__(self, ie=None, uf=None):
        
        self.estados = ['CE', 'AP', 'RJ'] 
        self.ie = ie
        self.soma = 0


    def EstipulaPesos(self, uf=None):
      """
      First digit indicates the number position, the second indicates
      it's weight.
      """

      if uf in ['CE', 'AP']:
            self.pesos = {8:9, 7:8, 6:7, 5:6, 4:5, 3:4, 2:3, 1:2}
      else:
            if uf in ['RJ']:
                self.pesos = {7:2, 6:7, 5:6, 4:5, 3:4, 2:3, 1:2}


    def CalculaSoma(self):

        ie = self.ie
        iter = len(ie)

        for i in ie:
            produto = i * self.pesos[iter]
            self.soma = self.soma + produto
            iter = iter - 1

----------------------------------------------------------------------

What I want is to sum the products of a certain 'ie' number where each
of its digits is multiplied by a certain weigth, depending on its
position inside the number. (This is for some documents verification
and 'ie' can vary in the amount of digits as well as its weigth can
vary from one place to another --- I know this is a mess, but it's the
way some Brazilian states did it...)

I don't know if I got it in the most pythonic way in my code. 


TIA,
-- 
Godoy.      <godoy at metalab.unc.edu>

Your analyst has you mixed up with another patient.  Don't believe a
thing he tells you.



From j.jakl at era.cz  Fri Apr 25 04:06:42 2003
From: j.jakl at era.cz (Jaroslav Jakl)
Date: Fri, 25 Apr 2003 10:06:42 +0200
Subject: event IDs for win32evtlogutil.ReportEvent
Message-ID: <b8aq99$1l79$1@news.vol.cz>

Hello all,
I need to write events to windows event log.
In the win32evtlogutil is writen that win32evtlog.pyd can be source of
messages,
but i don't know what messages are there and what are their IDs.
Can anybody help me.
Thanks in advance
Jaroslav Jakl





From juneaftn at REMOVETHIShanmail.net  Wed Apr 23 12:24:56 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Thu, 24 Apr 2003 01:24:56 +0900
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com>
Message-ID: <b86enq$kr2$1@news.hananet.net>

"Aahz" <aahz at pythoncraft.com> wrote in message
news:b86e9f$c15$1 at panix1.panix.com...
> In article <b86bqd$iu4$1 at news.hananet.net>,
> Changjune Kim <juneaftn at REMOVETHIShanmail.net> wrote:
> >"Aahz" <aahz at pythoncraft.com> wrote in message
> >news:b867d7$f53$1 at panix2.panix.com...
> >>
> >> While writing it, it occurred to me that I *ought* to be using unit
> >> tests to make sure that it was working correctly, but two things
stopped
> >> me.  Secondly, I was in a hurry.
> >
> >On the second:
> >
> >KentBeck : "If you write automated tests, you'll get more done every
day."
> >Johnny : "But we have too much to do. We don't have time to write
tests."
> >KentBeck : "No, no. You didn't understand. You'll get _more_ done, not
> >less."
> >Johnny : "Yes, but we don't have time for testing."
>
> Yes, I know; I didn't realize this would be a "program" rather than a
> "script" until I started writing it...
>
> Which brings up the question: do the unit-testers here write tests for
> throwaway scripts?

I do. :) More often than not I regret it after I write any piece of code
without tests, whether it's throw-away or not. Writing tests, it's more
enjoyable, less stressful and I learn more.




From jfromm at jasc.com  Thu Apr  3 17:06:06 2003
From: jfromm at jasc.com (Joe Fromm)
Date: Thu, 3 Apr 2003 16:06:06 -0600
Subject: Using Tkinter with Python embedded?
Message-ID: <3e8cb069$0$840$39cecf19@news.twtelecom.net>

I have an app on Windows that uses Python as a scripting language.  We have
both extended and embedded Python 2.2 in our application.

We can successfully use Tkinter in a script, but the problem is that windows
created by Tk very easily get sent behind the app.  What I am looking for is
a way to make a Tk window be a child of our application, so that it is not
allowed to go behind the application.  I could make it topmost easily
enough, but then it won't go behind any window.  Ideally I would make the Tk
window behave like a modal dialog, so that no interaction with the app is
possible until the dialog is dismissed.

I have tried getting the window handle with root.winfo_id, and then passing
that into the app and calling SetParent, but that didn't work.

I have tried creating a Toplevel widget with the use= option, passing in
window handles of either my app window or the handle of a dialog window.  No
joy there either, though possibly I gave up before I explored all the
alternatives.

I can't believe I'm the first person to have this problem, but a number of
google searches have come up dry.

Can anyone suggest an approach?

Joe






From pinard at iro.umontreal.ca  Tue Apr 15 21:31:42 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 15 Apr 2003 21:31:42 -0400
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <_l0na.16003$883.6891@nwrdny02.gnilink.net>
References: <b6pgcm$1t3$1@panix3.panix.com>
	<yfsbrzeh3kt.fsf@black132.ex.ac.uk>
	<9Qhla.7802$d8.3212@nwrdny02.gnilink.net>
	<yfswui1f3la.fsf@black132.ex.ac.uk>
	<9RDla.592$Gb1.345@nwrdny01.gnilink.net>
	<yfsk7dydmxj.fsf@black132.ex.ac.uk>
	<z0sma.14160$Gb1.10725@nwrdny01.gnilink.net>
	<m2n0itfrft.fsf@mycroft.actrix.gen.nz>
	<slrnb9l507.1q1.rmunn@localhost.localdomain>
	<m2fzokfgni.fsf@mycroft.actrix.gen.nz>
	<slrnb9oajn.493.rmunn@localhost.localdomain>
	<yfs3ckjcsi8.fsf@black132.ex.ac.uk>
	<_l0na.16003$883.6891@nwrdny02.gnilink.net>
Message-ID: <oqy92b2poh.fsf@alcyon.progiciels-bpi.ca>

> [...] Since rebinding sys.stdout is utterly stupid [...]

I've been following this thread at a distance, yet I'm noticing many
judgemental remarks like the one above, and others before it, which kills
interest somewhat.  The discussion is slipping.

You know, nothing is fully black nor fully white.  Not that we should not
have opinions, all the contrary, there is something "drabe" in a society
made of opinion-less people.  However, best opinions have some substance,
beyond mere emotions.  Emotions are good too, but substance helps! :-)

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From koenig at v-i-t.de  Tue Apr 15 16:31:54 2003
From: koenig at v-i-t.de (Mirko Koenig)
Date: Tue, 15 Apr 2003 22:31:54 +0200
Subject: MySQLdb: getting inserted data
Message-ID: <b7hq7q$v3e$1@news.tu-darmstadt.de>

Hi

I want to insert a new empty row into my table and get the value of the
primary key, so that i can identify it.
Because it may happen that there are other empty rows.
So i can't insert an empty row and then make a 

select....where <all values are empty>

So how to get the last inserted row?

Mirko Koenig



From aleax at aleax.it  Sun Apr 13 05:03:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 09:03:52 GMT
Subject: Creating a PyObject* from a C double*
References: <20cb355e.0304122315.c068128@posting.google.com>
Message-ID: <YJ9ma.6928$LB6.196819@news1.tin.it>

Joze wrote:

> Hi,
> 
> As part of a Python extension, i have an array of doubles created in a
> C function, which should return a corresponding PyObject* to Python.
> The array is meant to be used by Numeric/numarray, but apparently, the
> C function cannot directly return a PyArrayObject*.

It can _after casting the pointer_ -- is that what's blocking you?


> I have tried PyArray_FromDims(), PyArray_FromDimsAndData() but i can't

Those should only be used for static arrays -- ones that never will
be deallocated; it's unlikely that's what you want.

> seem to get it right. Can anyone tell me how this works? I'd be most
> grateful.

Here's a tiny working example just hacked together:


#include <Python.h>
#include <Numeric/arrayobject.h>

static PyObject*
tiny(PyObject* self, PyObject* args)
{
    PyArrayObject* parr;
    int dims[1];
    int i;
    double xx[] = {1.0, 2.3, 4.5, 6.7, 8.9};
    double *data;

    dims[0] = 5;
    parr = (PyArrayObject*) PyArray_FromDims(1, dims, 'd');
    if(!parr) return 0;

    data = (double*) parr->data;
    for(i=0; i<5; ++i)
        data[i] = xx[i];

    return PyArray_Return(parr);
}

static PyMethodDef tinyMethods[] = {
    {"tiny", tiny, METH_VARARGS, "example"},
    {0}
};

void
inittiny(void)
{
    Py_InitModule("tiny", tinyMethods);
    import_array();
}

anf after a python setup.py install with the obvious setup.py,

[alex at lancelot pronu]$ python2.2 -c 'import tiny; print tiny.tiny()'
array([1.0, 2.2999999999999998, 4.5, 6.7000000000000002, 
8.9000000000000004], 'd')
[alex at lancelot pronu]$


Alex




From bdelmee at advalvas._ZAPME_.be  Wed Apr  2 15:32:17 2003
From: bdelmee at advalvas._ZAPME_.be (Bernard Delmée)
Date: Wed, 2 Apr 2003 22:32:17 +0200
Subject: intro to OOP using Python, tutorial?:newbie
References: <3E8A5AE4.1080902@netscape.net> <b6el8v$7d7$2@nntp0.reith.bbc.co.uk>
Message-ID: <3e8b4828$0$6529$afc38c87@sisyphus.news.be.easynet.net>

>
> How to Think Like a Computer Scientist: Learning with Python
>
> ... is a gentle introduction to Python, OO and various other things.
> It's available as HTML, PDF, or LaTeX from:
>    http://greenteapress.com/thinkpython.html
>

<plug>
It's also available as a handy .CHM file if you're on Windows:
http://www.bdelmee.easynet.be/CHMs/thinkCSjava_chm.zip

Along with "Learning to Program" by Alan Gauld
http://www.bdelmee.easynet.be/CHMs/pyTutor.chm
</plug>





From meinmailfilter at gmx.de  Tue Apr 15 01:38:07 2003
From: meinmailfilter at gmx.de (HHaegele)
Date: 14 Apr 2003 22:38:07 -0700
Subject: class too big for one file! How can I split it?
References: <e956b5f1.0304100512.597654d3@posting.google.com>
Message-ID: <e956b5f1.0304142138.286e3bdf@posting.google.com>

Thank you for your answers.
I decided to make a extra file for the global(s) and a subclass as
suggested by Lee. So the globals can be accessed the same way in all
methods and the calling of the methods from outside doesn?t differ.

HHaegele



From do-not-spam-ben.hutchings at businesswebsoftware.com  Tue Apr  8 14:01:52 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 8 Apr 2003 18:01:52 GMT
Subject: Logging into CVS using Python
References: <ShBka.143$N5.26584@mencken.net.nih.gov>
Message-ID: <slrnb963kf.t8.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <ShBka.143$N5.26584 at mencken.net.nih.gov>, Mike wrote:
> Howdy All,
> 
> Thanks to everyone who replied to my previous post, even though my post was
> horribly worded. Now onto my question.
> 
> How do I use Python to login into CVS? Here's what I have tried, but it has
> failed. I'm not even sure that I'm using the correct methods.
<snip>

cvs login does little other than storing the supplied password in ~/.cvspass
(or some other chosen password file).  Subsequent operations automatically
retrieve the password from there.  It doesn't seem to be possible to pass
a password to cvs login programmatically, except perhaps with a pseudo-tty
under Unix.  What I ended up doing when I tried to automate CVS from Python
was to edit ~/.cvspass instead.  If you want to try this, read the source
files login.c and scramble.c for the details of the format of that file.



From hawkeye.parker at autodesk.com  Wed Apr  9 15:08:23 2003
From: hawkeye.parker at autodesk.com (haughki)
Date: 9 Apr 2003 12:08:23 -0700
Subject: windows-cmd shell doesn't reflect _winreg changes
References: <b6e56f7a.0304021609.521700e6@posting.google.com> <slrnb8p06l.174.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <b6e56f7a.0304091108.6030e453@posting.google.com>

Many thanks Ben!!




Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> wrote in message news:<slrnb8p06l.174.do-not-spam-ben.hutchings at tin.bwsint.com>...
> In article <b6e56f7a.0304021609.521700e6 at posting.google.com>, haughki wrote:
> > hi
> > 
> > i'm using _winreg to 'hard-change' and environmental variable in
> > windows.  the change is 'successful': regedit correctly reflects the
> > changes, and crtl panel > system > properties envronmental vars have
> > been correctly changed.  but, when i start a cmd shell (dos), the
> > shell evironment still reflects the old values until i manually open
> > the env vars cntrl panel and click 'ok'.  anyone?
> 
> This is really a Win32 programming issue rather than a Python one.
> However, I do vaguely remember this one.
> 
> Environmental variables are normally passed to a new process by the
> process that creates it, and they can never be updated from outside
> a process.  The exception to this is that Windows Explorer (and any
> other program that wishes to) will read those initial environment
> variables out of the registry when notified of a change in settings.
> The changes will then be passed on to any program started by Explorer
> afterwards.
> 
> Here's the Python code:
> 
>     import win32con, win32gui
>     win32gui.SendMessage(
>         win32con.HWND_BROADCAST, win32con.WM_SETTINGCHANGE, 0, 0)
> 
> You might want to use SendMessageTimeout instead of SendMessage.



From mickey at tm.informatik.uni-frankfurt.de  Tue Apr 29 09:07:03 2003
From: mickey at tm.informatik.uni-frankfurt.de (Michael 'Mickey' Lauer)
Date: 29 Apr 2003 13:07:03 GMT
Subject: Python on Sharp Zaurus SL-5500 PDA
References: <mailman.1047174074.9142.python-list@python.org>
Message-ID: <b8ltdn$baqnb$1@ID-192111.news.dfncis.de>

Lance Ellinghaus <lellinghaus at yahoo.com> wrote:
> Has anyone ported the latest version of Python to the Sharp Zaurus
> SL-5500 PDA? If so, where can I get it?

Besides the aforementioned feed from Phil Thompson, you may want to look
into my feed @ http://opie.net.wox.org/python which is differently structured.
It is labelled as Python for OpenZaurus, but it works also with the Sharp ROM.

Cheers,

Mickey.




From tim.one at comcast.net  Tue Apr  8 15:48:19 2003
From: tim.one at comcast.net (Tim Peters)
Date: Tue, 08 Apr 2003 15:48:19 -0400
Subject: Pathname problem
In-Reply-To: <3e9321b5$1@rmcnt03a.rmc.ca>
Message-ID: <BIEJKCLHCIOIHAGOKOLHAEGDFEAA.tim.one@comcast.net>

[bourassa]
> Hi there.  I am just trying to load files using PIL.  I get the following:
>
> im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
> Traceback (most recent call last):
>    File "<pyshell#13>", line 1, in ?
>      im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
>    File "C:\Python22\Lib\site-packages\PIL\Image.py", line 1479, in open
>      fp = __builtin__.open(fp, "rb")
> IOError: [Errno 2] No such file or directory:
> 'C:\\MATLAB6p1\toolbox\\pout.tif'
>
> I have stepped through and if the file in C:\ or C:\MATLAB6p1 it loads
> with no problem.  The minute I try to get it out of "toolbox" (which
> exists) I get the above error.  I'm using  PIL-1.1.4a3.win32 on a Win2K
> PIV.  TIA.

"\t" is a tab character, not a backslash followed by the letter t.  Easiest
is to use forward slashes:

    im = Image.open("C:/MATLAB6p1/toolbox/pout.tif")

You could instead use a raw string:

    im = Image.open(r"C:\MATLAB6p1\toolbox\pout.tif")

or double your backslashes:

    im = Image.open("C:\\MATLAB6p1\\toolbox\\pout.tif")

or paste path components together via os.path.join() for a wholly portable
approach.





From cswiger at mac.com  Sun Apr 13 16:47:28 2003
From: cswiger at mac.com (Chuck Swiger)
Date: Sun, 13 Apr 2003 16:47:28 -0400
Subject: popen(), sendmail: Success?
In-Reply-To: <3e99b0f8.10658656@news.muenster.de>
References: <3e95d63f.2635484@news.muenster.de> <3e99b0f8.10658656@news.muenster.de>
Message-ID: <3E99CCE0.5050200@mac.com>

Martin Bless wrote:
[ ... ]
> Unfortunately Puretec, a big german provider with several million
> accounts changed the rules recently when they moved to their new and
> wonderful computing centre. All of a sudden 
> S=smtplib.SMTP('any.smtp.server') would bring about error 111, 
> "connection refused".

Sounds like the ISP is blocking outbound SMTP via a firewall.

> Connecting to 'localhost' is an interesting idea I hadn't thought of
> - I'll give it a try. But I'm not very optimistic.

Ask Puretec's tech support which SMTP server you should use for your 
program to send mail.  Let your ISP worry about it from there.

[ ... ]
> (3) Looks like I only have to use popen2.popen3(...) if I want to
> retrieve sendmails stdout and stderr streams. Maybe I give it a try
> and do some experimentation on what's possible in this specific cgi
> environment.

Yes.  You might run "sendmail -v -t _msg_" by hand to see whether the 
additional debugging info generated is worth the coding effort.

-- 
-Chuck






From aleax at aleax.it  Tue Apr 15 18:59:12 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 22:59:12 GMT
Subject: book on Python for Fortran programmers
References: <3064b51d.0304150458.595afc30@posting.google.com> <vhbo9vgvsqjg8l7s0aluijrran56cg8mlm@4ax.com> <pan.2003.04.15.20.55.25.783394@mn.rr.com>
Message-ID: <490na.26787$T34.799064@news2.tin.it>

Anna wrote:

> On Tue, 15 Apr 2003 12:15:18 +0000, Michael Prager wrote:
> 
>> beliavsky at aol.com wrote:
>> 
>>>Python has been suggested by several posters to comp.lang.fortran as a
>>>scripting language worth learning. What are some good book(s) and web
>>>site(s) to get an experienced Fortran 95 programmer started in Python?
   ...
> Well, if you're looking for up-to-date, from the reviews I've seen,
> _Python in a Nutshell_ would be worth looking at for this. I've gotten the
> impression that it's particularly suited to experienced programmers.

I am, of course, quite biased in favour of the Nutshell (guess why;-),
but I do think it might be particularly helpful to Fortran programmers.

Being a Fortran programmer IS, after all, a substantial part of my own
background -- Fortran was my first language, the one I wrote my thesis in,
the main one I taught when I served as a university professor, etc, etc).

If you're interested in numeric programming, as many Fortran programmers
are likely to be, the chapter on Numeric might be helpful, for example.
(Do also look at the feedback about it that was posted on this NG -- I
haven't yet examined that feedback but it seems to be pointing out a
couple of important points that have to be added to that chapter!!!).

I would suggest having a look at the sample chapter, and table of
contents, that you can get from the O'Reilly site (or amazon.com, etc).

The chapter that the editor chose to make available as a sample is an
early but crucial one, summarizing the essentials of the language
itself (except for exceptions, modules/packages, and OO stuff, which
each have their own chapter) -- so, I think it might help you judge
whether the book can best meet your needs.


Alex




From michael at stroeder.com  Fri Apr 11 03:56:47 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Fri, 11 Apr 2003 09:56:47 +0200
Subject: popen(), sendmail: Success?
In-Reply-To: <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de>
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de>
Message-ID: <v29im-ia2.ln1@nb2.stroeder.com>

Carsten Gaebler wrote:
> If your code appears in a CGI script, you should catch the exception and
> print it to stdout because otherwise the error would go to stderr which
> will not be captured by the webserver.
 > [..]
 > except:
 >   print "something went wrong:", sys.exc_type, sys.exc_value
 > [..]

For security reasons I'd be scared of printing an arbitrary exception 
message to stdout in a web application since it is sent to the user's browser!

stderr *is* captured by the web server and most times written to an error 
log. That's where those error messages belong. If you don't have access to 
the error log redirect stderr to a local web application log file.

Ciao, Michael.




From sismex01 at hebmex.com  Wed Apr 16 10:36:14 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Wed, 16 Apr 2003 09:36:14 -0500
Subject: How many of you are Extreme Programmers?
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com>

> 
> So how many of you guys use XP processes?
> 

Interesting question.

Many of us have used XP techniques without actually knowing
that our methodology was labelled as "XP", it comes natural
in the scope of our tools and goals.

Tool:   Python.
Goal:   Write specs-conforming programs and modules.
Method: Create tests which verify the code.

I don't know; "OOP", "XP", "Structured Programming", etc.. are
all monickers, buzzwords, fads, which come and go; they come
into existance when some mass-market-oriented technical book
writer says "Hey, this programmer is successful, let's see
what he's doing"; once done, the writer creates some catchy
name for this guy's techniques, and sells his book.

I really dislike market speak, as you can tell. :-)

-gustavo




From lorenzo at mysurname.net  Wed Apr  9 09:38:19 2003
From: lorenzo at mysurname.net (Lorenzo Bolognini)
Date: Wed, 09 Apr 2003 13:38:19 GMT
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de>
Message-ID: <fnVka.21654$DT4.631328@twister1.libero.it>

"Johannes Eble" <skywalkerpackage at hotmail.com> ha scritto nel
messaggio news:3e93e50f.676072541 at news.online.de...

> I would like to draw UML diagrams.

Have a look at this
http://pyut.sourceforge.net/

Lorenzo





From alanmk at hotmail.com  Sat Apr  5 07:42:58 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Sat, 05 Apr 2003 13:42:58 +0100
Subject: What are the differences between _weak_ references and "normal" 
 references?
References: <169a7283.0304040901.4bc3531b@posting.google.com>
Message-ID: <3E8ECF52.CE3231FF@hotmail.com>

sdieselil wrote:

> I noticed a module weakref in Python standard library. What is weak
> reference and in what way it differs from "normal" reference? By the
> way, how can I implement usual reference in Python (I mean C++
> reference)?

A good explanation of weak references, albeit with reference to that
other VM-based language, Java (ugh! ;-) can be read at the Sun website.

http://developer.java.sun.com/developer/technicalArticles/ALT/RefObj/index.html

Reading this article should greatly assist your understanding.

I've found weak references very useful (in Java) when implementing cache
mechanisms. When caching an object for later reuse (say file contents),
to save yourself the time and effort of reading and parsing the file
again, reference the object with a weak reference. If the VM runs out of
memory, it can claim the memory belonging to your object, because it is
only referenced with a weak reference. When you go to actually reference
the (reclaimed) object, you then discover that the object is gone, and
that you have to reload it.

Can anyone else enlighten us as to other uses for weak references?

Regards,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From sholden at holdenweb.com  Thu Apr 17 08:02:31 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 12:02:31 GMT
Subject: PythonCard vs. Boa Constructor (was Re: Python for VB6 programmers)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com> <6e07b825.0304162200.34884d36@posting.google.com> <b7lhdg$q0f$1@panix3.panix.com> <Zvvna.38978$T34.1112133@news2.tin.it>
Message-ID: <rJwna.114821$It5.70735@news2.central.cox.net>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:Zvvna.38978$T34.1112133 at news2.tin.it...
> Aahz wrote:
>
> > In article <6e07b825.0304162200.34884d36 at posting.google.com>,
> > yaipa h. <yaipa at yahoo.com> wrote:
> >>
> >>Here in the SF Bay Area I have only met one VB programmer willing to say
> >>that he uses VB.net. The other VB programmers I find are talking about
> >>either sticking with VB6 (most), taking Java courses (some) or playing
> >>around with the C# idea (few).  Most share my feelings about the .net
> >>thing. That it will not be stable until release +3.x and there are
> >>little to no existing libraries. You know, batteries. ;-) So yes there
> >>is a window and it is wide open.
> >>
> >>With that said, without a slick, stable GUI builder with all the
> >>look and feel of the VB GUI builder it's a long way from here
> >>to there. I mean what do you expect them to do, pass in command-line
> >>arguments or hand assemble TK widgets. :^)
> >
> > Good point.  Anyone here willing to take a whack at claiming that
> > PythonCard, Boa Constructor, or something else would the best path for a
> > VB6 programmer?
>
> Personally I would suggest BlackAdder -- it's a commercial product,
> but not too costly, and while still Beta it's IMHO more stable than
> Boa (haven't tried PythonCard, natch).  The look and feel of each
> of these IDE's is obviously hugely different from that of VB, though.
>
> Not sure if ActiveState's "Visual Python" add-on for Microsoft Visual
> Studio .NET can provide sufficient functionality, as I haven't tried
> it yet -- at least the MS look and feel should be there, though,
> since it DOES work on top of MS's IDE...!-)
>

I'd be happier with that is there was ANY sign that work had been done on
BlackAdder in the past two years.While I don;t disagree with anything else
you wrote, I'm still waiting for the clock to start running on the year's
support I bought. This will apparently happen when it "comes out of beta".
Meantime, two years after my purchase, the product hasn't been updated.

The Kompany won't be getting any more of MY money, thank you very much.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From duncan at NOSPAMrcp.co.uk  Tue Apr 15 05:10:25 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 15 Apr 2003 09:10:25 +0000 (UTC)
Subject: Case-insensitive string comparison
References: <3e9b09bc@news.012.net.il> <Y%Dma.13754$LB6.370780@news1.tin.it> <3e9b0e03$1@news.012.net.il>
Message-ID: <Xns935E66BE3E498duncanrcpcouk@127.0.0.1>

"Avner Ben" <avner at skilldesign.com> wrote in
news:3e9b0e03$1 at news.012.net.il: 

>> >     Does standard Python have case-insensitive string comparison
>> >     and search?
>>
>> No.  You normally accomplish them either with regular expressions
>> (which CAN be compiled into case-insensitive mode) or by coercing
>> strings to one fixed case (e.g. with the .lower method of string
>> objects). 
>>
> 
>     I have tried the lower them all approach, but at the rate of
>     processing 
> I need, it has turned out to reduce application speed by a factor of
> four. What I need is an efficient built-in way of doing this.
> 

That seems very strange to me as lower casing the strings has the
potential to be much more efficient than a case insensitive compare even
when the compare is done by a C routine. Perhaps if you posted a general
outline of your code we could suggest ways to speed it up. The biggest 
potential for code speedup is often to use a different algorithm rather 
than trying to optimise individual operations.

Some ideas to consider:

Avoid lowercasing the same string multiple times. If you are doing more
than one comparison on a string, then lowercasing it once before doing
all the comparisons is better whether you are coding in Python or C. 

Are you using dictionaries enough? If you are looking for matching
strings then a dictionary lookup will beat multiple string comparisons
any day. 

Are you sorting the strings? If so, do your best to avoid using a
comparison function, use a decorated sort. This has two benefits, it
means you only do the lowercasing once, and it removes the Python call
overhead from inside the sort. 

-- 
Duncan Booth                                            
duncan at rcp.co.uk int month(char
*p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3" 
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure? 



From the_3_project at yahoo.com  Tue Apr 29 09:34:12 2003
From: the_3_project at yahoo.com (Joe Green)
Date: 29 Apr 2003 06:34:12 -0700
Subject: simplest question on win32com.client
Message-ID: <a5b86892.0304290534.31528b9e@posting.google.com>

I can't seem to get this code to run.


--
from win32com.client import Dispatch

xl = Dispatch("Excel.Application")
--

It comes back with:

The instruction at "0x1e039e79" referenced memory at "0x1e3840fc". The
memory could not be "read".

Click OK to terminate the program
Click on CANCEL to debug the program


Thanks



From ramen at lackingtalent.com  Tue Apr 15 22:05:58 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Wed, 16 Apr 2003 02:05:58 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <yfsistfb5n9.fsf@black132.ex.ac.uk>
Message-ID: <slrnb9pekp.vg8.ramen@lackingtalent.com>

In article <yfsistfb5n9.fsf at black132.ex.ac.uk>, Alexander Schmolck wrote:
> Well it seems to me that you're currently *really* keen on functional
> programming. You can do that (and learn about it) much better in scheme than
> in python, so even if you want to program in python for the rest of your life
> afterwards it might well be worth it. I mean you wouldn't want to learn OO in
> C, would you -- there a fewer good ressources and it's just much more of a
> pain (this comparison is of course overly pessimistic -- C is just horrible
> and python really isn't).

Point taken. I followed that lecture notes link, by the way. Thanks for
sending that. It's interesting how the author begins by describing a GIF
viewer as an interpreter.

>> That's true, but then I have to name "bar". And part of the joy of FP is not
>> having to name *everything*. Naming things is important if you need to call
>> them up later, but it's a hassle to name intermediate values
> 
> No argument here. I tend to think python would have been a better language if
> it had included something like smalltalk-like blocks from day 1 (I still
> really like python).

Are those anything like code blocks in Ruby? I think that's a cool idea...

> Yes but certainly not due to dynamic scoping (which python doesn't even have,
> as you can see from the raging Emacs thread). It now has non-crippled lexical
> scope :)

Thanks, I stand corrected. ;)

Peace,
Dave



From BPettersen at NAREX.com  Mon Apr 21 13:11:13 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Mon, 21 Apr 2003 11:11:13 -0600
Subject: Stackless 3.0 alpha 1 at blinding speed
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE030@admin56.narex.com>

> From: David Abrahams [mailto:dave at boost-consulting.com] 
> 
> "Bjorn Pettersen" <BPettersen at NAREX.com> writes:
> 
> >> From: David Abrahams [mailto:dave at boost-consulting.com] 
> >
> >> Also, "stacklessness" is really an implementation detail 
> >> AFAICT - it just says that all execution context is 
> >> stored in dynamically-allocated structures instead of 
> >> using local variables on the 'C' stack (via a recursive-
> >> calling implementation).
> >> 
> >> Have I misunderstood this whole thing?
> >
> > Which property exactly makes the C stack frames not dynamically
> > allocated? (the sequential property of the allocator, or the 
> > fixed size of the stack itself?)
> 
> Sorry; I guess they're dynamically allocated... on the stack.  IOW,
> not on the heap and with lifetime directly tied to the location of the
> program counter.  That's a more accurate way to say what I meant.

My apologies for being pedantic. Getting to this definition, it's much
easier to see that the lifetime issue is unrelated to the stack itself,
but rather to the algorithm (semantics) of function invocation (instead
of e.g. the lifetime [reachability] of the objects created there). In
other words, coming from a different background than the Algol lineage,
one can easier argue that there can exist other invocation semantics
that are more natural, although they don't map directly onto the HW
architecture...

-- bjorn




From exarkun at intarweb.us  Tue Apr 22 15:23:17 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 22 Apr 2003 15:23:17 -0400
Subject: Python Thread Question
In-Reply-To: <5b04353b.0304220351.2034a26@posting.google.com>
References: <bd993a2f.0304170534.245aa2d5@posting.google.com> <mailman.1050587609.6119.python-list@python.org> <b7meoo$mf6$1@panix2.panix.com> <5b04353b.0304220351.2034a26@posting.google.com>
Message-ID: <20030422192317.GA25856@meson.dyndns.org>

On Tue, Apr 22, 2003 at 04:51:20AM -0700, N.K wrote:
> aahz at pythoncraft.com (Aahz) wrote in message news:<b7meoo$mf6$1 at panix2.panix.com>...
> > 
> > This isn't quite true on a multi-CPU box.  Because I/O releases the GIL,
> > the OS can schedule threads to run on all available CPUs.  So it's
> > certainly possible for a threaded app to outrun non-blocking I/O.
> 
> 
>   Heard that new Intel Processors support Hyper Threading ( HT)
> Architecture from which multi-threaded apps can benefit from. Is that
> true?

  Some (All?) P4s have this technology.  Practically speaking, this means
applications see the system as having multiple processors.  More than one
thread can run at once.  So yes, multi-threaded apps benefit from this in
much the same way they benefit from traditional SMP systems.

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 33 days, 15:03, 2 users, load average: 0.39, 0.46, 0.44




From kseehof at neuralintegrator.com  Thu Apr  3 10:43:49 2003
From: kseehof at neuralintegrator.com (Ken Seehof)
Date: Thu, 03 Apr 2003 07:43:49 -0800
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304030705.524cf0fa@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com>
 <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>
Message-ID: <5.2.0.9.0.20030403074232.01edc548@pop.hurrah.com>

At 07:05 AM 4/3/2003 Thursday, Drew Moore wrote:
>"Anders J. Munch" <andersjm at dancontrol.dk> wrote in message 
>news:<3e8c083b$0$10386$edfadb0f at dread11.news.tele.dk>...
> > "Drew Moore" <drew at astro.pas.rochester.edu> wrote:
> > > Howdy!
> > >
> > > I submitted a pre-PEP on overloadable assignment.
> > >
> > > The idea is:
> > >
> > > In situations where the assignment token "=" appears, and an augmented
> > > assignment symbol (such as +=, -=, *=, etc..) would also be
> > > syntactically correct, the assigned-to object would be checked for an
> > > __assign__(self, other) method.
> >
> > You terminology is off.  There is no such thing as an "assigned-to"
> > object in Python.  Assignments change bindings, not objects.
>
>Right, the binding of the name always changes. When an augmented
>assignment operator is overloaded, the method must return a
>reference to an object, and this returned reference is bound to the
>name on the left hand side. Typical code inspects the "other" object,
>modifies self appropriately, and returns a reference to self.
>Regular assignment is just the trivial case of augmented assignment.
>
>Seems to me a += b -> __iadd__(a,b)
>is very similar a = b -> __assign__(a,b)
>
>the __assign__ method returns a reference, just like __iadd__
>this reference will be bound to the name on the left hand side.
>
> > Take a step back and tell us what problem you are trying to solve.
> > Whatever it is, I'm sure we can think of a better solution than having
> > assignment depend on whatever object, if any, happened to be
> > previously bound to the same name.
> >
> > python-is-not-c++-ly y'rs, Anders
>
>my original need?
>I wanted to create an object that controls a voltage and
>use it at the python command line.
>by overloading operators, I can do:
>
>voltage += 5  # (raise the voltage by 5 volts)
>voltage *= 2  # (double the current voltage)
>
>but when I do
>
>voltage = 3  # (set the voltage to 3 volts)
>
>my voltage object is clobbered with an integer object.
>I don't want to require the user to type any more than
>name = value # no name.val = newval, no name.setval(value)
>
>Overloadable augmented assignment provides a  nice framework that
>always rebinds the name, but gives the "about to be rebound" name
>some say in how this takes place. I'm at a loss to explain why
>the "most trivial case of augmented assignment" was denied this power.
>
>Thanks for responding.. Maybe posting on April 1st was a mistake,
>people might have thought I was joking around!!


You are thinking like a C programmer.  Python assignment is fundamentally
different from C assignment.  You probably know this.

In the statement a = b, the python interpreter simply does not look at the
previous value associated with `a' (except to decrement the reference count).
Therefore, it is not possible for the interpreter to check to see if `a' has an
__assign__ attribute.  In order for this to change, pythons concept of name 
binding
would have to completely change, and I doubt Guido would be up for that.
One effect would be to make python significantly slower (since all assignment
operations would have to be preceded by an extra method search for __assign__).

Historically, the addition of augmented assignment to python was controversial
because of the belief that it tends to confuse people by making assignment seem
to be more like other languages such as C.

Also, the notion of hiding information in order to simply reduce the amount of
typing is a dubious (though attractive) programming technique.  It is better to
have more typing if it means that the code is easier to read and debug.  I
recommend that you make voltage an property of a class.  This way you can
overload assignment any way you like.

- Ken






From cjw at sympatico.ca  Thu Apr 17 12:40:35 2003
From: cjw at sympatico.ca (Colin J. Williams)
Date: Thu, 17 Apr 2003 12:40:35 -0400
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net>
Message-ID: <3E9ED903.79B5F164@sympatico.ca>

user at domain.invalid wrote:

>  From the python documentation tutorial, section 8.5
> +++++++++++++++++++++++++++++++++++=
>  >>> class MyError:
> ...     def __init__(self, value):
> ...         self.value = value
> ...     def __str__(self):
> ...         return `self.value`
> ...
>  >>> try:
> ...     raise MyError(2*2)
> ... except MyError, e:
> ...     print 'My exception occurred, value:', e.value
> ...
> My exception occurred, value: 4
>  >>> raise MyError, 1
> Traceback (innermost last):
>    File "<stdin>", line 1
> __main__.MyError: 1
>
> +++++++++++++++++++++++++++++++++++=
>
> What do the back ticks in method "__str__" mean?

The back ticks, as in `self.value`,  are equivalent to repr(`self.value`).

ie. it provides the  ``official'' string representation of self.value.

Colin W.





From peter at engcorp.com  Fri Apr 18 09:25:09 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 18 Apr 2003 09:25:09 -0400
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com>
Message-ID: <3E9FFCB5.79AEA22E@engcorp.com>

Aahz wrote:
> 
> In article <mailman.1050587068.2733.python-list at python.org>,
> Jp Calderone  <exarkun at intarweb.us> wrote:
> >
> >--
> >It is practically impossible to teach good programming style to
> >students that have had prior exposure to BASIC: as potential
> >programmers they are mentally mutilated beyond hope of
> >regeneration.        -- Dijkstra
> 
> I like to believe that I'm a counter-example to this generalization.

I too learned BASIC first (from a book which I found in the public 
library, a book with lots of pictures and big fonts and stuff... 
I forget what it was called though :-( ).  I also hope my code is more 
an example of good style than of bad.

Nevertheless, we might give Dijkstra the benefit of the doubt and
accept that a *student* who had exposure to BASIC might be 
unteachable as far as improving her style goes.  Perhaps only once
she advances beyond the student stage, and has some years of actual
_practice_, will she discard the early mental mutilation of BASIC
and find her way to Good Style.

(Though I doubt Dijkstra intended to allow this loophole in his
potential "law". :-)

-Peter



From bourassa-m at rmc.ca  Thu Apr 10 14:19:02 2003
From: bourassa-m at rmc.ca (bourassa)
Date: Thu, 10 Apr 2003 14:19:02 -0400
Subject: Python compared toJava
In-Reply-To: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
Message-ID: <3e95b6fa$1@rmcnt03a.rmc.ca>

If you go to:

www.python.org

you should find a lot of info to put you in the picture of what Python 
is about.  You'll also be able to find details on the virtual machine.  HTH

majb

Blaktyger wrote:
> Hi!
> 
> I recently read that the Python interpreter is like the Java Virtual Machine.
> Is that true?




From bdesth at removethis.free.fr  Thu Apr 17 17:43:04 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Thu, 17 Apr 2003 21:43:04 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <871y01f2hw.fsf@deneb.enyo.de>
Message-ID: <3E9F1FE8.5030409@removethis.free.fr>

Florian Weimer wrote:
> laotseu <bdesth at removethis.free.fr> writes:
> 
> 
>>Lispers friends, let's have you're opinion on that point : is Common
>>Lisp a functionnal language ?
> 
> 
> It depends on your definition of "functional programming language".
> In the more general sense, any language with higher-order functions
> and functions as a first-class type is a functional programming
> language.  

Ok, the question coming from a 'Functional Python' point of view, the 
'more general sense' is enough for me.

And for you all dear Lispers, thanks for your opinions on that point. I 
know enough of lisp to know that it supports many paradigms and doesn't 
enforce a pure, religious FP style.

I notice that nobody here claimed that you could not do functional 
programming in Common Lisp.

 > In the strictest sense, the language has to be purely
 > applicative (or free of side effects).  Common Lisp satisfies the
 > first definition, but not the second.

Do you mean that it is *absolutely not possible* to write a whole 
program in CL whithout any side effect ? Or that many features in the 
language, that you can choose to use or not, are not side effect free ?

I might say that, in the first case, this does not prevent CL from being 
(also) a functional language in the 'strictest sens'.

Laotseu




From msg_2222 at yahoo.com  Mon Apr  7 13:16:24 2003
From: msg_2222 at yahoo.com (Rick Y)
Date: Mon, 7 Apr 2003 10:16:24 -0700 (PDT)
Subject: socket question
Message-ID: <20030407171624.59654.qmail@web20701.mail.yahoo.com>

how can i enable _sockt module in my solaris python?. 
i did not build it. Downloaded it from sunfreeware. 

./viewcvs-install
Traceback (most recent call last):
  File "./viewcvs-install", line 35, in ?
    import compat
  File "./lib/compat.py", line 20, in ?
    import urllib
  File "/usr/local/lib/python2.1/urllib.py", line 26,
in ?
    import socket
  File "/usr/local/lib/python2.1/socket.py", line 41,
in ?
    from _socket import *
ImportError: ld.so.1: python: fatal: libssl.so.0.9.6:
open failed: No such file or directory


__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com




From verlee at web.de  Wed Apr  2 15:02:02 2003
From: verlee at web.de (Verlee)
Date: Wed, 2 Apr 2003 22:02:02 +0200
Subject: Some newbie questions the FAQs didn't answer!
In-Reply-To: <6ffd97db.0304021132.62f86c7b@posting.google.com>
References: <6ffd97db.0304021132.62f86c7b@posting.google.com>
Message-ID: <18429949632.20030402220202@web.de>

Hi,

> [...] But then I read that in order to run a python program, the
> machine had to have the python VM installed, and > the user had to go
> to a command line and type "python <program name>>.py".

> What I had planned to do was write Win32 GUIs, but I would guess that
> most common users don't have Python installed.  Is there some way to
> write the program in C++ and then have it bring up the GUI written in
> Python without Python being installed?  Wouldn't that require the code
> being compiled into machine code?

first of all I should confess that I?m a newbie too, so I may be
wrong... anyhow, I?ll just try to help with that tiny little knowledge
of mine ;-)

There is a tool called py2exe that converts .py-files to .exe-files
which can be executed without having the Python interpreter installed.
I figure this is what you want to have?
(I?ve been worrying a lot about this, too...)
It can be downloaded at http://sourceforge.net/projects/py2exe/
Haven?t had time to try it yet, though. There may also be other tools
that do the same, but this is the (only) one I have heard of.

As for the C++ thing, I don?t have any idea...

Verlee





From janeaustine50 at hotmail.com  Sun Apr 13 12:50:30 2003
From: janeaustine50 at hotmail.com (Jane Austine)
Date: 13 Apr 2003 09:50:30 -0700
Subject: How to subclass datetime?
References: <ba1e306f.0304130339.3bece039@posting.google.com>
Message-ID: <ba1e306f.0304130850.1db6bf69@posting.google.com>

> <posted & mailed>
> 
> Jane Austine wrote:
> 
> > Running Python 2.3a2 on win XP.
> > 
> > I'm interested in the new datetime module and I want to subclass
> > date/datetime class to modify the behaviour. However, simply
> > inheritting the classes(like "class mydate(date):" ) didn't work.
> 
> Please define "didn't work"...:
> 
> >>> import datetime
> >>> class mydate(datetime.date):
> ...   pass
> ...
> >>> xx=mydate(2003,4,14)
> >>> print xx
> 2003-04-14
> >>>
> 
> Seems to be working just fine.  What exactly isn't working for you?
> 

For example,

>>> class mydate(datetime.date):
	def abc(self):
		print "abc"

		
>>> z=mydate(2003,4,14)
>>> z.abc()
Traceback (most recent call last):
  File "<pyshell#6>", line 1, in -toplevel-
    z.abc()
AttributeError: 'datetime.date' object has no attribute 'abc'
>>> type(z)
<type 'datetime.date'>
>>> z.__str__()
'2003-04-14'
>>> class mydate(datetime.date):
	def __str__(self):
		return "abc"

	
>>> z=mydate(2004,4,14)
>>> z
datetime.date(2004, 4, 14)
>>> print z
2004-04-14
>>> z.__str__()
'2004-04-14'

I reckon it is not subclassing the datetime.date class in
its proper definition of subclassing.

> 
> > Should I resort to delegation?
> 
> Maybe, but it's hard to say unless you can be more specific!
> 
> 
> Alex



From jeremy at jdyallop.freeserve.co.uk  Fri Apr 11 14:45:39 2003
From: jeremy at jdyallop.freeserve.co.uk (Jeremy Yallop)
Date: 11 Apr 2003 18:45:39 GMT
Subject: module for doing unix "find"-like things?
References: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com>
Message-ID: <b772gj$c11gs$1@ID-114079.news.dfncis.de>

Mark Harrison wrote:
> Is there such a thing?  I didn't see such a thing, but
> "find" is not the easist thing to search for.

os.path.walk is often useful in such circumstances.

> Specifically, I would like to replace an external call to
>     for i in `find $top -type l -maxdepth 1`; ...

With '-maxdepth 1' you might as well use os.listdir().  Perhaps
something like the following is what you're looking for:

  import os

  top = os.getenv('top')
  for i in [l for l in os.listdir(os.getenv('top')) 
                    if os.path.islink(os.path.join(top, l))]:
     ...
 
Jeremy.



From timbutterfield at mindspring.com  Sat Apr 19 19:08:04 2003
From: timbutterfield at mindspring.com (Tim Butterfield)
Date: Sat, 19 Apr 2003 18:08:04 -0500
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net>
Message-ID: <3EA1D6D4.7040001@mindspring.com>

Ron Stephens wrote:
 > Still, I want to point out what a good time I am having with <a
 > href = "http://pythoncard.sourceforge.net/">PythonCard</a> as an
 > easy gui-builder for Python scripts.

Thanks for the good intro. :)

Yesterday, I was reading an LWN Daily posting about All-American 
(division of Mayflower Transit) and their move to Linux.  This got me 
thinking (again) about an aspect of Linux that may be making it more 
difficult for some companies to make a similar move.

Linux is is not really an easy target for developing applications. 
That is not to say it is hard.  It can just be more time consuming 
than developing a similar app for Windows.  One thing that is missing 
from the Linux development mainstream is a good development 
environment.  Emacs and vi are great for editing code, but are not as 
good for laying out a gui.  As much as some people dismiss them, 
language based IDEs like VB and Delphi made development of Windows 
applications less time consuming.  Borland's Kylix may try to do 
something similar for Linux, but so far, I haven't seen widespread 
acceptance of it.

Perhaps you can guess where I am going with this.  Python is a very 
useful language, as I am sure most of you will agree.  However, even 
though it supports GUI development, it is not yet a rapid process.

Let me digress here just a bit.  If it could continue supporting what 
it already does, how would people feel if python were to be the next 
VB?  The language has the potential for it.  The cross platform nature 
of wxPython doesn't hurt either.

When looking for a similar development environment, PythonCard seems 
to come the closest so far with the property editors and resouce 
files.  What I envision is an IDE for python similar to VB and Delphi 
(PyDE?) that allows for those property editors and resource files to 
be integrated into an editor/debugger and builder.  Of course, being 
able to add custom widgets and tie in data support to those widgets 
would be needed.  Is such a thing desired?

--
Tim Butterfield
http://www.timbutterfield.com/





From grante at visi.com  Fri Apr 25 10:32:51 2003
From: grante at visi.com (Grant Edwards)
Date: 25 Apr 2003 14:32:51 GMT
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth> <slrnbahk9n.ero.grante@localhost.localdomain> <7h3fzo6luno.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3ea94713$0$198$a1866201@newsreader.visi.com>

In article <7h3fzo6luno.fsf at pc150.maths.bris.ac.uk>, Michael Hudson wrote:
> Grant Edwards <grante at visi.com> writes:
> 
>> In article <98imn-ust.ln1 at grendel.myth>, Jim Richardson wrote:
>> 
>> > I have come across a weirdness. I have a number of floats, (lat long
>> > pairs) which I stuff into a dict, but when I take them out, my nice %2f4
>> > number, (say, 47.5894) is something long and *almost* right, like
>> > 47.589333333333333339. What gives?
>> 
>> That's floats.  The best you can hope for is "almost right".
> 
> Well, in some circumstances you can hope for a bit more than that.

Yes, but only the timbot and three other people on the planet
(who refuse to be named) know what those circumstances are
ahead of time.  The rest of us settle for being able to figure
out what went wrong afterwards.

> But unless you're very careful, this is probably a good rule of
> thumb.

-- 
Grant Edwards                   grante             Yow!  FIRST, I'm covering
                                  at               you with OLIVE OIL and
                               visi.com            PRUNE WHIP!!



From uptonothing at yahoo.com  Thu Apr 10 10:46:28 2003
From: uptonothing at yahoo.com (Mike)
Date: Thu, 10 Apr 2003 10:46:28 -0400
Subject: pyserial newbie questions
References: <gilles.celli-943DC4.10083810042003@webmail.restena.lu>
Message-ID: <8tfla.154$N5.27502@mencken.net.nih.gov>

"Gilles" <gilles.celli at ecgs.lu> wrote in message
news:gilles.celli-943DC4.10083810042003 at webmail.restena.lu...
> Hello folks,
>
> Since I am very new to Python I have the following question:
>
> I try to write a program in Python (on RHat 8.0 Linux) with the
> "pyserial" module which communicates to a device via RS232 serial cable.
>
> The program should do:
>
> 1. Set Com port  to 300,n,8,1
> 2. Send 20 NULL characters Chr(0)
> 3. Wait 10 ms ( may vary in your case)
> 4. Send a string "HELLO"
> 5. Wait to receive ?READY? string  from device
> 6. If above string OK SEND   ?START?
>
> I figured out how to set the port, send the 20 NULL characters :)
>
> But I have no idea how to "wait for 10ms", send the HELLO string and
> read the READY string...
>
> Anyone has a clue or source code sample ?
>
> Thanks a lot.
>
> Gilles

I don't know if this is the answer you're looking for but you can use the
sleep(secs) function in the time module.

<code>
import time

time.sleep(.01)  # sleep 10 ms
...
</code>





From mirandacascade at yahoo.com  Sat Apr  5 23:18:58 2003
From: mirandacascade at yahoo.com (Miranda Evans)
Date: 5 Apr 2003 20:18:58 -0800
Subject: constructor inquiry
References: <59e5b87.0304040917.12c15907@posting.google.com> <mailman.1049564846.5425.python-list@python.org>
Message-ID: <59e5b87.0304052018.2beaba33@posting.google.com>

Steven Taschuk <staschuk at telusplanet.net> wrote in message news:<mailman.1049564846.5425.python-list at python.org>...
 
> > 3) is the 'xl' object in the demo() method an object of class xlEv? 
> > if so, how could the xl object come into existence without the
> > constructor getting invoked?  if not, how could the xl object have
> > visibility to the type attribute?
> 
> You can start learning the answers to these questions by adding
> 	print type(xl)
> 	print isinstance(xl, xlEv)
> at an appropriate point.  What do these lines produce?

Thank you for the tip about the built-in type() and isinstance()
functions.  They reveal that the xl object is of type instance and
that xl is not an instance of the xlEv class.  This answers the first
part of question 3.  I'm still, however, not seeing a connection. 
Specifically,

a) in the OnSheetBeforeDoubleClick method the statement
        self.type = "dbl click" 
assigns the string "dbl click" to an attribute named 'type'; use of
the isinstance() function within the OnSheetBeforeDoubleClick method
reveals that self is an instance of the xlEv class, so I'm assuming
that 'type' is an attribute of the xlEv class
b) in the demo() method, the type attribute of the xl object has the
value "dbl click"
c) the xl object in the demo() method is not an instance of the xlEv
class but it has a type attribute just like the xlEv class has a type
attribute

So, the xl object is not an instance of xlEv class, but it knows about
the same type attribute that was assigned to an instance of the xlEv
class...that's the connection I'm not making.



From BPettersen at NAREX.com  Wed Apr 16 02:22:46 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 16 Apr 2003 00:22:46 -0600
Subject: DB API for ADO/DAO?
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDF55@admin56.narex.com>

> From: Jimmy Retzlaff [mailto:jimmy at retzlaff.com] 
> 
> Steve Holden (sholden at holdenweb.com) wrote:
> > "John Burton" <john.burton at jbmail.com> wrote in message
> ...
> >> The problem with odbc is that you need to manually set up data
> >> sources using the control panel which isn't something I want 
> >> the users to have to do.

Agreed, but you don't have to even in ODBC :-) I'm assuming that what
you're really looking for is a way to call:

  hr = SQLConfigDataSource(
         NULL,
         ODBC_ADD_SYS_DSN, 
         LPCTSTR(driverDescription), // I believe for display purposes
only
	   "DSN=foo\0 DATABASE=DbName\0 TRUSTED_CONNECTION=YES\0
SERVER=ServerName\0\0 ");

which will create a new system data-source called "foo". I don't
remember seeing it exposed in either odbc or mxODBC, and it's too late
for me to remember all (not to mention the best) way to call a random C
function from Python, so I'm leaving that as an excercise for the reader
<wink>.

-- bjorn




From ssthapa at classes.cs.uchicago.edu  Thu Apr 10 19:30:01 2003
From: ssthapa at classes.cs.uchicago.edu (Suchandra Thapa)
Date: Thu, 10 Apr 2003 23:30:01 GMT
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com>
Message-ID: <Z7nla.45$W4.8475@news.uchicago.edu>

sebastien <s.thuriez at laposte.net> wrote:
> Is there any PCA analysis tools for python ?
> If it does, do you have any idea on how well it would scale ?
> 
> I have already seen PyClimate (but it is not available for Windows
> which will be one of the target). Is there some LAPACK like packages ?

Have you considered using R (www.r-project.com) and RS/Python
(http://www.omegahat.org/RSPython/).  R is a statistical programming language
with libraries to handle most common statistical methods.  Using RS/Python
you can create R objects, populate them, and pass them into R functions within
python or vice versa.

-- 
----------------------------------------------------------------------------
Suchandra Thapa                       
s-thapa-11 at NOSPAMalumni.uchicago.edu  
----------------------------------------------------------------------------



From gabnash at hotmail.com  Wed Apr 23 09:37:55 2003
From: gabnash at hotmail.com (David)
Date: 23 Apr 2003 06:37:55 -0700
Subject: capturing output from os.system() in Windows?
References: <f405d9e7.0304230003.63baab05@posting.google.com> <Xns9366626B53C4Eduncanrcpcouk@127.0.0.1>
Message-ID: <f405d9e7.0304230537.4f9a1d31@posting.google.com>

Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in message news:<Xns9366626B53C4Eduncanrcpcouk at 127.0.0.1>...
> gabnash at hotmail.com (David) wrote in 
> news:f405d9e7.0304230003.63baab05 at posting.google.com:
> 
> Us os.popen, os.popen2, os.popen3 or os.popen4 according to your exact 
> needs.
> e.g. lines = os.popen('net view').readlines()
> 

That worked well! Thanks very much.



From simo_salminen at iki.fi  Wed Apr 23 02:26:16 2003
From: simo_salminen at iki.fi (Simo Salminen)
Date: Wed, 23 Apr 2003 06:26:16 +0000 (UTC)
Subject: unused args in GUI callbacks: pychecker
References: <20030422145218.6e9e6549.gry@ll.mit.edu> <3EA5B0C4.7ACB349@alcyone.com>
Message-ID: <slrnbaccg8.du4.simo_salminen@pikkukorppi.cs.tut.fi>

* Erik Max Francis [Tue, 22 Apr 2003 14:14:44 -0700]
> george young wrote:
>> I recall that C++ had a way to tell the compiler this.
> 
> C++ is just the language, so it was just a particular compiler that you
> are referring to with that feature.
> 

I think he ment this:
int foo(int){}  <- no formal parameter name

-- 
simo salminen



From usenet at thinkspot.net  Thu Apr 24 02:43:35 2003
From: usenet at thinkspot.net (Sheila King)
Date: Wed, 23 Apr 2003 23:43:35 -0700
Subject: Web Hosting
References: <2rjna.171$hT2.182473@news2.news.adelphia.net> <6Bjna.111823$It5.95226@news2.central.cox.net> <e5os9vg7aj3gfek0abt05s1bjh9pfvfvb3@4ax.com> <0Puna.38530$T34.1097507@news2.tin.it>
Message-ID: <b878f8.18s.1@kserver.org>

On Thu, 17 Apr 2003 09:51:56 GMT, Alex Martelli <aleax at aleax.it> wrote in
comp.lang.python in article <0Puna.38530$T34.1097507 at news2.tin.it>:

> Afanasiy wrote:
>    ...
> > I was under the impression mod_python was not safe for shared
> > environments. If this is true, I don't think you will find a
> 
> I'm not aware of such concerns; could you please provide an
> URL discussing the issue, or what gave you this impression?

I believe the concern is that, mod_python runs under the web server's user
ID, whatever that may be. Thus, in a shared environment, user1 running
mod_python and user2 also running mod_python will have the same file
permissions on files created or read by such processes.

This is why PHP has safe_mode.

There are ways around this problem, but they are not easily implemented. In
many cases, hosts run PHP with safe_mode off, and no other security
restrictions in place to protect clients files, not even realizing the
security holes they have in place on their server.



-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org/



From Anish.Mehta at enst-bretagne.fr  Mon Apr 21 05:59:55 2003
From: Anish.Mehta at enst-bretagne.fr (Mehta, Anish)
Date: Mon, 21 Apr 2003 11:59:55 +0200
Subject: Explanation of ttx.py program
Message-ID: <3EA3C11B.8020609@antares.enst-bretagne.fr>

Hello Everyone!

I am interesting in making some changes in ttx.py program written in 
python. I am attaching the file with email. Can someone explain what 
each function is for here in this piece of code.

I will be waiting for reponse.

Thanks in advance.

Regards,

Anish MEHTA
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: ttx.py
URL: <http://mail.python.org/pipermail/python-list/attachments/20030421/64b85fab/attachment.ksh>

From tungwaiyip at yahoo.com  Wed Apr 23 17:17:38 2003
From: tungwaiyip at yahoo.com (Tung Wai Yip)
Date: Wed, 23 Apr 2003 14:17:38 -0700
Subject: Off topic: What usenet news reader to use?
References: <mailman.1051036295.31460.python-list@python.org> <d97bavkeagbpfkvm3649725tptqhkskds9@4ax.com>
Message-ID: <6h0eavc1dgmgd4rphnaodgvh9o6qigs62g@4ax.com>

On Tue, 22 Apr 2003 20:55:13 +0100, Dale Strickland-Clark
<dale at riverhall.NOTHANKS.co.uk> wrote:

>"Wai-Yip Tung \(wtung\)" <wtung at cisco.com> wrote:
>
>>I haven't used Usenet news for a long time. Now that I have subscribed
>>to python-list I find that email does not scale well to the high volumne
>>of discussion. Instead of deleting emails all day I think I should go
>>back to Usenet news. Is there any good Windows news reader you would
>>recommend. I'm happily using Opera as the browser. But it's
>>revolutionary news and email reader eludes me. What are you using?
>>
>>By the way I'm also looking for a free or low cost email client. My
>>company is using Outlook but I want to find one for personal use. Thank
>>you for you suggestions.
>
>You don't say whether you're interested in an offline or online reader
>so here's both. In no particular order, have you tried:
>
>* Outlook Express
>* Free Agent
>* Mozilla
>* XNews
>* And what's wrong with the one in Opera 7.10?

I get back Free Agent, which I have used years ago. It still work
well! The comp.lang.python looks much more usable compare to emails. I
have renew appreciation of the Usenet news.

Actually Opera's new reader works OK. Its M2 email client eludes me.
It miscategorized my email and I can't figure out how to orgranize.
After a while I gave up.

Wai Yip




From reply.in at the.newsgroup  Sat Apr  5 12:26:14 2003
From: reply.in at the.newsgroup (Rene Pijlman)
Date: Sat, 05 Apr 2003 19:26:14 +0200
Subject: Python - Database and Email
References: <3e8f0614.12738582@news.jaring.my>
Message-ID: <274u8vc961fr9cpaq3hsbfdo5uka5ibd5r@4ax.com>

Kenneth Gomez:
>I would like to send personalized emails out to client using 
>Python, MySQL

See
http://www.python.org/doc/current/lib/module-email.html
http://www.python.org/doc/current/lib/module-smtplib.html

>But it doesnt teach the integration with an email client?

You don't need to integrate with an email client. Just send the
mail from the Python program itself, using the above mentioned
libraries.

-- 
Ren? Pijlman



From sholden at holdenweb.com  Wed Apr 16 10:22:02 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 14:22:02 GMT
Subject: Python and Schools
References: <mailman.1050499405.10336.python-list@python.org>
Message-ID: <eGdna.110601$It5.106431@news2.central.cox.net>

"Andres Rosado" <arosado at softhome.net> wrote in message
news:mailman.1050499405.10336.python-list at python.org...
> On 05:06 AM 4/11/2003 -0400, python-list-request at python.org feed this fish
> to the sharks:
> >- if you learn programming with Python and later have to use Java or
> >Visual Basic, it will be the most frustrating experience for the young
> >fellows.
>
> Pardon my ignorance, but how could it be a frustating experience?
>

The frustration will arise from the non-availability of the various simple
features that Python provides to make programming easier.

For a real example, consider the contortions I have to go through when
writing VBScript ASP code for a client's project if I want a list (array) of
strings. Where in Python I can write:

    sarr = ['the', 'quick', 'brown', 'fox', 'jumps' 'over', 'the', 'lazy',
'dog]

in VBScript I must resort to one of [nontested code follows]:

    dim sarr(9)
    sarr(1) = 'the'
    sarr(2) = 'quick'
        ...
    sarr(8) = 'dog'

or possibly

    sarr = split("the quick brown fox jumps over the lazy dog")

Now imagine the further complexity when each element of sarr needs to have
structure itself, remembering that VBScript has no data structure constructs
except the array (well, I understand there's now a third-party module that
lets you create collections or some similar object, but there's still no way
to create an object with attributes).

Hope this makes it clearer.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From aahz at pythoncraft.com  Mon Apr 28 16:39:00 2003
From: aahz at pythoncraft.com (Aahz)
Date: 28 Apr 2003 16:39:00 -0400
Subject: Class attribute or instance attribute?
References: <vaqtlsqbi52j93@corp.supernews.com>
Message-ID: <b8k3h4$1br$1@panix2.panix.com>

In article <vaqtlsqbi52j93 at corp.supernews.com>,
Paul Watson <pwatson at redlinec.com> wrote:
>
>My biggest question is why self.v has a value of "Consequences" in the
>_init__() of f2?  Is self.v not an instance variable?

When an attribute lookup fails in an instance, Python searches the
namespace of the parent class and all of the parent's base classes (but
*not* the namespaces of the metaclasses).
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From bitbucket at safe-mail.net  Thu Apr 24 08:34:27 2003
From: bitbucket at safe-mail.net (poiboy)
Date: 24 Apr 2003 05:34:27 -0700
Subject: xmlrpc introspection
References: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com>
Message-ID: <af6b5bdd.0304240434.fd7e6ad@posting.google.com>

Quick answer: Register 'system.listMethods', 'system.methodSignature',
and 'system.methodHelp' in your server and make sure they perform as
stated in the docs (listMethods returns a list of strings, etc.). Yup
- just do it all by hand.

The output for these methods are typical <methodResponse>s (snippets
available at http://www.xmlrpc.org/spec) - which *should* be taken
care of automatically by the XMLRPC server; lists and/or strings
converted to something like

<?xml version="1.0"?>
<methodResponse>
    <params>
        <param>
            <value>
                <array>
                    <data>
                       
<value><string>remoteMethodOne</string></value>
                       
<value><string>remoteMethodTwo</string></value>
                       
<value><string>remoteMethodThree</string></value>
                    </data>
                </array>
            </value>
        </param>
    </params>
</methodResponse>

..for 'server.listMethods'. In other words, you *shouldn't* have to
worry about the raw XML. Verify your server's output democratically
via the debugger link above
(http://www.dscpl.com.au/xmlrpc-debugger.php - added to my bookmarks,
mahalo Graham!) or manually by testing 'server.listMethods' at
http://xmlrpc.usefulinc.com/demo/server.php.

A headstart on introspection can be found at
http://xmlrpc-c.sourceforge.net/hacks/xmlrpc_registry.py, and there's
a detailed account of a comprehensive 'system.describeMethods'
implementation at http://xmlrpc-epi.sourceforge.net/specs/rfc.system.describeMethods.php
which I recommend for practical documentation consideration.

Two unsolicited cents worth: Complimentary RPC test cases are
invaluable (list all methods as links, click to test individually or
click "test all"). Had to get that off my chest.

Aloha, the poiboy



From skip at pobox.com  Fri Apr  4 17:43:18 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 4 Apr 2003 16:43:18 -0600
Subject: Invalid format character in string
In-Reply-To: <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com>
        <mailman.1049493668.26508.python-list@python.org>
        <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>
Message-ID: <16014.2694.609037.280086@montanaro.dyndns.org>

    Nikolai> The problems arise where the HTML editor used (to produce the
    Nikolai> file in the first place) has put various expressions of %; and
    Nikolai> %>.  These are HTML codings the creator of the HTML file does
    Nikolai> not really know about.

Okay, how about a preprocessing step like this:

    data = open("htmlfile").read()
    # double up any % not followed by a left paren
    data = re.sub(r"%([^(])", r"%%\1", data)
    sys.stdout.write(data % valdict)

?  It's not perfect, but should be a lot better than the pain you're
experiencing now.

Skip




From puccio_13 at yahoo.it  Fri Apr 25 08:54:13 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Fri, 25 Apr 2003 14:54:13 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com>
Message-ID: <1waqa.26138$DN.735109@tornado.fastwebnet.it>

Cameron Laird wrote:

> In article <Dl9qa.26067$DN.734905 at tornado.fastwebnet.it>,
> Alessio Pace  <puccio_13 at yahoo.it> wrote:
>>"Martin v. L?wis" wrote:
>>
>>> Alessio Pace wrote:
>>> 
>>>> Hi, I'd like to make a GUI for a program written in python, meant to
>>>> run on mac, win and linux: which is the preferable solution? I saw
>>>> around tkinter, wxpython and pyqt....
>>> 
>>> I recommend Tkinter. It comes with the standard Python installation, and
>>> it works quite well atleast on Windows and X11 (I've never used it on
>>> the Mac).
>>> 
>>> Regards,
>>> Martin
>>
>>That's what I had in mind, but after reading some tutorials on Tkinter I
>>saw that wxpython and pyqt have more features still remaining
>>portable...mmm...
> .
> .
> .
> 1.  Each of the choices has serious advantages.
>     Even toolkits zealots recognize that no one
>     toolkits strictly dominates any other.
> 2.  Each toolkit has fans.
> 3.  It's relatively easy to download/install/
>     experience Python (or Perl or ...), and ac-
>     quire your own "feel" for each language.
>     The GUI toolkits are a bit heavier and
>     clumsier; it generally takes longer to gain
>     a proper understanding of them.
> 4.  Comparisons of GUI toolkits are popular.
>     Such discussions often erupt here and else-
>     where.  Abundant details are available, if
>     reading the words of others interests you.
> 5.  Tkinter is the easiest for a newcomer, in
>     its installation and "first launch".  It's
>     the best documented, in the sense of making
>     appearances in the most in-print published
>     books.
> 
> I think it's reasonable for you to start with Tkinter.

Thanks a lot. I've already starting reading some guides, and I saw it is not
difficult to develop with it.

-- 
bye
Alessio Pace



From goodger at python.org  Tue Apr 29 09:37:33 2003
From: goodger at python.org (David Goodger)
Date: Tue, 29 Apr 2003 09:37:33 -0400
Subject: POD vs. reST for standalone writing?
In-Reply-To: <20030429043815.GB13813@frobozz.local>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <mailman.1051584944.30800.python-list@python.org> <b8kuet$ftn$1@panix1.panix.com> <20030429043815.GB13813@frobozz.local>
Message-ID: <3EAE801D.4010005@python.org>

Andrew Bennetts wrote:
> Hmm, interesting.  I thought the :Foo: syntax was for definition lists, not
> inline in text?

As a block-level device, it's used for field lists:

     :Author: Me
     :Date: Today

It's also used in inline text, to extend interpreted text roles 
(explicit markup):

     Here is some `interpreted text`:role:.

Both uses attach a label or classifier to something.

> I guess what I'm asking is, what would be the ReST
> equivalent of:
> 
>     <p>To start the reactor, use the <code class="API"
>     base="twisted.internet">reactor.run</code> method.  To stop it, use
>     <code class="API" base="twisted.internet">reactor.stop</code>.</p>
>
> [The optional base attribute provides the fully qualified name to 
> allow the text to be linked to the right part of the API docs, but
> doesn't affect the final display.]

These examples, with the extra attributes, are not directly 
translatable: interpreted text roles are not currently parameterizable 
in reStructuredText.  There is some discussion underway on how to 
parameterize roles or define document-local roles, so it might be 
translated something like this:

     .. :api-ti: role:: api
        :base: twisted.internet

     To start the reactor, use the :api-ti:`reactor.run` method.  To stop
     it, use :api-ti:`reactor.stop`.

This is not yet implemented though, either because nobody has needed it, 
or not needed it enough to do the work.  Docutils/reStructuredText is 
also still undergoing development.

However, the above is not typical reStructuredText usage.  When a 
document gets to that level of complexity, it may be beyond what 
reStructuredText aims to support.  ReStructuredText doesn't aim to be a 
solution to every problem, just to the most common 80%-90%.

> And how would I make the ReST tools link that to my API documentation
> generated with pydoc, epydoc, or whatever?

Currently, with URL links.  Again, there's been some discussion about 
using other methods (perhaps XLink/XPath-based), but no implementation 
yet.  How does Lore link to outside docs?

> (Btw, don't get me wrong; I *like* ReST for some uses.  I just think it's
> better suited to wikis than to reference manuals and the like.  I don't
> really like ReST's liberal use of punctuation in large doses.)

The "liberal use of punctuation in large doses" only really kicks in for 
advanced uses.  Personally, I find occasional punctuation easier to 
ignore than <ubiquitous> <angle> <brackets>.  To each his own :-).

-- David Goodger






From ckea25d02 at sneakemail.com  Sun Apr  6 22:09:09 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 14:09:09 +1200
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304040445.a3752dd@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122> <3E8CF4B5.1E221DD7@alcyone.com> <9383481e.0304040445.a3752dd@posting.google.com>
Message-ID: <b6qm2p$7u18m$1@ID-169208.news.dfncis.de>

Drew Moore wrote:
>> still,
> I'd like to see this PEP die for a good technical reason.
> help me out here!!

PEPs don't need technical reasons to die; they only
need to fail to convince Guido that they're sufficiently
useful to be accepted.

And such a radical change as this has a VERY big threshold
of usefulness to reach before Guido will even start to
consider thinking about it...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From aleax at aleax.it  Mon Apr 14 15:55:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 19:55:46 GMT
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it> <yfshe92dmpt.fsf@black132.ex.ac.uk> <a8vma.11117$LB6.312431@news1.tin.it> <b7f0pb$d2g$1@slb9.atl.mindspring.net>
Message-ID: <6nEma.13875$LB6.373630@news1.tin.it>

Andrew Dalke wrote:

> Martelli-bot:
>> Still, it DOES mention, as the
>> one and only claimed advantage, that, without __slots__, "using a
>> separate dictionary to hold a single instance variable doubles the
>> memory used", and then proceeds to list the resulting limitations
>> as "tidbits and warnings".  Any different reading is clearly wishful
>> thinking, and the design intent seems pretty clear here.
> 
> I was recently looking at a Python package ported from Java.
> It was the first package I've reviewed which used __slots__.
> My impression is that __slots__ were used as a rough approximation
> to declaring instance attributes, and not because of an explicit desire
> to save space or improve lookup performance.  (That desire may
> be implicit, but given the early state of the code it would be
> premature to act on it.)
> 
> Given the style of the code, it appears that the author is not an
> experienced programmer (eg, used "if flag is True" rather than
> "if flag"), which I believe back my belief that people from statically
> type languages look for a way to declare attributes, find __slots__,
> and use it, not really knowing the appropriateness of this approach.

Yes, I fear this experience may indeed be typical.

> This is different than the design intent.  Should the documentation
> be updated to caution against using __slots__ except for cases
> which fit better the original intent? ....

I would vote for a big YES on that, if it were up for a vote.  But
where does "the documentation" (reference manuals) cover __slots__?


>> > 1. what's the point of not automatically including a slot for
>> > '__weakref__'
>> >    (hardly the unreasonable memory consumption such a slot would
> impose?)?
>>
>> Indeed definitely that -- increasing memory consumption by e.g. 50%
>> for an object with two "true" slots, just in case somebody in the
>> future MAY want to weakreference it, would be a weird choice.  If
>> you need your instances to be weakly referenceable include the slot
>> explicitly and pay the attendand memory price.
> 
> I don't know __slots__ well enough.  Does you mean that if I want
> weak references for these I can simply add that slot, as in
> 
>   __slots__ = [..., "__weakref__"]
> ?

Haven't seen it documented anywhere, actually.  But, e.g....:

>>> class W(object):
...   __slots__ = ['x', '__weakref__']
...   def __init__(self, x): self.x = x
...   def __repr__(self): return 'W(%d)'%self.x
...
>>> wlst = [W(i) for i in range(11, 33)]
>>> wd = weakref.WeakValueDictionary(zip(range(len(wlst)),wlst))
>>> for w in wlst[:]:
...     if w.x%5: wlst.remove(w)
...     del w
...
>>> wlst
[W(15), W(20), W(25), W(30)]
>>> wd.values()
[W(15), W(20), W(25), W(30)]
>>>

seems to work just as well as it would without __slots__.


>> Sure.  So what?  A trickily coded class could always throw introspection
>> based on instance __dict__ off -- __getattr__ could and often did "add"
>> attributes not otherwise visible.
> 
> Though it should have also implemented __methods__ and/or __members__,
> to return the list of otherwise invisible fields.
> 
> Very few __getattr__ implementation did this.  It's documented, but hard
> to tell when/why it's useful.

Indeed -- I don't think I ever remembered to do anything like that,
personally...;-).


Alex




From tebeka at cs.bgu.ac.il  Wed Apr 30 08:40:29 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 30 Apr 2003 05:40:29 -0700
Subject: plugin system
References: <WHIra.27759$su3.3307778@twister.nyc.rr.com>
Message-ID: <33803989.0304300440.7e21c863@posting.google.com>

Hello xam,

> I am interested in writing a plugin system where the main app is python but
> it is capable of accepting helper "plugins" which can be either python or
> C++. I guess I want some sort of an interface spec for the plugins that can
> be implemented in either language. Which technologies should I research? Are
> there other projects that I can refer to as a model -
You can just use __import__ to load new modules and use them.
Make you C++ modules compatible with Python C API (use boost.python
http://www.boost.org/libs/python/doc/index.html)

HTH.
Miki



From daniel.dittmar at sap.com  Tue Apr 29 12:09:01 2003
From: daniel.dittmar at sap.com (Daniel Dittmar)
Date: Tue, 29 Apr 2003 18:09:01 +0200
Subject: import problem
References: <mailman.1051631369.24390.python-list@python.org>
Message-ID: <b8m82u$52d$1@news1.wdf.sap-ag.de>

Vince Buonassisi wrote:
> from WVWService.util import XMLUtil
> from WVWService.test import HTTPClient
>
> the first import works just fine.  but, the second one i get the error
> "ImportError: No module named test".  i have checked for mispellings;
> i have checked to see if the directory test exists; i have checked the
> system path by printing out sys.path from w/i the script to see if the
> directory path is in sys.path (/home/vjb/workspace/wvwBase3/src/wvw).

Maybe a missing __init__.py in WVWService/test?

Daniel






From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Fri Apr 11 14:34:36 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Fri, 11 Apr 2003 20:34:36 +0200
Subject: distributed computing implementations
In-Reply-To: <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <d116fbae.0304100704.68ef4833@posting.google.com> <c48$3e960655$516049d2$3498@nf1.news-service.com> <3e96efac$0$49107$e4fe514c@news.xs4all.nl> <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk> <3e96f481$0$49099$e4fe514c@news.xs4all.nl> <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>
Message-ID: <3e970abc$0$49101$e4fe514c@news.xs4all.nl>

Robin Becker wrote:
> In article <3e96f481$0$49099$e4fe514c at news.xs4all.nl>, Irmen de Jong
> <irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> writes
> 
>>Sending lots of ints to the server
>>Time to send 5000 ints, 4.586772 (0.000917 per call)
>>
> 
> is that 5000 connections with new sockets? Seems really fast to me.

No, Pyro reuses socket connections for method calls.

--Irmen




From fredrik at pythonware.com  Fri Apr 25 15:59:06 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 25 Apr 2003 21:59:06 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <Kq9qa.9083$h%2.957726@read1.cgocable.net> <mailman.1051273778.5621.python-list@python.org> <fzo6emyy.fsf@python.net> <mailman.1051277378.13076.python-list@python.org> <oprn6944nxmu3s0q@news.cis.dfn.de>
Message-ID: <b8c40h$j03$1@main.gmane.org>

Klaus Meyer wrote.
> > And what about effbot's Tkinter 3000?
> How active is this project? Maybe the release date could
> be the year 3000 ;-)

the WCK implementation for Tkinter is perfectly usable (the
current version is 1.0b2, released in november 2002).

more here:

    http://effbot.org/zone/wck.htm

the windows version of the effnews RSS reader uses a WCK
implementation for native windows:

    http://effbot.org/zone/effnews-exe.htm

</F>







From spam at magnetic-ink.dk  Tue Apr  1 12:55:17 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Tue, 1 Apr 2003 17:55:17 +0000 (UTC)
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <d2c6a353-612a-4f45-8ddd-bddc04d59e6a@news.szn.dk> <slrnb8jfsq.9n8.philh@cabalamat.uklinux.net> <41d02ffe-a3d3-4bde-9466-d4b53309242a@news.szn.dk> <slrnb8jheo.5u1.ramen@lackingtalent.com>
Message-ID: <5d46e0af-6c2c-4e57-9861-611245099c60@news.szn.dk>

Dave Benjamin wrote:

>>>>#v+
>>> 
>>> What does this mean?
>> 
>> That's the beginning of a set of "verbatim marks" for the JED editor.
> 
> Really? How does that work?

Oh, sorry, it should have been "for the slrn newsreader", not "for the
JED editor".  What a silly mistake.  

Slrn? can optionally reformat articles with long lines, but when it sees
a set of #v+/#v- it will leave the lines inbetween unchanged - much like
the <pre></pre> HTML tags.  Also, it can display text within verbatim
marks in a different color.


  // Klaus

 ?) <http://www.slrn.org/>
-- 
 ><> 	unselfish actions pay back better



From sandysj at juno.com  Thu Apr 10 12:58:52 2003
From: sandysj at juno.com (Jeff Sandys)
Date: Thu, 10 Apr 2003 16:58:52 GMT
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de>
Message-ID: <3E95A2CC.A2EE9823@juno.com>

You can drive Visio with Pythom com module.

Johannes Eble wrote:
> 
> is there a library available for Python that 
> behaves similar to MS Visio?



From mwh at python.net  Tue Apr  8 10:20:23 2003
From: mwh at python.net (Michael Hudson)
Date: Tue, 8 Apr 2003 14:20:23 GMT
Subject: SystemError in weakrefobject.c
References: <mailman.1049808018.3607.python-list@python.org>
Message-ID: <7h31y0d5avt.fsf@pc150.maths.bris.ac.uk>

Etienne Labuschagne <elabuschagne at gmsonline.co.za> writes:

> Hi all,
> 
> I have a block of code within a try except that does message parsing
> work.  This block of code is never supposed to fail, it must log the
> error, reject the message which caused the parsing to fail and
> continue.  So, I catch all exceptions, log the error and continue
> normally with the rest of the messages.
> 
> I am however getting SystemErrors which seems to go past the except
> and then crashes the thread.  In fact, it not only crashes the thread,
> but the whole program.
> 
> Looking at the exception, it seems something more serious is happening
> and I can't figure out what is wrong.  Here are two of the exceptions
> which I have logged so far:
> 
> Traceback (most recent call last):
> File "C:\!Projects\Python\GlobalTrack\MessageDecoders.py", line 743, in decode
> '\n'.join(['%s = %s' %(k,dict[k]) for k in dict.keys()])))
> SystemError:
> D:\trentm\as\Apps\ActivePython-2_2\src\Core\Objects\weakrefobject.c:690:
> bad argument to internal function
> 
> and
> 
> Traceback (most recent call last):
> File "C:\!Projects\Python\GlobalTrack\MessageDecoders.py", line 80, in __call__
> return self.decode()
> File "C:\!Projects\Python\GlobalTrack\MessageDecoders.py", line 344, in decode
> cannedId = int(self._destinationAddress)
> SystemError:
> D:\trentm\as\Apps\ActivePython-2_2\src\Core\Objects\weakrefobject.c:690:
> bad argument to internal function
> 
> Any help would be greatly appreciated.

1) can you show the Python source that causes this?
2) what version of Python?  I'm guessing 2.2 -- can you try 2.2.2?
3) are you using any C extensions that might have bugs that could explain 
   this?

Otherwise I'll need to dust off that telepathy unit...

Cheers,
M.

-- 
  You owe The Oracle a TV with an 'intelligence' control - I've 
  tried 'brightness' but that didn't work.
                                      -- Internet Oracularity #1192-01



From bellman at lysator.liu.se  Fri Apr 11 14:13:58 2003
From: bellman at lysator.liu.se (Thomas Bellman)
Date: Fri, 11 Apr 2003 18:13:58 +0000 (UTC)
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de>
Message-ID: <b770l6$ejq$2@news.island.liu.se>

mb at muenster.de (Martin Bless) wrote:

> sendmail = os.popen("/usr/lib/sendmail -t", "w")
> sendmail.write(msg)
> sendmail.close()

> This works just fine. But is it enough? How do I know the operation
> succeeded?

Define "operation succeeded"...  Do you want to know if /usr/lib/sendmail
managed to queue the message for delivery?  Or do you want to know
if the mail reached its destination?  Or do you want to know if the
recipient has read the mail?

It is quite difficult to know if the mail reaches its destination.
You have to wait for several days and check for email bounces, i.e,
your program must be capable of *reading* mail too, not just
sending, and it must be able to differentiate between real bounces,
warnings that the mail hasn't been delivered for X hours (some mail
systems send a warning if it hasn't been able to pass the mail on
to the next hop within a few hours), and any other mail that might
arrive (like spam...).

Knowing whether the recipient has actually *read* the mail is even
more difficult...

However, if you just want to know that the mail has been placed in
the queue of the systems MTA, then it is much easier.  And others
have already answered that part.


-- 
Thomas Bellman,   Lysator Computer Club,   Link?ping University,  Sweden
"Life IS pain, highness.  Anyone who tells   !  bellman @ lysator.liu.se
 differently is selling something."          !  Make Love -- Nicht Wahr!



From cnetzer at mail.arc.nasa.gov  Tue Apr 15 20:16:09 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 15 Apr 2003 17:16:09 -0700
Subject: Big-O notation (was: Re: Python and Schools)
In-Reply-To: <roy-0FAA35.18584415042003@reader1.panix.com>
References: <qGqla.61$Tj7.91655@news.uswest.net>
	 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
	 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
	 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
	 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
	 <mailman.1050438927.20366.python-list@python.org>
	 <roy-0FAA35.18584415042003@reader1.panix.com>
Message-ID: <1050452168.11992.34.camel@sayge.arc.nasa.gov>

On Tue, 2003-04-15 at 15:58, Roy Smith wrote:

> We've got something which scans a directory (an O(N) process).  It's 
> supposed to happen once every time the program is run.  Somebody goofed 
> in coding, and moved it into the loop that happens once every input 
> file, not once per run.  Now it's an O(N^2) process.
> 
> Nobody noticed it in testing because we only tested it with maybe 100 
> files.  Then one of our customers ran it with 25,000 files (a perfectly 
> reasonable thing to do).  Whoops :-)

I have unit tests that actually try to detect these things using
regressions and t-tests.  For example, I have an implementation of a
numerical algorithm that I know is supposed to be O(N).  In my tests, I
time the result of the algorithm with varying input sizes (ie. different
N), and do a quadratic regression (like a linear regression, but with a
term for x**2).

My assumption is that the x**2 term should have a coefficient near zero
(ie. zero "curve" to the regression).  I then use a t-test to verify
this assumption, using a few runs on random data.  Set the confidence
level fairly low, and it seems to do a good job of detecting if I
accidentally make a O(N**2) algorithm, rather than the expected O(N)
(with very few false positives).

Fairly quick and easy to implement (given existing t-test and regression
toolks available in stats.py and Numeric)


-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From gupt_vive at hotmail.com  Sat Apr 19 07:28:37 2003
From: gupt_vive at hotmail.com (vivek kumar)
Date: Sat, 19 Apr 2003 16:58:37 +0530
Subject: widget like listview
Message-ID: <Law15-F114P2Bk87w4N000011a4@hotmail.com>

Hi all,

Is there any widget/mega widget available in Tk like ListView in windows (in 
Visual Studio). Actually i want to show a list of names with their addresses 
etc from a table and the user is allowed to choose a name from that list and 
the application will do processing for this name.

I have tried ScrolledList widget of PMW but it only shows the table, i can 
not select an entry from it. Moreover if the name etc in the table are not 
of same length then the format of the list is disturbed ( in windows , on 
linux it works fine ).

regards
Vivek Kumar

_________________________________________________________________
Windows Server 2003. Do more with less. 
http://www.msn.co.in/computing/winsvr2003/ Upgrade now!





From vze4rx4y at verizon.net  Thu Apr 24 23:57:26 2003
From: vze4rx4y at verizon.net (Raymond Hettinger)
Date: Fri, 25 Apr 2003 03:57:26 GMT
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth>
Message-ID: <Gm2qa.32861$ot1.23459@nwrdny02.gnilink.net>

> I have come across a weirdness. I have a number of floats, (lat long
> pairs) which I stuff into a dict, but when I take them out, my nice %2f4
> number, (say, 47.5894) is something long and *almost* right, like
> 47.589333333333333339. What gives? I am not doing anything special as
> far as I can tell, just stuffing the numbers into a dict, and recalling
> it later. This is with Python 2.2 on Linux, and no, it's not an original
> Pentium :)

This is clearly some form of steganography where the least order bits
are mutated in a seeming random fashion but, in fact, contain an
embedded message.  The message bits can be recovered by taking
the difference between the what was supposed to be in the dictionary
and the actual stored contents.  Historically, this was done by altering
the low order bits in pixels of an image.  This left the overall image
intact while hiding a message at a level not discernable to the eye.
However, the advent of lossy JPEG image compression resulted
in the lower bits being mangled.  As a result, steganographers have
been forced to the less visually appealing technique of altering
floats in a dictionary.  By leaving the high order bits intact, user
programs still produce correct answers to several significant
figures.  It is somewhat analogous to bank schemes that accumulate
fractions of a penny.

Since the most signficant digits are unaffected, you need not worry
about it.  Continue to use the round() function and print out your
answers using the % formatting operator.  Soon, you'll feel as
right as rain.


Raymond Hettinger


P.S.  As an example, note that the message above contains an
embedded reference to the movie, The Matrix.





From aleax at aleax.it  Sat Apr 26 05:41:20 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 09:41:20 GMT
Subject: True/False
References: <mailman.1051110756.23235.python-list@python.org> <2c60a528.0304240517.25f2fe29@posting.google.com> <a8b7f07a.0304251249.ad24819@posting.google.com> <2c60a528.0304252347.142f0548@posting.google.com>
Message-ID: <4vsqa.6347$3M4.180222@news1.tin.it>

Andrew Clover wrote:

> A. Lloyd Flanagan <alloydflanagan at attbi.com> wrote:
> 
>> This has the advantage of making True and False look more like builtins.
> 
> Interesting. I'm not sure about this; it seems to have the potential to
> confuse other, unrelated modules. For example some package might check for
> the existence of 'True' to see whether it needs to provide its own boolean
> type that's distinguishable from integers. If the builtin is fiddled it
> would look like there was a real boolean type but attempts to use it could

Such a hypothetical "some package" would disastrously fail under Python
2.2.2, which does define a built-in True BUT doesn't have a boolean type
that is at all distinguishable from integers.  That the OP's approach
shows up the "some package"'s fault just as 2.2.2 shows it doesn't seem
to me to be worth a condemnation of that approach (nor of 2.2.2) but
rather underlines that the "some package" is broken and needs to be fixed.

> fall over. Modules that affect unrelated modules merely by being imported
> can be tricky to debug.

Sure.  But something that (e.g.) runs on 2.2.1 and simply simulates closely
an effect that an upgrade to 2.2.2 would also have doesn't seem to present
any particularly high risk in this regard.


Alex




From aleax at aleax.it  Thu Apr 17 11:23:38 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 15:23:38 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050529915.4666.python-list@python.org>
Message-ID: <_Fzna.27993$LB6.643892@news1.tin.it>

Steven Taschuk wrote:
   ...
> trade-off as being between, say, an algorithm which runs in O(n)
> time but needs O(n^2) memory to do it, and an alternative which

I don't think there can be any such algorithm -- in O(n) time,
the algoritm cannot USE more than O(n) memory, so how could it
NEED it...?  The reverse, sure, but not this way, I think.


Alex




From peter at engcorp.com  Tue Apr 29 09:35:51 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 29 Apr 2003 09:35:51 -0400
Subject: simplest question on win32com.client
References: <a5b86892.0304290534.31528b9e@posting.google.com>
Message-ID: <3EAE7FB7.5512FDF8@engcorp.com>

Joe Green wrote:
> 
> I can't seem to get this code to run.
> 
> --
> from win32com.client import Dispatch
> 
> xl = Dispatch("Excel.Application")
> --
> 
> It comes back with:
> 
> The instruction at "0x1e039e79" referenced memory at "0x1e3840fc". The
> memory could not be "read".
> 
> Click OK to terminate the program
> Click on CANCEL to debug the program

Obviously this shouldn't happen. :-)  Maybe reinstalling Python
and win32all would be a good idea.  Otherwise, do you have some
alternate means of checking the COM interface to Excel?  Maybe
via VB (ugh)?

-Peter



From edcjones at erols.com  Fri Apr 11 10:20:16 2003
From: edcjones at erols.com (Edward C. Jones)
Date: Fri, 11 Apr 2003 10:20:16 -0400
Subject: Pyrex: wrapping #defines with a type name argument
Message-ID: <b76ivu$eri$1@bob.news.rcn.net>

A common type of C macro has a type as an argument. For example:

     #define GET_ELEM(type,x,n) (*(type*)((char*)(x) + n))

What is a good way to wrap this in Pyrex (if any)?




From mdanish at andrew.cmu.edu  Thu Apr 17 11:43:16 2003
From: mdanish at andrew.cmu.edu (Matthew Danish)
Date: Thu, 17 Apr 2003 11:43:16 -0400
Subject: Why functional Python matters
In-Reply-To: <1azna.115402$It5.22896@news2.central.cox.net>; from sholden@holdenweb.com on Thu, Apr 17, 2003 at
    02:49:33PM +0000
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr>
    <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk>
    <3E9EB3E9.1040208@removethis.free.fr> <87k7dt5m3j.fsf@interim.henrik-motakef.de>
    <1azna.115402$It5.22896@news2.central.cox.net>
Message-ID: <20030417114316.M13181@lain.cheme.cmu.edu>

On Thu, Apr 17, 2003 at 02:49:33PM +0000, Steve Holden wrote:
> I can see that, but wouldn't you agree that removing redundant features also
> removes cruft from the source, and eases the maintenance burden, in turn
> reducing the likelihood of implementation bugs? While Guido is known to
> regret the inclusion of lambda in Python, I don't think he would want to
> take it out on purely ideological grounds. Indeed, the suggestion was not so
> much that it would be taken out of the existing language, rather that it
> would not be implemented in some "futurePython", normally called Python 3 or
> Python 3000.
> 
> It would be nice to have One True Python, but that's about as likely as One
> True Lisp. The Python world is pragmatic enough to be able to live with "one
> *most obvious* way to do it", but if there are two similar features in the
> language they have to be maintained in lockstep as the language develops.

What's wrong with Python and CL!!!?  They have all these extraneous
features which can accomplish the same thing!!! All you need is the
<insert model of computability here>!!! Anything else is liable to
develop cruft and cause bugs in the single implementation of our ONE
PURE LANGUAGE!!!  I can't believe that people might want to model
problems any other way!!!

-- 
; Matthew Danish <mdanish at andrew.cmu.edu>
; OpenPGP public key: C24B6010 on keyring.debian.org
; Signed or encrypted mail welcome.
; "There is no dark side of the moon really; matter of fact, it's all dark."



From no at spam.invalid  Wed Apr 30 19:39:01 2003
From: no at spam.invalid (Russell E. Owen)
Date: Wed, 30 Apr 2003 16:39:01 -0700
Subject: Some oddities in 2.3b1, esp. with Tkinter
References: <b8mpp1$25hm$1@nntp6.u.washington.edu> <3EAEF1D6.30900@v.loewis.de> <b8p336$27pg$1@nntp6.u.washington.edu> <m3smrzg3q2.fsf@mira.informatik.hu-berlin.de>
Message-ID: <b8pmqm$32iu$1@nntp6.u.washington.edu>

In article <m3smrzg3q2.fsf at mira.informatik.hu-berlin.de>,
 martin at v.loewis.de (Martin v. Lowis) wrote:

>"Russell E. Owen" <no at spam.invalid> writes:
>
>> >> 1) A filter to remove fonts with non-ascii names (used to filter out 
>> >> stuff that I could not display in a pop-up menu) broke.
>> >>             try:
>> >>                 unicode(fontName, 'ascii', 'strict')
>> >>                 return True
>> >>             except UnicodeError:
>> >>                 return False
>[...]
>> This is not true on MacOS X. I found it out the hard way. The font names 
>> with unicode appear to be asian fonts, at least for the most part.
>
>I see. Please find out what fontName is, and submit a bug report to SF.

I don't think there's a bug. I realized on further investigation that my 
old test caught font names that were ordinary strings but with 
characters with ordinal not in range(128). This worked because in Python 
2.2.2/Tk 8.4.1 the list of font names solely consisted of ordinary 
strings (never unicode), though some names had unprintable characters.

What's new in Python 2.3b1/Tk 8.4.2 is that the list of font names 
returned by tkFont.Font contains some unicode "strings" (what is the 
right term for that?). Those cause the unicode function to raise 
TypeError.

If font names are never supposed to be unicode then I'll report it, but 
it seems like a good idea to me. Some of the asian font names certainly 
look as if they need unicode to be properly expressed.

>Ah, so the Aqua Tcl is multi-threaded. Bad choice. There is not much
>we can do: Tcl threading and Python threading are incompatible. A Tcl
>interpreter can be used only from the thread in which it is created.

I'm very confused about why this will hurt. My Python/Tkinter app will 
not call Tkinter from anything other than the main Python thread. (That 
used to be forbidden and it's a restriction I'm quite comfortable with). 

Also, I don't plan to do low level stuff with tcl from my Python/Tkinter 
app (such as launching additional tcl threads). At least not on purpose 
(perhaps something like createfilehandler violates that rule without my 
knowing).

So I'm hoping that my Python/Tkinter app will be doubly safe -- never 
launch a tcl thread and never call Tkinter from more than one thread -- 
and thus never get me into trouble.

However, a lot of different folks will be using my Python/Tkinter 
application on MacOS X, various flavors of unix and Windows. I'm sure 
some of them will have the threaded version of Tcl/Tk installed. I'm 
sure hoping that'll be OK.

>> >If your Tcl installation is threaded, you should use 
>> >tkapp.createfilehandler instead.
>> 
>> Hmmm...so I can first try my usual way of creating a file handler, and 
>> if that fails, then try this? 
>
>I recommend to *always* use
>tkapp.createfilehandler. _tkinter.createfilehandler is only supported
>for backwards compatibility, and cannot be supported in threaded Tcl.

That sounds perfect. One call to handle both cases!

Unfortunately, so far I've not been able to find tkapp. It doesn't seem 
to be a module in Tkinter or in Tkinter.Tk nor an attribute of root = 
Tkinter.Tk(). And none of these has a function createfilehandler. (I did 
find a Tkinter file tkappinit.c but reading it was not enlightening).

-- Russell



From jarausch at skynet.be  Sat Apr 26 07:07:08 2003
From: jarausch at skynet.be (Helmut Jarausch)
Date: Sat, 26 Apr 2003 13:07:08 +0200 (CEST)
Subject: Linux SUID
Message-ID: <200305042021.h44KLUa0004826@sojef.skynet.be>

Many thanks for your hint,

Helmut.

-- 
Helmut Jarausch

Lehrstuhl fuer Numerische Mathematik
RWTH - Aachen University
D 52056 Aachen, Germany




From max at alcyone.com  Tue Apr  1 17:28:27 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 01 Apr 2003 14:28:27 -0800
Subject: Precidence of unary plus/minus relative to exponentiation
References: <mailman.1049218990.10515.python-list@python.org>
Message-ID: <3E8A128B.132F091C@alcyone.com>

"Warnes, Gregory R" wrote:

> The current operator precedence places unary plus and minus higher
> than
> exponentiation.  This leads the counterintuitive result:
> 
>         >>> -2.0**2.0
>         -4.0
> 
> Python's operator predicence is causing this to be interpreted as:
> 
>         >>> -(2.0**2.0)
> 
> rather than the normal mathematical precidence, which would give
> 
>         >>> (-2.0)**2.0
> 
> Is there a good reason for the Python order of precedence?

In programming languages, there's often variance on whether
exponentiation has higher precedence than unary negation.  The curious
thing about your question is that Python's choice _is_ the way that it's
done in mathematics; when you write

	  b
	-a

this means -(a**b), not (-a)**b [using Python's exponential operator
**].  This can easily be seen if you deal with polynomial, for instance;
in normal type, the hyphen for the unary negation operator is large, and
the superscribed exponent is small, so the negation stands out much more
strongly.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ A good indignation brings out all one's powers.
\__/ Ralph Waldo Emerson
    Esperanto reference / http://www.alcyone.com/max/lang/esperanto/
 An Esperanto reference for English speakers.



From robin at jessikat.fsnet.co.uk  Thu Apr 17 19:23:33 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Fri, 18 Apr 2003 00:23:33 +0100
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org>
 <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk> <3E9EA707.4DB53582@engcorp.com>
 <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk>
 <b7n4b8$maq$1@slb0.atl.mindspring.net>
Message-ID: <IX3TiVA1dzn+Ewpk@jessikat.fsnet.co.uk>

In article <b7n4b8$maq$1 at slb0.atl.mindspring.net>, Andrew Dalke
<adalke at mindspring.com> writes
>Robin Becker:
>> interestingly genetic programming et al seem to perform worse when
>> attempts are made to 'direct' the evolution.
>
>Pardon?
>
>I thought the point of the cost function was to 'direct' the
>evolution.
>
>                    Andrew
>                    dalke at dalkescientific.com
>
>
I think the *direction* P Hansen is talking here is in the solutions
that are selected; the programmers know how to choose the candidate
solutions.

In GP if you bias the choices then, almost always, the results are
inferior.

You are correct in saying that the cost function implicitly directs the
evolution; it does not determine how candidates are chosen. In simple
terms the candidates should be drawn from a set of non zero measure (ie
as large as possible in the set of all solutions). Conceptually GP is
direct search using an evolution like process for the candidate
selection so ideally we would like to cover all possible solutions, but
given the time constraint we cannot.
-- 
Robin Becker



From max at alcyone.com  Thu Apr 17 18:57:11 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 17 Apr 2003 15:57:11 -0700
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net> <3E9D35FC.FCAD17C9@alcyone.com> <b7mp2u$3k8$0@216.39.172.122> <3E9EEB54.682751BC@alcyone.com> <b7n5le$srb$0@216.39.172.122>
Message-ID: <3E9F3147.8C70A9C8@alcyone.com>

Bengt Richter wrote:

> Yes, I'm aware. I thought we were talking about `...` in Python ;-)

Sorry, I thought you were asking a more open-ended question.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Covenants without the sword are but words.
\__/ Camden
    Bosskey.net: Aliens vs. Predator 2 / http://www.bosskey.net/avp2/
 A personal guide to Aliens vs. Predator 2.



From mgerrans at mindspring.com  Mon Apr 28 00:29:04 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Sun, 27 Apr 2003 21:29:04 -0700
Subject: Sorting files
References: <ATZqa.667651$S_4.718911@rwcrnsc53>
Message-ID: <b8ianf$oth$1@slb4.atl.mindspring.net>

You can sort them using the optional compare function parameter of sort:

Images.sort( lambda a,b: cmp(os.stat(a)[stat.ST_MTIME],
os.stat(b)[stat.ST_MTIME]) )

(This isn't the most readable, you might want to define a separate compare
function and use that)

> How can I get the listing of files of a particular type in a directory,
> sorted by date? For example, I using the following call to get all the
> JPEG's in a dir:
>
> Images = glob.glob1(MyPath+"\\images", "*.jpg")
>
> I want to do just this, but have them sorted by date. Any ideas?






From eric.brunel at pragmadev.com  Mon Apr  7 12:02:08 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Mon, 07 Apr 2003 18:02:08 +0200
Subject: Parsing data
References: <b6s5lt$5c1$1@usenet.otenet.gr>
Message-ID: <b6s71b$9j3$1@news-reader12.wanadoo.fr>

Antonis Kaklis wrote:
> hello to everybody...
> 
> A have a file with values ordered by columns:
> Say (e.g force values, in time steps of 0.0200sec)
> 
>    .2161473E-05   .2183854E-05   .2177804E-05   .2174099E-05   .2254829E-05
>    .2145163E-05   .2322683E-05   .2112049E-05   .2092585E-05   .2069618E-05
>    .2052460E-05   .2572402E-05   .2625725E-05   .2701999E-05   .2771539E-05
>    .1884220E-05   .1864868E-05   .1835972E-05   .3150803E-05   .3245766E-05
>    .3343058E-05   .3437211E-05   .3543562E-05   .1512718E-05   .1483842E-05
>    .1440723E-05   .1408216E-05   .1363429E-05   .1340211E-05   .1291136E-05
>    .1260124E-05   .4968635E-05   .5089609E-05   .5228032E-05   .5364493E-05
>    .7569464E-06   .7437904E-06   .5947033E-05   .5593801E-06   .6339938E-05
>    .6471278E-05   .2300983E-06   .1903562E-06   .1526042E-06   .7414982E-05
>      ...            ...              ...          ...              ...
> 
> how can i parse the values (in order to proccess or plot or whatever else i
> want)
> of a structure like that?

for line in file('data.txt'):
   print [float(x) for x in line.split()]

?

(NB: for older Python versions, you'd better do:

for line in open('data.txt').xreadlines():
   print [float(x) for x in line.split()]
)

Was that what you wanted? What you want to do with your data is not quite clear 
from your post...
-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From xxdigitalhellxx at hotmail.com  Wed Apr 16 10:30:58 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Wed, 16 Apr 2003 08:30:58 -0600
Subject: =?ISO-8859-1?Q?R=E9ception_et_d=E9codage_d=27une__?=
 =?ISO-8859-1?Q?_trame_UDP?=
In-Reply-To: <3E9CA7C2.2C9730EA@engcorp.com>
References: <3e9c9b1e$0$1509$7a628cd7@news.club-internet.fr> <3E9CA7C2.2C9730EA@engcorp.com>
Message-ID: <b7jpgn$g2q$1@heckle.cc.uregina.ca>

Peter Hansen wrote:
> Benoit BESSE wrote:
> 
>>Bonsoir tout le monde,
>>
>>qui a d?j? ?crit un module qui ?coute, re?oit, d?code et insert dans une BD
>>une trame UDP.
>>Je voudrais d?coder la trame UDP ainsi et l'ins?rer dans une table :
>>    - Port Source
>>    - Port Destination
>>    - Longueur
>>    - Somme de contr?le
>>    - Donn?es
>>
>>Merci par avance.
> 
> 
> Hmm... almost looks like another "Twisted has that" but I'm 
> not sure what "BD" is so I can't be sure...  ;-)

Well, according to acronymfinder.com, this seems like the most likely 
definition, given the situation:

BD  	Bandes Dessin?es (French: Comic Books)  	

:)

-- 
Ryan




From oblivious at web.de  Wed Apr  9 14:45:57 2003
From: oblivious at web.de (Dominic)
Date: Wed, 09 Apr 2003 20:45:57 +0200
Subject: What are the differences between SWIG and Boost.Python?
In-Reply-To: <mailman.1049741353.6320.python-list@python.org>
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org> <XSXja.346441$3D1.186570@sccrnsc01> <mailman.1049741353.6320.python-list@python.org>
Message-ID: <b71pmn$ibu$03$1@news.t-online.com>

> But it's especially nice to wrap C with Pyrex because it VERY easy to adapt
> interface to pythonic way.

Pyrex is a very nice tool :) I like it very much.

I have used it to glue xmms-control with a
python script. The char* strings returned
by xmms need to be freed which can be
easily done with Pyrex and a
two line utilitiy function.
My previous swig wrapped functions
simply leaked the memory :(

Ciao,
  Dominic




From just at xs4all.nl  Tue Apr  8 15:33:20 2003
From: just at xs4all.nl (Just)
Date: Tue, 08 Apr 2003 21:33:20 +0200
Subject: Pathname problem
References: <3e9321b5$1@rmcnt03a.rmc.ca>
Message-ID: <just-BA0B09.21332008042003@news1.news.xs4all.nl>

In article <3e9321b5$1 at rmcnt03a.rmc.ca>, bourassa <bourassa-m at rmc.ca> 
wrote:

> Hi there.  I am just trying to load files using PIL.  I get the following:
> 
> im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
> Traceback (most recent call last):
>    File "<pyshell#13>", line 1, in ?
>      im=Image.open("C:\MATLAB6p1\toolbox\pout.tif")
>    File "C:\Python22\Lib\site-packages\PIL\Image.py", line 1479, in open
>      fp = __builtin__.open(fp, "rb")
> IOError: [Errno 2] No such file or directory: 
> 'C:\\MATLAB6p1\toolbox\\pout.tif'
> 
> I have stepped through and if the file in C:\ or C:\MATLAB6p1 it loads 
> with no problem.  The minute I try to get it out of "toolbox" (which 
> exists) I get the above error.  I'm using  PIL-1.1.4a3.win32 on a Win2K 
> PIV.  TIA.

Try to use forward slashes or raw strings (r"..."). "\toolbox" actually 
expands to "<TAB>oolbox" ;-)

Just



From theller at python.net  Tue Apr  1 15:14:50 2003
From: theller at python.net (Thomas Heller)
Date: 01 Apr 2003 22:14:50 +0200
Subject: Implementing a "classmethod" in an extension type
References: <mailman.1049226972.15149.python-list@python.org>
Message-ID: <he9infet.fsf@python.net>

"Mark Franklin" <mfranklin at tcsi.com> writes:

> Thanks Thomas,
> 
> The key was the DoClassMethods() call at the end of your code. Once
> I'd copied that over, it worked perfectly:)
> 

It seems I forgot to include the code for DoClassMethods() in my reply,
sorry for that. But you found it anyway ;-)

Thomas



From adonisv at earthlink.net  Mon Apr 28 20:30:51 2003
From: adonisv at earthlink.net (Adonis)
Date: Tue, 29 Apr 2003 00:30:51 GMT
Subject: Tkinter widget question.
Message-ID: <%Ijra.707$5C2.204@tornadotest1.news.pas.earthlink.net>

I have a function that creates N amount of widgets (Entry()), I wish to keep
track of them, looking into the Tkinter documentation there is a mention of
a 'name' option is available for the widgets, but then it goes on to say
that it can not be accessed by a dictionary interface? How can I access it?

Or any other idea on how to track my created widgets.

I am guilty of not googling, as I already had my news client open, and was
lazy to do so...

Any help is greatly appreciated.

Adonis





From h_schneider at marketmix.com  Sat Apr 26 01:57:04 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Sat, 26 Apr 2003 07:57:04 +0200
Subject: AW: FreeBSD: Python exits on signal 11
In-Reply-To: <Pine.OS2.4.44.0304260951350.28674-100000@tenring.andymac.org>
Message-ID: <00e101c30bb8$aadb69a0$0100a8c0@tbird>

OK - thanks for this information. I'll testrun on 4.8.

--Harald


> -----Urspr?ngliche Nachricht-----
> Von: Andrew MacIntyre [mailto:andymac at bullseye.apana.org.au] 
> Gesendet: Samstag, 26. April 2003 00:59
> An: Harald Schneider
> Cc: python-list at python.org
> Betreff: Re: FreeBSD: Python exits on signal 11
> 
> 
> On Fri, 25 Apr 2003, Harald Schneider wrote:
> 
> > I get this error on FreeBSD5 with Python 2.2.2, when I run 
> an application
> > that does its job without errors on NT and Linux platform. 
> It uses threads
> > and sockets as well (xqmlrpclib). It just stops after a 
> while, displays this
> > error and does a core dump.
> 
> Keep in mind that FreeBSD 5.0 is effectively a pre-release, 
> especially as
> the threads support has been re-written to use the new KSE facilities,
> and is thus still relatively immature.  If your code runs on 
> FreeBSD 4.8
> Ok, I suggest that you need to take this to a FreeBSD forum.
> 
> Last I heard, the FreeBSD team don't expect 5.x to become the stable
> branch until 5.2; until then the 4.x branch continues as the stable
> branch.
> 
> --
> Andrew I MacIntyre                     "These thoughts are 
> mine alone..."
> E-mail: andymac at bullseye.apana.org.au  | Snail: PO Box 370
>         andymac at pcug.org.au            |        Belconnen  ACT  2616
> Web:    http://www.andymac.org/        |        Australia
> 





From oblivious at web.de  Thu Apr 10 14:01:11 2003
From: oblivious at web.de (Dominic)
Date: Thu, 10 Apr 2003 20:01:11 +0200
Subject: Macros in Python?
In-Reply-To: <slrnb9b01m.108.do-not-spam-ben.hutchings@tin.bwsint.com>
References: <b71qde$q71$07$1@news.t-online.com> <slrnb9b01m.108.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <b74bem$oi3$02$1@news.t-online.com>

> Alternatively, you can curry the above function:
> 
>     def with_root_privs(func):
>         def call(func, *args, **kwargs):
>             save_euid = os.geteuid()
>             os.seteuid(0)
>             try:
>                 func(*args, **kwargs)
>             finally:
>                 os.seteuid(save_euid)
>         return lambda *args, **kwargs: call(func, *args, **kwargs)
> 
> You can then write something like the following:
> 
>     def activate_user(x):
>         if x in user:
>             active.append(x)
>         ...
>     activate_user = with_root_privs(activate_user)
>     ...
>     activate_user(x)

Thanks. This one is also very good. :-)
Interesting idea to use a closure to build
a function wrapper.

Ciao,
  Dominic




From ecastro at calarts.edu  Tue Apr 22 17:21:26 2003
From: ecastro at calarts.edu (Eduardo Castro)
Date: Tue, 22 Apr 2003 14:21:26 -0700 (PDT)
Subject: Installing MySQL-python-0.9.2 on Solaris 2.9
In-Reply-To: <b84ald$k8k$1@biggoron.nerim.net>
Message-ID: <Pine.GSO.4.21.0304221420430.21478-100000@muse.calarts.edu>

Thanks a lot. 

the: ln to the libmysqlclient worked ok.  But I got another
message after this:

ld: fatal: relocations remain against allocatable but non-writable
sections
collect2: ld returned 1 exit status
error: command 'gcc' failed with exit status 1


Any ideas?

Thanks
Eduardo.


On Tue, 22 Apr 2003, Gilles Lenfant wrote:

> "Eduardo Castro" <ecastro at calarts.edu> a ?crit dans le message de news:
> mailman.1051034858.6946.python-list at python.org...
> >
> > Hi, has anyone had problems installing MySQL-python-0.9.2 on Solaris 2.9?
> >
> > I am getting the following error when I run 'setup.py build'
> >
> > ld: fatal: library -lmysqlclient: not found
> >
> >
> > This is part of the error messages:
> >
> > building '_mysql' extension
> > skipping _mysql.c (build/temp.solaris-2.9-sun4u-2.2/_mysql.o up-to-date)
> > gcc -shared build/temp.solaris-2.9-sun4u-2.2/_mysql.o -L/usr/lib/mysql
> > -L/usr/local/lib/mysql -L/usr/loca/mysql/lib -L/usr/local/mysql/lib/mysql
> > -L/usr/local/mysql/include -L/usr/bin
> > -Wl,-R/usr/local/lib:/usr/openwin/lib:/usr/dt/lib -lmysqlclient -lz -o
> > build/lib.solaris-2.9-sun4u-2.2/_mysql.so
> > ld: fatal: library -lmysqlclient: not found
> > ld: fatal: File processing errors. No output written to
> > build/lib.solaris-2.9-sun4u-2.2/_mysql.so
> > collect2: ld returned 1 exit status
> > error: command 'gcc' failed with exit status 1
> >
> >
> > Thanks
> >
> > Eduardo Castro
> > California Institute of the Arts.
> >
> 
> Fell on this on a FreeBSD box
> 
> $ locate libmysqlclient
> 
> should find something like...
> 
> /usr/local/lib/mysql/libmysqlclient.so.10
> 
> but perhaps somewhere else for you.
> 
> Just cd to a place of your ld path (/usr/local/lib/mysql may be a good
> place)
> 
> $ ln -s /usr/local/lib/mysql/libmysqlclient.so.10 libmysqlclient.so
> 
> If "locate" doesn't find anything, you must install MySQL client package
> before building MySQL-python.
> 
> HTH
> 
> --Gilles
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 





From ipellew at pipemedia.co.uk  Mon Apr 14 21:36:39 2003
From: ipellew at pipemedia.co.uk (Ian Pellew)
Date: 14 Apr 2003 18:36:39 -0700
Subject: Newbie - C API Question
Message-ID: <30875970.0304141525.3f0af2d5@posting.google.com>

Hi;

This is my first attempt at getting to and from my C world.

I cannot get past Python Reference Manual(2.2.1) section 1.2.1.1 :-

h.c:-
#include "Python.h"
main() {
    int x;
    PyObject *t;
    t = PyTuple_New(3);
    x = Py_file_input;
    printf("Hiya, x = %d\n", x);
}   

My compile gives:-
[ `gcc -I/usr/local/include/python2.1 -o h h.c` ] && { ./h; }
Undefined           first referenced
 symbol                 in file
PyList_New                          /var/tmp//ccNWkGt5.o
ld: fatal: Symbol referencing errors. No output written to h
collect2: ld returned 1 exit status

I am on Sun Solaris 5.8 
Python 2.1.1 (#1, Aug 25 2001, 04:19:08) 
[GCC 3.0.1] on sunos5

Any pointers/advice please;

Regards
Ian



From patrick at fraley.de  Fri Apr 18 05:12:00 2003
From: patrick at fraley.de (Patrick W. Fraley)
Date: Fri, 18 Apr 2003 11:12:00 +0200
Subject: Zope Sequences and Python Scripting
References: <pan.2003.04.17.12.34.15.777976@fraley.de> <qt6u9vs13houri99i5fmsen20oof07ile0@4ax.com>
Message-ID: <pan.2003.04.18.09.11.54.233792@fraley.de>

On Thu, 17 Apr 2003 22:26:29 +0100, Dale Strickland-Clark wrote:

> "Patrick W. Fraley" <patrick at fraley.de> wrote:
> 
>>Hi,
>>
>>first off:  I am pretty new to python and have just started with Zope, so
>>please excuse and fopars on my side :)
>>
> Does that count as a faux pas?
> 
>   ;-)

That is what I was talking about ;)



From sholden at holdenweb.com  Wed Apr 30 10:19:41 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 30 Apr 2003 14:19:41 GMT
Subject: Temporary Files
References: <mailman.1051682327.24987.python-list@python.org>
Message-ID: <1YQra.32153$WQ4.15012@news2.central.cox.net>

"Tony Meyer" <ta-meyer at ihug.co.nz> wrote in message
news:mailman.1051682327.24987.python-list at python.org...
> [me]
> > Is there perhaps a different method of copying?  One that
> > doesn't take filenames, but file objects?
>
> Ack.  Yes, there is shutil.copyfileobj().
>
> *sigh*  Rephrased in light of this rather belated discovery; is this
> what I should be doing?
>
> >>> import shutil
> >>> import tempfile
> >>> out = tempfile.TemporaryFile("w")
> >>> out.write(data)
> >>> out.flush()
> >>> # shutil.copyfile(out.name, permanent_name)  # not this!
> >>> permanent_file = file(permanent_name)
> >>> shutil.copyfileobj(out, permanent_file)
> >>> out.close()
> >>> permanent_file.close()
>
> TIA.
>

Well, just to interrupt this internal dialog, would it help if you actually
*did* that and then tell us what you see? ;-)

You might need a "w" second argument for permanent_file =
file(permanent_name), by the way ... you are going to write the file, after
all.

I recently wrote a utility to scan several filestore directories and copy
the content into a date-time hierarchical directory structure, with the new
files named "hhmmss." plus the original name. Here are the extracts that
show another attempt (which I didn't try on Linux yet). You'll notice that
like you I'm too idle to spend an hour looking for shutil.copyfileobj() when
I could do it in Python in a few lines ;-)

...
import mx.DateTime as dt
...
for lang, lv in curs.fetchall():
    ...
    new_recordings = glob(os.path.join(intakedir, lang, "*.wav"))
    for f in new_recordings:
        fnam = os.path.basename(f)    # just the name, no path stuff
        crtime = os.stat(f)[7]        # creation date & time
        ...                           # next line gets time tuple items
        yyyy, mm, dd, HH, MM, SS = time.localtime(crtime)[:6]
        fnmake(archivedir, yyyy, mm, dd)  # ensures directory exists
        inf = file(f, "rb")           # open read and close for data
        data = inf.read()
        inf.close()
        localname = os.path.join(str(yyyy), str(mm), str(dd),
"%02d%02d%02d.%s" % (HH, MM, SS, fnam))
        outf = file(os.path.join(archivedir, localname), "wb")
        outf.write(data)
        outf.close()
        curs.execute("INSERT INTO Recording (RcdDateTime, RcdLang, RcdPath)
VALUES (?, ?, ?)",
                     (dt.DateTimeFromTicks(crtime), lv, localname))
        connection.commit()
        os.unlink(f)

I'm rather pleased with it, especially as it also logs the files into a
relational database to allow web-based call processing to begin. Though I
magine it will be now be picked over for hygiene and sanity at endless
length. I confess I originally omitted the binary attribute in file open
modes, which is all too easily done. Unlike omitting "w" when you want to
write, it can silently modify the data rather than raising an exception.

Anyway, a shot in the dark. It's been too long since I posted some actual
code.

One nice point: having been a good drone and provided the client with a
reliable-though-not-inspired Dreamweaver-based VBScript web solution (rather
than fight with them to get to do it in Python) they have now trusted me to
install 2.2.2 on their database server, and I anticipate being able to write
Python web applications as well soon. Yea! Woo, hoo!

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From jon+usenet at unequivocal.co.uk  Wed Apr 16 09:12:44 2003
From: jon+usenet at unequivocal.co.uk (Jon Ribbens)
Date: 16 Apr 2003 13:12:44 GMT
Subject: Pythonic way of web-programming
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <B8Pma.20554$T34.656139@news2.tin.it> <mailman.1050471334.22000.python-list@python.org> <0gqdnfw_XuYm3gCjXTWcow@speakeasy.net>
Message-ID: <slrnb9qlmb.djs.jon+usenet@snowy.squish.net>

In article <0gqdnfw_XuYm3gCjXTWcow at speakeasy.net>, A.M. Kuchling wrote:
>> I don't think FastCGI has much relation to CGI.  I don't entirely
>> understand FastCGI, but from what I can tell it just sucks -- it's
>> easier to do the same thing with a simpler socket connection ...

You're wrong - FastCGI is excellent. It's a reasonably simple,
standard, well-supported, platform-independent way of speeding up your
CGI scripts. It provides nearly all the advantages (and in some ways,
more advantages) of something like mod_python without any of the
disadvantages of having to run your scripts in the same process as the
web server.

> The FastCGI protocol is really complicated because it supports multiplexing 
> several different transactions over one socket.  Data is broken up into
> multiple packets, each with a unique transaction ID, and the process on the
> other end can then reassemble them and hand them out to different threads.
> However, I'm not sure the Apache mod_fastcgi supports this and few client
> modules seem to, so all of that complexity buys nothing in most cases.

You are correct in that mod_fastcgi does not support this
multiplexing. I am not aware of any clients that do support it either,
except the one I wrote (jonpy) ;-). I don't agree that FastCGI is
particularly complex though, and even if you think it is it makes no
difference if you simply use a ready-made library to take care of all
the FastCGI work for you.



From jbublitzNO at SPAMnwinternet.com  Sat Apr  5 03:49:22 2003
From: jbublitzNO at SPAMnwinternet.com (Jim)
Date: Sat, 05 Apr 2003 08:49:22 GMT
Subject: ANN: PyKDE-3.5-1 Released
Message-ID: <3E8E973F.9060707@SPAMnwinternet.com>

The most recent version of PyKDE (PyKDE-3.5-1) is now
available at:

http://www.riverbankcomputing.co.uk/pykde

PyKDE is a set of Python bindings for the basic KDE
libraries (kdecore, kdeui, kio, kfile, kparts, khtml
and others). It allows you to build KDE applications
in Python and use KDE widgets, dialogs and other
features as well. It's also compatible with PyQt
which allows use of the Qt widget set. PyKDE and PyQt
cover nearly all features, classes and methods from
their respective widget sets.

PyKDE-3.5-1 supports any KDE version from 2.1.1 through
3.1.1, using  Qt from 2.3.0 through 3.1.1, and Python
from 1.5.2 through 2.2.2. PyKDE-3.5-1 *requires* that
you have sip 3.5 and PyQt-x11-gpl-3.5 installed. Both
are available on the riverbankcomputing site. Debian
packages and rpms will probably be forthcoming over
the next few weeks. rpm packages for previous versions
are available at:

http://sourceforge.net/projects/pykde

This release is a "not quite everything I wanted to
get done" release but slightly more complete than a
development snapshot. The full release should be
available when Python 2.3 and sip 3.6 are released.

This release will *not* build with the recent sip 3.6
development snapshots without a minor fix. You can
find the fix or post questions, comments or bug reports
on the PyKDE mailing list at:

http://mats.gmd.de/mailman/listinfo/pykde

Future releases of PyKDE will be dropping support for
KDE 2.x.x, Qt 2.x.x and some older Python versions.

Jim




From staschuk at telusplanet.net  Tue Apr 29 17:11:56 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 15:11:56 -0600
Subject: Another itertool function?
In-Reply-To: <C0ura.19455$3M4.563854@news1.tin.it>; from aleax@aleax.it on Tue, Apr 29, 2003 at 12:14:26PM +0000
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <45pra.20060$xw4.19534@nwrdny01.gnilink.net> <C0ura.19455$3M4.563854@news1.tin.it>
Message-ID: <20030429151156.A1725@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
> Raymond Hettinger wrote:
>    ...
> > Here's another idea that was submitted:
> > 
> > def windowed(iterable, group=2):
> >     "returns (i[0], i[1], ... i[group-1]), (i[1], i[2], ... i[group]),
> 
> I think this would be perfect if the overlap between successively
> returned windows was also a parameter, rather than being fixed to 1.

Alexander Schmolck proposed such a function back in February [1];
his version included a "step" argument doing what you suggest.

In the discussion of whether such a function was useful enough to
be included in itertools, I provided several use cases [2], but
was unable to find a convincing one which needed a step other than
one.  Do you have one in mind?

[1] <http://groups.google.com/groups?selm=yfsbs172bjw.fsf%40black132.ex.ac.uk>

[2] <http://groups.google.com/groups?selm=mailman.1045851191.21005.python-list%40python.org>

-- 
Steven Taschuk             "[W]e must be very careful when we give advice
staschuk at telusplanet.net    to younger people: sometimes they follow it!"
                             -- "The Humble Programmer", Edsger Dijkstra




From aahz at pythoncraft.com  Wed Apr  9 21:00:53 2003
From: aahz at pythoncraft.com (Aahz)
Date: 9 Apr 2003 21:00:53 -0400
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1049901964.20840.clpa-moderators@python.org>
Message-ID: <b72fo5$bq6$1@panix1.panix.com>

In article <mailman.1049901964.20840.clpa-moderators at python.org>,
Anders J. Munch <andersjm at dancontrol.dk> wrote:
>
>Capon 0.1
>---------
>Capon is a software build framework.
>Or you might say, a Make replacement.
>Or you might say, an alternate notation for Python.
>Or you might say, a delegation-based object-oriented programming
>language.
>
>Why Capon?
>----------
>Capon blurs the line between a conventional make utility and a
>general-purpose programming language to provide the best of both
>worlds.

So how does Capon differ from SCons?
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From max at alcyone.com  Wed Apr 16 19:23:28 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 16:23:28 -0700
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <3E9DE5F0.623D4CD8@alcyone.com>

Christopher Blunck wrote:

> Was just reading a thread about how python implements protected and
> private methods.  I found JP's response quite interesting (this is
> something I've heard numerous friends of mine say of the language):

I certainly do extreme programming.  I always do my best programming
while parachuting out of planes, bungee off of towers, and rappeling
into ceiling light caves.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Would a harvest from my heart / Find its season
\__/ Sandra St. Victor
    Python chess module / http://www.alcyone.com/pyos/chess/
 A chess game adjudicator in Python.



From goodger at python.org  Wed Apr 30 00:57:44 2003
From: goodger at python.org (David Goodger)
Date: Wed, 30 Apr 2003 00:57:44 -0400
Subject: POD vs. reST for standalone writing?
In-Reply-To: <7gy91ssyrl.fsf@pikespeak.metacarta.com>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <mailman.1051590229.26633.python-list@python.org> <7gy91ssyrl.fsf@pikespeak.metacarta.com>
Message-ID: <3EAF57C8.10401@python.org>

eichin at metacarta.com wrote:
> The particular writability issue that comes to mind is the use of
> underlining characters, quoting from quickstart.html:
...
> and the description "The underline/overline must be at least as long
> as the title text."  My first problem with this is that if changing
> the content causes the markup to need changing, the markup changes
> will be a bit more error prone.

If you're liable to edit your titles a lot, just use extra-long underlines:

     Title
     ==============================================================

That gives you plenty of room for growth.

> The next is that level for a given
> punctuation type is based on, if I'm not misinterpreting this, what
> punctuation usage has been seen *in this document*, which makes moving
> selections among a small set of documents a little harder, unless you
> have an explicit style guide (in which case, why not represent it
> outright in the code?)
 >
> POD here would just say
> 
> =head1 Chapter 1 Title
> 
> =head2 Section 1.1 Title
> 
> =head3 Section 1.1.1 Title

There is no one best way, IMO.  Given good enough tools, the problem 
disappears, whatever scheme is used.  There's some Emacs code to aid 
with title underline style changes when editing reStructuredText files.

Explicitly specified levels have their own shortcomings.  They have to 
be edited when the doc structure changes (change all "head2" to "head3", 
but don't forget to change all "head3" to "head4" first [etc etc] or 
else the doc structure will be screwed).

With underline styles, doc structure can be altered by adding new 
styles.  For example, a 1st-level section can become a 2nd-level section 
by adding a new top-level title.

> Perhaps this is just one corner of the notation - but I'm looking for
> a case that simply doesn't have all that much notation, or that much
> to explain to the end user.

If the end user in question is the *reader* (which is the case with 
reStructuredText), I think the underline-style method is most 
transparent.  In other words, the reader doesn't notice the markup or care.

-- David Goodger






From gustav at morpheus.demon.co.uk  Mon Apr 14 13:41:29 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Mon, 14 Apr 2003 18:41:29 +0100
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <mailman.1050336802.11273.python-list@python.org>
Message-ID: <vfxhrn7a.fsf@morpheus.demon.co.uk>

Skip Montanaro <skip at pobox.com> writes:

>     Paul> To use a concrete example, I'd like to print the Euro symbol. A
>     Paul> Unicode string for this is u'\20a0'
>
>     >>>> unicodedata.name(u'\u20a0')
>     Paul> 'EURO-CURRENCY SIGN'
>
>     ...
>     >>>> print u'\u20a0'.encode("latin-15")
>     ...
>     >>>> print u'\u20a0'.encode("latin-1")
>     ...
>     >>>> print u'\u20a0'.encode("iso8859_15")
>     ...
>     >>>> print u'\u20a0'.encode("cp1258")
>     ...
>     >>>> print u'\u20a0'.encode("mbcs")
>     ...
>     Paul> [BTW, in case it's relevant - I know it probably is - the output
>     Paul> from "chcp" at the console prompt is "Active code page: 1252"]
>
> Seems that cp1252 is the most likely encoding you didn't try.  What happens
> if you execute
>
>     print u'\u20a0'.encode("cp1252")

Doh. Dumb typo. But no joy with that:

>chcp 1252
Active code page: 1252

>python
Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> print u'\u20a0'.encode("cp1252")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "C:\Apps\Python\lib\encodings\cp1252.py", line 18, in encode
    return codecs.charmap_encode(input,errors,encoding_map)
UnicodeError: charmap encoding error: character maps to <undefined>

I don't know if this makes me feel better or worse...

Paul.
-- 
This signature intentionally left blank



From staschuk at telusplanet.net  Thu Apr  3 23:42:07 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 3 Apr 2003 21:42:07 -0700
Subject: How to compile a python script to *.pyo
In-Reply-To: <3407.192.168.2.15.1049424214.squirrel@iamlearning.com>; from danny@iamlearning.com on Thu, Apr 03, 2003 at 08:43:34PM -0600
References: <2977.143.166.82.240.1049308819.squirrel@iamlearning.com> <3E8C8A57.9080603@gatwick.westerngeco.slb.com> <3407.192.168.2.15.1049424214.squirrel@iamlearning.com>
Message-ID: <20030403214207.C2938@tibia.amotlpaa.bogus>

Quoth danny at iamlearning.com:
  [...]
> For Martin,
> Thanks for your advice. That's what I need.
> If this mailing list got more people like you, there are less junk
> messages to read.

I wonder whether you intended to imply that the previous responses
were "junk".  If you did, I urge you to reconsider; the previous
posters were actually being helpful.  It is unfortunate (for you)
if you don't realize that.

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From ckea25d02 at sneakemail.com  Tue Apr 15 23:47:15 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 15:47:15 +1200
Subject: Why functional Python matters
In-Reply-To: <mailman.1050459866.18751.python-list@python.org>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <slrnb9p3vo.vg8.ramen@lackingtalent.com> <20030416004537.GB7550@meson.dyndns.org> <mailman.1050459866.18751.python-list@python.org>
Message-ID: <b7ij37$1aacq$1@ID-169208.news.dfncis.de>

Jack Diederich wrote:
> I was a little disapointed that the implementaion in bltinmodule.c doesn't
> special case lambdas but instead does a regular function call.

How exactly would you propose to special-case lambdas?

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From exarkun at intarweb.us  Sun Apr  6 01:19:50 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 6 Apr 2003 01:19:50 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
Message-ID: <20030406061950.GA25203@meson.dyndns.org>

On Sun, Apr 06, 2003 at 04:32:59AM +0000, Neil Schemenauer wrote:
> Ian Bicking <ianb at colorstudy.com> wrote:
> > The next step I can envision would be for Twisted to become the
> > canonical infrastructure, where higher-level web frameworks are all
> > built ontop of it.
> 
> That's not feasible.  Some applications would require extensive
> modification to work on top of an event driven framework.

  That's a shame.  I think this indicates that they are poorly designed.  ;)

  Seriously, though.  The input source of an application should *not* be
very tightly integrated with the actual logic of the application.  I can see
how this is sometimes unavoidable, though, and how other times it simply is
more (perceived) work than it's (perceived) worth.  Perhaps Ian should have
said "canonical event-driven infrastructure".

  Jp

-- 
In the days when Sussman was a novice Minsky once came to him as he sat
hacking at the PDP-6. "What are you doing?" asked Minsky. "I am training a
randomly wired neural net to play Tic-Tac-Toe." "Why is the net wired
randomly?" asked Minsky. "I do not want it to have any preconceptions of how
to play." Minsky shut his eyes. "Why do you close your eyes?" Sussman asked
his teacher. "So the room will be empty." At that moment, Sussman was
enlightened.
 -- 
 up 17 days, 2:00, 5 users, load average: 1.13, 1.17, 1.09




From ckea25d02 at sneakemail.com  Mon Apr  7 23:16:52 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 15:16:52 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net> <b6qs04$7eqi1$1@ID-169208.news.dfncis.de> <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
Message-ID: <b6tedc$8rgif$1@ID-169208.news.dfncis.de>

Carl Banks wrote:
> Second, given that the keyboard focus can only be active in one buffer
> at a time, the idea of a current buffer is absolutely necessary.

There certainly needs to be a notion of which buffer
has the keyboard focus. But I was thinking of the way
that lots of elisp code uses current-buffer as a
kind of implicit parameter, when it really has
nothing to do with keyboard focus. The fact that
there exist things like with-buffer is a testament
to this.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From tdelaney at avaya.com  Mon Apr 28 20:42:55 2003
From: tdelaney at avaya.com (Delaney, Timothy C (Timothy))
Date: Tue, 29 Apr 2003 10:42:55 +1000
Subject: problem with string.lower function
Message-ID: <338366A6D2E2CA4C9DAEAE652E12A1DE4AC159@au3010avexu1.global.avaya.com>

> From: Don Low [mailto:mt at open2web.com]
> 
> Thanks, you're right. The line I added used spaces for 
> indentation while
> the other lines used tabs. 
> 
> >>         for eachLine in allLines:
> >>         test = string.lower(eachLine) // change is here
> >>         if string.find(test, word) > -1:
> >>                 count = count + 1
> >>                 print eachLine,

For this reason you should always invoke python as

    python -tt script.py

while developing.

Tim Delaney




From claird at lairds.com  Fri Apr 25 08:11:15 2003
From: claird at lairds.com (Cameron Laird)
Date: Fri, 25 Apr 2003 12:11:15 -0000
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it>
Message-ID: <vai9f3ug54144@corp.supernews.com>

In article <Dl9qa.26067$DN.734905 at tornado.fastwebnet.it>,
Alessio Pace  <puccio_13 at yahoo.it> wrote:
>"Martin v. L?wis" wrote:
>
>> Alessio Pace wrote:
>> 
>>> Hi, I'd like to make a GUI for a program written in python, meant to run
>>> on mac, win and linux: which is the preferable solution? I saw around
>>> tkinter, wxpython and pyqt....
>> 
>> I recommend Tkinter. It comes with the standard Python installation, and
>> it works quite well atleast on Windows and X11 (I've never used it on
>> the Mac).
>> 
>> Regards,
>> Martin
>
>That's what I had in mind, but after reading some tutorials on Tkinter I saw
>that wxpython and pyqt have more features still remaining portable...mmm...
			.
			.
			.
1.  Each of the choices has serious advantages.
    Even toolkits zealots recognize that no one
    toolkits strictly dominates any other.
2.  Each toolkit has fans.
3.  It's relatively easy to download/install/
    experience Python (or Perl or ...), and ac-
    quire your own "feel" for each language.
    The GUI toolkits are a bit heavier and 
    clumsier; it generally takes longer to gain
    a proper understanding of them.
4.  Comparisons of GUI toolkits are popular.
    Such discussions often erupt here and else-
    where.  Abundant details are available, if
    reading the words of others interests you.
5.  Tkinter is the easiest for a newcomer, in
    its installation and "first launch".  It's
    the best documented, in the sense of making
    appearances in the most in-print published
    books.

I think it's reasonable for you to start with Tkinter.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From mb at muenster.de  Thu Apr 24 13:51:22 2003
From: mb at muenster.de (Martin Bless)
Date: Thu, 24 Apr 2003 17:51:22 GMT
Subject: Bug? Pythonwin editor trouble
References: <3ea58721.1845828@news.muenster.de> <mMspa.7849$8K2.70671@news-server.bigpond.net.au>
Message-ID: <3ea823e5.749718@news.muenster.de>

["Neil Hodgson" <nhodgson at bigpond.net.au>]

>   OK, this was fixed in a more recent (1.48 or later) version of Scintilla.
>Eventually this will be bundled with Pythonwin.

... and more.

Ok, thank you. Good to know!

Martin




From guy at obstruction-no-spam.com  Fri Apr 18 16:21:08 2003
From: guy at obstruction-no-spam.com (Guy Middleton)
Date: Fri, 18 Apr 2003 20:21:08 GMT
Subject: Pythonic way to sum n-th list element?
Message-ID: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>

What is the most Pythonic way to sum the n-th element of a list of lists (or
of tuples)?

The obvious (to me) way is to use reduce, a list comprehension, and a lambda:

	$ python
	Python 2.2.1 (#1, Jul  4 2002, 15:23:46) 
	[GCC 2.95.3 20010315 (release) [FreeBSD]] on freebsd4
	Type "help", "copyright", "credits" or "license" for more information.
	>>> x = (("a", 1), ("b", 2), ("c", 3))
	>>> # want to sum 1+2+3 = 6
	... 
	>>> reduce(lambda a, b: a+b, [y[1] for y in x])
	6
	>>>
	$

But I don't really like lambda, is there a better way to do this?

 -Guy



From theller at python.net  Wed Apr 16 05:16:49 2003
From: theller at python.net (Thomas Heller)
Date: 16 Apr 2003 11:16:49 +0200
Subject: A better unittest
Message-ID: <3ckier9a.fsf@python.net>

I'm moving more and more to TDD (test driven development [1]), which
makes tests an integral part of the development process.  Failing tests
are ocurring very often with this development style, because you write
the tests first and then implement the functionality.

Having said this, I'm not really happy with the output of
unittest. Consider the following testcase:

#-----
class FailingTests(unittest.TestCase):
    def test_exception(self):
        return 1/0

    def test_fail(self):
        self.fail("fail on purpose")

    def test_failIf(self):
        self.failIf(1, "failIf on purpose")

if __name__ == '__main__':
    unittest.main()
#-----

Running this code, you get this output:

-----start output-----
EFF
======================================================================
ERROR: test_exception (__main__.FailingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 32, in test_exception
    return 1/0
ZeroDivisionError: integer division or modulo by zero

======================================================================
FAIL: test_fail (__main__.FailingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 35, in test_fail
    self.fail("fail on purpose")
  File "C:\Python22\lib\unittest.py", line 254, in fail
    raise self.failureException, msg
AssertionError: fail on purpose

======================================================================
FAIL: test_failIf (__main__.FailingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 38, in test_failIf
    self.failIf(1, "failIf on purpose")
  File "C:\Python22\lib\unittest.py", line 258, in failIf
    if expr: raise self.failureException, msg
AssertionError: failIf on purpose

----------------------------------------------------------------------
Ran 3 tests in 0.000s

FAILED (failures=2, errors=1)
-----end output-----

It can be seen that the codeline in the unittest module where the
failure is raised is included in the output.

It has been discussed before that raising an exception in the callers
frame is not currently possible, but it is possible to change the
unittest output to not print the innermost stack frame.

I've patched unittest to do this, and now the output looks like this:

-----start output-----
EFF
======================================================================
ERROR: test_exception (__main__.FailingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 32, in test_exception
    return 1/0
ZeroDivisionError: integer division or modulo by zero

======================================================================
FAIL: test_fail (__main__.FailingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 35, in test_fail
    self.fail("fail on purpose")
AssertionError: fail on purpose

======================================================================
FAIL: test_failIf (__main__.FailingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 38, in test_failIf
    self.failIf(1, "failIf on purpose")
AssertionError: failIf on purpose

----------------------------------------------------------------------
Ran 3 tests in 0.000s

FAILED (failures=2, errors=1)
-----end output-----

The big advantage is that if I execute the tests from within XEmacs with
C-c C-c, the output ocurrs in a *Python Output* buffer, and the cursor
in the source buffer jumps to the real cause of the error: the
self.failxxx() call.

What do people think? Is this a useful change, and should I submit a
patch?

Thomas

[1] Kent Beck: Test Driven Development: by example. Addison Wesley, 2002.
    ISBN 0-321-14652-0



From afriere at yahoo.co.uk  Tue Apr 29 00:50:41 2003
From: afriere at yahoo.co.uk (Asun Friere)
Date: 28 Apr 2003 21:50:41 -0700
Subject: Class attribute or instance attribute?
References: <vaqtlsqbi52j93@corp.supernews.com> <2259b0e2.0304281329.12ac184b@posting.google.com>
Message-ID: <38ec68a6.0304282050.3c17fc75@posting.google.com>

mis6 at pitt.edu (Michele Simionato) wrote in message news:<2259b0e2.0304281329.12ac184b at posting.google.com>...
> "Paul Watson" <pwatson at redlinec.com> wrote in message news:<vaqtlsqbi52j93 at corp.supernews.com>...
> > When is an attribute a class attribute or an instance attribute?
> 
> I give you a subtle example:
> 
On the other hand here is a totally obvious example illustrating the same ;-)

class Foo (object) :
    foo = 'spam'
    def __init__ (self) :
        print Foo.foo, self.foo
        self.foo = 'ham'
        print Foo.foo, self.foo

>>> f = Foo()
spam spam
spam ham
>>> print f.foo, Foo.foo
ham spam



From ckea25d02 at sneakemail.com  Thu Apr  3 20:39:36 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 04 Apr 2003 13:39:36 +1200
Subject: Roman assignments (Re: Overloadable Assignment PEP)
In-Reply-To: <a8b7f07a.0304031432.3c585657@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <a8b7f07a.0304031432.3c585657@posting.google.com>
Message-ID: <b6in8i$600fi$1@ID-169208.news.dfncis.de>

A. Lloyd Flanagan wrote:
> Oh, good, if this had turned out to be a joke, I'd be afraid I won't
> get my Roman numeral literals after all.

The Roman numerals PEP seems unlikely to get passed
in any case, but perhaps we can rescue something from
it to help with this problem: we could apply
augmented assignment semantics when and only when
the name being assigned to resembles a Roman numeral.

This very neatly solves the "voltage" problem,
because then

   V = 3

would be an augmented assignment, whereas

   Voltaire = 3

would just be an ordinary assignment.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ckea25d02 at sneakemail.com  Mon Apr 14 22:32:34 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 15 Apr 2003 14:32:34 +1200
Subject: Adding item in front of a list
In-Reply-To: <slrnb9l4od.1fg.do-not-spam-ben.hutchings@tin.bwsint.com>
References: <mailman.1050038237.25551.python-list@python.org> <ZUXla.7029$T34.191442@news2.tin.it> <mailman.1050176305.12881.python-list@python.org> <CE0ma.5922$LB6.169206@news1.tin.it> <b7deoe$devs8$1@ID-169208.news.dfncis.de> <slrnb9l4od.1fg.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <b7fqbl$geod$1@ID-169208.news.dfncis.de>

Ben Hutchings wrote:
> 3. The non-keyword arguments are assigned to the first (m - n) unmatched
>    optional parameters, and the n unassigned required parameters, in
>    order of parameter declaration.

Oh, man! I had to do some serious thinking to get my
brain around that.

> Then range would be range(start=0, stop, step=1).  But having existing
> arguments match different parameters when you add arguments to the end
> of the list could be confusing.

Well, it wouldn't be any more confusing than what
range() already does.

Having optional parameters in the *middle* of the
list would be interesting, however...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From jimmy at retzlaff.com  Tue Apr 15 11:41:57 2003
From: jimmy at retzlaff.com (Jimmy Retzlaff)
Date: Tue, 15 Apr 2003 08:41:57 -0700
Subject: DB API for ADO/DAO?
Message-ID: <A7B5C22F20EE3C4E95744E2F03F1887F0F1D36@MAYOR.retzlaff.com>

Steve Holden (sholden at holdenweb.com) wrote:
> "John Burton" <john.burton at jbmail.com> wrote in message
...
>> The problem with odbc is that you need to manually set up data
sources
>> using the control panel which isn't something I want the users to
have
>> to do.
...
>As well as the other responses pointing you to ADO code, also note that
>mxODBC allows you to use DSN-less data sources.
...
>     DriverConnect('FILEDSN=test.dsn;UID=test;PWD=test')
>
>This is sometimes useful when you want to dynamically setup a data
source,
>e.g. a MS Access database."""

The current version of the win32all ODBC driver has this same
capability. It is spelled:

odbc('FILEDSN=test.dsn;UID=test;PWD=test')

To connect to the master database of SQL Server on the same machine:

odbc('Driver={SQL
Server};Server=(local);Database=master;UID=test;PWD=test')

Jimmy




From max at alcyone.com  Thu Apr 17 04:06:31 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 17 Apr 2003 01:06:31 -0700
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr>
Message-ID: <3E9E6087.7F84E805@alcyone.com>

Jean-Paul Roy wrote:

> - Python is bad : tail recursion is not iterative (also astonished at
> that, I can understand with an interpreter, but compiler ?)

I think the reason for this is that there simply hasn't been enough
pressure to do so.  I doubt anyone thinks it would be _bad_, it's just
not enough people have insisted it is worth doing.

> Aside from that points, a personal question as a Schemer who reads the
> Python primer:
> - how do you ask if a list is empty ? Is the following ok ?
>       if L == [] :

Even simpler:

	if L: ...

> - how do you ask if the list has only one element, with time 0(1) ?
> The
>   Scheme/Lisp analog is (null? (cdr L))

I suspect the length of a Python list is cached in the list object, but
don't know that for a fact (a simple test would verify it easily
enough).  I would think that

	if len(L) == 1: ...

would be fine.

> - more generally, is the length of a list a builtin ?

I think so but have not checked.

> Is the L[k]
 access
>   time O(1) with a list ? Probably not. And with a tuple ?

Python lists and tuples are both implemented as computer science
"vectors," so indexing access time is indeed O(1).

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Can I walk with you / 'Till the day that my heart stops beating
\__/ India Arie
    Polly Wanna Cracka? / http://www.pollywannacracka.com/
 The Internet resource for interracial relationships.



From jkraska at san.rr.com  Wed Apr 23 01:07:00 2003
From: jkraska at san.rr.com (Courageous)
Date: Wed, 23 Apr 2003 05:07:00 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <3EA2C43F.2080802@tismer.com> <20030421121843.A596@tibia.amotlpaa.bogus> <3EA4A843.2050507@tismer.com> <20030422202417.A1462@tibia.amotlpaa.bogus> <mailman.1051064984.17837.python-list@python.org>
Message-ID: <fi7cavkndoi0hhel77d8occtjiqm6r002o@4ax.com>

>Continuations aren't ...

But what continuations *are* is a mechanism that can be
very useful to those experimenting with metalinguistics
in computer programming language. In that sense, they are
very lispish. It's certainly true that continuations have
enabled me to build a framework that resembles something
that otherwise is best implemented in a first class language
(and also a temptation to do just that).

C//




From jcribbs at twmi.rr.com  Wed Apr  9 00:21:45 2003
From: jcribbs at twmi.rr.com (Jamey Cribbs)
Date: Wed, 09 Apr 2003 04:21:45 GMT
Subject: [ANNOUNCE] KirbyBase 1.0
Message-ID: <tdNka.76495$0X.14720956@twister.columbus.rr.com>

Announcing the initial release of KirbyBase.

It can be downloaded from http://www.netpromi.com/files/kirbybase1.0.zip

What is KirbyBase?
---------------------

KirbyBase is a simple, pure-python, flat-file database management 
system, with a few twists.

1. All data is kept in plain-text, newline delimited files that can be 
edited by hand.  This gives you the ability to make changes by just 
opening the file up in a text editor, or you can use another programming 
language to read the file in and do things with it.

2. KirbyBase can either be used as an embedded database or in a 
client/server, multi-user mode.  To switch from one mode to the other, 
you only have to change one line in your program.  Included in the 
distribution are two sample database server scripts.  One is 
single-threaded, blocking, and pretty safe.  The other one is an 
experimental mult-threaded, non-blocking script.  So far it has worked 
well, but it probably needs more testing.

3. KirbyBase is not an "in-memory" database.  The entire database is NOT 
read into memory at initialization and then completely written back out 
to disk when it is closed.  KirbyBase tables are kept on disk and 
accessed from disk when selecting, updating, inserting, and deleting.  
Changes to the database are written immediately to disk.  This has the 
advantage of accomodating larger databases, because you aren't taking up 
lots of memory to hold the entire database.  Additionally, since changes 
are immediately written to disk, the chance of lost data due to power 
interruptions, crashes, etc.  should be smaller.

4. KirbyBase allows you to specify field types, then it  allows you to 
use Python's regular expression syntax to search by string fields and 
Python expressions to search by numeric fields.  So, for exampe, to 
search a table containing info on WWII fighter planes and select US 
planes that had a maximun speed higher than 400 mph, you would say 
something like:

     db.select('plane.tbl', ['country', 'speed'], ['USA', '>400'])

I like doing this more than building SQL query strings in Python.

That's about it.  There are a few other goodies, like automatic 
auto-incrementing primary keys on each record and a kind of interesting 
way to keep updates quick by trying to write back to the same spot in 
the physical file, but it's mostly internal stuff.


What's it good for?
-------------------

I'm not sure.  :)  If you are looking for an easy to use dbms that has a 
syntax that is a little more "python friendly" than SQL, has an internal 
data format that is easily accessible (text files), doesn't take up much 
room (one file, 34K), that can be used embedded or client/server, and 
that allows you to change how it works (source code is there for you to 
see/modify), then you might find KirbyBase interesting.  I guess I see 
it fitting somewhere in between pysqlite and plain text files for your 
dbms needs.


Where did KirbyBase come from?
-----------------------------------

A lot of the ideas and concepts for KirbyBase came from a tcl dbms 
script written by Dave Maggiano.  Dave included this script on the cdrom 
that accompanied his book "CGI Programming with Tcl".  I bought the 
book, which is excellent, and, in the course of reading it, discovered 
his dbms script.  I have a soft-spot for small, simple, cleverly 
designed database management systems, and Dave's was one of the coolest 
I have seen.  The icing on the cake was that I could actually understand 
and follow the source code!  I'm no genius at programming, so that was 
definitely a plus.

Well, I used his dbms in a few Tcl projects at work and it performed 
flawlessly.  However, when I started using other languages like Python 
and Ruby, I had to leave Dave's tcl-based dbms behind.  I told myself 
that one day I would convert it into another language.

Well, lately I have been learning Python while writing a gui program 
using pygtk.  I really like the combination of Python and Gtk, so I 
plunged in and decided to port Dave's dbms to Python.  As I started to 
do this, I noticed a lot of things that could be done differently 
because, of course, Python is a different language.  Also, because Dave 
and I are unique human beings, I also took the opportunity to add some 
functionality that Dave's script didn't have, and take some stuff away 
that I didn't find as important.  The end result is a program that uses 
a lot of Dave's concepts, but has some ideas of its own.

I have tried, in vain, to contact Dave and get his blessing/permission 
for all of this.  I have looked through his book, source code, and cdrom 
to try to find out what license his script is under, but I have been 
unable to find anything.  I have tried three different emails that were 
supposed to be his, but received no response.  I even posted a message 
to the Tcl mailing list asking if anyone knew how I could contact Dave, 
but no one did.  So, Dave, if you read this...please contact me.  I want 
to, first of all, thank you for such an awesome piece of software, and, 
secondly, make sure you are ok with what I am doing.  I'm not directly 
using any code you wrote, but I am definitely borrowing some concepts 
from you and I want to make sure that is alright with you.


Where to go from here?
------------------------

KirbyBase can be downloaded from 
http://www.netpromi.com/files/kirbybase1.0.zip I don't know if it will 
prove of value to anyone, but me.  I do hope some of you could take a 
look at it and give me feedback.  I am sure that there are a lot of code 
improvements that can be made and I would welcome any suggestions.  
Right now, it performs pretty well on small databases, but starts to 
drop off when doing large queries against 50,000+ record databases.  
Mostly, writing KirbyBase fulfilled a promise I made to myself.  I also 
am using it for a few projects at work.  If nothing else, I have learned 
a LOT about Python, just by writing KirbyBase, and, it was actually kind 
of fun doing it. 


Jamey Cribbs
jcribbs at twmi.rr.com




From BPettersen at NAREX.com  Sat Apr 19 18:36:11 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sat, 19 Apr 2003 16:36:11 -0600
Subject: Is this a super bug?
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE00A@admin56.narex.com>

M:\python>python
Python 2.3a2 (#39, Feb 19 2003, 17:58:58) [MSC v.1200 32 bit (Intel)] on
win32
Type "help", "copyright", "credits" or "license" for more information.
>>> class ff(str):
...    def foo(self):
...      print 'str.__getitem__', str.__getitem__(self, 0)
...      print 'super(ff, self).__getitem__', super(ff,
self).__getitem__(0)
...      print 'super(ff, self)[0]', super(ff, self)[0]
...
>>> f = ff('asdf')
>>> f.foo()
str.__getitem__ a
super(ff, self).__getitem__ a
super(ff, self)[0]
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "<stdin>", line 5, in foo
TypeError: unsubscriptable object
>>>




From tjreedy at udel.edu  Mon Apr 28 12:46:30 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Mon, 28 Apr 2003 12:46:30 -0400
Subject: Forgetting "()" when calling methods
References: <mailman.1051540303.31725.python-list@python.org>
Message-ID: <BfScnR2GHIkjxzCjXTWcoQ@comcast.com>

"Tim Peters" <tim.one at comcast.net> wrote in message
news:mailman.1051540303.31725.python-list at python.org...
> No, the current state makes no sense ... [and] ...is an inconsistent
mish-mash

OK, we agree on that.  It a result of various local optimazations and
compromises and =needs ground up redesign.  My point in relation to ()
is that I would hate to see a similar mishmash created in relation to
truth values, especially not for the local optimazation of allowing
people to forget to type '()' in boolean context.

Back to the tangent.

> Guido doesn't want
>     2 < "42"
> to pass silently any more than he wants
>     2 + "42"
> to pass silently.

I agree.  The question is whether .sort() should come with at least an
option to work with a heterogenous items and not raise an exception -- 
without the user having to write a custom order function  -- so
programmers who know what they are doing when trying to sort such a
list can do so nearly as easily as they used to.

Terry





From gbrunet at nospamsempersoft.com  Mon Apr 21 15:16:31 2003
From: gbrunet at nospamsempersoft.com (Greg Brunet)
Date: Mon, 21 Apr 2003 14:16:31 -0500
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com>
Message-ID: <va8gtps1kj813d@corp.supernews.com>

"Denis Agar" <canrocks at yahoo.ca> wrote in message
news:25a672ee.0304210914.1cc2049c at posting.google.com...
> I'm blabbering! Anyway, my problem is this. I tried to compile a
> simple Hello World program into something runnable on non-python
> computers, and it came out being 7MB in size. Is there something I'm
> doing wrong? Just so you know, I do know somewhat what I'm doing. I
> know that it's incorporating all of the GUI. So is there a way to not
> incoporate all of the gui, and still have a functional program? I'm
> thinkin in the 800k range for a simple program? The reason I ask is
> because I'm thinking of porting something I made in VB to Python
> before I develop it any further.
>
> PS: The program is 92k in Visual Basic.

Realize that VB has a runtime as well, and if you packaged all of the
supporting files with it that are needed, (and probably already on your
computer), it would likely be in the same range as your 7 MB.  To more
closely compare apple to apples, I would recommend doing the following:
two install/setup programs.  The first with Python and your GUI support
files (in my case wxPython), the second with the required scripts
(packaged using Py2exe or comparable tool).

The first setup program would be comparable to VBRun60.exe which
contains the VB runtime files (though it doesn't include 3rd party OCX's
which can also get quite large).  My version of this has: python22.dll,
wxc.pyd, & wxmsw233h.dll. (or uh.dll) Packaged in a setup program
(compressed), they are about 2 MB.  As long as you don't change versions
of python of wxPython, this setup file could be shared among your
applications.

The second setup program would have your script (converted to an EXE so
that people can just run it without having to load the entire python
environment).  Even with all of the supporting scripts that will come
along in the package, this should only be 1/2 meg in size, maybe less.
In my tests, a fair amount of this size comes from the supporting
wxpython PYC files, which, with a little extra work, could be packaged
into the first setup program.

If you look at it this way, you'll end up with something comparable to
your VB experience.  For a more detailed example of all this, check out:
http://wiki.wxpython.org/index.cgi/SmallApp

-- 
Greg





From pinard at iro.umontreal.ca  Tue Apr  1 08:24:02 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 01 Apr 2003 08:24:02 -0500
Subject: Parsing C
In-Reply-To: <Pine.GSO.4.44.0303311743400.19083-100000@amy4.Stanford.EDU>
References: <Pine.GSO.4.44.0303311743400.19083-100000@amy4.Stanford.EDU>
Message-ID: <oqd6k6bbbh.fsf@alcyon.progiciels-bpi.ca>

[Daniel Timothy Bentley]

> I find myself having to write a C interpreter in Jython.  Really, I'm
> looking for a way to get a C compiler and mildly painlessly build it
> to create the AST that my program will look for.  Oh, and the lexer
> in addition to the parser would be nice, but I bet it's all easy and
> fully of re-goodness in {p|j}ython.

I am not sure if I'm really replying to your desire.

I have this project where I needed to read C source code from Python,
and for it, I recycled a yacc/lex grammar for C, making it available to
Python through a Pyrex interface.  That project not being so important for
me, I merely put it aside unfinished when other emergencies popped up,
with the vague intent of revising it later, one of these days...  Yet,
it might help you getting started, who knows.  Just tell me if you feel
adventurous enough to peruse this previous effort of mine. :-)

A word of warning, though.  I really program for Python and happily rely
on its timely finalisation.  This is my choice.  So, you would need some
revision work if you wanted to use my code in a Jython environment.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From alanmk at hotmail.com  Wed Apr 30 05:20:47 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Wed, 30 Apr 2003 10:20:47 +0100
Subject: HTTP application streams a response to the client gradually - how  
 deal with it?
References: <Xns936CB5697A6D4cpl19ghumspamgourmet@62.153.159.134> <3EAEC97C.1EADEBF7@hotmail.com> <Xns936D814A89Fcpl19ghumspamgourmet@62.153.159.134>
Message-ID: <3EAF956F.CF1D9A26@hotmail.com>

[Harald]

> There is a existing Chat server, using HTTP / HTML. Currently it only
> works with Internet-Explorer as a client. Lynx is unlucky, Mozilla works
> only half.
> 
> So I want to do a Chat Client in Python.

Ah, I see better now what you're trying to do.

Is the following the reason why you cannot use urllib and httplib:
because they block your client from doing anything but connecting to the
server? I.E. You want to be able to read the user input channel at the
same time as retrieving updates from the server. And because urllib and
httplib encapsulate the entire request/response inside one method call,
your client is blocked while the server call is executing, and thus very
unresponsive?

There is a simple solution to that problem: use multiple threads. One
thread reads input from the user, another thread does nothing but
contact the server periodically for updates.

If the latter is the case, post again, and I'll post sample code or
links to same.

> Perhaps translating from the
> HTTP-Stream to another protocol.
>
> So your feedback was valuable - especially the part about Request /
> Response. It looks as if I have to go the socket way. Maybe a second try
> with twisted.

Given what you've said, I'd say it's probably overkill to go down to the
socket level. If you do that, then you'll find yourself both creating
and parsing HTTP messages, in order to interact with the HTTP server,
which is a non-trivial exercise. Urllib and httplib have all the
functionality you need, and (mostly) bug free, having been used by
thousands of people over the years. Rolling your own HTTP protocol would
be an enormous amount of work, due to the wide range of weird and
wonderful ways in which HTTP is used and abused.

regards,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From xtian at toysinabag.com  Tue Apr  1 00:21:31 2003
From: xtian at toysinabag.com (xtian)
Date: 31 Mar 2003 21:21:31 -0800
Subject: Question on passing function as argument
References: <b6abiv$b6t$1@lust.ihug.co.nz>
Message-ID: <d111eafa.0303312121.22bb903@posting.google.com>

matthew <matthew at newsgroups.com> wrote in message news:<b6abiv$b6t$1 at lust.ihug.co.nz>...
> Hi,
> 
> Howto pass a func as an argument to a list?
> 
> event.add_listener('linear', [calc_position(params)])
> 
> the function seems to execute 'in-place' and the event_list[event] shows 
> None.

calc_position(params) is calling calc_position with params before
calling event.add_listener with the result (which is None, in this
case).

To pass a one element list containing the function object named
calc_position to your add_listener method (you might want to rename
that to add_listeners, by the way), you'd do the following:

event.add_listener('linear', [calc_position])

Leaving off the () means that the function isn't called, it's just
passed in.
Now, this isn't quite what you want - it doesn't pass the parameters.

If params is accessible from your notify method, you could just pass
it in when calling the function (guessing at the implementation of the
notify method...):

def notify(self, event, params):
    for listener in self.event_list[event]:
        listener(params)

If you won't be able to access params from notify, you'll need to
package the arguments with the function . You could do this in a
number of ways:

1) defining a local function:

def local_calc():
    calc_position(params)
event.add_listener('linear', [local_calc])

2) defining a local anonymous function:

event.add_listener('linear', [lambda: calc_position(params)])

3) by currying the function:

class SimpleCurry:
    def __init__(self, func, *args, **kw):
        self.func = func
        self.args = args
        self.kw = kw
    def __call__(self):
        return self.func(*self.args, **self.kw)

event.add_listener('linear', [SimpleCurry(calc_position, params)])

If you use one of these, the notify method changes to:

def notify(self, event):
    for listener in self.event_list[event]:
        listener()

Note - the first two options use nested scopes - I'm not sure whether
they're on by default in 2.1, but I don't think so. And the
func(*args, **kw) syntax used in 3 isn't available until then either.
If you're using an earlier version of Python (before 2.1), these
options will be spelt slightly differently, although using the same
basic concepts.

Hope that helps!
xtian



From jason at tishler.net  Mon Apr  7 09:55:51 2003
From: jason at tishler.net (Jason Tishler)
Date: Mon, 07 Apr 2003 09:55:51 -0400
Subject: win98 results (was Re: Cygwin Python regression testers wanted)
In-Reply-To: <3E8E1E83.4020209@yahoo.com>
References: <mailman.1049385789.4709.python-list@python.org>
 <3E8E1E83.4020209@yahoo.com>
Message-ID: <20030407135551.GG2536@tishler.net>

Hal,

On Fri, Apr 04, 2003 at 04:08:35PM -0800, Hal Wine wrote:
> Jason Tishler wrote:
> >3. post any failures and/or hangs other than test_commands
> 
> Final command:
[snip]

Is the above for 98 or 98SE?

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From ckea25d02 at sneakemail.com  Sun Apr  6 22:57:40 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 14:57:40 +1200
Subject: how to test when correct output is "random"?
In-Reply-To: <mailman.1049564865.5908.python-list@python.org>
References: <20030404215312.GE7895@unpythonic.net> <mailman.1049564865.5908.python-list@python.org>
Message-ID: <b6qoto$7v6a1$1@ID-169208.news.dfncis.de>

Steven Taschuk wrote:
> Consider two implementations of this function:
> 	def permute(seq):
> 		"""Returns a copy of seq, randomly permuted."""
> 		# ...

In that kind of situation, what you should probably
be checking is not that a particular permutation is
produced, but just that the output is *some* permutation
of the input.

You may also want to call it a few times and check
that it returns a different permutation each time.

That's probably good enough to pick up whether you've
broken anything.

If you're worried that the output might pass those
tests but still not be random enough, then you're
into statistical testing, as Mr. Tasch suggested.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mirandacascade at yahoo.com  Wed Apr  2 18:29:13 2003
From: mirandacascade at yahoo.com (Miranda Evans)
Date: 2 Apr 2003 15:29:13 -0800
Subject: event handling
Message-ID: <59e5b87.0304021529.b044026@posting.google.com>

In the following code, the while loop is written such that it
terminates when either the double click event in the Excel object
happens or ten seconds has elapsed...whichever comes first.

import time, win32event, win32com.client, pythoncom
class xlEv:
    def OnSheetBeforeDoubleClick(self, sh, target, cancel):
       win32event.SetEvent(self.event)

def demo():
    xl=win32com.client.DispatchWithEvents("Excel.Application", xlEv)
    xl.event = win32event.CreateEvent(None, 0, 0, None) # work around
    xl.Visible=1
    xl.Workbooks.Add()
    start_time = time.time()
    while 1:
       rc = win32event.MsgWaitForMultipleObjects((xl.event,),
            0, 1000, win32event.QS_ALLEVENTS)
       if rc == win32event.WAIT_OBJECT_0:
          print 'event signalled'
          break
       elif rc == win32event.WAIT_OBJECT_0+1:
          pythoncom.PumpWaitingMessages()
       # check for timeout
       if time.time() - start_time > 15:
          print 'timed out'
          break

demo()

Suppose I want to augment the demo() method so that it continues to do
what it's already doing (waiting for either the double click event or
for 10 seconds to elapse) plus one more thing: if the double click
event is detected, display the values of the sh, target and cancel
arguments.

How would one modify the demo() method to do that?



From tebeka at cs.bgu.ac.il  Mon Apr 28 06:17:16 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 28 Apr 2003 03:17:16 -0700
Subject: problems with compiling and finding modules
References: <oprobt7xfxx3zh5c@news.mindspring.com>
Message-ID: <33803989.0304280217.22d2c2de@posting.google.com>

Hello Steven,

> I use a programmer's editor called "Visual SlickEdit".  I'm trying to set 
> it up to work with Python.  So, how do I get it to compile a python file.  
> I can't seem to find a command line method for either compiling a .py file 
> to a .pyc file or even just do a syntax check for finding errors in the 
> editor.  I found a file called py_compile, but this just seems to add a 
> function called compile which I can execute from the interpreter but not 
> the command line.  The file compileall, will compile everything in 
> directory or entire tree, but not just a single file.  I would asume there 
> must be a way short of writing my own single file compile script.  I 
> understand I'm going to have to do something like:
> 
>   python compile_command myfile.py
> 
> but what do I put in for compile_command?
Have a look at: http://mail.python.org/pipermail/python-list/2002-June/107625.html
It's for VIm but you should get the general idea.

> Also, how do I change the working directory from within IDLE?  If I open my 
> .py file from the Windows Explorer using a right click and choose 'Edit 
> with IDLE', everything works fine, but if I 'Open...' a .py file from 
> within IDLE, it can't find modules I import from.  I assume this is because 
> it isn't in the current directory.
Python find modules according to what's in sys.path so:
import sys
sys.path.append('my/python/dir')
should do the trick.

HTH.
Miki



From exarkun at intarweb.us  Tue Apr  8 14:08:04 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 8 Apr 2003 14:08:04 -0400
Subject: Cross-reference 'import' in a class hierarchy
In-Reply-To: <slrnb9580p.7og.marcus@infa.abo.fi>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <mailman.1049778853.27420.python-list@python.org> <slrnb9580p.7og.marcus@infa.abo.fi>
Message-ID: <20030408180803.GA3694@meson.dyndns.org>

On Tue, Apr 08, 2003 at 01:12:07PM +0200, Marcus Alanen wrote:
> >> I continue to be amazed that there's simply no way to say
> >> 
> >>   import "../foo/ook.py"
> >> 
> >> as should be possible to implement any well-architected class hierarchy.
> >
> >  It's simple.  Code should not depend directly on code higher than it in
> >the package hierarchy.  If it does (and it isn't one of a few extremely rare
> >cases), it is misdesigned.
> >
> >  Code shallow in the hierarchy should manage required references to objects
> >required by code deep in the hierarchy by passing them around explicitly. 
> >Any other solution is just bunk.
> 
> Could you explain why using import via PYTHONPATH is different than
> import via current directory?
> 

  Because of how Python determines module identity.

  This line:

    "from foobar import t2 as t1"

  causes "foobar" to be imported, adding "foobar" to sys.modules, then
causes "foobar.t2" to be imported, adding "foobar.t2" to sys.modules, then
binds the local "t1" to that module object.

  This line:

    "import t2"

  causes "t2" to be imported.  Before it is -actually- imported, sys.modules
is checked for the string "t2" to see if the module has already been loaded. 
If it has, that module object is returned, but if it hasn't (and as far as
Python knows, it hasn't, only "foobar.t2" has been imported, and that is
clearly a different module), "t2" is added to sys.modules and the local "t2"
is bound to the new module object.

> [snip]
> 
> Why are the id:s different? That is, why do I get two copies of the A
> class?

  Hopefully you see now that there are two copies of A, because there are
two copies of the module.


> This leads (IMHO) to the fact that one _never_ should use "import X",
> (witness isinstance(t1.A(), t2.A) returning 0...) everything should go via
> PYTHONPATH. Which in turn means that you have to set PYTHONPATH for all
> programs. ?

  One could look at it that way.  One could look at it another way, too:
don't mix imports in such a way that the same module may be referred to by
more than one name.

  This doesn't always mean putting the base of every application in
PYTHONPATH, it just means being consistent about how you import modules.

> 
> Some insight into why this is done as it is would be helpful.
> 

  The whys I can only speculate on.  Maybe someone who had a hand in the
implementation or who has discussed this with such a person could give us a
hint.

  Jp

-- 
Somewhere, something incredible is waiting to be known.
                -- Carl Sagan
-- 
 up 19 days, 15:01, 4 users, load average: 0.10, 0.12, 0.23




From andrew-pythonlist at puzzling.org  Wed Apr 16 10:33:46 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Thu, 17 Apr 2003 00:33:46 +1000
Subject: Python and Schools
In-Reply-To: <eGdna.110601$It5.106431@news2.central.cox.net>
References: <mailman.1050499405.10336.python-list@python.org> <eGdna.110601$It5.106431@news2.central.cox.net>
Message-ID: <20030416143346.GI27886@frobozz.local>

On Wed, Apr 16, 2003 at 02:22:02PM +0000, Steve Holden wrote:
[...]
> 
> For a real example, consider the contortions I have to go through when
> writing VBScript ASP code for a client's project if I want a list (array) of
> strings. Where in Python I can write:
> 
>     sarr = ['the', 'quick', 'brown', 'fox', 'jumps' 'over', 'the', 'lazy',
> 'dog]
> 
> in VBScript I must resort to one of [nontested code follows]:
> 
>     dim sarr(9)
>     sarr(1) = 'the'
>     sarr(2) = 'quick'
>         ...
>     sarr(8) = 'dog'
> 
> or possibly
> 
>     sarr = split("the quick brown fox jumps over the lazy dog")

Or, if my memory isn't failing me:

    sarr = Array("the", "quick", ... "dog")

-Andrew.





From tchur at optushome.com.au  Mon Apr 14 03:28:46 2003
From: tchur at optushome.com.au (Tim Churches)
Date: 14 Apr 2003 17:28:46 +1000
Subject: Sawa sa Maynila (Python in Manila)
Message-ID: <1050305338.4350.119.camel@amihan>

Yet to meet any Python users here in Manila, but there were plenty of
Python books in evidence in the IT sections of the National Book Store
(a large book retail chain). Anyone know of any Python user groups here?
-- 
>From the Manila desk of Tim C






From aleax at aleax.it  Mon Apr 14 03:29:33 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 07:29:33 GMT
Subject: Inserting Python commands into HTML documents
References: <zpqdne0Kztcb0AejXTWcow@comcast.com>
Message-ID: <xrtma.10655$LB6.301836@news1.tin.it>

<posted & mailed>

Em wrote:

> Hello,
> 
> I am an extremely new member of the Python family. I had a job interview
>   where they charged me with the task of creating a (fake) corporate
> website using Python in ten days--with two left now.

Ouch!  What happened to the other eight days?!


> The problem is that what I have read seems to say Python is a CGI
> language... Am I remiss in believing that my website is going to be
> completely run from inside a couple .PY files in the CGI-BIN?

CGI is probably fastest to program, for simple cases, if you know
what you're doing -- and Python's indeed quite suitable for writing
CGI scripts.  But, you do have oodles of other possibilities.


> Is there a way for it to act more like PHP, as in run Python commands
> from inside the (index|contacts|about|etc).html or at least simulate
> that behavior? (This is on a linux server and I am not to use PHP, FYI.)

Sure, there are many ways in which you can choose to use this
debatable "let's mix logic right inside presentation so we can't
easily separate them any more" approach (perhaps excusable when
one's in a deuced hurry, if one's used to it and thus can deploy
it fastest).  My suggestion would be to use "Webware for Python"
and specifically its Webkit component -- I think it will be easiest
for you to download and deploy, particularly because, if need be,
you can drive Webware from any server via CGI, while still leaving
you with a choice of programming models for your Python scripts
("PSP" embedding Python inside HTML, servlets a la Java, templates
[ideally with Cheetah, but other templating approaches are also
well integrable with Webware]...).


Of course, the richness of choices -- which is a blessing in many
respects -- may mean you need TIME to evaluate then and choose
the one most suitable for you... just what you don't have right
now with 2 days' left!  So I suggest you just go for Webware.
Use it's "oneshot" interfacing when you're developing to ensure
all mods you're doing to the pages or servlets ARE picked up
automatically...


Alex




From ramen at lackingtalent.com  Tue Apr 15 15:51:37 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 15 Apr 2003 19:51:37 -0000
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <slrnb9ooms.sho.ramen@lackingtalent.com>

In article <12240-3E9C5F83-187 at storefull-2371.public.lawson.webtv.net>,
W S wrote:
> Are the nutshell and cookbook O'Reilly Python books easier to read?

I just got my copy of the Python Cookbook over the weekend, and I love it so
far. I don't think I'd recommend it for a beginner, but I do find it easy to
read. ;)

> I'm fairly new to programming but trying to learn. I have read Learning
> Python, Core Python, Python Bible and a bunch of perl books

Learning Python is good book, very helpful but a bit outdated. I've heard
good things about the Nutshell book, but I haven't read that one yet. I will
say that Nutshell books in general aren't good introductory texts either,
they're much better as references.

Steve Holden's "Python Web Programming" has a great introduction to the
Python language, as well as relational database design and web applications,
if you're into web stuff. I was surprised at how much introductory material
he covered, and his writing style is easy for me.

Also, don't forget, sometimes the best things in life are free:
http://www.ibiblio.org/obp/thinkCSpy/

Peace,
Dave



From robin at jessikat.fsnet.co.uk  Sun Apr 13 18:02:36 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Sun, 13 Apr 2003 23:02:36 +0100
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk>
 <m38yuepdva.fsf@mira.informatik.hu-berlin.de>
 <LtFzMJA4ZZm+EwDh@jessikat.fsnet.co.uk> <3kjma.11793$T34.382463@news2.tin.it>
 <m3vfxi15xg.fsf@mira.informatik.hu-berlin.de>
Message-ID: <sp06dVA85dm+EwG8@jessikat.fsnet.co.uk>

In article <m3vfxi15xg.fsf at mira.informatik.hu-berlin.de>, Martin v.
L?wis <martin at v.loewis.de> writes
>Alex Martelli <aleax at aleax.it> writes:
>
>> The simplest solution is to have pyRXP return a tree of *LISTS*,
>
>Why is that a solution?
>
>>>> x=[]
>>>> import weakref
>>>> weakref.ref(x)
>Traceback (most recent call last):
>  File "<stdin>", line 1, in ?
>TypeError: cannot create weak reference to 'list' object
>
>Regards,
>Martin
I was going to say that, but you weren't down at the pub :)

-- 
Robin Becker



From mrsinger at usa.net  Tue Apr 15 21:27:11 2003
From: mrsinger at usa.net (Matt Singer)
Date: 15 Apr 2003 18:27:11 -0700
Subject: FTP
Message-ID: <e62b26d3.0304151727.6de2e2e4@posting.google.com>

I have an application with a PYQT interface that uploads large files. 
Is there an async method to get how much data has been sent with
storlines to update a progress bar?  Or is there an alternate ftp
interface?



From greg.bland at valpo.edu  Tue Apr 22 20:52:38 2003
From: greg.bland at valpo.edu (Mitch MF Cummstain)
Date: Wed, 23 Apr 2003 00:52:38 GMT
Subject: pattern match
Message-ID: <220420031949394828%greg.bland@valpo.edu>

               I just started using python and I need help with a
pattern match.
      
   m = re.match("(.{30})(word)(.{30})",paragraph);

I need to match 30 characters on each side of word.  The problem word
is getting treated as a string not a variable like I need it to.  How
can I make it treat word as a variable ?



From mfranklin at tcsi.com  Tue Apr  1 08:59:16 2003
From: mfranklin at tcsi.com (Mark Franklin)
Date: Tue, 1 Apr 2003 05:59:16 -0800
Subject: Implementing a "classmethod" in an extension type
Message-ID: <51CE884FE0BAEE47A12058CC902BDEE718A242@xmail2ca.tcsi.com>

Hi,

I've got an built-in extension type, implemented in C. I want to be able to implement the equivalent of a "classmethod" in that type. 

I thought I might be able to do something like this, which fairly closely mimics what you do in Python:

	PyTypeObject myType = {
		// initialised in the usual way. It's tp_methods field includes a method definition for "MyMethod".......
	};

	PyObject *method = PyObject_GetAttrString(myType, "MyMethod");
	// method is a PyMethodDescrObject
	PyObject *classMethod = PyClassMethod_New(method);
	PyObject_SetAttrString(myType, "MyMethod", classMethod);
	Py_DECREF(method);

	PyType_Ready(myType);


However, when I do this I get a TypeError exception from the PyObject_SetAttrString call: "can't set attributes of built-in/extension type 'MyType' ".

So, how do I get the class method into the type object?


Thanks,
Mark





From gbrunet at nospamsempersoft.com  Wed Apr  2 20:44:51 2003
From: gbrunet at nospamsempersoft.com (Greg Brunet)
Date: Wed, 2 Apr 2003 19:44:51 -0600
Subject: Some newbie questions the FAQs didn't answer!
References: <6ffd97db.0304021132.62f86c7b@posting.google.com>
Message-ID: <v8n5gunvv7teea@corp.supernews.com>

"Jason C." <stigwood4 at yahoo.com> wrote in message
news:6ffd97db.0304021132.62f86c7b at posting.google.com...
> I was thinking about taking Python up this summer.  I was initially
> drawn towards it because I had heard how easy it was to program GUIs
> using the various packages.  But then I read that in order to run a
> python program, the machine had to have the python VM installed, and
> the user had to go to a command line and type "python <program
> name>.py".
>


I wrote up http://wiki.wxpython.org/index.cgi/SmallApp a few months
back, and it covers not only writing a small app, but doing things not
normally shown in tutorials, but what I would consider 'standard' in
writing an application, like accelerator keys on menus, creating an icon
for the app, etc.  At the end of the page, I talk about how to build a
windows EXE, and finally wrapping it in an Install program.  You might
want to check that page out for some tips.  Good luck,

-- 
Greg




From adalke at mindspring.com  Mon Apr 14 15:09:11 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Mon, 14 Apr 2003 13:09:11 -0600
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it> <yfshe92dmpt.fsf@black132.ex.ac.uk> <a8vma.11117$LB6.312431@news1.tin.it>
Message-ID: <b7f0pb$d2g$1@slb9.atl.mindspring.net>

Martelli-bot:
> Still, it DOES mention, as the
> one and only claimed advantage, that, without __slots__, "using a
> separate dictionary to hold a single instance variable doubles the
> memory used", and then proceeds to list the resulting limitations
> as "tidbits and warnings".  Any different reading is clearly wishful
> thinking, and the design intent seems pretty clear here.

I was recently looking at a Python package ported from Java.
It was the first package I've reviewed which used __slots__.
My impression is that __slots__ were used as a rough approximation
to declaring instance attributes, and not because of an explicit desire
to save space or improve lookup performance.  (That desire may
be implicit, but given the early state of the code it would be
premature to act on it.)

Given the style of the code, it appears that the author is not an
experienced programmer (eg, used "if flag is True" rather than
"if flag"), which I believe back my belief that people from statically
type languages look for a way to declare attributes, find __slots__,
and use it, not really knowing the appropriateness of this approach.

This is different than the design intent.  Should the documentation
be updated to caution against using __slots__ except for cases
which fit better the original intent? ....

> > 1. what's the point of not automatically including a slot for
> > '__weakref__'
> >    (hardly the unreasonable memory consumption such a slot would
impose?)?
>
> Indeed definitely that -- increasing memory consumption by e.g. 50%
> for an object with two "true" slots, just in case somebody in the
> future MAY want to weakreference it, would be a weird choice.  If
> you need your instances to be weakly referenceable include the slot
> explicitly and pay the attendand memory price.

I don't know __slots__ well enough.  Does you mean that if I want
weak references for these I can simply add that slot, as in

  __slots__ = [..., "__weakref__"]
?

> Sure.  So what?  A trickily coded class could always throw introspection
> based on instance __dict__ off -- __getattr__ could and often did "add"
> attributes not otherwise visible.

Though it should have also implemented __methods__ and/or __members__,
to return the list of otherwise invisible fields.

Very few __getattr__ implementation did this.  It's documented, but hard
to tell when/why it's useful.

Regarding Alexander Schmolck's post upon which Alex Martelli is
commenting -- Peering at the __dict__ is also problematical because
it didn't find class attributes.

>>> class Spam:
...  spam = 1
...
>>> dir(Spam())
['__doc__', '__module__', 'spam']
>>> Spam().__dict__
{}
>>>

                    Andrew
                    dalke at dalkescientific.com





From waldbie at attglobal.net  Sat Apr  5 16:08:24 2003
From: waldbie at attglobal.net (Carl Waldbieser)
Date: Sat, 5 Apr 2003 13:08:24 -0800
Subject: Q: The roles of Python
Message-ID: <5LEja.3680$945.10664@tor-nn1.netcom.ca>

It seems obvious to me how Python can be extremely useful on the server side
of a distributed application, but I was wondering if it has been widely used
on the client side of such applications or in desktop applications in
general.  Are there some good articles on the issues associated with using
Python in this way?  If you wrote a desktop application in Python, are there
any examples of how you would distribute Python with your application?  How
transparent would (or could?) the process be to the end user?

I think Python offers a lot of advantages over the other two big choices in
the Windows scripting world (JScript and VBScript), notably the fact that it
seems a lot easier to organize Python scripts into discrete modules.  It has
been my experience that in big applications, the other scripting engines I
have used tend to have problems with colliding names, and information tends
to be passed between scripts using global variables.  The net result is that
they tend to get very messy and hard to debug very quickly.

I think that Python scripts could be a lot easier to manage than other
scripting enviornments, but I have not heard much about Python being used
much behind the scenes in desktop scenarios.

Also, I would be very interested in hearing about client-side and desktop
applications where the application was written primarily in Python rather
than simply being the scripting engine.

Thanks for any insights and information in advance!

Carl Waldbieser
waldbie at attglobal.net






From kyle at lakeofburningfire.org  Tue Apr 15 19:51:07 2003
From: kyle at lakeofburningfire.org (Kyle Yancey)
Date: Tue, 15 Apr 2003 23:51:07 GMT
Subject: Problems with pso
Message-ID: <pv5p9v0rlsbj6uomjndku6as5vjvqpeoko@4ax.com>

Has anyone here been sucessful in installing pso with mod_python. I'm
working on a website, and I decided on pso because of the clean
template syntax and clear seperation of interface and code.  It's the
only templating system I've seen that I really like.  Unfortunately,
it seems to have a few bugs.  I installed it on Windows XP with
mod_python 3.0.3, python 2.2.1 and apache 2.0.43.  I've noticed
several bugs. All of them are already on the buglist at sourceforge,
so I see no point in reiterating them.  My initial reaction was that
the kinks really needed to be ironed out.  I can't believe the system
is unworkable, since I've heard so many remarks to the contrary.  What
I'm fishing for is some of your own practical experience with pso.

What did you add to your conf file to get everything working?
Are there any gotchas that you could point out now?
The documentation for the individual modules seems to be broken links
right now.  Would anyone know where an offline copy is right now?

I'm not looking for you to install this for me, but any practical
advice would be appreciated.



From blunck at gst.com  Tue Apr 29 11:54:33 2003
From: blunck at gst.com (Christopher Blunck)
Date: 29 Apr 2003 08:54:33 -0700
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com> <mailman.1051628445.12681.python-list@python.org>
Message-ID: <1147e466.0304290754.3dd43fcb@posting.google.com>

Skip Montanaro <skip at pobox.com> wrote in message news:<mailman.1051628445.12681.python-list at python.org>...
>     -c> Any insight you can offer would be appreciated.
> 
> Lead by example.  Create a test subdirectory in your source code repository
> and start writing test cases.  Use those test cases to demonstrate that your
> code functions as required.  

Funny that you mention this because it's exactly what I've done.  I've
written a good number of tests for my deliverables as well as the bugs
I find and fix.  But what I'm really trying to do now is to get others
to do the same.  That's the real trick.  :)

I suppose if I knew the answer to that question I'd be at a higher
level than what I am now.  :)


-c



From kp at kyborg.dk  Tue Apr  8 02:11:02 2003
From: kp at kyborg.dk (Kim Petersen)
Date: Tue, 08 Apr 2003 08:11:02 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk>
Message-ID: <3E9267F6.9020605@kyborg.dk>

Alexander Schmolck wrote:
> Kim Petersen <kp at kyborg.dk> writes:
 >> (snip)
>>function in elisp - do a shallow copy of local variable space (dictionary) and
>>send it as the local vars - you can safely throw away the dict after function
>>returns).
> 
> Ahem, there are several thousand variables kicking about in global in emacs,
> you want to copy them all for each function call?

look above - *Local* variable space ... globals stay globals.

> 
> How do your propose to handle the following efficiently:
> 
> (defun foo()
>   (message "when foo is called folding is: %s" case-fold-search))
> (defun bar()
>   (message "when bar is called folding is: %s" case-fold-search)
>   (let ((case-fold-search t))
>     (foo))
>   (message "now folding is: %s again" case-fold-search))
> (bar)

let me propose the following:

    * in python a function (lispname) is called as follows:
        def lispfun(arg0,arg1,...,rest,locals,buffer):

        - arg0,arg1 is filled with the first n arguments from cons,
          rest is filled with the rest of the cons:

          def lispadd(rest,locals,buffer): pass
          def cons(car,cdr,locals,buffer): pass
          def car(cons,locals,buffer): pass
          def while(expr,body,locals,buffer): pass


        - locals is the local variable space
        - buffer is the buffer variable space
        - (globals could be defined - but no need)

    * now (let ...) should be easy to handle... in either of two ways.
         1. make dictionary copy of locals
         2. push and pop old definitions from a variable stack.

> 
> when bar is called folding is: nil
> when foo is called folding is: t
> now folding is: nil again
> 
> alex



-- 
Med Venlig Hilsen / Regards

Kim Petersen - Kyborg A/S
IT - Innovationshuset
Havneparken 2
7100 Vejle
Tlf. +4576408183 || Fax. +4576408188




From catharsis at nospam.com  Wed Apr 30 13:07:30 2003
From: catharsis at nospam.com (Joel Jensen)
Date: Wed, 30 Apr 2003 17:07:30 GMT
Subject: Python / Zope Flash Remoting
Message-ID: <MPG.1919beb84d09f671989680@netnews.attbi.com>

I am trying to convert some perl/php code to python for use in zope.
I have never done anything like this. 

If anyone would like to help, please post a response.


Flash Remoting is a way to share datastructures between Cold Fusion and 
flash,make a dataset, varible, or array in cold fusion/java, retreve it 
as a map,variable,array in flash via flashs' OO language actionscript.
http://www.macromedia.com/software/flashremoting/


This protocol is referred to by macromedia as AMF
These guys:
http://sourceforge.net/projects/amfphp
http://www.simonf.com/flap/

have hacked the protocol and written scripts to send datastructures from 
perl and php.
Make an array in PHP, call it from flash..

It would be nice if there was a version of this code available in 
python.

Zope is a Python based Object Oriented web application platform.
www.zope.org 
If flash could share datasets with ZOPE, a developer could make a Zope 
site to contain the data structure, searches, and navigation structure, 
and flash designers could access these datastructures to easily build 
dynamic flash sites.

Zope and flash already support xml-rpc, however AMF is far better.
http://www.zope.org/Members/teyc/howxmlrpc

thanks

/joel



From news at snakefarm.org  Wed Apr 16 08:53:28 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 16 Apr 2003 12:53:28 GMT
Subject: Python 2.3a2: sockettimeout() buggy?
Message-ID: <slrnb9qki8.7p2.news@cipw08.physik.uni-bonn.de>

Hi there,

I just tried out the new socket timeouts in Python 2.3a2. Seems that the
timeout does not work if you make a file from the socket. Here's what I
tried:

import socket
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
s.bind(('', 12345))
s.listen(socket.SOMAXCONN)
(conn, addr) = s.accept()
conn.settimeout(10)


Now, when I say 'conn.recv(4096)' it raises an exception after 10 seconds
if there's no data to read as expected. But when I say:

f = conn.makefile('w+')
f.readline()

then the readline() call will raise that exception immediately. Is this a bug
or feature? Or did I miss something?

cg.



From mis6 at pitt.edu  Mon Apr 21 11:50:00 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 21 Apr 2003 08:50:00 -0700
Subject: Is this a super bug?
References: <mailman.1050817233.1530.python-list@python.org>
Message-ID: <2259b0e2.0304210750.5eaf5df0@posting.google.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1050817233.1530.python-list at python.org>...
> > From: Bjorn Pettersen 
> 
> Hate to reply to myself, but...
> 
> If this isn't the   getattribute   bug I read about, it needs to be
> documented that super only supports  direct  access to attributes, but
> not properties (although the Python version does), of a class in the
>   mro  , but will not access any attribute the language defines as
> "hooks" for various functionality indirectly... (i.e. all the   xx  
> methods -- I could understand   getattribute   would be a problem, but
> even   getattr   should have been possible to implement).
> 
> Sort of a special purpose proxy for the superclass... not what I
> expected from the name. And there doesn't seem there is a way to
> implement it in Python either. If I understand correctly, these
> attributes are looked for directly in the instance's class, skipping
> lookup in the instance itself, which is why the super instance's
>   getattr   or   getattribute   aren't called. Question: are   getattr  
> and   getattribute   ever called on the metaclass? Why not/why not in
> this case?
> 
> -- bjorn
> 

I think you got confused, here. Nothing related to metaclasses.
You should give a look to the attributes of super(ff,f):

>>> super(ff,f)
['__class__', '__delattr__', '__doc__', '__get__', '__getattribute__',
'__hash__', '__init__', '__new__', '__reduce__', '__repr__',
'__self__', '__setattr__', '__str__', '__thisclass__']

You see that super(ff,f) has no string attributes and no __getattr__.
It has a non-trivial __getattribute__, however, a __get__ attribute 
(thus it is a descriptor) and two interesting attributes __self__ and
__thisclass__ with clear meaning.

The output of 'dir' in my implementation of Super is more misleading, 
since dir invokes __getattribute__ and retrieves the attribute of the
str class:

>>> Super(ff,f)
['__add__', '__class__', '__contains__', '__delattr__', '__dict__',
'__doc__', '__eq__', '__ge__', '__getattribute__', '__getitem__',
'__getslice__', '__gt__', '__hash__', '__init__', '__le__', '__len__',
'__lt__', '__mod__', '__module__', '__mul__', '__ne__', '__new__',
'__reduce__', '__repr__', '__rmod__', '__rmul__', '__setattr__',
'__str__', '__weakref__', 'capitalize', 'center', 'count', 'decode',
'encode', 'endswith', 'expandtabs', 'find', 'foo', 'index', 'isalnum',
'isalpha', 'isdigit', 'islower', 'isspace', 'istitle', 'isupper',
'join', 'ljust', 'lower', 'lstrip', 'replace', 'rfind', 'rindex',
'rjust', 'rstrip', 'split', 'splitlines', 'startswith', 'strip',
'swapcase', 'title', 'translate', 'upper', 'zfill']

This is confusing, since actually my Super does not define all those 
methods, it is only a proxy to them, like the standard one. The 
dict of Super gives you the truth:

>>> for k,v in vars(type(Super(ff,f))).items(): print k,':',v #my
Super
__dict__ : <attribute '__dict__' of 'Super' objects>
__getattribute__ : <function __getattribute__ at 0x4032b764>
__weakref__ : <attribute '__weakref__' of 'Super' objects>
__doc__ : Invoked as Super(cls,obj).meth returns the supermethod of
    cls with respect to the MRO of obj. If obj is a subclass of cls,
    it returns the unbound supermethod of obj; otherwise, if obj is an
    instance of some subclass of cls, it returns the obj-bound method.
__init__ : <function __init__ at 0x4032b72c>

Somebody more knowledgeable than me should explain how the built-in
super
can override __getattribute__ without effects on the output of 'dir' !

It is also possible my understanding of how super is working is
incorrect;
I had much more trouble in understanding super the metaclasses ;)

Cheers,


                                                      Michele



From bdesth at removethis.free.fr  Thu Apr 17 10:02:17 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Thu, 17 Apr 2003 14:02:17 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk>
Message-ID: <3E9EB3E9.1040208@removethis.free.fr>

Alexander Schmolck wrote:
> laotseu <bdesth at removethis.free.fr> writes:

(snip : someone here posted about his fear that the few functionnal 
features in Python could be deprecated one day)

> 
>>Paul Foley wrote:
>>
>>>On Tue, 15 Apr 2003 23:49:37 +0000, laotseu  wrote:
>>>
>>>
>>>>Even for OO programmers, functionnal features in Python are IMHO a
>>>>great plus, and BTW functionnal and OO paradigm does not have to
>>>>conflict (that would be functionnal vs imperative and object vs
>>>>procedural). CLOS is one of the great system objects out here, and
>>>>it's been implemented on top of a functionnal language,
>>>
>                                       ^^^^^^^^^^^^^^^^^^^^
> 
> You go and say that in comp.lang.lisp :)
> 
> 'as

Ok, cross-posted and fu2 positionned.

Lispers friends, let's have you're opinion on that point : is Common 
Lisp a functionnal language ?

<Paul>
Your turn now : *you* go and say in comp.lang.lisp that Common Lisp *is 
not* a functionnal language !-)
</Paul>

Laotseu




From BPettersen at NAREX.com  Sun Apr 20 00:07:30 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sat, 19 Apr 2003 22:07:30 -0600
Subject: Is this a super bug?
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE00B@admin56.narex.com>

> From: Terry Reedy [mailto:tjreedy at udel.edu] 
> 
> "Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message
>
> >>> class ff(str):
> ...    def foo(self):
[...]
> ...      print 'super(ff, self)[0]', super(ff, self)[0]
> ...
> >>> f = ff('asdf')
> >>> f.foo()
> super(ff, self)[0]
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "<stdin>", line 5, in foo
> TypeError: unsubscriptable object
> ----------
> I suspect that printing type(super(ff,self)) might answer your
> question.

I'm afraid not :-)  I allready did that, and from what I can tell, this
is the sequence that "should" happen:

1  super(ff, f)[0]  # f == self
2  superinst(type:<class ff>, obj:f)[0]
3  si.__getitem__(0)
4    si.__getattr__('__getitem__')
5    .. return str.__dict__['__getitem__'].__get__(f)
6    boundmethod(str.__getitem__, f)
7  bmeth(0)
8   'a'

well, the transition from 3 to 4 doesn't happen which surprised me a
little. What surprised me more was that I couldn't get __getattribute__
to be called either (neither in the class or its metaclass... I've only
got 2.3a2, and I read a note about this not working before a4. Is this
what is happening? If not, the only other reason I could see would be
from the refMan 5.3.2, "The primary must evaluate to an object of a
sequence or mapping type."  If that is the case, then I still think it's
a super bug, i.e. there is no reason super should assume that it's not
"going to be" a sequence, in fact the direct superclass of ff _is_ a
sequence. To fix it simply add:

  def __getitem__(self, v):
      m = self.__getattr__('__getitem__')
      return m(v)

to the definition of super. This will change the exception to
AttributeError if a superclass isn't a sequence. To keep the current
semantics, add the lookup to only instances that are sequence proxies
(example below, refactor between __getattr__ as desired...)

-- bjorn

class Super(object):
    def __init__(self, type, obj=None):
        self.__type__ = type
        self.__obj__ = obj

        if isinstance(self.__obj__, self.__type__):
            starttype = self.__obj__.__class__
        else:
            starttype = self.__obj__
        mro = iter(starttype.__mro__)
           
        for cls in mro:             
            if cls is self.__type__:
                break               

        for cls in mro:
            if '__getitem__' in cls.__dict__:
                def getitem(v):
                    m = self.__getattr__('__getitem__')
                    return m(v)
                self.__possibleGetitem = getitem
                break

    def __possibleGetitem(self, v):
        raise TypeError("unindexable object")

    def __getitem__(self, v):
        return self.__possibleGetitem(v)




From mertz at gnosis.cx  Sat Apr  5 13:30:14 2003
From: mertz at gnosis.cx (David Mertz)
Date: Sat, 05 Apr 2003 13:30:14 -0500
Subject: POP3 and selective download/deletion
References: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com>
 <pan.2003.04.05.16.22.31.832230.878@thomas-guettler.de>
 <luCja.11925$It5.5735@news2.central.cox.net>
Message-ID: <2Cyj+kKkX0PQ092yn@gnosis.cx>                       

"Steve Holden" <sholden at holdenweb.com> wrote previously:
|Unfortunately the POP3 protocol doesn't allow you download the headers of a
|message without downloading the whole thing.

Nonsense.  This is part of the script I use to filter SPAM based only
on headers (so I don't download the large virus/spam message bodies).  There is
a bit of cruft in that I display some progress information and log some
actions, but it shows a realistic program.

    def check_pop_box(host, user, passwd, log, blist, wlist, screen_setup=""):
        freerows = ROWS-6   # 6 rows used by decorations
        try:
            mbox = poplib.POP3(host)
            say(screen_setup)
        except socket.gaierror:
            return None
        mbox.user(user)
        mbox.pass_(passwd)

        spams = {}
        for i in range(1, mbox.stat()[0]+1):
            # messages use one-based indexing
            headerlines = mbox.top(i, 0)[1]    # No body lines
            subject, from_ = '<NONE>','<NONE>'
            whitelisted = 0
            header = []
            for ln in headerlines:
                header.append(ln)
                if ln.upper().startswith('SUBJECT: '): subject = ln[9:]
                elif ln.upper().startswith('FROM: '):  from_ = ln[6:]
                for fld, pat in blist:  # blacklisted
                    if match_header(ln, fld, pat):
                        spams[i] = spams.get(i,'')+ln.strip()+'\n'
                for fld, pat in wlist:  # whitelisted
                    if match_header(ln, fld, pat):
                        whitelisted = 1
            # check the header as a whole for spamminess (& log headers)
            header = "\n".join(header)
            spam_prob = is_spam(header)
            summary = summary_line(from_, subject, spam_prob, whitelisted)
            if not whitelisted and spam_prob > 0.98:
                spams[i] = spams.get(i,'')+'+'+summary[:73].strip()+'\n'
            if spams.has_key(i):
                open('new-spam.headers','a').write("%s\n" % header)
                spams[i] += summary
            display(summary)
            freerows -= 1

        say(e('30;1m'),' '*80,'\n',e('36;1m'))  # gray line, then cyan
        for i,fld in spams.items():
            lines = fld.split('\n')
            summary, criteria = lines[-1], lines[:-1]
            freerows -= len(criteria)
            try:
                print "SPAM:", '\n'.join(criteria)
                print >> log, time.strftime("[%y-%m-%d]"), summary
                mbox.dele(i)
            except poplib.error_proto:
                print "Bad response on: 'mbox.del(%d)'" % i
        mbox.quit()
        log.flush()

        say(e('30;1m'),' '*80,'\n')  # gray line, then gray
        for ln in open('spam.log').readlines()[-freerows:]:
            print ln[:80].strip()


--
    _/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: Postmodern Enterprises _/_/_/
   _/_/    ~~~~~~~~~~~~~~~~~~~~[mertz at gnosis.cx]~~~~~~~~~~~~~~~~~~~~~  _/_/
  _/_/  The opinions expressed here must be those of my employer...   _/_/
 _/_/_/_/_/_/_/_/_/_/ Surely you don't think that *I* believe them!  _/_/





From sholden at holdenweb.com  Tue Apr 15 19:38:02 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 15 Apr 2003 23:38:02 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
Message-ID: <uJ0na.105066$It5.96339@news2.central.cox.net>

"Greg Brondo" <greg at brondo.com> wrote ...
> Aahz wrote:
[...]
> > Knowing the difference between O(N) and O(N^2) is critical to writing
> > even the simplest programs.  OTOH, I do agree with you that focusing on
> > programming as a craft is more important to being a programmer than
> > learning CS.
>
> I understand why the code is bad (copying the string each time in
> memory as it grows).  However, I've never understood the notation's
> O(N) and O(N2)?
>
These notations are to specify the "order of magnitude" of the run time of
an algorithm. If an algorithm is O(N) and your code runs in 10 seconds on a
list of 100 elements then you might expect it to run in 30 seconds on a list
of 300 elements - in other words, the run time is proportional to N, the
input size.

If the algorithm were O(N^2) then the run time could be expected to increase
with the SQUARE of the input size, meaning that an input of 200 elements
would take 40 seconds and an input of 300 elements would take 90 seconds.

The O() notation is a way of describing algorithmic performance without
getting stuck in the real world: it's assumed that for large enough N the
order effects will dominate, leaving any "trivial" variations in timing lost
in the noise. Thus it's a useful way to predict *roughly* how a given
algorithm will scale.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From sjmachin at lexicon.net  Sat Apr 19 20:15:19 2003
From: sjmachin at lexicon.net (John Machin)
Date: Sun, 20 Apr 2003 00:15:19 GMT
Subject: When not to use an RE -- an example
References: <c76ff6fc.0304191459.2adf73a@posting.google.com> <yfs7k9qjbmx.fsf@black132.ex.ac.uk>
Message-ID: <3ea1e522.18284281@news.lexicon.net>

On 20 Apr 2003 00:44:22 +0100, Alexander Schmolck <a.schmolck at gmx.net>
wrote:

>sjmachin at lexicon.net (John Machin) writes:
>
>> I needed a check for strings consisting of repeated characters -- like
>> when users type "ZZZZZZZ" instead of "UNKNOWN" into a database field.
>> After implementing the obvious overlapping-substring comparison, I got
>> to thinking how this could be done with REs. The following resulted:
>> 
>> import re
>> repeats1 = re.compile(r"^(?:(.)(?=\1))+\1\Z", re.DOTALL).match
>> def repeats2(s):
>>    return len(s) > 1 and s[1:] == s[:-1]
>> for testvalue, expected in zip(
>>    ['','x','xx','xxx','xxxxxx','xy','xxy','xyy','\n\n\n','aaa\n'],
>>    [0,  0,  1,   1,    1,       0,   0,    0,    1,       0     ]):
>>    print repr(testvalue), not not repeats1(testvalue),
>> repeats2(testvalue), expected
>> 
>> Pergly/phugly, eh? Note the effort required to ensure the newline
>> cases worked.
>
>
>What's wrong with:
>
>>>> matchRepetition = re.compile(r'(.)\1+\Z', re.DOTALL).match
>>>> map(bool, map(matchRepetition,
>...  ['','x','xx','xxx','xxxxxx','xy','xxy','xyy','\n\n\n','aaa\n']))
>[0, 0, 1, 1, 1, 0, 0, 0, 1, 0]
>
>?

(1) Too simple. (2) I didn't think of it. (3) Probably runs a lot
faster, too.

I'll get back into my box now. :-)




From info at BizBulk.com  Wed Apr 23 18:51:38 2003
From: info at BizBulk.com (info at BizBulk.com)
Date: Wed, 23 Apr 2003 15:51:38 -0700
Subject: Python Professional Services...
Message-ID: <GYNNETADK5Ovah8sVew00003a0a@gynnetad.gynnnet.com>

We recently visited your web site and we saw that you are involved in
the MAILING SERVICES business.

We are putting together a business-to-business mailing list and would
like to place you on our list. There is no charge to you. Our
business-to-business mailing list is designed to provide businesses in
the same or similar category the ability to communicate with one
another. Our goal is to expand business communications within particular
business sectors.

If you would like to be included in our list and make yourself available
to businesses, worldwide, please select your title and hit the OPT IN
button below. 

If you would not like to be included in our list, simply ignore this
message and you will not be contacted further.

Thank your for your time.



Select your title from the list: Assistant Corp SecretaryAssistant
TreasurerAssistant VPAssociateAuditorBranch ManagerCashierChairmanChief
Financial OfficerChief Information OfficerChief Operating OfficerChief
Technology OfficerCo-FounderConsultantControllerCorporate
SecretaryCSODeveloperDirectorDivision ManagerDivision
PresidentEditorEngineerExecutive Vice PresidentFounderGeneral
CounselGroup Vice PresidentInvestment OfficerManagerManaging
DirectorMedia ContactMedical OfficerMember of Advisory BoardMember of
Board of DirectorsOtherOwnerPartnerPersonal
AssistantPresidentPrincipalPublisherSales RepSenior Vice
PresidentSupervisorTreasurerTrust OfficerVice ChairmanVice
PresidentVice-PrincipalWeb Master

  _____  

www.BizBulk.com <http://www.BizBulk.com>  - your online business
communications tool.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030423/09aa0fdf/attachment.html>

From jepler at unpythonic.net  Wed Apr 30 15:52:58 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Wed, 30 Apr 2003 14:52:58 -0500
Subject: Problem in scalling............
In-Reply-To: <20030430001533.3975.qmail@web41708.mail.yahoo.com>
References: <20030430001533.3975.qmail@web41708.mail.yahoo.com>
Message-ID: <20030430195257.GE22792@unpythonic.net>

canvas.scale actually changes the objects themselves, not the coordinate
system of the canvas.

example:
>>> c = Tkinter.Canvas()
>>> i = c.create_rectangle((0,0,100,100))
>>> c.coords(i)
[0.0, 0.0, 100.0, 100.0]
>>> c.scale(i, 0, 0, 2, 2)
>>> c.coords(i)
[0.0, 0.0, 200.0, 200.0]

Jeff




From rmunn at pobox.com  Mon Apr 21 17:40:22 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 21 Apr 2003 21:40:22 GMT
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr> <mailman.1050568530.28331.python-list@python.org>
Message-ID: <slrnba84ii.shc.rmunn@localhost.localdomain>

Andrew Bennetts <andrew-pythonlist at puzzling.org> wrote:
> On Thu, Apr 17, 2003 at 09:55:46AM +0200, Jean-Paul Roy wrote:
> [..]
>> Aside from that points, a personal question as a Schemer who reads the 
>> Python primer:
[snip one question]
>> - how do you ask if the list has only one element, with time 0(1) ? The
>>   Scheme/Lisp analog is (null? (cdr L))
> 
> len(L) == 1
> 

List length is cached with the list object, so this is indeed O(1).
Another way would have been:

    try:
        L[1]
    except IndexError:
        print "L has at most one element"

Note that this will not tell you whether the list has exactly one
element, so it answers a slightly different question. But it's always
useful to know of different ways of solving a problem.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From peter at engcorp.com  Sun Apr 20 10:15:31 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 20 Apr 2003 10:15:31 -0400
Subject: Creating a stand alone Python Service (win32)
References: <PA$jhlyBDHA.3088@lime.oak-park.us>
Message-ID: <3EA2AB83.B4A95A14@engcorp.com>

Michael Steuer wrote:
> 
> I tried to register the PythonService.exe on the client, but (after adding
> the DLLs it needs) I got the following message when running
> "PythonService.exe /register" (enclosed in *'s to separate it from the rest
> of this message):
> 
> *********************************
> Registering the Python Service Manager ...
> 'import site' failed; use -v for traceback
> *********************************

Are you certain that it didn't work anyway?  That message comes when
the python.exe (or your service?) is run in a context where it is
unable to find the file site.py.  It's just a warning though, not
a serious failure, unless something later depends on site.py
and its normal functioning.

(To see this in action, set PYTHONHOME somewhere incorrect, then
run python using the full path.)

If you want to make it go away, you might be able to include a
dummy site.py in your package, or perhaps somehow invoke the -S
option on startup, although to be honest I'm not sure how all
this interacts with py2exe so at this point I'm rambling....

-Peter



From bokr at oz.net  Thu Apr 17 17:53:25 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 21:53:25 GMT
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr>
Message-ID: <b7n7ol$153$0@216.39.172.122>

On Thu, 17 Apr 2003 09:55:46 +0200, Jean-Paul Roy <roy at unice.fr> wrote:
[...]
>- Pyhton is bad : recursion limit (I am astonished at that one)
I think it's just there to warn you, so you don't have to wait so long
or get potentially confusing OOM symptoms before you realize you've got
a lot of recursion. You can adjust the limit.

 >>> import sys
 >>> help(sys.setrecursionlimit)
 Help on built-in function setrecursionlimit:

 setrecursionlimit(...)
     setrecursionlimit(n)

     Set the maximum depth of the Python interpreter stack to n.  This
     limit prevents infinite recursion from causing an overflow of the C
     stack and crashing Python.  The highest possible limit is platform-
     dependent.

 >>> help(sys.getrecursionlimit)
 Help on built-in function getrecursionlimit:

 getrecursionlimit(...)
     getrecursionlimit()

     Return the current value of the recursion limit, the maximum depth
     of the Python interpreter stack.  This limit prevents infinite
     recursion from causing an overflow of the C stack and crashing Python.

 >>> sys.getrecursionlimit()
 1000

The latter is the default on the 2.2.2 windows version.

Regards,
Bengt Richter



From aleax at aleax.it  Sat Apr 12 11:26:27 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 15:26:27 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>
Message-ID: <DeWla.6508$T34.173122@news2.tin.it>

Mads Orbesen Troest wrote:

> Hi Pythoners;
> 
> If I create a class which I want to be able to print like this ...
> tst = myclass()
> exp = r"test: %(intepol)s";
> print exp % { 'intepol':tst }
> 
> ... it seems I have two choices, either of which appear to work. I can
> overload the __str__ or the __repr__ method.
> 
> Which is the most correct to use; or should I even overload both? The

I've found that "both" is usually the answer -- if you define __repr__
and not __str__, all stringification of instances of your class uses the
__repr__ you defined, so that may suffice if there's only one way in
which you want such instances to be stringified.

> documentation says __str__ is for the, quote, informal representation;
> whereas __repr__ is for the, quote, official representation. Furthermore,

Yes, normally -- HOWEVER, unfortunately, when what you ask for an
"informal representation" of is a list, tuple or dict containing
instances of your class, it's your class's __repr__ -- NOT its
__str__ -- that ends up being used...:

>>> class Z:
...   def __str__(self): return 'nice!'
...
>>> print Z()
nice!
>>> print [Z()]
[<__main__.Z instance at 0x81c9e1c>]
>>>

> __reptr__ should, quote, look like an expresseion. I'm not quite sure what
> is meant by these distinctions. Is it that __repr__ is used by pickle and
> the like, and - thus - should represent the entire object state? Given

Not used by pickle, but ideally it should indeed represent the entire
object state -- eval(repr(x))==x is the goal, though it's not always
feasible to get anywhere near it (while remaining somewhat redable to a
programmerexamining the repr-stringified object).

> that I simply want to determine what should happen when one tries to print
> an instance, do I assume correctly when I gather that __str__ would be the
> more appropriate of the two - or is there some other way I should achieve
> this?

If you only care about print myclass() and not about print [myclass()], then
yes, __str__ suffices.  If you find yourself often printing lists, tuples
and dicts containing instances of your class, then __repr__ too will often
end up being used, so you may want to define both (or just __repr__, since
if __str__ isn't defined then print myclass() will use myclass.__repr__).


Alex




From SBrunning at trisystems.co.uk  Tue Apr 29 04:15:16 2003
From: SBrunning at trisystems.co.uk (Simon Brunning)
Date: Tue, 29 Apr 2003 09:15:16 +0100
Subject: GUI testing with python  (SendKeys)
Message-ID: <31575A892FF6D1118F5800600846864D01200884@intrepid>

> From:	Dave Brueck [SMTP:dave at pythonapocrypha.com]
> I've been happy with AutoIt (http://www.hiddensoft.com/AutoIt/) as it
> simulates keypresses, mouse clicks, and a few other things. There is no
> Python binding but it does have an ActiveX interface you can use, so that
> works just as well, e.g.:
> 
> from win32com.client import Dispatch
> c = Dispatch('AutoItX.Control')
> c.WinActivate('SomeApp', '')
> c.LeftClick(400, 200)
 
You might find "Driving win32 GUIs with Python" interesting -
<http://www.brunningonline.net/simon/blog/archives/000652.html>.

Cheers,
Simon Brunning
TriSystems Ltd.
sbrunning at trisystems.co.uk




-----------------------------------------------------------------------
The information in this email is confidential and may be legally privileged.
It is intended solely for the addressee. Access to this email by anyone else
is unauthorised. If you are not the intended recipient, any disclosure,
copying, distribution, or any action taken or omitted to be taken in
reliance on it, is prohibited and may be unlawful. TriSystems Ltd. cannot
accept liability for statements made which are clearly the senders own.




From donn at u.washington.edu  Thu Apr 17 16:58:28 2003
From: donn at u.washington.edu (Donn Cave)
Date: 17 Apr 2003 20:58:28 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org>
Message-ID: <b7n4hk$2g1u$1@nntp6.u.washington.edu>

Quoth Steven Taschuk <staschuk at telusplanet.net>:
| Quoth Tim Peters:
|> [...] What to do with each
|> method is in large part art, and the teachable seem to do OK learning by
|> example.  [...]
|
| I wonder, then, whether the docs would benefit from being
| augmented with a few examples.
|
| Donn -- would that improve their clarity in your view?

No, but if _replaced_ with a few examples, maybe.  I mean, if you
accept my contention that the documentation misrepresents the
actual usage of these functions, it's hard to remedy that with
an example.  Unless maybe hostile examples, like a repr that
spews out pages of state, but I assume that would not be the idea.

But then it's hard to describe their actual usage without examples,
so in the end, examples would certainly help, once the issues with
the text have been resolved.

	Donn Cave, donn at u.washington.edu



From hs2l2ATyahoodotcom at yahoo.com  Thu Apr 17 06:34:08 2003
From: hs2l2ATyahoodotcom at yahoo.com (Haran Shivanan)
Date: 17 Apr 2003 10:34:08 GMT
Subject: inheriting variables
Message-ID: <b7lvuu$29aj9$1@ID-148859.news.dfncis.de>

This seems like a fairly straight forward thing but I'm not sure about how
to implement it.
I have a base class from which I'm deriving several new classes:
class A:
	mem_a=0
	mem_b=1
class B(A):pass

I want B to inherit mem_a and mem_b from A without my having to explicit do
something like:

self.mem_a = 0

in the constructor for class B.
Basically, when I want to add a membership variable to all classes that
inherit from A, and I want to do it without manually editing all the class
definitions\constructors.

Currently I'm doing:
class B(A):
	def __init(self):
		for k in A.__dict__.keys():
			self.__dict__[k] = A.__dict__[k]
	...

This is somewhat crude. Also it does wierd things with methods declared in
A. Is there some really obvious clean implementation I'm missing?

Thanks.
-- 
Happiness is not a fish you can catch



From rmunn at pobox.com  Mon Apr  7 16:39:13 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 07 Apr 2003 20:39:13 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain> <6j8u8v0hjh3n9dm1jjflu3a970ada6qho4@4ax.com>
Message-ID: <slrnb933p7.46e.rmunn@localhost.localdomain>

carroll at tjc.com <carroll at tjc.com> wrote:
> On Sat, 05 Apr 2003 18:11:50 GMT, Robin Munn <rmunn at pobox.com> wrote:
>>You may not be doing type-checking, in which case this advice doesn't
>>apply to you, but type-checking is the wrong way to go about that. Do
>>interface checking instead, e.g.:
>>
>>    # Is it a file-like object?
>>    try:
>>        output_file.write
>>        print "output_file is a file-like object"
> 
> Just curious; why is this preferred?  Serious question; I'm relatively
> new to Python, and still learning a lot.
> 
> I'm thinking it might be two different philosophies.  The one in the
> present nntplib will try to invoke write() in whatever object was
> passed to it, and if the caller invoked it with the wrong object type,
> he'll get an error.  The one above seems to change the spec to support
> any object that has a method named "write."

The two approaches you describe in the paragraph above are really the
same approach, in a *slightly* different disguise. When you say "if the
caller invoked it with the wrong object type, he'll get an error," I'm
guessing that your idea of the "wrong object type" is an object that's
not a subclass of "file". But in Python, the "wrong object type" would
be... (drumroll)... Any object type that didn't have a method named
"write"!

The key concept to understand is this: in Python, types usually don't
matter. What matters is interfaces. If you have a function that expects
to be passed a file object, for instance, someone could pass it a
StringIO object (http://www.python.org/doc/2.2/lib/module-StringIO.html)
instead. And it would just work, without your function needing to know
anything about the StringIO class -- because StringIO implements the
write(), read(), readline(), etc. methods. That's why in many places in
the documentation you'll see "Parameter foo is a file-like object..."

In statically-typed languages like Java, you would have to jump through
some hoops to get this kind of functionality. For example, you might
have an Interface called IFileLike. Your functions would expect a
parameter that satisfied IFileLike, and your classes like StringIO would
implement IFileLike. In Python, though, all you have to do is try to use
the write() method of the object you've been passed. That's it. The Java
paradigm, with formally defined interfaces, could be called Look Before
You Leap (LBYL), while the Python paradigm could be called Easier to Ask
Forgiveness than Permission (EAFP).

For more reading on this subject, look at the following recipe in the
Python Cookbook online:

    http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/52291

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From Scott.Daniels at Acm.Org  Thu Apr  3 22:40:49 2003
From: Scott.Daniels at Acm.Org (Scott David Daniels)
Date: Thu, 03 Apr 2003 19:40:49 -0800
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <slrnb8orsl.174.do-not-spam-ben.hutchings@tin.bwsint.com>
References: <mailman.1049296805.12278.python-list@python.org> <b6fvj9$518n3$1@ID-169208.news.dfncis.de> <slrnb8orsl.174.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <3e8cfec2$1@nntp0.pdx.net>

Ben Hutchings wrote:
...
> The problem of confusion with existing identifiers can be dealt
> with once Python supports Unicode source files, when we can use
> this simple expression:
> 
>    \u216F{0,3}((\u216E?\u216D{0,3}|\u216D[\u216E\u216F])(\u216C?\u2169{0,3}|\u2169[\u216C\u216D])(\u2164?\u2160{0,3}|\u2160[\u2164\u2169])|\u216E?\u216D{0,4}\u216C?\u2169{0,4}\u2164?(\u2160{0,4})
> 
> Of course we would also need to deal with overbars for 5000,
> 10000, and so on.

Now to make this a lot more readable:
u"""
\N{ROMAN NUMERAL ONE THOUSAND}{0,3}
  ((\N{ROMAN NUMERAL FIVE HUNDRED}? \N{ROMAN NUMERAL ONE HUNDRED}{0,3}
   | \N{ROMAN NUMERAL ONE HUNDRED}
     [ \N{ROMAN NUMERAL FIVE HUNDRED} \N{ROMAN NUMERAL ONE THOUSAND} ]
   )
   ( \N{ROMAN NUMERAL FIFTY}? \N{ROMAN NUMERAL TEN}{0,3}
   | \N{ROMAN NUMERAL TEN}
     [ \N{ROMAN NUMERAL FIFTY} \N{ROMAN NUMERAL ONE HUNDRED} ]
   )
   ( \N{ROMAN NUMERAL FIVE}? \N{ROMAN NUMERAL ONE}{0,3}
   | \N{ROMAN NUMERAL ONE}
     [ \N{ROMAN NUMERAL FIVE} \N{ROMAN NUMERAL TEN} ]
   )
  |(\N{ROMAN NUMERAL FIVE HUNDRED}?
    \N{ROMAN NUMERAL ONE HUNDRED}{0,4}
    \N{ROMAN NUMERAL FIFTY}?
    \N{ROMAN NUMERAL TEN}{0,4}
    \N{ROMAN NUMERAL FIVE}?
    \N{ROMAN NUMERAL ONE}{0,4}
   ))
"""

or (possibly better):
(u'%(M)s{0,3}'
  u'((%(D)s?%(C)s{0,3}|%(C)s[%(D)s%(M)s])(%(L)s?%(X)s{0,3}'
  u'|%(X)s[%(L)s%(C)s])(%(V)s?%(I)s{0,3}|%(I)s[%(V)s%(X)s]))'
  u'|(%(D)s?%(C)s{0,4}%(L)s?%(X)s{0,4}%(V)s?%(I)s{0,4})') % {
     u'M': u'\N{ROMAN NUMERAL ONE THOUSAND}',
     u'D': u'\N{ROMAN NUMERAL FIVE HUNDRED}',
     u'C': u'\N{ROMAN NUMERAL ONE HUNDRED}',
     u'L': u'\N{ROMAN NUMERAL FIFTY}',
     u'X': u'\N{ROMAN NUMERAL TEN}',
     u'V': u'\N{ROMAN NUMERAL FIVE}',
     u'I': u'\N{ROMAN NUMERAL ONE}'}

Makes perl look better, doesn't it?

Anyhow, I'm suggesting that constants like
u'\N{ROMAN NUMERAL TEN}' beat u'\2169' for readability.  The
cost is relatively low as well: the lookup for the name only
happens when the .pyc (or .pyo) is generated.

-Scott David Daniels
Scott.Daniels at Acm.Org




From aleax at aleax.it  Sat Apr 26 06:55:02 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 10:55:02 GMT
Subject: MD5 to long?
References: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de> <b8dnlo$8sco1$1@ID-126101.news.dfncis.de>
Message-ID: <aAtqa.6585$3M4.185018@news1.tin.it>

Axel Grune wrote:

> Shoot me! eval works well;-)
> 
>  >>> longnumber = eval( '0x' + 'fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8'
>  >>> )

Sure, but calling long works MUCH better...:

[alex at lancelot Lib]$ python timeit.py 
-s'x="fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8"' 'ln=eval("0x"+x)'
10000 loops, best of 3: 31.4 usec per loop

[alex at lancelot Lib]$ python timeit.py 
-s'x="fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8"' 'ln=long(x,16)'
100000 loops, best of 3: 2.13 usec per loop
[alex at lancelot Lib]$

A speedup of 15 times as a result of using simpler and more direct
code seems to me to be worth pursuing.


Alex




From Franck.BUI-HUU at gemplus.com  Thu Apr 24 07:49:39 2003
From: Franck.BUI-HUU at gemplus.com (Franck Bui-Huu)
Date: Thu, 24 Apr 2003 13:49:39 +0200
Subject: eval function
Message-ID: <3EA7CF53.1080100@gemplus.com>

Hy,

I would like to use eval function with my own dicts.
For instance :

class my_dict(dict):
    def __getitem__(self, key):
        print 'using my_dict'
        dict.__getitem__(self, key)

 >>> dict = my_dict()
 >>> dict['x'] = 1
 >>> eval('1+x', dict)
2
 >>>

but eval seems to not use __getitem__ method of my_dict object because
"using my_dict" is not printed.
How can I print "using my_dict" when calling eval function ?

Thanks





From mertz at gnosis.cx  Mon Apr  7 21:56:08 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Mon, 07 Apr 2003 21:56:08 -0400
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com>
 <mailman.1049389032.4904.python-list@python.org>
 <u8bka.651$001.591065@monger.newsread.com>
 <mailman.1049722991.18682.python-list@python.org>
 <v938h26mdl9s8f@corp.supernews.com> <3E921B51.5030401@hotmail.com>
Message-ID: <4wik+kKkXMuc092yn@gnosis.cx>         

msw <markscottwright at hotmail.com> wrote previously:
|Maybe, but since I'm the only one here who's read the requirements
|document, I submit that I'm the only one qualified to suggest a software
|design, no matter how limited my intelligence.

I'm sure I speak for all the other posters in saying that it is not that
we think you are wrong in your ability to read the requirements
document--it is the requirements document that is wrong in the
presuppositions it makes.

I've read enough of those things in my time to know that a requirments
document never says what it means, nor ever means what it says.  You
might not get fired for writing a program that meets spec, but you sure
as heck won't get promoted for doing so.

In any case, a number posters have provided you with way more than the
amount of information needed for you to develop "protected code."
You're right that no one is volunteering to do your (paid?) job for
you... but it's not exactly *difficult* to modify the behavior of
__import__() a little bit.

Yours, Lulu...

--
 mertz@  _/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: \_\_\_\_    n o
gnosis  _/_/             Postmodern Enterprises            \_\_
.cx    _/_/                                                 \_\_  d o
      _/_/_/ IN A WORLD W/O WALLS, THERE WOULD BE NO GATES \_\_\_ z e





From jason at tishler.net  Fri Apr  4 09:53:35 2003
From: jason at tishler.net (Jason Tishler)
Date: Fri, 04 Apr 2003 09:53:35 -0500
Subject: Cygwin Python regression testers wanted
In-Reply-To: <slrnb8qjnc.i81.rmunn@localhost.localdomain>
References: <mailman.1049385789.4709.python-list@python.org>
 <slrnb8qjnc.i81.rmunn@localhost.localdomain>
Message-ID: <20030404145334.GD1336@tishler.net>

Robin,

On Fri, Apr 04, 2003 at 03:21:12PM +0000, Robin Munn wrote:
> Jason Tishler <jason at tishler.net> wrote:
> > If you are willing to help, then perform the following steps:
> 
> I'd be willing to help with 95 OSR2, 98, and 98SE -- but it will have
> to wait until this weekend, I'm afraid.

No problem -- I'm willing to wait.

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From a.schmolck at gmx.net  Thu Apr 24 20:19:38 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 25 Apr 2003 01:19:38 +0100
Subject: string to variable name
References: <b89qi8$2oij$1@agate.berkeley.edu> <yfsist3wij7.fsf@black132.ex.ac.uk>
Message-ID: <yfsel3rwhr9.fsf@black132.ex.ac.uk>

Alexander Schmolck <a.schmolck at gmx.net> writes:

> >   FILE = open("<mytextfile");
> >   while (<FILE>){
> >     /^(\w+)\s+(\w+)\s+(\w+)$/;
> >     inst.$2 = $1;
> >   }
> > }
> 
> 
> > Is there any way to (easily) achieve what I want to do in python?
> 
> Sure: you can do it even shorter and almost as unreadably with (UNTESTED):
> 
> for line in file("mytextfile"):
>     inst.__dict__.update(dict([line.split()[:2]]))

make that ``split()[1::-1]`` (you have key second, value first, right?)

(this will only work in python2.3, though -- but then setattr is much better
here anyway)

'as



From vf at interna.ru  Wed Apr 30 10:31:12 2003
From: vf at interna.ru (vadim)
Date: 30 Apr 2003 07:31:12 -0700
Subject: ActiveScripting: import email => no module named _codecs
Message-ID: <db81c492.0304300631.28ec1f36@posting.google.com>

I am doing rather simple thing after installing win32com 
extensions (IE):

<script language="Python" >
import email.Parser
</script>

and getting following error:
...
File "<ScriptBlock> ", line 2, in ?
  import email.Parser
SystemError: Failed to load the builtin codecs: No module named _codecs

Does anyone could help what it means?



From andrew at acooke.org  Mon Apr 28 12:43:42 2003
From: andrew at acooke.org (andrew cooke)
Date: Mon, 28 Apr 2003 12:43:42 -0400 (CLT)
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
In-Reply-To: <yu99sms2k35f.fsf@europa.research.att.com>
References: <mailman.1051532982.25352.python-list@python.org> 
     <UXara.22428$K35.745468@news2.tin.it> 
     <yu99sms2k35f.fsf@europa.research.att.com>
Message-ID: <1230.10.2.0.2.1051548222.squirrel@127.0.0.1>

Andrew Koenig said:
[...]
> So, for example, to express the integers 4, 5, 6, 7, you can write [4,
> 7], (3, 7], [4, 8), or (3, 8).  Here, the [ ] mean that the range
> *includes* the number next to the bracket, and the ( ) mean that the
> range *excludes* the number next to the parenthesis.  The question,
> then, is which kind of range to prefer.
[...]
> From the foregoing argument, I conclude that it's easiest
> overall to represent *all* ranges asymmetrically: the first
> element of the range and the first element beyond the range.

using [) consistently also gives you nice things like correct addition of
intervals:

[1,4)+[4,7) = [1,7)

isn't there some (fairly old) popular math book by a famousish person
(knuth?  hofstadter?) that develops maths (at least arithmetic) based on
defining sets of points from the "number line"?  i wonder what that says
about this?  (i should admit that i picked up the above from one of
celko's db books).

andrew

-- 
http://www.acooke.org/andrew




From db3l at fitlinxx.com  Mon Apr  7 13:24:14 2003
From: db3l at fitlinxx.com (David Bolen)
Date: 07 Apr 2003 12:24:14 -0500
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com>
Message-ID: <uk7e6i8cx.fsf@fitlinxx.com>

msw <markscottwright at hotmail.com> writes:

> I think it ought to be relatively easy to distribute a hacked version
> of the python interpreter with your application that substitutes the
> standard "import" code with code that first looks for .pys files (or,
> what the hell, .tcl - let's not make this any easier) that are
> des-encrypted .py files.  Compilation to .pyo would have to be
> disabled, which would increase start-up time.  The hacked interpreter
> (it's pretty easy to embed the python interpreter in your
> application), could also check some sort of hardware token before
> starting.  I have no idea if this would violate python's license.

No real need to hack the Python interpreter proper for this if you
don't want.  Most packaging tools will have some sort of front end
code that they use to create the final exe that you ought to be able
to tune to your liking.  And you'd need to use a packaging tool or
else someone could just download the standard Python interpreter and
avoid your modified version entirely.

For example, with Gordon McMillan's Installer, there is a run.exe or
runw.exe that is prepended to the resulting executable, and source is
included.  So you could tweak that loader to do whatever you might
like prior to handing control over to the interpreter.  Likewise, the
import handling code can be adjusted if you'd like to have it do more
strenuous checking or decoding of the packaged files.

I know that I'm considering following this path as a minor adjustment
to the tool to permit a single Python runtime environment to exist for
multiple packaged applications.

> But unfortunately, if anyone's done this, you'll never find out about
> it on c.l.p.  :-)

I don't see that anyone here would have any objections to sharing this
information if it had been done.  It may actually be the case that
nobody has needed to yet, or that they have found the existing tools
suitable - for example, just packaging the application up as an
executable may be considered a reasonable level of protection against
casual snooping.

--
-- David
-- 
/-----------------------------------------------------------------------\
 \               David Bolen            \   E-mail: db3l at fitlinxx.com  /
  |             FitLinxx, Inc.            \  Phone: (203) 708-5192    |
 /  860 Canal Street, Stamford, CT  06902   \  Fax: (203) 316-5150     \
\-----------------------------------------------------------------------/



From Simon.Wittber at perth.maptek.com.au  Thu Apr  3 22:42:23 2003
From: Simon.Wittber at perth.maptek.com.au (Simon Wittber (Maptek))
Date: Fri, 4 Apr 2003 11:42:23 +0800
Subject: mp3 or wave  ---> midi , in python , how?
Message-ID: <10F0E58C0018054484E329DC494C4D7F9490B4@mexper1>

Frank wrote:

>Hi, All:
>
> Who can tell me how to do this in python ,are there any modules
existing
> can help me?

Wave files (mp3s) are digitised sound waves. Midi files are sequences of
notes which are fed into a synthesizer, usually your soundcard, which 
reproduces the music. No doubt you already know this...

I have seen and used software (some time ago) which attempts to convert 
wave files to midi data, however it only produced one channel of midi
which
is essentially useless unless you are a composer trying to extract a 
sequence of notes from a small piece of music.

I think you will find it very difficult to find an application which
does 
this properly, let alone a python module.

Simonw.




From tjreedy at udel.edu  Mon Apr 28 02:09:15 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Mon, 28 Apr 2003 02:09:15 -0400
Subject: Forgetting "()" when calling methods
References: <mailman.1051497342.21537.python-list@python.org>
Message-ID: <cHadnX5vTdzoWDGjXTWcpw@comcast.com>

"Tim Peters" <tim.one at comcast.net> wrote in message
news:mailman.1051497342.21537.python-list at python.org...
> I'm pretty sure that if Python had never supported meaningless
(although
> consistent within a single run) comparisons, a compelling case for
adding
> them couldn't have been made.

Nor can I imagine a compelling case for exactly the current rules if
they were not whay they currently are.  Allowing 1j to compare to '1'
but not 1 is pretty exoteric.  The argument seem to be this: cmp(1j,
'1') is so meaningless that it is ok to give an answer while cmp(1j,1)
might be seen as meaningful (and lexicographically, it is) so therefor
it must be forbidden.

Terry





From max at alcyone.com  Sat Apr  5 05:11:39 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sat, 05 Apr 2003 02:11:39 -0800
Subject: In memory File
References: <b6m741$dqo$1@news-reader12.wanadoo.fr> <3E8EA197.3AE807EB@alcyone.com> <3E8EA9E4.60704@wanadoo.fr>
Message-ID: <3E8EABDB.9AA000A0@alcyone.com>

Salvatore wrote:

> I wanted to redirect sys.stdout to StringIO object but il doesn't seem
> to work

>>> import StringIO
>>> import sys
>>> sys.stdout = StringIO.StringIO()
>>> print "Told you"
>>> print >> sys.stderr, sys.stdout.getvalue()
Told you

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Did you ever love somebody / Did you ever really care
\__/ Cassandra Wilson
    Alcyone Systems / http://www.alcyone.com/
 Alcyone Systems, San Jose, California.



From riemer at lincvs.org  Mon Apr  7 02:41:01 2003
From: riemer at lincvs.org (Tilo Riemer)
Date: Mon, 7 Apr 2003 08:41:01 +0200
Subject: PYTHON EDITOR!!
In-Reply-To: <03040711314900.15230nanthan@siswa.utm.my>
References: <03040711314900.15230nanthan@siswa.utm.my>
Message-ID: <20030407064101.GA8292@ppprs1.phy.tu-dresden.de>

Hello Nanthan, 


> have been working in windows, but my development actually is on 
> Linux, so can u suggest me the best editor(i mean use friendly 
> and graphical,i dont want command line editor)for python?
> 
> is cooledit the right choice.

cooledit is nice... or SciTE or Emacs.


Best regards, Tilo




From peter at engcorp.com  Wed Apr  2 15:32:32 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 15:32:32 -0500
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com> <mailman.1049298845.7655.python-list@python.org> <3E8B4094.F61515B3@engcorp.com> <ZpHia.23$Sd.4@fed1read01>
Message-ID: <3E8B48E0.C786D192@engcorp.com>

Tim Hochberg wrote:
> 
> I believe what was discussed was making shadowing of builtins from
> outside of a module forbidden. So:
> 
>     range = my_custom_range_function
> 
> would be OK. While,
> 
>     some_random_module.range = my_custom_range_function
> 
> would not.

Hmm.... that is exactly what is needed to make this effective in 
unit testing.  One imports the module under test, then fakes the
open method in that other module:

import unittest
import module_under_test as mut

from mock_files import mock_open

class MyCase(unittest.TestCase):
    def setUp(self):
        mut.open = mock_open  # shadow builtin

    def tearDown(self):
        del mut.open  # remove shadow

    def test(self):
        mut.doSomething()


I'm not sure what other approach could be as simple, elegant, and 
effective and I worry about losing it.  Surely there are many other
people doing similar things who would be hurt by this.

-Peter



From mwh at python.net  Mon Apr  7 06:44:31 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 7 Apr 2003 10:44:31 GMT
Subject: Handling slices
References: <mailman.1049676910.30971.python-list@python.org>
Message-ID: <7h3ptnyio34.fsf@pc150.maths.bris.ac.uk>

Francois Pinard <pinard at iro.umontreal.ca> writes:

> Hi, Python friends.
> 
> I'm currently using Python 2.2.1.  How does one support slices when given
> to __getitem__?  Suppose I get a slice object.  How can I use it to get
> a slice of a local array?  In other words, how one does apply a slice?

In 2.2.1, by hand, I think.

def __getitem__(self, item):
    if isinstance(item, types.SliceType):
       if item.step != 1:
          raise ValueError
       if item.start is None:
          start = 0
       else:
          start = item.start
       if item.stop is None:
          stop = len(self.local_array)
       else:
          stop = item.stop
       return self.local_array[start:stop]
    ...

I presume you don't need to be told that this is all much nicer in
2.3...

Cheers,
M>

-- 
  On the other hand, the following areas are subject to boycott
  in reaction to the rampant impurity of design or execution, as
  determined after a period of study, in no particular order:
    ...                              http://www.naggum.no/profile.html



From a.schmolck at gmx.net  Mon Apr 28 00:54:25 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 28 Apr 2003 05:54:25 +0100
Subject: overloading print behavior
References: <vapb1f5mcqf0cd@corp.supernews.com>
Message-ID: <yfsbryrusqm.fsf@black132.ex.ac.uk>

Cere Davis <cere at u.washington.edu> writes:

> So I have a function like:
> 
> class C:
>     def __init__(self):
>        self.a="asdf"
>        self.b="123"
> c=C()
> st=["abc",c]
> 
> When I:
> 
> print st
> 
> I get:
> 
> ['abc', <__main__.C instance at 0x40288fac>]
> 
> Which is what I expect, but I would like it if print or perhaps the str()
> coercion that is happening behind the scenes would walk through my C instance
> 
> and resolve the name/value paris within the C class and print the values.
> Something like:
> 
> 
> print st
> 
>  > 'abc',<C instance>['a=asdf','b=123']

> Could someone enlighten me on this simple concept?

Sure:

class C:
    def __init__(self):
        self.a="asdf"
        self.b="123"
    def __repr__(self):
        return "<%s instance with a=%r, b=%r>" % (self.__class__,
                                                  self.a, self.b)

>>> C()
<__main__.C instance with a='asdf', b='123'>


The place to look these things up is the language reference (all "magical"
methods used to customize instance behavior are listed there).


'as



From sholden at holdenweb.com  Sat Apr  5 10:19:45 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sat, 05 Apr 2003 15:19:45 GMT
Subject: POP3 and selective download/deletion
References: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com> <pan.2003.04.05.16.22.31.832230.878@thomas-guettler.de>
Message-ID: <luCja.11925$It5.5735@news2.central.cox.net>

"Thomas Guettler" <pan-newsreader at thomas-guettler.de> wrote in message
news:pan.2003.04.05.16.22.31.832230.878 at thomas-guettler.de...
> On Sat, 05 Apr 2003 12:12:26 +0200, Dino Levy wrote:
>
> > Hello!
> > I'm making a mail program and I have a little problem: I don't know how
> > to achieve the selective downloads/deletion from POP3 servers.
>
> Maybe this recipe from the python cookbook does help you to get started:
>   http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/134945
>
> It prints all messages which are biger then N bytes and ask if it should
> be deleted.
>

Unfortunately the POP3 protocol doesn't allow you download the headers of a
message without downloading the whole thing. IMAP4 is much more capable, but
it's also a very much more complex protocol.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From aleax at aleax.it  Fri Apr 18 17:54:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 21:54:46 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net>
Message-ID: <Gu_na.34381$LB6.779929@news1.tin.it>

Tim Gahnstr?m /Bladerman wrote:

> 
> "Evan Simpson" <evan at 4-am.com> wrote in
>> Guy Middleton wrote:
>> > >>> reduce(lambda a, b: a+b, [y[1] for y in x])
>> >
>> > But I don't really like lambda, is there a better way to do this?
>>
>> What's wrong with:
>>
>> t = 0
>> for y in x:
>>      t += y[1]
>>
>> The extra variable?  The number of lines?
> 
> I am really curious about that to, I would most definitley say that this
> is the most pythonic way. It is simple and easily readabel by anyone and I
> am sure it is just as fast any of the other way.

Quite.  Performance is easy to measure, rather than guess at:

[alex at lancelot python2.3]$ python -O timeit.py -s"lst=[('a',1), ('b',2), 
('c',3)]*100" -s'import operator' 'x=reduce(lambda x,y: x+y[1], lst, 0)'
1000 loops, best of 3: 235 usec per loop

[alex at lancelot python2.3]$ python -O timeit.py -s"lst=[('a',1), ('b',2), 
('c',3)]*100" -s'import operator' 'x=reduce(operator.add,[y[1] for y in 
lst])'
1000 loops, best of 3: 277 usec per loop

[alex at lancelot python2.3]$ python -O timeit.py -s"lst=[('a',1), ('b',2), 
('c',3)]*100" -s'import operator' 'x=0' 'for y in lst: x+=y[1]'
10000 loops, best of 3: 103 usec per loop

Using lists 10 or 100 times as long seems to show O(N) behavior for
each case, as could be expected.  So, the simplest (and I would agree
thereby most Pythonic) approach is over twice as fast as the fancy ones
(measurements on 2.2.2 seem to show similar results, all approaches
being about 20% slower than in 2.3, uniformly).


Alex





From tjreedy at udel.edu  Thu Apr 10 11:29:41 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Thu, 10 Apr 2003 11:29:41 -0400
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <MZOcnSSTJ_cvEAijXTWcoA@comcast.com>

"Alex Martelli" <aleaxit at yahoo.com> wrote in message
news:b73ql8011te at enews4.newsguy.com...
> I _know_ there's a decent algorithm to solve the following problem,
> but I don't recall its name and thus can't rapidly google for the
> details... can somebody help?

Linear discriminant analysis.  I have done several over my career.

> The problem: I need to design a "decision criterion" to classify
> "observations".  For each observation I measure the values of a
> number N of features, x1, x2, ... xN; the desired design criterion
> is a set of weights w1, w2, ... wN such that for any observation I
> will then just compute a weighted sum
>   S = w1*x1 + w2*x2 + ... + wN*xN

+ constant term C

> and classify the observation as Black if S<=1, White if S>1.

For k groups, the procedure results in k classifier functions
(weighted sums + constant). To classify an unknown, calculate all k
sums and assign to group corresponding to highest.  For two groups,
you can subtract functions to get one function and then compare to 0.

>  To
> train my classifier I have a large corpus of observations already
> made and annotated with "ground-truth" data about whether each
> given observation should have been classified as B or W, and an
> error-cost value for each kind of classification (Ebw is the cost
> of erroneously classifying a feature as W when it should be B,
> Ewb is that of classifying it as B when it should be W).  So,
> what's the algorithm to estimate the weights given the corpus of
> observation and ground-truth data, and the error-costs?

The added constant(s) is/are the fudge factor(s).  They are usually
adjusted according to specified a priori probabilities for being in
each group.  I do not know how to go from error costs to adjusted a
prioris, but perhaps someone has worked this out and published it.

The basic matrix equations are available in several statistics books.
However, there is a practical problem of over-fitting.  Some of your
features may be highly correlated and redundant.  Others may be
useless for discrimination even when the effect other variables is
accounted for.  Wasting input information to estimate
redundant/useless coefficients (which thereby constitute noise) can
and likely will result in more classification errors, not less.   So,
for best performance in classifying unknowns, you want a minimal set
of jointly significant features.  This can be developed by a stepwise
procedure, which is *much* messier.  The Computation Methods appendix
of the BMDP Statistical Software Manual is one source of the gory
details.  If this is not a toy project and you actually care about the
results, you will probably do better with professionally written
statistics software that has this procedure.

Terry J. Reedy






From kennethg at pd.jaring.my  Sat Apr  5 23:47:52 2003
From: kennethg at pd.jaring.my (Kenneth Gomez)
Date: Sun, 06 Apr 2003 04:47:52 GMT
Subject: Python - Database and Email
References: <3e8f0614.12738582@news.jaring.my> <274u8vc961fr9cpaq3hsbfdo5uka5ibd5r@4ax.com>
Message-ID: <3e8fb139.5845025@news.jaring.my>

Thank you very much Rene.
Cheers,
Kenneth Gomez.



From mis6 at pitt.edu  Mon Apr 28 17:07:45 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 14:07:45 -0700
Subject: metaclasses, how you might actually use 'em
References: <mailman.1051458882.30400.python-list@python.org> <2259b0e2.0304280514.5ca68d50@posting.google.com> <mailman.1051549553.26987.python-list@python.org>
Message-ID: <2259b0e2.0304281307.7e9b3a89@posting.google.com>

Jack Diederich <jack at performancedrivers.com> wrote in message news:<mailman.1051549553.26987.python-list at python.org>...

> <snip> 
> I could never get the __new__ to do what I wanted.  I would always end up
> recursively calling __new__.  Doing everything in __init__ works, so I just
> kept with that.
> 
> -jackdied
> http://jackdied.com/code/

Yes, __init__ is less tricky than __new__. In order to avoid the
recursion you must do something like this:

class M(type):
    def __new__(meta,name,bases,dic):
        # do something
        return type.__new__(meta,name,bases,dic)

However, this is not cooperative. It is better to write

class M(type):
    def __new__(meta,name,bases,dic):
	# do something
	return super(M,meta).__new__(meta,name,bases,dic)

BTW, you should write your __init__ in this way:

class M(type):
    def __init__(cls,name,bases,dic):
	super(M,cls).__init__(name,bases,dic)
        # do something

This ensures safeness against multiple inheritance of metaclasses, which
soon or later you will be forced to use (to avoid metatype conflicts).


                      Michele



From theller at python.net  Tue Apr 29 13:39:19 2003
From: theller at python.net (Thomas Heller)
Date: 29 Apr 2003 19:39:19 +0200
Subject: ctypes 0.6.0 released
Message-ID: <fzo12ofs.fsf@python.net>

Overview

    'ctypes' is a Python package to create and manipulate C data types
    in Python, and to call functions in dynamic link libraries/shared
    dlls. It allows wrapping these libraries in pure Python.

    It works on Windows, Linux and MacOS X (the latter two require that
    your machine is supported by libffi).

News

    *ctypes version 0.6.0 has been released*, and the documentation has
    been updated.

    There have been some (also incompatible) changes, but I'm quite
    pleased with the result: ctypes is much more consistent than before.

COM

    A proof-of-concept COM framework has been implemented, and added to
    the package. It is in an early stage, although usable, and I would
    like experienced COM developers to look at it, and submit feedback.

Homepage

    http://starship.python.net/crew/theller/ctypes.html

Thomas





From jeremy at zope.com  Thu Apr 10 13:33:24 2003
From: jeremy at zope.com (Jeremy Hylton)
Date: 10 Apr 2003 13:33:24 -0400
Subject: ssl error "PRNG not seeded"
In-Reply-To: <61957B071FF421419E567A28A45C7FE59AF3AD@mailbox.nameconnector.com>
References: 	 <61957B071FF421419E567A28A45C7FE59AF3AD@mailbox.nameconnector.com>
Message-ID: <1049996004.20054.95.camel@slothrop.zope.com>

On Thu, 2003-04-10 at 11:13, Geoffrey Talvola wrote:
> sslerror: (1, 'error:24064064:random number generator:SSLEAY_RAND_BYTES:PRNG
> not seeded')

> Is this likely to be a problem in the Python SSL implementation or an
> OpenSSL problem?

The error quoted about is generated in OpenSSL.  The Python SSL code
just wraps the error message in a Python exception and propagates it.

Jeremy






From anton at vredegoor.doge.nl  Sat Apr 26 15:14:21 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Sat, 26 Apr 2003 21:14:21 +0200
Subject: MD5 to long?
References: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de>
Message-ID: <b8eluj$knb$1@news.hccnet.nl>

Axel Grune <NineOfSix at gmx.de> wrote:

>Hi, is there any function which can convert a MD5 or SHA1 sum
>(e.g. fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8) to a long integer?

MD5 or SHA do have digest() and also hexdigest(). I couldn't find sum.
Digest returns bytes and hexdigest returns a string of hexadecimal
digits. It seems *very* unlikely that the sequence you are giving is a
digest and others have already presented solutions based on the
hypothesis that it is a hexdigest. However, since it is not 100%
certain that it is not a digest, and because possibly a more flexible
approach can still be interesting even if it is a hexdigest, I would
like to present a class which just ranks a list. It is unlikely that a
MD5 (check?) sum will be used for integer arithmetic so it shouldn't
matter that the mapping is unconventional. The class is halfway on its
way to its final release so please test it extensively but don't use
it for production code.

Anton

class Rank:
    
    """ 
        ranks and unranks lists, it has to be initialized with a 
        reference list wherein each item occurs once
    """
    
    def __init__(self,items):
        #initialize with reference list
        self.items, self.base = items, len(items)

    def unrank(self,n):
        #create the list with rank n
        result, base, items = [], self.base, self.items
        while n >= 0:
            i = (n-base)/base
            result.append(items[n-base*(i+1)])
            n = i
        result.reverse()
        return result

    def rank(self,L):
        #compute the rank of list L
        index, base = self.items.index, self.base
        i = index(L[0])
        for x in L[1:]:  i =  index(x)+base*(i+1)
        return i

def test():
    str = """I am the giant Caraculiambro, lord of the island of
Malindrania, vanquished in single combat by the never sufficiently
extolled knight Don Quixote of La Mancha, who has commanded me to
present myself before your Grace, that your Highness dispose of me
at your pleasure"""

    import md5
    R = Rank([chr(i) for i in range(256)])
    dgst = md5.new(str).digest()
    idx = R.rank(dgst)
    assert  ''.join(R.unrank(idx))==dgst
    print "%s\n\n%s" %(str,idx)

if __name__=='__main__':
    test()




From jjl at pobox.com  Tue Apr 29 18:07:15 2003
From: jjl at pobox.com (John J. Lee)
Date: 29 Apr 2003 23:07:15 +0100
Subject: Finding unused code
References: <pan.2003.04.29.21.07.45.44830@effectif.com>
Message-ID: <86y91tarfw.fsf@pobox.com>

"Graham Ashton" <graham at effectif.com> writes:

> I'm involved in maintaining a rather large Python project at work, and am
[...]
> I'd like to write a script to analyse the source code and tell me which
> bits haven't been used, but I'm at a bit of a loss for where to start.
> 
> Any ideas? Or has somebody else already written such a tool?

Google for coverage python.


John



From nagylzs at freemail.hu  Fri Apr  4 13:58:01 2003
From: nagylzs at freemail.hu (=?ISO-8859-1?Q?Nagy_L=E1szl=F3_Zsolt?=)
Date: Fri, 04 Apr 2003 20:58:01 +0200
Subject: My Object pool isnt working
References: <LNBBLJKPBEHFEDALKOLCGEJDECAB.tim.one@comcast.net>
Message-ID: <3E8DD5B9.501@freemail.hu>


>>Hashtables in Java are thread safe, but I'm not sure if the built in
>>dict type in python is.
>>    
>>
>
>It is.
>  
>
Whew! I always thought that I must use locks to access any object in 
Python before I modify it from a thread.
Is it possible to append a list or a dictionary from a thread without 
locking it? What does thread-safe mean here?
For example, what would happen if I walk through a list in a for loop 
while another thread modifies the list?
Thank you in advance.

  Laci

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030404/ad60edbd/attachment.html>

From adalke at mindspring.com  Sat Apr 19 20:13:24 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Sat, 19 Apr 2003 18:13:24 -0600
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <b7q72i$nvh$1@slb6.atl.mindspring.net> <roy-416201.21303118042003@reader1.panix.com>
Message-ID: <b7sog8$pqj$1@slb6.atl.mindspring.net>

Roy Smith:
> I agree that all those things are good things, but by the time you've
> made that many changes to the language, what's left that it can still
> call itself BASIC?

Put it this way.  When I got QuickBasic, I had already programmed
on TI Basic, TI Extended Basic, GW-Basic, Atari Basic, Apple Basic,
TRS-80 Basic, and a few more.

My working definition of BASIC then is "could I sit down and
start writing programs in it with little help"  (eg, some syntax changes,
graphics calls, etc, were different).  And the answer is yes, I could
start being productive in QuickBasic right away.

Then take away line numbers.  (Whoo-hoo!  No need to 'renum'!)
Then replace "gosub" with a real function.
  ...

So yes, QB should be called a BASIC.

                    Andrew
                    dalke at dalkescientific.com





From duncan at NOSPAMrcp.co.uk  Thu Apr 24 08:38:20 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 12:38:20 +0000 (UTC)
Subject: eval function
References: <mailman.1051185097.26743.python-list@python.org>
Message-ID: <Xns93678A9E034CFduncanrcpcouk@127.0.0.1>

Franck Bui-Huu <Franck.BUI-HUU at gemplus.com> wrote in
news:mailman.1051185097.26743.python-list at python.org: 

> but eval seems to not use __getitem__ method of my_dict object because
> "using my_dict" is not printed.
> How can I print "using my_dict" when calling eval function ?

You can't. See http://www.python.org/2.2.2/descrintro.html
"Unifying types and classes in Python 2.2" by Guido van Rossum which
explains: 

> We can also use the new type in contexts where classic only allows
> "real" dictionaries, such as the locals/globals dictionaries for the
> exec statement or the built-in function eval(): 
> 
>     >>> print a.keys()
>     [1, 2]
>     >>> exec "x = 3; print x" in a
>     3
>     >>> print a.keys()
>     ['__builtins__', 1, 2, 'x']
>     >>> print a['x']
>     3
>     >>> 
> 
> However, our __getitem__() method is not used for variable access by
> the interpreter: 
> 
>     >>> exec "print foo" in a
>     Traceback (most recent call last):
>       File "<stdin>", line 1, in ?
>       File "<string>", line 1, in ?
>     NameError: name 'foo' is not defined
>     >>> 
> 
> Why doesn't this print 0.0? The interpreter uses an internal function
> to access the dictionary, which bypasses our __getitem__() override. I
> admit that this can be a problem (although it is only a problem in
> this context, when a dict subclass is used as a locals/globals
> dictionary); it remains to be seen if I can fix this without
> compromising performance in the common case. 

A workaround which might work for you would be to compile the expression 
first (using the 'compile' builtin) and then check the co_names attribute 
of the code object. Then you need to build a real dictionary with just 
those names that can be accessed by the expression. This won't, of course, 
help if your expressions depend on short circuiting logical operators.

>>> c = compile('1+x', '!', 'eval')
>>> c.co_names
('x',)
>>> eval(c, {'x':3})
4
>>> 




-- 
Duncan Booth                                            
duncan at rcp.co.uk int month(char
*p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3" 
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure? 



From graham at effectif.com  Tue Apr 29 17:07:45 2003
From: graham at effectif.com (Graham Ashton)
Date: Tue, 29 Apr 2003 22:07:45 +0100
Subject: Finding unused code
Message-ID: <pan.2003.04.29.21.07.45.44830@effectif.com>

I'm involved in maintaining a rather large Python project at work, and am
curious about how many of the classes, methods and functions that are
defined aren't actually used. There's been a fair amount of refactoring
over the last couple of years, so I wouldn't be surprised if there's
rather a lot of it.

I'd like to write a script to analyse the source code and tell me which
bits haven't been used, but I'm at a bit of a loss for where to start.

Any ideas? Or has somebody else already written such a tool?

--
Graham Ashton



From evan at 4-am.com  Thu Apr 17 17:22:40 2003
From: evan at 4-am.com (Evan Simpson)
Date: Thu, 17 Apr 2003 16:22:40 -0500
Subject: Zope Sequences and Python Scripting
In-Reply-To: <pan.2003.04.17.19.26.22.654130@fraley.de>
References: <pan.2003.04.17.12.34.15.777976@fraley.de> <GN6dnaRxi5queAOjXTWcpw@august.net> <pan.2003.04.17.19.26.22.654130@fraley.de>
Message-ID: <3E9F1B20.4020306@4-am.com>

Patrick W. Fraley wrote:
> Also after reading up a little more on Zope doc's I get the feeling that
> I should be doing this with page templates instead. Any comments,
> suggestions?

Yep - your feeling is correct, and you'll get *much* better answers to 
Zope questions on the Zope mailing lists than here on c.p.g.

Translation to ZPT:

<tal:block repeat="row here/filter_row">
    <span tal:replace="item/filter_id">Id</span><br />
    <span tal:replace="item/bez_ref">Ref</span><br />
    <span tal:replace="item/col_span">Span</span><br />
    <hr />
</tal:block>

(assuming that 'filter_row' is a Script or something that DTML is 
acquiring).

Cheers,

Evan @ 4-am






From aleax at aleax.it  Fri Apr 25 12:58:18 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 16:58:18 GMT
Subject: pythong referencing/aliasing
References: <ylcqa.9116$h%2.963225@read1.cgocable.net> <mailman.1051284819.25791.python-list@python.org> <vaioqs7e92v14d@corp.supernews.com>
Message-ID: <KOdqa.7010$K35.195585@news2.tin.it>

<posted & mailed>

Cameron Laird wrote:

> In article <mailman.1051284819.25791.python-list at python.org>,
> Jeff Epler  <jepler at unpythonic.net> wrote:
> .
> [other true stuff]
> .
> .
>>give you a clearer explanation.  I can only add one other piece of
>>information: I rarely am hurt by the fact that Python doesn't have
>>references of the type you're talking about.
> .
> .
> .
> I'll make a positive assertion of this:  in C and
> C++, stylish programmers frequently use references

Well in C they use POINTERS of course, since references
aren't in the language (they are in C++).  And yes, of
course EVERY C programmer uses pointers extensively --
it's hard to write any substantial C program without
such extensive use.  In C++, having both references and
pointers, one uses both, for different roles.

> to good effect.  Python rarely, very rarely, needs
> to do the same.  It might be instructive to choose
> a typical C++ reference idiom, and illustrate how
> its typical use case is better served in Python be
> a derived return, or dictionary construction, or ...

...or direct use of the fact that Python also passes
'references' -- but they're references to VALUES, not
to NAMES.  So, a typical C++ use such as:

void transfer_sum(Dollars x, Account& from, Account& onto)
{
    from.withdraw(x);
    onto.deposit(x);
}

translates right into Python

def transfer_sum(x, from, onto):
    from.withdraw(x)
    onto.deposit(x)

assuming in each case that a class Account has appropriate
methods withdraw and deposit.

In C++, you have to specify that from and onto are passed
by reference, otherwise you'll get COPIES of the from and
onto objects (if class Account supports copy-assigment at 
all, otherwise, more likely, an exception).  In Python, you
don't get copies unless you explicitly ask for copies, so
the issue just doesn't arise.

The only C++ usage that needs *translation* is that where
a function ASSIGNS to its by-reference argument.  In C++,
variables are "boxes" with values inside -- a reference is
a reference to the "box".  In Python, variables are "labels"
("post-it notes") weakly attached to values -- a reference
is a reference to the *value* (there is no such thing as
a "reference to the variable", i.e., "to the NAME").

Java's a lot like Python in this way, incidentally (it has
some complications for a few types, such as numbers, which
use a different assignment and argument-passing approach
than all 'normal' types -- but in practice, since the
equivalent Python types are immutable, Python and Java
behaviors end up being very close anyway in this respect).


Alex





From haribeau at gmx.de  Wed Apr 23 15:06:47 2003
From: haribeau at gmx.de (Clemens Hermann)
Date: 23 Apr 2003 12:06:47 -0700
Subject: dbm concurrency
References: <6c492956.0304230623.74b92d7f@posting.google.com> <mailman.1051108959.9593.python-list@python.org>
Message-ID: <6c492956.0304231106.202555ca@posting.google.com>

Hi Chuck,

thanks for your answer!

> If it's a problem that the C program may cache old user/regexp 
> combinations after they've been updated 

should not be an issue. The C program is an external_acl validator
(http://devel.squid-cache.org/external_acl/) which is called for each
request. So there should not occure any caching because the dbm is
opened again for each call to the C program and closed later.

might there still occur concurrency problems in this scenario?

> You could also look into something 
> like the BerkeleyDB 4 from sleepycat.com, and what sendmail does for the 
> hash or btree map types like /etc/mail/aliases or mailertable or access.

thanks for the hints, I'll have a closer look at these

/ch



From aleax at aleax.it  Thu Apr 17 13:53:47 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 17:53:47 GMT
Subject: More
References: <teBna.9109$pK2.11130@news.indigo.ie>
Message-ID: <LSBna.28781$LB6.660244@news1.tin.it>

kpop wrote:

> Hi
> I usualy print strings the normal way
> print x, but if the string is too big it will
> go by the screen too fast and i wont be
> able to see anything. I notice the python
> internel help system hasa thing like the
> unix prgram more, where you press enter
> to go down. How can i use that in my own
> programs? Is there a function or a module.
> I am using windows so i dont have more.
> Thanks
> 
> reposted because i got the date wrong last time

Perhaps

import pydoc

pydoc.pager(thelongstring)

comes close to doing what you want.  However,
that will only help you if thelongstring has
many LINES, not necessarily if it's e.g. just 
one line of humungous length (as pydoc itself
doesn't need to struggle with that...).  I
cannot think, offhand, of a builtin way to
solve that problem -- normally, Python's own
formatting tools wouldn't break strings except
at word boundaries, which might not be enough
for you.  If you do observe this problem and
can't find a built-in solution for it, then
it's not hard to program one yourself, of
course, e.g. (warning, untested):

def splitby(s, L=78):
    lines = []
    while s:
        lines.append(s[:L])
        s = s[L:]
    return '\n'.join(lines)


Alex




From bfordham at socialistsushi.com  Thu Apr 10 21:01:27 2003
From: bfordham at socialistsushi.com (Bryan L. Fordham)
Date: Thu, 10 Apr 2003 20:01:27 -0500 (EST)
Subject: Various IE events
Message-ID: <1763.130.254.28.203.1050022887.squirrel@socialistsushi.com>

I recently pulled out some code I'd been working on a while back for
controlling Internet Explorer using win32com in python.  I would like to
be able to respond to events, but am having a bit of trouble.

I can handle the webbrowser events (ie OnNavigate, OnVisible, etc) with no
problem.  What I would like is to be able to tell, for instance, which
link was pressed on a page.  Or, when the user clicks a button on a form,
be able to see the values of each field and tell which button was clicked.

All the example code I can find, however, deals with the OnNavigate, et al
events.  Any suggestions?

thanks
--B

socialistsushi.com
sushi for everyone






From achrist at easystreet.com  Fri Apr 25 14:13:22 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Fri, 25 Apr 2003 11:13:22 -0700
Subject: Python Package Index @ python.org down?
References: <3ea96bce$0$49106$e4fe514c@news.xs4all.nl> <mailman.1051292020.13758.python-list@python.org>
Message-ID: <3EA97AC2.E2196DC7@easystreet.com>

I've seen them a few times recently, once this morning.  I waited
five seconds and clicked again, and it worked fine.

Al

Gerhard H?ring wrote:
> 
> Irmen de Jong wrote:
> > I get a Python stacktrace when trying to access the Python Package Index,
> > http://www.python.org/pypi
> >
> > "OperationalError: database is locked"
> >
> > Tried three times. Anything wrong over there?
> 
> I did get these as well yesterday.
> 
> See http://mail.python.org/pipermail/catalog-sig/2003-April/000376.html
> 
> -- Gerhard



From jason at tishler.net  Wed Apr  2 13:35:04 2003
From: jason at tishler.net (Jason Tishler)
Date: Wed, 02 Apr 2003 13:35:04 -0500
Subject: Tkinter under cygwin (was Re: converting from shell script to
 python)
In-Reply-To: <3E8B0E04.7000203@yahoo.com>
References: <mailman.1049248805.1309.python-list@python.org>
 <3E8B0E04.7000203@yahoo.com>
Message-ID: <20030402183504.GB3064@tishler.net>

Hal,

On Wed, Apr 02, 2003 at 08:21:24AM -0800, Hal Wine wrote:
> I.e. the official cygwin packages don't have Tkinter included. 

Except for python-2.2.2-6, Cygwin Python has included Tkinter since
python-2.1.1-2:

    http://cygwin.com/ml/cygwin-announce/2003-03/msg00017.html

> What all did you have to compile,

No compiling is necessary...

> and from where did you fetch it?

If you are not running python-2.2.2-7, then upgrade using Cygwin's
setup.exe.

Please read the README:

    http://www.tishler.net/jason/software/python/python-2.2.2.README

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From exarkun at intarweb.us  Tue Apr 15 13:26:58 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 13:26:58 -0400
Subject: Speeding up: s += "string"
In-Reply-To: <b7h53f$cf3$1@heckle.cc.uregina.ca>
References: <mailman.1050393203.4704.python-list@python.org> <UoPma.20694$T34.658757@news2.tin.it> <v9ntbr3ioli8d3@corp.supernews.com> <b7h53f$cf3$1@heckle.cc.uregina.ca>
Message-ID: <20030415172658.GB32515@meson.dyndns.org>

On Tue, Apr 15, 2003 at 08:29:35AM -0600, sik0fewl wrote:
> Cameron Laird wrote:
> >I'm not optimistic.  I'll be more precise:  the problem
> >is real, and good solutions are sensitive to data char-
> >acteristics.  It's easy to improve performance of the
> >s1 += s2 operations that characterize e-mail filtering
> >while simultaneously degrading unacceptably the same
> >operation embedded in an XML filter.  I am close to
> >asserting the conjecture that every solution harms too
> >many existing legitimate users.
> 
> Just adding my two cents..
> 
> People should keep in mind that s += a is shorthand for s = s + a. If you 
> look at it the second way you can clearly see that a is appending to s and 
> then s is *assigned* that value. In both cases two operations are being 
> done.
> 

  While your first point is correct (but *only* for immutable s), I don't
see how it matters that there are two operations being performed.

  Binding a name is practically free.  It's an extra trip through the
interpreter loop, yes, but those are pretty fast.  The opcode isn't called
STORE_FAST for nothing ;)

> Perhaps we could introduce .= as a true assign/append operation. Although 
> this would still be a mutable operation, which contradicts the documents.

  And would hence require a new string type.  I somewhat like Alex's idea of
wrapping std::rope.  I've never had a practical use for it myself, but I've
always *wanted* to, because it's such a nice approach :)

  Jp

-- 
No, `Eureka' is Greek for `This bath is too hot.'
                -- Dr. Who
-- 
 up 26 days, 13:02, 2 users, load average: 0.00, 0.02, 0.04




From chad.scherrer at pnl.gov  Wed Apr 16 15:16:55 2003
From: chad.scherrer at pnl.gov (Chad Scherrer)
Date: 16 Apr 2003 12:16:55 -0700
Subject: compatibility problem
Message-ID: <1f4c702e.0304161116.5b0f6565@posting.google.com>

I'm having a compatibility problem, and I'm wondering if anyone might
be able to point me in the right direction. I'm running some code was
written for python 2.1, and I'd like to make it run under 2.2, so I
can subclass built-in types. The code is about 15,000 lines, with lots
of different classes running around, but the gist of the problem is
finding a solution set to an equation based on input from a file.
Python 2.1 gives the correct solution, but somehow 2.2.2 is only
giving me a subset. Not only that, but the order is different! The
ordering is arbitrary anyway, but I would still expect it to be
consistent between the two.

So I'm wondering what I should look for to make this work under 2.2.2.
My first thought was nested scopes, but the PEP I read suggests that
Python 2.1 gives a warning for any case that might cause a problem in
2.2. I don't think it could be the method resolution order, because
object (or any built-in type) is never subclassed. Is there anything
else I should be looking out for?

Thanks,
Chad Scherrer



From adalke at mindspring.com  Wed Apr  9 21:12:57 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 9 Apr 2003 19:12:57 -0600
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain> <6j8u8v0hjh3n9dm1jjflu3a970ada6qho4@4ax.com> <slrnb933p7.46e.rmunn@localhost.localdomain> <0au59v0n82d016vs3mai634d4088uf9f0k@4ax.com> <slrnb98hor.8ee.rmunn@localhost.localdomain>
Message-ID: <b72g6v$g0j$1@slb2.atl.mindspring.net>

Robin Munn:
> But that is not the Python philosophy. The Python philosophy is "if it
> quacks like a duck, then treat it like a duck". :-)

Rather, if you only need something which quacks, don't worry
about checking that it's also a duck.

                    Andrew





From jepler at unpythonic.net  Tue Apr 15 14:21:42 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Tue, 15 Apr 2003 13:21:42 -0500
Subject: pread/pwrite (was Re: files: direct access - NOT sequential)
In-Reply-To: <200304151925.57736.aleax@aleax.it>
References: <mailman.1050398560.8881.python-list@python.org> <S_Sma.22945$T34.690529@news2.tin.it> <20030415161936.GA30739@unpythonic.net> <200304151925.57736.aleax@aleax.it>
Message-ID: <20030415182141.GC30739@unpythonic.net>

On Tue, Apr 15, 2003 at 07:25:57PM +0200, Alex Martelli wrote:
> I think the positioning argument[s] should be the same as for .seek --
> ONE argument (unless it's a tuple specifying origin and offset) would
> not suffice.  Consider a frequent case: "I want to write these bytes
> at the END of the file" - no matter how long it's now and where it's
> currently positioned; or, ditto but "at the START of the file"; I don't
> think the Pythonic convention of -1 to meand end would work here,
> i.e. to write _after the end_ -- I think that, like for seek, offset and
> indicator of origin (start, end, current) would in general be needed.

Well, pread()/pwrite() don't have 'whence' arguments.  I think the
reasoning is that only SEEK_SET makes sense.  SEEK_CUR doesn't, because
the idea is to be independent of a file position set in a different
thread.  SEEK_END doesn't, because again you'd have to have some sort of
mutex on the file to know just where the end *is* when you make your
pread() call.  (if you *always* want to append to the file, there's a way
to do that now with write(), isn't there? opening with os.open(,O_APPEND)
/ file(,"a").  Hm, what does pwrite() do on a file opened with
O_APPEND?)

Some prototypes:
       off_t lseek(int fildes, off_t offset, int whence);
       ssize_t pread(int fd, void *buf, size_t count, off_t offset);
       ssize_t pwrite(int fd, const void *buf, size_t count, off_t offset);

[snippage]
> I do think that the key issue is defining the syntax, semantics and
> use cases, as opposed to implementing via a patch

I was really thing in "wrap C routines to Python" mode, not "what would
be a useful way to expose this kind of functionality to Python" mode...

Frankly, I'm less and less sure this is actually useful in Python.
(Yeah, I know, I'm the fellow who brought it up)  If you're going to use
an existing library, you'll wrap it (and not pread/pwrite directly).
If you're a real masochist, and want to interoperate with something using
that library in another thread you'd want this.  Blech.  But if you're
writing a Python program you'd probably never use something as primitive
as a single file object shared in multiple threads.

Jeff




From soporte_cali at terra.com.co  Mon Apr 21 06:38:35 2003
From: soporte_cali at terra.com.co (soporte_cali)
Date: Mon, 21 Apr 2003 20:38:35 +1000
Subject: "pythonic"
Message-ID: <d36a03d375fc.d375fcd36a03@terra.com>

I am quite new to python so please forgive me if my question is too 
newby. I am running a phyton exe file, linux red hat, it is an indexing
system that takes a flat file and index it. the py program once runned
is giving  me this:

 ./index.py:57: SyntaxWarning: import * only allowed at module level
  def createIndex(dbname,idtokenNum=1,keySize=10,idtokenChar=' '):
error: unknown library " <nameoflibrary.py> " does not exit 
Library  nameoflibrary.py  Does not exist

so where should I place the name of this library? I mean is there any
special place to put it? libnames? well thanks in advance.. 

_______________________________________________________________________
Te recomendamos:
Participa en la F?rmula Terra 2003: http://co.terra.com/formula1/
Consulta a diario tu hor?scopo: http://www.terra.com.co/horoscopo/






From aahz at pythoncraft.com  Tue Apr 29 10:06:10 2003
From: aahz at pythoncraft.com (Aahz)
Date: 29 Apr 2003 10:06:10 -0400
Subject: POD vs. reST for standalone writing?
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <mailman.1051584944.30800.python-list@python.org> <b8kuet$ftn$1@panix1.panix.com> <mailman.1051591198.24028.python-list@python.org>
Message-ID: <b8m0si$bus$1@panix3.panix.com>

In article <mailman.1051591198.24028.python-list at python.org>,
Andrew Bennetts  <andrew-pythonlist at puzzling.org> wrote:
>On Tue, Apr 29, 2003 at 12:18:37AM -0400, Aahz wrote:
>> In article <mailman.1051584944.30800.python-list at python.org>,
>> Andrew Bennetts  <andrew-pythonlist at puzzling.org> wrote:
>>>
>>>  - ReST isn't extensible enough; how do you add markup for, e.g. API
>>>    references?  How do you plug that into your docutils installation?  You
>>>    can't.  (The Lore markup for APIs is 
>>>        <code class="API">twisted.internet.reactor.run</code>
>>>    )
>> 
>> You're wrong::
>> 
>>     :API:`twisted.internet.reactor.run`
>> 
>> Granted that isn't currently trivial to add, but it's definitely
>> possible.
>
>Hmm, interesting.  I thought the :Foo: syntax was for definition lists,
>not inline in text?

The inline context uses ``:Foo:`` as the role attribute for interpreted
text.  For example, in my writing, I have a :chapter:`Foo` role.

>I guess what I'm asking is, what would be the ReST
>equivalent of:
>
>    <p>To start the reactor, use the <code class="API"
>    base="twisted.internet">reactor.run</code> method.  To stop it, use
>    <code class="API" base="twisted.internet">reactor.stop</code>.</p>
>
>[The optional base attribute provides the fully qualified name to allow
>the text to be linked to the right part of the API docs, but doesn't
>affect the final display.]

Adding in the base attribute, I don't think that can be done, but I
might be wrong.  You might be able to twist around it through
auto-generation of roles:

    :twisted.internet:`reactor.run`

but David Goodger would probably kill me for suggesting it.  ;-)

>And how would I make the ReST tools link that to my API documentation
>generated with pydoc, epydoc, or whatever?

You'd have to write code that figures out the context.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From LogiplexSoftware at earthlink.net  Mon Apr 21 21:05:18 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 21 Apr 2003 18:05:18 -0700
Subject: "pythonic"
In-Reply-To: <20030417123557.B1553@localhost.localdomain>
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
	 <qait9vsttagq0c8gk5f40ou991bitistfe@4ax.com>
	 <20030417123557.B1553@localhost.localdomain>
Message-ID: <1050973518.18603.13.camel@software1.logiplex.internal>

On Thu, 2003-04-17 at 09:35, Jack Diederich wrote:

> I know you were being tounge-in-cheek, but to be a Communist in the 50s was to
> advocate the violent overthrow of the US government.  Today it just means
> you like hanging out in espresso bars.  McCarthy might have been a self 
> promoting dick, but there really _was_ a secret cabal out to get us. The fact
> that they didn't succeed doesn't mean they didn't exist.

How can you protect the implementation of something (in this case, the
government) by destroying its foundations (the Constitution and Bill of
Rights)?

Confused'ly,

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From jhs_bkk at nospam.yahoo.com  Sat Apr 12 09:16:27 2003
From: jhs_bkk at nospam.yahoo.com (Jason Smith)
Date: Sat, 12 Apr 2003 20:16:27 +0700
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <3e96b088_2@127.0.0.1> <xPzla.367$xm2.20@news02.roc.ny.frontiernet.net> <mailman.1050070881.6673.python-list@python.org>
Message-ID: <3e9813e4_2@127.0.0.1>

Bryan L. Fordham wrote:

> It would make perfect sense for a teacher to require you to use
> something more than list.sort() if you're studying sorting... but after
> that, why make you write your own sorting routines if what's provided
> works just fine?

I agree.

Still, what's wrong with this?

  a.the_sort_that_i_learned_last_week()

Anyway, this decision is all up to the teacher of course.  So then we see
that Python gives a teacher flexibility depending on the focus of study. 
And that sounds like another big plus to me.


----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---



From aleax at aleax.it  Tue Apr  1 04:26:32 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 09:26:32 GMT
Subject: Is Python overhyped (just like Java)?
References: <mailman.1049147749.15994.python-list@python.org> <u65pz3vnl.fsf@boost-consulting.com>
Message-ID: <cXcia.53505$i26.1636187@news2.tin.it>

David Abrahams wrote:

>> Also, "read all the lines from a file" is not a primitive in Python -
>> it's a method call ... *and* it is written in C.
> 
> That's what I mean by "a primitive".  All of Python is written in C,
> so that's as primitive as things get.  "Read all the lines from a
> file" can be specifically optimized for Python, whereas when you write
> the "obvious" code for it in C++, you're probably seeing overheads
> which reflect, e.g., buffering which makes sense for general-purpose
> I/O but which is a waste when reading all lines from a file.

Yeah, but as you'll see in my code in the other post I wasn't using
readlines anyway -- just looping on the file, i.e. reading a line at
a time with buffering.  That's pretty close to the loop calling
getline I have in the C++ version, one would hope.


>> Obviously there could be as good an implementation in C++ on the
>> same platform - whether there could be a *better* implementation is
>> another question.
> 
> And another question is whether there could be a better
> implementation *for that particular job*.

Sure!  Hey, 2 minutes' twiddling with the Python code (not even
bothering to profile etc, just on general principles) already
speeded it up a lot -- of course there must be even speedier ways.
Nor did I ever claim otherwise!


Alex




From mwh at python.net  Fri Apr 25 09:52:21 2003
From: mwh at python.net (Michael Hudson)
Date: Fri, 25 Apr 2003 13:52:21 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net>
Message-ID: <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk>

Thomas Heller <theller at python.net> writes:

> Michael Hudson <mwh at python.net> writes:
> 
> > mis6 at pitt.edu (Michele Simionato) writes:
> > 
> > > The meta-type conflict is rather annoying, indeed. According to the
> > > metaclass book, the language should generate the correct metaclass
> > > that solves the conflict automatically, without burden for the
> > > programmer. Python is not that magic (yet) therefore you have to
> > > solve the conflict yourself.
> > 
> > I guess a potential problem is that if you do
> > 
> > class C(object):
> >     __metaclass__ = A
> > 
> > class D(object):
> >     __metaclass__ = B
> > 
> > class E(C, D):
> >     pass
> > 
> > class F(C, D):
> >     pass
> > 
> > you'd like both E and F to have the *same* (rather than merely
> > equivalent) metaclasses.  This would seem to be tricky to arrange.
> 
> Wouldn't a caching class factory be able to do this?

Yes.  I only said "tricky", not impossible.

class MetaGit(type):
    def __eq__(self, other):
        # aha! this makes class objects unhashable, cackle!
        return self is other

I don't know whether you want to cope with things like that.

Cheers,
M.

-- 
  Just point your web browser at http://www.python.org/search/ and
  look for "program", "doesn't", "work", or "my". Whenever you find
  someone else whose program didn't work, don't do what they
  did. Repeat as needed.    -- Tim Peters, on python-help, 16 Jun 1998



From rmunn at pobox.com  Wed Apr  2 11:52:51 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 16:52:51 GMT
Subject: Newbie - C API Question
References: <30875970.0303310546.27111833@posting.google.com> <mailman.1049119604.1230.python-list@python.org> <90e3791.0303311642.45f0752d@posting.google.com>
Message-ID: <slrnb8lgbb.ccp.rmunn@localhost.localdomain>

Natalia C. <n4t4lia at yahoo.com> wrote:
> I wrote this code that I got somewhere from a book 
> 
[snip sample embedded-Python code]
> 
> ********************************************************
> and I just do a simple cc testProj.c
> and it gives me this compiling error which I don't know what to do
> with
> 
> Undefined                       first referenced
>  symbol                             in file
> Py_BuildValue                       testProj.o
> PyArg_ParseTuple                    testProj.o
> ld: fatal: Symbol referencing errors. No output written to a.out
> 
> I don't really understand what do you mean (and how do you do) the
> loading of libpython.a??

When you compile a program, the compiler takes your source code (in this
case the testProj.c file) and produces an "object file". The internal
details of object files are complicated, but basically it contains
multiple sections: one section contains the binary, compiled version of
your code, and another section contains a list of all the symbols
(names) that your code tries to reference. This list includes things
like printf, strcpy, PyArg_ParseTuple... For each of those names, there
is a library file (usually ending in .a or .so) somewhere on your disk
that contains the code that defines those names. printf and strcpy, for
instance, are defined in the standard C library, which is usually found
in /lib with a filename like /lib/libc.so.# (where # is the version
number of your C library; on my system it's 6). The names starting in Py
would be defined in the Python library, which is usually called
libpython.a or libpython.so; I would expect to find that file in
/usr/lib, or maybe in /usr/local/lib.

Now: once a compiler has built an object file, it then needs to "link"
that object file with the appropriate libraries in order to produce an
executable. The process of "linking" resolves all those name references,
so that when your code calls printf(), the executable actually has a
function call that jumps to the right printf() code. If there are names
referenced in your object file that the linker can't find any
definitions for, then it will refuse to produce an executable. The
standard C library is usually linked in by default, so you don't have to
specify it explicitly on your command line. But the Python library is
not linked by default, so you would specify it by giving the "-lpython"
parameter to your compiler. "-lfoo" tells the compiler than in the
linking stage, it should look for a file named libfoo.a or libfoo.so in
the standard library search path, which usually includes /lib, /usr/lib,
/usr/local/lib, and sometimes a few other directories as well.

So you should do "cc testProj.c -lpython".

I hope this wasn't a long-winded explanation of something you understood
perfectly well already. If so, I'm sorry for having wasted your time.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From lists at ghaering.de  Tue Apr  1 13:17:09 2003
From: lists at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 01 Apr 2003 20:17:09 +0200
Subject: PySqlite for the Python 2.3a
In-Reply-To: <5B880349CA70D4119B7500D0B74C62D9010F8015@EXCHANGE>
References: <5B880349CA70D4119B7500D0B74C62D9010F8015@EXCHANGE>
Message-ID: <3E89D7A5.4090300@ghaering.de>

[Spamming all involved lists, too]

Driver, David wrote:
> "I don't plan to make an "official" package for 2.3 alphas, but I could
> provide an "unofficial" one."
> 
> At this point I would gladly use an unofficial install. [...]

David,

I've now uploaded *unsupported* binaries for Python 2.3 alpha2 on 
Windows of the following to my homepage:

- CVS snapshot of PySQLite as of today
- CVS snapshot of pyPgSQL as of today
- build of egenix mxDateTime 2.1beta5 (zip file)

You can grab the goodies at:

http://ghaering.de/python/unsupported/

"Unspported" means I'm only interested in bug reports against released 
versions of Python (2.1.x or 2.2.x) or PySQLite/pyPgSQL in combination 
with a released mxDateTime version. Bugs in the combination of a 
pre-release Python, pyPgSQL/PySQLite *and* mxDateTime are just too hard 
to track.

Otherwise, have fun playing with the new 2.3 features :-)

-- Gerhard




From gh at ghaering.de  Fri Apr 25 09:35:45 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Fri, 25 Apr 2003 15:35:45 +0200
Subject: GUI libs
In-Reply-To: <fzo6emyy.fsf@python.net>
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
	<Kq9qa.9083$h%2.957726@read1.cgocable.net>
	<mailman.1051273778.5621.python-list@python.org> <fzo6emyy.fsf@python.net>
Message-ID: <3EA939B1.4080503@ghaering.de>

Thomas Heller wrote:
> Gerhard Haering <gh at ghaering.de> writes:
>>Mark Nenadov wrote:
>>
>>>Alessio Pace wrote:
>>>
>>
>>>>Hi, I'd like to make a GUI for a program written in python, meant to run
>>>>on mac, win and linux: which is the preferable solution? I saw around
>>>>tkinter, wxpython and pyqt....
>>>
>>>I'd strongly recommend wxPython. wxPython is quite a rich GUI
>>>library and I have found it to be more advanced/modern/feature-rich
>>>than tkinter.
>>
>>
>>You'll get a Python interface with the error behaviour of the wrapped
>>C++ library. The default is to crash. In my limited experience, anyway.
> 
> Not for me, but I can only speak for Windows.

On Windows, try this:

C:\>python
Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
 >>> from wxPython.wx import *
 >>> f = wxFrame(NULL, 1, "foo")
 >>> f.Show()

This will segfault Python. IIRC, this was what I did in my very first 
exploration of wxPython.

I believe the reason of this crash is that the wxWindows main loop isn't 
initialized, yet. What I've done may be not very clever, but in my 
opinion, a Python app should *never* crash. The whole point of me using 
a highly dynamic language like Python is that it will give me exceptions.

The high likelyhood of a wxPython crashing my Python interpreter pretty 
much kills the idea of interactively epxloring the framework for me. 
Which might not be very useful for a GUI framework, but it's the way I 
usually start learning a new Python library.

All this aside, I still prefer wxPython over the alternatives.

-- Gerhard




From edcjones at erols.com  Mon Apr 28 10:43:49 2003
From: edcjones at erols.com (Edward C. Jones)
Date: Mon, 28 Apr 2003 10:43:49 -0400
Subject: Idea for improving Pyrex "C name specifications"
Message-ID: <b8jeol$hcm$1@bob.news.rcn.net>

I am writing a program which outputs wrappers for some complex C header 
files. Obviously, I would prefer to wrap each declaration independently 
of all the others. But I have "namespace" problems.

Here are two C structs from "astruct.h":

typedef struct ABC
{
     int abc;
     int* xyz;
} ABC;

typedef struct DEF
{
     ABC* abcp;
} DEF;

Here is part of the code for wrapping these:

cdef extern from "astruct.h":
     ctypedef struct c_ABC "ABC":
         int abc
         int* xyz

cdef extern from "astruct.h":
     ctypedef struct c_DEF "DEF":
         c_ABC* abcp

Note the "c_ABC" in the last line. When I wrap "DEF", I need to know 
that "ABC" is wrapped in the same Pyrex file. This can be done by 
processing the header file twice. It would be nicer if I could do 
something like:

use c_ABC for ABC:
     cdef extern from "astruct.h":
         ctypedef struct c_ABC:
             int abc
             int* xyz

use c_DEF for DEF:
     cdef extern from "astruct.h":
         ctypedef struct c_DEF:
             ABC* abcp

Note that this removes the unpythonic c_ABC "ABC". I am no language 
designer so I will not defend the "use ... for ..." syntax.




From spahievi at vega.bg  Sat Apr  5 16:46:59 2003
From: spahievi at vega.bg (Niki Spahiev)
Date: Sun, 6 Apr 2003 00:46:59 +0300
Subject: What are the differences between SWIG and Boost.Python?
In-Reply-To: <roy-C473C1.08401305042003@reader2.panix.com>
References: <169a7283.0304040821.9ad0976@posting.google.com>
 <b6l0eq$tf0$1@main.gmane.org> <mailman.1049539029.20263.python-list@python.org>
 <roy-C473C1.08401305042003@reader2.panix.com>
Message-ID: <1763992290.20030406004659@vega.bg>

4/5/2003, 16:40:13, Roy Smith wrote:


RS> Plus I saw a recent reference to something called SIP?

RS> There was a recent item on slashdot lamenting how much effort is wasted 
RS> in the free software community on reinventing the wheel, and how much 
RS> better off we would be if there was less NIH going on and more people 
RS> cooperating on fewer projects.  Do we really need a half-dozen ways to 
RS> make Python talk to C/C++?

Unfortunately SIP can't handle smart pointers.

Niki Spahiev





From johnson at equadriga.com  Sat Apr 19 03:22:27 2003
From: johnson at equadriga.com (Elester Johnson)
Date: Sat, 19 Apr 2003 12:52:27 +0530
Subject: Python and SSL sockets
Message-ID: <LIELKILHEABECBGHPOOJCEEHCIAA.johnson@vdc.equadriga.com>



Hi Friends,

    How do I transfer an HTML file from a windows machine to a linux machine
(server), using SSL sockets.

Kind Rgds,
    Elester Johnson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030419/d5657993/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Notebook.jpg
Type: image/jpeg
Size: 2950 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030419/d5657993/attachment.jpg>

From stuffduff at cox.net  Sat Apr 26 23:26:09 2003
From: stuffduff at cox.net (stuffduff)
Date: Sun, 27 Apr 2003 03:26:09 GMT
Subject: Class problem: loosing my mind
References: <v0fmavkm9iknef7hit5n17d4m0qi6tpnia@4ax.com> <mailman.1051411241.22289.python-list@python.org>
Message-ID: <diimavsd60audlnjm80m6dcan0e8b67hae@4ax.com>

Yikes!

It WORKED!

I may still have a job come Monday!  ;)

Seems pretty wierd that:

import gV

fails while:

from gV import gV

works.

Thanks a 10E6!

P.S. Any of you 'language lawyers' feel free to enlighten me as to WHY
this works and the other doesn't, I'm all ears!


On Sun, 27 Apr 2003 12:39:25 +1000, Andrew Bennetts
<andrew-pythonlist at puzzling.org> wrote:

>On Sun, Apr 27, 2003 at 02:21:32AM +0000, <stuffduff> wrote:
>> I'm 'doing the class pickle thing.'
>> 
>> I understand that both the module doing the pickling and the module
>> doing the unpickling need to reference the same class definition.
>> 
>> So I go and make a directory under site-packages named gV and in the
>> directory i add a file gV.py class gO and a completely empty
>> __init__.py
>> 
>[..]
>> 
>> At the python prompt I enter import gV, no error.
>> 
>> Next I try and create an instance:
>> 
>> x=gV.gO()
>> 
>> and I get AttributeError: 'gV' module has no attribute 'gO'
>> 
>> What am I missing here?
>
>Try printing gV.  You'll see that it's the package, i.e. the directory with
>the __init__, not the module in it.
>
>You want to do "from gV import gV", then gV.gO() should work.
>
>-Andrew.
>




From llimllib at f2o.org  Tue Apr  8 02:56:36 2003
From: llimllib at f2o.org (Bill Mill)
Date: 7 Apr 2003 23:56:36 -0700
Subject: How to wait for a click?
Message-ID: <af26e1d6.0304072256.2567f2e6@posting.google.com>

Hello,
I am making a Go game in wxPython. I completed the GUI with few 
problems; originally it only allowed human players, the graphical
board was tied directly to the logical board, so the click handler
passed the move directly to my class representing a Go board. All 
was well. However, I am now trying to adopt a more elegant structure.

In the new structure, a referee class calls the genmove() function 
of a HumanGoPlayer class and expects as a return value a move. Herein
lies the problem; how do I have genmove() (which has a reference to 
the board) wait for a click and return the move? I know I could have
it check some variable in the board class every x milliseconds, but
that's very inelegant. Is there a better way? Am I thinking about the 
problem wrong?

Thanks
Bill Mill
llimllib at f2o.org
Peace.



From aleaxit at yahoo.com  Thu Apr 10 08:54:00 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: Thu, 10 Apr 2003 14:54:00 +0200
Subject: newbie: strange python problem
References: <pan.2003.04.10.12.27.06.600319@somedomain.com>
Message-ID: <b73ph80bbt@enews3.newsguy.com>

maxx wrote:

> Simple code (I think):
> 
> ---
> #!/usr/bin/python
> 
> import os
> 
> for line in os.popen('ls -la /').readlines():
>     direntry = line.split()
>     print direntry
>     print direntry[0]
>     print direntry[1]
>     print direntry[2]
> ---
> 
> No list-items with an index above 1 are printed. They produce a
> 'IndexError: list index out of range' error... which is strange, because
> the list 'direntry' is 9 items large, so listitems up to index 8 should be
> printed... right?
> 
> maxx

Not necessarily.  When I run it on my system, ls -la / produces
output that starts with a two-words line, such as
    
total 108

or the like.  So, running this script obviously does produce an
IndexError, and specifically:

bash-2.05b$ python pap.py
['total', '108']
total
108
Traceback (most recent call last):
  File "pap.py", line 10, in ?
    print direntry[2]
IndexError: list index out of range


Is this the symptom you're observing?  Apparently not, since you say
"the list 'direntry' is 9 items large", while in this case it's most
obviously only two...


Alex




From peter at engcorp.com  Mon Apr 28 16:58:58 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 28 Apr 2003 16:58:58 -0400
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com> <gkuqavo4cgrfuljhokpli5q41pdlvf8v76@4ax.com>
Message-ID: <3EAD9612.1162B8B5@engcorp.com>

Dale Strickland-Clark wrote:
> 
> Watch out for Livewires, though. It's blatant religious evangelistic
> twaddle wrapped in a veneer of respectable education.
> 
> Too much of that stuff and your intellect takes a dive. Telltale signs
> include your spelling of simple words like 'stuff', 'because' and
> 'cool' goes all to Hell.

Hmmm... I can't tell whether that site would affect a child's spelling,
but judging by the pages on Python, it seems your assertion is somewhat
unwarranted as a generalization.  Check out pages like this:

  http://www.livewires.org.uk/python/spec-v1.txt

or their licence at:

  http://www.livewires.org.uk/python/licence.html

These pages seem no different than pages with similar purposes on other 
sites which have no (apparent) religious affiliation whatsoever.

Dale, you might have a stick up your butt about religion(*), but please
don't steer someone away from a good Python resource without more 
to say against it than that you don't think highly of the methods of
the organization behind it.

-Peter

(*) I'm about as atheistic as they come, so please don't bother attacking 
me on that front either....



From junk at aol.com  Mon Apr 21 16:53:03 2003
From: junk at aol.com (junk at aol.com)
Date: Mon, 21 Apr 2003 15:53:03 -0500
Subject: regular expressions
Message-ID: <b81lq4$17n8$1@galaxy.us.dell.com>

Can anybody recommend some good resources for learning regular
expressions.  Primarly web sites.



From aahz at pythoncraft.com  Wed Apr 23 19:18:08 2003
From: aahz at pythoncraft.com (Aahz)
Date: 23 Apr 2003 19:18:08 -0400
Subject: .pyc write location
References: <3EA6FBBA.50701@cs.cmu.edu>
Message-ID: <b876vg$hgv$1@panix1.panix.com>

In article <3EA6FBBA.50701 at cs.cmu.edu>,
John Kominek  <jkominek at cs.cmu.edu> wrote:
>
>I have a question about the location of compiled objects. By default, 
>when you compile a source file to a pyc (or pyo) file, the output file 
>is in the same directory as your source. I've always found this behavior 
>untidy. Maybe it's just me, but I like to build object files in a 
>directory separate from my source code.

PEP 304 -- Controlling Generation of Bytecode Files
http://www.python.org/peps/pep-0304.html
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From b.maryniuk at forbis.lt  Fri Apr 25 10:18:14 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Fri, 25 Apr 2003 16:18:14 +0200
Subject: What means exactly "Memory error"?
In-Reply-To: <slrnbai52u.8cv.hat@se-46.wpa.wtb.tue.nl>
References: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net> <mailman.1051198958.16669.python-list@python.org> <slrnbai52u.8cv.hat@se-46.wpa.wtb.tue.nl>
Message-ID: <200304251618.14933.b.maryniuk@forbis.lt>

On Friday 25 April 2003 12:56, Albert Hofkamp wrote:
> Do you need all data DOMmed at once?

That what I doing right now. Well, it not so easy to make an app with high 
perfomance where to parse it *twice* and re-collect data from the disk 
storage. But I see there is no other way...

-- 
Regards, Bogdan

We have enough youth, how about a fountain of SMART?





From info at dohao.org  Wed Apr 23 23:02:45 2003
From: info at dohao.org (Wenshan Du)
Date: 23 Apr 2003 20:02:45 -0700
Subject: Zope and FreeBSD
References: <e786d63d.0304201806.4c5293a6@posting.google.com> <mailman.1051115619.9208.python-list@python.org>
Message-ID: <f130eb8f.0304231825.69b51525@posting.google.com>

thanks, Dylan, I'll try.



From tyler at cg1.org  Fri Apr 25 18:19:51 2003
From: tyler at cg1.org (Tyler Eaves)
Date: Fri, 25 Apr 2003 22:19:51 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <vuiqa.3522$%_3.2810696@newssrv26.news.prodigy.com>
Message-ID: <bwiqa.242$5C2.185@tornadotest1.news.pas.earthlink.net>

dsavitsk wrote:

> 
> <achrist at easystreet.com> wrote in message
> news:3EA97D61.3CC9FE74 at easystreet.com...
>> Any suggestions of the shortest path from a Python program to RTF
>> output?  RTF seems to be the best (reasonable combination of
>> features and compatability) common denominator for exporting
>> data so that (windows) users can bring it into their own word
>> processing program, whatever that may be.  (Any problem with
>> that?)
>>
>> Requirements are about like this:
>>
>> 1. Don't need an edit control, just want to write a file, but if
>> I have to open an edit window, I can.
>>
>> 2. Don't want to do too much fancy.  Text, headings, bulleted
>> lists,  maybe some variations in font size.
>>
>> 3. Development platform is WinNT, target platform is any reasonably
>> recent Windows machine (eg either anything newer than Win95 release 1
>> or Win98 or better).
> 
> Your requirements suggest to me that html is the best common denominator,
> and it's easier to write.
> 
> -d

Except for that fact that most wordproccesors make a complete pigs ear of
it...




From nilsgri at idi.ntnu.no  Fri Apr 11 05:14:07 2003
From: nilsgri at idi.ntnu.no (Nils Grimsmo)
Date: Fri, 11 Apr 2003 11:14:07 +0200
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no> <ufzouhs8b.fsf@fitlinxx.com>
Message-ID: <pan.2003.04.11.09.14.07.31448@idi.ntnu.no>

On Mon, 07 Apr 2003 18:12:36 -0400, David Bolen wrote:
> The only time that the GUI will be updated is when events are being
> processed, which is normally blocked while you are processing a
> current event.  You can use wxYield() to process any pending events
> while still inside a current event handler if necessary to get GUI
> updates without returning from your current event handler.

thanks. this is what i was looking for. i am making a little board game,
and i like the board to be updated before the game i over when the
computer plays against itself :)


klem fra nils

-- 
nils grimsmo
nilsgri at idi.ntnu.no




From jeffbernstein at yahoo.com  Wed Apr  9 14:26:40 2003
From: jeffbernstein at yahoo.com (Jeff Bernstein)
Date: 9 Apr 2003 11:26:40 -0700
Subject: Adding item in front of a list
References: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de>
Message-ID: <f1e1cae2.0304091026.25f82c56@posting.google.com>

How about the following?

>>> x = [2,3,4]
>>> x
[2, 3, 4]
>>> x[:0] = [1]
>>> x
[1, 2, 3, 4]

Jeff


Thomas Guettler <pan-newsreader at thomas-guettler.de> wrote in message news:<pan.2003.04.09.14.42.36.965880.811 at thomas-guettler.de>...
> Hi!
> 
> I think it would be nice if
>  l=[2, 3, 4]
>  l.insert(1)
> 
> would insert in front of the list, as opposed to append() which adds new
> elements at the end.
> 
> l.insert(1) looks better than l.insert(-1,1)
> 
>  thomas



From jubafre at brturbo.com  Tue Apr  8 19:01:18 2003
From: jubafre at brturbo.com (jubafre at brturbo.com)
Date: Tue, 8 Apr 2003 20:01:18 -0300 (BRT)
Subject: STRING TO LIST
Message-ID: <10862677.1049842878351.JavaMail.nobody@webmail2.brturbo.com>

Hello, 

I need to change a variable from string to list. 
> 
> >>> x=str("[[12,13],[14,15]]") 
> >>> x 
> '[[12,13],[14,15]]' 
> 
> i have x and i want to change to a list, but not in this way: 
> 
> >>> x=list(x) 
> >>> x 
> ['[', '[', '1', '2', ',', '1', '3', ']', ',', '[', '1', '4', ',', '1', '5', ']', ']'] 
> 
> must be a list like this: 
> 
> >>> x=[[12,13],[14,15]] 
> >>> x 
> [[12, 13], [14, 15]] 
> 
> someone know??? 


Juliano Freitas

From pobrien at orbtech.com  Sun Apr  6 22:01:47 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 21:01:47 -0500
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <m2fzov5alw.fsf@localhost.localdomain>

Ben Finney <bignose-hates-spam at and-zip-does-too.com.au> writes:

> On Sun, 06 Apr 2003, Erik Max Francis wrote:
> >  Ben Finney wrote:
> > > In ork.py I want:
> > >     import eek
> >  
> > If the toplevel package is put in a place where Python can find it,
> >  	import monty.foo.eek
> 
> It's not.  Even if I make it so on my computer, it's not going to be so
> on someone else's computer when they download my package tree.

It will be if you tell them that's what's required.  That's how most
Python packages are distributed.  Put an __init__ in monty and foo,
and require users to put monty somewhere on their PYTHONPATH.  The
path to monty doesn't have to be directly in PYTHONPATH -- monty just
needs to be a child of some directory that is in PYTHONPATH.  For
example, I have a directory, ~/Code (C:\Code on Windows) that is in my
PYTHONPATH and then I just put all the various packages I work on as
subdirectories of ~/Code.  Other packages, like wxPython, IDLE, PIL,
Gtk, etc., install themselves in python2.2/site-packages, which is
also on the PYTHONPATH.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From Kyler at news.Lairds.org  Thu Apr  3 13:23:02 2003
From: Kyler at news.Lairds.org (Kyler Laird)
Date: Thu, 03 Apr 2003 18:23:02 GMT
Subject: Zope style (was: tuning our site (was: help! advocacy resources needed fast))
References: <mailman.1046819836.25704.python-list@python.org> <irdbl-1ds.ln1@news.lairds.org> <v8mjffpuph1oc2@corp.supernews.com> <4ufsl-ki3.ln1@beastie.ix.netcom.com> <v8oa2oi9094a0c@corp.supernews.com>
Message-ID: <d8aul-jto.ln1@news.lairds.org>

claird at lairds.com (Cameron Laird) writes:

>Interesting!  I want to be clear on the work-flow; you'd edit
>and preview in HomeSite, then switch back to the Zope manage-
>ment screens, and use the "Upload File" selection to update 
>the site itself--do I have that right?  Have I just demon-
>strated my ignorance of some far slicker interface? 

Yes, WebDAV.  I don't know about HomeSite, but at Purdue (where
I am no longer) there was work toward allowing users to do such
things using Macromedia/MS products.  (FTP is also an option,
but I refused to stoop to using FTP.)

>To me, that prospect seems frustratingly clumsy.

Using WebDAV and an "HTML editor" with which someone is already
familiar is a cheap way to get some users to maintain their own
content.  Using Page Templates allows dynamic pieces to be
added and not be destroyed in the editing process.

...or so I'm told...

--kyler



From andrew-pythonlist at puzzling.org  Sun Apr 27 01:47:06 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 15:47:06 +1000
Subject: Forgetting "()" when calling methods
In-Reply-To: <698f09f8.0304262115.386c2571@posting.google.com>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com>
Message-ID: <20030427054706.GK26530@frobozz.local>

On Sat, Apr 26, 2003 at 10:15:03PM -0700, Jeremy Fincher wrote:
> Frantisek Fuka <fuka at fuxoft.cz> wrote in message news:<b8cadp$1fj9$1 at ns.felk.cvut.cz>...
> > When I try to call methods, I sometimes forget to include the 
> > parentheses. Instead of:
> > 
> > if object.isGreen():
> > 	do something...
> > 
> > i sometimes write:
> > 
> > if object.isGreen:
> > 	do something...
> 
> Perhaps the answer is to remove the __nonzero__ method on
> functions/methods.  Logically, what purpose does it serve?

As I posted earlier in this thread, it allows this idiom:

        # Call the handler for this, if one exists
        handler = getattr(self, 'handler_' + name, None)
        if handler:
            handler(data)

I can see similar uses for another type of callable: classes.  E.g.:

    class XylophonePacket(Packet):
        ...

    class AccordionPacket(Packet):
        ...

    packetTypes = {
        'XY': XylophonePacket, 
        'AC': AccordionPacket,
        'PI': None   # Ignore pianos
    }

    class MusicalNetworkProtocol(Protocol):
        def dataReceived(self, data):
            try:
                packetType = packetTypes[data[:2]]
            except KeyError:
                raise BadPacketType, data[:2]
            
            if packetType:
                packet = packetType(data)
                ...

That's a contrived example off the top of my head -- but I have used code
very similar to that in the past.

There are certainly alternate ways to do these things that don't rely on
functions, methods and classes having a __nonzero__ (e.g. testing against
"is not None", or using a "lambda x: None" default argument), but there is
certainly existing code that relies upon this behaviour, so I think this
can't be changed until Python 3000.  

Even for Python 3000, I'm not sure it's a good idea.  It's currently
commonplace to simply say "if foo" to check if something isn't "empty", for
lack of a better word.  It's usually not any clearer to say 
    
    if len(myList) == 0:
        print "I have a list of stuff"

than

    if myList:
        print "I have a list of stuff"

I think the same applies to testing to see if a variable holds a callable,
just as much as it applies to checking the length of something.  Removing
this feature will lead to unnecessarily longer code for minimal benefit.

-Andrew.





From bdesth at removethis.free.fr  Wed Apr 16 19:37:58 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 16 Apr 2003 23:37:58 +0000
Subject: How many of you are Extreme Programmers?
References: <mailman.1050513152.32206.python-list@python.org> <3E9DABA4.B2B53153@engcorp.com> <3e9daec3$0$133$e4fe514c@dreader9.news.xs4all.nl>
Message-ID: <3E9DE956.8000001@removethis.free.fr>

Boudewijn Rempt wrote:


(snip)

> With the wrong partner, it can be hell. Curiously enough, wrong didn't have
> anything to with territoriality (snip)  but experience. And coding mentality. When your
> partner refuses to learn even the language you're supposedly programming in
> the fun ends pretty quickly.

Bad partner. Send partner to junk and find another one !

Laotseu




From hjwidmaier at web.de  Mon Apr 14 07:40:22 2003
From: hjwidmaier at web.de (Hans-Joachim Widmaier)
Date: 14 Apr 2003 04:40:22 -0700
Subject: Terminal Emulation
References: <pan.2003.04.13.07.40.04.733053@web.de> <mailman.1050260060.17218.python-list@python.org>
Message-ID: <6e990e29.0304140340.5e2c9153@posting.google.com>

Jp Calderone <exarkun at intarweb.us> wrote in message news:<mailman.1050260060.17218.python-list at python.org>...
[is there a terminal widget?]

>   Paul Swartz has done some work in this area, based off the Tk Canvas
> widget, for a "GUI" ssh client he wrote.  The relevant bits are distributed
> with Twisted, in the conch/ui package.
> 
>   It is very slow, but I think it supports a lot of vt100.  It might be
> worth basing your work off of.

Thanks for the suggestion, I've glimpsed at it. As I see it, it
doesn't do much more than color escapes. And "very slow" is not
exactly a virtue I really want it to have. But it's still a good
starting point (the best I have). Maybe a good idea to use the canvas
widget.

It'll be some time until I really have to start with that program.

Thanks again,
Hans-Joachim



From MFranklin1 at gatwick.westerngeco.slb.com  Thu Apr  3 14:24:07 2003
From: MFranklin1 at gatwick.westerngeco.slb.com (Martin Franklin)
Date: Thu, 03 Apr 2003 20:24:07 +0100
Subject: How to compile a python script to *.pyo
In-Reply-To: <2977.143.166.82.240.1049308819.squirrel@iamlearning.com>
References: <2977.143.166.82.240.1049308819.squirrel@iamlearning.com>
Message-ID: <3E8C8A57.9080603@gatwick.westerngeco.slb.com>

danny at iamlearning.com wrote:
> I'm new on this, please bear with we.
> 
> Let's say I have this script:
> 
> --------------------------
> #!/usr/bin/env python
> x = 'Hello Python'
> 
> print x,
> 
> --------------------------
> 
> How can I compile this script to *.pyo?

I'll ignore the why's and wherefore's and give you what I consider a 
straight answer.....

pydoc compileall

  Python Library Documentation: module compileall

NAME
     compileall - Module/script to "compile" all .py files to .pyc (or 
.pyo) file.

FILE
     /usr/local/lib/python2.2/compileall.py

DESCRIPTION
     When called as a script with arguments, this compiles the directories
     given as arguments recursively; the -l option prevents it from
     recursing into directories.

     Without arguments, if compiles all modules on sys.path, without
     recursing into subdirectories.  (Even though it should do so for
     packages -- for now, you'll have to deal with packages separately.)

     See module py_compile for details of the actual byte-compilation.

FUNCTIONS
     compile_dir(dir, maxlevels=10, ddir=None, force=0, rx=None)
         Byte-compile all modules in the given directory tree.


.....  snip .....



I use this for yes you've guessed it compiling my applications, on the fly
as they are installed.



Regards
Martin









From aleax at aleax.it  Sun Apr 27 06:48:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 10:48:39 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <mailman.1051423062.28410.python-list@python.org> <bOLqa.13827$K35.455183@news2.tin.it> <mailman.1051430877.5518.python-list@python.org>
Message-ID: <bAOqa.14719$K35.482952@news2.tin.it>

Andrew Bennetts wrote:

> On Sun, Apr 27, 2003 at 07:38:47AM +0000, Alex Martelli wrote:
   ...
>> See?  PyChecker CAN and DOES warn about "this case"!  But OF COURSE
>> it can't help with the following erro1.py:
> 
> Sure, but I said: "removing __nonzero__ from functions still doesn't help
> this case".  You can catch that today, with PyChecker, as you've

I.e., it doesn't help with a case that doesn't really need help -- is THAT
the "point" you thought it was worth a post to make (and DIDN'T mention!)?

There are a zillion cases where no "help" is particularly needed nor
warranted.  The fact that any proposed change "doesn't help" with any
subset of the cases that don't particularly need help is so clearly
irrelevant to any assessment of that proposed change that I _still_
don't see why ever one would want to start enumerating such cases.


> It's still not a justification for breaking __nonzero__ on functions. 

Would it damage you if that __nonzero__ raised a warning that is
suppressed by default, but can easily be (selectively) enabled?

> There are plenty of cases where "if func:" is perfectly valid, and no
> amount of "SINERGY" is going to convince me that making that fail (either
> in PyChecker or at runtime) is a good idea :)

I apologize for the misspelling, but it doesn't affect my argument.

The Romans said "experientia docet stultos" -- implying that some
people, presumably "the wise", can learn from the experience of
OTHERS, while others, "the foolish", only learn from their own.  I
find that few enough people learn even from their own experiences, 
that I consider this proverb to be wishful thinking.  But still, it
does seem to apply to the case in point.

You mentioned that you have no extensive experience teaching newbies,
particularly ones that come from the many languages where trailing
empty parentheses are not necessary to call an argument-less function,
such as Pascal and its descendants, as well as Visual Basic.  Yet it
is exactly to help such newbies that I'm considering making __nonzero__
raise an optional warning for function and method objects -- help them,
together with PyChecker for the diagnosis of other cases, get info
ASAP about the fact that once again they've forgotten a trailing '()',
rather than having to deduce it from subtle errors arising later and
quite possibly in quite separate points of their code.

If you have no experience with such teaching, nor any empathy for the
needs of support of such newbies, then my posts surely can't give it.
Just about the only thing I can do, is wish and hope that you will
find yourself thrust into such a teaching position, have to live with
the resulting frustration, and hopefully reflect that it could in
some part have been diminished if these warnings had been added.

They won't be, I think -- because of what was posted on this thread
by exactly one of the people I think such an addition would help.  He
argued, instead, that there will be no learning, ever, of the fact that
you need trailing parentheses to call callables without arguments; AND 
that the diagnosis should be performed in PyChecker (i.e. by static
checking), which is simply unfeasible -- implying that it would not help 
to have instead the optional early, precise warning that WOULD be feasible.

I think he's dead wrong on both accounts, but my disagreement in this
case has to be of a very different order from the one I have with you: I
can hardly claim I have MORE experience than he does at BEING a newbie,
after all.  In other words, my conviction that he's dead wrong is based
on arrogance -- on believing that I know what's best for HIM, better
than he does himself.  Sometimes (e.g. to act as a teacher) one does
need some touch of that kind of arrogance -- to do what one's deeply
convinced is best for the student, rather than what the student very
explicitly asks for -- but nevertheless I detest acting on such a basis.

Arrogant people, and teachers in particular, do much more harm than good
in this world.  I cannot erase my own arrogance just by wishing it gone,
but I can, at least, avoid acting on it when I have an option -- and in
this case, I do, and the "creative non-action" choice of doing nothing
at all on this subject thereby becomes attractive.


Alex




From srfarley at yahoo.com  Thu Apr 10 23:09:48 2003
From: srfarley at yahoo.com (Steven F.)
Date: 10 Apr 2003 20:09:48 -0700
Subject: Can I define "static" (like in C++) methods in Python?
References: <169a7283.0304101430.2bec922b@posting.google.com>
Message-ID: <86d3b75b.0304101909.342d2bcb@posting.google.com>

If you are using Python 2.2, see
http://www.python.org/2.2/descrintro.html#staticmethods
++Steve



From frank at niessink.com  Thu Apr 10 10:05:23 2003
From: frank at niessink.com (frank at niessink.com)
Date: Thu, 10 Apr 2003 16:05:23 +0200
Subject: ASTVisitor
In-Reply-To: <b73pks$t0t$1@swifty.westend.com>
References: <b73pks$t0t$1@swifty.westend.com>
Message-ID: <20030410140523.GA808@obstler>

On Thu Apr 10 14:58:36 2003, Peter Maas wrote:
> I am trying to employ ASTVisitor to convert a python expression to
> postfix form. Following the documentation, I created my own visitor
> class with methods called visit<class>, where <class> represents a
> syntax node like And, Or, ...:
> 
[snip]
> 
> only the Expression node was handled, i.e. the code above didn't
> descend down the tree. According to the docs the return value
> of visit... should decide wether to visit the node's subtree but
> my code didn't bother.

I agree the docs aren't very clear. Apparently, you have to restart 
the traversal after each match. Below is an excerpt of some of my 
code that visits python code and translates it into XML:


# from base.py:

import compiler

class Visitor(compiler.visitor.ASTVisitor):
    def walk(self, node):
        compiler.walk(node, self)


# from class_.py:

class ClassVisitor(base.Visitor, mixin.ImportVisitor):
    def __init__(self, doc, sourcefile = None):
        self.doc = doc
        self.sourcefile = sourcefile or pythonfile.PyFile()

    def visitSuperclasses(self, bases):
        superclassVisitor = SuperclassVisitor()
        for base in bases:
            superclassVisitor.walk(base)
            self.doc.insertSuperclass(superclassVisitor.name)

    def visitClass(self, classNode):
        self.doc.insertClass(classNode.name)
        self.visitSuperclasses(classNode.bases)
        self.doc.insertSourceFile(self.sourcefile.path())
        self.walk(classNode.code)

    def visitFunction(self, functionNode):
        visitor = function.FunctionVisitor(self.doc)
        visitor.walk(functionNode)

    def visitAssName(self, assnameNode):
        self.doc.insertClassField(assnameNode.name)


class SuperclassVisitor(base.Visitor, mixin.NameVisitor):
    pass



Hope this helps,

Frank
-- 
"Look," said Arthur, "would it save you a lot of time if I just gave up and 
went mad now?"
	-- Douglas Adams, 'The Hitch Hiker's Guide to the Galaxy'




From staschuk at telusplanet.net  Fri Apr 18 13:07:50 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 18 Apr 2003 11:07:50 -0600
Subject: Classes in Python
In-Reply-To: <3EA01A9B.4030209@vnet.net>; from wcmartin@vnet.net on Fri, Apr 18, 2003 at 10:32:43AM -0500
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <20030418110750.D709@tibia.amotlpaa.bogus>

Quoth Bill Martin:
> I'm wondering about the value of allowing a class definition like this:
> 
>              class C:
>                pass
>
> [...] I know there must be a 
> reason that python allows this kind of thing that I described, but I 
> can't figure out what that reason could be. [...]

One slightly different form which I find very useful:

    class HullBreachError(Exception):
        pass

Everything I need is inherited from Exception, with the (ahem)
exception that I want this type of error to be a distinct class,
so try: ... except HullBreachError: ... works.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"What I find most baffling about that song is that it was not a hit."
                                          -- Tony Dylan Davis (CKUA)




From reply.in at the.newsgroup  Mon Apr  7 03:50:31 2003
From: reply.in at the.newsgroup (Rene Pijlman)
Date: Mon, 07 Apr 2003 09:50:31 +0200
Subject: PYTHON EDITOR!!
References: <mailman.1049690233.24009.python-list@python.org>
Message-ID: <2db29vkr03tkr5avg6fnlvm3831go0pgmm@4ax.com>

Pathmenanthan RamaKrishna:
>im Nanthan, a postgraduate student from Malaysia.im very new to 
>python language but im already getting use of it.Currently,i 
>have been working in windows, but my development actually is on 
>Linux, so can u suggest me the best editor(i mean use friendly 
>and graphical,i dont want command line editor)for python?

Have a look at
http://www.python.org/cgi-bin/moinmoin/PythonEditors

-- 
Ren? Pijlman



From mlh at furu.idi.ntnu.no  Sun Apr 13 15:18:19 2003
From: mlh at furu.idi.ntnu.no (Magnus Lie Hetland)
Date: Sun, 13 Apr 2003 19:18:19 +0000 (UTC)
Subject: PEP305 csv package: from csv import csv?
References: <mailman.1049814434.25284.python-list@python.org>        <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>  <mailman.1050010580.22697.python-list@python.org>
Message-ID: <slrnb9jdvr.se1.mlh@furu.idi.ntnu.no>

In article <mailman.1050010580.22697.python-list at python.org>, Skip
Montanaro wrote:
>
>    Magnus> As far as I can see, csv.csv.__all__ contains the relevant
>    Magnus> names, so simply adding
>
>    Magnus>   from csv import *
>
>    Magnus> to csv.__init__.py should do the trick nicely.
>
> Done.

Great!

Now I can be 100% happy about the new cvs^H^Hsv package! Yay!

(Yes, I really did mis-spell csv above... Oh, well. :)

When will it be available in the main Python cvs? (Just curious...)

> Thanks.
> 
> BTW, I also make the csv/cvs typo periodically.  I think we'll all
> just have to grin and bear it...

Yup. No problem.

>Skip

-- 
Magnus Lie Hetland               "Nothing shocks me. I'm a scientist." 
http://hetland.org                                   -- Indiana Jones



From BPettersen at NAREX.com  Fri Apr 18 05:09:06 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Fri, 18 Apr 2003 03:09:06 -0600
Subject: question on buildinh /demo/embed
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFD8@admin56.narex.com>

> From: Martin v. L?wis [mailto:martin at v.loewis.de] 
> 
> "Jing (George) Cao" <jcao at cs.uchicago.edu> writes:
> 
> > I try to build the example in /Demo/embed, but there is 
> > several compiling errors. It is said in README that I 
> > should modify $LIBS as in /Modules/Makefile.  But I 
> > can't find any Makefile in the /Modoles directory.
> > 
> > I am lost here, anybody can help me?
> 
> If you want to get help, you should report the precise errors you are
> getting, and the precise version of the software you are using.

.. which Martin knows (but Jing might not) would also include the compiler the OS and potentially the hardware (is there any OSen left that run on two hw configs?)

-- bjorn

ps: yes, my first job was in tech support, so we'll need your registration number too <wink>.




From max at alcyone.com  Tue Apr 15 19:49:04 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 15 Apr 2003 16:49:04 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>	 <mailman.1049530929.9225.python-list@python.org>	 <a3b5b608.0304051049.7048d76@posting.google.com>	 <mailman.1049571010.25625.python-list@python.org>	 <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>	 <mailman.1049653811.11491.python-list@python.org>	 <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
Message-ID: <3E9C9A70.5A0D94C9@alcyone.com>

Andy Robinson wrote:

> First, let's note the half-a-dozen different templating schemes
> available.  DTML, Woven, Quixote, YAPTU, preppy.

EmPy, too.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The tremor of awe is the best in man.
\__/ Goethe
    REALpolitik / http://www.realpolitik.com/
 Get your own customized newsfeed online in realtime ... for free!



From max at alcyone.com  Wed Apr 16 19:10:26 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 16:10:26 -0700
Subject: Problem Cron-Job using Python-Script
References: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de> <mailman.1050493585.1066.python-list@python.org> <b7jhfs$1i2r1$1@ID-15245.news.dfncis.de>
Message-ID: <3E9DE2E2.968432A2@alcyone.com>

Mathias Waack wrote:

> Gerhard Haering <gh at ghaering.de> wrote:
> 
> > Cron normally doesn't set any environment variables.
> 
> Thats not right, cron sets at least HOME, SHELL and LOGNAME.

It varies from cron daemon to cron daemon.  It's true, though, that cron
rarely sets anything like the full environment that you're used to in an
interactive shell.

If you take a program that relies on external programs and put it in a
cronjob, you should expect to have problems due to the environment
differences.  Most notably, PATH will either be nonexistent or severely
limited.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Would a harvest from my heart / Find its season
\__/ Sandra St. Victor
    Python chess module / http://www.alcyone.com/pyos/chess/
 A chess game adjudicator in Python.



From skip at pobox.com  Tue Apr 22 14:42:34 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 22 Apr 2003 13:42:34 -0500
Subject: True/False
In-Reply-To: <1051036194.697.177.camel@lothlorien>
References: <1051036194.697.177.camel@lothlorien>
Message-ID: <16037.36122.922342.405340@montanaro.dyndns.org>

    Ian> True, False = 1==1, 0==1

Why not

    try:
        True
    except NameError:
        True, False = 1==1, 0==1

?  Should work anywhere without complaints.

    Ian> Is there any reason why reassigning True and False will be a
    Ian> problem in the future?

The builtins may get "tightened up" in the future.  This particular idiom
probably won't be affected, but you may not be able to (easily) shadow a
builtin in another module's global namespace, e.g.:

    import a
    a.True, a.False = 1==1, 0==1

might not be possible.  The idea here is to allow the byte compiler to know
all the possible builtin shadowing that goes on in a module by examining
just the module itself.  This, in turn, would allow access to builtins (and
globals to some extent I think) to be optimized.

    Ian> Should I tell people who are putting in version tests that it isn't
    Ian> necessary?

Testing against particular versions is generally not a good idea.  I'd test
against the presence of the object in question instead.

Skip




From mertz at gnosis.cx  Mon Apr 28 12:11:23 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Mon, 28 Apr 2003 12:11:23 -0400
Subject: Forgetting "()" when calling methods
References: <mailman.1051497342.21537.python-list@python.org>
 <ge5ra.13872$3M4.378491@news1.tin.it> <F+Ur+kKkX04C092yn@gnosis.cx>
Message-ID: <rKVr+kKkXwyA092yn@gnosis.cx>         

|Tim Peters wrote:
|> For years, ints came
|> out "less than" lists, which in turn came out "less than" longs, so that
|>     [10, [9], 7L, 8]
|> was considered to be "in sorted order" as-is.  It was something like 7
|> years before someone finally noticed how nuts that was.

Alex Martelli <aleax at aleax.it> wrote previously:
|*Blink* -- I don't get it.  Even if int < list < long, how CAN the
|two int's here be ``considered in sorted order''???

I had foolishly imagined that the Timbot had a slip-of-the-finger (or
whatever organ a robot has) in the example.  But in Python 1.5.1 (and
presumably earlier), the stated example is correct:

    Python 1.5.1 (#0, Sep 21 1998, 08:28:40) [VisualAge C/C++] on os2
    Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam
    >>> l = [10, [9], 7L, 8]
    >>> l.sort()
    >>> l
    [10, [9], 7L, 8]

This seems natural once you consider that:

    >>> 7L < 8
    1
    >>> [9] < 7L
    1
    >>> 10 < [9]
    1

:-)

Yours, Lulu...

--
    _/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: Postmodern Enterprises _/_/_/
   _/_/    ~~~~~~~~~~~~~~~~~~~~[mertz at gnosis.cx]~~~~~~~~~~~~~~~~~~~~~  _/_/
  _/_/  The opinions expressed here must be those of my employer...   _/_/
 _/_/_/_/_/_/_/_/_/_/ Surely you don't think that *I* believe them!  _/_/





From max at alcyone.com  Tue Apr 22 17:18:23 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 14:18:23 -0700
Subject: True/False
References: <mailman.1051036959.11378.python-list@python.org>
Message-ID: <3EA5B19F.D0D0E091@alcyone.com>

sismex01 at hebmex.com wrote:

> No problem, because (1==1) will evaluate to True [or bool(1)]
> and (1==0) will evaluate to False [or bool(0)].

It can be if Python starts generating warnings about rebinding True and
False.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From bignose-hates-spam at and-zip-does-too.com.au  Thu Apr 10 18:43:55 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Thu, 10 Apr 2003 22:43:55 GMT
Subject: Python compared toJava
References: <mailman.1049992396.31167.python-list@python.org>
Message-ID: <slrnb9bt6r.ou.bignose-hates-spam@rose.localdomain.fake>

On Thu, 10 Apr 2003 11:26:40 -0500, sismex01 at hebmex.com wrote:
> > I recently read that the Python interpreter is like the Java Virtual
> > Machine.
>  
>  Almost but not quite totally unlike.

Foul!  Offside reference to Hitch-Hiker's Guide in a Python forum!

-- 
 \      "Teach a man to make fire, and he will be warm for a day.  Set |
  `\    a man on fire, and he will be warm for the rest of his life."  |
_o__)                                               -- John A. Hrastar |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From Dick.Zantow at lexisnexis.com  Tue Apr  1 16:32:45 2003
From: Dick.Zantow at lexisnexis.com (rzed)
Date: Tue, 1 Apr 2003 16:32:45 -0500
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <3E89EC01.F0CF6D31@engcorp.com>
Message-ID: <b6d0jm$n7j$1@mailgate2.lexis-nexis.com>

Peter Hansen wrote:
> Mike Meyer wrote:
>>
>> Abstract
>>
>>     This PEP proposes adding Roman numerals as a literal type.  It
>>     also proposes the new built-in function "roman", which
>>     converts an object to an integer, then converts the integer to
>>     a string that is the Roman numeral literal equivalent to the
>> integer.
>>
>> Rationale
>>
>>     Roman numerals are used in a number of areas, and adding them
>>     to Python as literals would make computations in those areas
>>     easier. For instance, Superbowls are counted with Roman
>>     numerals, and many older movies have copyright dates in Roman
>> numerals.
>

As an enhancement, I'd like to suggest two-way conversion to the
similar Pig-Latinate numeric system of the form XIA (9), CIXA (91), or
IIVA (7).

> In addition, any calculations involving the Roman calendar
> (cf http://www.clubs.psu.edu/aegsa/rome/romec.html) would clearly
> benefit from such a module.
>
> My kalend.py module, currently running at 87,543 lines of code,
> could be shrunk substantially if it could do its calculations using
> the new roman type.  *Thank you*, Mike, for starting the process of
> getting this valuable feature into Python.
>
> (I've also considered whether we should be including the roman
> alphabet in the source encodings (PEP 263) so that, for example,
> names like "MAIOR" and "IVSTINIANUS" would be properly interpreted
> as "major" and "Justinians".  Unfortunately, this requires coupling
> that PEP with PEP 235 (Migration to Case-Insensitive Python), and
> coupling is always a bad idea in computing.)
>

Isn't that already covered in LATIN-1?

> Oh, and a Festive Veneralia to you all. :-)
>

I'm all for this proposal, but I'm afraid their may be some confusion
with proto-PEP 7334(Romansch Numerals) and PEP 2/14 (Romantic
numerals), not to mention PEPS 3.95/# (Romaine Minerals) and 5/$1
(Ramen noodles).

 --
rzed





From Jeffrey.S.Whitaker at noaa.gov  Wed Apr 16 18:41:57 2003
From: Jeffrey.S.Whitaker at noaa.gov (Jeffrey.S.Whitaker at noaa.gov)
Date: Wed, 16 Apr 2003 16:41:57 -0600
Subject: Standard way to express a date in a real number?
In-Reply-To: <420ced91.0304141131.73a6206e@posting.google.com>
References: <420ced91.0304141131.73a6206e@posting.google.com>
Message-ID: <Pine.OSX.4.53.0304161637310.801@mac10.cdc.noaa.gov>

On Mon, 14 Apr 2003, sdhyok wrote:

> I am looking for a way to express a date in a real number
> in Python.
>
> For instance, in MatLab,
> n = datenum(2001,12,19,18,0,0) returns n = 731204.75.
>
> This kind of expression turned out to be very helpful
> when handling or storing dates.
>
> Is there any funtion for the purpose in Python (e.g. in datetime module)?
>
> Daehyok Shin
>

Daehyok:

The mx.DateTime module will do what you ask
(http://www.egenix.com/files/python/mxDateTime.html).  For example, if you
want hours since Jan 1 1900:

import mx.DateTime
year = 1900
month = 1
day = 1
hour = 0
refdate = mx.DateTime.Date(year,month,day,hour)
curdate = mx.DateTime.now()
datediff = curdate - refdate
print datediff.hours


-Jeff

 --
Jeffrey S. Whitaker         Phone : (303)497-6313
NOAA/OAR/CDC  R/CDC1        FAX   : (303)497-6449
325 Broadway                Web   : http://www.cdc.noaa.gov/~jsw
Boulder, CO, USA 80305-3328 Office: Skaggs Research Cntr 1D-124



From andrew.henshaw at gtri.gatech.edu  Thu Apr 17 08:10:00 2003
From: andrew.henshaw at gtri.gatech.edu (Andrew Henshaw)
Date: Thu, 17 Apr 2003 08:10:00 -0400
Subject: HTML form pre-populating functionality or hints on building my own
References: <mailman.1050550654.3607.python-list@python.org>
Message-ID: <v9t6cvq9mmka8f@corp.supernews.com>

Scott Chapman wrote:

> Hi!
> I'm wanting to find (or make) a module that will take a HTML file, locate
> the input-related form tags (<INPUT, <TEXTAREA, <SELECT, <OPTION) and
> modify them by putting in values from a dictionary (i.e. pre-populate the
> form). I want a module that is implemented 100% in python and does _not_
> use any special tags to do this (such as webware, albatross, spyce, etc.)
> 
> The only thing I've seen that does this is EmbPerl (I'm tired of Perl) and
> Integratis' DFP module (implemented in C and not a straight-forward
> install).
> 

Try looking at PyMeld <http://www.entrian.com/PyMeld>.  Very slick.  You may 
need to check the license, however.


--Andy Henshaw



From john.p.brown at caramail.com  Tue Apr 15 17:15:51 2003
From: john.p.brown at caramail.com (John P. Brown)
Date: 15 Apr 2003 14:15:51 -0700
Subject: dll load failed: cannot find file specified
References: <mailman.1050255863.19330.python-list@python.org>
Message-ID: <4848e0e2.0304151315.1f77b490@posting.google.com>

John Hunter <jdhunter at ace.bsd.uchicago.edu> wrote in message news:<mailman.1050255863.19330.python-list at python.org>...

> Normally, I would think this meant the DLLs from the C lib weren't in
> the PATH environment variable, but they are: I can run the C examples
> and I can load all the pygtkglext *.pyd files in Dependency Walker
> without errors.  I have installed the GTK runtime, those libs are in
> my PATH, and pygtk 1.99.16 is running fine.  But something is going
> wrong when I try and import pygtkglext into python.
> 
> Any ideas what can be causing this message, or suggestions for how to
> debug what is going wrong, or other advice?

You might want to try Pe Explorer from http://www.heaventools.com


WBR,
JPB



From MCollins at co.seminole.fl.us  Wed Apr  9 16:13:01 2003
From: MCollins at co.seminole.fl.us (MCollins at co.seminole.fl.us)
Date: Wed, 9 Apr 2003 16:13:01 -0400
Subject: BEGINNER: converting integer to short month name?
Message-ID: <OF47A94545.30289BFA-ON85256D03.006E4C80-85256D03.006FA823@seminole.fl.us>

hi,

i've just been introduced to python.  i started out as with asp, migrated
to php, and now here i am with python.

anyway, i'm working on a utility that will download all the log files i
need every month for all the servers we monitor.  some of these files are
ftp'd and some are pulled down via http.

the http files have names such as ex0303.log or exMar03.log, depending on
the server admin's decision.

so i have a function that will take the current month and year, and back it
up to the previous month to build the correct logfile name (i.e., in april
2003, i'm downloading 0303.log).

that works fine for the numerical month names.  now i'm trying to convert
the numerical month into it's short month name equivalent to grab those
files.

i could always use a long if , elif, elif to do the job, but i was
wondering if there was built-in function already capable of that.  i was
just looking for something a little more elegant.

any suggestions?

current code:

month_year = strftime("%m,%y", gmtime())
      split_month_year = month_year.split(",")

      the_year = split_month_year[1]
      the_month = split_month_year[0]
      the_year = int(the_year)
      the_month = int(the_month)

      if the_month == 1:
            the_year = the_year - 1
            the_month = 12
      else:
            the_month = the_month - 1


      the_year="0" + str(the_year)
      the_month="0" + str(the_month)

      getMonthYear = fileString + the_year + the_month + ".log"
      return getMonthYear












From mathias-usenet at valpo.de  Wed Apr 23 04:07:25 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 23 Apr 2003 08:07:25 GMT
Subject: i18n in a library
Message-ID: <b85hjt$6l2na$1@ID-15245.news.dfncis.de>

I'm using some different applications which all use the same package.   
Lets assume there are app "A" and "B" and a package "L". Now I've got 
a question regarding the translation of these apps: 

It's easy to prepare translation using gettext.py. But I don't want 
to include all strings from the library in each app. So I get 3 
files: A.po, B.po and L.po. Is it possible to setup the gettext 
module to use both A.po and L.po to lookup strings for app A? 

I think there must be a simple solution because this should be a 
common problem for each library writer. But I can't find a solution 
in the python docs:(

Mathias



From garrett at bgb.cc  Sat Apr 19 01:14:07 2003
From: garrett at bgb.cc (Don Garrett)
Date: Sat, 19 Apr 2003 05:14:07 GMT
Subject: Python Enumeration class (revisited, again)
Message-ID: <zW4oa.521940$F1.74261@sccrnsc04>

   I wrote a Python enumeration class for my own benifit, and only 
afterwards, did I go searching and see how many other people had done it.

   I'm posting my main class here, just because I'd like to hear 
comments about style, etc.

   I'm also wondering if anyone has considered taking one of the many 
different pure Python Enumeration classes that have been proposed, and 
add it to the standard 'Batteries'?

   A zip file of this class including a test program can be found here:

http://www.bgb.cc/garrett/originals/Enumeration.zip

   But the main classes are here:

class Element(object):
     """Internal helper class used to represent an ordered abstract
     value.

     The values have string representations, have strictly defined
     ordering (inside the set) and are never equal to anything but
     themselves.

     They are usually created through the Create factory method as values
     for Enumerations.

     They assume that the enumeration member will be filled in before any
     comparisons are used. This is done by the Enumeration constructor.
     """

     def __init__(self, name, value):
         self.__name = name
         self.__value = value
         self.enumeration = None # Will be filled in later

     def __repr__(self):
         return self.__name

     def __cmp__(self, other):
         """We override cmp only because we want the ordering of elements
         in an enumeration to reflect their position in the enumeration.
         """
         if (self.__class__ == other.__class__ and
             self.enumeration is other.enumeration):
             # If we are both elements in the same enumeration, compare
             #   values for ordering
             return cmp(self.__value, other.__value)
         else:
             # Otherwise, fall back to the default
             return NotImplemented

class Enumeration(tuple):
     """This class represents an enumeration. You should not normally
     create multiple instances of the same enumeration, instead create
     one with however many references are convenient.

     The enumeration is a tuple of all of the values in it. You can
     iterate through the values, perform 'in' tests, slicing, etc. It
     also includes functions to lookup specific values by name, or
     names by value.

     You can specify your own element values, or use the Create factory
     method to create default Elements. These elements are unique
     system wide, and are ordered based on the order of the elements
     in the enumeration. They also are _repr_'d by the name of the
     element, which is convenient for testing, debugging, and generation
     text output.
     """

     def __new__(self, names, values):
         """This method is needed to get the tuple parent class to do the
         Right Thing(tm).
         """
         return tuple.__new__(self, values)

     def __init__(self, names, values):
         """The arguments needed to construct this class are a list of
         element names (which must be unique strings), and element values
         (which can be any type of value). If you don't have special
         needs, then it's recommended that you use Element instances for
         the values.

         This constructor is normally called through the Create factory
         (which will create Elements for you), but that is not a
         requirement.
         """

         assert len(names) == len(values)

         # We are a tuple of our values, plus more....
         tuple.__init__(self, values)

         self.__nameMap = {}
         self.__valueMap = {}

         # Store each enum value as a named attribute
         self.__slots__ = names

         for i in xrange(len(names)):
             name = names[i]
             value = values[i]

             # Tell the elements which enumeration they belong too
             if type(value) == Element:
                 value.enumeration = self

             # Prove that all names are unique
             assert not name in self.__nameMap

             # Create mappings from name to value, and vice versa
             self.__nameMap[name] = value
             self.__valueMap[value] = name

         # Create named attributes on ourself.
         self.attachValuesTo(self)

         # Force this instance to be read only
         def noset(name, value): raise "No assignments allowed"
         self.__setattr__ = noset

     def attachValuesTo(self, other):
         """Create named values on an object to match our elements.

         This is appropriate for use in a constructor, against a class
         object, or a module object.

         Cannot be used on a new style class instance after initialization
         is complete.
         """
         for name, value in self.__nameMap.items():
             setattr(other, name, value)

     def valueFromName(self, name):
         """Look up the enumeration value for a given element name.
         """
         return self.__nameMap[name]

     def nameFromValue(self, value):
         """Look up the name of an enumeration element, given it's value.

         If there are multiple elements with the same value, you will
         only get a single matching name back. Which name is undefined.
         """
         return self.__valueMap[value]

     __call__ = valueFromName


def Create(*elements):
     """Factory method for Enumerations. Accepts of list of values that
     can either be strings or (name, value) tuple pairs. Strings will
     have an Element created for use as their value.

     Example:  Enumeration.Create('fred', 'bob', ('joe', 42))
     """
     names = []
     values = []

     for element in elements:
         if type(element) == tuple:
             assert len(element) == 2

             names.append(element[0])
             values.append(element[1])

         elif type(element) in (str, unicode):
             names.append(element)
             values.append(Element(element, len(names)))

         else:
             raise "Unsupported element type"

     return Enumeration(names, values)



--
Don Garrett                     http://www.bgb-consulting.com/garrett/
BGB Consulting                                        dgarrett at acm.org






From skip at pobox.com  Wed Apr  2 17:13:41 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 2 Apr 2003 16:13:41 -0600
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
In-Reply-To: <3E8B55B2.D8D7D2F2@engcorp.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
        <b68jti$2mj66$1@ID-169208.news.dfncis.de>
        <RbCia.6782$9I.555309@news20.bellglobal.com>
        <mailman.1049298845.7655.python-list@python.org>
        <3E8B4094.F61515B3@engcorp.com>
        <ZpHia.23$Sd.4@fed1read01>
        <3E8B48E0.C786D192@engcorp.com>
        <mailman.1049316186.31356.python-list@python.org>
        <3E8B55B2.D8D7D2F2@engcorp.com>
Message-ID: <16011.24725.434698.210954@montanaro.dyndns.org>

    >> How about calling new.module() to create a new module object from the
    >> one you want to test which has suitably munged globals?

    Peter> Interesting idea... but ...
    ...
    Peter> # install mock open again
    Peter> mungedMut.open = mockOpen

    Peter> What I don't understand is the difference between the final line
    Peter> above and what I was originally trying to do, except that this is
    Peter> a newly created module instead of one that was imported.

    Peter> Okay, I guess I must have missed a key point and I'm way off the 
    Peter> track here.  Please clarify.

No, it turns out I was.  I was thinking you passed a bunch of stuff (like a
dict to use as globals()) to new.module().  I guess I was thinking of
new.code() and new.function().  You might be able to poke elements directly
into the module's __dict__.  Perhaps new.module can grow an optional dict
argument to handle this sort of thing.

I suggest you post your use case to python-dev as something for people there
to think about.  In fact, new.module() would become pretty useless if you
weren't able to do what you can do today, so I think it will probably need a
dict argument if the "no modifying another module's globals" restriction is
implemented.

Skip





From ianb at colorstudy.com  Tue Apr  8 00:32:46 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 07 Apr 2003 23:32:46 -0500
Subject: ANN: SQLObject 0.3
Message-ID: <1049776366.759.96.camel@lothlorien>

SQLObject 0.3 has been release:
  http://sqlobject.org

Changes
-------

* Automatic table creation based off class.
* Create classes from database.
* Dynamically add and remove columns to live classes and databases.
* anydbm-based backend (queryable database without SQL).
* Better caching.
* Some bugs fixed.

What is it?
-----------

SQLObject is an object-relational mapper.  It allows you to translate
RDBMS table rows into Python objects, and manipulate those objects to
transparently manipulate the database.

SQLObject emphasizes convenience.  It's easy to learn and easy to use,
and doesn't require too much typing or use of any tools outside of your
editor and your Python source.


-- 
Ian Bicking  ianb at colorstudy.com  http://colorstudy.com
4869 N. Talman Ave., Chicago, IL 60625  /  773-275-7241
"There is no flag large enough to cover the shame of 
 killing innocent people" -- Howard Zinn





From danb_83 at yahoo.com  Tue Apr  1 17:10:53 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 1 Apr 2003 14:10:53 -0800
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <ad052e5c.0304011410.54ff985a@posting.google.com>

Mike Meyer <mwm at mired.org> wrote in message news:<x7y92uhzpw.fsf at guru.mired.org>...
> PEP: 313
> Title: Adding Roman Numeral Literals to Python
[major snippage]

Not a bad idea.  I frequently do date calculations with copyright
dates and Super Bowl numbers, only to have to go through the hassle of
converting to Arabic numerals.  This PEP will save me a lot of time.

Speaking of time, it would also be nice to have sexagesimal literals
so we could could more easily write mixed hours-minutes-seconds
values.  For example, the literal 0s123456 would have the value 12 *
3600 + 24 * 60 + 56, or 45296.  Or, considering the intended
application, perhaps datetime.timedelta(0, 45296) would be better.



From martin at v.loewis.de  Thu Apr 17 13:31:30 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 17 Apr 2003 19:31:30 +0200
Subject: [Que]Do any SGML lib good for Python?
References: <2778697.1050595278@dbforums.com>
Message-ID: <m3of35ni8d.fsf@mira.informatik.hu-berlin.de>

dolephi9080 <member27604 at dbforums.com> writes:

> I search  the SGML Lib for python, but it's hard to find. even the
> python contain a sgml parser itself. But that's a limite parser.
> Does some one know it? pls tell me. Thx

I find that parsing the output of nsgmls is quite powerful and
effective.

Regards,
Martin




From jack at performancedrivers.com  Sun Apr 13 17:16:31 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Sun, 13 Apr 2003 17:16:31 -0400
Subject: Question regarding indirect function calls
In-Reply-To: <pan.2003.04.13.20.13.21.504295@maxwell.com>; from hammer@maxwell.com on Sun, Apr 13, 2003 at 04:13:21PM -0400
References: <pan.2003.04.13.20.13.21.504295@maxwell.com>
Message-ID: <20030413171631.I1039@localhost.localdomain>

On Sun, Apr 13, 2003 at 04:13:21PM -0400, Maxwell Hammer wrote:
> In the program I'm trying to write I have a module called "plugins"
> containing a number of functions. I only know at runtime the name of a
> function to call from "plugins", i.e. need to call a function indirectly.
> The problem is that assigning the name of one of plugin's functions to a
> variable, I cannot then call the function. IOW:
>  func=function1
>  plugins.func() ..........will fail
>  plugins.function1() .......will pass
> 
<snipped code>

What you are actually doing is this
  func = "function1"
  plugins.func()
which doens't work in the same way this doesn't work
  plugins."function1"()

Try one of these
  getattr(plugins, "function1")()
  plugins.__dict__["function1"]()


hope that helps,

-jackdied




From revanna at mn.rr.com  Tue Apr 22 23:39:10 2003
From: revanna at mn.rr.com (Anna)
Date: Wed, 23 Apr 2003 03:39:10 GMT
Subject: A few errors in "Python in a Nutshell" - Numeric related
References: <b6frd5$uq$1@peabody.colorado.edu> <pan.2003.04.03.00.14.15.232608@mn.rr.com>
Message-ID: <pan.2003.04.23.03.38.26.30560@mn.rr.com>

On Wed, 02 Apr 2003 18:14:15 +0000, Anna wrote:

Stuph that already got posted...
Sorry - my newsreader hiccuped. 

Bad newsreader. Go sit in the corner!

Anna



From sross at connectmail.carleton.ca  Sat Apr 26 13:16:39 2003
From: sross at connectmail.carleton.ca (Sean Ross)
Date: Sat, 26 Apr 2003 13:16:39 -0400
Subject: Some Python Code
Message-ID: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com>

Hi.
I've just finished a semester of university where I've been fortunate enough
to be able to use my language of choice, Python, for all of my programming
projects/assignments. I've implemented a general purpose genetic algorithm
package and a little circuit description programming language, plus a few
other AI related items of differing quality. I don't really want to do any
maintenance on these projects, so I'll not be bothering to put them up in
some place like sourceforge or pypi. However, the code may be useful to
someone else, perhaps as a starting point for their own projects, so I'd
like to make it available. For anyone who may be interested, I can be
contacted at sross at connectmail.carleton.ca or frobozz_electric at hotmail.com.
The code is, of course, free and open source although it has not been
licensed - except the programming language which contains a SPARK license.
OK. Thanks for your time.
Sean





From xxdigitalhellxx at hotmail.com  Tue Apr 15 11:05:53 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Tue, 15 Apr 2003 09:05:53 -0600
Subject: ANN: Capon 0.1 - a software build framework
In-Reply-To: <3e9c069f_1@spamkiller.newsgroups.com>
References: <mailman.1050389907.24670.python-list@python.org> <3e9c069f_1@spamkiller.newsgroups.com>
Message-ID: <b7h77f$dc6$1@heckle.cc.uregina.ca>

Brad Clements wrote:
> _
> "David LeBlanc" <whisper at oz.net> wrote in message
> news:mailman.1050389907.24670.python-list at python.org...
> 
>>>-----Original Message-----
> 
> 
>>>>   aahz> So how does Capon differ from SCons?
>>>>
>>>>Well, the name has funnier connotations, for one.   :)
>>>
>>>You deserve to be roasted for that comment.
> 
> 
>>A fate of many capons ;)
>>
>>First they have "the" operation, then they grow up and get put on a spit
> 
> and
> 
>>thence over a slow fire. Life as a chicken is hell ;)
>>
> 
> 
> My neighbhor, a 70+ year old former federal meat inspector (before that he
> was a Vet for the Italian Cavalry ) raises and does his own capons.
> 
> He showed me how to do "the operation".  Quite amazing to see him slice open
> a rooster's behind and reach right in. This one was a bit large, so he had
> to "go in from behind" instead of under the wing as he usually does. Then he
> just sews them up with dental floss. No antiseptic! he says chickens don't
> get staph infections. He says when he learned how to do "the operation" he
> killed 5 of them right off. Apparently there's a nack to not cutting some
> important arteries.
> 
> 
> He says at a store it's hard to find real capons. Back in the 70s they used
> estrogen to make capons, but that turned out to give people cancer.  The FDA
> banned that practice, however many "unscrupulous" poultry plants continue to
> use estrogen, but they also cut the rooster to give it "the scar".. So meat
> inspectors look, think the operation has been done, but it really hasn't.
> 
> Enjoy!
> 
> (btw, home grown capons are yummy!)

While I can't say for sure, I don't think this is the kind of discussion the 
OP was hoping for :)

-- 
Ryan




From DavidMLewis.NO at SPAM.mac.com  Wed Apr 30 19:55:02 2003
From: DavidMLewis.NO at SPAM.mac.com (David Lewis)
Date: Wed, 30 Apr 2003 23:55:02 GMT
Subject: 2.3b1 PackageManager fails to initialize, Mac OS X
Message-ID: <300420031655007042%DavidMLewis.NO@SPAM.mac.com>

I've built Python 2.3b1 from the tarball. I've got WASTE installed, so
it built the PythonIDE, etc. If I select PackageManager from the
PythonIDE file menu, I get a traceback from urllib2.py, when it tries
to get
'http://homepages.cwi.nl/~jack/pimp/pimp-darwin-6.5-Power_Macintosh.plis
t'. I can see that page in a browser, so that's not the problem. I can
manually execute the following code in the PythonIDE interactive
window:

   import urllib2

   f = urllib2.urlopen(
'http://homepages.cwi.nl/~jack/pimp/pimp-darwin-6.5-Power_Macintosh.plis
t')

   for line in f:
      print line

and that all works. Any ideas?

   David



From alloydflanagan at attbi.com  Wed Apr 30 10:53:26 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 30 Apr 2003 07:53:26 -0700
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <3EA9B64D.3208D730@alcyone.com> <roy-D5BE38.18462125042003@reader1.panix.com>
Message-ID: <a8b7f07a.0304300653.589b54ec@posting.google.com>

Roy Smith <roy at panix.com> wrote in message news:<roy-D5BE38.18462125042003 at reader1.panix.com>...
> Erik Max Francis <max at alcyone.com> wrote:
> > One thing that PyChecker, at least, is more likely to catch is "code has
> > no effect" statements like
> > 
> > 	sys.exit
> > 
> > Without the parentheses, this clearly can't do anything useful
> 
> In a language as dynamic as Python, you have to be really careful with 
> words like "clearly".
> 
> Invoking the name of an object to see if you raise a NameError could be 
> useful in some situations.  Within the past week or so, there was a 
> serious discussion on this group where code along the lines of:
> 
>    try:
>       True
>    except NameError:
>       do stuff
> 
> was suggested as a way to tell if the version of Python being run had 
> boolean support.

Just an FYI:  pychecker does indeed flag code like that as having 'no
effect', including the boolean test.  Very useful if you call a
function as a statement and forget the parentheses, easy to ignore if
you meant to do it.



From cnetzer at mail.arc.nasa.gov  Tue Apr  8 15:03:35 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 08 Apr 2003 12:03:35 -0700
Subject: How to protect python code ?
In-Reply-To: <Ohska.730$001.647543@monger.newsread.com>
References: <3E8BE455.4000904@novagrid.com>
	 <mailman.1049389032.4904.python-list@python.org>
	 <u8bka.651$001.591065@monger.newsread.com>
	 <mailman.1049722991.18682.python-list@python.org>
	 <3E921A4C.1020908@hotmail.com>
	 <mailman.1049769313.31229.python-list@python.org>
	 <Ohska.730$001.647543@monger.newsread.com>
Message-ID: <1049828614.601.10.camel@sayge.arc.nasa.gov>

On Mon, 2003-04-07 at 21:32, msw wrote:
>  The fact that your post is the first one that 
> references any actual code, that it occurs about 20 posts into this 
> thread, and that it is snide and insulting indicates that even if my 
> first suggestion wasn't very good, my second one was spot on.

OH, oh I'm sorry, but this is "abuse".  If you want an "argument", that
is 12A, just down the hall.

*plonk*









From ngo at cartan.de  Fri Apr 18 18:33:57 2003
From: ngo at cartan.de (Nils Goesche)
Date: 19 Apr 2003 00:33:57 +0200
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <b7m546$2a83$1@news.peterlink.ru> <lyptnlxq1n.fsf@cartan.de> <m24r4xdyd6.fsf@mycroft.actrix.gen.nz> <lyd6jlxkpi.fsf@cartan.de> <usft9v01l8fcdg8erdl2kishg0cnr1t9dp@4ax.com> <lyadepp1ck.fsf@cartan.de> <1lt0avslergps3bdv1lpvr48kef6t03640@4ax.com>
Message-ID: <87d6jjjuzu.fsf@darkstar.cartan>

Thaddeus L Olczyk <olczyk at interaccess.com> writes:

> While it's true that in math and CS, many definitions evolve
> over time ( ones that come to mind right now are
> function/procedure [which were distinct, now considered the
> same thing], fiber bundles, sheaves, topological space ... ),

Actually, the resp. definitions of fiber bundles, sheaves and
topological spaces /haven't/ changed much over time.  If they
have changed at all, once they were called by that names.

> that's all just bullshit thrown out to distract people from
> *your stupidity*.

Darn, I had hoped I'd get away with it one more time.

> Thaddeus L. Olczyk, PhD

I was somewhat hoping that now, having finally obtained your
degree, which is seemingly very important to you, you'd have
become a bit more relaxed.  I was wrong.

Farewell,
-- 
Nils G?sche
Ask not for whom the <CONTROL-G> tolls.

PGP key ID #xD26EF2A0



From tug at wilson.co.uk  Tue Apr 29 12:27:10 2003
From: tug at wilson.co.uk (John Wilson)
Date: Tue, 29 Apr 2003 17:27:10 +0100
Subject: Problem subclassing tuple
References: <F7DB8D13DB61D511B6FF00B0D0F062330489D3F8@mail.hebmex.com>
Message-ID: <094e01c30e6c$2edfc6a0$640a0b0a@handel>

----- Original Message ----- 
From: <sismex01 at hebmex.com>
To: <python-list at python.org>
Sent: Tuesday, April 29, 2003 5:16 PM
Subject: RE: Problem subclassing tuple


[snip]

> Most prob'ly you've already grasped it; the problem with
> the current implementation is that many internal classes
> don't "allow" subclassing, because the interpreter's
> internal mechanics access the C code for those classes
> (and subclasses), so your Python code doesn't get called.
>
> Bummer, really. :-/

No doubt this will be fixed in the future. I'll just create a class which
delegates the calls I need to support.

Many thanks for your help - at least I know it's not my brain softening :-))

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From richardd at hmgcc.gov.uk  Wed Apr  2 04:58:51 2003
From: richardd at hmgcc.gov.uk (Richard)
Date: Wed, 2 Apr 2003 10:58:51 +0100
Subject: Python equivalent of _tempnam?
Message-ID: <3e8ab45b$1@mail.hmgcc.gov.uk>

Hi,

I am writing an application which produces file output. Files are outputted
to a directory and I want to ensure that if an output filename is given that
already exists in that directory a number is affixed to the filename to
ensure that the existing file is not appended to or overwritten. i.e.
output_0.txt, then output_1.txt etc...

I have been told that a way of doing this in C is by using _tempnam( ).
However I want to write this in Python, so is there an equivalent function?
My application will only be executed occasionally so I am unable to keep an
internal record of file numbering. I assume that I need to find the highest
number file (if exists) in the directory and then increment that by one
before creating the new file.

Any ideas?

Cheers

Richard





From peter at engcorp.com  Wed Apr  2 10:02:15 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 10:02:15 -0500
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to 
 Python.Good idea? Bad Idea? PEP?
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com>
Message-ID: <3E8AFB77.38D0EDDF@engcorp.com>

WP wrote:
> 
> Greg Ewing wrote the following Pyrex snippet:
> >     for i from 0 <= i < n:
> >       result.append(a[i])
> 
> This pyrex snippet reminds me (because of its orthogonality to C) of one
> thing, more than tab indents, that I have always thought was bogus in Python,
> which is the handling of integer loops.  range(0,n) sucks, xrange(0,n) is an
> interesting hack which obscures the readability of code, and post 2.x versions
> of python, with generators, and other 'inverted' ways of accomplishing the
> obvious procedural effect of an integer for loop, are all just ways of
> obfuscating otherwise perfectly readable python code. Okay, so I'm being
> opinionated. I know that. But does anyone else thing that for the extremely
> common (and therefore bytecode optimizeable, syntax customizeable) case where
> you need to do a for loop over a range (possibly large range) of integers,
> that a language level keyword construct (such as the above) would be
> preferable 

I just scanned a few tens of thousands of lines of our code, and as near
as I can tell only 13 cases out of over 160 instances of "for" loops are
using range() or xrange().  This backs up my personal feeling that in
Pythonic code, it's actually rare to iterate over an integer range 
specified with simple upper and lower bounds like that, and certainly not
common enough to justify special syntax.

I can't agree with the desire to make this a special case, when range()
provides such a simple, clean answer.

-Peter



From tj_scarlet at yahoo.com  Wed Apr 23 18:30:18 2003
From: tj_scarlet at yahoo.com (Tj)
Date: 23 Apr 2003 15:30:18 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <1azna.115402$It5.22896@news2.central.cox.net> <tyfhe8xqhmy.fsf@pcepsft001.cern.ch> <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com> <b86lc3$tnd$1@plover.com>
Message-ID: <ccc7084.0304231430.6b81effb@posting.google.com>

mjd at plover.com (Mark Jason Dominus) wrote in message news:<b86lc3$tnd$1 at plover.com>...
> I don't believe that's correct.  At an early Open Source Conference,
> perhaps around 1998 or so, I was present at a conversation between
> Guido and Tom Christiansen.  Tom asked Guido why Python didn't have
> closures, and Guido's reply (and I think this is an exact quote) was
> "Closures aren't important."
> 
> I remember it because I was so astonished by the obtuseness of the answer.

Well, it seems the burden of proof is on you.  Please explain the
usability advantages of closures vs. Python's current scope system.

Tj



From mustafa.celik at sympatico.ca  Fri Apr 11 11:07:15 2003
From: mustafa.celik at sympatico.ca (Mustafa Celik)
Date: Fri, 11 Apr 2003 11:07:15 -0400
Subject: collecting data from file
References: <3E9671A3.2050004@sympatico.ca> <3E96B6BC.E9DBE74F@engcorp.com>
Message-ID: <3E96DA23.6070500@sympatico.ca>

The file that I'm analyzing guarantees to have >4 columns, and lines 
contain string, integer, float..., they are seperated by spaces.

An example line is as below:

  1850099.32 HELLO 0xfce 6 OTTAWA stree_number_200 exit_Metcalfe 
take_hw_417 arrive_at_Ottawa_Airport

2nd column will tell me if someone has arrive, 3rd column will tell me 
how much $ they'll bring with them, and 4th column will tell who he is 
(a hex guy).

I will find out;
   * how many people arrived and did not leave,
   * how many have arrived and left
   * how much we have in Ottawa at the end
   * how much transient money flow have occured

My file may contain some lines that only leaves Ottawa, so I should 
ignore them (they probably arrived earlier, and should not decrement my 
dollar gain)  - This condition is not very important, if it'll 
complicate things, I have other ways to prevent this.

I found a good awk document  last night (can't remember the web site), I 
was planning to dig in it today, but you guys have supported me with my 
initial plan with Python. I'm not a beginner in Python, this is probably 
not the best project to start with, but I need to have this done by Awk 
or Python, ...

Thanks,
   Mustafa





Peter Hansen wrote:
> Mustafa Celik wrote:
> 
>>I want to scroll thru a file;
>>   * find lines that match a string (e.g. HELLO) on the 2nd column
>>   * add up the 4th column (an integer) on each matching line, say the
>>variable is TOTAL
>>   * subtract the 4th column from TOTAL if another is string (e.g. BYE)
>>is hit on 2nd column of a line
>>
>>Any tips?
> 
> 
> Yes, post an example showing what you intend.  The instructions are
> decent, but implementable only by making a few assumptions about what
> you mean.  Also, are there any error conditions to be handled, or do
> you guarantee the input is always perfect (e.g., always at least four
> columns, and so on)?
> 
> -Peter




From dave at pythonapocrypha.com  Wed Apr  2 16:16:05 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Wed, 2 Apr 2003 13:16:05 -0800 (PST)
Subject: Comments
In-Reply-To: <b6fc6e$1fe$1@panix2.panix.com>
References: <435d40e0.0304010014.503f9f8f@posting.google.com>
 <435d40e0.0304011457.56e32bd0@posting.google.com> <sKSdnfwAUb0mrxejXTWcog@comcast.com>
 <435d40e0.0304021052.736293b3@posting.google.com> <b6fc6e$1fe$1@panix2.panix.com>
Message-ID: <Pine.LNX.4.51.0304021312170.17774@ns0.tbye.com>

On Wed, 2 Apr 2003, Roy Smith wrote:

> >>> What are the reasons for not having multi-line comments?
>
> Multi-line comments are evil.  I grep up on C's /* */ comments, so
> this is not a case of "you like what you grew up with".
>
> The first time you spend a half-hour trying to understand a block of
> ugly code and finally get to the end and find a "*/", you'll
> understand why they're evil.
>
> The first time you accidentally type "* /" instead of "*/" and wonder
> why your program still compiles but is suddenly doing strange stuff,
> you'll understand why they're evil.

Roy, may I interest you in a syntax-highlighting editor? <0.5 wink>

Seriously though, those problems you describe have pretty much been
eradicated by modern editors because the color of the text just screams
"this has been commented out!!".

-Dave




From LogiplexSoftware at earthlink.net  Tue Apr 22 13:13:55 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 22 Apr 2003 10:13:55 -0700
Subject: Twisted 1.0.4
In-Reply-To: <20030422121836.GA8567@frobozz.local>
References: <20030421110438.4991f713.itamar@itamarst.org>
	 <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local>
	 <341410000.1051007646@leeloo.intern.geht.de>
	 <20030422121836.GA8567@frobozz.local>
Message-ID: <1051031635.18957.24.camel@software1.logiplex.internal>

On Tue, 2003-04-22 at 05:18, Andrew Bennetts wrote:
> On Tue, Apr 22, 2003 at 12:34:06PM +0200, Marc Recht wrote:
> > >to Twisted and Twisted's users.  Similarly, Twisted has some C extensions
> > >which are typically built with gcc, but gcc's licence (GPL) also has no
> > >relevance to Twisted.
> > IIRC then gcc libs are LGPL'ed, so it doesn't affect Twisted. I wouldn't be 
> > so sure if they were GPL'ed.
> 
> No, gcc is GPLed.  I'm not sure what you mean by "gcc libs"?  I'm pretty
> sure libgcc1 is part of gcc and thus GPLed also.

No, it's LGPL'd.  GCC itself is GPL, the libraries are LGPL'd.

> The point is that a C program or library compiled with gcc is not a derived
> work of gcc, anymore than a thesis written in and printed with MS Word is a
> derived work of MS Word.  Python isn't GPLed, despite often being compiled
> with gcc.

You are forgetting that by compiling your source with gcc, you are most
likely linking against LGPL'd libraries (even in rather trivial
programs).  That means that your executable will *contain* LGPL'd code. 
The LGPL was created just for this purpose: to allow non-GPL'd software
to link against (L)GPL'd software.


-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From mwh at python.net  Thu Apr 10 10:53:10 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 10 Apr 2003 14:53:10 GMT
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <7h3u1d6z9nt.fsf@pc150.maths.bris.ac.uk>

Alex Martelli <aleaxit at yahoo.com> writes:

> I _know_ there's a decent algorithm to solve the following problem,
> but I don't recall its name and thus can't rapidly google for the 
> details... can somebody help?

AdaBoost?  Not sure, but it's not totally off the map.

Cheers,
M.

-- 
  > I'll just wait for the community to discuss the proposal and 
  > agree on some way to count votes, then count them.
  The most recent estimates by astronomers is that the age of the 
  universe is essentially unbounded.        -- GvR and Martijn Faassen



From aleax at aleax.it  Sat Apr 12 10:57:28 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 14:57:28 GMT
Subject: distutils module
References: <b79863$dil$1@news-reader10.wanadoo.fr>
Message-ID: <sPVla.6373$T34.168242@news2.tin.it>

Salvatore wrote:

> Hello,
> 
> I've just installed Debian Knoppix,and i was surprised
> to see that distutils module could not be found.
> 
> Does anybody experienced the same problem ?

Not the specific one you mention, but I've often seen various
Linux distros package up Python in several and not necessarily
well-conceived sub-packages.  Personally, I always install Python
from sources -- I believe this reduces the amount of problems.


Alex




From martin at v.loewis.de  Tue Apr 29 17:02:13 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Tue, 29 Apr 2003 23:02:13 +0200
Subject: Determining Unicode encoding.
In-Reply-To: <ef7a350b.0304291242.42c9b024@posting.google.com>
References: <ef7a350b.0304291242.42c9b024@posting.google.com>
Message-ID: <b8mp8b$ttp$05$1@news.t-online.com>

Sean wrote:

> The problem is that I don't know how to determine what
> the *right* encoding to use on a particular string is.  

Do you have this problem when reading a byte string, or when
writing it?

If you are given a byte string, and you are supposed to interpret
the bytes as characters, there is, in general, no good way to do
so - that's why people came up with the idea of a universal character
set in the first place, to overcome the problems with multiple
character sets.

That said, you can make educated guesses on the data you read.

1. Perhaps the data you read has some file format which specifies
    the encoding, or allows parametrization, such as XML or HTML.
    You will need to look *into* the file to find out what its
    encoding is.

2. Perhaps the data has some fixed encoding, as part of the file
    format specification. For many files, this is US-ASCII.

3. Perhaps this is a plain text file, and you should use the encoding
    that the user's text editor is most likely to use (of course, you
    don't know what text editor the user uses, nor what encoding that
    editor uses). locale.getdefaultlocale()[1] offers you some guess;
    python 2.3's locale.getpreferredencoding() gives a better guess.

> Is there any way to
> determine, from the unicode string itself, what encoding I need to use
> to prevent data loss?

That sounds you have the problem when *writing* Unicode strings.

In that case, you can invoke .encode: it will give a UnicodeError if
the encoding is not supported. At some point, you need to make up your
mind what encoding to use for a certain file - if you then get an error,
all you can do is to inform the user, and

a) perhaps ignore the bad characters, replacing them with appropriate
    replacement characters (usually '?'), or

b) go back and recode the output so far in a different encoding.


> Am I even asking the right questions?  I'm really pretty lost and my
> O'Reilly books arn't helping very much.

Don't worry. These things are inherently difficult. Organizations like
W3C have essentially given up, and say that XML is UTF-8 by default
(knowing that this will support arbitrary characters). If people 
absolutely want XML in different encodings, they can do that, but they
are left alone with the issue of encoding unsupported characters
(for XML, they can actually use character references).

You will have to make explicit choices: either support only UTF-8
(and accept that it will be tedious for some users to produce the proper
files), or support arbitrary encodings (and accept that some encodings
cannot represent all characters, and that you may not have the codecs
available to read the data, and that a mechanism must be provided to
determine the encooding), or support only a few non-UTF-8 encodings
(restricting the data format to a subset of all living languages).

Regards,
Martin




From roy at panix.com  Fri Apr 18 21:30:31 2003
From: roy at panix.com (Roy Smith)
Date: Fri, 18 Apr 2003 21:30:31 -0400
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <b7q72i$nvh$1@slb6.atl.mindspring.net>
Message-ID: <roy-416201.21303118042003@reader1.panix.com>

"Andrew Dalke" <adalke at mindspring.com> wrote:
> One thing to note, which Dijkstra didn't predict, is that I stopped
> using a 'traditional' BASIC back in '86, when I got a copy of
> QuickBasic for my birthday.  No line numbers, real functions,
> better control structures, compiled code - while editing in the
> IDE, so the compile-link-run cycle is really fast.

I agree that all those things are good things, but by the time you've 
made that many changes to the language, what's left that it can still 
call itself BASIC?



From aahz at pythoncraft.com  Sat Apr  5 10:25:38 2003
From: aahz at pythoncraft.com (Aahz)
Date: 5 Apr 2003 10:25:38 -0500
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <b6kfjv$4cc$1@panix2.panix.com> <mailman.1049537111.4041.python-list@python.org>
Message-ID: <b6mshi$190$1@panix1.panix.com>

In article <mailman.1049537111.4041.python-list at python.org>,
Bo M. Maryniuck <b.maryniuk at forbis.lt> wrote:
>On Friday 04 April 2003 19:32, Aahz wrote:
>>
>> Probably the "standard" answers these days for the basic building block
>> of a web application is either Zope (if you want a full-blown framework
>> for content management) 
>
>Zope ain't just a "full-blown framework for content management". This what 
>CMF-like or KONTENTOR-like products does. Zope is an application server, 
>where you can build anything for web you want, even intergrate it with J2EE 
>via SOAP/XML-RPC.

While true, my impression has been that most people choose solutions
other than Zope unless they need at least some content management.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From aleax at aleax.it  Tue Apr 15 08:33:54 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 12:33:54 GMT
Subject: files: direct access - NOT sequential
References: <mailman.1050398560.8881.python-list@python.org>
Message-ID: <S_Sma.22945$T34.690529@news2.tin.it>

Patrick Carabin wrote:

> i search a means to get " direct access " inside files, eg mixing
> read & write operaions for the same file
> example:
> given a file with contents : "qwerty" , positionning on the "w", and
> writing a "N" , so the file becomes "qNerty"
> How can i achieve this in Python ?

thefile = open('thefile', 'r+')    # must be already-existing
thefile.seek(1)
thefile.write('N')
thefile.close()

Here, in fact, we're NOT mixing read and write operations -- the
only _operation_ is a write -- but, if we wanted to, we might (that
is what mode 'r+' means).


Alex




From nhodgson at bigpond.net.au  Wed Apr 23 05:10:42 2003
From: nhodgson at bigpond.net.au (Neil Hodgson)
Date: Wed, 23 Apr 2003 09:10:42 GMT
Subject: Bug? Pythonwin editor trouble
References: <3ea58721.1845828@news.muenster.de>
Message-ID: <mMspa.7849$8K2.70671@news-server.bigpond.net.au>

Martin Bless:

> (2) Problem:
> The editor causes trouble once you start using german "Umlaute".
> When you type (each line here representing a keypress and what you get
> and a comment):
>
> " -> "     # ok
> ? -> "?     # ok
> " -> "?     # " is missing

   OK, this was fixed in a more recent (1.48 or later) version of Scintilla.
Eventually this will be bundled with Pythonwin.

> (3) Can somebody confirm the error?

   Yes.

> (4) What can be done about it?

   The simplest option would be to use another editor until Pythonwin is
updated, perhaps using Pythonwin just for debugging.

   Neil





From aahz at pythoncraft.com  Tue Apr 15 00:40:20 2003
From: aahz at pythoncraft.com (Aahz)
Date: 15 Apr 2003 00:40:20 -0400
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1049901964.20840.clpa-moderators@python.org> <b72fo5$bq6$1@panix1.panix.com> <87ptnuxryk.fsf@theasylum.dyndns.org>
Message-ID: <b7g2fk$g1u$1@panix3.panix.com>

In article <87ptnuxryk.fsf at theasylum.dyndns.org>,
Tim Lavoie  <tlavoie at acm.org> wrote:
>>>>>> "aahz" == aahz  <aahz at pythoncraft.com> writes:
>
>    aahz> So how does Capon differ from SCons?
>
>Well, the name has funnier connotations, for one.   :)

You deserve to be roasted for that comment.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From duncan at NOSPAMrcp.co.uk  Thu Apr 17 04:52:47 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 17 Apr 2003 08:52:47 +0000 (UTC)
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com> <TUhna.928$cq2.100423716@newssvr13.news.prodigy.com> <f7199550.0304161720.702ac0f6@posting.google.com>
Message-ID: <Xns936062E30ED4Bduncanrcpcouk@127.0.0.1>

pc451 at yahoo.com (Peter Clark) wrote in 
news:f7199550.0304161720.702ac0f6 at posting.google.com:

>     w += [chr(176) + scale.strip()]
> UnicodeError: ASCII decoding error: ordinal not in range(128)
> 
>     I've tried changing its encoding to 'utf-8' as well with
> chr(176).encode('utf-8'), but that just returns the same error. This
> doesn't make sense: if 'print chr(176)' works fine, why doesn't it
> work later on? In a nut shell, all I want to do is add the degree
> character to a string contained within a list. That doesn't seem hard
> at all, but I'm completely stumped.
> 

A latin1 string containing chr(176) is a degree symbol. This string is 
already encoded (as latin-1), but Python doesn't know that until you tell 
it. If you were to do 'print chr(176)' in a command window on the machine 
I'm typing on, you get a funny graphic character because the output is 
using codepage 437 not latin1.
 
If you want to change its encoding, e.g. to utf-8, you must first DECODE it 
from latin1 (giving you a unicode string) and then ENCODE it to utf8.

>>> chr(176).decode('latin1').encode('utf8')
'\xc2\xb0'

(Or in my case I can do "print chr(176).decode('latin1').encode('cp437')" 
and see a degree symbol.)

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From mlichten at harris.com  Fri Apr 25 12:25:00 2003
From: mlichten at harris.com (Lichtenstein, Mark (Contractor))
Date: Fri, 25 Apr 2003 12:25:00 -0400
Subject: Getting values from python interpreter from Java
Message-ID: <EFAA78F77E74D411A0F700508BDFEAD607D3E897@mlbmx9.corp.harris.com>

I would like to see how to get the value of a local variable from Java.
In Java, I am saving the values of the local dictionary in a PyObject Java
object. 
Using a PythonInterpreter object called "interp", I use setLocals and getLocals
methods to 
get and restore the local dictionary. 

interp.setLocals(localsObject);
interp.getLocals(localsObject); 

My question is from Java how does someone go into the PyObject contatining the
locals dictionary to get the value of a local dictionary variable?


Thank you,

Mark Lichtenstein





From piet at cs.uu.nl  Wed Apr  2 09:59:00 2003
From: piet at cs.uu.nl (Piet van Oostrum)
Date: 02 Apr 2003 16:59:00 +0200
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com>
Message-ID: <wzel4l3pzf.fsf@nono.cs.uu.nl>

>>>>> sdieselil at yahoo.com (sdieselil) (s) wrote:

s> Hi
s> How can I implement double-linked lists and other complex tree-like
s> structures in Python?

use a class with a prev and a next attribute. Or parent, child and so on
for trees.
-- 
Piet van Oostrum <piet at cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.van.Oostrum at hccnet.nl



From peter at norvig.com  Tue Apr 15 04:23:07 2003
From: peter at norvig.com (Peter Norvig)
Date: 15 Apr 2003 01:23:07 -0700
Subject: Palindrome finder: help find ways to speed up?
References: <080420031230508991%camel@oasis.com>
Message-ID: <da69ff6a.0304150023.6a019cc3@posting.google.com>

http://www.norvig.com/pal-alg.html explains how you can generate 200
word palindromes in 5 seconds, and 10,000 word palindromes in about an
hour.

The code follows:

import string, random, os, re, bisect

"""Produce Panama-ish Palindromes. Copyright (C) 2002, Peter Norvig.
See http://www.norvig.com/license.html and
http://www.norvig.com/pal-alg.html"""

def is_panama(p):
    "Test if p is a Panama-ish palindrome."
    def is_unique(seq): return len(seq) == len(dict(zip(seq, seq)))
    return (p.endswith('Panama') and is_palindrome(p)
            and is_unique([s.strip() for s in p.split(',')]))

def is_palindrome(phrase):
    "Test if a phrase is a palindrome."
    cphrase = canonical(phrase)
    return cphrase == reverse(cphrase)

def canonical(word, sub=re.compile('[^A-Za-z0-9]').sub):
    "The canonical form for comparing: lowercase alphanumerics."
    return sub('', word).lower()

def read_dict(filename='npdict.txt'):
    "Read the file into global variables _fw and _bw and _truename."
    global _fw, _bw, _truename
    _fw, _bw, _truename = [], [], {'': ''}
    for word in open(filename).read().splitlines():
        w = canonical(word)
        _fw.append(w)
        _bw.append(reverse(w))
        _truename[w] = word
    _fw.sort(); _bw.sort()
    return len(_fw), len(_bw), len(_truename)

def update(obj, **entries): obj.__dict__.update(entries); return obj

class PalDict:
    """A dictionary from which you can find canonical words that start
or end
    with a given canonical substring, and find the true name of a
    canonical word."""
    def __init__(self, fw=None, bw=None, truename=None):
        update(self, fw=fw or _fw, bw=bw or _bw, truename=truename or
_truename)
    def startswith(self, prefix, k=100):
        """Return up to k canonical words that start with prefix.
        If there are more than k, choose from them at random."""
        return k_startingwith(k, self.fw, prefix)
    def endswith(self, suffix, k=100):
        """Return up to k canonical words that end with suffix.
        If there are more than k, choose from them at random.
        Both the suffix and the word returned are reversed."""
        return k_startingwith(k, self.bw, suffix)

def k_startingwith(k, words, prefix):
    """Choose up to k words that match the prefix (choose randomly if
> k)."""
    start = bisect.bisect(words, prefix)
    end = bisect.bisect(words, prefix + 'zzzz')
    n = end - start
    if k >= n:
        results = words[start:end]
        random.shuffle(results)
    else: # Should really try to avoid duplicates
        results = [words[random.randrange(start, end)] for i in
range(k)]
    return results

class Panama:
    def __init__(self, L='A man, a plan', R='a canal, Panama',
dict=None):
        left = [canonical(w) for w in L.split(', ')]
        right = [canonical(reverse(w)) for w in reverse(R.split(',
'))]
        update(self, left=left, right=right, dict=dict or PalDict(),
best=0,
               seen={}, diff=len(''.join(left)) - len(''.join(right)))
        for word in left + map(reverse, right):
            self.seen[word] = 1
    def missing(self, k=20):
        """Return the substring that is missing, and candidate
words."""
        if self.diff >= 0: # Left is longer, missing on right
            substr =  self.left[-1][-self.diff:]
            return substr, self.dict.endswith(substr, k)
        else: # Right is longer, missing on left
            substr =  self.right[-1][self.diff:]
            return substr, self.dict.startswith(substr, k)
    def search(self, k=200):
        "Search for palindromes; consider at most k words at each
level."
        self.stack = [self.missing(k)]
        while self.stack:
            substr, words = self.stack[-1]
            if is_palindrome(substr):
                self.report()
            if words:
                self.extend(words.pop(), k)
            elif not self.backtrack():
                return
    def extend(self, word, k):
        "Add a new word (unless we've already seen it)."
        if self.diff >= 0: # Left is longer, add to right
            fword = reverse(word)
            if fword in self.seen: return
            self.diff -= len(fword)
            self.seen[fword] = 1
            self.right.append(word)
            self.stack.append(self.missing(k))
        else: # Right is longer, add to left
            if word in self.seen: return
            self.diff += len(word)
            self.seen[word] = 1
            self.left.append(word)
            self.stack.append(self.missing(k))
    def backtrack(self):
        "Remove the last word added; return 0 if can't backtrack"
        if self.diff >= 0: # Left is longer, pop from left
            if not self.left: return 0
            word = self.left.pop()
            self.diff -= len(word)
            del self.seen[word]
        else: # Right is longer, pop from right
            if not self.right: return 0
            word = self.right.pop()
            self.diff += len(word)
            del self.seen[reverse(word)]
        self.stack.pop()
        return 1
    def report(self):
        "Write current state to log file."
        if len(self) > self.best + 200:
            self.best = len(self)
            print self.best
            self.bestphrase = str(self)
            assert is_panama(self.bestphrase)
            f = open('pallog%d.txt' % os.getpid(), 'w')
            f.write(self.bestphrase + '\n')
            f.close()
    def __len__(self):
        return len(self.left) + len(self.right)
    def __str__(self):
        truename = self.dict.truename
        lefts = [truename[w] for w in self.left]
        rights = [truename[reverse(w)] for w in
reverse(self.right[:])]
        return ', '.join(lefts + ['*****'] + rights)

def reverse(x):
    "Reverse a list or string."
    if type(x) == type(''):
        return ''.join(reverse(list(x)))
    else:
        x.reverse()
        return x

if __name__ == '__main__': read_dict(); p = Panama(); p.search()

dromedary <camel at oasis.com> wrote in message news:<080420031230508991%camel at oasis.com>...
> I've written a palidrome finder two ways (see below), ...



From ramen at lackingtalent.com  Thu Apr 24 12:27:19 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Thu, 24 Apr 2003 16:27:19 -0000
Subject: methods and vars in zope objects
References: <c42e062e.0304211805.4ced1175@posting.google.com>
Message-ID: <slrnbag44a.4e4.ramen@lackingtalent.com>

In article <c42e062e.0304211805.4ced1175 at posting.google.com>,
Erik Lechak wrote:
> Is there a simple script, document, manual, or button that will list
> the methods and member vars available to a certain object?  Something
> like dir().

How about "dir()"? =)

Try this:
 1. Create a file in your Extensions directory called "dir.py".
 2. In this file, write the single line:
    dir_ = lambda x: dir(x)
 3. Create an External Method:
    id: dir
    module: dir
    function: dir_
 4. Use context.dir(obj) from your PythonScript.

In general, this technique works for exposing Python library functions to
the Zope environment. Mind the associated security risks, of course.

Peace,
Dave



From jane.doe at acme.com  Sat Apr  5 12:47:05 2003
From: jane.doe at acme.com (Jane Doe)
Date: Sat, 05 Apr 2003 19:47:05 +0200
Subject: Why use a GUI framework?
Message-ID: <1g5u8vc1k8t97jg8vsm2gi81sfnu7cm1hf@4ax.com>

Hi,

	I know a bit more about the multiple options available to
build a GUI-based application (Windows, in our case), but I haven't
found information on the point of using another tool since the
win32all extension is available to use the native Win32 widgets. 

Do wxWindows et al. just make it easier, or is it a must-use for
people who wish to write OS-independant Python applications?

If win32all is good enough, does anybody have a good tutorial on using
this add-on to build GUIs?

Thx much
JD.



From mwh at python.net  Tue Apr 15 06:02:46 2003
From: mwh at python.net (Michael Hudson)
Date: Tue, 15 Apr 2003 10:02:46 GMT
Subject: Complexity of standard Python data structures
References: <mailman.1050343651.17799.python-list@python.org> <slrnb9m2qj.9r2.marcus@infa.abo.fi>
Message-ID: <7h3fzokrsca.fsf@pc150.maths.bris.ac.uk>

marcus at infa.abo.fi (Marcus Alanen) writes:

> >Note that it's not true that Python's list is really a vector/array -- the
> >language doesn't define its implementation, and doesn't want to.  It is true
> 
> Hmm. So the Python language is more about specifying _what_ will
> happen, instead of _how_, with no concern for actual speed or memory?

I'm afraid your line of questioning slightly confuses me.  The Python
language is not specified, in any precise sense of the word, anywhere.
For one thing, it's evolving.  If some super new implementation of
hash tables appeared[1] that had somewhat different algorithmic
behaviour but were considered superior, they might well turn up in
Python 2.5 (say) and *if* the closest thing Python has to a spec (the
language reference) described the algorithmic nature of various
operations (which it doesn't) it would then be changed.

On the other hand, you can take it as read that the Python developers
will make a good faith effort to ensure that each new version of
Python is actually an improvement over its predecessors.

What kind of answers are you expecting to such a question?

Cheers,
M.

[1] Please ignore the unlikelyness of this...

-- 
  > say-hi-to-the-flying-pink-elephants-for-me-ly y'rs,
  No way, the flying pink elephants are carrying MACHINE GUNS!
  Aiiee!!  Time for a kinder, gentler hallucinogen...
                               -- Barry Warsaw & Greg Ward, python-dev



From cpl.19.ghum at spamgourmet.com  Mon Apr 28 04:39:10 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Mon, 28 Apr 2003 10:39:10 +0200
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <Xns936B6C09F44Fcpl19ghumspamgourmet@62.153.159.134>

Steven,

wouldn't be printing out "Dive into Python" maybe qualify as a textbook?

I'm not informed about the American high-school system, but working as a 
freelance trainer I definitly would consider to use this book.

Harald



From vlindberg at verio.net  Wed Apr  9 19:18:30 2003
From: vlindberg at verio.net (VanL)
Date: Wed, 09 Apr 2003 23:18:30 GMT
Subject: HTML -> HTMLGen (or similiar)
Message-ID: <aT1la.8930$8Z5.56436@sea-read.news.verio.net>

Hello,

I have a bunch of existing HTML that I need to integrate into a new web 
app.  I would ideally like to use one of the existing HTML generation 
systems (HTMLgen, HyperText, any others) but I don't want to have to 
build up an equivalent representation for the many pages I need to work 
with.

The pages validate as HTML 4.01, so no strange measures are necessary. 
However, does anyone know of a package that can take existing (valid) 
HTML and make it into an object that I can manipulate?

What I have in mind is something similar to the python 2.2+ email module:

import email
msg = email.message_from_string(email_message)
[modify here]
print str(message)  [Serialize out to text]


instead, an html module:

import html
page = html.page_from_string(html_page)
[modify here]
print str(page)


Thanks,

VanL




From martin at v.loewis.de  Sun Apr 13 12:21:29 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 18:21:29 +0200
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk>
Message-ID: <m38yuepdva.fsf@mira.informatik.hu-berlin.de>

Robin Becker <robin at jessikat.fsnet.co.uk> writes:

> What's the correct incantation to get a weak reference to a tuple? The
> rather obvious wrapper approach fails as the wrapper dies immediately
> 
> ie weakref.ref(Wrapper((....)))

What is 'Wrapper'?

Tuples are not weakly-referencable, unless you change the source code
of tupleobject.c.

Regards,
Martin



From juliagoolia301 at hotmail.com  Sun Apr 20 19:13:34 2003
From: juliagoolia301 at hotmail.com (Julia Goolia)
Date: 20 Apr 2003 16:13:34 -0700
Subject: python dynamic scoping question
Message-ID: <79ad5955.0304201513.6ffd4f83@posting.google.com>

Hi,

I have a question regarding scoping in python.  I wish to have some
mechanism which puts variables at a function level scope into the
global scope, dynamically.  I cannot alter my design.

For example, I want this to work:

def f():
  def g():
    x = 5
  g()
  print x
f()

I want to be able to see x in f, event though it was assigned in a
lower scope, g.

Thank you for your time.



From andrew-pythonlist at puzzling.org  Tue Apr 22 10:08:47 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 23 Apr 2003 00:08:47 +1000
Subject: coverting list back to string
In-Reply-To: <3cf72d52.0304211132.799d8afb@posting.google.com>
References: <3cf72d52.0304211132.799d8afb@posting.google.com>
Message-ID: <20030422140847.GB8567@frobozz.local>

On Mon, Apr 21, 2003 at 12:32:11PM -0700, scn wrote:
> my code:
> 
> import xreadlines
> 
> input = open('c:/python22/programs/linetest.txt', 'r') # 3 lines of
> text
> 
> output = open('c:/python22/programs/output.txt', 'w')
> 
> for line in input.xreadlines():
>     remove = line.split()
>     remove[0:1] = []
>     output.write(remove)
> 
> output.close()
> input.close()
> 
> my quest:  i know that output.write(remove) is not correct.  you
> cannot use the 'write' method on a list.  is there a better way to
> convert the list to a string after the first element of the list has
> been removed and then output the strings to a file?  i'm a newbie
> programmer and need guru help. thank you.

Simply don't split more than you need to:

    # ...

    for line in input.xreadlines():
        # Split on whitespace a maximum of one time
        discard, rest = (line.split(None, 1) + [''])[:2]
        output.write(rest)
    
    # ...

-Andrew.





From donn at u.washington.edu  Fri Apr 25 19:38:05 2003
From: donn at u.washington.edu (Donn Cave)
Date: 25 Apr 2003 23:38:05 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <20030417205405.A10672@tibia.amotlpaa.bogus> <mailman.1051310385.1948.python-list@python.org>
Message-ID: <b8cgst$20ho$1@nntp6.u.washington.edu>

Quoth Steven Taschuk <staschuk at telusplanet.net>:
...
| I've submitted a documentation patch for the __str__ and __repr__
| docs, following ideas discussed in this thread:
|     <http://www.python.org/sf/727789>
| I solicit comments.  (The __repr__ blurb is much the same as that
| in the previous post; I've fiddled more extensively with the
| __str__ blurb.)

I still think that the normal usage of __str__ can be directly
described as a data conversion to type string.  I know you don't
find that completely illuminating on its own, but I think it does
convey an essential idea that is otherwise likely to be missed.
See for example the __str__ and __repr__ definitions in Cookie.py.
Since this apparently isn't as obvious to you as it is to me, you
may be in a better position to explain it if I can convince you
that there's something to it.  If we can decide on some central
notion behind normal __str__ usage, the examples will make more
sense and it won't be necessary to describe __str__ as "not __repr__".

	Donn Cave, donn at u.washington.edu
----------------------------------------
| __str__(self)
|
|     Called by the str() built-in function and by the print statement
|     to compute a string representation of an object.  The return value
|     must be a string object.
|
|     This differs from __repr__() in that it does not specifically
|     serve the purpose of providing a programmer with detailed
|     information about the object; it may provide a string
|     representation of the object for other purposes, as appropriate to
|     the object.  Such purposes include display to an application
|     end-user, reference in an application log for administrators, or
|     providing a string to other code which requires one -- in general,
|     purposes for which the programmer-oriented information of
|     __repr__() is not appropriate.



From sidlonDoesntLikeSpam at yahoo.com  Tue Apr 29 19:25:48 2003
From: sidlonDoesntLikeSpam at yahoo.com (Michael McFarland)
Date: Tue, 29 Apr 2003 23:25:48 GMT
Subject: weird behavior in 'for line in sys.stdin'
Message-ID: <0SDra.65118$xB4.1544318@news2.east.cox.net>

   Here's basic three ways of printing lines from standard input.  If
standard input is coming from a file or other program, all three behave
properly.  When run interactively from a bash prompt, however, there are
three different behaviors.  Method #3's behavior seems especially odd.  Is
there a bug here?

METHOD #1:
while 1:
  line = sys.stdin.readline()
  if not line:
    break
  line = line.rstrip()
  print "LINE: ", line

METHOD #2:
for line in sys.stdin.readlines():
   line = line.rstrip()
   print "LINE: ", line

METHOD #3:
for line in sys.stdin:    [or sys.stdin.xreadlines()]
   line = line.rstrip()
   print "LINE: ", line


---
Here's what happens when I type 'a', 'b', 'c', then 'ctrl-d' interactively
with each:

  METHOD #1:
a
LINE:  a
b
LINE:  b
c
LINE:  c
$>

  METHOD #2:
a
b
c
LINE:  a   <printed after crtl-D>
LINE:  b
LINE:  c
$>

  METHOD #3:
a
b
c
LINE:  a   <printed after crl-D>
LINE:  b
LINE:  c
<doesn't return to shell>
d
e
f
LINE:  d   <printed after a 2nd ctrl-D>
LINE:  e
LINE:  f
< and so on, until ctrl-D is hit twice in succession >
$>


   I can see why method #2 will obviously not print anything until
"readlines()" finishes.  But I would generally have expected method #3 with
its lazy evaluation to work like #1.  Can anyone shed some light on what's
really going on w/ 'for line in sys.stdin'?

  - Michael





From noamr at users.sourceforge.org.replace.org.with.net  Thu Apr  3 18:08:15 2003
From: noamr at users.sourceforge.org.replace.org.with.net (Noam Raphael)
Date: Fri, 04 Apr 2003 01:08:15 +0200
Subject: state bits of an event  - on unix and mac
Message-ID: <b6ibad$376$1@news.iucc.ac.il>

Hello,
   I'm writing in Python a wrapper over Tkinter widgets, that will, for 
each event, call all appropriate functions and not just the most 
specific one.
   For this I need to interpret the state variable which is generated 
for each event. The problem is that I've seen no documentation on the 
meaning of the bits of event, and I want my class to work on unix, win 
and mac, so finding it from my experience only won't be too good.
   I think that the mask 2**0 (the least significant bit) is always the 
Shift modifier, 2**1 is always Lock, 2**2 is always Control, 2**8 is 
always B1 ... 2**12 is always B5. The problem is with the Alt and Meta 
modifiers. From what I saw, on Windows Alt is 2**17, but for unix it 
depends on the X modifier map. Do you know how can the script find which 
out of mod1..mod5 is Alt and which is Meta (if there is a Meta)? I would 
prefer using a tcl/python command over calling an external program. 
Also, can you tell me what happens with modifiers on Mac? I don't have 
one, so I have no clue.

Thank you,
Noam Raphael




From abpillai at lycos.com  Mon Apr  7 06:57:03 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 7 Apr 2003 03:57:03 -0700
Subject: I'm new to programming
References: <oprm4ogjb9acxezr@news.datafast.net.au> <mailman.1049671991.8387.python-list@python.org>
Message-ID: <bd993a2f.0304070257.4c3468ec@posting.google.com>

Why not Tinker with existing Apps ? 
Find some bugs and fix them.

Try vaults of Parnassus.
http://www.vex.net/parnassus/

Anand 


"Raseliarison nirinA" <nirina at mail.blueline.mg> wrote in message news:<mailman.1049671991.8387.python-list at python.org>...
> ----- Original Message ----- 
> From: "Casey Jones" <cajones at datafast.net.au>
> Newsgroups: comp.lang.python
> To: <python-list at python.org>
> Sent: Saturday, April 05, 2003 3:20 AM
> Subject: I'm new to programming
> 
> > Inspiration would be appreciated. 
> 
> follow the thread 'need help with game'
> i've posted some food for your imagination.
> 
> nirinA



From john.abel at pa.press.net  Sat Apr  5 04:49:38 2003
From: john.abel at pa.press.net (John Abel)
Date: Sat, 5 Apr 2003 10:49:38 +0100
Subject: segfault with imp.load_module()
References: <1049534591.20683.22.camel@musicbox>
Message-ID: <000701c2fb58$c8065ff0$8201fd0a@hallows>

Could this not have to do with your use of file as a variable?

Regards

John
----- Original Message ----- 
From: "Rob Brown-Bayliss" <rob at zoism.org>
To: "Python List" <python-list at python.org>
Sent: Saturday, April 05, 2003 10:22 AM
Subject: segfault with imp.load_module()


> 
> Hi, I can do this in interactive python:
> 
> >>> f,n,d = imp.find_module("oda",["./items/"])
> >>> print f,n,d
> <open file './items/oda.py', mode 'r' at 0x82c6db8> ./items/oda.py
> ('.py', 'r', 1)
> >>> it = imp.load_module("oda",f,n,d)
> >>> print it
> <module 'oda' from './items/oda.pyc'>
> 
> but this code:
> file, filename, description = imp.find_module(item, ["./items/"])
> print file, filename, description
> item_module = imp.load_module(item,file, filename, description)
> file.close()
> 
> gets this:
> 
> [rob at musicbox BHH-V]$ python utils.py
> <open file './items/oda.py', mode 'r' at 0x81c2af8> ./items/oda.py
> ('.py', 'r', 1)
> Segmentation fault
> 
> 
> -- 
> 
> *  Rob Brown-Bayliss
> *  ================= 
> *      zoism.org
> 
> 
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 




From matt at teaphoo.fsnet.co.uk  Tue Apr  8 19:27:09 2003
From: matt at teaphoo.fsnet.co.uk (Matthew Russell)
Date: Wed, 9 Apr 2003 00:27:09 +0100
Subject: usage of assert and -O
Message-ID: <005e01c2fe26$608868c0$0ba587d9@jaws>

Hi all,

I understand that when you invoke the interpreter with the -O flag assert statments are not executed - is this 100% correct? (as i belive it to be)

and if so, shouldn't the use of assert (if not assert itself) be depricated?

I note that in unittest.py assert is not use,
we have failIf or assert_ instead - thus if you avoided using assert in your own tests you can saftly run all your unittests optimised.... 

..or can you...?

As some code in the standard library use assert,
and since most python users' applications use library modules.....

(sorry,  I cannot remmeber specificly what, but there are quite a few instances - asyncore.py is one i think)


I would be interested to know what others think about usage of assert in context with optimised python code, in particular in releation to unittesting, as this is where i think it would be useful - if not critical to have one way of doing it (i.e not useing assert or -O [ one or the other ])

Matt Russell


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030409/e92cbb23/attachment.html>

From johnroth at ameritech.net  Wed Apr 16 10:11:59 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 16 Apr 2003 10:11:59 -0400
Subject: A better unittest
References: <3ckier9a.fsf@python.net>
Message-ID: <v9qot54nmk87cb@news.supernews.com>

"Thomas Heller" <theller at python.net> wrote in message
news:3ckier9a.fsf at python.net...
> I'm moving more and more to TDD (test driven development [1]), which
> makes tests an integral part of the development process.  Failing
tests
> are ocurring very often with this development style, because you write
> the tests first and then implement the functionality.
>
> Having said this, I'm not really happy with the output of
> unittest. Consider the following testcase:

[snip]

> What do people think? Is this a useful change, and should I submit a
> patch?

I don't think I'd find it useful for two reasons. One is that it doesn't
get rid of enough clutter; removing the stack trace cuts the output
down by maybe 50%, and leaves in a good deal of poorly formatted
visual clutter.

The other reason is that it's really hard to figure out when the error
is expected so the stack trace is redundant, and when it's not expected
so that the stack trace includes useful information.

In any case, I really only expect *one* test to fail at any point
in time, and I suppose that I'm of the "more information is better"
school of thought. If more than one test fails, I suspect that I've
done something really stupid on the last change.

-1

John Roth

>
> Thomas
>
> [1] Kent Beck: Test Driven Development: by example. Addison Wesley,
2002.
>     ISBN 0-321-14652-0





From anton at vredegoor.doge.nl  Wed Apr 30 05:54:21 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Wed, 30 Apr 2003 11:54:21 +0200
Subject: Problem in scalling............
References: <20030430001533.3975.qmail@web41708.mail.yahoo.com> <mailman.1051665946.14826.python-list@python.org>
Message-ID: <b8o6kj$pnn$1@news.hccnet.nl>

Chad Netzer <cnetzer at mail.arc.nasa.gov> wrote:

>Lots of people, inclusing myself, have learned that the Tk Canvas sucks
>for doing things like you want to (anything requiring lots of scaling,
>rotating, etc.).  In particular, the kinds of transformations you can do
>to objects is limited, and those the canvas almost non-existant
>(translations only).

It's possible to define a virtual world and to use the Tk Canvas as a
viewport on that world. There are certain advantages in creating ones
own transformations but I don't think transparency is one of them.
However if interested in my explorations, here's a noddy (learn a new
term everyday ;-) file at:

http://home.hccnet.nl/a.vredegoor/cartesian

and here's a bigger project that is using the same principle
(a viewer for sgf-files for the game of go, sgf is a format for
storing game records):

http://home.hccnet.nl/a.vredegoor/pygobo/

I agree with Chad that if you want to conform to a standard -which is
good if someone else is going to use your code- it might be better to
try something else. Tk is in the standard distribution however, which
is good if you're aiming for end users to use your script.

Anton



From altis at semi-retired.com  Mon Apr 14 15:10:48 2003
From: altis at semi-retired.com (Kevin Altis)
Date: Mon, 14 Apr 2003 12:10:48 -0700
Subject: ANN: Python 11 - OSCON 2003 early bird registration deadline
Message-ID: <mGDma.144$yO.96884@news.uswest.net>

The deadline for Python 11 - OSCON 2003 early bird registration is May 23rd,
2003. The conference will be held July 7-11, 2003 in Portland, Oregon, USA.

  http://conferences.oreillynet.com/os2003/

  http://conferences.oreillynet.com/pub/w/23/register.html

Save up to $400 when you register by May 23rd for the Python 11/OSCON 2003
conference. Additional discounts are available.

I'm including the text of the special discounts available for the conference
below since they are quite significant; see the URL above for further info
and mailto links. All of the discounts listed at the bottom of the
registration page are in addition to the early bird discount. The early bird
discount is (obviously) time-limited, but the others are not.

Special Discounts
=================
Early Bird Pricing - Save up to $400 when you register by May 23, 2003.
SPECIAL TUTORIAL OFFER - Buy 2 tutorials, get 2 free (when you purchase a
full conference pass).

Full Time Student - 65% off with proof of status, a copy of ID & class
schedule. Fax to 707-829-1342. Please use os03st in discount field.

Academic Instructor - 50% off with proof of full time academic instructor
status on organization letterhead. Fax to 707-829-1342. Please use os03ac in
discount field.

Academic Staff - 25% off with proof of full time academic employee status on
organization letterhead. Fax to 707-829-1342. Please use os03sa in discount
field.

Company Team Discount - 10% off per person if you register 3 or more people
from one company. Please use os03team in discount field. Proof of status of
employment for each attendee must be faxed to 707-829-1342.

Federal Government Discount - 20% off with proof of agency connection. Use
discount code, os03gov, and fax to 707-829-1342.

User Group Members - UG discounts are available. For discount information or
to see if your group is eligible please email Marsee Henon for details. To
sign your group up with the O'Reilly User Group Program please go to
http://ug.oreilly.com/.

Alumni - Have you attended one of our O'Reilly conferences? If so, you are
eligible for a 30% alumni discount. Please email Linda Holder for
verification and your special discount code.

ka
---
Kevin Altis, Python 11 co-chair
altis at semi-retired.com
http://www.pycs.net/altis/
http://www.pythoncard.org/





From spam at magnetic-ink.dk  Thu Apr  3 02:31:47 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Thu, 3 Apr 2003 07:31:47 +0000 (UTC)
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org>
Message-ID: <20ce5436-bb06-4862-a55e-b5401610f413@news.szn.dk>

danny at iamlearning.com wrote:

> How can I compile this script to *.pyo?

See the py_compile and compileall modules.


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From ianb at colorstudy.com  Fri Apr 25 01:54:38 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 25 Apr 2003 00:54:38 -0500
Subject: Overloading and?  was <RE: Should I prefer an external
	database>
In-Reply-To: <b8afri$oqt$1@slb4.atl.mindspring.net>
References: <mailman.1051023275.29338.python-list@python.org>
	 <b8420j$629$1@slb3.atl.mindspring.net>
	 <mailman.1051043856.10118.python-list@python.org>
	 <b86gim$60r$1@slb5.atl.mindspring.net>
	 <mailman.1051127503.25664.python-list@python.org>
	 <b87il2$3rf$1@slb6.atl.mindspring.net>
	 <mailman.1051211247.5278.python-list@python.org>
	 <b8afri$oqt$1@slb4.atl.mindspring.net>
Message-ID: <1051250078.700.66.camel@lothlorien>

On Fri, 2003-04-25 at 00:11, Andrew Dalke wrote:
> Can that object be saved?  Eg, what does this do?
> 
> class Spam:
>     def __and__(self, other):
>       self.saved = other
>       return other()
> 
> spam = Spam()
> try:
>   a = spam and 1/0:
> except ZeroDivisionError:
>   pass
> spam.saved()

Arrr matey... there be continuations.  Dark waters indeed!

> For some languages, like Smalltalk and Ruby, this makes sense.
> Those languages support code blocks, which can be passed
> around and executed.  But Python does not.

I suspect that in Smalltalk you cannot actually overload and:.  It is a
quiet optimization that Smalltalks do, that turn particular methods into
special bytecode.  Things like ifTrue:ifFalse:, and to:do: are not
compiled to method calls, and I suspect the same is true of and:.  It's
just too inefficient.

I don't actually know how Ruby does and...?  Syntactically distinct like
Python, or as a method call like Smalltalk?  I expect more like
Python... either way I'm sure they all do the same thing deep down.

> Were Python to have code blocks, then it should be applied
> more generally to Python code.  Consider the if/else PEP of
> a couple months ago.  With code blocks, the following would
> be possible
> 
> ifelse(a != 0, 1/a, None)
> 
> which would pass the 1/a and None as independent code
> blocks, to be executed based on the evaluation of a != 0.

Of course, ifelse would be best implemented as a macro, not with code
blocks (which presumably would require some syntax, like 
ifelse(a != 0, `1/a`, `None`)).  Macros are of course compile-time, not
runtime, and Python deliberately does very little of interest during the
compile.  I think there are actually some serious problems introduced
because of its runtime nature, but nevertheless that's how Python is.

  Ian






From bellman at lysator.liu.se  Fri Apr 11 14:04:40 2003
From: bellman at lysator.liu.se (Thomas Bellman)
Date: Fri, 11 Apr 2003 18:04:40 +0000 (UTC)
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de> <mailman.1050021261.25010.python-list@python.org>
Message-ID: <b7703o$ejq$1@news.island.liu.se>

Chuck Swiger <cswiger at mac.com> wrote:

> If you plan on sending any significant amount of mail, it is recommended 
>   that your code talk SMTP to port 25 and pipeline all of the messages 
> you need to send in one continuous stream.  If you fork a sendmail 
> subprocess for each message you send, that will create a wastefully huge 
> amount of system load.

If you write your application to talk SMTP directly, instead of
using the proper Unix API for sending mail (i.e, calling
/usr/lib/sendmail), you have a number of problems.

First, which machine should you talk SMTP with?  Localhost?  Not
portable.  There are lots of Unix machines which do not have any
MTA running listening to port 25.  Every machine I administrate
run Sendmail in queue mode only (except for the mail servers, of
course).  Thus, you need your application to be configurable for
the mail server to use.  If the system changes, and another mail
server should be used, the sysadmin might not remember, or know,
that that particular application must be reconfigured.  It
becomes even worse if it is an application where each user on the
machine has their own configuration, like Netscape or Mozilla...

Second, you better be prepared to queue the mails in case of
network or mail server problems and retry sending them at a later
time.  For example, when the latest Sendmail hole was announced,
I turned the mail server at work off for more than 12 hours
waiting for a patch from RedHat (after that time I decided I
couldn't wait for RedHat any longer and made my own patched
RPM).

If you talk SMTP directly, you run the risk of skipping any local
policy for sending email, like masquerading as a specific domain
instead of as the FQDN of the local host, or a number of other
settings.

There are many subtleties involved in the game of being, or
pretending to be, an MTA, many of which are not handled by the
smtplib Python module, and many of them are subject to local
policy, decided by the system administrator.  Sendmail, Postfix
and Qmail are written to take care of those; you are unlikely to
do a better job when writing an application whose primary purpose
*isn't* being an MTA.

I know many other sysadmins that run their systems similarly to
the way I run mine.  Any application that talks SMTP on their own
instead of calling /usr/lib/sendmail to do it, is considered
broken by us...

So, when you write any program that sends email, at *least* give
the option of configuring it to use /usr/lib/sendmail, and
preferably to use that by default.  It is at least as likely to
work out of the box as any other setting is.

I wouldn't mind if there were some standard Python API for
sending email, that did the right thing on all platforms,
though.


-- 
Thomas Bellman,   Lysator Computer Club,   Link?ping University,  Sweden
"Adde parvum parvo magnus acervus erit"       ! bellman @ lysator.liu.se
          (From The Mythical Man-Month)       ! Make Love -- Nicht Wahr!



From alloydflanagan at attbi.com  Tue Apr 22 14:49:11 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 22 Apr 2003 11:49:11 -0700
Subject: Overloading and?  was <RE: Should I prefer an external database>
References: <mailman.1051023275.29338.python-list@python.org>
Message-ID: <a8b7f07a.0304221049.3e26a569@posting.google.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1051023275.29338.python-list at python.org>...
> > From: Ian Bicking [mailto:ianb at colorstudy.com] 
> [... + SQLObject example]
> 
> > for paper in Paper.select(AND(Paper.q.title == 'foo',
> >                               Paper.q.author.startswith('Bob'))):
> 
> Does anyone know the reason for not allowing an overload of the and
> operator? (of course, the only use case I've seen is the one above
> [thrice, first time in 97] and alternate logical calculi [e.g. fuzzy
> logic] -- so not much different from some of the other operators? :-)
> 
> -- bjorn

I'm going to guess it's because a) it's a boolean operation, not an
arithmetic one, and therefore b) overloading it could cause incredible
mischief with the language.  Imagine:
class C:
  def __and__(self, some_other_thing):
     return (self or some_other_thing)

Now suddenly if statements don't work for some objects, all kinds of
internal code probably fails for no apparent reason, etc.
Other reasons would be performance impact on a fundamental language
construct, and possibly that no one could think of a good reason why
you'd want to do that.



From vze4rx4y at verizon.net  Tue Apr  1 00:45:53 2003
From: vze4rx4y at verizon.net (Raymond Hettinger)
Date: Tue, 01 Apr 2003 05:45:53 GMT
Subject: Regular expression help
References: <2FGdnaUSmIhkghSjXTWc-g@speakeasy.net>
Message-ID: <lI9ia.2600$tr2.543@nwrdny03.gnilink.net>

"Stephen Boulet" <stephen at theboulets.net> wrote in message
news:2FGdnaUSmIhkghSjXTWc-g at speakeasy.net...
> How do I use the sub function to replace my match with my match
> prepended by a newline?

Make your match into a group by enclosing it in parenthesis.
Then, reference that group with a backslash and the group
number.  If there is only one group, the group number is 1:

>>> import re
>>> s = 'the quick brown fox jumped over the lazy dog'
>>> print re.sub('(j....d)', r'\1\n', s)
the quick brown fox jumped
 over the lazy dog


Raymond Hettinger





From peter at engcorp.com  Wed Apr 16 10:47:23 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 10:47:23 -0400
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <3E9D6CFB.829F5217@engcorp.com>

Christopher Blunck wrote:
> [snip]
> That got me to thinking:  continuous testing is a cornerstone of the
> XP methodology.  <generalization>And Python programmers typically
> write lots of tests</generalization> (at least more than their Java
> counterparts from what I've seen).  That being said, can that
> generalization be extended to "Python programmers subscribe to the XP
> methodology"?
> 
> So how many of you guys use XP processes?

My team at Kaval Wireless is an XP team, with (currently) nine 
developers using two-week iterations, test-driven development (TDD),
daily scrums, no overtime, on-site customer, The Works...

Most development is done with Python, and we definitely find a 
*very strong* synergy between Python and the XP processes, especially
on the testing front.

We also do some Javascript, and some embedded C work (realtime
kernel, 16-bit micros), both of which we are slowly wrapping with
a framework allowing tests to be written in Python using PyUnit
(a.k.a. unittest.py).

Our evolution to being an XP team (over about two years) came about 
at the same time as our adoption of Python, and I believe it would be 
next to impossible to distinguish between the two in terms of which 
came first.  I believe it was a positive feedback loop, with each 
feeding a growing appreciation for and dependence on the other.

As the man says, "Python is an agile language!" :-)

(See http://www.agilealliance.org for more on that term.)

-Peter



From john_taylor_1973 at yahoo.com  Wed Apr 16 20:39:25 2003
From: john_taylor_1973 at yahoo.com (John Taylor)
Date: 16 Apr 2003 17:39:25 -0700
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <fc30bff.0304161639.2969734a@posting.google.com>

Can someone please describe to me XP in a few sentences?  I have heard
the terminology before, but never an explanation.  Also, any good
links or books would be greatly appreciated.  Python is such a joy to
program in/with, I'd love to hear about making the process even
better!

Thanks,
John Taylor


blunck at gst.com (Christopher Blunck) wrote in message news:<1147e466.0304160630.5b510724 at posting.google.com>...
> Was just reading a thread about how python implements protected and
> private methods.  I found JP's response quite interesting (this is
> something I've heard numerous friends of mine say of the language):
> 
> [In Python] everything is permissible, but not
> everything is opportune.  And that is very much Python's philosophy:
> rather than focusing on trying to make some things impossible (and
> generally failing -- a simple cast in most implementations of C++ lets
> you blast away any "protected" that a silly library designer may have
> tried to impose on you;-), Python empowers and trusts the programmer.
> 
> 
> The natural response a non-Python programmer has to this statement is,
> "what?!  you __trust__ the programmer?!  that doesn't work in most
> environments."  Despite the ignorance this response demonstrates, I
> none-the-less thought about the statement and the answer I came up
> with was "we write lots of really thorough tests to demonstrate
> functionality."
> 
> That got me to thinking:  continuous testing is a cornerstone of the
> XP methodology.  <generalization>And Python programmers typically
> write lots of tests</generalization> (at least more than their Java
> counterparts from what I've seen).  That being said, can that
> generalization be extended to "Python programmers subscribe to the XP
> methodology"?
> 
> So how many of you guys use XP processes?



From carsten at sarum.dk  Tue Apr 29 04:28:14 2003
From: carsten at sarum.dk (Carsten Gehling)
Date: Tue, 29 Apr 2003 10:28:14 +0200
Subject: Function to convert fetchone() result from a list to a dictionary
Message-ID: <BHEFKIOIMEDGILGDACKDCEAPDFAA.carsten@sarum.dk>

I'm fairly new to Python, and have yet to discover alle the little
details, that may make my programs look leaner and perhaps become
faster...

Using the DB API to access databases, I've stumbled upon what I would
call a shortcoming in the API (IMHO). All the fetch commands return
the results as lists instead of dictionaries.

So I've made my own conversion function, using the cursor.description
attribute. It's pretty straight-forward, so I was wondering, if
there's a smarter way to do this?

The function is like this:
-------------------------------
def cursor_fetchone_dict(cursor):
	result = cursor.fetchone()
	if result == None: return None
	result_dict = {}
	for i in range(len(result)):
		result_dict[cursor.description[i][0]] = result[i]
	return result_dict
-------------------------------

Example of use:
-------------------------------
import MySQLdb

def cursor_fetchone_dict(cursor):
	result = cursor.fetchone()
	if result == None: return None
	result_dict = {}
	for i in range(len(result)):
		result_dict[cursor.description[i][0]] = result[i]
	return result_dict

con = MySQLdb.connect(host='localhost', port=3306, user='xxxxx',
passwd='xxxxxx', db='xxxxxx')

cursor = con.cursor()
sql = "select * from sometable "
cursor.execute(sql)

result = cursor_fetchone_dict(cursor)
print result
-------------------------------

Any comments are welcome, TIA

- Carsten

 




From rjkuhns at geetel.net  Wed Apr  2 09:24:04 2003
From: rjkuhns at geetel.net (Richard Kuhns)
Date: Wed, 2 Apr 2003 09:24:04 -0500
Subject: A question about help() and docstrings
Message-ID: <20030402092404.7a466b0c.rjkuhns@geetel.net>

OK, I'm feeling rather dense right now.  I'm obviously not understanding
something, so if anyone would like to hit me with a clue stick please feel
free.

Given a file dummy.py as follows:
====
def dummy_function(one, two):
    """This is a rather stupid function.
    """
    print "1 + 1 = 3"
====

I can run

: molin$/var/tmp; python
Python 2.2.1 (#1, Aug 30 2002, 12:15:30) 
[GCC 3.2 20020822 (Red Hat Linux Rawhide 3.2-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
Importing sys and os...
>>> import dummy
>>> help('dummy.dummy_function')
Help on function dummy_function in dummy:

dummy_function(one, two)
    This is a rather stupid function.

>>> 

The doc string shows up just fine.  However, if I use 'from dummy import
*' help() no longer knows anything about it.

: molin$/var/tmp; python
Python 2.2.1 (#1, Aug 30 2002, 12:15:30) 
[GCC 3.2 20020822 (Red Hat Linux Rawhide 3.2-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
Importing sys and os...
>>> from dummy import *
>>> help('dummy_function')
no Python documentation found for 'dummy_function'

>>> dir(dummy_function)
['__call__', '__class__', '__delattr__', '__dict__', '__doc__', '__get__',
'__getattribute__', '__hash__', '__init__', '__name__', '__new__',
'__reduce__', '__repr__', '__setattr__', '__str__', 'func_closure',
'func_code', 'func_defaults', 'func_dict', 'func_doc', 'func_globals',
'func_name']
>>> print dummy_function.__doc__
This is a rather stupid function.
    
>>>

Why doesn't help() work in the second case?
-- 
Richard Kuhns			rjkuhns at geetel.net




From home at dunlavy.net  Sun Apr 20 13:12:16 2003
From: home at dunlavy.net (Howard Dunlavy)
Date: Sun, 20 Apr 2003 17:12:16 GMT
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org>
Message-ID: <QxAoa.540832$3D1.296696@sccrnsc01>

"Tim Butterfield" <timbutterfield at mindspring.com> wrote

>>>>>>
What I envision is an IDE for python similar to VB and Delphi
(PyDE?) that allows for those property editors and resource files to
be integrated into an editor/debugger and builder.  Of course, being
able to add custom widgets and tie in data support to those widgets
would be needed.  Is such a thing desired?
<<<<<<

There is project that I have come across called Boa Constructor which lists
as its goal the creation of the equivalent of Delphi for Python.  It can be
found at Source Forge also.

-Howard





From aleaxit at yahoo.com  Thu Apr 10 03:04:41 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: Thu, 10 Apr 2003 09:04:41 +0200
Subject: [Resend] A few errors in "Python in a Nutshell" - Numeric related
References: <b721ql$akr$1@peabody.colorado.edu>
Message-ID: <b735291r7m@enews1.newsguy.com>

Fernando Perez wrote:

> This is just a resend since it seems A Martelli is back, and I'm pretty
> sure he did a massive delete of c.l.py, considering the volume this group
> generates :)  I did drop a note on the O'reilly site about this stuff, so
> this is just a heads up.

This is highly appreciated -- thanks!  Unfortunately I'm NOT "back" yet
(PythonUK is over but I'm traveling for other work-related stuff), just
occasionally at a machine where I can poke into c.l.py from;-).  But I'll
forward your mail to myself and look at it carefully (just like at all
the other errata for the Nutshell, but this one is subtler!-) next week,
when at last I _should_ indeed be "back"!-)


Alex




From dave at boost-consulting.com  Mon Apr 21 07:55:56 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Mon, 21 Apr 2003 07:55:56 -0400
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org>
 <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de>
 <mailman.1050725072.21332.python-list@python.org>
 <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl>
 <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl>
 <erk5avkh423eh8bc13ctvtbjqt9f6kt4kv@4ax.com>
Message-ID: <84smscm5df.fsf@boost-consulting.com>

Courageous <jkraska at san.rr.com> writes:

>>Thanks. It seems Tim is making a heroic effort to explain the new in
>>terms of the old. This necessarily must fail, but what other option is
>>there instead of rebuilding all terminology out of gotos?
>
> "non local goto" describes it well, but another way of looking at
> it is as a "first class, user controlled context switch". That's
> actually the meat of what it is. In unix, theoretically one could
> set up the heavy weight equivalent of a continuation by the use of
> setjmp and

Really?  I thought longjmp could only go back up the call stack,
losing any execution context from inner frames.  The whole point of
continuations, AFAICT, is that:

       1. execution context is preserved
       2. You can transfer control up, down, or between call stacks.

> In conjunction with a "stackless" environment, a continuation is
> actually a very lightweight state save coupled with the ability
> to restore to the last point the continuation was set.

That sounds right, but I don't see how it relates to setjmp/longjmp.

Also, "stacklessness" is really an implementation detail AFAICT - it
just says that all execution context is stored in
dynamically-allocated structures instead of using local variables on
the 'C' stack (via a recursive-calling implementation).

Have I misunderstood this whole thing?

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From tim at bladerman.com  Thu Apr  3 14:12:02 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Thu, 03 Apr 2003 19:12:02 GMT
Subject: wxSplitterWindow to fill up its avalible space
Message-ID: <6I%ia.3871$Du.12407@newsc.telia.net>

Hi
I have put a wxSplitterWindow in a wxBoxSizer together with a panel.
My problem is that I must set the splitterWindowsize by hand when I create
it

    wxSplitterWindow(self, -1, size=(300,300))

Otherwise it gets no size at all. If I just put two panels there instead,
they fill upp al avaliable space nicely and I am sure there is an easy way
to get the splitterWindow to do the same but I cannot seem find it.

I can add the splitterWindow with a large proportionparameter like this

    sizerBox.Add(splitterWin , 10)

I works somewhat, it makes the avalible space larger but it doesnt scale up
the splitterwindow. Must the scaling of the splitterWindow be done manually
in repaint or something?

Thanks
Tim





From mwh at python.net  Fri Apr 25 08:49:02 2003
From: mwh at python.net (Michael Hudson)
Date: Fri, 25 Apr 2003 12:49:02 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com>
Message-ID: <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk>

mis6 at pitt.edu (Michele Simionato) writes:

> The meta-type conflict is rather annoying, indeed. According to the
> metaclass book, the language should generate the correct metaclass
> that solves the conflict automatically, without burden for the
> programmer. Python is not that magic (yet) therefore you have to
> solve the conflict yourself.

I guess a potential problem is that if you do

class C(object):
    __metaclass__ = A

class D(object):
    __metaclass__ = B

class E(C, D):
    pass

class F(C, D):
    pass

you'd like both E and F to have the *same* (rather than merely
equivalent) metaclasses.  This would seem to be tricky to arrange.

Cheers,
M.

-- 
  Problem: Printer is missing.
  Solution: Printer has been stolen.  Buy new one.
                                      -- Internet Oracularity #1306-03



From noemail at forme.com  Mon Apr  7 00:33:29 2003
From: noemail at forme.com (Andrew H - Lisasoft)
Date: Mon, 07 Apr 2003 14:03:29 +0930
Subject: OpenFlow, Help Please!!!
Message-ID: <K87ka.16$DD6.2390@vicpull1.telstra.net>

Hi thanks for reading this.

I'm trying to use OpenFlow (running on Zope). I've no idea how to even 
open the application. Please help me. Is there a tutorial or some kind 
of documentation that will help get me started? I've looked everywhere 
but can't find anything.


Any help would be GREATLY appreciated.




ps please respond in the newsgroup, so others can have this information.




From skip at pobox.com  Tue Apr 22 15:22:38 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 22 Apr 2003 14:22:38 -0500
Subject: True/False
In-Reply-To: <1051038682.702.191.camel@lothlorien>
References: <1051036194.697.177.camel@lothlorien>
        <16037.36122.922342.405340@montanaro.dyndns.org>
        <1051038682.702.191.camel@lothlorien>
Message-ID: <16037.38526.105041.678508@montanaro.dyndns.org>

    Ian> This was the sort of issue I was thinking about.  But another
    Ian> change is:

    Ian> Python 2.3a2+ (#2, Mar 21 2003, 22:13:05) 
    Ian> [GCC 3.2.3 20030316 (Debian prerelease)] on linux2
    Ian> Type "help", "copyright", "credits" or "license" for more information.
    >>>> None = 10
    Ian> <stdin>:1: SyntaxWarning: assignment to None

    Ian> Will this happen to True/False?  Probably not, because that would
    Ian> *really* mess up backwards compatibility (since try:except:
    Ian> wouldn't even work).

It might.  There was a thread on python-dev or c.l.py about this particular
item within the last couple months.  I don't recall if any conclusions were
reached.  A google search should turn it up.

Skip




From alet at librelogiciel.com  Wed Apr 30 16:47:56 2003
From: alet at librelogiciel.com (Jerome Alet)
Date: Wed, 30 Apr 2003 22:47:56 +0200
Subject: [ANNOUNCE] PyKota v1.05 is out
Message-ID: <pan.2003.04.30.20.47.56.647397@librelogiciel.com>

Hi,

I'm please to announce PyKota v1.05

PyKota is a python-written, internationalized, complete, centralized and
extensible print quota solution for CUPS (http://www.cups.org) and now LPRng
(http://lprng.sourceforge.net).

PyKota is distributed under the terms of the GNU GPL from :

  http://www.librelogiciel.com/software/

Changes from version 1.03 are VERY important :

  - LPRng is now supported too ! This makes PyKota the first
    free print quota and accounting solution to support both
    CUPS and LPRng.

  - External accounting methods are now available !
    This means that people who can't use SNMP or AppleTalk,
    or prefer to do accounting prior to each job being sent
    to the printer, instead of always being one job late,
    can now use WHATEVER COMMAND THEY SEE FIT to count pages
    in print jobs.

  - A PyKota-vs-PrintBill web page was set up, for people who wants
    a quick and fair overview of the two products. This was checked
    by PrintBill's author so it is believed to be honest.

  - Pluggable accounting methods for people who want to contribute
    internal accounting methods to the project. A 'stupid' example
    is included.

  - Several tries are done during one minute when direct printer querying
    is used, this lets the time to warm up for some printers which don't
    answer immediately when idle.

  - Default print policy is now DENY instead of ALLOW for users unknown to
    the Print Quota system.

  - The usual bug fixes.

You can see the main features, view a screenshot of a print quota
report, give a look at the recent changes, download the CVS version,
or purchase an official package from :

  http://www.librelogiciel.com/software/PyKota/action_Presentation

You can see a sample report at :

  http://servmediup.unice.fr/cgi-bin/printquota.cgi

Please send bug reports, flames, comments to alet at librelogiciel.com

Thanks for reading

Jerome Alet



From aleax at aleax.it  Thu Apr 24 17:33:20 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 21:33:20 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <3ea434fd$0$180$a1866201@newsreader.visi.com> <e9570f37.0304220625.11312687@posting.google.com> <K2qdnXJ_ZeZMVTujXTWcqg@speakeasy.net> <F3Epa.51023$LB6.1152908@news1.tin.it> <e9570f37.0304240556.50465c66@posting.google.com>
Message-ID: <AKYpa.3501$K35.79126@news2.tin.it>

Graham Fawcett wrote:

> Alex Martelli <aleax at aleax.it> wrote in message
> news:<F3Epa.51023$LB6.1152908 at news1.tin.it>...
>> A.M. Kuchling wrote:
>>         ...
>> > Money does not make interested developers materialize,
>> 
>> Surely you're joking, Mr Kuchling?
> 
> Alex, are you demonstrating the materialization of an interested
> developer? ;-)

Just asserting it;-).  Put ENOUGH money [or goods and services
worth money, as are many of those you list] as the reward for
some activity, and people interested in that activity WILL magically
materialize, just like that.  Many of them will be people that
would normally still be interested, but not have time due to their
engagement in other gainful activities (because e.g. raising a
family does cost a lot of money, and to a lesser extent do so many
creature comforts).

> and of course,
> the way that girls go crazy for Python contributors.

Well, THAT, too (I didn't *know*, when I fell in love with Python,
that it would help me connect with the most wonderful woman in the
world -- but, had I known, I'd surely have jumped into it sooner).


> Of course we have no infrastructure for this (at least for the ones
> that cost money). But I'm sure we could create one if the need arose.

I'm not so sure, myself -- but it's sure worth trying.

> If it would help to move our goals along, I would gladly pay an annual
> Python tax, and/or participate in some kind of token economy. It would
> be a small price to pay in return for the *tremendous* advantages (and
> joys!) that the Python community provides to me.
> 
> In short, I think we should all send $20 a year to Guido, in return

That's the kind of thing the Python Sofware Foundation is going to
aim for -- collect small (or large!-) donations (tax-deductible, too,
for US citizens), and use them to fund Python development (and/or
other such Pythonic activities).  Stay tuned for further announcements...


Alex




From pinard at iro.umontreal.ca  Tue Apr  8 17:10:09 2003
From: pinard at iro.umontreal.ca (Francois Pinard)
Date: 08 Apr 2003 17:10:09 -0400
Subject: Is there any way to make repr(aString) use double quotes?
In-Reply-To: <698f09f8.0304081057.1ba7041f@posting.google.com>
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
Message-ID: <oqhe98ofv2.fsf@alcyon.progiciels-bpi.ca>

[Jeremy Fincher]

> The topic really says it all -- I'd like to *force* repr(aString) to
> use double quotes.  Is that possible?  If it's not, then is it possible
> to achieve the same thing (everything escaped properly, etc.) through
> other means?

Pymacs once needed exactly this, from the that double quoted strings
in Python and in Emacs were "compatible".  But after Emacs changed how
hexadecimal escapes are represented, the trick did not work anymore,
and I had to abandon it for something more explicit.

Let me seek in obsolete Pymacs releases...  [dig, dig, dig...]  OK,
found it!  The kludge below was found with the kind collaboration of the
Python community, mixing the ideas that were discussed at the time.


# Python delimits a string it by single quotes preferably, unless
# single quotes appear within the string while double quotes do
# not, in which case it uses double quotes for string delimiters.
# Checking the string contents, the C code stops at the first NUL.
# We prefix the string with a single quote and a NUL, this forces
# double quotes as delimiters for the whole prefixed string.  Then,
# we get rid of the representation of the single quote and the NUL.

write('"' + repr("'\0" + value)[6:])

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From kossay_omary at yahoo.com  Wed Apr 30 20:12:50 2003
From: kossay_omary at yahoo.com (Kossay Omary)
Date: Wed, 30 Apr 2003 17:12:50 -0700 (PDT)
Subject: connecting to a modem from python
Message-ID: <20030501001250.43594.qmail@web20512.mail.yahoo.com>

I am working in a linux environment. I have an
external modem connected to one of the serial ports.
Is there any python module which I can use, which
already enables me to send commands to the modem from
python, check status, etc...

My final application would require a python code which
will recognize that an incoming call is hapenning and
will display the phone number (assuming the modem and
the line have caller-id capability).

Can some one give me some direction as to how to
implement this using existing code or where to look
for.
I wouldn't mind writing the code myself, just that I
do not have much background in dealing with modems
and/or serial interfaces, so I appreciate your help...

Thanks,
Kossay


__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com




From andersjm at inbound.dk  Thu Apr  3 17:17:01 2003
From: andersjm at inbound.dk (Anders J. Munch)
Date: Fri, 4 Apr 2003 00:17:01 +0200
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <x7of3qjh4g.fsf@guru.mired.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com> <wBIia.2000$y3.310439@news010.worldonline.dk> <x7adf7ibk6.fsf@guru.mired.org>
Message-ID: <Kp2ja.2590$y3.426173@news010.worldonline.dk>

"Mike Meyer" <mwm at mired.org> wrote:
> "Anders J. Munch" <andersjm at inbound.dk> writes:
> > For a Python example, see the Decorate-Sort-Undecorate pattern.  In
> > C++ the straightforward and fast solution is to use a suitable
> > comparison predicate.  In Python you can do the same, but that's slow,
> > and you are left with a choice between clarity and speed.
>
> The statement about the preferred method in C is simply false. Anytime
> the "suitable comparison predicate" is sufficiently complex - for
> instance, what the unix sort(1) command allows - a DSU that allows
> simple comparisons will be faster than a sort with complex
> comparisons.

The typical case is a simple projection.  And when it's not you can
turn it into one by storing or caching the projected value in the
objects themselves.

> In fact, any time you can simplify the comparison with DSU, the DSU
> sort will always perform faster on n objects for sufficiently large n.

A comparison predicate in C++ (who said anything about C?) that
forwards to instance variables simplifies the comparison similarly.

- Anders






From tim.one at comcast.net  Mon Apr 28 14:05:45 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 28 Apr 2003 14:05:45 -0400
Subject: More rants on a bad sort
In-Reply-To: <0MVr+kKkXUES092yn@gnosis.cx>
Message-ID: <BIEJKCLHCIOIHAGOKOLHCELOFHAA.tim.one@comcast.net>

[Lulu of the Lotus-Eaters]
> ,,,
> The sorting algorithm winds up only performing allowed comparisons in:
>
>     [u'x', 'x', chr(255)].sort()

That's so because u'x' <= 'x' and 'x' <= chr(255).  Note that it's
impossible to guess whether a given user would *consider* this to be a
sorted list, though, because without specifying the intended encoding it's
impossible to know which Unicode character(s) is(are) encoded by chr(255).
The decoded Unicode string corresponding to that may be <, ==, or > than
u'x'.

> Well... OK.  We have a similar story with complex numbers:
>
>     >>> '1' < 1, 1j < '1'
>     (0, 1)
>     >>> 1 < 1j
>     Traceback (most recent call last):
>       File "<stdin>", line 1, in ?
>     TypeError: cannot compare complex numbers using <, <=, >, >=
>
> Given all that, is it POSSIBLE to construct a list containing a string,
> int, and complex, that will not crash on a sort?  I don't know!

It isn't, but it's hard to guess why anyone would care <wink>.

> I couldn't find one--and I THINK the relative order of complex and strings
> precludes it... but actually proving this one way or another is
> decidedly non-obvious.

It can't happen because (conceptually; the implementation is slicker) the
list is first sorted by alphabetical order of type name, where in order to
prevent cases like

    10 < [9] < 8L < 10

the "type name" of each numeric type is artificially considered to be the
empty string.  So if the list has a complex, an int, and a string, the order
is first conceptually reduced to one of these two cases:

    int, complex, string
    complex, int, string

Then an attempt to compare the int and complex values will be made (they
tied on the "primary sort key" -- they both pretended to have "" as their
type name), and that will complain.

Now you can blame that on that Python complains in some senseless cases.  I
blame it on that Python doesn't complain in all senseless cases <wink>.





From nirina at mail.blueline.mg  Thu Apr 17 06:53:31 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Thu, 17 Apr 2003 14:53:31 +0400
Subject: scrollbar
References: <b7kh3g$51u$1@nemesis.news.tpi.pl>
Message-ID: <042701c304cf$fb5b1e60$f6a2383e@raseliar>

"zyziu" wrote:

> Hello. I made lanChat in Python. I have problem with Text widget. I want
to
> text from Entry goes to Text widget and scrollbar goes down, because I
can't
> see what write another user. Now I must do it manually. How do it to
> scrollbar goes down automatically? Thanks for answer. Zyziu from Poland.

Maybe you want Text widget see method.
Also, you can try using ScrolledText instead of Scrollbar

# ----begin------
from ScrolledText import *
root = Tk()
st = ScrolledText(root)
st.pack()
st.insert('insert','hello\n')

for i in range(200):
    st.insert('insert',str(i)+'\n')
st.insert('insert','This is the end')

# the see method
st.see('end')
st.pack()
Button(root,text='quit',command=root.quit).pack()
root.mainloop()
# ----end--------

Hope this helps
nirinA







From pobrien at orbtech.com  Wed Apr 16 12:02:59 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 16 Apr 2003 11:02:59 -0500
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050394215.20671.python-list@python.org> <b7i26g$17t5m$1@ID-169208.news.dfncis.de>
Message-ID: <m24r4yh1l8.fsf@orbtech.com>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> writes:

> Christian Tanzer wrote:
> > This has nothing to do with touching point but how cursor
> > movements behave. For example, searching doesn't work properly in some
> > cases unless you let `inhibit-point-motion-hooks` to a true value.
> 
> Er... if it has nothing to do with touching point, how
> come you have to fiddle with some global that has 'point'
> in its name?
> 
> > But there isn't that much difference between an attribute of a buffer
> > object (in the Python sense) and a buffer-locale variable (in the
> > Emacs sense).
> 
> The difference is that the buffer (or some other data structure
> from which the buffer can be easily reached) would be an explicit
> argument.

I suppose one could bind attributes (current frame, current buffer,
point) to a module and then import that module everywhere that needed
access to these attributes, rather than passing the attributes as
parameters.  So the module would act as a global, but with a structure
that could be navigated.  Just a thought.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From staschuk at telusplanet.net  Thu Apr 24 13:49:11 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 24 Apr 2003 11:49:11 -0600
Subject: serveur TFTP
In-Reply-To: <e4d8f90f.0304240622.64cf7f8e@posting.google.com>; from naoaqualis@aol.com on Thu, Apr 24, 2003 at 07:22:26AM -0700
References: <e4d8f90f.0304240622.64cf7f8e@posting.google.com>
Message-ID: <20030424114911.A252@tibia.amotlpaa.bogus>

Quoth alex:
> j'ai recup?r? une librairie TFTP ecrit en python mais j'avoue avoir un
> peu de mal ? l'utiliser.comment faire un serveur TFTP avec? j'essai
> d'identifier les methodes que je pourrais appliquer a un objet
> mon_serveur par exemple :
> 
> import tftplib
> mon_serveur = tftplib.TFTPServer()
> 
> ...et apr?s je suis un peu perdu.

Je sais seulement un peu de Fran?ais, et peut-?tre je ne comprends
pas votre question.  Mais je pense que vous voulez:
    import tftplib
    server = tftplib.TFTPServer()
    server.forever()

-- 
Steven Taschuk                            staschuk at telusplanet.net
"Our analysis begins with two outrageous benchmarks."
  -- Clinger et al., "Implementation strategies for continuations"




From cpl.19.ghum at spamgourmet.com  Fri Apr 11 04:48:49 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Fri, 11 Apr 2003 10:48:49 +0200
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
Message-ID: <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>

> [cut]
> I think python is a very good language for education. You could get in
> contact with local schools.

Thomas, I am not so sure about this.

Yes, I love Python, and it is the most elegant programming language I 
ever met. 

Python gives you objectoriented, functionorientet and imperative 
programming paradigm in one fitting package.

Python does not bother beginners with a million typecasts. Very great.

For education there are small drawbacks:

- 50% of all exercises "to teach programming" are already solved in the 
standard-library or are a method of a buildin object / are allready build 
in the core language

- 45% of these exercises are solved within "the python cookbook", 
available also online

- if you learn programming with Python and later have to use Java or 
Visual Basic, it will be the most frustrating experience for the young 
fellows.


Anyway, I regret not having met Python earlier in my programming career.




From rando-15696825.5.reiman at xoxy.net  Fri Apr 18 12:27:08 2003
From: rando-15696825.5.reiman at xoxy.net (Charlie Reiman)
Date: 18 Apr 2003 09:27:08 -0700
Subject: A new and very robust method for doing file locking over NFS?
References: <lcu1cw525a.fsf@gaffa.mit.edu>
Message-ID: <5j65pbai03.fsf@xoxy.net>

Douglas Alan <nessus at mit.edu> writes:

> I'd like to do file locking over NFS without using lockd.  The reason
> I want to avoid using lockd is because many lockd implementations are
> too buggy.
> 
> It is fairly easy to avoid using lockd -- just avoid using lockf() to
> lock a file.  Instead of using lockf(), lock a file by creating a lock
> file that you open with the O_CREAT | O_EXCL flags.  To unlock the
> file, you merely unlink the lock file.  This method is fairly
> reliable, except that there is a small chance with every file lock or
> unlock that something will go wrong.
> 
> The two failure symptoms, as I understand things, are as follows:
> 
> (1) The lock file might be created without the client realizing that
> it has been created if the file creation acknowledgement is lost due
> to severe network problems.  The file being locked would then remain
> locked forever (until someone manually deletes the lock) because no
> process would take responsibility for having locked the file.  This
> failure symptom is relatively benign for my purposes and if needed it
> can be fixed via the approach described in the Red Hat man page for
> the open() system call.
> 
> (2) When a process goes to remove its file lock, the acknowledgement
> for the unlink() could be lost.  If this happens, then the NFS driver
> on the client could accidentally unlink a lock file created by another
> process when it retries the unlink() request.  This failure symptom is
> pretty bad for my purposes, since it could cause a structured file to
> become corrupt.
> 
> I have an idea for a slightly different way of doing file locking that
> I think solves problem #2 (and also solves problem #1).  What if,
> instead of using a lock file to lock a file, we rename the file to
> something like "filename.locked.hostname.pid"?  If the rename()
> acknowledgement gets lost, the client will see the rename() system
> call as having failed due to the file not existing.  But in this case
> it can then check for the existence of "filename.locked.hostname.pid".
> If this file exists, then the process knows that the rename() system
> call didn't actually fail--the acknowledgement just got lost.  Later,
> when the process goes to unlock the file, it will rename the file back
> to "filename".  Again, if the rename system call appears to fail, the
> process can check for the existance of "filename.locked.hostname.pid".
> If the file no longer exists, then it knows the rename call really did
> succeed, and again the acknowledgement just got lost.
> 
> How does this sound?  Is this close to foolproof, or am missing
> something?
> 
> I'm not much of an NFS expert, so I am a bit worried that there are
> details of NFS client-side caching that I don't understand that would
> prevent this scheme from working without some modification.
> 
> |>oug

I like this idea a lot but it has one drawback. In the traditional
lockfile approach, you can have multiple readers even if someone is
writing. Now that you rename the file, eveytime someone is writing the
filename is a moving target. Depending on how you intend to use the
file, this might make your solution a no-go from the get-go.

But as long as your locking behavior applies to both reading and
writing, it seems quite foolproof.

FWIW, I've never had any problems with NFS for traditional lock
files. NFS works pretty hard to get messages where they need to
be. If I were that deeply concerned or running across an unreliable
network, I'd probably write my own lock daemon and run it over TCP.

Charlie.





From peter at engcorp.com  Fri Apr 25 09:09:20 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 09:09:20 -0400
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth> <+xKq+ks/K/pQ089yn@the-wire.com> <Yv2qa.32862$ot1.31621@nwrdny02.gnilink.net>
Message-ID: <3EA93380.883EFB08@engcorp.com>

Raymond Hettinger wrote:
> 
> [Jim]
> > >I have come across a weirdness. I have a number of floats, (lat long
> > >pairs) which I stuff into a dict, but when I take them out, my nice %2f4
> > >number, (say, 47.5894) is something long and *almost* right, like
> > >47.589333333333333339. What gives?
> 
> [Mel]
> >    Chances are that the float hasn't changed, but that
> > you're printing it in a different way.
> 
> The Python FAQ has a pretty good explanation:
> 
> http://www.python.org/cgi-bin/faqw.py?req=show&file=faq04.098.htp

I looked for the FAQ entry to offer as well, but I wasn't satisfied
with that one.  It refers to apparent errors from numerical calculations,
not primarily to the representation of the float when printed.

Is the OP's question deserving of a separate entry in the FAQ, or
at least an addendum to the one Raymond linked above?

-Peter



From peter.maas at mplusr.de  Thu Apr 10 07:25:20 2003
From: peter.maas at mplusr.de (Peter Maas)
Date: Thu, 10 Apr 2003 13:25:20 +0200
Subject: design patterns in german (object-oriented)
In-Reply-To: <b73asl$m64$1@newsreader2.netcologne.de>
References: <b73asl$m64$1@newsreader2.netcologne.de>
Message-ID: <b73k60$pcn$1@swifty.westend.com>

Thomas M. wrote:
> hi,
> are there any good german sides or books with design patterns for 
> object-oriented programming?

Enter "design patterns" in Google an request german language sites
only. You have to decide whether these sites fit your needs.
Example:

http://www.datec.at/dp.html

Mit freundlichen Gr??en,

Peter Maas

-- 
-------------------------------------------------------------------
Peter Maas, M+R Infosysteme, D-52070 Aachen, Hubert-Wienen-Str. 24
Tel +49-241-93878-0 Fax +49-241-93878-20 eMail peter.maas at mplusr.de
-------------------------------------------------------------------




From jcribbs at twmi.rr.com  Tue Apr 15 23:50:27 2003
From: jcribbs at twmi.rr.com (Jamey Cribbs)
Date: Wed, 16 Apr 2003 03:50:27 GMT
Subject: ANNOUNCE:  KirbyBase 1.01 is now available
Message-ID: <7q4na.96766$0X.23006457@twister.columbus.rr.com>

I don't know if this really warrants an announcement, but I wanted to 
say that I have posted version 1.01 of KirbyBase to my website.

I have added a few methods, including a "drop" method to delete tables, 
but the most important change is that I added some (hopefully) helpful 
documentation.

You can download Kirbybase 1.01 at: 
www.netpromi.com/files/kirbybase1.01.zip.

You can read about KirbyBase at:  www.netpromi.com/kirbybase.html.


Jamey Cribbs




From sjmachin at lexicon.net  Sat Apr  5 16:26:37 2003
From: sjmachin at lexicon.net (John Machin)
Date: 5 Apr 2003 13:26:37 -0800
Subject: Input from a file as a command line argument
References: <3E8EDBDA.7020701@pbezone.net>
Message-ID: <c76ff6fc.0304051326.151edfe9@posting.google.com>

Teemu Luojola <teemu.luojola at pbezone.net> wrote in message news:<3E8EDBDA.7020701 at pbezone.net>...
> If I make an executable script in linux environment, how can I pass 
> input to that script from a file on command line? Like the following:
> 
> $ script.py input.txt

import sys
file_name = sys.argv[1]
file_handle = file(file_name)
for line in file_handle:
   print line
file_handle.close()

> or perhaps
> 
> $ script.py < input.txt

import sys
for line in sys.stdin:
   print line

Notes: (1) This will work with Windows too. (2) The above is written
for Python version 2.2 (the latest production release) -- if you have
only an earlier version, (like 1.5.2), you may like to read back
through the "What's new in Python x.y" series on www.python.org to see
what you are missing. (3) Unless you are forced to by inability to
upgrade your Python version, or a need to read another's code, you
don't need to know about xreadlines() -- the "for line in file_object"
paradigm is the thoroughly modern way.

HTH,
John



From cpitaar at yahoo.com.ar  Tue Apr 29 11:50:42 2003
From: cpitaar at yahoo.com.ar (Carlos P.)
Date: 29 Apr 2003 08:50:42 -0700
Subject: Raising objects
Message-ID: <453d10b2.0304290750.222fb453@posting.google.com>

Hi all,

I have this short question: can the operator 'raise' raise exceptions
which are instances of object? I'm getting the following behaviour
from python 2.2.2:

>>> class F: pass

>>> raise F()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
__main__.F: <__main__.F instance at 0x8136bf4>

>>> class E(object): pass

>>> raise E()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: exceptions must be strings, classes, or instances, not E

Thank you,
Carlos



From hupp at cs.wisc.edu  Tue Apr 22 21:31:53 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Tue, 22 Apr 2003 20:31:53 -0500
Subject: pattern match
In-Reply-To: <220420031949394828%greg.bland@valpo.edu>
References: <220420031949394828%greg.bland@valpo.edu>
Message-ID: <20030423013153.GA11844@upl.cs.wisc.edu>

On Wed, Apr 23, 2003 at 12:52:38AM +0000, Mitch MF Cummstain wrote:
>                I just started using python and I need help with a
> pattern match.
>       
>    m = re.match("(.{30})(word)(.{30})",paragraph);
> 
> I need to match 30 characters on each side of word.  The problem word
> is getting treated as a string not a variable like I need it to.  How
> can I make it treat word as a variable ?

word = "foo"

pat = "(.{30})(%s)(.{30})" % word

m = re.match(pat, paragraph)


The '%' operator on a string acts like sprintf, it inserts it's
right hand side into the string at the point '%s' (s for string, d for
integer, etc).  

It's not necessary to put a semicolon at the end of a statement, BTW.

-Adam




From achim.domma at syynx.de  Wed Apr 30 02:49:34 2003
From: achim.domma at syynx.de (Achim Domma)
Date: Wed, 30 Apr 2003 08:49:34 +0200
Subject: running a PyUnit TestSuite
Message-ID: <b8nrlv$dvn$07$1@news.t-online.com>

Hi,

I have a TestSuite which depends on the execution order of the tests. If I
use the 'main' object provided by unittest.py I can not specify the
execution order, so I set up a TestSuite by hand. But how should I run it? I
found no easy way to pass a TestSuite for unittest.main. Any hint?

regards,
Achim





From srijit at yahoo.com  Sat Apr 12 09:16:51 2003
From: srijit at yahoo.com (srijit at yahoo.com)
Date: 12 Apr 2003 06:16:51 -0700
Subject: Stackless python
References: <221d8dbe.0304050531.107bc603@posting.google.com> <mailman.1049558408.987.python-list@python.org>
Message-ID: <221d8dbe.0304120516.4d543fb@posting.google.com>

Hello Tismer,

> This is temporary, the new version which I'm about to upload
> tomorrow does not have that external file, since it is all moced into
> the core dll, for simplicity.

You mention about a new version to be uploaded on 06th April, 2003.
But I do not see any new version in www.stackless.com

I am not comfortable with CVS. At least can there be a small tutorial
and some examples as HTTP/FTP download?

Without good documentation and examples it is very difficult to take
off on a new topic like Stackless Python. One of the most important
reasons for Python being so popular is excellent documentation with
every release.

After going through StacklessEuroPy.ppt, I feel it is worth learning
to use Stackless Python. But (I repeat!) I am afraid without good
documentation my attempt cannot last long.

With best wishes,
Srijit

Christian Tismer <tismer at tismer.com> wrote in message news:<mailman.1049558408.987.python-list at python.org>...
> srijit at yahoo.com wrote:
> > I am interested to try Stackless Python. I have downloaded python22.dll.zip. 
> > How do I install this zip file? I already have standard python 2.2.2 installed.
> 
> You either install python22.dll in a local folder. When you start
> Python from there, it will be stackless.
> If you want to install it permanently, rename the dll in
> windows/system32 and copy the new dll there.
> 
> In the version you have, there is a stacklessmodule.pyd which
> should be in the local folder or be copied into /python22/DLLs.
> This is temporary, the new version which I'm about to upload
> tomorrow does not have that external file, since it is all moced into
> the core dll, for simplicity.
> 
> > I also do not see any documentation like tutorials,
> > example programs, demos etc.
> 
> There are a few test programs, if you download the CVS tree.
> There is no public release, yet, this is still pre-alpha
> stuff.
> You can get all internal documentation by
>  >>> help(stackless)
>  
> > Is it 100% compatible with standrad Python 2.2.2?
> 
> Yes, it should be. If not, let me know.
> 
> ciao - chris



From peter at engcorp.com  Sun Apr  6 18:10:35 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 06 Apr 2003 18:10:35 -0400
Subject: PYTHONPATH question
References: <mailman.1049624230.16151.python-list@python.org>
Message-ID: <3E90A5DB.E3CF1C8E@engcorp.com>

Ali Dada wrote:
> 
> i want to add an entry to
> PYTHONPATH and i don't want to use sys.path because i want the new entry to
> remain permanently.

Consider creating a .pth file in the python folder itself (or wherever
they go under Linux).  See python/lib/site.py for details.  

-Peter



From lost at soul.net  Sun Apr 13 17:01:40 2003
From: lost at soul.net (lost)
Date: Sun, 13 Apr 2003 22:01:40 +0100
Subject: Help with CGI
References: <pan.2003.04.13.19.41.42.420658@soul.net> <z2kma.284$8c4.159@tornadotest1.news.pas.earthlink.net>
Message-ID: <pan.2003.04.13.21.01.40.565182@soul.net>

On Sun, 13 Apr 2003 20:48:31 +0000, Tyler Eaves wrote:

> On Sunday 13 April 2003 15:41, lost wrote:
> 
>> form = cgi.FieldStorage
> 
> There's your problem. 
> That line should read: 
> 
> form = cg1.FieldStorage()
> 
> The () is the difference between what you want, and simply getting a copy of 
> the class.
meh! can't believe i didn't see that one :)
thanks, Tyler!




From warren.postma at sympatico.ca  Mon Apr 21 01:29:38 2003
From: warren.postma at sympatico.ca (Warren Postma)
Date: Mon, 21 Apr 2003 01:29:38 -0400
Subject: Why functional Python matters
In-Reply-To: <a8b7f07a.0304181341.5e4b69f3@posting.google.com>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com> <RoacnSAeGeS6CACjXTWcoQ@comcast.com> <a8b7f07a.0304181341.5e4b69f3@posting.google.com>
Message-ID: <ilLoa.1695$Qg2.126929@news20.bellglobal.com>


> I'll second that.  Maybe we could get func added as a synonym for lambda, at least.


Maybe we should translate all language keywords into some other popular 
languages than english.  In which case, Lambda can remain as is, for our 
greek friends, and the rest of the keywords can change to greek words.

My-big-fat-greek-python-ly-yr's

Warren




From andrew-pythonlist at puzzling.org  Mon Apr 28 22:54:36 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Tue, 29 Apr 2003 12:54:36 +1000
Subject: POD vs. reST for standalone writing?
In-Reply-To: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
Message-ID: <20030429025435.GA13813@frobozz.local>

On Mon, Apr 28, 2003 at 10:21:59PM -0400, eichin at metacarta.com wrote:
> So, after a recent bad experience with LyX and version skew, I was
> looking at alternatives for software end-user and administrator
> documentation.  The constraints are pretty simple:
> 
> * produce good html and PDF
>   ** via linux tools, and highly preferably free ones
> * be editable on multiple platforms
>   ** linux first, but also, osx, windows
> * support "literal computer output" and screen shots

I think you want Twisted's Lore format (which I am the maintainer of).

It is essentially a subset of XHTML, with a number of a pre-defined classes.
Take a look at the docs at http://twistedmatrix.com/documents/howto for an
example of Lore's output (and for Lore's documention!).

E.g., here is the source to one of the documents:
    http://twistedmatrix.com/documents/howto/defer.html
and here is the Lore'd HTML output:
    http://twistedmatrix.com/documents/howto/defer
and there are PS and PDFs too.

Let me know if you want any more information or help (or ask on the Twisted
list: twisted-python at twistedmatrix.com).  We developed this format and it's
tools for much the same reasons that are motivating you to look around:
  - LyX/LaTeX have too high a learning curve and the tools are difficult to
    obtain and install on some platforms.
  - DocBook has an equally bad learning curve and toolset.  We want to
    *encourage* contributions, not scare them away!
  - LaTeX also has crummy HTML rendering.
  - ReST isn't extensible enough; how do you add markup for, e.g. API
    references?  How do you plug that into your docutils installation?  You
    can't.  (The Lore markup for APIs is 
        <code class="API">twisted.internet.reactor.run</code>
    )

But consider Lore:
  - Lore's source format is mostly viewable directly, being XHTML (although
    it looks better after run through Lore, of course!), and it also comes
    with a "lint" tool, to identify markup errors in your documents.
  - Lore is extensible; it's easy to write your own Lore input plugin if you need
    to define markup that isn't available by default.  You can write your
    own output plugin, too -- e.g. the "slides" input plugin has a
    corresponding magicpoint output plugin.
  - Lore is easy to write, assuming your authors already know some HTML...    
  - Code samples don't have to be included inline in HTML -- they can be
    referenced from live .py (or .c, or .java, or ...) files that you can
    be run and test directly, but the source will be inlined (and
    pretty-printed, if the source format is supported) in the Lore output.
    There's support for simple inline code examples too, (just use e.g. <pre
    class="python">), and even then the lint tool will syntax check them for
    you (if they are of Python code) :)
  
Lore is still a fairly new system, and is still undergoing development, but
it's rapidly maturing.  Twisted finds it invaluable.  I recommend you take a
look!

Regards,

-Andrew.





From rmunn at pobox.com  Wed Apr  2 10:34:17 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 15:34:17 GMT
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com>
Message-ID: <slrnb8lbo0.ccp.rmunn@localhost.localdomain>

sdieselil <sdieselil at yahoo.com> wrote:
> Hi
> 
> How can I implement double-linked lists and other complex tree-like
> structures in Python?

Do you *need* doubly-linked lists? Is there something that Python's
builtin list type can't do for you? Consider whether it's worth taking
the time to code up a data structure, with all the attendant edge cases,
when you've already got such nice useful builtin types...

Having said that, if you really need to code up a data structure that
doesn't already exist in Python, I'd recommend the same thing that
others have said: code up a class. E.g.,

class BinaryTreeNode(object):
    def __init__(self, data=None):
        self.data = data
        self.left = None
        self.right = None

class BinaryTree(object):
    def __init__(self):
        self.root = None

    def insert(self, data):
        if self.root is None:
            self.root = BinaryTreeNode(data)
        else:
            cur = self.root
            parent = None
            while cur is not None:
                cmp_result = cmp(cur.data, data)
                if cmp_result == 0:
                    raise DuplicateDataError
                elif cmp_result > 0:
                    # cur.data > data
                    parent = cur
                    cur = cur.right
                else:
                    # cur.data < data
                    parent = cur
                    cur = cur.left
            # Found the right place, now insert a node
            if cmp_result > 0:
                assert(parent.right is None)
                parent.right = BinaryTreeNode(data)
            else:
                assert(parent.left is None)
                parent.left = BinaryTreeNode(data)

    def search(self, data):
        # Etc.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From Frank at no.spam  Thu Apr  3 03:29:28 2003
From: Frank at no.spam (Frank)
Date: Thu, 03 Apr 2003 03:29:28 -0500
Subject: Persistent server with GUI interactive functions ?
Message-ID: <mgqn8v01utuhhioq5vq32rm2r2oadtrm8i@4ax.com>

I need to write a GUI app that will serve as a peer to peer
client/server.  At the highest level it would be similar to a chat
client, which has a server component (waiting for any incoming
connections) as well as the client part (initiating a connection,
executing user commands).

Each of those function appears to involve entering an "endless loop"
as the last logical step to initiate processing (i.e. "mainloop()" for
the user interface and "serve_forever()" for the server).  I haven't
yet grasped the logical structure that will allow me to have two
non-returning calls and maintain control of the program.

Is there an example of this around somewhere, or does anyone know of
documentation that will explain this?
(Note:  I will need to have the app run under Win, Linux, and
Solaris.)

Thanks,
Frank




From aldo at nullcube.COM  Wed Apr  2 22:27:30 2003
From: aldo at nullcube.COM (Aldo Cortesi)
Date: Thu, 3 Apr 2003 13:27:30 +1000
Subject: ANNOUNCE: pyGDChart2 alpha
In-Reply-To: <3E8BA830.277CCDA@engcorp.com>
References: <mailman.1049336408.10652.python-list@python.org> <3E8BA830.277CCDA@engcorp.com>
Message-ID: <20030403032730.GA21391@nullcube.com>

Thus spake Peter Hansen (peter at engcorp.com):

> Aldo Cortesi wrote:
> > 
> > pyGDChart2 itself can be found at:
> > 
> >         http://www.nullcube.com/software/pygdchart2.html
> 
> This link is on that page, but is missing:
> 
> http://www.nullcube.com/software/software/pygdchart2.html

Fixed!


Cheers,

Aldo

-- 
Aldo Cortesi
aldo at nullcube.com
http://www.nullcube.com




From mgerrans at mindspring.com  Mon Apr 21 12:44:51 2003
From: mgerrans at mindspring.com (Matt Gerrans)
Date: Mon, 21 Apr 2003 09:44:51 -0700
Subject: looping and deleting
References: <3cf72d52.0304210801.2128eca2@posting.google.com> <b815os$618$1@panix1.panix.com>
Message-ID: <b8176u$a2e$1@slb9.atl.mindspring.net>

"Aahz" wrote:
> Are you trying to modify the file in-place?  If so, look at the
> fileinput module.  If not:

I thought the fileinput module was just for gathering all the files on the
command line into one virtual stream (that is, reading them all without
respect to filename, as one long series of lines).

I don't know if this is the best or most Pythonic way, but this is what I
usually do to modify a text file in place (unless I know it will be a really
big file):

lines = file( filename ).readlines()
f = file( filename, 'w' )
for line in lines:
   f.write( modifyTheLine(line) )
f.close()





From shalehperry at attbi.com  Sun Apr  6 18:47:03 2003
From: shalehperry at attbi.com (Sean 'Shaleh' Perry)
Date: Sun, 6 Apr 2003 14:47:03 -0800
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2u1dcw22s.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <200304061547.03783.shalehperry@attbi.com>

On Saturday 05 April 2003 22:51, Patrick K. O'Brien wrote:
> Here is the background.  I'm working on a file editor (called
> PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> PyCrust code base.  I've got a basic version working, and it supports
> editing multiple files and all the usual, basic features.  But I want
> it to be as customizable and extensible as Emacs.  In fact, I want to
> model it after Emacs, so that elisp code could be rewritten in Python
> and work with PyAlaMode.  To do that, I've got to support the same
> primitives as Emacs; expose files, buffers, regions, and windows the
> same way Emacs does; handle the Emacs keybindings and ability to remap
> keys, the same way Emacs does, etc.
>

I was actually pondering this earlier in the week as I read more of the elisp 
and emacs docs.

One thing about emacs is that even though it is based on elisp most people can 
blissfully ignore that and just use it.  Please, if you make a newer, better, 
GUI editor with a Python backend do not limit it or even aim it at the Python 
community.  Make it the great new thing for ALL text editing.  This is why I 
use emacs over vi -- I can teach the editor about whatever it is I am editing 
and decrease the amount of time I spend typing in text. 6 years of vi and now 
a year and a half with emacs i can't imagine going back.  To get me to leave 
emacs will require something with all of emacs' might and more modern 
features.

real pop up hints, suggestive programming, etc as seen in commercial IDEs 
would be nice.  As much as the IDEs have improved they are still one trick 
ponies that can only be used to edit source so a real editor like emacs still 
wins overall.  Support for refactoring regardless of the language.  World 
peace.  you name it.




From ramen at lackingtalent.com  Tue Apr 15 18:55:50 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 15 Apr 2003 22:55:50 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <mailman.1050438685.19213.python-list@python.org>
Message-ID: <slrnb9p3g9.vg8.ramen@lackingtalent.com>

In article <mailman.1050438685.19213.python-list at python.org>,
Jp Calderone wrote:
>   While they take getting used to, you can't say that list comprehensions
> aren't functional.  They're not traditional-lisp functional, but that isn't
> the only kind :)

Certainly not, especially considering where they supposedly came from. I
think that they are certainly functional, but only representative of
particular (but common) functional idioms related to list processing.

For instance, there has yet to be a dictionary comprehension. This is
something I think would be nice to add, because I currently do things like
this: 

dict([(x['key'], x['value']) for x in data])

Where it might be nice to say instead:

{x['key']: x['value'] for x in data}

But I'm getting off track here. I'm not arguing that we should have
map/filter instead of list comprehensions, or the other way around. I think
both are useful and operate on different levels, similar to the relationship
between relational algebra and relational calculus. Just because they can
both do the same things doesn't mean one should be abolished in lieu of the
other.

And in a broader sense, I'm not even talking about built-in functions or
language constructs so much as the direction of Python's evolution. I don't
necessarily think that Python needs to become another LISP, though I agree
with Paul Graham on many of his observations of language evolution. I think
that Python is just fine being Python, but I don't want to see it become
another Java either. The day Python tells me "top-level functions have been
deprecated" is the day I find a new pet language. ;)

>> [snip - my plea for lambda love]
>
>   No worries, there are enough people who would shout loudly enough were
> this to be removed that its existence is safe, at least for the near future.
> 
>   Just make sure you give a holler anytime someone suggests discarding them.

Well, I suppose that's what I'm now attempting to do. =)

>   Lambdas are an excellent tool for obfuscation, that's for sure.  You're
> right not to recommend overusing them in production code.  Generally, I try
> to limit myself to the simplest expressions, on the order of "x and y" or
> "result[index]", and translate to fully fledged functions for larger
> expressions.  It's just easier to maintain.

This has certainly crossed my mind, and I understand that this is one of the
main reasons why Guido dislikes the construct. He seems to put readability
at a very high priority, which is why Python is often so elegant. But
readability is one of many things that are important to me. Another is
mathematical conciseness. If you have represented an idea very formally and
concisely, you can always expand it so that all functions and variables are
named and traditional if/while blocks are used. Doing the reverse is often
much more difficult.

For an even greater exercise in obfuscation, see my functional types recipe.
Warning: you may lose respect for me after you read this. ;)
http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/161403

I have successfully produced a custom XML-generation script that translates
a Java-based object model into XML using Jython. I began by writing only
lambdas, and almost all of them were one-liners. Then, I turned the lambdas
into proper "def"ined functions. Then, I observed the patterns in argument
passing to determine what data could be shared, and used this observation to
determine what classes to create. When I was done, I had something that
looked like it had been an object-oriented design from the start, but it was
very logical, concise, fast, and readable. On top of all of that, I had
*confidence* that my code worked properly because it was designed from
small, easily-defined parts that I could (attempt to) prove the correct
operation of. Proving the correctness of objects isn't so easy.

>   Also, note that lambdas *aren't* side-effect free, because Python isn't
> side-effect free.  You can write ones that are, but then again, you can
> write functions with "def" that are side-effect free too.

This is true. And you can write classes that are side-effect free, too. The
string object is a perfect example. It's also an example of how performance
can suffer when you enforce being side-effect free (the += operator). It's
not cut-and-dry by any means. However, because of the way Guido designed
lists and dictionaries, mutators typically return None, so it's awkward at
best to write lambdas that use them. This encourages side-effect-free lambdas.

See this page for a good laugh:
http://p-nand-q.com/lambda.htm

>   Statements are, by definition, things which cannot appear in lambdas ;)
> No worries here, I think.

Except that there are some that consider lambda "broken", and that it should
either support statements or be removed entirely. This I disagree with.

>   Luckily, map() and filter() are implementable in terms of list
> comprehensions.  So even in Python3k, when they're finally removed from the
> language, you can implement them yourself in Python, and gain the
> performance benefits of using a JIT on python bytecode ;)

So they are scheduled to be removed! I knew it! =) This is why I'm paranoid!

>   The inverse of zip is already available (and with one less character than
> unzip! ;).  x == zip(*zip(x)) (ignoring sequence types, of course).

Cool! Thanks for the tip...

>   I don't think Python is headed the way of Unlambda, don't worry :)

Hehe... it does need a COME FROM construct though. Who wants to help me
write the PEP for that one?

>   What I think is an important distinction to make here is the language vs
> the library.  Right now, there are overlapping language features.  I'd much
> prefer it if some of the things, even the useful things like map(), were
> re-implemented in terms of other Python features and moved out of the
> language and into the library.  This gives the best of both worlds - a
> language simple to learn and use that also provides an expressive set of
> features through its stdlib.

I agree with your premise, but I've been feeling lately that there isn't
such a clear distinction between language and library. For instance, when
the string object and the string module merged, the main difference was that
my sentences were rearranged and I didn't have to say "import string"
anymore. But I never really thought in terms of a string *library*, so on the
surface it was part of the language all along.

If I just have to say "from functional import *" to get an environment
friendly to FP, I will happily quiet down and mind my own business. I just
want to make sure FP is being represented. ;)

Thanks for your reply,
Dave



From peter at engcorp.com  Wed Apr 16 15:01:23 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 15:01:23 -0400
Subject: How many of you are Extreme Programmers?
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com> <mailman.1050504987.840.python-list@python.org> <3E9D8A80.FD76BDE9@engcorp.com> <b7k3kj$4gj$1@news.hananet.net>
Message-ID: <3E9DA883.CE0C38B4@engcorp.com>

Changjune Kim wrote:
> 
> Check out the Standish Group CHAOS report.
> 
> http://www.softwaremag.com/archive/2001feb/CollaborativeMgt.html (refers to
> 2001 report)
> http://www.pm2go.com/sample_research/chaos_1994_1.php

The first link redirects to http://www.softwaremag.com/magazine.css,
which is not found (404).

-Peter



From jhs_bkk at nospam.yahoo.com  Fri Apr 11 08:04:12 2003
From: jhs_bkk at nospam.yahoo.com (Jason Smith)
Date: Fri, 11 Apr 2003 19:04:12 +0700
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <3e96b088_2@127.0.0.1>

Harald Massa wrote:

> For education there are small drawbacks:
> 
> - 50% of all exercises "to teach programming" are already solved in the
> standard-library or are a method of a buildin object / are allready build
> in the core language
> 
> - 45% of these exercises are solved within "the python cookbook",
> available also online

Holey moley!  And all this time, I had it in my Safari queue!

> - if you learn programming with Python and later have to use Java or
> Visual Basic, it will be the most frustrating experience for the young
> fellows.

I don't think so.  Particularly with computer science courses, I think it's
often the case that students are required to re-implement features that
come with a library.  For example, consider a course where students
implement their own data structures in C++, and then they implement
iterators to mimic the STL.  (I actually took this course, so there's your
existence proof.)  If a student uses code from a book, or say some GNU
code, then that's simply academic dishonesty; and any good teacher would or
should catch these things as part of their general grading techniques.

Where I think python would be great is at the lower educational levels.  I
think we all agree (heh) that programming as part of a general computer
literacy course should be necessary for any high-school diploma today, just
like Chemistry or such.  And in that arena, as a student's first exposure
to programming, Python shines.

I mean, if we're all studying sorting, and I turn in

  a = [ 'this', 'is', 'a', 'list' ]
  a.sort()

well then clearly I just got an F.  But it's not Python's fault.


----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---



From ckea25d02 at sneakemail.com  Tue Apr  8 21:20:39 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 09 Apr 2003 13:20:39 +1200
Subject: What are the differences between SWIG and Boost.Python?
In-Reply-To: <mailman.1049824396.32551.python-list@python.org>
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org> <XSXja.346441$3D1.186570@sccrnsc01> <mailman.1049741353.6320.python-list@python.org> <tntka.364035$F1.57656@sccrnsc04> <mailman.1049824396.32551.python-list@python.org>
Message-ID: <b6vrv3$9d6pt$1@ID-169208.news.dfncis.de>

Mike Rovner wrote:
> "Bryan" <belred1 at yahoo.com> wrote in message
> news:tntka.364035$F1.57656 at sccrnsc04...
> 
>>how is pyrex at handling a c varargs function?
> 
> Never tried myself, but ToDo.txt says:
> '''
> DONE - Varargs C func defs and calls.
> 
> Varargs argument traversal.
> '''

Just to expand on that a bit, you can declare external
C functions which take variable parameters, and you can
call them.

What there isn't any explicit support for yet is using
the varargs facilities (va_start, va_arg etc) to step
through variable argument lists and process them.

That's not to say it can't be done -- chances are you
could declare these as though they were functions, and
use typecasts liberally. I haven't tried, though.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From puccio_13 at yahoo.it  Wed Apr 23 05:45:46 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Wed, 23 Apr 2003 11:45:46 +0200
Subject: python lexer
References: <8Wspa.23568$DN.669072@tornado.fastwebnet.it>
Message-ID: <qxtpa.23645$DN.670112@tornado.fastwebnet.it>

Alessio Pace wrote:

> Hi, I was wondering which is a good lexer written in Python, I heard about
> SPARK and PLY. I need only to tokenize texts, not to build trees on their
> structure.
> string.split() I guess it is not enough, because I can't use reg exp as
> token separators..How could I do otherwise?
> Thanks
> 

Ehm.... it seems like I solved...

import re

tokens = re.split("reg_exp", string_text)

:-)



-- 
bye
Alessio Pace



From mike at bindkey.com  Fri Apr 18 13:28:18 2003
From: mike at bindkey.com (Mike Rovner)
Date: Fri, 18 Apr 2003 10:28:18 -0700
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de>
Message-ID: <b7pchr$ig0$1@main.gmane.org>

"Martin v. L?wis" <martin at v.loewis.de> wrote in message

> > Is there a short way to say why Guido didn't want continuations?
>
> He didn't like the implementation strategy; I believe he was neutral
> about the feature per se.

Second that.
AFAIK one of the points was/is the ability to implement it in Jython,
which is/was impossible while you are mangling with C stack frames.

> > Or should I go digging in the archives?
>
> Yes.

It depends on the level of details you want to know. ;)

Mike







From garycramblitt_NOSPAM at comcast.net  Tue Apr  1 17:39:02 2003
From: garycramblitt_NOSPAM at comcast.net (Gary Cramblitt)
Date: Tue, 01 Apr 2003 17:39:02 -0500
Subject: How to obtain python file path
References: <20030401163212631-0500@news.covad.net> <slrnb8jusc.5u1.ramen@lackingtalent.com> <20030401165546102-0500@news.covad.net>
Message-ID: <d9idnZaiLMUViBejXTWcoQ@comcast.com>

On Tue, 01 Apr 2003 21:58:12 +0000, Dan Grassi wrote:

> Not quite, that only provides the path to the initial .py file.
> 
> With a directory structure:
> fa
>     a.py
>     fb
>         b.py
> 
> cd .../fa
> execute a.py
> where a adds fb to sys.path
> a imports b.py
> 
> I need b.py to report it's path which would be .../fa/fb
> 
> 
I might have seen something on the python cookbook website.  As I recall,
the basic technique is to force an exception, then use the info in the
exception object to get the path to the .py file causing the exception.



From python at dylanreinhardt.com  Thu Apr  3 13:20:32 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 03 Apr 2003 10:20:32 -0800
Subject: Zope style (was: tuning our site (was: help! advocacy
	resources needed fast))
In-Reply-To: <v8oh3m1f28gl74@corp.supernews.com>
References: <mailman.1046819836.25704.python-list@python.org>
	 <irdbl-1ds.ln1@news.lairds.org> <v8mjffpuph1oc2@corp.supernews.com>
	 <mailman.1049323265.22242.python-list@python.org>
	 <v8oh3m1f28gl74@corp.supernews.com>
Message-ID: <1049394031.6719.243.camel@ida.dylanreinhardt.com>

On Thu, 2003-04-03 at 06:25, Cameron Laird wrote:
> acquisition is very much like object orientation, in its
> calculus of inheritance (although I think I have a theorem that
> acquisition is harder to get right--more on that, later), and,
> of course, notationally.  acquisition is also at least as dif-
> ficult to design correctly.  

The comparison is an apt one, and it's clear what inspired acquisition. 
It's a bit like composition by context.

But there is a bit of a conceptual dead end you're going to hit by
defining acquisition in terms of inheritance.  Acquisition can just as
easily be thought of as a way of providing abstraction and polymorphism.

Thinking of acquisition that way, the expressive power of templates
becomes greater.  Instead of thinking about what logical operations
should be performed, it is possible to think in terms of contextual
behavior.  This can shift the balance between logic and presentation
enough that templating is sufficient for a greater range of tasks.

Say my_method uses some markup we've saved seperately as my_widget. 
It's pretty straightforward to set up my_method such that it checks for
an overriding my_widget and uses the standard my_widget if no other is
found.  It is now trivially easy to implement custom changes for
clients, as each change requires only that the differing code be stored
in a client-specific Folder.  One glance at each client's root folder is
sufficient to display what customizations have been made on their behalf
and hardly any logical operations are used at all.

If you get more that a handful of clients on one box, this can be a huge
improvement over burying the implementation details in the code of each
method.  Being able to view customizations *by client* instead of *by
method* can be very helpful.

Of course, you can do the same thing with Python scripts... but I find
aggressive use of acquisition tipping the scales towards DTML more often
than not.


> it's always--in my experience--more costly to
> set up a rational acquisition structure than to work with Python
> scripts (themselves occasionally boosted by acquisition).

It's been my experience that acquisition structures are often needlessly
complex.  Then again, I make heavy use of custom products, which makes
it quite a bit easier to simplify your folder hierarchy.

> The symptom of acquisition's liabilities is complexity.  It appears
> to me simply to take too much discipline to keep DTML-oriented
> acquisition hierarchies from becoming incomprehensible.

Well... clarity and strength of design are issues with software in
general.  It takes a lot of discipline to come up with a good design and
manage the changes that are made to it during implementation.  You can
make spaghetti sites with DTML quite easily.  On the other hand, design
discipline pays off in most software tasks... why shouldn't it in Zope?

> 
> Understand, I'm skeptical about the fine divisions of labor that
> appear to prevail on some projects:  a Web application designer,
> a Web architect, a Web security administrator, several Web visual
> designers, and so on.  

As with most group software projects, design control and revision
control are critical.  It's slightly more difficult to do revision
control on Zope objects than it is on source files... but not so
difficult that it isn't worth doing.


> Maybe the real issue is just the asymmetry I feel; I much prefer to
> add """-coded presentation to a script, than <dtml-ish algorithms
> to DTML.

For me, it depends on what's being done.  I find inserting multiple
values into a page of markup is more readable in DTML.  Some HTML --
particularly form controls -- seem better suited for Python.

Examples:

--------

Hello, <dtml-var fname>.  This is demonstration of MySoft version
<dtml-var software-version>.  Just think, <dtml-var fname>, now 
you can shift paradigms and think outside the box at 
<dtml-var company> without leaving your office.

vs.

my_output = """Hello, %s.  This is demonstration of MySoft
         version %s.  Just think, %s, now you can shift 
         paradigms and think outside the box at %s without 
         leaving your office.""" % (fname, version, fname, company)
return my_output

--------

I find the former to be far more readable and more easily maintained...
particularly as the size of the sample increases.  Inserting additional
variables or moving around existing ones is easy in the first sample and
must be done carefully in the second.

On the other hand:

--------
<select name="<dtml-var select_name>">
<dtml-in my_key_value_pairs prefix=my>
    <option value="<dtml-var my_key>"
    <dtml-if "my_key==my_selected_option"> selected</dtml-if>
    ><dtml-var my_value></option>
</dtml-in>
</select>

vs.

output = ['<select name=%s>' % select_name]
selected = {1:' selected', 0: ''}
option_tag = '<option value=%s%s>%s</option>'
for kv_pair in key_value_pairs:
    output.append(option_tag % (kv_pair[0], 
                                selected[kv_pair[0]==selcted_key], 
                                kv_pair[1]))
output.append('</select>')
return ''.join(output)

- or -

<input type=text name=foo value="<dtml-var "my_method('argument')">">

vs.

return '<input type=text name=foo value="%s">' % my_method('argument')

-------

I often go with Python for form controls since Python handles logic and
quoting with more grace. But many times it's a toss-up which is better
for a given situation.  Really, they're *both* good tools to have at
your disposal.

Dylan





From olczyk at interaccess.com  Sat Apr 19 01:39:10 2003
From: olczyk at interaccess.com (Thaddeus L Olczyk)
Date: Sat, 19 Apr 2003 05:39:10 GMT
Subject: Why functional Python matters
References: <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <b7m546$2a83$1@news.peterlink.ru> <lyptnlxq1n.fsf@cartan.de> <m24r4xdyd6.fsf@mycroft.actrix.gen.nz> <lyd6jlxkpi.fsf@cartan.de> <usft9v01l8fcdg8erdl2kishg0cnr1t9dp@4ax.com> <lyadepp1ck.fsf@cartan.de> <1lt0avslergps3bdv1lpvr48kef6t03640@4ax.com> <87d6jjjuzu.fsf@darkstar.cartan>
Message-ID: <bdh1avcltqircf31cnkfro4thkafe2pamq@4ax.com>

On 19 Apr 2003 00:33:57 +0200, Nils Goesche <ngo at cartan.de> wrote:

>Thaddeus L Olczyk <olczyk at interaccess.com> writes:
>
>> While it's true that in math and CS, many definitions evolve
>> over time ( ones that come to mind right now are
>> function/procedure [which were distinct, now considered the
>> same thing], fiber bundles, sheaves, topological space ... ),
>
>Actually, the resp. definitions of fiber bundles, sheaves and
>topological spaces /haven't/ changed much over time.  If they
>have changed at all, once they were called by that names.
>
When Gauss spoke of topological spaces they were more
like smooth manifolds ( smooth Hausdorff manifolds to be more
precise ). The definition changed over a period of time till
it settled on it's present definition mid 40's ( IIRC).

Steenrod discusses the problems of varying definitions
of fibre bundles in his book.

I'll skip sheaves since I don't remember the details.
>> that's all just bullshit thrown out to distract people from
>> *your stupidity*.
>
>Darn, I had hoped I'd get away with it one more time.
>
Don't you mean twice, as it's clear that this reply is 
another attempt at distraction, totally ignoring the
issues.

>> Thaddeus L. Olczyk, PhD
>
>I was somewhat hoping that now, having finally obtained your
>degree, 
Sorry to disillusion you, but I got my doctorate many years ago.
I stopped keeping track when it passed ten.

Perhaps you will learn from this mistake. Making unfounded
assumptions will often cause you to make dumb errors. Keep
this in mind when you start doing serious programming, as
it wil save you from a lot of  errors in your code.

>which is seemingly very important to you, 
Is it? I suppose so. A doctorate is not something just anyone
can accomplish. It takes a dedicated effort over a period of years,
something most people don't have the patience for.

Most of the time I really don't think about it. It's just something
I did in the past.


>you'd have
>become a bit more relaxed.  I was wrong.
You seem to be wrong quite a bit of the time. Perhaps you would
not be wrong as much if you did not make unfounded assumptions.
--------------------------------------------------
Thaddeus L. Olczyk, PhD
Think twice, code once.



From rmunn at pobox.com  Wed Apr  2 11:36:38 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 16:36:38 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <3E88C1D8.3070305@motorola.com> <3e8a57d0.55986091@news-server>
Message-ID: <slrnb8lfcv.ccp.rmunn@localhost.localdomain>

NOTE: Post re-arranged so quoting order makes sense. Please don't top-post.

BBFoto <bbfoto2002 at netscape.net> wrote:
> 
> On Mon, 31 Mar 2003 16:31:52 -0600, Stephen Boulet
><stephen.boulet at motorola.com> wrote:
> 
>>BBFoto wrote:
>>> This question is so basic and I am so lacking in this area, that I
>>> will completely understand if I am ignored.  
>>> 
>>> I received a cd from PC World with a game called PySol, among other
>>> things.  It says I need Python to play it.  It gives a reference, so I
>>> get Python and I load the game to my HD.  (Win98 platform)   Nowhere
>>> in the game does it tell me how to get it to run.  It refers to "drop
>>> it on Python."  I have no idea what that means.  
>>> 
>>> I run the Python executable and type in "PySol" at the prompt on the
>>> black screen and get "name not defined."  I can't find a .bat or .exe
>>> file for PySol and I am totally unfamiliar with Python.
>>> 
>>> If anyone could give me a clue, it would be greatly appreciated!
>>> 
>>> BBfoto
>>
>>You could fire up a "cmd" console, and type in:
>>
>>   C:\python22\python "C:\program files\pysol\pysol.py"
>>
>>Adjust the paths to fit your setup. Or you could associate ".py" files with 
>>the python executable.
>>
>>-- Stephen
>
> I copied your statement, changed the path and tried it.
> 
> "Syntax error" was the result.
> 
> Thanks for trying to help.
> 
> BB

What is the exact path to your Python executable?

What is the exact path to your pysol.py file?

What is the exact command you ran, and the exact error message you
received? A copy-and-paste from your DOS prompt window would be
perfect as an answer to this question.

If you can give us this information, we should be able to give you the
exact command you need to run. Until then, we're guessing blind.

BTW, please don't top-post. It makes the question-answer-response order
really confusing. See http://www.caliburn.nl/topposting.html for more
explanations on why you should bottom-post instead of top-post.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From grante at visi.com  Wed Apr 16 17:26:12 2003
From: grante at visi.com (Grant Edwards)
Date: 16 Apr 2003 21:26:12 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com> <3e9d9bb9$0$151$a1866201@newsreader.visi.com> <a8b7f07a.0304161314.361a80fa@posting.google.com>
Message-ID: <3e9dca74$0$170$a1866201@newsreader.visi.com>

In article <a8b7f07a.0304161314.361a80fa at posting.google.com>, A. Lloyd Flanagan wrote:

>> > python << DONE
>> > print 2
>> > for i in (1,4):
>> >      print i
>> > DONE
>> 
>> But how do you do it in a Makefile?
> 
> Err.  Put it in a script file, and call the script file from make?

That's one solution, but now you've got another file to keep
track of.

> I tried sticking it in a Makefile, can't get it to work.  So far.

How 'bout this:

all:
        python2 -c $$'for i in range(5):\n print i\n\n'
        

I think that the $'<string>' notation is a bash-ism.  If that's
a problem, then

        echo -e "for i in range(5):\n print i\n\n" | python2
        
should also work.  Adjust the echo options appropriately to get
interpretation of \n to work.

There may be other ways to get Make to interpret \n in a
string...

-- 
Grant Edwards                   grante             Yow!  I want to dress you
                                  at               up as TALLULAH BANKHEAD and
                               visi.com            cover you with VASELINE and
                                                   WHEAT THINS...



From peter at engcorp.com  Sun Apr 20 06:05:21 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 20 Apr 2003 06:05:21 -0400
Subject: xmlrpclib, connection refused? revisited
References: <ARi5a.9343$Ua.656611@news2.west.cox.net> <mailman.1045817162.16116.python-list@python.org> <3E562CCF.749B6D18@engcorp.com> <3ea24a46$1_3@omega.dimensional.com>
Message-ID: <3EA270E1.AD831A99@engcorp.com>

Mark Gibson wrote:
> 
> Peter Hansen <peter at engcorp.com> wrote:
> : John Abel wrote:
> :>
> :> I'm guessing your using SimpleXMLRPCServer?  What interface is the
> :> server bound too?  It needs to be bound to the IP, rather than
> :> localhost.  When specifiying the host/port, try using:
> :>
> :> (socket.gethostbyname( socket.gethostname() ), 8888 )
> 
> : Better yet, use ('', 8888) as your address to listen on *all*
> : interfaces, including the loopback interface (localhost).
> 
> : -Peter
> 
> I had the same problem, and using '' fixed it.  Can someone explain it what's
> happening?

A listening port can be bound to one or more interfaces on a
computer.  Some computers have only one physical network interface,
but generally there is also a "loopback" interface, which is 
always numbered 127.0.0.1.  Connections made from the computer
to 127.0.0.1 or localhost will never go out over the network,
but will be "looped-back" inside the lower levels of the network
software *inside* the computer.

If you bind only to socket.gethostname(), your socket is *not*
listening on the loopback interface.  This means connections 
made to "localhost" will not be answered!

If you bind only to localhost, connections made from outside
the computer, across the network, will not be answered.

Listening to '' causes the network software to bind the socket
to *all* interfaces, thereby allowing you to establish both local
and remote connections.

Googling for "loopback interface" and "localhost" and "TCP/IP"
or some similar set of words would doubtless give more information,
as would another read of the documentation for the socket module.

-Peter



From nanthan at siswa.utm.my  Sun Apr  6 23:31:49 2003
From: nanthan at siswa.utm.my (Pathmenanthan RamaKrishna)
Date: Mon, 7 Apr 2003 11:31:49 +0800
Subject: PYTHON EDITOR!!
Message-ID: <03040711314900.15230nanthan@siswa.utm.my>

hi,

im Nanthan, a postgraduate student from Malaysia.im very new to 
python language but im already getting use of it.Currently,i 
have been working in windows, but my development actually is on 
Linux, so can u suggest me the best editor(i mean use friendly 
and graphical,i dont want command line editor)for python?

is cooledit the right choice.

i hope u'll reply soon

thanks

Nanthan.R









From gh at ghaering.de  Thu Apr 24 05:24:59 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Thu, 24 Apr 2003 11:24:59 +0200
Subject: Using the MySQLdb module
In-Reply-To: <slrnbaf93p.had.news@cipw08.physik.uni-bonn.de>
References: <3EA6E75E.7070305@cis.ohio-state.edu> <slrnbaf93p.had.news@cipw08.physik.uni-bonn.de>
Message-ID: <3EA7AD6B.90007@ghaering.de>

Carsten Gaebler wrote:
> In Michael Mayhew wrote:
> 
>>_mysql_exceptions.OperationalError: (1054, "Unknown column 'atr' in 
>>'field list'")
>>
>>sql = "insert into refFlatHum values ( %s, %s, %d, %d )" % ('atr', 
>>'chr1', 90810, 92190)
>>
>>c.execute(sql)
> 
> 
> What you need here is documentation on SQL, not on MySQLdb. :-) If you want to
> use strings in a query you need to enclose them in single or double quotes but
> the query string in your example actually looks like this:
> 
> "insert into  refFlatHum values ( atr, chr1, 90810, 92190)"
> 
> The correct syntax would be:
> 
> "insert into  refFlatHum values ( 'atr', 'chr1', 90810, 92190)"

So far, so good.

> Note the single quotes around atr and chr1. To achieve this in your code you
> will have to say:
> 
> sql = "insert into refFlatHum values ( %s, %s, %d, %d )" % \
> ("'atr'", "'chr1'", 90810, 92190

Please don't construct statements manually like this. You'll save 
yourself a lot of potential trouble if you let the DB-API module do the 
quoting, like I described in my other post:

cu.execute("insert into refFlatHum values ( %s, %s, %s, %s )",
	("atr", "chr1", 90810, 92190))

Note the use of %s for all types. The DB-API module will quote each 
supported type in the correct way. For numbers, the correct way is to 
just insert the str() representation in the query. For strings, the 
correct way is to enclose them in single quotes, and escape any existing 
single quotes with another single quote. For BLOBs and whatnot, the 
correct way is another one, but I'll leave these details to the DB-API 
module :-)

-- Gerhard





From bob_tommy at hotmail.com  Mon Apr 28 19:23:28 2003
From: bob_tommy at hotmail.com (tommys)
Date: 28 Apr 2003 16:23:28 -0700
Subject: Jython, getting instance name from the interpreter
Message-ID: <f795454b.0304281523.3d543e22@posting.google.com>

Does anyone know to get the instance names from the interpreter?
I want the getMyName method to return "instance" in this case, or a
list of names if many...

Thanks for any help!!

Tommy

public class InterPreter() {
	private static PythonInterpreter interp = new PythonInterpreter();
	private static String myscript = "instance = MyClass()";
	public InterPreter() {interp.exec(myscript);}


	public static String getInstanceName(Myclass mycl) {

	//hmm, how to do this?

	}
}
public class MyClass() {


public String getMyName() {
	return InterPreter.getInstanceName(this);
}



From spam at mattbergin.co.uk  Mon Apr 14 07:28:17 2003
From: spam at mattbergin.co.uk (Matt Bergin)
Date: Mon, 14 Apr 2003 12:28:17 +0100
Subject: Reference Tracking
In-Reply-To: <z7lma.10014$LB6.267626@news1.tin.it>
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it>
Message-ID: <b7e60e$re$1@wisteria.csv.warwick.ac.uk>

Alex Martelli wrote:
> Matt Bergin wrote:
>    ...
>>I'm just worried that if an exception occurs in a thread then the cursor
>>will never be returned, using the 'no black magic' method which appeals
>>to me the most. Could I prevent this by careful use of try: finally?

> Yes.  And you can *in addition* keep the cursors you've handed out
> in a list and ALSO periodically do the scan checking refcounts just
> to help you debug user threads that aren't returning cursors properly
> (or use other tricks as above but for debugging-help rather than
> functional purposes).

Just out of interest, when I run a very simple script, i.e.

import sys
a=1
print sys.getrefcount(a)

Why is the refcount of a so high (50 from the command line)? Where are 
these references lurking? Should I check be checking for changes in the 
refcount of an object rather than its absolute value?

-- 
Matt Bergin




From staschuk at telusplanet.net  Wed Apr 16 22:37:38 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 20:37:38 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <3E9DF377.E94BFCB8@engcorp.com>; from peter@engcorp.com on Wed, Apr 16, 2003 at 08:21:11PM -0400
References: <mailman.1050525028.7854.python-list@python.org> <b7km3s$238m$1@nntp6.u.washington.edu> <mailman.1050537782.19056.python-list@python.org> <3E9DF377.E94BFCB8@engcorp.com>
Message-ID: <20030416203738.C6846@tibia.amotlpaa.bogus>

Quoth Peter Hansen:
  [...]
> I've always considered __repr__ to be for use in *debugging* the
> application, where a *debugging* log needs to display information,
> or in a traceback.  Since it's next to impossible for most interesting
> objects to be reproduced easily from the output of __repr__ if it's
> also desirable that the output be meaningful and useful to a human 
> reader, I don't buy into the 'eval(repr(x)) == x' theory very much,
> other than to note that it's the most obvious thing to do for 
> primitive data types.

There are, I admit, lots of objects for which the eval(repr(x)) ==
x theory doesn't make sense.  (File objects, for example.)  But in
my (admittedly not very broad) experience, these tend to be
objects for which == doesn't make sense in the first place.

Imho the notion that eval(repr(x)) == x theory makes lots of sense
for value objects generally.  (See, e.g., the timbot's example of
datetime.date.)

> I've never had any confusion in my own mind about when it's 
> appropriate to use one or the other.  Perhaps this perspective 
> should be used in the documentation, rather than the focus on 
> reproducing objects from their __repr__.  

*shrug*  As quoted in another post, the docs explicitly state that
repr "is typically used for debugging".  I suppose it could be
more emphasized.

-- 
Steven Taschuk                                                   w_w
staschuk at telusplanet.net                                      ,-= U
                                                               1 1




From cpl.19.ghum at spamgourmet.com  Tue Apr 29 18:49:23 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Wed, 30 Apr 2003 00:49:23 +0200
Subject: HTTP application streams a response to the client gradually - how  deal with it?
References: <Xns936CB5697A6D4cpl19ghumspamgourmet@62.153.159.134> <3EAEC97C.1EADEBF7@hotmail.com>
Message-ID: <Xns936D814A89Fcpl19ghumspamgourmet@62.153.159.134>

Alan,

> If you disagree that HTTP is not suitable for a chat application,
> please feel free to say why.

I don't want to write a Chat Server via HTTP. Chat-Servers are just three 
imports away using twisted - and IRC, Jabber.... are all designed to be 
Chat-Protocols.


There is a existing Chat server, using HTTP / HTML. Currently it only 
works with Internet-Explorer as a client. Lynx is unlucky, Mozilla works 
only half.

So I want to do a Chat Client in Python. Perhaps translating from the 
HTTP-Stream to another protocol.

I think it's not a good idea to use HTTP for Chat - but it is used. 
 
> If your issue is with the use of port 80 (or whatever HTTP port is in
> use), often being the only port that goes through firewalls, then 
>that's  a different issue from trying to use HTTP for what it wasn't 
>designed for.
The issue is: HTTP is used for this kind of stuff. And as you mention - 
it was not designed for it. 

So your feedback was valuable - especially the part about Request / 
Response. It looks as if I have to go the socket way. Maybe a second try 
with twisted.

Harald



From staschuk at telusplanet.net  Wed Apr 16 20:02:08 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 18:02:08 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <b7km3s$238m$1@nntp6.u.washington.edu>; from donn@u.washington.edu on Wed, Apr 16, 2003 at 10:39:56PM +0000
References: <mailman.1050525028.7854.python-list@python.org> <b7km3s$238m$1@nntp6.u.washington.edu>
Message-ID: <20030416180208.A6543@tibia.amotlpaa.bogus>

Quoth Donn Cave:
  [...]
> It's clear that the docs promote this idea, but they also hint at
> its half-baked quality - repr() evals to the original value, unless
> it doesn't.  I don't think the idea is clear at all.
> [...] The only kind of clarity this documentation possesses is
> when it lets you read into it what you already thought.  Not the fault
> of the documentation, though, but of the unsound ideas behind it.

I'm of the opinion that the ideas are sound; if there is a
problem, I think it must lie in the documentation.

The docs use the nice phrase "information-rich and unambiguous"
for the output of __repr__.  The notion that ideally we'd have
eval(repr(x)) == x follows from this; there can be no more
information-rich and unambiguous representation than one which
would cause the runtime to reconstruct the object.

But information-rich and unambiguous representations, however
useful for (e.g.) debugging, are not always easy to read or
understand; for the humans then, we'd like an "informal",
easier-to-read representation, whence __str__.

What's unsound here?

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From graham__fawcett at hotmail.com  Wed Apr  2 10:50:38 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 2 Apr 2003 07:50:38 -0800
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de> <e9570f37.0304011121.1873dac2@posting.google.com> <pan.2003.04.02.08.34.27.966495.810@thomas-guettler.de>
Message-ID: <e9570f37.0304020750.645a8c86@posting.google.com>

Hi Thomas,

Thomas Guettler wrote:

>I just saw your message called "Quixote Server on top of Twisted
Python"
>in Quixote-Users of last week.
>
>What is your intend of doing this? Do you want to use PTL, or do you
want
>use it to use ZODB? I haven't looked at twisted in detail, but it
should
>be possible to use twisted with ZODB without quixotte, too.
>

The short answer: modularity. I want to build systems with components
that each do one thing, and do that thing well.

Now the long answer... ;-)

I look at Twisted as a Web server written in Python, and nothing more.
Twisted is an excellent general-purpose networking framework. Twisted
certainly is  more than that, for example, it has its own Web
application framework (the twisted.web package) that provides form
"widgets" and other conveniences for application programmers. But, I
don't want it to be all of that -- that's serving multiple purposes.

That's not good separation of concerns, IMHO. If I write an
application based on the twisted.web framework, then I am commited to
running that application on a Twisted Web server -- I can't swap out
Twisted and swap in another Web server if the need arises. For
example, if I needed to move to HTTPS, the easiest way for me to do it
would be to move the application to an Apache server that already had
HTTPS installed and configured. But I couldn't easily do this because
of my application's dependency on Twisted as a *Web server*, not just
as an *application framework*.

Quixote is not a Web server, and it doesn't care what Web server you
use. There is a bias toward Apache, I think mainly because that's what
the principal developers use. But this bias hasn't affected the
framework, it's truly server-agnostic.

Andrew Kuchling added a "driver" to let Quixote run on the Medusa Web
server. My script is (hopefully) the starting point of a similar
"driver" to let Quixote run on top of the Twisted Web server. (There
is another, probably better, way to do this: see the other discussion
on the quixote-users list in March re: the SCGI approach.)

So:

Twisted gives me a Web server. I disregard its Web application
framework. So I have a component that does a single thing well.

Quixote gives me a Web application framework. It interoperates with
(conceivably) any Web server, giving me great portability. Again, a
single thing well.

Lastly, the application I wrote makes use of ZODB. I love not having
to worry about persistence, and the design and performance of ZODB are
suitable for my application. And again, ZODB does persistence -- and
nothing else -- very well.

That approach minimizes my application's dependencies on other
people's code. Yes, my app depends on Quixote, but it doesn't depend
on any given Web server: I can use any one that suits my requirements.
Similarly it depends on ZODB, but the back-end of ZODB could be a
FileStorage, a BerkeleyDB, or anything else ZODB supports.

In support of Quixote, it is the cleanest Web app framework I've seen.
It's simple. It lets me do my work, and doesn't get in my way while
I'm doing it. It vanishes into the background, like any good framework
should.

Lastly -- and very importantly, I think -- you can read the Quixote
source in a couple of hours, and understand it from top to bottom,
inside and out. I tried that once with Zope2, and the nightmares
haven't gone away yet. Hopefully Zope3 will be more readable, but I
think you'd need a bit more time to understand what it's doing
underneath the surface.

Hope that helps,

-- Graham



From nessus at mit.edu  Mon Apr 21 17:29:48 2003
From: nessus at mit.edu (Douglas Alan)
Date: Mon, 21 Apr 2003 17:29:48 -0400
Subject: A new and very robust method for doing file locking over NFS?
References: <lcu1cw525a.fsf@gaffa.mit.edu> <mailman.1050665494.2487.python-list@python.org>
Message-ID: <lcznmjo7xv.fsf@gaffa.mit.edu>

Andy Jewell <andy at wild-flower.co.uk> writes:

> It depends on how other processes are playing with your files; what
> if another program is just scanning directories and opening files at
> random?  If this program doesn't know your locking scheme, it'll
> just walk all over your process...

This is just business as usual for Unix, so I'm used to living with
this worry.  File locking on Unix is typically only advisory, even
when locking files exclusively on local filesystems, and I've seen on
occasion, for instance, Linux distributions where the mail delivery
agent and the mail reading agent were compiled to use different
advisory locking methods.  When this happens the result, of course, is
not so robust.

In any case, I am glad to hear that you have had success with a very
similar strategy as mine.  It gives me some confidence that my
approach will work well.

|>oug



From exarkun at intarweb.us  Wed Apr  9 18:56:39 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 9 Apr 2003 18:56:39 -0400
Subject: Macros in Python?
In-Reply-To: <b71qde$q71$07$1@news.t-online.com>
References: <b71qde$q71$07$1@news.t-online.com>
Message-ID: <20030409225639.GB10833@meson.dyndns.org>

On Wed, Apr 09, 2003 at 08:58:04PM +0200, Dominic wrote:
> 
> 
> (defmacro with-root-privs (() &body body)
>   (let ((oldidsym (gensym)))
>     `(let ((,oldidsym (geteuid)))
>        (seteuid 0)
>        (unwind-protect
> 	   (progn , at body)
> 	 (seteuid ,oldidsym)))))
> 
> Is there an easy way to produce similar code
> in Python which gurantees that certain things
> happen around some code? (Maybe by using
> generators?)

  Yay metaclasses yay!

    class WrapperMetaclass:
        def __init__(self, prefix, wrapper):
            self.prefix = prefix
            self.wrapper = wrapper
        def __call__(self, name, bases, attrs):
            for k in attrs:
                if k.startswith(self.prefix):
                    attrs[k] = self.wrapper(attrs[k])
            return type(name, bases, attrs)

> [snip]
> 
> Is something like this feasible?
> 
> def server(a,b):
>   if cond:
>       ...
>   else:
>     With_Root_Privs:
>        if x in user:
>           active.append(x)
>          (...)
> 

    def acquireRoot():
        # Whatever
    def releaseRoot():
        # Whatever
    def superuserWrapper(f):
        def function(*args, **kw):
            acquireRoot()
            try:
                f(*args, **kw)
            finally:
                releaseRoot()
        return function

  class Server(object):
      __metaclass__ = WrapperMetaclass('asroot_', superuserWrapper)

      def asroot_addUser(self, x):
          active.append(x)

      def checkCondition(self, a, b):
          if cond:
              ....
          else:
              self.asroot_addUser(x)

  Jp

> -- 
> http://mail.python.org/mailman/listinfo/python-list

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 20 days, 19:01, 1 user, load average: 0.21, 0.15, 0.05




From Padraig at Linux.ie  Tue Apr  1 08:14:48 2003
From: Padraig at Linux.ie (Padraig at Linux.ie)
Date: Tue, 01 Apr 2003 14:14:48 +0100
Subject: spawning a separate process
In-Reply-To: <b6c2c6$lr7$1@aspen.sucs.soton.ac.uk>
References: <b6c2c6$lr7$1@aspen.sucs.soton.ac.uk>
Message-ID: <3E8990C8.4020804@Linux.ie>

Mark Light wrote:
> Hi,
>        I have a python program from which I wish to spawn an os.system call
> (a bash script) that may take several minutes to run - whilst the rest of
> the script carries on. I have tried using bits of code from books - but it
> never seem to work quite right

how about using the popen2.Popen3 class?
something like:

import popen2
child=popen2.Popen3("mycommand")
while 1:
     if child.poll() == -1:
         #child not finished yet
     do other stuff

note the module is implemented in python
so you can look at it in your distribution.

P?draig.




From exarkun at intarweb.us  Tue Apr 15 16:05:20 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 15 Apr 2003 16:05:20 -0400
Subject: Python and Schools
In-Reply-To: <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
Message-ID: <20030415200520.GA5198@meson.dyndns.org>

On Tue, Apr 15, 2003 at 07:07:27PM +0000, Greg Brondo wrote:
> Aahz wrote:
> >In article <v9e21mhemm8c2 at corp.supernews.com>,
> >Paul Watson <pwatson at redlinec.com> wrote:
> >
> >>We need to teach students correct design priciples to build something
> >>greater than what already exists.  We will never get very far if we
> >>require everyone to start with a quark or atom.  Yes, of course we need
> >>some people who design silicon and create microcode.  They will learn
> >>the low-level details what they need to know as they need it.  Knowing
> >>great design and organization principles will enable them to make the
> >>most of it.
> >
> >
> >While there's some truth to that, try explaining why the following code
> >is a Bad Idea to someone who has no basic understanding of CS:
> >
> >    s = ''
> >    for i in range(1000000):
> >        s += str(i)
> >
> >Knowing the difference between O(N) and O(N^2) is critical to writing
> >even the simplest programs.  OTOH, I do agree with you that focusing on
> >programming as a craft is more important to being a programmer than
> >learning CS.
> 
> I understand why the code is bad (copying the string each time in
> memory as it grows).  However, I've never understood the notation's
> O(N) and O(N2)?

  Big-O is just a way to formally describe the complexity of an algorithm. 
It tends to only be applied to the most complex (and therefore most
expensive) portion, and all the lesser parts are ignored.  It is almost
always in terms of some variable, usually (but not always) one with an
obvious meaning.  N above refers to the number of times the concatenation is
performed.  So O(N) simply means the algorithm has linear complexity - that
is, the runtime cost of running this algorithm grows directly with the
number of concatenations performed.  O(N**2) would indicate quadratic
complexity - the runtime costs grow with the square of the number of
concatenations performed.

  It also occassionally makes sense to talk about the best case, the average
case, and the worst case complexity.  For example, quick-sorting an almost
sorted list can take much longer than quick-sorting one ordered such that
the partitions are always exactly at the center of each segment being
examined.  The former of these cases is the worst case and the latter is the
best case.  In practical programming, as I believe someone else has
mentioned in this thread, often only the average case is generally
considered.

  Also note that hidden things can affect the complexity of algorithms. 
With a malloc-dependant algorithm, such as string concatenation, while your
algorithm may be O(N) or O(N**2), the algorithms you depend on may increase
the overall complexity.  While you would still talk about your algorithm
having some complexity, the actual runtime performance might vary
significantly from this, and vary significantly from environment to
environment, where the hidden complexity changes due to different library
implementations.  For this reason, the practical use of Big-O notation is
fairly limited (IMHO).

  Hope this helps,

  Jp

-- 
http://catandgirl.com/view.cgi?90
-- 
 up 26 days, 15:02, 4 users, load average: 0.29, 0.10, 0.02




From belred1 at yahoo.com  Tue Apr  8 01:47:05 2003
From: belred1 at yahoo.com (Bryan)
Date: Tue, 08 Apr 2003 05:47:05 GMT
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org> <XSXja.346441$3D1.186570@sccrnsc01> <mailman.1049741353.6320.python-list@python.org>
Message-ID: <tntka.364035$F1.57656@sccrnsc04>

how is pyrex at handling a c varargs function?

bryan


"Mike Rovner" <mike at bindkey.com> wrote in message
news:mailman.1049741353.6320.python-list at python.org...
> "Bryan" <belred1 at yahoo.com> wrote in message
> news:XSXja.346441$3D1.186570 at sccrnsc01...
> > mike,
> >
> > you mentioned small c libs you use pyrex.  what about small but complex
> data
> > types?  we have a structures that contain pointers to structures that
>
> Sure Pyrex has its limits. My picture was general. In each particular case
I
> weight pro and contra.
>
> But it's especially nice to wrap C with Pyrex because it VERY easy to
adapt
> interface to pythonic way.
> Say I had a lib with (void*) handle parameter to pass almost every data
> structure and a bunch of different real structures.
> With Pyrex an ability to redefine was a plus - hard things like swig and
c++
> will not do in that case (nicely I mean).
> Complex (struct inside struct) data was no problem.
> I also use a script to convert an .h file to Pyrex - it was simple
15-liner
> __to run once__:
>
> #!/usr/bin/env perl
> /^\s*{\s*$/ && next;
> /^\s*} T.*$/ && next;
>
> s/;//;
> s/^\t//;
>
> s|/\*|\#|;
> s|^ *\*|\#|;
> s| *\*/$||;
>
> s/^\s*typedef +struct +SSX([A-Z_]+)/cdef struct TSX$1:/; $n=$1 if $1;
> s/^\s*struct/\ncdef struct $n:/;
> s/^\s*}/\t$n/;
>
> print;
>
> WBR,
> Mike
>
>
>
>





From skip at pobox.com  Thu Apr 10 10:10:47 2003
From: skip at pobox.com (Skip Montanaro)
Date: Thu, 10 Apr 2003 09:10:47 -0500
Subject: newbie: strange python problem
In-Reply-To: <pan.2003.04.10.12.27.06.600319@somedomain.com>
References: <pan.2003.04.10.12.27.06.600319@somedomain.com>
Message-ID: <16021.31591.803175.116228@montanaro.dyndns.org>

    maxx> #!/usr/bin/python

    maxx> import os

    maxx> for line in os.popen('ls -la /').readlines():
    maxx>     direntry = line.split()
    maxx>     print direntry
    maxx>     print direntry[0]
    maxx>     print direntry[1]
    maxx>     print direntry[2]
    maxx> ---

    maxx> No list-items with an index above 1 are printed. They produce a
    maxx> 'IndexError: list index out of range' error... which is strange,
    maxx> because the list 'direntry' is 9 items large, so listitems up to
    maxx> index 8 should be printed... right?

Note that you are munch on the precise output of ls -la /, which includes
the summary line at the start of the listing, not just the directory entries
themselves.

Try os.listdir("/") instead:

    >>> print os.listdir("/")
    ['.DS_Store', '.hidden', '.Trashes', '.vol', 'Applications',
    'Applications (Mac OS 9)', 'automount', 'backup', 'bin', 'cores',
    'Desktop (Mac OS 9)', 'Desktop DB', 'Desktop DF', 'Desktop Folder',
    'dev', 'Developer', 'Documents', 'etc', 'include', 'Installer Log File',
    'Library', 'mach', 'mach.sym', 'mach_kernel', 'Network', 'private',
    'sbin', 'Shutdown Check', 'STV0680 Camera Snapshots', 'sw', 'System',
    'System Folder', 'Temporary Items', 'TheVolumeSettingsFolder', 'tmp',
    'Trash', 'Users', 'usr', 'var', 'Volumes'] 

That will only give you the names, but you can then call os.stat for each
entry to get the other information in the ls listing.

Skip




From usenet at microtonal.co.uk  Thu Apr 10 05:22:15 2003
From: usenet at microtonal.co.uk (Graham Breed)
Date: 10 Apr 2003 02:22:15 -0700
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com> <3E935B75.64968B9@engcorp.com> <b71s3u$lf3$01$1@news.t-online.com> <3E947BBB.48B85048@engcorp.com>
Message-ID: <1d22f12e.0304100122.2c779dc2@posting.google.com>

If you do want to return that lone final character:

>>> re.findall('.{1,2}', 'hello world')
['he', 'll', 'o ', 'wo', 'rl', 'd']



From boudewijn at tryllian.com  Mon Apr  7 09:02:54 2003
From: boudewijn at tryllian.com (Boudewijn Rempt)
Date: Mon, 07 Apr 2003 15:02:54 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <3e91777f$0$49105$e4fe514c@news.xs4all.nl>

Patrick K. O'Brien wrote:

> If you were crazy enough to think that you could create a program
> along the lines of Emacs, but written in Python, how would you go
> about it?  How would you design the domain model for files, buffers,
> windows, and frames?  How would you allow the same level of
> customizability?  How would you map functions (or methods or whatever)
> to keys?  Any thoughts?
> 

About a year ago I began something like that (mostly to teach myself how
an editor really worked). I extended Kalam -- the example application
in my PyQt book which used the Qt edit control -- with a backend that
represented buffers and a redisplay written in Python that painted the
glyphs and cursor on a widget with QPainter. Was quite fun until I lost
the source just as I had the redisplay working.

The best resource for someone implementing an editor from scratch is:

The Craft of Text Editing, or Emacs for the Modern World by Craig A. Finseth

http://www.finseth.com/~fin/craft/

-- 
Boudewijn | http://www.valdyas.org



From a.schmolck at gmx.net  Tue Apr  1 17:49:17 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 01 Apr 2003 23:49:17 +0100
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com>
Message-ID: <yfs65pxn89e.fsf@black132.ex.ac.uk>

"Max Khesin" <max at cNOvSisiPonAtecMh.com> writes:

> > Yes, but humans aren't good optimizing compilers.
> So who writes optimizing compilers? :).

Humans. If they were good optimizing compilers themselves why would they
bother?

Whether e.g.  assembler code can theoretically achieve better performance than
fortran code is irrelevant. What matters is whether it is humanly possible to
hand-code such assembler code.

For most cases it isn't and thus your argument (closer to machine level ->
automatically faster) breaks down.

(I'm not making any claims about the relative performance merrits of C++ and
python).

alex



From schull at digitalgoods.com  Tue Apr  1 22:09:24 2003
From: schull at digitalgoods.com (Jon Schull)
Date: 1 Apr 2003 19:09:24 -0800
Subject: recursion confusion
Message-ID: <2621b014.0304011909.d56a12a@posting.google.com>

Sometimes recursion befuddles me.  This is one of those times.

I want to convert a graph structure into a nested string 
For example I want this...

g={0:[], 1:[11], 2:[21,  22], 22:[221,222]}

...to become this
	"0, 	1[11], 2 [21,  22[221, 222]] "

(with 221 and 222 nested under 22 which is also nested under 2)

Of course, I want this to work for arbitrary nesting, and for strings
as well as numbers.

(Why? in order to create graphviz "record" shapes.)

I suspect there are concise recursive, and not-so-concise
understandable, ways to do this.  I'd be grateful for either, but
would prefer the latter.

Thanks very much...



From jepler at unpythonic.net  Tue Apr  1 19:34:38 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Tue, 1 Apr 2003 18:34:38 -0600
Subject: stop exeuction and give my a prompt
In-Reply-To: <b6dara$5mv$1@news.asu.edu>
References: <b6dara$5mv$1@news.asu.edu>
Message-ID: <20030402003438.GC30045@unpythonic.net>

On Tue, Apr 01, 2003 at 05:28:26PM -0700, David Bear wrote:
> I would like to have the ability to insert a statement in my code that would 
> stop program execution at that point and give me a python prompt.  I don't 
> know what to call this. 
> 
> What I would like to do is to stop program execution and inspect the state 
> of certain dictionaries --- I don't want to insert a bunch of print 
> statements and watch output fly of the screen.  I'd also like to be able to 
> fiddle with these dictionaries when my program stops.  
> 
> Any pointers?
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 

python -i gives you a prompt when your program would otherwise have
exited.

You should also be able to use something derived from code.interact() to
drop into an interactive console at any time, but exactly how you'd do
this is fuzzy. (by default, code.interact() works in a fresh namespace,
__console__)

Jeff




From mc at mclaveau.com  Wed Apr 16 17:19:45 2003
From: mc at mclaveau.com (Michel Claveau)
Date: Wed, 16 Apr 2003 23:19:45 +0200
Subject: wmi ?
Message-ID: <b7khea$kt2$1@news-reader13.wanadoo.fr>

Hi !            (sorry for my bad english)

I like pythonwin32Com & Idispatch.

But, i need of use an equivalent of  this VBS term :
       Set objWMIService = GetObject("winmgmts:\\server")

Can everybody give a way for this ?

Thanks in advance !

@-salutations
--
Michel Claveau





From glenfant at NOSPAM.bigfoot.com  Wed Apr  9 09:20:57 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Wed, 9 Apr 2003 15:20:57 +0200
Subject: PythonWin Open(file) path bug?
References: <3E93B4E8.6020003@netscape.net>
Message-ID: <b71603$18pq$1@biggoron.nerim.net>

Hi,

PythonWin files and misc resources are hosted at sourceforge now.

https://sourceforge.net/projects/pywin32/

You should use the appropriate tracker from there to write your bug
scenario.

Don't forget to say what build you use (menu Help > About PythonWin...)

--Gilles

"vector" <Vector180W at netscape.net> a ?crit dans le message de news:
3E93B4E8.6020003 at netscape.net...
> newbie,pythonwin,2.2.2
> im simply trying to open a file
> script snip...
> cfg_file = open("lptctrl.cfg", "rb")     #open  config text file
> end snip..
>
>   File "E:\Md_prj\md_422\Code\Python\LptCtrl_104.py", line 20, in __init__
>      cfg_file = open("lptctrl.cfg", "rb")     #open  config text file
> IOError: [Errno 2] No such file or directory: 'lptctrl.cfg'
>
> the file is in the same folder as the python script
> i get the same error if opened from the interactive window
> import sys
> print sys.path
> indeed shows the script dir is on the path so why cant it find the
> config file?
>
> If i call the script direct from explorer (no IDE) it works fine
> if I open lptctrl.cfg within pythonWin and save As it then runs and
> finds the file(it must be save as, even modifing and just save dosent
> change the error)
>
> any ideas
> a bug or am i dong something wrong
> I dont want absolute addressing, the file will always be in the "current
> dir"
>




From max at cNOvSisiPonAtecMh.com  Tue Apr 29 15:06:12 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Tue, 29 Apr 2003 19:06:12 GMT
Subject: stupid q-n (wxPy)
References: <Ukxra.33100$7M5.3263531@twister.nyc.rr.com> <3EAEB850.50D9DEB9@engcorp.com> <7hzra.27540$su3.3180311@twister.nyc.rr.com> <ullxt6tae.fsf@fitlinxx.com>
Message-ID: <E2Ara.27587$su3.3185854@twister.nyc.rr.com>

gotcha. good advice. thanks.


this copy-paste blooper was brought to you by...
--
========================================
Max Khesin, software developer -
max at cNvOiSsPiAoMntech.com
[check out our image compression software at www.cvisiontech.com, JBIG2-PDF
compression @
www.cvisiontech.com/cvistapdf.html]


"David Bolen" <db3l at fitlinxx.com> wrote in message
news:ullxt6tae.fsf at fitlinxx.com...
> "Max Khesin" <max at cNOvSisiPonAtecMh.com> writes:
>
> > The only detail I have is that the script was copy-pasted out straight
of
> > wxPython demo app. There were no specific instructions in the readme for
> > running the code so I just gave it a shot. Since wxPython is so popular
I
> > fugured someone would just tell me "didn't you bla-bla, (idiot)?".
> > anyway, thanks.
>
> Ah - it's the point about copying it from the "demo" that would have
> helped mentioning in your first post (as opposed to just saying
> "examples").  As it was, it's hard to know why the code snippet you
> gave was referencing the run module, since it's not a standard module.
>
> The wxPython demo is built around a framework that automates tying
> each demo module into the overall demo.  If you look in the same demo
> directory where you found this example, you'll find the run.py that
> you probably didn't copy over to where you are experimenting.
>
> Note that run.py is not part of standard Python, nor is it something
> that is documented as part of wxPython.  It's just part of the
> wxPython demo code base.
>
> My guess is that you don't really want to use this code (nor the
> run.py module) since that's geared specifically towards the way the
> wxPython demo presents itself.  Instead, run the wxPython demo, find
> the module that you're interested in, and then borrow the actual code
> that implements the behavior to put in your own module, and ignore the
> boilerplate framework that ties the module into the demo display.
>
> --
> -- David
> --
> /-----------------------------------------------------------------------\
>  \               David Bolen            \   E-mail: db3l at fitlinxx.com  /
>   |             FitLinxx, Inc.            \  Phone: (203) 708-5192    |
>  /  860 Canal Street, Stamford, CT  06902   \  Fax: (203) 316-5150     \
> \-----------------------------------------------------------------------/





From tweedgeezer at hotmail.com  Thu Apr 24 04:53:55 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 24 Apr 2003 01:53:55 -0700
Subject: SGMLParser questions
References: <igmbavg5ughls6ksiden2rgnael1s0ub8q@4ax.com> <m3n0ihdb80.fsf@mira.informatik.hu-berlin.de>
Message-ID: <698f09f8.0304240053.1b9e38e9@posting.google.com>

martin at v.loewis.de (Martin v. L?wis) wrote in message news:<m3n0ihdb80.fsf at mira.informatik.hu-berlin.de>...
> > 3. handle_starttag() turn tags into lower case. This work perfectly
> > for HTML. But can it works case sensitively?
> 
> No.
> 
> To get these features, you could subclass SGMLParser. However, you
> probably have to copy a lot of code into your subclass.

It seems it'd be easy enough to add a "case sensitive" keyword
argument to __init__ and sprinkle a few if statements before the
.lower()s on strings.  Would such a patch be accepted?

Jeremy



From santacruz at southern.edu  Mon Apr 14 11:51:46 2003
From: santacruz at southern.edu (dansan)
Date: 14 Apr 2003 08:51:46 -0700
Subject: A python server
Message-ID: <15f445fb.0304140751.170d410e@posting.google.com>

Hello all!

I have read and re-read the PY Sockets HOWTO, and I guess I'm still in
the dark about how to do some stuff.

I'm trying to write a little server that will allow connections via
telnet-like clients to play a board game.  The server would wait for a
wile for connections, and at some point, it would read a command from
the CLI on the server that says <startTournament>.  The server would
then generate a tournament schedule and start the games.  Since these
are turn-based games, the server would have to give each player its
<go> command (and start a timer) in the proper sequence.

I guess I'm having dificult with the higher-level desing of such a
system.  Right now my desing is very primitive, and has a LOT of
issues:

 # start
 # loop-waiting-for-connections (can't break out of this from cli yet)
 # make tournament schedule
 # run-games

I have program in other programming languages before, and lately I've
been doing some python.  So far I have enjoyed it a lot, so I wanted
to give this project a try in py.  Any help would be greatly
appreciated, in higher level desing as well as in lower-level code
snippets.  URL to documentation are also invaluable :)

Thanks a ton.
Dansan



From s.keim at laposte.net  Tue Apr  1 07:41:35 2003
From: s.keim at laposte.net (=?ISO-8859-1?Q?S=E9bastien_Keim?=)
Date: Tue, 1 Apr 2003 14:41:35 +0200
Subject: Slicing in Python 2.3a2
Message-ID: <466A3D6D-643F-11D7-9A2A-00306558A12A@laposte.net>

I have some troubles to understand how work slicing in python 2.3.
In my examples, I will use the following list:
 >>> A = range(10)

Python 2.3 add a new argument to slice wich can for sample be used to
get only the items at positive indexes:
 >>> A[::2]
[0, 2, 4, 6, 8]
This works also for subslices: A[x:y:z] is equal to A[x:y][::z]
 >>> A[1::2]
[1, 3, 5, 7, 9]

With a negative value you can reverse a list:
 >>> A[::-1]
[9, 8, 7, 6, 5, 4, 3, 2, 1, 0]

But I don't understand the result for a subslice with a negative third 
argument:
 >>> A[1::-1]
[1, 0]
Intuitively, I expected to get:
 >>> A[1:][::-1]
[9, 8, 7, 6, 5, 4, 3, 2, 1]

The numarray package has the same behavior, so I guess it's intentional.
 >>> import numarray
 >>> numarray.array(range(10))[1::-1]
array([1, 0])

Then what is the real semantic for this third argument?





From linux_pt2003 at netcabo.pt  Tue Apr  8 10:32:29 2003
From: linux_pt2003 at netcabo.pt (Carlos Sousa)
Date: Tue, 08 Apr 2003 15:32:29 +0100
Subject: PyQT
Message-ID: <pan.2003.04.08.14.32.29.728319@netcabo.pt>

Hi

I've some troubles in install PyQt with eric3, can someone help me ?

I install sip, qscintilla, Qt..but not work!


Thanks 



From lists at ghaering.de  Sat Apr 12 11:34:03 2003
From: lists at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Sat, 12 Apr 2003 17:34:03 +0200
Subject: distutils module
In-Reply-To: <b79a5t$gue$1@news-reader12.wanadoo.fr>
References: <b79863$dil$1@news-reader10.wanadoo.fr>
	<sPVla.6373$T34.168242@news2.tin.it> <b79a5t$gue$1@news-reader12.wanadoo.fr>
Message-ID: <3E9831EB.3020109@ghaering.de>

Salvatore wrote:
> Alex Martelli wrote:
>> Salvatore wrote:
>>> Hello,
>>>
>>> I've just installed Debian Knoppix,and i was surprised
>>> to see that distutils module could not be found.
>>>
>>> Does anybody experienced the same problem ?
>>
>> Not the specific one you mention, but I've often seen various
>> Linux distros package up Python in several and not necessarily
>> well-conceived sub-packages.  Personally, I always install Python
>> from sources -- I believe this reduces the amount of problems.
>>
>  Thanks Alex, I will do so

Alternatively, install the python-devel package.

-- Gerhard




From staschuk at telusplanet.net  Tue Apr 29 17:33:15 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 15:33:15 -0600
Subject: Determining Unicode encoding.
In-Reply-To: <ef7a350b.0304291242.42c9b024@posting.google.com>; from sean@activeprime.com on Tue, Apr 29, 2003 at 01:42:40PM -0700
References: <ef7a350b.0304291242.42c9b024@posting.google.com>
Message-ID: <20030429153315.B1725@tibia.amotlpaa.bogus>

Quoth Sean:
> I'm really new to dealing with unicode, so please bear with me.  I'm
> trying to add unicode support to a program I'm working on, and I'm
> getting stuck a little when printing a unicode string to a file.  I
> know I have to encode the string using an encoding (UTF-8, UTF-16,
> latin-1, etc).  The problem is that I don't know how to determine what
> the *right* encoding to use on a particular string is.  The way I
> understand it, utf-8 will handle any unicode data, but it will
> translate characters not in the standard ASCII set to fit within the
> 8-bit character table.  [...]

Actually characters outside ASCII get turned into multibyte
sequences by UTF-8.  To use an example that came up here a little
while back:

    >>> u = u'\N{DEGREE SIGN}' # U+00B0; not in ASCII
    >>> u
    u'\xb0'
    >>> u.encode('utf-8') # two-byte sequence
    '\xc2\xb0'

UTF-8 is capable of representing any Unicode character without
information loss, so if you need to deal with arbitrary Unicode
it's a good choice.  (It also has other pleasant properties.)

> [...] My problem is I'm handling data from a lot of
> different encodings (latin, eastern, asian, etc) and I can't allow
> data in the strings to be changed.  I also can't (at least I don't
> know how to) determine what encodings the strings are using.  IE, I
> don't know what strings are from what languages.  [...]

I think I detect a confusion or two here.

To encode is to turn (a sequence of) characters into (a sequence
of) bytes, and to decode is the reverse.  An encoding is a scheme
for doing these things; it need not be strongly associated with
any particular language.  (Though often an encoding can only
represent certain characters, and as a result is only useful for
those languages which use just those characters.)

A Unicode string (that is, a Python object of type 'unicode', such
as u'foo') is a sequence of characters, not bytes.  It therefore
is not in any particular encoding.

A normal string (that is, a Python object of type 'str', such as
'foo') is a sequence of bytes, not characters.  It can be
interpreted as a sequence of characters only if an encoding is
used to decode it.  (Usually ASCII is assumed if you do not
specify one explicitly.)

So, if you have a bunch of normal strings and don't know what
encodings they're in, you're hooped.  But if you have a bunch of
Unicode strings, it doesn't make sense to ask what encodings
they're in.

Now, as for not allowing the data in your strings to be changed:
If you mean you need to preserve the same sequence of characters,
then it's okay to change the encoding.  You'll almost certainly
want the file you produce to be all in one encoding, so you'll
want an encoding which can represent any character you might
encounter -- UTF-8, for example.

But if you mean you need to preserve the exact byte sequences,
you're hooped.  (Unless you'd be happy to have the output file be
a mishmash of different encodings and so virtually unusable.)

> [...] Is there any way to
> determine, from the unicode string itself, what encoding I need to use
> to prevent data loss?   Or do I need to find a way to determine
> beforehand what encoding they are using when they are read in?

You will have to know the encoding at the input stage; use that
information to decode the bytes into a Unicode string.  Then
assemble the Unicode strings to be output and encode them in, for
example, UTF-8.

In principle you might be able to look at the characters in a
Unicode string and determine some "least encoding" which could
deal with all the characters in it.  But there's not much point to
this, imho; just use UTF-8, which can handle anything.

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From aleax at aleax.it  Wed Apr 30 08:01:30 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 12:01:30 GMT
Subject: looping through a file
References: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au>
Message-ID: <uWOra.24347$3M4.683692@news1.tin.it>

<posted & mailed>

Psybar Phreak wrote:

> hi all - ive just started with python (after having done java and c) and
> am
> having a little bit of trouble with the script.  Im using as a login
> script.
> 
> There's a users.dat file in the format
> user1
> user1password
> user2
> user2password
> ... etc

So far so good.

> 
> at the moment - the code i work, only checks the first and second lines
> (ie.
> details of user1).  but it appears that python doesn't have a do...
> while... until loop.

Right, it doesnt't -- and indeed you don't need it for this purpose.


> #!/usr/local/bin/python
> 
> import cgi, string
> 
> form = cgi.FieldStorage()
> 
> if form.has_key('usernameField') and form.has_key('passwordField'):
>          users_username = string.strip(form['usernameField'].value)
>          users_password = string.strip(form['passwordField'].value)
> 
> InFile = open('users.dat', 'r')
> 
> file_username = string.strip(InFile.readline())
> file_password = string.strip(InFile.readline())

If the InFile is empty or has only one line this sets file_username
and/or file_password to empty strings (just as it does for all-white
lines, e.g. lines consisting just of a \n ["empty" ones]).  OK.

You're also using the string module -- that was necessary back in
the dark ages of Python 1.5.2, but with modern versions of Python
you should best use string methods instead.

> 
> authorised = 0
> 
> while file_username and file_password:
>        if file_username == users_username:
>               if file_password == users_password: # correct
>              authorised = 1
>       break
>               else:                               # login matched but not
> password
>                     file_username = InFile.readline()

You're omitting the strip calls here -- different semantics
from before.

>              file_password = InFile.readline()
>        else: # neither match
>               file_username = InFile.readline()
>               file_password = InFile.readline()

How complicated.  If the InFile can be expected to have a sensible
dimension, say no more than a few hundreds of KB at most (and in this
case that seems only reasonable!), just read all of its lines as
once with the readlines call and then index into the resulting list
of lines.  For example:

InFile = open('users.dat', 'r')
lines = InFile.readlines()
InFile.close()

for i in range(0, len(lines), 2):
    file_username = lines[i].strip()
    file_password = lines[i+1].strip()
    if file_username==users_username and file_password==users_password:
        authorized = 1
        break
else:
    authorized = 0

This may have slightly different semantics when InFile has an odd
number of lines, but I doubt you need to worry about that anyway.

I think this will most likely solve your practical problems.  However,
supposing for the sake of argument that InFile could be SO large that
reading it all at once with readlines would be a problem (say, tens
of megabytes), you could solve that too, for example, using again
modern Python (2.2 or better in this case):

InFile = open('users.dat', 'r')

for line in InFile:
    file_username = line.strip()
    file_password = InFile.next().strip()
    if file_username==users_username and file_password==users_password:
        authorized = 1
        break
else:
    authorized = 0

InFile.close()

Once again different semantics (an exception...) for odd numbers
of lines in users.dat -- you can use try/except StopIteration: around
the explicit call to InFile.next to deal with this as you wish.


Alex




From Norbert.Klamann at klamann-software.de  Thu Apr 10 02:32:12 2003
From: Norbert.Klamann at klamann-software.de (Norbert Klamann)
Date: 9 Apr 2003 23:32:12 -0700
Subject: Is Google sufficient ? was  no Python mail archiving utility?
References: <joe-17FB5F.10302009042003@orngca-news02.socal.rr.com> <mailman.1049911160.17811.python-list@python.org>
Message-ID: <65c27c6b.0304092232.23cebdc6@posting.google.com>

Skip Montanaro <skip at pobox.com> wrote in message 
[discussion snipped]

> via http, for most lists (those which are public) I suspect Google searching
> is sufficient, which lowers the demand for an offline search facility.
> 
> Skip

I don't think that google is sufficient to search mail archive pages.
There are a lot false positives due to the references to next and
previous mail which are found on these pages.

A search utility which knows about the structure of an e-mail would be
more powerful.

Norbert



From Curzio.Basso at unibas.ch  Mon Apr 14 05:41:04 2003
From: Curzio.Basso at unibas.ch (Curzio Basso)
Date: Mon, 14 Apr 2003 11:41:04 +0200
Subject: PCA principal component analysis
In-Reply-To: <3E96D39B.4C60CA87@sympatico.ca>
References: <b1ab591.0304090601.302cc4b6@posting.google.com>
 <hfYka.2414$h%2.256728@read1.cgocable.net>
 <a3b5b608.0304100901.58ced140@posting.google.com>
 <3E96D39B.4C60CA87@sympatico.ca>
Message-ID: <1050305123.21846.13.camel@libya.cs.unibas.ch>

On Fri, 2003-04-11 at 16:39, Colin J. Williams wrote:

> From my recollection, PCA can be applied to either the covariance 
> matrix or the correlation matrix. 

PCA of data is done finding the eigenvectors of the covariance matrix,
but this is best obtained from the SVD of the data matrix.

That is:

X is the data matrix, sample vectors are the columns
C = X*X' is the covariance matrix
C*P = L*P is the eigenproblem, the P matrix stores the components, L is
a diagonal matrix storing the eigenvalues
If X=U*S*V' through SVD, then

C=X*X'=(U*S*V')*(U*S*V')'=(U*S*V')*(V*S'*U')=U*(S*S')*U'

(because V'*V=I from the property of SVD)

=> C*U=(S*S')*U

which means P=U and L=(S*S').
So, the columns of U are the principal components, and the square of the
diagonal elements of S (which is a diagonal matrix) are the "weights"
which scale the components to the input space.

I hope I did not mess things up...

-- 
Curzio Basso <Curzio.Basso at unibas.ch>





From tim.one at comcast.net  Mon Apr 21 16:19:35 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 21 Apr 2003 16:19:35 -0400
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <yu99adel3thi.fsf@europa.research.att.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCEEMJEDAB.tim.one@comcast.net>

[Andrew Koenig]
> By way of further context for continuations, it might be worth
> mentioning that Scheme (an MIT dialect of Lisp) is based on
> continuations in a fundamental way.

It's also worth mentioning that SICP ("Structure and Interpretation of
Computer Programs" -- an excellent book all MIT Schemers know by heart)
doesn't mention Scheme's call/cc.

> Also, the Bell Labs/Princeton implementation of ML (``Standard ML of New
> Jersey'') is based internally on continuations, although this fact is
> pretty much invisible to user programs.

Par for the continuation course <wink>.  More recently, REBOL 1.0 supported
first-class continuations, but I *think* they were removed after Joe
Marshall stopped work on the REBOL interpreter.  Joe wrote a detailed
account of how continuations in REBOL 1.0 were used to nuke tail recursion:

    http://ll1.ai.mit.edu/marshall.html

There's a list of pros and cons at the end, the last of which is especially
telling:

    7. You need a degree in Comp. Sci. from one of a handful of places
       to understand this.  This turned out to be one of the biggest
       hurdles.  No one else in the company could extend the interpreter.
       Nonetheless, I don't see an easy way of achieving tail-recursion
       otherwise.

REBOL posed special problems, though, particularly in that you don't know at
compile-time where REBOL argument lists begin and end.





From mk at net.mail.penguinista  Mon Apr  7 10:11:48 2003
From: mk at net.mail.penguinista (=?UTF-8?B?0LTQsNC80ZjQsNC9INCzLg==?=)
Date: Mon, 07 Apr 2003 16:11:48 +0200
Subject: cx_Oracle and Number(11) 
Message-ID: <3e918724@news.mt.net.mk>

I have a column in the Oracle database that is defined as 
  ORDER_ID NUMBER ( 11),
but when I select that table with cx_Oracle 2.5 the values in the column
show up as floats in the python program.



-- 
??????                                  (jabberid:damjan at bagra.net.mk)

      There are three kind of lies: lies, damn lies and benchmarks.



From pinard at iro.umontreal.ca  Sun Apr 20 22:05:52 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 20 Apr 2003 22:05:52 -0400
Subject: Pyrex - C function declaration in extension type
In-Reply-To: <b7vgi2$e7g4$1@as201.hinet.hr>
References: <b7vgi2$e7g4$1@as201.hinet.hr>
Message-ID: <oqhe8sfvun.fsf@titan.progiciels-bpi.ca>

[Haris Bogdanovic]

> Pyrex doesn't let me define C function inside extension type (cdef class).
> Why is that and is there another way to acomplish the same thing (have C
> callable functions inside extension type) ?

C functions (Pyrex cdef's) are not directly callable from Python.  From
Python, one calls a Pyrex def, and from there, cdef's may be called.  So
from Python, only `def' within extension types may be used as methods.

Now, when I want a `cdef' associate with an extension type (call it
Shubbery, say!), instead of writing, _within_ Shubbery:

---------------------------------------------------------------------->
    cdef my_C_method(self, etc.
----------------------------------------------------------------------<

I write this _outside_ Shubbery:

---------------------------------------------------------------------->
cdef my_C_method(Shubbery self, etc.
----------------------------------------------------------------------<

That is, I use a typed `self' argument, and write the function body as if it
was a method from the extension type.  This is an acceptable compromise for
my needs, usually.  However, I have to call this function from within Pyrex
(it cannot be called directly from Python anyway) as a simple function,
passing it an explicit Shubbery as an actual argument for the formal `self',
avoiding the more comfortable `.' notation for methods.

On the other hand, I agree with you that it might be interesting being able
to declare `cdef's within extension types, for being later able to use the
`.'  notation for calling them, would it only be from within Pyrex.  But
this would be sugar, nothing fundamental nor essential so far that I see.

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From gh at ghaering.de  Wed Apr 23 16:58:20 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 23 Apr 2003 22:58:20 +0200
Subject: Using the MySQLdb module
In-Reply-To: <3EA6E75E.7070305@cis.ohio-state.edu>
References: <3EA6E75E.7070305@cis.ohio-state.edu>
Message-ID: <3EA6FE6C.2020602@ghaering.de>

Michael Mayhew wrote:
> Greetings,
> 
>     I am trying to use the MySQLdb module on a simple database local to 
> my personal Linux machine. However, I keep running into trouble [...]
> 
> sql = "insert into refFlatHum values ( %s, %s, %d, %d )" % ('atr', 
> 'chr1', 90810, 92190)
> 
> c.execute(sql)

Let the DB-API interface quote the values for you:

c.execute("insert into refFlatHum values ( %s, %s, %d, %d )",
	('atr', 'chr1', 90810, 92190))

Also see the DB-API 2 specification and the documentation and examples 
included in the MySQLdb source distribution.

-- Gerhard




From anton at vredegoor.doge.nl  Tue Apr  8 09:32:57 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Tue, 08 Apr 2003 15:32:57 +0200
Subject: Making python module using Borland C++ 5
References: <b6u7fp$d2c$1@news.cc.tut.fi>
Message-ID: <b6ujd4$pe5$1@news.hccnet.nl>

Rami Saarinen <rs55862 at NOSPAMyahoo.com> wrote:

>ImportError: dynamic module does not define init function (initpyclips)

Please have a look at:

http://www.cyberus.ca/~g_will/pyExtenDL.shtml

Anton.




From ckea25d02 at sneakemail.com  Tue Apr 15 23:37:27 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 15:37:27 +1200
Subject: Why functional Python matters
In-Reply-To: <mailman.1050453385.17091.python-list@python.org>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <mailman.1050453385.17091.python-list@python.org>
Message-ID: <b7iigt$1d3jt$1@ID-169208.news.dfncis.de>

Jp Calderone wrote:
>   I think you meant "partial lexical scoping" ;)  Dynamic scoping is what
> Python used to do.

No, it didn't!

There isn't any widely understood name for what Python used
to do, because no other language has ever been quirky
enough to do it that way. It was lexical, but with only
three levels visible at any one time.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From rmunn at pobox.com  Tue Apr 15 17:45:02 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 15 Apr 2003 21:45:02 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <m2fzokfgni.fsf@mycroft.actrix.gen.nz>
Message-ID: <slrnb9oajn.493.rmunn@localhost.localdomain>

Paul Foley <see at below.invalid> wrote:
> On Mon, 14 Apr 2003 16:50:40 GMT, Robin Munn wrote:
> 
>> What I really want to know is why you would consider sample #1 "broken
>> and fragile code". I can't think of any circumstances where sample #1
>> would fail to achieve its desired effect. Can you?
> 
> What if another thread (B) does the same thing, and changes sys.stdout
> while this thread (A) is still inside the "try" block?  And if thread
> A runs its "finally" clause and resets sys.stdout while thread B is
> still in its "try" block, they screw each other.

This is the classic problem that arises when two threads need to use any
shared resource, isn't it? I wasn't considering threads at all when I
wrote those two code samples, so of course I didn't use locks to ensure
good behavior.

Actually, now that I start thinking about it, any kind of
save-and-restore mechanism would start to get pretty hairy around
threads, wouldn't it? Let's look at a hypothetical "with" construct:

    with sys.stdout = cStringIO.StringIO():
        function_that_prints()
        parse_text_from(sys.stdout)

Let's assume we have two threads executing this code, and that the
"with" statement uses thread-local storage to save the original values.
If the "with" statement implements some kind of locking mechanism, this
is what you'll get:

----------------------------------------------------------------------
Thread A:                               Thread B:

(saves sys.stdout)
function_that_prints()
parse_text_from(sys.stdout)
(restores sys.stdout)
                                        (saves sys.stdout)
                                        function_that_prints()
                                        parse_text_from(sys.stdout)
                                        (restores sys.stdout)
----------------------------------------------------------------------

No problems here, of course. But what if this execution order happens?

----------------------------------------------------------------------
Thread A:                               Thread B:

(saves sys.stdout)
                                        (saves sys.stdout)
function_that_prints()
                                        function_that_prints()
parse_text_from(sys.stdout)
(restores sys.stdout)
                                        parse_text_from(sys.stdout)
                                        (restores sys.stdout)
----------------------------------------------------------------------

Thread B's parse_text_from() will be trying to read from thread A's
original sys.stdout, which will probably fail.

I don't really see any good solutions other than a lock around the whole
with: statement, which is something that would be easy to do in a try
... finally block. But maybe you can come up with something.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From blunck at gst.com  Wed Apr 16 12:04:33 2003
From: blunck at gst.com (Christopher Blunck)
Date: Wed, 16 Apr 2003 12:04:33 -0400
Subject: How many of you are Extreme Programmers?
In-Reply-To: <1050508138.3e9d7b6a1d520@mcherm.com>
References: <1050508138.3e9d7b6a1d520@mcherm.com>
Message-ID: <20030416160433.GC21292@homer.gst.com>

On Wed, Apr 16, 2003 at 08:48:58AM -0700, Michael Chermside wrote:
> Christopher writes:
> > So how many of you guys use XP processes?
> Well, I'm not sure if I qualify or not. I really LIKE most of the points
> that XP makes, but I find many of its proponants overly exuberant. I
> think exhaustive, automated tests are a wonderful idea. I believe in
> refactoring. I think pair programming works very well most of the time.
> I think super-short release cycles (with code that grows feature-by-
> feature) is a great idea. I think having a customer on-site is extremely
> helpful.

It's quite funny, but if you sit an XP and RUP guy down at the same table
you'll find that they agree with most principals.  They'll disagree on some
things (like design), but I've never seen them disagree on iterative 
development, frequent testing, frequent integration, etc.  They may disagree
on the frequency of "frequent testing" (XP favoring daily testing), or on
the implementation details of iterative development.  But for the most part,
I've found would-be opposites to share very similar views.


> But none of that really matters, because I don't get to set the policy
> where I work. We don't do unit tests (except for the ones that _I_ write)
> because "they would take too long". We deliver functionality in a single
> big release. We don't do pair programming except for very brief periods.
> Perhaps someday I can change those things, but for now, I don't get to
> practice much of this, except in my at-home-on-my-own-time projects
> (basically non-existant since my son was born 4 months ago).

join 'da club


-c

-- 
 11:55am  up 21:07,  1 user,  load average: 0.14, 0.11, 0.15




From alloydflanagan at attbi.com  Fri Apr 18 17:28:53 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 18 Apr 2003 14:28:53 -0700
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net> <s1Vna.126364$It5.107742@news2.central.cox.net> <mailman.1050692370.715.python-list@python.org>
Message-ID: <a8b7f07a.0304181328.745890d5@posting.google.com>

Andy Jewell <andy at wild-flower.co.uk> wrote in message news:<mailman.1050692370.715.python-list at python.org>...

> In all my days playing with rope, I've never managed to shoot myself with
>  it.  
> Hang myself, maybe, but not shoot myself.  :-)
> 

There's an excellent book by Allen Holub about ways to avoid trouble
in C and C++ programming.  It's called "Enough Rope To Shoot Yourself
In The Foot".  If you program in C or C++ and don't have it, get it.



From bkc at Murkworks.com  Tue Apr 15 09:18:20 2003
From: bkc at Murkworks.com (Brad Clements)
Date: Tue, 15 Apr 2003 09:18:20 -0400
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1050389907.24670.python-list@python.org>
Message-ID: <3e9c069f_1@spamkiller.newsgroups.com>

_
"David LeBlanc" <whisper at oz.net> wrote in message
news:mailman.1050389907.24670.python-list at python.org...
> > -----Original Message-----

> > >    aahz> So how does Capon differ from SCons?
> > >
> > >Well, the name has funnier connotations, for one.   :)
> >
> > You deserve to be roasted for that comment.

> A fate of many capons ;)
>
> First they have "the" operation, then they grow up and get put on a spit
and
> thence over a slow fire. Life as a chicken is hell ;)
>

My neighbhor, a 70+ year old former federal meat inspector (before that he
was a Vet for the Italian Cavalry ) raises and does his own capons.

He showed me how to do "the operation".  Quite amazing to see him slice open
a rooster's behind and reach right in. This one was a bit large, so he had
to "go in from behind" instead of under the wing as he usually does. Then he
just sews them up with dental floss. No antiseptic! he says chickens don't
get staph infections. He says when he learned how to do "the operation" he
killed 5 of them right off. Apparently there's a nack to not cutting some
important arteries.


He says at a store it's hard to find real capons. Back in the 70s they used
estrogen to make capons, but that turned out to give people cancer.  The FDA
banned that practice, however many "unscrupulous" poultry plants continue to
use estrogen, but they also cut the rooster to give it "the scar".. So meat
inspectors look, think the operation has been done, but it really hasn't.

Enjoy!

(btw, home grown capons are yummy!)






From exarkun at intarweb.us  Wed Apr 23 10:35:43 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 23 Apr 2003 10:35:43 -0400
Subject: Testing if string contains a substring
In-Reply-To: <70846faf-b437-47ee-8a2c-7f685b00425b@news.szn.dk>
References: <b865dq$28gi$1@ns.felk.cvut.cz> <70846faf-b437-47ee-8a2c-7f685b00425b@news.szn.dk>
Message-ID: <20030423143543.GA10150@meson.dyndns.org>

On Wed, Apr 23, 2003 at 02:23:24PM +0000, Klaus Alexander Seistrup wrote:
> 
> [snip]
> 
> >>> contains = lambda haystack, needle: haystack.find(needle) > 0

  The above should be >= 0, otherwise the first example below fails.

> >>> contains('foo bar fly', 'foo')
> 1

  Jp

-- 
"One World, one Web, one Program." - Microsoft(R) promotional ad
"Ein Volk, ein Reich, ein Fuhrer." - Adolf Hitler
-- 
 up 34 days, 11:03, 4 users, load average: 1.96, 1.78, 1.70




From mykbourassa at cogeco.ca  Wed Apr  9 09:57:56 2003
From: mykbourassa at cogeco.ca (myk)
Date: Wed, 09 Apr 2003 09:57:56 -0400
Subject: Newbie: Python + C Question
In-Reply-To: <3E93586E.D4AB1531@engcorp.com>
References: <3e92ebe8$1@rmcnt03a.rmc.ca> <3E933C8D.5020700@removethis.free.fr> <3E93586E.D4AB1531@engcorp.com>
Message-ID: <OHVka.946$Jl.386744@read2.cgocable.net>

Thanks for the comments.  I'm not a programmer and am trying to switch 
from Matlab (8 yrs exlusive use) to a "real" programming language for 
flexibility and cost reasons.  I was reading through "Learning Python" 
but the answer to the question of where the compiler comes from (esp. in 
Windows) I eventually found, partly, here:

http://www.python.org/dev/doc/devel/inst/tweak-flags.html

Cheers.

Peter Hansen wrote:

>laotseu wrote:
>  
>
>>bourassa wrote:
>>    
>>
>>>When Python uses the C extension does it just process
>>>it in an interpreted fashion
>>>      
>>>
>>Python is not a C interpreter.
>>
>>    
>>
>>>or does the code written in C run like a
>>>compiled C program?
>>>      
>>>
>>newbie to programming, uh ?-)
>>
>>How do you think this all works ?
>>    
>>
>
>Actually, to be fair to the OP, this is a valid and even interesting
>question.  Where, after all, does it say that the C code that is
>in the extension is actually pre-compiled into an executable and
>is run as machine code when called by Python?  (Maybe it's there
>in the docs, but maybe it's not or maybe it's unclear.  Have you
>checked?)
>
>I could picture someone believing that Python had the ability to
>execute C code interpretively via some fancy C-interpreter
>extension, which would allow the existing code-base of C to be 
>used within an otherwise-pure-Python program.  Definitely not
>very likely to happen as a general feature, but still an interesting
>concept.
>
>-Peter
>  
>




From nramchandani at harveynash.com  Thu Apr 24 07:19:30 2003
From: nramchandani at harveynash.com (Neeta Ramchandani)
Date: Thu, 24 Apr 2003 12:19:30 +0100
Subject: OO + Python Developers Wanted
Message-ID: <sea7d661.069@lon_nw_9.harveynash.com>

I am looking for an OO Scriptor with Python Scripting experience.
You must have at least 2 years Java experience with Unix and Proper Pyhton development expertise.
This is a role for an Investment Bank in the City, for duration of 3-6 months.
Rates are OPEN!!

Neeta Ramchandani
Key Account Manager
Harvey Nash IT
Investment Banking / Finance Team
DD:  020 73331518
Fax:  020 73332657
E-mail: nramchandani at harveynash.com
Website:  www.harveynash.com



*****************************************************************
IMPORTANT NOTICE

The information in this e-mail and any attached files is CONFIDENTIAL and may be legally privileged or prohibited from disclosure and unauthorised use.  The views of the author may not necessarily reflect those of the Company.

It is intended solely for the addressee, or the employee or agent responsible for delivering such materials to the addressee.  If you have received this message in error please return it to the sender then delete the email and destroy any copies of it.  If you are not the intended recipient, any form of reproduction, dissemination, copying, disclosure, modification, distribution and/or publication or any action taken or omitted to be taken in reliance upon this message or its attachments is prohibited and may be unlawful.

At present the integrity of e-mail across the Internet cannot be guaranteed and messages sent via this medium are potentially at risk.  All liability is excluded to the extent permitted by law for any claims arising as a result of the use of this medium to transmit information by or to the Harvey Nash Group plc.
*****************************************************************





From michael at stroeder.com  Tue Apr  8 07:01:07 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Tue, 08 Apr 2003 13:01:07 +0200
Subject: Announce: apm.py 0.1.0
In-Reply-To: <mailman.1049750436.1395.python-list@python.org>
References: <mailman.1049750436.1395.python-list@python.org>
Message-ID: <komam-r2d.ln1@nb2.stroeder.com>

Tilo Riemer wrote:
> 
> Because I don't have dozens of laptops and operating systems I need help! It
> would be great to have a Python modul which covers APM more or less
> completely for many platforms one day.

What about ACPI? That became standard on some Linux distros (e.g. SuSE 8.1+).

Maybe it would be better to call the module powerman or similar to subsume 
APM, ACPI and any other standard popping up.

Ciao, Michael.




From and-google at doxdesk.com  Thu Apr 24 09:17:30 2003
From: and-google at doxdesk.com (Andrew Clover)
Date: 24 Apr 2003 06:17:30 -0700
Subject: True/False
References: <mailman.1051110756.23235.python-list@python.org>
Message-ID: <2c60a528.0304240517.25f2fe29@posting.google.com>

Anthony Baxter <anthony at interlink.com.au> wrote:

> Wasn't there also some talk of making None, True and False keywords (for
> possible performance gains)?

In this case I guess we're looking at something like:

  try:
    True
  except NameError:
    exec 'False, True= 0, 1'
    def bool(c):
      if c:
        return True
      return False

Any objections?

-- 
Andrew Clover
mailto:and at doxdesk.com
http://www.doxdesk.com/



From dave at pythonapocrypha.com  Fri Apr 18 11:56:19 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Fri, 18 Apr 2003 08:56:19 -0700 (PDT)
Subject: Mem "leak" w/ long-running network apps?
In-Reply-To: <LNBBLJKPBEHFEDALKOLCIECLEHAB.tim_one@email.msn.com>
References: <LNBBLJKPBEHFEDALKOLCIECLEHAB.tim_one@email.msn.com>
Message-ID: <Pine.LNX.4.51.0304180838560.23375@ns0.tbye.com>

On Fri, 18 Apr 2003, Tim Peters wrote:

> > I'd notice that after my tests ran for a long time I'd stop them and
> > memory usage would drop down after awhile by a few megabytes, and upon
> > starting my tests again (without restarting my app) mem usage would drop
> > down some more but not all the way down and then it would gradually grow
> > again to a new high-water mark, so that overnight my process had hundreds
> > of MB of RAM.  With my recv-replacement I'm holding steady at about
> > 30 MB, which is normal.
>
> If you want to know more, scour your platform docs for whatever debug info
> your local malloc can deliver.

Okie dokie, I'll do that (I'm running on Linux).

> > Here's the function. I can get away with using a single large buffer for
> > all my receives because while the server is a mixture of threading and
> > poll-based, all the I/O happens sequentially in one thread against sockets
> > that are known to be ready for I/O.
>
> Heh.  That sounds regrettable, eventually <wink>.

It might be, eventually. :) For now though it's much faster and scales
much better than asyncore and the like, but I'm always looking to make it
better...

Thank you *so* much for taking the time to respond and for helping me
clean up my replacement function - I wasn't sure I really understood the
problem (or even why the fix worked), but it's encouraging to know I was
at least heading in the right general direction.

-Dave




From graham__fawcett at hotmail.com  Sun Apr  6 13:44:54 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 6 Apr 2003 10:44:54 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net>
Message-ID: <e9570f37.0304060944.79564f88@posting.google.com>

Neil Schemenauer <nas-usenet at arctrix.com> wrote in message news:<%5Oja.10776$7w2.9784 at nwrddc01.gnilink.net>...
> Ian Bicking <ianb at colorstudy.com> wrote:
> > The next step I can envision would be for Twisted to become the
> > canonical infrastructure, where higher-level web frameworks are all
> > built ontop of it.
> 
> That's not feasible.  Some applications would require extensive
> modification to work on top of an event driven framework.
> 
>   Neil

Neil, I may just be suffering from
daylight-savings-adjustment-syndrome, but what kinds of applications?
We are talking only about Web application frameworks, which are
request handlers and are therefore event-driven by definition.

Unless, as JP sugested, you're suggesting bad design, like an
application that is too tightly bound to a specific networking
framework?

I'm not a Twisted fanatic, just trying to understand.

-- Graham



From spahievi at vega.bg  Fri Apr  4 14:20:04 2003
From: spahievi at vega.bg (Niki Spahiev)
Date: Fri, 4 Apr 2003 22:20:04 +0300
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304031917.1d0100bf@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com>
 <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>
 <9383481e.0304030705.524cf0fa@posting.google.com>
 <mailman.1049388384.29614.python-list@python.org>
 <9383481e.0304031917.1d0100bf@posting.google.com>
Message-ID: <1588057776.20030404222004@vega.bg>

4/4/2003, 6:17:19, Drew Moore wrote:

DM> # modify the voltages, in a very simple, readable syntax.
DM> vdduc = -3.8 # try a reduced supply
DM> vreset = -3.2 # and increased reset drive...
DM> vbias = 2.3 # and decreas the bias a little..

DM> # do some more stuff with new voltages, etc etc...

vdduc [:]= -3.8 is valid Python and can be made to work now. Maybe new
augmented assignment a := b to mean a[:] = b.

Niki Spahiev





From mertz at gnosis.cx  Tue Apr  8 16:18:00 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 16:18:00 -0400
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
 <+6xk+kKkXQeJ092yn@gnosis.cx> <mailman.1049829736.3719.python-list@python.org>
Message-ID: <45yk+kKkXAlY092yn@gnosis.cx>         

"Mike C. Fletcher" <mcfletch at rogers.com> wrote previously:
|>    '"%s"' % repr(aString)[1:-1]
|Except when there's a " character in the string ;) .

Ahh... what I MEANT to write was, of course:

    '"%s"' % repr(aString)[1:-1].replace('"',r'\"')

Emphasizing the "not pretty" part further.  (I still think it is
prettier than Mike's though :-)).

Btw. The OP wrote "string", so obviously I am not interested in
repr(u"x").

Yours, Lulu...

--
    _/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: Postmodern Enterprises _/_/_/
   _/_/    ~~~~~~~~~~~~~~~~~~~~[mertz at gnosis.cx]~~~~~~~~~~~~~~~~~~~~~  _/_/
  _/_/  The opinions expressed here must be those of my employer...   _/_/
 _/_/_/_/_/_/_/_/_/_/ Surely you don't think that *I* believe them!  _/_/





From skip at pobox.com  Mon Apr 14 14:15:35 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 13:15:35 -0500
Subject: parser stack size limit
In-Reply-To: <20030414172211.GA20921@latt.if.usp.br>
References: <20030414172211.GA20921@latt.if.usp.br>
Message-ID: <16026.64199.203154.450236@montanaro.dyndns.org>

    Francisco> """
    Francisco> s_push: parser stack overflow
    Francisco> MemoryError
    Francisco> """

    ...

    Francisco> Any ideas on how to avoid this? Could it be as simple as a
    Francisco> file to modify and then recompile python?

Yes, I believe you can boost the value of MAXSTACK in Parser/parser.h.  This
question came up recently.  I believe the long-term solution is to
dynamically allocate the parser's stack.

Skip




From glenfant at NOSPAM.bigfoot.com  Mon Apr  7 06:23:57 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Mon, 7 Apr 2003 12:23:57 +0200
Subject: Customize rlcompleter
Message-ID: <b6risd$giq$1@biggoron.nerim.net>

Hi,

I'd like to use "rlcompleter" with a custom vocabulary for a personal
command line interpreter based on the "cmd" package. But the doc for
rlcompleter is quite sparse, and I can't see how to tailor it to my own
usage.

Thanks in advance for any hint, pointer, RTFM (...).

--Gilles




From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr  4 12:07:48 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 4 Apr 2003 17:07:48 GMT
Subject: Simulating Qubits?
References: <yyyu1deuvd2.fsf@adams.patriot.net>
Message-ID: <slrnb8rev4.rs.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <yyyu1deuvd2.fsf at adams.patriot.net>, Nick Vargish wrote:
<snip> 
> I'm thinking that it might be interesting to write a module that
> simulates the computational abilities of qubits. The Economist article
> specifically mentions object-oriented approaches as key to integrating
> qubit logic with traditional computing programs.
> 
> Okay, at this point I'm just musing aloud. Has anyone else been
> thinking about this, or perhaps done some actual work on the problem?
<snip>

Perl 6 may well have superpositions:
<http://www.perl.com/pub/a/2002/11/p6pdigest/20021121.html#superpositions_and_laziness>



From djc at object-craft.com.au  Tue Apr 29 20:12:24 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 30 Apr 2003 10:12:24 +1000
Subject: Web programming and a different "type" problem
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com> <slrnbarhfb.67t.jon+usenet@snowy.squish.net> <b8lj2q$l14$1@newshost.accu.uu.nl> <m3fzo1v561.fsf@ferret.object-craft.com.au> <mailman.1051654786.4587.python-list@python.org>
Message-ID: <m365owlu6v.fsf@ferret.object-craft.com.au>

> On Tue, 2003-04-29 at 07:50, Dave Cole wrote:
> > Albatross solves this yet another way.  As input tags are processed by
> > the template interpreter the form tag accumulates a summary of all the
> > input fields.  When the form is closed the summary is then pickled,
> > MD5 signed with a secret, compressed, and base64 encoded in a hidden
> > field.
> 
> You mean, it takes the expected names?  Like a form:
> 
> Object 1: <input type=checkbox name="id*" value=1>
> ...
> Object 4: <input type=checkbox name="id*" value=4>
> 
> You create a hash of (in this small form) just "id*", and so confirm
> the user isn't sending extra fields, or changing field names?

Sort of.

> Seems a little odd... or are you doing something else?  There's
> always the issue of the security of hidden fields, but that's kind
> of a different issue.

With Albatross templating the form would look like this:

<al-form>
 <al-input type="checkbox" name="id" value="1" list>
 <al-input type="checkbox" name="id" value="4" list>
</al-form>

Just to restate the problem:

When the browser sends a response to the above form (assuming you
have a submit input somewhere) it can send 0, 1, or 2 values for the
"id" field.  With the base cgi.py handling you have to do something
like this:

   fields = cgi.FieldStorage()
   if fields.has_key('id'):
       if type(fields['id']) is type([]):
           for field in fields['id']:
               process(field.value)
       else:
           process(fields['id'].value)

What you really want to be able to do is something like this
regardless of how many values were submitted by the browser:

   for value in request.fields['id']:
       process(value)

So what Albatross does is allow (require) you to specify when a field
should return a list.  If you build a form which contains multiple
inputs with the same name ("radio" is an obvious exception) then it
will raise an exception if you do not use the "list" attribute on all
of the instances of the field in the form.

In our experience you can get subtle and annoying bugs in your
application as a result of incorrect handling of fields which
sometimes return multiple values.

How Albatross handles the problem:

When the template interpreter sees the <al-form> it starts recording
the types of all <al-input> tags.  Each <al-input> is checked for
multiple instances and consistent declaration of "list" attribute.
Any error raises and exception.  This allows programming errors to be
caught during development, not when the application goes live.

When the </al-form> tag is processed, the record of fields is encoded
into a hidden field called __albform__.  The field value is MD5 signed
by combining the encoded value with a server side secret and the sign
is combined with the __albform__ value.

When the browser posts a response to the form the __albform__ value is
checked for tampering.  If the check fails the __albform__ value is
discarded.  If the __albform__ value is OK it is used to process the
rest of the fields in the browser request.

If the request was OK fields which were declared with the "list"
attribute will always yield a list of values.  In the above example
the 'id' field will be set to a list of 0, 1, or 2 values.  Fields
which were not declared as "list" will be set to None if they are not
present in the browser request.

Currently one downside to this is that when the __albform__ value is
not present, the request merging falls back to a mode where it merges
all values present in the request in the form that they are present.
This means that it is more a convenience feature than a security
feature :-)

- Dave

-- 
http://www.object-craft.com.au



From ssthapa at classes.cs.uchicago.edu  Sat Apr 12 16:33:16 2003
From: ssthapa at classes.cs.uchicago.edu (Suchandra Thapa)
Date: Sat, 12 Apr 2003 20:33:16 GMT
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <Z7nla.45$W4.8475@news.uchicago.edu> <7oZla.5396$LB6.150919@news1.tin.it>
Message-ID: <gK_la.81$W4.16625@news.uchicago.edu>

Alex Martelli <aleax at aleax.it> wrote:
>> Have you considered using R (www.r-project.com) and RS/Python
> 
> Hmmmm -- this site's most succint homepage's in English, but all
> links from it seem to be trying to show me lots and lots of Japanese
> text.  What's the right way to get R docs & other downloads these
> days for the non-Japanese-speaker...?

My mistake, it's acutally http://www.r-project.org/ the PCA component is in the
package pcurve.   Once R is installed install.package('pcurve') should get the
pcurve package, compile and install it on your system.

-- 
----------------------------------------------------------------------------
Suchandra Thapa                       
s-thapa-11 at NOSPAMalumni.uchicago.edu  
----------------------------------------------------------------------------



From tim.one at comcast.net  Mon Apr 28 10:00:42 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 28 Apr 2003 10:00:42 -0400
Subject: Python-2.3b1.exe install error on W98se
In-Reply-To: <b8ja52$k77$1@news.hccnet.nl>
Message-ID: <BIEJKCLHCIOIHAGOKOLHMEKEFHAA.tim.one@comcast.net>

[Anton Vredegoor]
> ... lets not forget that this really happening on three W98se machines
> here. No one else has confirmed this result yet, so I can't be sure
> whether it's a local issue or not.

You can:  immediately after Guido announces a release, I download it from
the site and try it.  Works fine.  Try using a different method for
downloading (if using http, try ftp, or vice versa; if using Netscape, try
IE, or vice versa).  Since I wasn't looking over your shoulder, I'm afraid I
can't point at anything insane you may have done <wink>.





From martin at v.loewis.de  Fri Apr 18 12:14:06 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 18 Apr 2003 18:14:06 +0200
Subject: How can I read event logs with python?
In-Reply-To: <b7p48d$15pt$1@galaxy.us.dell.com>
References: <b7p48d$15pt$1@galaxy.us.dell.com>
Message-ID: <b7p88b$jm3$07$1@news.t-online.com>

junk at aol.com wrote:

> I just want python to look through my event logs and print warnings and
> errors.  I know that win32evtlog should allow me to but I'm just not
> able to get it to work.

So what is the code that you have tried to use win32evtlog, what was 
your expected outcome of that code, and what happened instead? Don't 
make us guess what your problem is.

Regards,
Martin




From blachner at gmx.de  Tue Apr 22 14:44:03 2003
From: blachner at gmx.de (Bernd Lachner)
Date: Tue, 22 Apr 2003 20:44:03 +0200
Subject: Problem with PyQt on SuSE 8.2
Message-ID: <b842hl$dsa$1@online.de>

Hello,

we try ro run several Python programs which are using PyQt on SuSE 8.2.
Every time the same error message occurs:

Traceback (most recent call last):
  File "./kdevpy.py", line 16, in ?
    from qt import *
  File "/usr/lib/python2.2/site-packages/qt.py", line 39, in ?
    import libqtc
ImportError: /usr/lib/python2.2/site-packages/libqtcmodule.so: undefined
symbol: _ZNK16QAssistantClient9classNameEv

We tried the original PyQt packages which comes with SuSE 8.2. We also have
compiled PyQt from the source RPMs, but the same error occurs.

Is ther e anybody who can help us? Thanks in advance.

Best regards

Bernd & Birgit Lachner



From nessus at mit.edu  Mon Apr 21 17:44:56 2003
From: nessus at mit.edu (Douglas Alan)
Date: Mon, 21 Apr 2003 17:44:56 -0400
Subject: A new and very robust method for doing file locking over NFS?
References: <lcu1cw525a.fsf@gaffa.mit.edu> <5j65pbai03.fsf@xoxy.net>
Message-ID: <lcvfx7o78n.fsf@gaffa.mit.edu>

Charlie Reiman <rando-15696825.5.reiman at xoxy.net> writes:

> I like this idea a lot but it has one drawback. In the traditional
> lockfile approach, you can have multiple readers even if someone is
> writing. Now that you rename the file, eveytime someone is writing the
> filename is a moving target. Depending on how you intend to use the
> file, this might make your solution a no-go from the get-go.

Yes, that is an issue to consider.  In the application I am
considering this for, it is adequate for the locks to be completely
exclusive.

If one wanted to use my approach for non-exclusive locks, I think that
other processes could glob for "filename.locked.*.*", or the locking
process could leave a symbolic link behind.  If a process wants to
read a file and it isn't there, it would wait in a sleep loop for
either "filename" or "filename.locked.*.*" to appear (or for the
symbolic link to appear, depending on which method we chose).  If
neither appears after some amount of time, then the process would
conclude that the file has gone missing and generate an error.

It is possible, although unlikely, that if many processes were locking
and unlocking the file, that one of the processes might see neither
"filename" nor "filename.locked.*.*" every time it looks for them,
even after many attempts, and even though the file is really there.
In that case we would get a spurious "file missing error", but this
would not result in any corrupted data, and if we make the locking
processes persistent and patient enough, this will probably never
occur in our lifetimes.

> FWIW, I've never had any problems with NFS for traditional lock
> files. NFS works pretty hard to get messages where they need to
> be. 

This is somewhat encouraging to hear.  Does your positive experience
extend to Linux NFS servers and clients?

> If I were that deeply concerned or running across an unreliable
> network, I'd probably write my own lock daemon and run it over TCP.

That's certainly another reasonable approach.  I'd prefer not to have
a centralized daemon, however, unless absolutely necessary, because it
makes the software more difficult to install and administer.

|>oug



From jepler at unpythonic.net  Fri Apr 25 16:26:41 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 25 Apr 2003 15:26:41 -0500
Subject: Hierarchy of types ?
In-Reply-To: <37cbc1de.0304251216.6068146f@posting.google.com>
References: <37cbc1de.0304251216.6068146f@posting.google.com>
Message-ID: <20030425202640.GI30690@unpythonic.net>

Comparison of objects of the same type depends on the type:
[...]
* Most other types compare unequal unless they are the same object; the
  choice whether one object is considered smaller or larger than another one
  is made arbitrarily but consistently within one execution of a program.

  -- http://python.org/doc/ref/comparisons.html

Jeff




From martin at v.loewis.de  Wed Apr  2 16:41:04 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 02 Apr 2003 23:41:04 +0200
Subject: Compiled Python on Solaris8  at 2.2.2 - Problem with (_)socket
References: <30875970.0304020715.73954878@posting.google.com>
Message-ID: <m3vfxwk26n.fsf@mira.informatik.hu-berlin.de>

ipellew at pipemedia.co.uk (Ian Pellew) writes:

> Can someone please advise what to do.

I recommend you recompile the socket module without SSL support,
by editing Modules/Setup.

Regards,
Martin



From tjreedy at udel.edu  Tue Apr 22 13:02:09 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 22 Apr 2003 13:02:09 -0400
Subject: Overloading and?  was <RE: Should I prefer an external database>
References: <mailman.1051023275.29338.python-list@python.org>
Message-ID: <wU-dnQVJgtHr6zijXTWcow@comcast.com>

> Does anyone know the reason for not allowing an overload of the and
operator?

I presume because 'and' and 'or' are already defined for all Python
objects (that have not disabled logic by raising an exception when
asked for truth value).

TJR





From rmunn at pobox.com  Fri Apr  4 10:19:07 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 04 Apr 2003 15:19:07 GMT
Subject: computer science with python
References: <m3istvtodl.fsf@moroni.pp.asu.edu>
Message-ID: <slrnb8qjjf.i81.rmunn@localhost.localdomain>

David Bear <david.bear at asu.edu> wrote:
> 
> I purchased the OReilly book "Computer Science with Perl" (or
> something similar and found it very fun -- even though I really hate
> Perl.  It made me think that python would be a terrific language to
> demonstrate some key computer science concepts.
> 
> Anyone know if a book has already been done?  I find that many CS
> language concepts are debated/discussed here and have to be careful
> when I look at this group because I become too interested in reading
> all the conceptual stuff rather than getting a specific question
> answered.  Seems this list would provide a good starting point for
> such a book.  I image it could be many volumes...

I've never read _Computer Science with Perl_, so I don't know how
similar or dissimilar this book might be, but _How to Think Like a
Computer Scientist: Learning with Python_ has to be pretty close to what
you're looking for, at least. You can find it at:

    http://www.ibiblio.org/obp/thinkCSpy/

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From skip at pobox.com  Tue Apr 29 10:59:25 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 29 Apr 2003 09:59:25 -0500
Subject: Where did you learn to unit test?  How did you learn?
In-Reply-To: <1147e466.0304290638.287687f6@posting.google.com>
References: <1147e466.0304290638.287687f6@posting.google.com>
Message-ID: <16046.37709.757941.43279@montanaro.dyndns.org>

    -c> I've noticed in my development that Python programmers (in general)
    -c> write more unit tests than their java counterparts.  If I download a
    -c> project off of jakarta.apache.org, it very rarely has unit tests. 
    -c> However, when I download a python module, it frequently has unit
    -c> tests.

That's (I think) because:

    * In general writing Python is easier than writing Java.  That holds
      true for test code as well as application code.  Note that Perl
      packages tend to also have plenty of test cases.

    * The PythonLabs folks set an excellent example.  Take a look at the
      Lib/test directory in your Python distribution.  The tests there are a
      mixture of the newer unittest-based stuff and a couple older schemes.

    -c> What advice can you offer on spreading the testing bug around?  

Bugs.  When you discover a bug, write a test case which tickles the bug.
Then fix the bug and use the test case to demonstrate that the bug is gone.
It's simple and objective.

    -c> What success stories do you have?  

Python itself works as a pretty good success story.  The presence of a large
suite of test cases increases confidence that new code doesn't break things,
and when it does, that fact is generally discovered very quickly.

    -c> Any insight you can offer would be appreciated.

Lead by example.  Create a test subdirectory in your source code repository
and start writing test cases.  Use those test cases to demonstrate that your
code functions as required.  

Skip




From DennisR at dair.com  Tue Apr 22 17:18:25 2003
From: DennisR at dair.com (Dennis Reinhardt)
Date: Tue, 22 Apr 2003 21:18:25 GMT
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com> <mailman.1050978337.8048.python-list@python.org> <87wuhm2rsu.fsf@pobox.com>
Message-ID: <Bkipa.1154$Hm6.104194064@newssvr14.news.prodigy.com>

JpC> >   Is the app that generates these available (commercially or for
free)?
JJL> Probably py2exe or McMillan Installer, I imagine.

Nope.... neither one.  I used Python itself and other programs, as described
in separate posting.

--

Dennis Reinhardt

DennisR at dair.com   http://www.spamai.com?ng_clp





From matt at teaphoo.fsnet.co.uk  Fri Apr  4 17:42:08 2003
From: matt at teaphoo.fsnet.co.uk (Matt Russell)
Date: 4 Apr 2003 14:42:08 -0800
Subject: Socket timeouts for HTTP connections (httplib)?
References: <9df3d134.0304020550.7e32bf6b@posting.google.com>
Message-ID: <459f9b68.0304041442.58cf392e@posting.google.com>

I have recently posted a patch to sf regarding ftplib, which has the
same problem (no way to use ne wtimeout functionailty).


I have recently found that for classes ftplib.FTP, poplib.POP3,
imaplib.IMAP, telnetlib.Telnet, smtplib.SMTP and
httplib.HTTPConnection all have very similar connect methods. Assuming
my patch will be taken on board, it will allow access to the socket
timeout from each of these classes via optional parameters to connect
and settimeout/gettimeout on each class in question.

Would being able to  do:

conn  = httllib.HTTP(url, timeout=n)
conn.settimeout(n)
...
conn.send("<data>")

solve your problem?

do u need this functionailty now?


Matt Russell


nick_bower at yahoo.com.au (nick) wrote in message 
news:<9df3d134.0304020550.7e32bf6b at posting.google.com>...
> I'm using httplib in a threaded HTTP load imparting test suite for our
> portal.  A problem is that HTTP GETs often hang, blocking my
> requesting threads.  I was highly disapointed to find, after writing
> my test suite, that no high-level HTTP timeout functionality exists in
> Python.
> 
> I have neither an expert understanding of socket programming,
> threading or asyncronous IO, yet can get by using threading.Thread and
> httplib.  I am therefore looking for a timeout at a fairly high level
> (ie the httplib api).
> 
> There were rumblings about Guido implementing this at a socket level
> in June 2002, but I've not been able to find anything since. 
> Presumably this has not yet risen up into httplib either because the
> python.org documentation doesn't mention it.
> 
> Has/is someone involved with this?  timeoutsocket simply doesn't work
> for me and throws continuous "connection reset by peer" exceptions. 
> So I'm completely stuck for options.
> 
> Thanks, Nick



From puccio_13 at yahoo.it  Wed Apr 16 10:55:19 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Wed, 16 Apr 2003 16:55:19 +0200
Subject: reading Unix mailbox
References: <hZbna.17434$DN.480144@tornado.fastwebnet.it> <Z5dna.110590$It5.98167@news2.central.cox.net> <xdena.17614$DN.485153@tornado.fastwebnet.it>
Message-ID: <Tiena.17622$DN.485322@tornado.fastwebnet.it>

> 
> import email.Message
> import mailbox
> 
> mbox_file = file('../spam1.txt')
> mbox_obj = mailbox.PortableUnixMailbox(mbox_file, email.Message)
> msg = mbox_obj.next()           # <-----what's wrong with this??
> 
> But I got:
> 
> Traceback (most recent call last):
>   File "testMailBox.py", line 6, in ?
>     msg = mbox_obj.next()
>   File "/usr/lib/python2.3/mailbox.py", line 35, in next
>     return self.factory(_Subfile(self.fp, start, stop))
> TypeError: 'module' object is not callable
> 
> I am not a Python guru as you see...
> thanks

I solved putting as factory: email.message_from_file
Thanks again, it seems to be the solution I needed

-- 
bye
Alessio Pace



From none at none.com  Wed Apr  2 16:32:15 2003
From: none at none.com (Iwan van der Kleyn)
Date: Wed, 02 Apr 2003 21:32:15 +0000
Subject: DVD on Mdk 9.1: VideoLan works, but what about Xine?
Message-ID: <3e8b3dbc$0$2228$e4fe514c@dreader6.news.xs4all.nl>

I've installed the VideoLan client to play DVD (including encrypted) on 
Mandrake 9.1. I've been using it for the last year and a half on several 
distrubutions and it works reliably. Check out 
http://www.videolan.org/vlc/download-mandrake.html

However, from several earlier posts I understand that it is possible to 
use Xine to play encrypted DVD's. Not out of the box of course, but 
through one or more plugins. However, I've not been able to get it 
working. Can anyone point to a HOW-TO or other resource which describes 
how to enable Xine for DVD playback?

Regards,

Iwan




From haris.bogdanovic at zg.hinet.hr  Sun Apr 20 21:12:32 2003
From: haris.bogdanovic at zg.hinet.hr (Haris Bogdanovic)
Date: Mon, 21 Apr 2003 03:12:32 +0200
Subject: Pyrex - C function declaration in extension type
Message-ID: <b7vgi2$e7g4$1@as201.hinet.hr>

Pyrex doesn't let me define C function inside extension type (cdef class).
Why is that and is there another way to acomplish the same thing (have C
callable functions inside extension type) ?

Thanks
Haris





From carroll at tjc.com  Sat Apr  5 14:09:48 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Sat, 05 Apr 2003 11:09:48 -0800
Subject: nntplib: how to access NNTP response codes on exceptions?
References: <cirn8vsv4bhtp82n1tfil3e2kkounpgvav@4ax.com> <m3brzn27z3.fsf@telia.com> <hn2s8vs4p97bl4jcskrn34mgi7eartj4h6@4ax.com> <m3he9crfum.fsf@telia.com>
Message-ID: <t9au8vst2p8ks59v5ubuf5aeokvl92nkpm@4ax.com>

On Sat, 05 Apr 2003 17:53:03 GMT, Andrew Markebo
<andrew.markebo at telia.com> wrote:

>Oh.. you read the docs?? Eh? ;-) I do the code ;-) 

Heh, yeah, I'm still new enough to Python that I use the docs.
Although I did find this one by looking at the code.

(Actually two days ago, I didn't know how to use nntplib, and now I'm
already modifying it and planning to submit my mod as a patch.)




From exarkun at intarweb.us  Fri Apr 11 12:35:31 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 11 Apr 2003 12:35:31 -0400
Subject: Dynamically loading python classes
In-Reply-To: <9f363ad9.0304110813.3778cc8a@posting.google.com>
References: <9f363ad9.0304110813.3778cc8a@posting.google.com>
Message-ID: <20030411163531.GB31416@meson.dyndns.org>

On Fri, Apr 11, 2003 at 09:13:18AM -0700, Brian Bull wrote:
> I am writing an application that has to dynamically load classes from
> other modules based on certain criteria.  I load the class initially
> like this:
> 
>  self.dict["handleUnknown"] =
> new.classobj("handleUnknown",(handleUnknown.handleUnknown,),{})
> 
> I later initialize this class by calling
> 
> runningClass = self.dict[clas](self.pluginLog)
> 
> where clas is the name of the class I want to load such as
> "handleUnknown" in this example
> 
> I then execute methods against the class object and delete it when I
> am finished.
> 
> What I am looking for is a way to load the class "fresh" each time I
> instantiate it.  For instance, if I make a change to the loaded class,
> I want that change to be activated the next time the class is loaded
> without having to restart the main program that loads the classes.

  The key here is probably to reload() the module you're getting the update
information from.  This might be as easy as checking the last-modified
timestamp on the file it is loaded from in whatever routine you use to
import it.

> 
> I also need to be able to load the module and class from strings
> instead of actually using the module/class name.  ie.
> 
> runningClass = handleUnknown.handleUnknown()
> 
> This is what I want to be able to do (something similar to this)
> className = "handleUnknown"
> runningClass = className.className()
> 

  I'm not sure what the point of the above lines of code are, but from your
description it sounds like you will probably be interested in the
__import__() and getattr() builtin functions.

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 22 days, 13:01, 5 users, load average: 0.99, 0.98, 0.99




From mykbourassa at cogeco.ca  Wed Apr  9 12:47:56 2003
From: mykbourassa at cogeco.ca (myk)
Date: Wed, 09 Apr 2003 12:47:56 -0400
Subject: PCA principal component analysis
In-Reply-To: <b1ab591.0304090601.302cc4b6@posting.google.com>
References: <b1ab591.0304090601.302cc4b6@posting.google.com>
Message-ID: <hfYka.2414$h%2.256728@read1.cgocable.net>

Don't really need a tool.  Just:

a)  "centre" your data (translate for zero mean and scale for unity 
variance);
b)  do svd (in NumPy I think) on the data set resulting from a);
c)  eigenvectors are columns of U and eigenvalues are diagonal of S.  
PCA scores are just the values in the columns of U i.e. 1st column is 
first PC scores, etc.

You have much more flexibility if you don't use a "package".  HTH

majb

sebastien wrote:

>Hi,
>
>Is there any PCA analysis tools for python ?
>If it does, do you have any idea on how well it would scale ?
>
>I have already seen PyClimate (but it is not available for Windows
>which will be one of the target). Is there some LAPACK like packages ?
>
>Thanks for reading this mail.
>
>Sebastien.
>  
>




From martin at v.loewis.de  Fri Apr 18 04:49:02 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 18 Apr 2003 10:49:02 +0200
Subject: question on buildinh /demo/embed
References: <mailman.1050635669.2708.python-list@python.org>
Message-ID: <m3d6jk9on5.fsf@mira.informatik.hu-berlin.de>

"Jing (George) Cao" <jcao at cs.uchicago.edu> writes:

> I try to build the example in /Demo/embed, but there is several compiling
> errors. It is said in README that I should modify $LIBS as in
> /Modules/Makefile.  But I can't find any Makefile in the /Modoles
> directory.
> 
> I am lost here, anybody can help me?

If you want to get help, you should report the precise errors you are
getting, and the precise version of the software you are using.

Regards,
Martin




From just at xs4all.nl  Tue Apr  8 06:00:06 2003
From: just at xs4all.nl (Just)
Date: Tue, 08 Apr 2003 12:00:06 +0200
Subject: Embedded MacPython in a Multi-Threaded App
References: <4ecaae91.0304071500.2f66da7a@posting.google.com>
Message-ID: <just-2AA8B3.12000608042003@news1.news.xs4all.nl>

In article <4ecaae91.0304071500.2f66da7a at posting.google.com>,
 b_orcutt at pobox.com (Bill Orcutt) wrote:

> I'm writing an C application for OS 9 and I need/want to call some 
> python code from inside of an MP Task Thread. Possible? I've had a 
> look at the official Python thread documentation and googled and 
> gleaned as much as I can from the web, but I still have yet find 
> something that works.

Please ask on the MacPython mailing list:
  http://mail.python.org/mailman/listinfo/pythonmac-sig

Just



From rmunn at pobox.com  Tue Apr 29 13:03:30 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 29 Apr 2003 17:03:30 GMT
Subject: file transfer question
References: <mailman.1051564904.25073.python-list@python.org>
Message-ID: <slrnbasnat.jnm.rmunn@localhost.localdomain>

Ali Dada <afd00 at aub.edu.lb> wrote:
> hi all:
> 
> i want to read a file from a network pc and use it in a local python script. i
> thought of an easy way: i log on from within the python script to the host pc
> using ssh then copy the file to my working directory using scp. below is the code:
> 
> import os
> os.popen('ssh username at hostname')
> os.popen('scp username at hostname:~/requiredFile ./copiedFile')
> 
> but what's happening is that when i execute the code, i get prompted for the
> password by ssh, so i enter it, and then the program freezes and i can do
> nothing.  how do you think i should alter the code? thanks in advance..

Why are you using os.popen() for this -- and why are you throwing its
return code away? os.popen() creates a pipe to the program you're
running so that you can read its output. It returns the opened pipe as a
file-like object, so that you can call read() and close() on it. But you
don't need to read the output from ssh or from scp; you should use
os.system() instead.

Furthermore, you don't need to call ssh here at all; scp will do the
logging-in required. Try:

    import os
    os.system('scp username at hostname:~/requiredFile ./copiedFile')

You should still get prompted for a password.

BTW, if you want to automate this script, you might need to get rid of
the password prompting; look into SSH's features for public/private SSH
keys. The manpage for ssh-keygen might be a good place to start.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From max at alcyone.com  Tue Apr 22 17:10:39 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 14:10:39 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <f6a31cb8.0304220810.547f5acd@posting.google.com>
Message-ID: <3EA5AFCF.DF8792F0@alcyone.com>

Ted Drain wrote:

> I think I'd rather see the behavior change when you do:
> >>> s
> To call __str__ instead of __repr__.  It seems to me that repr is a
> specialized functionality that should be explicitly asked for when
> needed.  __str__ is the more normal output and should be the default
> in 'print s' or 's'.

I disagree.  In the interactive interpreter, I want to see the repr of
evaluations of the expressions I type in, not the str.  If I see

>>> someInvolvedExpression
2

Is that the numeric 2? the string '2'? the string '2 \t '? some custom
type whose __str__ is defined to return 2?

When I'm using the interactive interpreter, I'm essentially in a
"debugging" mentality, whether or not I'm actually debugging.  I want to
see the most unambiguous and detailed information about what I'm
evaluating.  After all, if I wanted the __str__, I'd just write

>>> str(someInvolvedExpression)

or

>>> print someInvolvedExpression

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From gh at ghaering.de  Wed Apr 16 09:40:40 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 16 Apr 2003 15:40:40 +0200
Subject: Hi
In-Reply-To: <FA13E89D0804CB4DB5FB420B70E2C0CC0ADD952E@kecmsg07.ad.infosys.com>
References: <FA13E89D0804CB4DB5FB420B70E2C0CC0ADD952E@kecmsg07.ad.infosys.com>
Message-ID: <3E9D5D58.7090607@ghaering.de>

201005- M.S.Shivakumaraswamy wrote:
> Hi
> 
> I am working on windows
> 
> But there is no curses module for the windows

There are alternatives to curses on Windows, like /F's console module [1].

> Can any one suggest where can I get curses enabled python??

Currently your only option is to use Cygwin's Python instead.

-- Gerhard

[1] http://effbot.org/zone/console-index.htm





From max at alcyone.com  Mon Apr 14 05:13:05 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 02:13:05 -0700
Subject: conversion of INTEGER into string
References: <a01b8ad6.0304140110.2b055605@posting.google.com>
Message-ID: <3E9A7BA1.AEEB589B@alcyone.com>

Chris wrote:

> I want to convert an integer, e.g. 2000 into a string (i.e. "2000")
> As far as I know, the chr() funktion only works with integers  in the
> range of -256 to + 256
> 
> So, what is the right method ??

You want the str function/type:

>>> str(2000)
'2000'

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Bachelors have consciences, married men have wives.
\__/ H.L. Mencken
    WebVal / http://www.alcyone.com/pyos/webval/
 URL scanner, maintainer, and validator in Python.



From uptonothing at yahoo.com  Tue Apr  8 10:47:13 2003
From: uptonothing at yahoo.com (Mike)
Date: Tue, 8 Apr 2003 10:47:13 -0400
Subject: Logging into CVS using Python
Message-ID: <ShBka.143$N5.26584@mencken.net.nih.gov>

Howdy All,

Thanks to everyone who replied to my previous post, even though my post was
horribly worded. Now onto my question.

How do I use Python to login into CVS? Here's what I have tried, but it has
failed. I'm not even sure that I'm using the correct methods.

import popen2

pipe_in, pipe_out = popen2.popen2("cvs -d
:pserver:mike at server.com:1000/path/to/CVS login")

s = pipe_in.readline()

print s

pipe_out.write("PASSWORD")

pipe_out.flush()

pipe_in.close()
pipe_out.close()

The code just results in "cvs login: authorization failed." Any help would
be appreciated.

Thanks,
Mike





From mwh at python.net  Wed Apr 16 05:47:10 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 09:47:10 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it> <slrnb9o9ob.493.rmunn@localhost.localdomain> <eX%ma.18570$G
Message-ID: <7h31y02srj6.fsf@pc150.maths.bris.ac.uk>

Carl Banks <imbosol-1050441179 at aerojockey.com> writes:

> First, you're basing it on a bad example.  Rebinding sys.stdout is bad
> programming, period.

Whoah there!  Says who?  Would you like to try using IDLE or any other
IDE without binding sys.stdout?

Cheers,
M.

-- 
  I really hope there's a catastrophic bug insome future e-mail
  program where if you try and send an attachment it cancels your
  ISP account, deletes your harddrive, and pisses in your coffee
                                                         -- Adam Rixey



From syver-en+usenet at online.no  Tue Apr  1 17:40:04 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 02 Apr 2003 00:40:04 +0200
Subject: Creating instance
References: <b6d191$j7s$1@news-reader12.wanadoo.fr>
Message-ID: <u3cl17sfv.fsf@online.no>

Salvatore <salvatore at wanadoo.fr> writes:

> Hello,
> 
> I would like to create an instance of class
> from the string name of the class
> 
> class Foo:
> 	pass
> 
> classname = 'Foo'

Does this suit you?
 
instance = eval(classname)()

-- 

Vennlig hilsen 

Syver Enstad



From max at alcyone.com  Fri Apr 11 23:56:48 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 11 Apr 2003 20:56:48 -0700
Subject: mirroring text window to file
References: <3E9783E3.6010108@videotron.ca>
Message-ID: <3E978E80.AC7BA047@alcyone.com>

Marcel Achim wrote:

> I was wondering if there's an existing solution other than writing
> to both a file and an output text window (Tk text widget), as in :
> 
>    text_window.insert('end', output)
>    log_file.write(output)

Write a wrapper class that does both.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ When you talk to her / Talk to her
\__/ India Arie
    WebVal / http://www.alcyone.com/pyos/webval/
 URL scanner, maintainer, and validator in Python.



From duncan at NOSPAMrcp.co.uk  Thu Apr 10 05:40:38 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 10 Apr 2003 09:40:38 +0000 (UTC)
Subject: Python/C API
References: <mailman.1049956515.20986.python-list@python.org>
Message-ID: <Xns93596B25D1EC7duncanrcpcouk@127.0.0.1>

Jp Calderone <exarkun at intarweb.us> wrote in 
news:mailman.1049956515.20986.python-list at python.org:

>   How does one write generators in C, or does one not?
> 
You don't, but creating iterators has the same effect.

You need to create an iterator object with two methods __iter__ returning 
self and __next__ calculating the next value and returning it. All of the 
state must be saved in the object between calls as there is no way to save 
the C stack frame. You probably also need an __init__ method to set things 
up.

Use Pyrex or Boost in preference to the C api if you can. If you must use 
the C api directly then you are probably best looking at something liike 
itertoolsmodule.c and plagiarising the code for one of the iterator types.

Here's a pyrex example. It could be calling C functions directly, although 
for this example I didn't bother. Although it looks like Python the 
expressions all use genuine C ints.

---- Seq.pyx ---
cdef class MyGen:
    cdef int current

    def __init__(self, int initial):
        self.current = initial

    def __iter__(self):
        return self

    def __next__(self):
        cdef int current
        
        current = self.current
        if current == 1:
            raise StopIteration

        if current%2 == 0:
            self.current = self.current / 2
        else:
            self.current = 3 * self.current + 1

        return current
---- test.py ---
import Seq

for v in Seq.MyGen(7):
    print v,
print

print list(Seq.MyGen(23))
---- end ---

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From davec999 at yahoo.co.uk  Mon Apr  7 06:14:48 2003
From: davec999 at yahoo.co.uk (David Currie)
Date: 7 Apr 2003 03:14:48 -0700
Subject: Zope large PDF question
Message-ID: <e09988ac.0304070214.4212e2e9@posting.google.com>

Hi everyone.

I am having some problems serving large(ish) PDF files from Zope. 
Basically, Acrobat Reader seems to hang trying to open a PDF directly
from a Zope server when used as an IE plugin.  Downloading the file
with right-click/save-as works fine, opening the file in the plugin
when it is stored on an Apache server on the same machine and opening
the file from the same place also works without any problems.  So it
looks like the Reader plugin does not play well with Zope.  Does
anyone have any ideas on why this could be happening?  Does Zope have
problems transferring large files under some circumstances?

Any ideas/fixes would be much appreciated. :)

Thanks.

David Currie



From knight at baldmt.com  Tue Apr  1 12:45:21 2003
From: knight at baldmt.com (Steven Knight)
Date: Tue, 1 Apr 2003 11:45:21 -0600 (CST)
Subject: ANNOUNCE:  SCons 0.13 (build tool in Python) is now available
Message-ID: <Pine.LNX.4.44.0304011144290.22256-100000@theopera.home.baldmt.com>

SCons is a software construction tool (build tool, or make tool) written
in Python.  It is based on the design which won the Software Carpentry
build tool competition in August 2000.

Version 0.13 of SCons has been released and is available for download
from the SCons web site:

	http://www.scons.org/

Or through the download link at the SCons project page at SourceForge:

	http://sourceforge.net/projects/scons/

RPM and Debian packages and a Win32 installer are all available, in
addition to the traditional .tar.gz and .zip files.


WHAT'S NEW IN THIS RELEASE?

Release 0.13 contains an important fix for a critical bug found shortly
after SCons version 0.12 was released.

SCons version 0.12 has a bug that occurs when a build directory
is used without duplicating source files in the build tree
("BuildDir(duplicate=0)") and any subsidiary SConscript file includes
another SConscript file.

This bug has been fixed in version 0.13, and a test has been added to
the regression test base to catch this situation in the future.

Release 0.13 also contains an unrelated fix to support fetching files
using earlier versions of Perforce that don't use the Windows registry.


ABOUT SCONS

Distinctive features of SCons include:

  - a global view of all dependencies; no multiple passes to get
    everything built properly
  - configuration files are Python scripts, allowing the full use of a
    real scripting language to solve difficult build problems
  - a modular architecture allows the SCons Build Engine to be
    embedded in other Python software
  - the ability to scan files for implicit dependencies (#include files);
  - improved parallel build (-j) support that provides consistent
    build speedup regardless of source tree layout
  - use of MD5 signatures to decide if a file has really changed; no
    need to "touch" files to fool make that something is up-to-date
  - easily extensible through user-defined Builder and Scanner objects
  - build actions can be Python code, as well as external commands

An scons-users mailing list is available for those interested in getting
started using SCons.  You can subscribe at:

	http://lists.sourceforge.net/lists/listinfo/scons-users

Alternatively, we invite you to subscribe to the low-volume
scons-announce mailing list to receive notification when new versions of
SCons become available:

	http://lists.sourceforge.net/lists/listinfo/scons-announce


ACKNOWLEDGEMENT

Special thanks to Charles Crain for making this release possible by
finding and fixing the critical bug fixed in this release.

On behalf of the SCons team,

	--SK





From xnedra at aa-ha.com  Wed Apr 23 06:19:36 2003
From: xnedra at aa-ha.com (Xnedra)
Date: 23 Apr 2003 03:19:36 -0700
Subject: Setting the Title of the script window
Message-ID: <1814331d.0304230219.3c6a82de@posting.google.com>

Hi there,

is there a possibility to set the title of the window the python
script was started in?
I know it is possible with dos with these lines:

if os.name == "nt":
  cmdline = 'TITLE TestTitle'
  os.system(cmdline)

but i would like to set it independent of the operating system. Is
there a python command for this or does anyone know the corresponding
commandline for unix?

Thanks a lot,
              Katrin / Xnedra



From rmunn at pobox.com  Wed Apr  2 10:42:45 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 15:42:45 GMT
Subject: else: w/o except: - why not?
References: <mailman.1049156943.12744.python-list@python.org> <v8j9pb3v2cnra6@corp.supernews.com>
Message-ID: <slrnb8lc7s.ccp.rmunn@localhost.localdomain>

Cameron Laird <claird at lairds.com> wrote:
> I want to be clear about the judgment that "a bare except is bad
> practice".  Are you talking about any unqualified except?  Is
>   try:
>       f1()
>   except ZeroDivisionError:
>       f2()
>   except:
>       f3()
> an instance of "a bare except" in the sense that you intend here?

For myself, I'd say yes. Unless f3() is a function specifically written
to handle any and all kinds of exceptions, e.g. by returning an HTML
traceback over CGI instead of printing the traceback to stderr, then the
above code is bad practice.

> And, for you, does
>   try:
>       f1()
>   except:
>       f2()
>       raise
> become *not* bad practice?

Yes; the above code is *not* bad practice, IMHO, because it allows
exceptions (which presumably are *not* handled by f2()) to propagate
upwards. The problem with bare except is that there are almost always
some exceptions, like SystemExit, which you *want* to propagate upwards.
Catching those without re-raising them is usually a bug.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From mcfletch at rogers.com  Sat Apr  5 23:48:41 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sat, 05 Apr 2003 23:48:41 -0500
Subject: properties and get/set methods
In-Reply-To: <ad052e5c.0304052016.4c67355c@posting.google.com>
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com>
Message-ID: <3E8FB1A9.20507@rogers.com>

Dan Bishop wrote:

>Timothy Grant <tjg at craigelachie.org> wrote in message news:<mailman.1049581810.25706.python-list at python.org>...
>  
>
...

>>Is it possible to determine which attribute was used to as the trigger for the 
>>call to setvar() or getvar()? so that set_appropriate_attribute() or 
>>appropriate_attribute() can set the correct attribute or get the correct 
>>attribute?
>>    
>>
>
>Why not just use two different getvar and setvar functions?
>

When you're doing generic "framework" level programming, it's just not 
acceptable to have to define 2 (really 3) new functions for every 
property.  Compare:

    def setMyProperty( self, value ):
       do_the_generic_thing( self, 'myProperty', value)
       self.__dict__["myProperty"] = value
    def getMyProperty( self ):
       do_the_generic_thing( self, 'myProperty', value)
       return self.__dict__["myProperty"]
    def delMyProperty( self ):
       del self.__dict__[ "myProperty" ]
    myProperty = property( getMyProperty, setMyProperty, delMyProperty, 
"""Some documentation""")

versus:

    myProperty = StringProperty(
       "myProperty", """Some documentation""",
       defaultValue = "this value",
    )

In these cases, you're using properties to simplify the modelling of 
domain objects, or to provide some *generic* functionality over and 
above a regular value get/set.  You don't want to have to specify that 
functionality for every property, you just want to say "this is one of 
_those_ properties".  The code bloat (not to mention mindless 
duplication of code) when you get up to 15 or 20 properties for an 
object would be ridiculous.

As for how the problem ((a deficiency in the property API IMO) of not 
knowing the name being set/get within the get/set methods) gets worked 
around.  I (as seen above) do some comparatively minor (but still 
annoying) duplication of code by passing an explicit string version of 
the name to the property sub-class's constructor and then use that name 
when the property instances are called.  Technically, the string value 
can be anything, and certain sub-classes of property just ignore it, 
while in any case, it can be something different than the "public" name 
through which the property API calls were made :( .

i.e. this works, storing the value as instance.__dict__['_myProperty']:

    myProperty = StringProperty(
       "_myProperty", """Some documentation""",
       defaultValue = "this value",
    )

but the property doesn't know what name it's _really_ being accessed as 
in this case :( .

For the particular case of the original poster, where the 
number/complexity of things being done can be modelled with a single 
function that dispatches on the property name, consider overriding 
__getattribute__ and __setattr__, doing the generic operations where 
appropriate, then call the superclass' implementation

Enjoy,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From tweedgeezer at hotmail.com  Sun Apr 27 01:15:03 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 26 Apr 2003 22:15:03 -0700
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
Message-ID: <698f09f8.0304262115.386c2571@posting.google.com>

Frantisek Fuka <fuka at fuxoft.cz> wrote in message news:<b8cadp$1fj9$1 at ns.felk.cvut.cz>...
> When I try to call methods, I sometimes forget to include the 
> parentheses. Instead of:
> 
> if object.isGreen():
> 	do something...
> 
> i sometimes write:
> 
> if object.isGreen:
> 	do something...

Perhaps the answer is to remove the __nonzero__ method on
functions/methods.  Logically, what purpose does it serve?

Jeremy



From peter at engcorp.com  Mon Apr 21 14:44:11 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 21 Apr 2003 14:44:11 -0400
Subject: printing email parts of type message/external-body??
References: <ZKzoa.602320$sj7.24103068@Flipper> <3EA2CFF2.1440294E@engcorp.com> <iAWoa.649943$sj7.26068978@Flipper>
Message-ID: <3EA43BFB.7FE1FF84@engcorp.com>

Magnus Therning wrote:
> 
> Peter Hansen <peter at engcorp.com> wrote:
> > Magnus Therning wrote:
> >>
> >> I am having problems with the following piece of code, would appreciate
> >> any insight that can be offered!
> >>
> >> ---- test.py ------------
> >> from email.Message import Message
> >>
> >> meb_message = Message()
> >> meb_message.add_header('Content-Type', 'message/external-body',
> >>     access_type='local-file')
> >> meb_message.set_payload('name="test"\n')
> >> print meb_message
> >> -------------------------
> >> magnus at die:[test].$ python test.py
> >>
> >> Traceback (most recent call last):
> >>    g.flatten(msg.get_payload(0), unixfrom=False)
> >>  File "/usr/lib/python2.2/email/Message.py", line 179, in get_payload
> >>    raise TypeError, i
> >> TypeError: 0
> >
> > Almost looks like a bug.  According to email.Message's documentation,
> > a TypeError will be raised by get_payload() if an index parameter is
> > provided but the payload is a string, as in this case.
> 
> But no index is given, so that would indicate a bug, right?

Actually, if you look at the g.flatten() line in the traceback above,
you'll see that g.flatten() _is_ providing an index value of 0.  I
think that means somebody is not properly handling the documented 
special case where the TypeError is thrown.

-Peter



From mwh at python.net  Mon Apr 14 15:09:22 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 14 Apr 2003 19:09:22 GMT
Subject: parser stack size limit
References: <mailman.1050341002.21628.python-list@python.org>
Message-ID: <7h3of38sxp9.fsf@pc150.maths.bris.ac.uk>

Francisco Borges <frandebo at latt.if.usp.br> writes:

> There are some bogus lists inside it I know, but as soon as I start
> using this for more complicated sentences (linguistics being done here)
> the number of nested lists will increase.
> 
> Any ideas on how to avoid this? Could it be as simple as a file to
> modify and then recompile python? 

Probably, yes.  I'd look inside the Parser directory of the source;
grepping for the error message should point you on your way.

I'd also consider a saner way of building data structures...

Cheers,
M.

-- 
  ARTHUR:  Why should he want to know where his towel is?
    FORD:  Everybody should know where his towel is.
  ARTHUR:  I think your head's come undone.
                    -- The Hitch-Hikers Guide to the Galaxy, Episode 7



From peter at engcorp.com  Tue Apr 29 15:36:49 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 29 Apr 2003 15:36:49 -0400
Subject: pdb and threads
References: <slrnbas092.qkk.matloff@laura.cs.ucdavis.edu>
Message-ID: <3EAED451.2A5E379B@engcorp.com>

Norm Matloff wrote:
> 
> Is there a not-too-complicated workaround to get pdb to execute within
> threads?

I'm not sure if the above describes a well-known problem to others, but
I haven't had any problems using pdb (in our extremely limited fashion)
within threads. 

The only thing I ever do with it is "import pdb; pdb.set_trace()" and 
that works fine within threads so far.

What's the symptom of the problem you encounter?

-Peter



From zhitenev at cs.vsu.ru  Tue Apr 22 05:25:19 2003
From: zhitenev at cs.vsu.ru (Lexy Zhitenev)
Date: Tue, 22 Apr 2003 13:25:19 +0400
Subject: problem starting python interpreter in emacs (win98)
References: <V65pa.4289$AB5.573029@news02.tsnz.net>
Message-ID: <b831pv$2d65$1@serv.vrn.ru>

"selwyn" <selwyn at home.net.nz> wrote in message:
news:V65pa.4289$AB5.573029 at news02.tsnz.net...
> system: win98
> python version 2.2.2
> emacs  21.2.1
> ..........
> hi all,

<snip>

> My .emacs file is as follows:
>
> (setq load-path (cons "C:\Python22" load-path))


Not sure about ELisp, but try to use escaped backslash:

(setq load-path (cons "C:\\Python22" load-path))


Regards, Lexy





From cben at techunix.technion.ac.il  Fri Apr 11 08:29:20 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Fri, 11 Apr 2003 15:29:20 +0300 (IDT)
Subject: Macros in Python?
In-Reply-To: <b71qde$q71$07$1@news.t-online.com>
References: <b71qde$q71$07$1@news.t-online.com>
Message-ID: <Pine.GSO.4.51.0304111457190.15707@techunix.technion.ac.il>

Dominic wrote on 2003-04-09:

> (defmacro with-root-privs (() &body body)
>    (let ((oldidsym (gensym)))
>      `(let ((,oldidsym (geteuid)))
>         (seteuid 0)
>         (unwind-protect
> 	   (progn , at body)
> 	 (seteuid ,oldidsym)))))
>
> Is there an easy way to produce similar code
> in Python which gurantees that certain things
> happen around some code? (Maybe by using
> generators?)
>
To complete the list of solutions, there is an evil way to abstract
try..finally that is not recommended due to it's unreadability to
anybody that doesn't know the trick.  It also only works in CPython
with reference counting that invokes __del__ immediately.

The basic idea was to use a for loop and a generator:

def with_root_privs():
    oldid = geteuid()
    seteuid(0)
    yield None # dummy
    seteuid(oldid)

for _ in with_root_privs:  # `_` is dummy
    # Dohere whatever you wanted.  The generator will cause this to be
    # executed exactly once, with pre- and post- code excuted...

The problem, of course is that the for loop's body can raise an
exception, or exit the loop in other evil ways.  Luckily in CPython,
exiting the loop immediately drops the only reference to the
generator, so hooking it's __del__ would do the trick.  Writing it as
an iterator class instead of a generator would be be ugly but we can
only do it once, by a universal wrapper taking a naive generator and
returning a safe one.  Google for "abstracting try..finally with
generators" in c.l.py if you are interested.  Again, it's hard to
recommend this hack.  Instead, consider pushing PEP 310 :-).

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>

Never spend several weeks on a single homework - or you will miss a
nice Linux conference ;-(.




From b.maryniuk at forbis.lt  Wed Apr 30 12:37:18 2003
From: b.maryniuk at forbis.lt (Bo M. Maryniuck)
Date: Wed, 30 Apr 2003 18:37:18 +0200
Subject: MQ
Message-ID: <200304301837.18459.b.maryniuk@forbis.lt>

Hello.

Here we had debate with a lot of buzzwords, such as XML, Java, SOAP to all 
about JMS and Message Queues. 

Now, is here some already done stuff in Python? Well, I would not like to see 
java-bloatware and would like to deal with a Python. 

What do I need exactly? I need to receive a messages (it is what JMS all 
about) and queue it somewhere, then convert/parse and put to the Oracle 
database or vice versa: Oracle queues messages (like a buffering) and other 
stuff fetches it.

Does anybody know anything about what MQ/JMS is and how to fit it with a 
Python? I found this one: http://www.hare.demon.co.uk/pymqi/pymqi.html but it 
is for IBM MQ Series. It will be better to find something to deal with like 
SonicMQ or so... 

Any ideas or wishes are appreciated! ;-) Thank you.

--
Regards, Bogdan

"Besides, I think [Slackware] sounds better than 'Microsoft,' don't you?"
(By Patrick Volkerding)





From te0006 at web.de  Thu Apr 17 05:30:49 2003
From: te0006 at web.de (Thilo Ernst)
Date: 17 Apr 2003 02:30:49 -0700
Subject: PYTHON EDITOR!!
References: <mailman.1049690233.24009.python-list@python.org>
Message-ID: <95cc3fc4.0304170130.120efcbf@posting.google.com>

nanthan at siswa.utm.my (Pathmenanthan RamaKrishna) wrote in message news:<mailman.1049690233.24009.python-list at python.org>...
> 
> Linux, so can u suggest me the best editor(i mean use friendly 
> and graphical,i dont want command line editor)for python?
> 

SciTE (http://www.scintilla.org/SciTE) remains my favourite. Reasons:

- syntax highlighting for Python and ~20 other languages
- copes well with Python's special requirements w.r.t. indentation,
  spaces & tabs
- powerful search and replace (string or regex-based, as the situation
  necessitates), even find-in-files is there
- bookmarks
- syntax-based folding (if you have three class of function defs A,B,C
  and want to see A side-by-side to C, just reduce B to a single line 
  with a mouse click)
- call tips (limited but useful)
- good GUI but everything is available via keystrokes as well
- basic Windows-style keyboard bindings (C-c=copy etc.)
- execution from within editor, output visible in special console 
  subwindow, mouseclick on exception brings you to source location
  (actually SciTE offers two actions "build" and "execute" for 
  compiled languages, I bound "build" to pychecker in the config file)
- easily configurable via X11 style config file
- available for both Linux and Windows

These features in my view make SciTE more than just an editor;
it can reasonably be called a basic free, cross-platform,
multi-language mini-IDE. Something to always have on your memory stick
(it's small, too).

- Thilo



From john.burton at jbmail.com  Wed Apr 16 06:17:45 2003
From: john.burton at jbmail.com (John Burton)
Date: Wed, 16 Apr 2003 11:17:45 +0100
Subject: DB API for ADO/DAO?
References: <mailman.1050474265.16331.python-list@python.org>
Message-ID: <9kaj7b.56q.ln@jbmail.com>

Bjorn Pettersen wrote:

> > From: Jimmy Retzlaff [mailto:jimmy at retzlaff.com] 
> > 
> > Steve Holden (sholden at holdenweb.com) wrote:
> > > "John Burton" <john.burton at jbmail.com> wrote in message
> > ...
> > >> The problem with odbc is that you need to manually set up data
> > >> sources using the control panel which isn't something I want 
> > >> the users to have to do.

> Agreed, but you don't have to even in ODBC :-) I'm assuming that what
> you're really looking for is a way to call:

>   hr = SQLConfigDataSource(
>          NULL,
>          ODBC_ADD_SYS_DSN, 
>          LPCTSTR(driverDescription), // I believe for display purposes
> only
> 	   "DSN=foo DATABASE=DbName TRUSTED_CONNECTION=YES
> SERVER=ServerName ");

> which will create a new system data-source called "foo". I don't
> remember seeing it exposed in either odbc or mxODBC, and it's too late
> for me to remember all (not to mention the best) way to call a random C
> function from Python, so I'm leaving that as an excercise for the reader
> <wink>.


I didn't know you could do that. It shouldn't be too hard for me to code
up a little c extension to do this if all else fails.

And thanks for all the other replies here and in email. Some useful
information there.




From mt at open2web.com  Mon Apr 28 14:22:07 2003
From: mt at open2web.com (Don Low)
Date: 28 Apr 2003 18:22:07 GMT
Subject: problem with string.lower function
Message-ID: <b8jrgf$ashp7$1@ID-145503.news.dfncis.de>

I'm learning Python from the "Core Python" book by Wesley Chun. In
chapter 3 is a script called fgrepwc.py, which I understand well enough.
At the end of the chapter, he asks the reader to modify the script so
that it performs a search in a case-insensitive manner.

Here's an excerpt from the script:

def filefind(word, filename):
	count = 0
	try:
		fh = open(filename, 'r')
	except:
		print filename, ":", sys.exc_info()

	allLines = fh.readlines()
	fh.close()

	for eachLine in allLines:
	if string.find(eachLine, word) > -1:
		count = count + 1
		print eachLine,

This is how I modified the script to make it search in a
case-insensitive manner:

def filefind(word, filename):
	count = 0
	try:
		fh = open(filename, 'r')
	except:
		print filename, ":", sys.exc_info()

	allLines = fh.readlines()
	fh.close()

	for eachLine in allLines:
	test = string.lower(eachLine) // change is here
	if string.find(test, word) > -1:
		count = count + 1
		print eachLine,

When I run the script, however, I get the following:

 File "2bin/./fgrepwc.py", line 37
 		 if string.find(eachLine, word) > -1:
	     ^
SyntaxError: invalid syntax

Note: I imported string so the problem isn't with a missing module.

Any help is greatly appreciated.

-- 
Thanks,

Mark



From dale at riverhall.NOTHANKS.co.uk  Wed Apr 16 09:59:29 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 16 Apr 2003 14:59:29 +0100
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com>
Message-ID: <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com>

danra at walla.co.il (Danra) wrote:

>def f(fileobj):
>    if not hasattr(fileobj,'write'):
>        raise TypeError, "Given arg must have the 'write' attribute."
>    .
>    .
>    fileobj.write('...')

You want assert, you do:

def f(fileobj):
    assert hasattr(fileobj, 'write'), 'Bang! Sort your code out.'

-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From scott_list at mischko.com  Wed Apr 16 22:39:13 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Wed, 16 Apr 2003 19:39:13 -0700
Subject: HTML form pre-populating functionality or hints on building my own
Message-ID: <200304161939.13004.scott_list@mischko.com>

Hi!
I'm wanting to find (or make) a module that will take a HTML file, locate the 
input-related form tags (<INPUT, <TEXTAREA, <SELECT, <OPTION) and modify them 
by putting in values from a dictionary (i.e. pre-populate the form). I want a 
module that is implemented 100% in python and does _not_ use any special tags 
to do this (such as webware, albatross, spyce, etc.)

The only thing I've seen that does this is EmbPerl (I'm tired of Perl) and 
Integratis' DFP module (implemented in C and not a straight-forward install).

I'm working on bringing some of EmbPerl's functionality to Python.

If I should build this myself, I'd like to know, would be better to use a text 
parsing module such as Plex to do the parsing?  I've never used any text 
parsing tool in Unix so it'll be all new to me.  I want to keep it simple as 
a standalone module to only populate form elements from a variable 
dictionary, so I'm thinking a parser might be overkill? If you think a parser 
would be the way to go on this, which one would you recommend? (Simple and 
quick is best IMHO).

Thanks!
Scott





From tim.one at comcast.net  Fri Apr 25 11:57:12 2003
From: tim.one at comcast.net (Tim Peters)
Date: Fri, 25 Apr 2003 11:57:12 -0400
Subject: Reading a non-standard floating point format
In-Reply-To: <24944f78.0304250631.1778654e@posting.google.com>
Message-ID: <BIEJKCLHCIOIHAGOKOLHEEFNFHAA.tim.one@comcast.net>

[Fidtz]
> I'm trying to do a database extraction program. Part of this is
> reading a large number of files with a 48bit floating point format
> that I think from the definition and searching on google, is unlike
> any on any other system (at least current ones anyway).

Sounds like a safe bet.  32-bit and 64-bit float formats are most common
now; 80-bit and 128-bit formats have some current presence.

> While I think I can see the way to do this with various bit-grubbing
> techniques, the potential for error is massive, esp considering my non
> comp-sci background.

Post a link to a definition of this format, and I bet someone will help.

> Does anyone know of a library in python or even (fairly) easily
> wrappable C that might let me specify the format in a more high level
> way?

Precise English will translate into Python easily enough <wink>.





From robin at jessikat.fsnet.co.uk  Sun Apr 13 07:51:45 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Sun, 13 Apr 2003 12:51:45 +0100
Subject: tuple weakref
Message-ID: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk>

What's the correct incantation to get a weak reference to a tuple? The
rather obvious wrapper approach fails as the wrapper dies immediately

ie weakref.ref(Wrapper((....)))

or we make a non anonymous wrapper which is what we want to avoid.
-- 
Robin Becker



From python at rcn.com  Fri Apr 11 17:36:21 2003
From: python at rcn.com (Raymond Hettinger)
Date: Fri, 11 Apr 2003 17:36:21 -0400
Subject: [Python-Dev] RE: Adding item in front of a list
References: <BIEJKCLHCIOIHAGOKOLHGEIJFEAA.tim.one@comcast.net>
Message-ID: <005201c30072$65a527a0$cb10a044@oemcomputer>

> [Andrew Koenig, starting with l=[2, 3, 4]]
> > ...
> > I would have thought that after l.insert(-1, 1), l would be
> > [2, 3, 1, 4], but it doesn't work that way.
> 
> Alas, list.insert() existed before sequence indices were generalized to give
> a "count from the right end" meaning to negative index values.  When the
> generalization happened, it appears that list.insert() was just overlooked.
> 
> I'd like to change this.  If I did, how loudly would people scream?

+1 from me.


Raymond Hettinger




From ckea25d02 at sneakemail.com  Fri Apr  4 01:06:22 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 04 Apr 2003 18:06:22 +1200
Subject: mp3 or wave  ---> midi , in python , how?
In-Reply-To: <b6isqq$23hs$1@mail.cn99.com>
References: <b6isqq$23hs$1@mail.cn99.com>
Message-ID: <b6j6sl$5m5dh$1@ID-169208.news.dfncis.de>

Frank Zheng wrote:
 > Subject: mp3 or wave  ---> midi , in python , how?
> 
>    Who can tell me how to do this in python ,are there any modules existing
> can help me?

My first thought was that this sounds almost impossible,
Python or no. But it seems there are things around that
do it, e.g.

    http://www.simtel.net/pub/pd/55886.html

Don't know about doing it from Python, though...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From david at no.westcontrol.spam.com  Mon Apr  7 09:25:53 2003
From: david at no.westcontrol.spam.com (David Brown)
Date: Mon, 7 Apr 2003 15:25:53 +0200
Subject: pyparallel errors
References: <3E8CD3AA.7070704@netscape.net>
Message-ID: <b6rtpf$teu$1@news.netpower.no>

"vector" <Vector180W at netscape.net> wrote in message
news:3E8CD3AA.7070704 at netscape.net...
> Thanks chris but I get this error any ideas?
> im running xp
> i wonder if I have to tell xp to release the printer
>
> PythonWin 2.2.2 (#37, Nov 26 2002, 10:24:37) [MSC 32 bit (Intel)] on
win32.
> Portions Copyright 1994-2001 Mark Hammond (mhammond at skippinet.com.au) -
> see 'Help/About PythonWin' for further copyright information.
>
>
>  >>> import parallel
> Traceback (most recent call last):
>    File "<interactive input>", line 1, in ?
>    File "G:\Development\Python\Lib\site-packages\parallel\__init__.py",
> line 13, in ?
>      from parallelwin32 import *
>    File
> "G:\Development\Python\Lib\site-packages\parallel\parallelwin32.py",
> line 53, in ?
>      import _pyparallel
> ImportError: Couldn't access giveio device
>

(Note - I've skipped this newsgroup for a while, so this might be a repeat
of something you've already read.)

I've only used giveio with NT and W2K, but I think XP is much the same.  You
don't have to do anything with the printer - when a program uses giveio to
access the parallel port, it simply walks all over any other program using
it - I freely mix a number of debuggers and hardware that uses the parallel
port (although obviously I only run one at a time).  But it sounds much more
like you don't have giveio installed, or you don't have it running - try
"net start giveio".






From BPettersen at NAREX.com  Mon Apr 21 11:38:50 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Mon, 21 Apr 2003 09:38:50 -0600
Subject: Stackless 3.0 alpha 1 at blinding speed
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE025@admin56.narex.com>

> From: David Abrahams [mailto:dave at boost-consulting.com] 
> 
> Courageous <jkraska at san.rr.com> writes:
> 
> >
> > "non local goto" describes it well, but another way of looking at
> > it is as a "first class, user controlled context switch". That's
> > actually the meat of what it is. In unix, theoretically one could
> > set up the heavy weight equivalent of a continuation by the use of
> > setjmp and
> 
> Really?  I thought longjmp could only go back up the call stack,
> losing any execution context from inner frames.  The whole point of
> continuations, AFAICT, is that:
> 
>        1. execution context is preserved
>        2. You can transfer control up, down, or between call stacks.

All true, except I would probably call them invocation chains :-) I do
like the term "first class user controlled context switch" though, even
it it only implies continued execution :-)

[...]

> Also, "stacklessness" is really an implementation detail AFAICT - it
> just says that all execution context is stored in
> dynamically-allocated structures instead of using local variables on
> the 'C' stack (via a recursive-calling implementation).
> 
> Have I misunderstood this whole thing?

Which property exactly makes the C stack frames not dynamically
allocated? (the sequential property of the allocator, or the fixed size
of the stack itself?)

-- bjorn




From schulz.online at gmx.de  Thu Apr 24 09:55:13 2003
From: schulz.online at gmx.de (Birgit Lachner)
Date: Thu, 24 Apr 2003 15:55:13 +0200
Subject: Reimplementation of a slot?
Message-ID: <b88qc9$ie$1@online.de>

Hi ...
I created a ui-file with QtDesigner and translated it to Python with pyuic.
The new .py-file is then imported to the file with the main-function.
Everything is working. But I want to reimplement the existing slot in the
pyuic-generated file.
How can I do this? I search through the internet and found many documents
about Python, Qt and slots but nothing about how to reimplement slots.

Can you help me?

Here is the main-function (copied from a auto-executable file created by
pyuic :
####################################################
import sys
from hallo import *

def slotDruck():
    ???.setText("Hallo Welt") ###### not working :-(

if __name__ == "__main__":
   a = QApplication(sys.argv)
   QObject.connect(a,SIGNAL("lastWindowClosed()"),a,SLOT("quit()"))
   w = Form1()
   a.setMainWidget(w)
   w.show()
   a.exec_loop()
####################################################

Bye ... Birgit Lachner.



From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr  4 11:47:41 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 4 Apr 2003 16:47:41 GMT
Subject: newbie os.stat question
References: <AV_ia.7382$4P1.570905@newsread2.prod.itd.earthlink.net> <slrnb8p1ut.174.do-not-spam-ben.hutchings@tin.bwsint.com> <7h3vfxvjrxy.fsf@pc150.maths.bris.ac.uk>
Message-ID: <slrnb8rdpc.rs.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <7h3vfxvjrxy.fsf at pc150.maths.bris.ac.uk>, Michael Hudson wrote:
> Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> writes:
> 
>> In article <AV_ia.7382$4P1.570905 at newsread2.prod.itd.earthlink.net>,
>> Tipton Bandy wrote:
>> > 
>> > I've got my first python utility working pretty good, but
>> > I just need help on one point.
>> > 
>> > I need to test a filename and return true or false based on 
>> > whether the file is readable and writable by the user running
>> > my utility. Oh, yes, this is on Linux:)
> 
> [...]
> 
>> os.access
> 
> is almost never what you want, because it uses the real user id, not
> the effective user id.
<snip>

Well that's what you are most likely to want to check!  But writing a
setuid program should be done with extreme caution and I wouldn't want
to try doing it in Python.

Anyone contemplating doing so should read /Advanced Programming in the
Unix Environment/ <http://www.kohala.com/start/apue.html> and /Setuid
Demystified/ <http://www.cs.berkeley.edu/~daw/papers/setuid-usenix02.pdf>
(treat the latter as corrections for the former) and should use
misc/setuid-prog.c as a wrapper for the Python script.



From peter at engcorp.com  Wed Apr 16 10:31:06 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 10:31:06 -0400
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com>
Message-ID: <3E9D692A.3B6A3DA@engcorp.com>

Danra wrote:
> 
> Here's an idea I've been playing with for a syntax for type checking
> in python. 
>[snip]
> The advantages of the second definition are obvious - The fact that
> the argument passed doesn't have the write method would be detected
> immediately, rather than when trying to call the it.
> 
> Why is that so important? Well, for starters in many cases it isn't.
> In a simple script or application it really is sufficient to get the
> exception only when calling write.
> 
> But - what if:
> 
> 1) The stuff you do before calling write takes a lot of time, which
> could be saved if the error is detected when the function is called.

Uh, wouldn't this happen only *once*, anyway, since this represents
a fundamental flaw in the code?  On top of that, if you are doing proper 
testing, you would never release this code, so who cares that it takes
lots of time before it fails?

> P.S. I am not a expert on python or anything of the sort. It could be
> that all of the above is just balderdash in your opinions, in which
> case feel free to say so :)

Well, I wouldn't use a nasty word like "balderdash" (oh shit, I just did!)
in this family forum, but issues very similar to this _have_ been 
raised and discussed repeatedly in the newsgroup and the idea rarely
catches much "traction".

-Peter



From martin at v.loewis.de  Thu Apr 10 16:47:51 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 10 Apr 2003 22:47:51 +0200
Subject: pickle and security
References: <pan.2003.04.10.18.52.25.740636@physics.purdue.edu>
Message-ID: <m3wui2gjuw.fsf@mira.informatik.hu-berlin.de>

"Jason Miller" <jmiller at physics.purdue.edu> writes:

> As I understand it, the major security issue with unpickling untrusted
> sources is that it may cause python to instantiate objects, calling
> constructors that could do just about anything.  If I only want to
> unpickle objects that are not class instances, and (using cPickle) I set
> find_global to None, are there any security concerns that remain?

Depending on your version of cPickle, it also contains a call to eval,
to unpickle a string. This is believed to be safe (as only safe
strings are passed to eval), but you may want to review that specific
fragment of code.

Regards,
Martin




From do-not-spam-ben.hutchings at businesswebsoftware.com  Mon Apr 28 14:41:50 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 28 Apr 2003 18:41:50 GMT
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com> <7xvqa.4236$b67.1131392@read2.cgocable.net>
Message-ID: <slrnbaqtfe.ao.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <7xvqa.4236$b67.1131392 at read2.cgocable.net>, Mark Nenadov wrote:
> Anna wrote:
> 
>> My daughter just asked me about this "python stuph". So I pulled up
>> livewires.org.uk and showed her some basics.
>> 
>> She just coded her first while loop and for loop... she thought it was
>> kewl! In fact, we had to kick her off the computer and send her to bed cuz
>> she's got so much to do tomorrow...
>> 
>> FWIW, she's 12...
> 
> Wow! That is excellent!
> 
> I introduced by nephew to the Python language when he was about 8 or so. We 
> never get much further than the 'print' statement, though :>

I wrote my first programs when I was 6, in BASIC.  If I remember correctly,
they converted temperatures in degrees celsius to degrees fahrenheit, and
vice versa.  If Python had existed then, I might have been able to write
something interesting and wouldn't have got stuck in that stupid language
for so long.



From adalke at mindspring.com  Tue Apr  1 14:17:34 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Tue, 1 Apr 2003 12:17:34 -0700
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <b6coda$uog$1@slb2.atl.mindspring.net>

Mike Meyer:
>     Any literal composed entirely of M, D, C, L, X, V and I characters
>     that does not follow this format will raise a syntax error,
>     because explicit is better than implicit.

Shouldn't that be a ValueError?

And you should probably add that roman(1000000) will
raise an OverflowError.  But what does roman(0) do?

                    Andrew
                    dalke at dalkescientific.com





From tomchance at gmx.net  Sun Apr 13 15:23:36 2003
From: tomchance at gmx.net (Tom Chance)
Date: Sun, 13 Apr 2003 20:23:36 +0100
Subject: resizeable windows in pyqt
Message-ID: <B8jma.1829$WY.308@newsfep3-gui.server.ntli.net>

Hullo,

I know this sounds really daft, but I can't get my windows to be nicely
resizeable in PyQt. I'm using Qt Designer to make the GUI, and I start with
a MainWindow project, without any of the toolbars/menus (so a blank slate).
I then add the various widgets in grids, and when I preview it, the window
resizes but none of the widgets resize with it. I've set the sizePolicy as
"preferred", and played around with other likely looking options, but I
can't work it out! How do I do it?

Thanks,
Tom



From ckea25d02 at sneakemail.com  Mon Apr 14 22:36:31 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 15 Apr 2003 14:36:31 +1200
Subject: Python FAQ (Was: Re: Confusion: assignments create object references.)
In-Reply-To: <slrnb9kveb.1q1.rmunn@localhost.localdomain>
References: <mailman.1050080058.23175.python-list@python.org> <3E970241.483698C7@engcorp.com> <slrnb9kveb.1q1.rmunn@localhost.localdomain>
Message-ID: <b7fqj0$geod$3@ID-169208.news.dfncis.de>

Robin Munn wrote:
> and one for "Here are more questions that crop up from time to time; 

An OAQ (Occasionally Asked Quesions)?

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From grante at visi.com  Fri Apr 25 18:23:05 2003
From: grante at visi.com (Grant Edwards)
Date: 25 Apr 2003 22:23:05 GMT
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com> <3ea947e1$0$198$a1866201@newsreader.visi.com> <mailman.1051301620.25438.python-list@python.org> <3ea9ac4c$0$179$a1866201@newsreader.visi.com> <b8camh$g65$05$1@news.t-online.com>
Message-ID: <3ea9b549$0$175$a1866201@newsreader.visi.com>

In article <b8camh$g65$05$1 at news.t-online.com>, Martin v. L?wis wrote:
> Grant Edwards wrote:
>> But those C runtime library routines aren't redundant. they're
>> providing functionality that Python doesn't implement. 
> 
> They are *also* providing functionality that Python doesn't
> implement. There is quite some redundancy, too. E.g. a Posix C
> library supports regular expressions, which Python also does,
> and Python doesn't use the C implementation.

If Python doesn't use a particular library module, then it
won't get linked in.  The C run-time library isn't a single
monolithic object.

> > OTOH, having both Tcl and Python interpreters seems a bit
> > redundant.
> 
> "A bit" redundant would be fine, I think.
> 
>> I have used Scheme with Tk bindings, and that didn't use Tcl.
>> The Scheme interpreter was bound directly to the Tk widgets.
>> That results in several advantages to the application
>> programmer, but disadvantages to the binding maintainer. 
> 
> What are those advantages?

The big advantage was that Tk "understands" Scheme variables.
If you hook a Scheme variable to a Tk widget, then changes to
that variable are reflected in the widget, and changes to the
widget also change the variable.  IOW you don't have to use
"wrapped" string/int/etc. objects and then use accessor
methods.

Another advantage was that parameters passed to and and return
values from Tk methods all use native Scheme data types.

> Aren't they also obtained by exposing the Tcl object interface?

Don't know.

-- 
Grant Edwards                   grante             Yow!  Just imagine you're
                                  at               entering a state-of-the-art
                               visi.com            CAR WASH!!



From pwatson at knightsbridge.com  Tue Apr 22 12:59:53 2003
From: pwatson at knightsbridge.com (Paul Watson)
Date: Tue, 22 Apr 2003 11:59:53 -0500
Subject: Multiple assignment?
References: <mailman.1050918753.29031.python-list@python.org>
Message-ID: <3ea57490$1_1@themost.net>

The os.path.split() method returns a tuple of two strings.  It yields the
result of the UNIX commands dirname and basename to isolate the directory
name and filename.  Try experimenting with the interactive mode of Python.

>>> x,y=os.path.split("now/is/the time/for/all.txt")
>>> x
'now/is/the time/for'
>>> y
'all.txt'
>>> z=os.path.split("now/is/the time/for/all.txt")
>>> z
('now/is/the time/for', 'all.txt')
>>> for item in z:
...   print item
...
now/is/the time/for
all.txt
>>> type(z)
<type 'tuple'>


"Mehta, Anish" <Anish.Mehta at enst-bretagne.fr> wrote in message
news:mailman.1050918753.29031.python-list at python.org...
> Can someone tell me the what is being done here in this line of code.
>
> dir, file = os.path.split(input)
>
> Waiting for reply
>
> Thanks
>
> Anish MEHTA
>
>





From yaipa at yahoo.com  Fri Apr 11 13:26:49 2003
From: yaipa at yahoo.com (yaipa h.)
Date: 11 Apr 2003 10:26:49 -0700
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net> <3E96D270.B8E99B27@engcorp.com>
Message-ID: <6e07b825.0304110926.59c9d12c@posting.google.com>

>  
> >>> import this
> 
OT: Sure glade the reader to your reply knows what you are talking about.
    Otherwise he might be a little taken aback. ;^))



Peter Hansen <peter at engcorp.com> wrote in message news:<3E96D270.B8E99B27 at engcorp.com>...
> Mads Orbesen Troest wrote:
> > 
> > On Fri, 11 Apr 2003 10:48:49 +0200, Harald Massa wrote:
> > 
> > > Yes, I love Python, and it is the most elegant programming language I
> > > ever met.
> > 
> > I agree it's mostly elegant, but the OO features are clumsy in my oppinion
> > (coming as a professional developer from other languages). Not as bad as
> > Perl's bless-a-dictionary-into-being-a-class retro-fitting hack, of course,
> > but really; all the self me here and self me there is - to me at last -
> > exceedingly clumsy and a constant source of frustration.
>  
> >>> import this
> 
> 
> -Peter



From skip at pobox.com  Fri Apr  4 17:02:08 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 4 Apr 2003 16:02:08 -0600
Subject: how to test when correct output is "random"?
In-Reply-To: <20030404215312.GE7895@unpythonic.net>
References: <20030404215312.GE7895@unpythonic.net>
Message-ID: <16014.224.480191.124970@montanaro.dyndns.org>

    Jeff> I'd like to use this project as an opportunity to teach myself
    Jeff> something about testing methodology, but since the output is based
    Jeff> on a random number generator, how can the normal "unit test" or
    Jeff> "doctest" tools work for me?

Seed the rng with the same starting value.  I believe you will always get
the same sequence:

    >>> random.seed(42)
    >>> random.random()
    0.63942679845788375
    >>> random.random()
    0.025010755222666936
    >>> random.random()
    0.27502931836911926
    >>> random.seed(42)
    >>> random.random()
    0.63942679845788375
    >>> random.random()
    0.025010755222666936
    >>> random.random()
    0.27502931836911926

Skip




From aleax at aleax.it  Sun Apr 27 03:38:47 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 07:38:47 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <mailman.1051423062.28410.python-list@python.org>
Message-ID: <bOLqa.13827$K35.455183@news2.tin.it>

Andrew Bennetts wrote:
   ...
> For that matter, removing __nonzero__ from functions still doesn't help
> this case:
> 
>     sys.exit
> 
> So it's of pretty minimal benefit.

Not really.  Write the following erro.py file:

import sys
sys.exit

and now run:

[alex at lancelot booksa]$ pychecker erro.py
Processing erro...

Warnings...

erro.py:2: Statement appears to have no effect
[alex at lancelot booksa]$

See?  PyChecker CAN and DOES warn about "this case"!  But OF COURSE
it can't help with the following erro1.py:

import sys
if sys.exit:
    print 'ha!'

[alex at lancelot booksa]$ pychecker erro1.py
Processing erro1...
ha!

Warnings...

None
[alex at lancelot booksa]$

So, _if_ we could avoid functions and methods being usable as bools,
we'd help get (ASAP, at runtime) some errors that PyCheccker *CANNOT*
get through its static analysis.

This may not be doable if we need to support the given getattr idiom
f=getattr(sys,'whatever',None)
if f: ...
rather than requiring an "if f is None:".  And it may not be worth
doing, since the people who I think would be helped by this appear to
insist PyChecker should do it (apparently getting a prompt runtime
error isn't sufficient for them -- it must be PyChecker) and PyCheker
just can't.  But, for those of us that see PyChecker and runtime errors
in SINERGY (i.e., things are reasonably fine as long as AT LEAST ONE
of these catches an error early), the benefit wouldn't be minimal, IMHO.


Alex




From gupt_vive at hotmail.com  Mon Apr 21 04:48:03 2003
From: gupt_vive at hotmail.com (vivek kumar)
Date: Mon, 21 Apr 2003 14:18:03 +0530
Subject: widget like listview
Message-ID: <Law15-F45e4wk2UfxhU00007102@hotmail.com>

dear richard,

  THX for the info. But this MCScrolledListBox has two drawbacks (according 
to my project)..

1. Once you have fixed the widths of the columns the user can not chnage the 
widths of the columns as per his/her preference.

2. What about linux. As this module uses the win32all extensions. the code 
no more remains portable. Currently whatever i have coded i have coded 
keeping in mind that i have to run the application on both windows and 
linux.

>If what you require is a multi-column list widget, then try
>MCScrolledListBox widget from the PmwContribD package.

>See http://pmwcontribd.sourceforge.net/
>and http://sourceforge.net/projects/pmwcontribd/


>You need to have Pmw installed to be able to use it.

This was very much helpful for me. if you haven't provided this i would have 
been doing some crazy stuff to debug the code..

>There is a bug when used with Pmw 1.1 due to a name clash. A simple
>workaround is to replace the 2 occurrences of 'vertscrollbar' in
>MCScrollListBox.py with a different name e.g. 'mcvertscrollbar'.

THX,
Vivek Kumar

_________________________________________________________________
Gizmo freak? See what's the hottest. http://www.msn.co.in/Computing/Gizmos/ 
Read product reviews.





From hp at rentec.com  Wed Apr  9 18:21:52 2003
From: hp at rentec.com (HP)
Date: 9 Apr 2003 15:21:52 -0700
Subject: limitation  in urllib.urlretrieve(url, fn) ??
Message-ID: <10caf2e2.0304091421.2ec9bcbc@posting.google.com>

For example,
url = 'http://www.londonstockexchange.com/techlib/word/ser/Live8102.doc';
fn  = '/local/disk/file_name';

urllib:urlretrieve(url, fn) hangs in the air !!

This happens for many of the files in the above 
web site.  Those files have one thing in common:
their file size is larger than 200K.

Is there a limit on the file size in the function:
urllib:urlretrieve(url, fn) ??

(I am running python.2.2.2 on Sun Solaris 5.8).

HP



From drew at astro.pas.rochester.edu  Thu Apr  3 10:05:26 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 3 Apr 2003 07:05:26 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>
Message-ID: <9383481e.0304030705.524cf0fa@posting.google.com>

"Anders J. Munch" <andersjm at dancontrol.dk> wrote in message news:<3e8c083b$0$10386$edfadb0f at dread11.news.tele.dk>...
> "Drew Moore" <drew at astro.pas.rochester.edu> wrote:
> > Howdy! 
> > 
> > I submitted a pre-PEP on overloadable assignment. 
> > 
> > The idea is:
> > 
> > In situations where the assignment token "=" appears, and an augmented
> > assignment symbol (such as +=, -=, *=, etc..) would also be
> > syntactically correct, the assigned-to object would be checked for an
> > __assign__(self, other) method. 
> 
> You terminology is off.  There is no such thing as an "assigned-to"
> object in Python.  Assignments change bindings, not objects.

Right, the binding of the name always changes. When an augmented
assignment operator is overloaded, the method must return a
reference to an object, and this returned reference is bound to the
name on the left hand side. Typical code inspects the "other" object,
modifies self appropriately, and returns a reference to self.
Regular assignment is just the trivial case of augmented assignment.

Seems to me a += b -> __iadd__(a,b)
is very similar a = b -> __assign__(a,b)

the __assign__ method returns a reference, just like __iadd__
this reference will be bound to the name on the left hand side.

> Take a step back and tell us what problem you are trying to solve.
> Whatever it is, I'm sure we can think of a better solution than having
> assignment depend on whatever object, if any, happened to be
> previously bound to the same name.
> 
> python-is-not-c++-ly y'rs, Anders

my original need?
I wanted to create an object that controls a voltage and
use it at the python command line.
by overloading operators, I can do:

voltage += 5  # (raise the voltage by 5 volts)
voltage *= 2  # (double the current voltage)

but when I do 

voltage = 3  # (set the voltage to 3 volts)

my voltage object is clobbered with an integer object.
I don't want to require the user to type any more than
name = value # no name.val = newval, no name.setval(value)

Overloadable augmented assignment provides a  nice framework that
always rebinds the name, but gives the "about to be rebound" name
some say in how this takes place. I'm at a loss to explain why
the "most trivial case of augmented assignment" was denied this power.

Thanks for responding.. Maybe posting on April 1st was a mistake,
people might have thought I was joking around!!



From tim.one at comcast.net  Sun Apr 20 20:34:24 2003
From: tim.one at comcast.net (Tim Peters)
Date: Sun, 20 Apr 2003 20:34:24 -0400
Subject: Good Python generators example?
In-Reply-To: <7%Eoa.1624$hT2.1255148@news2.news.adelphia.net>
Message-ID: <LNBBLJKPBEHFEDALKOLCCEKAEDAB.tim.one@comcast.net>

[Robert Oschler, asks for generator examples; Andrew Koenig gives an
 N-Queens solver]
> Thanks!  Interesting choice of examples, last time I solved that
> problem was with Prolog.  With a stack of Python generators you can
> pretty  well emulate Prolog backtracking, except for the automatic
> unbinding of variables.

Check out test_generators.py too, in your Python Lib/test/ directory.  There
are three implementations of a conjoin() generator in there, which combines
other generators (passed as a list of arguments) in a backtracking fashion.
One application of conjoin() in test_generators is an efficient N-Queens
solver.  Another is a Knight's Tour solver.  One of the conjoin()
implementations is non-recursive, and can be used to solve N-Queens or
Knight's Tour problems for boards much larger than Python's recursion limit
(which varies across platforms, but which is always limited by the platform
C stack size).  (Note that an N-Queen's solver normally uses one recursion
level per row, but Knight's Tour needs a level for each square.)

An interesting feature of the Knight's Tour solution is that it uses
different generator functions for the first, second, and last moves than it
uses for the other m*n-3 moves (which latter all use the same generator
function).  This allows the generator function used most often to skip
conditional tests unique to the endpoints, and so speeds it significantly;
it also allows, e.g., the first move to be special in a natural way (since a
Knight's Tour is a cycle, the first move may as well pick on the upper left
square -- any solution can be viewed as starting there).

There is, as you note, no automatic backtracking of binding states in
Python.  The examples illustrate by-hand data backtracking via saving and
restoring state via a combination of local vars, lexical closures, and
instance vars.





From roy at panix.com  Wed Apr 16 08:50:42 2003
From: roy at panix.com (Roy Smith)
Date: Wed, 16 Apr 2003 08:50:42 -0400
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com>
Message-ID: <roy-92A049.08504216042003@reader1.panix.com>

Peter van Kampen <news at datatailors.xs4all.nl> wrote:
> I think I almost 'get it'. Except who or what decides what Big-O a
> certain algorithm has. Is that by 'hunch' or by running benchmarks or
> are there still other ways to recognize 'Big-O's

You do it by analysis.  Basicly, you find the "inner loop", i.e. the bit 
of core code that gets run over and over while the algorithm does its 
work.  Then, you figure out (by looking at the code) how many times that 
inner loop will run given that the input is a given size.

Sometimes, in a high level language like Python, there are implied loops 
in what look like atomic operations.  This can obscure the real 
algorithmic complexity of a program, unless you know enough to look 
deeper.  Take the classic Python example:

s1 = ""
for s2 in stringList:
   s1 += s2

On the surface, it looks like this should be an O(N) program, where N is 
the number of strings in stringList.  Certainly, for N strings, the loop 
gets executed N times.  But, the string addition has an couple of 
inherent loops of its own.  The most obvious is that each character in 
s2 has to be copied someplace, so it really looks like this:

s1 = ""
for s2 in stringList:
   for c in s2:
      s1 += c

Now it looks like it takes O(k*N) steps, where N is again the number of 
strings in stringList, and k is the average number of characters in a 
string.

Here's the first tricky part -- for the purpose of big-O notation, the k 
is meaningless.  It's essentially a constant factor.  For a given 
universe of strings (say, words in a dictionary), k is not going to 
change as you add more strings.  Sure, it's going to have minor 
variations as the average bops around, but once you've got a statisticly 
valid sample of input strings, it'll pretty much stay put.  k may be a 
function of the kinds of strings you're giving it (k might be bigger for 
German compared to English), but it's not a function of the number of 
strings you're feeding the program.  For big-O purposes, O(k*N) where k 
is a constant is the same as O(N).  What we're trying to describe is the 
shape of the growth curve, not the scaling factors.

And, here's the second tricky part -- since strings in Python are 
immutable, the way you do s1 += s2 is to figure out how long each of the 
two components are, allocate a new string big enough to hold the sum of 
those, and copy the two original strings into the new one.  So, let's 
tear that apart a bit and see what it looks like, complexity wise.

We need to figure out how long each component string is.  On the 
surface, that's O(N), where N is the length of the string.  In C, it 
looks something like:

length = 0;
while (*string != NULL) {
   length++;
   string++;
}

This takes 2*N steps (it does two additions on each of N trips through 
the loop), but big-O says that O(2*N) is the same as O(N).

Fortunately for us, Python doesn't need to do that.  Python stores 
strings in a way where the length is pre-computed and stored.  All it 
really needs to do to get the length is something like "return 
string.length", which is a constant-time operation (i.e. it takes the 
same amount of time regardless of the length of the string).  That's 
called O(1).

So, we could rewrite the whole thing as something like:

s1 = ""
for s2 in stringList:
   len1 = len(s1)
   len2 = len(s2)
   newLen = len1 + len2
   temp = allocate buffer for newLen characters
   for c in s1:
      temp += c
   for c in s2:
      temp += c
   s1 = temp

Now, we analyse the complexity of each bit of that.  It does the outer 
loop N times, but how much work is involved in each iteration of the 
outer loop?  Each of the len()'s we already figured out is constant 
time, so they don't add anything complexity-wise.  Adding the two 
lengths is also constant time.

For the sake of argument, I'm going to assume that memory allocation is 
constant time, although I don't really know that for sure.  To really do 
the analysis right, I'd have to know the details of how Python memory 
management works, and I don't, but I'm reasonably confident that it's 
not going to be a determining factor here.

Now we come to the first interesting part.  The "for c in s1" loop takes 
as many iterations as s1 is long.  Well, how long is s1?  It varies as 
the program progresses, but after we've processed M strings, it's 
approximately k*M characters long (where k is again the average string 
length).  We again toss out the constant k, and come up with the 
complexity of this particular loop being O(M).  But, what's M?  It 
varies from 1 to N, and on average, it's N/2.  But, the constant factor 
of 1/2 doesn't mean anything to big-O, so O(M) = O(N/2) = O(N).

The "for c in s2" loop we've already discussed and decided it was O(k), 
which is the same as O(1), i.e. constant time.

And finally, the rebinding of s1 to the new temporary string is constant 
time.  One minor hitch is that this causes the old s1 to fall out of 
scope and become a candidate for garbage collection.  Again, without 
knowing the details of Python's memory management, I can't say for sure 
what the algorithmic complexity of freeing a string is, but for now I'm 
going to assume it's O(1).

So, overall, we do an O(N) inner loop (plus a lot of constant-time stuff 
that we're ignoring), O(N) times, giving us an overall algorithmic 
complexity of O(N^2).  This is often called quadratic behavior, and is 
generally considered evil and something to be avoided if at all possible.

How evil?  Well, I read recently that the new Python 2.3 release is 
supposed to be 10% to 20% faster than 2.2.  That sounds impressive (and, 
don't get me wrong, it is impressive), but it's bupkis compared to 
algorithm tuning.  Imagine you've got a program which runs in O(N^2) 
time.  If you could find a way to reduce it to O(N), for an input set of 
just 10 items, you would have speeded it up by a factor of 10!  Compared 
to upgrading your Python interpreter, you did 50-100 times better tuning 
the algorithm.  Now try to consider the implications of going from 
O(N^2) to O(N) with 25,000 input items!

BTW, this is perhaps one of the few arguments against teaching Python 
(or any high-level language) as a first programming language.  So much
algorithmicly complex stuff gets pushed down into atomic language 
constructs that there's no longer much of a correlation between code 
complexity and algorithmic complexity.  This, of course, is what makes 
high level languages so useful in the first place, but I fear we may be 
training a new generation of programmers for whom algorithmic complexity 
is not as familiar a concept as it used to be.

I'm not saying we should still be teaching C (or assembler) as a first 
language.  On the other hand, I suspect analysis of algorithms probably 
needs to be emphasized more (and earlier) in the curriculum.  When I 
first learned this stuff, it was mostly just a codification and rigorous 
analysis of concepts I'd already explored experimentally in the normal 
course of writing C programs.  I suspect now it's rather completely new 
ground, and may seem rather esoteric and theoretical and not 
particularly relevant to real-world problems like designing java 
animiations for web pages :-)



From jidanni at dman.ddts.net  Thu Apr 24 00:30:05 2003
From: jidanni at dman.ddts.net (Dan Jacobson)
Date: Thu, 24 Apr 2003 12:30:05 +0800
Subject: python -c "help('while')"
Message-ID: <878yu0h60i.fsf@jidanni.org>

Is this the easiest way to get help on things like "while" in one
shell command, outside the interpreter: 
$ python -c "help('while')"
-- 
http://jidanni.org/ Taiwan(04)25854780



From gerrit.muller at embeddedsystems.nl  Tue Apr 29 10:32:58 2003
From: gerrit.muller at embeddedsystems.nl (Gerrit Muller)
Date: Tue, 29 Apr 2003 16:32:58 +0200
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org> <1bf4033f.0304290627.40f7d831@posting.google.com>
Message-ID: <3EAE8D1A.8090209@embeddedsystems.nl>

Kent wrote:
> "Roman Yakovenko" <romany at actimize.com> wrote in message news:<mailman.1051603275.12037.python-list at python.org>...
> 
>>). I decided to propose 
>>python instead of VB. For this purpose I wrote small VB-Python compare.
> 
> 
> Managers always fear new development tools because of the productivity
> hit while we get up to speed on them.  Here's how I'm starting to
> convince them that Python is a good tool (and therefore worth the
> investment of time):
> 
> 1.  Show that Python can be used to implement our requirements.  This
> shows that Python is right for _that_ job.
> 
> 2.  Show that Python can be used to implement anything that can be
> implemented in the other environments.  This shows that it's not going
> to be a long-term hinderance, and that it's a real language, not just
> a toy.  The available GUI toolkits, and the immense standard library
> (with all its Internet protocol support) really show this point.
> 
<...snip...>

You bring a very important message here: avoid religious discussion, 
work instead by showing real value. If you are able to show a few 
relevant and not totally trivial functions up and running, including the 
actual hours spent on implementation and testing, than you have a 
self-selling case!

Working by example costs in general less time and is much more 
convincing and hence effective.

regards Gerrit

-- 
Gaudi systems architecting:
http://www.extra.research.philips.com/natlab/sysarch/




From thedustbustr at aol.com  Sun Apr 20 09:25:41 2003
From: thedustbustr at aol.com (TheDustbustr)
Date: 20 Apr 2003 13:25:41 GMT
Subject: PyOpenGL glClearColor() trouble
References: <mailman.1050806732.26162.python-list@python.org>
Message-ID: <20030420092541.18878.00000270@mb-m19.aol.com>

>use a real GUI library ;)
Which library do you suggest?

>I'm rather surprised if you care, it's very >seldom that such a call winds up
having a >significant overhead in an application
I *believe* that this is also the source of another problem (my textures don't
map to their poly's, my glColor3f() calls don't do anything), is this possible?
 I'm also using an event driven interface, could that be a problem?



From logistix at cathoderaymission.net  Fri Apr 18 17:18:18 2003
From: logistix at cathoderaymission.net (logistix)
Date: Fri, 18 Apr 2003 17:18:18 -0400
Subject: How to Import dll inside other directory
In-Reply-To: <ea0e2298.0304171906.37cb0b8a@posting.google.com>
Message-ID: <000001c305f0$08926e50$20bba8c0@XP>


> -----Original Message-----
> From: python-list-admin at python.org 
> [mailto:python-list-admin at python.org] On Behalf Of Alan Pong
> Sent: Thursday, April 17, 2003 11:07 PM
> To: python-list at python.org
> Subject: How to Import dll inside other directory
> 
> 
> a C extension dll (myext1.dll) compiled by vc++ and placed in c:\
> 
> run python 2.3.exe under c:\ in dos prompt, 
> >>from myext1 import *
> ok
> 
> then, move the dll to c:\temp, run python 2.3.exe again under c:\,
> >>from temp.myext1 import * (according to the help doc.)
> error:
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> ImportError: No module named temp.myext1
> 
> Does anybody know how to import dll inside other directory? 
> Thanks. Rgds. Alan
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 

The directory needs to be in pythonpath for it to be found.  Use "import
sys;print sys.pythonpath" to see what this is.  Assuming "C:\temp" is in
your pythonpath, you'd still only "import myext" not "import temp.myext"





From philh at cabalamat.org  Tue Apr  1 11:34:34 2003
From: philh at cabalamat.org (phil hunt)
Date: Tue, 1 Apr 2003 17:34:34 +0100
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <d2c6a353-612a-4f45-8ddd-bddc04d59e6a@news.szn.dk>
Message-ID: <slrnb8jfsq.9n8.philh@cabalamat.uklinux.net>

On Tue, 1 Apr 2003 08:39:43 +0000 (UTC), Klaus Alexander Seistrup <spam at magnetic-ink.dk> wrote:
>Chris wrote:
>
>> I would like to request that Python support multi-line
>> comments (e.g. /* */ in C).
>
>You could always enclose comments in triple-quotes:
>
>#v+

What does this mean?

>comment = """
>Sed ut perspiciatis, unde omnis iste natus error sit voluptatem
> [...]

Or just:

"""
This is a comment
"""

To make it more obvious where the comment begins/ends, I sometimes 
use:

"""*****
This is a multi-line comment.
Here is some more text, etc, etc, etc.
*****"""

Incidently, do other people often use comments to denote the data-type of 
parameters or instance variables, e.g.:

def filenameToHtml(fn):
   """ Remove any extension on a filename, and replace it with an .html
       extension.
   @param fn [string] = a filename
   @return [string] = a filename    
   """
   rightmostDot = string.rfind(fn, ".")
   if rightmostDot == -1:
      return fn + ".html"
   else:
      return fn[:rightmostDot] + ".html"   

def timeToStr(t):
   """ return a string representing a time 
   @param t [int]  = time in seconds since start of Unix Epoch
   @return [string]
   """
   fmt = "%a, %Y %b %d, %H:%M UCT"
   return time.strftime(fmt, time.gmtime(t))



-- 
Philip Hunt <philh at cabalamat.org>

Interested in adventure holidays in Spain? 
Look at: <http://www.cabalamat.org/advcon/>



From FBatista at uniFON.com.ar  Tue Apr  1 14:54:56 2003
From: FBatista at uniFON.com.ar (Batista, Facundo)
Date: Tue, 1 Apr 2003 16:54:56 -0300 
Subject: Precidence of unary plus/minus relative to exponentiation
Message-ID: <A128D751272CD411BC9200508BC2194D01885360@escpl.tcp.com.ar>

#- 	>>> -2.0**2.0
#- 	-4.0
#- 	
#- Python's operator predicence is causing this to be interpreted as:
#- 
#- 	>>> -(2.0**2.0)

It's OK, because the minus and the plus are the last verified signs, in
Python and everywhere else, as far I know.


#- rather than the normal mathematical precidence, which would give
#- 
#- 	>>> (-2.0)**2.0

No, no, no. In "everyday math" if you want to elevate the -2 to 2, you need
to put the brackets explicitly.


.elm.




From missive at frontiernet.net  Mon Apr  7 21:03:44 2003
From: missive at frontiernet.net (Lee Harr)
Date: Tue, 08 Apr 2003 01:03:44 GMT
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com> <b6soav$7p2$1@panix1.panix.com>
Message-ID: <Qdpka.2889$Zu2.2627@news01.roc.ny.frontiernet.net>

In article <b6soav$7p2$1 at panix1.panix.com>, Aahz wrote:
> In article <fd1c664e.0304071056.31a0d377 at posting.google.com>,
> Danra <danra at walla.co.il> wrote:
>>
>>Here's my experience with 'super':
>>
>>>>> class b:
>>	pass
> 
> No time to test, but try:
> 
>     class b(object):
>         pass


>From the super docs ;o)  it says you might do this:

    class C(B):
        def meth(self, arg):
            super(C, self).meth(arg)


but why not just do:

    class C(B):
        def meth(self, arg):
            B.meth(self, arg)



Is there any difference?

Any advantage either way?




From peter at engcorp.com  Tue Apr  1 07:45:14 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 07:45:14 -0500
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de>
Message-ID: <3E8989DA.924C4985@engcorp.com>

Thomas Guettler wrote:
> 
> I think about moving from Zope to
> ZODB and (Quixote or Twisted).
> 
> I looked at zope3 but it seems very bloated.
> 
> Up to now Zope seems the best http server for ZODB.
> 
> Has anyone experience with serving an application
> which uses ZODB as database backend but does not use Zope?

What goes in the database in the case you envision?  Is it
*content* as in content management as in how Zope did it,
or are you looking at a more traditional web server approach
where the database is just used to store, for example, user
data and such?

We've successfully connected Twisted to ZODB, as a quick
experiment (it takes, like, five minutes, right?) but haven't
grokked enough of Twisted yet to conclude what approach should
be taken for content management.

-Peter



From donn at u.washington.edu  Thu Apr 17 20:31:25 2003
From: donn at u.washington.edu (Donn Cave)
Date: 18 Apr 2003 00:31:25 GMT
Subject: Using __repr__ or __str__ for own printable class?
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org>
Message-ID: <b7nh0t$33li$1@nntp6.u.washington.edu>

Quoth Steven Taschuk <staschuk at telusplanet.net>:
...
| Looking back over your posts in this thread, I see a few kinds of
| objection to the current documentation.
|
| First, there's the description of __repr__ and __str__ in terms of
| fuzzy properties their output should have -- "informal" vs
| "official", "convenient" vs "complete", and the like; you feel
| that these descriptions are unclear.  For this problem, a few
| examples would certainly help to clarify the intended meanings of
| these terms.  It might also help to give examples of where these
| representations would normally be used -- for debugging output,
| for normal logging, for display to the user of the application.

I think the terminology is worse than unclear.  The lessons drawn
from examples that appear along with this text are going to be
a function of the text, and hence also misguided.

| Second, I have some sense that you feel these terms are not merely
| unclear but misleading or wrong.  Is this true?  If so, do you
| have alternative fuzzy terms in mind (the "data" vs the "object"),
| or would you prefer that the documentation avoid fuzzy terms
| entirely, and only give examples?

The data vs. object distinction is the best I can do so far.  It
isn't all that fuzzy, either.  It's vague, in that it applies to
a very high level of generality, but it does not seem ambiguous
to me.

| Third, you (and others) specifically object to the notion that if
| possible one should have eval(repr(x)) == x.  Would you want this
| notion to be omitted entirely?  Mentioned as possible and
| sometimes good?  Described as highly desirable, though sometimes
| infeasible?  (Imho the docs currently do this last, though perhaps
| they could do so more clearly.)

It would be fine to document current practice.  __repr__ sometimes
supports that application, on purpose.  The problem is that the
current documentation tries to make it a requirement, as if there
might be some penalty for not supporting it.  On the contrary,
there is little reason to actually do this, and plenty of reasons
not to (can't, awkwardly large, requires introspection that breaks
abstractions, etc.)  I don't think the documentation needs to get
into that, just document current practice.

	Donn Cave, donn at u.washington.edu



From peter at engcorp.com  Wed Apr 16 10:22:18 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 10:22:18 -0400
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3E9D671A.C2D1CE4D@engcorp.com>

Michael Hudson wrote:
> 
> Peter Hansen <peter at engcorp.com> writes:
> > One of the most important things to learn as a programmer is this,
> > as stated by various people but most popularly attributed to Knuth:
> > "Premature optimization is the root of all evil in programming."
> 
> And Knuth really did say it!  In "Structured Programming with goto
> Statements", on page 268 of the rather naff scan[*] I have of it, we
> find:
> 
>  "We /should/ forget about small efficiencies, say about 97% of the time:
>  premature optimization is the root of all evil."
> 
> So even you have managed to misquote him, unless he (or someone else)
> said it differently somewhere else.
> 
> [*] you can get your own naff scan easily enough with google; I can't
>     remember where mine came from.

The last time I used Google (of which I'm well aware of course), it
told me that Knuth was in fact merely paraphrasing someone else 
(google can tell you who it was, if you wish ;-), and since I was
not directly quoting him, but merely paraphrasing "various people"
myself, I made no claims that he said precisely that.  It is, after all, 
the sentiment that matters, more than the specific words used to convey 
them.

-Peter



From avner at skilldesign.com  Mon Apr 14 12:53:32 2003
From: avner at skilldesign.com (Avner Ben)
Date: Mon, 14 Apr 2003 18:53:32 +0200
Subject: an ugly file-reading pattern
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net> <mailman.1050201322.13220.python-list@python.org> <slrnb9h7u4.it3.rmunn@localhost.localdomain> <3e9a857b$1@news.012.net.il> <slrnb9ksaa.1q1.rmunn@localhost.localdomain>
Message-ID: <3e9ae78f$1@news.012.net.il>

"Robin Munn" <rmunn at pobox.com> wrote in message
news:slrnb9ksaa.1q1.rmunn at localhost.localdomain...

> I must have been dealing with too much spam recently: when I saw "I have
> tried this and it works," I thought the next thing I was going to see
> would be instructions for sending $5 to these five addresses... :-)


    Good thing they don't put filters on messages sent to newsgroups...


> You're right, I didn't close the file explicitly. But explicit is better
> than implicit, so IMHO it's better to give a name to the file object so
> that you can close it explicitly:
>
>     input_file = file("somefile")
>     for line in input_file:
>         process(line)
>     input_file.close()


    The C# language has a nice construct: using (variable = whatever) { /*
code */ }. The variable in the using header is guaranteed to be
garbage-collected at the end of the block (or so I think). Could be useful
here.

        Avner.





From peter at engcorp.com  Thu Apr 17 13:02:25 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 17 Apr 2003 13:02:25 -0400
Subject: Audit tools
References: <mailman.1050590157.29243.python-list@python.org>
Message-ID: <3E9EDE21.476A7190@engcorp.com>

Bharath Sreekanth wrote:
> 
> Hello Everyone,
> I want to do an architectural audit on python. Can you suggest me some tools 
> that are ideal. Also, is it possible to get the design documents for python.

It's called "the source code"....  or maybe you want to see the unit tests
for Python, in the test folder?

What's the purpose of this audit?  Is it based on some standard, or just
a made-up internal type of thing that your company does to exclude highly
effective and robust software from being used.  ;-)

-Peter



From ovid at easystreet.com  Sat Apr 19 11:52:53 2003
From: ovid at easystreet.com (Ovid)
Date: Sat, 19 Apr 2003 11:52:53 -0400
Subject: Testing methods
In-Reply-To: <mailman.1050774511.9020.python-list@python.org>
References: <9d8ac130.0304190938.755d4b4c@posting.google.com> <mailman.1050774511.9020.python-list@python.org>
Message-ID: <va36jkmvptr8ee@corp.supernews.com>

Jp Calderone wrote:
>   Sometimes this is undesirable, but then the right answer depends on the
> particulars of why it is undesirable.  There are other approaches, but none
> that work as reliably.  For example:
> 
>     if hasattr(foo, 'some_method') and callable(foo.some_method):
>         ...
> 
>   callable() may return false positives, though.  I don't think it will ever
> return false negatives.

Thank you.  I was actually just playing around with something like that. 
  The main problem is that the method name is determined at run time. 
I've wound up with something like:

     def can(object, method):
         if not hasattr(object, method):
             return 0
         else:
             return callable(getattr(object,method))

While that seems to work okay, you did mention that callable() can 
return false positives.  Can you give some examples?

Cheers,
Ovid






From awheat at imaging.robarts.ca  Wed Apr 23 13:34:49 2003
From: awheat at imaging.robarts.ca (Andrew Wheatley)
Date: 23 Apr 2003 10:34:49 -0700
Subject: pmw listbox selection loses focus
Message-ID: <d44e8ee7.0304230934.37035ed@posting.google.com>

I've created a Pmw.ScrolledListBox on Redhat Linux 8.0.  If I mouse
click on a Terminal window or Emacs window for example (but not its
decoration) after selecting a selection, the selection is turned off
graphically.  However, the selectioncommand is not called.  How can I
prevent the selection(s) from being deselected when I click inside
other windows?

Andrew



From usenet at rob.dqd.com  Thu Apr 10 13:42:32 2003
From: usenet at rob.dqd.com (Rob Mayoff)
Date: 10 Apr 2003 10:42:32 -0700
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>	 <6ee58e07.0304061144.35c5ab92@posting.google.com>	 <m2k7e78j5s.fsf@localhost.localdomain>	 <mailman.1049662570.29853.python-list@python.org>	 <yfsr88fkppr.fsf@black132.ex.ac.uk> <mailman.1049687898.18892.python-list@python.org>
Message-ID: <e0caaad7.0304100942.192cf0c6@posting.google.com>

Ian Bicking <ianb at colorstudy.com> wrote in message news:<mailman.1049687898.18892.python-list at python.org>...
> The quasi-modal stuff
> is an idea, but not a successful one.

If you mean that quasimodes haven't been successful in popular text
editors - true.  But quasimodes are actually used in several popular
programs.  In Adobe Photoshop, holding down the space bar switches you
to the "hand" tool (which moves the image around); releasing the space
bar reverts you to the previous tool.  Holding down command-spacebar
gives you the zoom-in tool, and holding down option-spacebar gives you
the zoom-out tool.  (The modifier keys are different on Windows.) 
These are quasimodes.



From aahz at pythoncraft.com  Mon Apr 28 12:45:59 2003
From: aahz at pythoncraft.com (Aahz)
Date: 28 Apr 2003 12:45:59 -0400
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net> <b8jbrt$41$1@panix1.panix.com> <zAara.22169$K35.739509@news2.tin.it>
Message-ID: <b8jls7$qve$1@panix2.panix.com>

In article <zAara.22169$K35.739509 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>Aahz wrote:
>> In article <b8hc64$6n2$1 at slb9.atl.mindspring.net>,
>> Andrew Dalke <adalke at mindspring.com> wrote:
>>>
>>>        i = i + 1
>> 
>> I'm a little curious why you didn't write this as
>> 
>>     i += 1
>> 
>> Any reason (i.e. still need to support 1.5.2) or just habit?  If the
>> latter, why haven't you changed your habit?
>
>There is no special reason to "change that habit", e.g. not performance:
>  [...]
>No statistically significant / measurable difference.  The tiny difference
>in conciseness is hardly compelling one way or another, either.

However, I do think the difference in readability is somewhat compelling.
Using augmented assignment, I can think to myself, "Ah, this is mutating
an object."  (This aside from the fact that numbers can't actually be
mutated -- it's a convenient mental shorthand.)

And, of course, the readability scales up rapidly with longer names.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From scott_list at mischko.com  Sun Apr 13 01:20:07 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Sat, 12 Apr 2003 22:20:07 -0700
Subject: pydoc doesn't find related modules?
Message-ID: <200304122220.07479.scott_list@mischko.com>

Hi!

I have pydoc running on RH8.0 Python 2.2.1 using the web interface.

I installed plenty of site-packages, one being the MoinMoin wiki and it's 
site-package directory MoinMoin.

When I browse pydoc, it gives me errors in some cases:

problem in /usr/lib/python2.2/site-packages/MoinMoin/macro/Form.pyc - 
ImportError: cannot import name wikiutil

wikiutil.py is in /usr/lib/python2.2/site-packages/MoinMoin/.

problem in /usr/lib/python2.2/site-packages/MoinMoin/wikiutil.pyc - 
ImportError: cannot import name config 

config.py is in the same directory as wikiutil.pyc.

MoinMoin (the wiki) works great so I know it's finding this stuff correctly.

I'm fairly sure this has to do with the pythonpath not being set correctly but 
I'm not sure why MoinMoin works at all in that case. I think pydoc doesn't 
handle the path by recursing into subdirs or else the import statements are 
done differently in MoinMoin than they are in other modules (i.e. Plex) and 
this is confusing pydoc.

Can anyone help me get pydoc to working correctly?

Thanks!
Scott





From claird at lairds.com  Tue Apr 29 15:43:42 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 29 Apr 2003 19:43:42 -0000
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com> <b8lpof$q0f$1@pegasus.csx.cam.ac.uk> <Xns936C8AB67B739duncanrcpcouk@127.0.0.1>
Message-ID: <vatlfeb16jlled@corp.supernews.com>

In article <Xns936C8AB67B739duncanrcpcouk at 127.0.0.1>,
Duncan Booth  <duncan at rcp.co.uk> wrote:
			.
			.
			.
>The easiest solution is to leave the program in Python and just call the 
>relevant functions from C. Possibly the easiest way to do this would be to 
>use Pyrex.
			.
		[persuasive illustration]
			.
			.
ctypes <URL: http://starship.python.net/crew/theller/ctypes.html >
should be even easier to use.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From mwh at python.net  Sun Apr  6 13:10:47 2003
From: mwh at python.net (Michael Hudson)
Date: Sun, 6 Apr 2003 17:10:47 GMT
Subject: segfault with imp.load_module()
References: <mailman.1049534708.5967.python-list@python.org> <pan.2003.04.05.13.23.43.694594.12686@metaslash.com> <mailman.1049561469.28124.python-list@python.org>
Message-ID: <7h3znn3imaw.fsf@pc150.maths.bris.ac.uk>

Rob Brown-Bayliss <rob at zoism.org> writes:

> On Sun, 2003-04-06 at 01:23, Neal Norwitz wrote:
> 
> > What version of python, what platform, what compiler, what options?
> > Assuming item = 'oda' this code works for me.
> 
> Redhat 8
> [rob at musicbox rob]$ python
> Python 2.2.1 (#1, Aug 30 2002, 12:15:30)
> [GCC 3.2 20020822 (Red Hat Linux Rawhide 3.2-4)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> 
> 
> If item = "abc" then I get (as expected) ImportError: No module named
> abc
> 
> So any way, I did some more playing and my module had a line:
> 
> from bhh import verbose
> 
> But the bhh module is not in the path, it's in the parent of the
> ./items/ directory.  If I remove that line then there is no segmentation
> fault.

I'm not sure I follow (which may be my fault) but could you come up
with a small, self-contained example?  This certainly looks like a bug
from here.

Cheers,
M.

-- 
  This is an off-the-top-of-the-head-and-not-quite-sober suggestion,
  so is probably technically laughable.  I'll see how embarassed I
  feel tomorrow morning.            -- Patrick Gosling, ucam.comp.misc



From sholden at holdenweb.com  Thu Apr 17 07:32:55 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 11:32:55 GMT
Subject: HTML form pre-populating functionality or hints on building myown
References: <200304161939.13004.scott_list@mischko.com> <mailman.1050560546.21370.python-list@python.org>
Message-ID: <Hhwna.114818$It5.51036@news2.central.cox.net>

"Ian Bicking" <ianb at colorstudy.com> wrote in message
news:mailman.1050560546.21370.python-list at python.org...
> On Wed, 2003-04-16 at 21:39, Scott Chapman wrote:
> > Hi!
> > I'm wanting to find (or make) a module that will take a HTML file,
locate the
> > input-related form tags (<INPUT, <TEXTAREA, <SELECT, <OPTION) and modify
them
> > by putting in values from a dictionary (i.e. pre-populate the form). I
want a
> > module that is implemented 100% in python and does _not_ use any special
tags
> > to do this (such as webware, albatross, spyce, etc.)
>
> Huh... for some reason it's never occured to me to do it this way, I
> always emit the tags with the values filled in, rather than parsing the
> HTML.  An interesting idea.
>
> Of course someone may correct me, but I haven't seen anything like this
> for Python, and I at least try to keep an eye out for such stuff.
>
I'd be happy to :-). See

    http://wwwsearch.sourceforge.net/ClientForm/

It's a very interesting module, and appears (from my own limited testing) to
work well.

[...]

>From the documentation: """ClientForm is a Python module for handling HTML
forms on the client side, useful for parsing HTML forms, filling them in and
returning the completed forms to the server. It developed from a port of
Gisle Aas' Perl module HTML::Form, from the libwww-perl library, but the
interface is not the same"""

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From da5id at LUVSPAMwolfe.name  Wed Apr 30 09:33:30 2003
From: da5id at LUVSPAMwolfe.name (David Wolfe)
Date: Wed, 30 Apr 2003 09:33:30 -0400
Subject: file transfer question
References: <mailman.1051564904.25073.python-list@python.org> <slrnbasnat.jnm.rmunn@localhost.localdomain>
Message-ID: <hbmdnZdp6s_9TDKjXTWcpA@comcast.com>

> > but what's happening is that when i execute the code, i get prompted
> > for the password by ssh...

This is normal.  You could use a passphraseless key, but if you care
enough to use SSH in the first place, you probably shouldn't.  Try
using an 'authentication agent' to store your (enencrypted) keys in
memory; that way you only have to type in the passphrase once per
session.  OpenSSH comes with a program called ssh-agent (man ssh-agent)
that will do the trick.  On Windows, try Pageant/PuTTY:

http://www.chiark.greenend.org.uk/~sgtatham/putty/download.html

HTH...





From forodejazz at yahoo.es  Sun Apr  6 06:18:50 2003
From: forodejazz at yahoo.es (=?ISO-8859-1?Q?David_Asorey_=C1lvarez?=)
Date: 6 Apr 2003 03:18:50 -0700
Subject: Python & Music: Performance question
Message-ID: <1759581e.0304060218.164d4cd4@posting.google.com>

Hello, group.

I'm developing a small music program, and I have some doubts.

   I want define a class, Note, which represents a note. It has two
attributes, pitch and duration (quarter note, half note, etc). There
will be a lot of instances of this object, maybe hundreds of notes,
and the program needs access to all of this objects and read its
attributes.
   Which implementation will be more efficient?

Implementation A:

class Note:
   def __init__(self, pitch, duration):
      self.pitch = pitch
      self.duration = duration
   def change_pitch(factor): pass
   def change_duration(factor): pass
   # more methods ...

Implementation B:

class Note:
   def __init__(self, pitch, duration):
      self["pitch"] = pitch
      self["duration"] = duration
   def change_pitch(factor): pass
   def change_duration(factor): pass
   # more methods ...

Regards.

David.

P.S.: Excuse my poor English!



From mwh at python.net  Tue Apr 29 09:55:55 2003
From: mwh at python.net (Michael Hudson)
Date: Tue, 29 Apr 2003 13:55:55 GMT
Subject: Determining the volume ID for a file/drive?
References: <uv7savkms6lsgchcv6jh8ahtccu9lon3oi@4ax.com> <JZtra.19431$3M4.563286@news1.tin.it> <zlvra.19900$3M4.571480@news1.tin.it>
Message-ID: <7h3vfwxh0he.fsf@pc150.maths.bris.ac.uk>

Alex Martelli <aleax at aleax.it> writes:

> Alex Martelli wrote:
>    ...
> > Quick and dirty...:
> > 
> > import os
> > x = os.popen('C:\\')
> 
> OOPS -- TOO quick -- sorry, I meant os.popen('DIR C:\\')!!!

I was wondering what that did!

Cheers,
M.

-- 
  Two things I learned for sure during a particularly intense acid
  trip in my own lost youth: (1) everything is a trivial special case
  of something else; and, (2) death is a bunch of blue spheres.
                                             -- Tim Peters, 1 May 1998



From tim at bladerman.com  Sat Apr  5 10:11:18 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Sat, 05 Apr 2003 15:11:18 GMT
Subject: Calling C libraries from a Python script
References: <e81e0f3d.0304050639.37b59655@posting.google.com>
Message-ID: <qmCja.4148$Du.14013@newsc.telia.net>

"Joshua Goodlett" <joshua at goodish.org> wrote in message
> I'm trying to find examples of calling C library functions from within
> a Python script.  Any advice or pointers on where to look for examples
> would be appreciated.


There is a whole chapter in the docs dedicated to the subject,

Have a look at

http://www.python.org/doc/current/ext/ext.html

Good luck

Tim





From glenfant at NOSPAM.bigfoot.com  Mon Apr  7 14:36:01 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Mon, 7 Apr 2003 20:36:01 +0200
Subject: Customize rlcompleter
References: <b6risd$giq$1@biggoron.nerim.net> <80ab95a6.0304070828.4697e4b9@posting.google.com>
Message-ID: <b6sfn0$1aav$1@biggoron.nerim.net>

Eric,

Many thanks for the enlightenment. Gonna make my own Completer class.

--Gilles

"Eric Price" <eprice at tjhsst.edu> a ?crit dans le message de news:
80ab95a6.0304070828.4697e4b9 at posting.google.com...
> Gilles Lenfan wrote:

[...]




From footech at get2net.dk  Mon Apr 14 13:01:22 2003
From: footech at get2net.dk (Mikkel Rasmussen)
Date: Mon, 14 Apr 2003 19:01:22 +0200
Subject: PYTHON EDITOR!!
References: <mailman.1049690233.24009.python-list@python.org> <2db29vkr03tkr5avg6fnlvm3831go0pgmm@4ax.com>
Message-ID: <zNBma.88$C75.79@news.get2net.dk>

Based on that list I found EmEditor v3 (Windows) which seems good.

It's fast, has Python syntax colouring from the start and most importantly:
is fairly easy to get to show Unicode characters correctly (as opposed to
TextPad).

It's not free, but fairly cheap shareware.

I'm used to TextPad, but this editor seems to be able to most the same
things.

Best wishes
Mikkel Rasmussen
www.mellon.dk

>
> Have a look at
> http://www.python.org/cgi-bin/moinmoin/PythonEditors
>
> --
> Ren? Pijlman





From skip at pobox.com  Tue Apr 22 13:03:52 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 22 Apr 2003 12:03:52 -0500
Subject: Paintshop Pro 8
In-Reply-To: <000801c308ee$e4d07050$0c02a8c0@Arts>
References: <000801c308ee$e4d07050$0c02a8c0@Arts>
Message-ID: <16037.30200.857420.134378@montanaro.dyndns.org>

    Art> Have some theories about why Python fits so well with graphics
    Art> work, which I will spare the group.

    Art> It would, however, be great to see the graphics related Python
    Art> reesources collected in one place.

    Art> "Computer Graphics with Python" - my book, I thought of it first.

Over in the marketing-python mailing list I brought up the idea of
soliciting a group of white papers focused on Python's applicability to
different application domains.  If you're really serious about writing a
book, a thorough white paper on the subject might be a good thing to throw
at a potential publisher.  For details on the thread, start here:

    http://wingide.com/pipermail/marketing-python/2003-April/thread.html

and search for "white paper".

Skip




From ecastro at calarts.edu  Thu Apr 24 14:58:26 2003
From: ecastro at calarts.edu (Eduardo Castro)
Date: Thu, 24 Apr 2003 11:58:26 -0700 (PDT)
Subject: Installing ExpectPy Error message
In-Reply-To: <200304241917.47911.b.maryniuk@forbis.lt>
Message-ID: <Pine.GSO.4.21.0304241155070.1324-100000@muse.calarts.edu>

	
Hi

I am getting this error message when installing ExpectPy-1.8 on Solaris2.9
and Python-2.2

When type:  

./configure I get:

checking for python... none
configure: error: Cannot find python interpreter
configure: error: ./configure failed for test

has anyone has similar problem?

Thank





From exarkun at intarweb.us  Wed Apr 30 22:39:23 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 30 Apr 2003 22:39:23 -0400
Subject: plain object()
In-Reply-To: <slrnbb0el9.2iks.missive@localhost.my.domain>
References: <slrnbb0el9.2iks.missive@localhost.my.domain>
Message-ID: <20030501023923.GA32266@meson.dyndns.org>

On Thu, May 01, 2003 at 01:03:25AM +0000, Lee Harr wrote:
> 
> 
> Hi;
> 
> Can someone explain this?
> 
> 
> >>> a = object()
> >>> a
> <object object at 0x8149bc0>
> >>> a.x = 3
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> AttributeError: 'object' object has no attribute 'x'
> >>> class B(object):
> ...   pass
> ...
> >>> b = B()
> >>> b
> <__main__.B object at 0x81194cc>
> >>> b.x = 3
> >>> b.x
> 3
> >>>
> 

    print b.__dict__
    print a.__dict__

  Jp

-- 
C/C++/Java/Perl/Python/Smalltalk/PHP/ASP/XML/Linux (User+Admin)
Genetic Algorithms/Genetic Programming/Neural Networks
Networking/Multithreading/Legacy Code Maintenance/OpenGL
See my complete resume at http://intarweb.us:8080/
-- 
 up 41 days, 23:04, 2 users, load average: 0.19, 0.08, 0.11




From sholden at holdenweb.com  Tue Apr 15 09:11:00 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 15 Apr 2003 13:11:00 GMT
Subject: Python compared toJava
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com> <3e9bdf88$0$24726$edfadb0f@dread14.news.tele.dk>
Message-ID: <ExTma.98936$It5.95989@news2.central.cox.net>

"Frithiof Andreas Jensen" <frithiof.jensen at remove_this.jensen.tdcadsl.dk>
wrote ...
>
> Blaktyger <blaktyger at hotmail.com> skrev ...
> > Hi!
> >
> > I recently read that the Python interpreter is like the Java Virtual
> Machine.
> > Is that true?
>
> Not quite - Python actually *delivers* on "write once, run everywhere" ;-)
>

This is a little ingenuous: some modules are only available on some
platforms, and others have different functionality on each different
platform. I'm as anxious as anyone else that the world gets a good
impression of Python, but distorting the truth will only lead to
disappointment.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From mis6 at pitt.edu  Mon Apr 28 09:14:37 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 06:14:37 -0700
Subject: metaclasses, how you might actually use 'em
References: <mailman.1051458882.30400.python-list@python.org>
Message-ID: <2259b0e2.0304280514.5ca68d50@posting.google.com>

Jack Diederich <jack at performancedrivers.com> wrote in message news:<mailman.1051458882.30400.python-list at python.org>...

> <snip evangelism to metaclasses>

All your usages of metaclasses seems to me quite typical and "good"
(to see a "bad" usage, google for my "black magic trick" ;)

> If this is known stuff or there is a better standard way just mark 
> this thread as ignored and enjoy your Sunday. 

Do you know that the standard 'type'
metaclass (and therefore all metaclasses derived from it, which means
ALL metaclasses) contains a __subclasses__ method ?

>>> dir(type)
['__base__', '__bases__', '__basicsize__', '__call__', '__class__', 
 '__cmp__', '__delattr__', '__dict__', '__dictoffset__', '__doc__', 
 '__flags__', '__getattribute__', '__hash__', '__init__', '__itemsize__', 
 '__module__', '__mro__', '__name__', '__new__', '__reduce__', '__repr__', 
 '__setattr__', '__str__', '__subclasses__', '__weakrefoffset__', 'mro']

It works this way:

>>> class C(object): pass
...
>>> class D(C): pass
...
>>> C.__subclasses__() # means type(C).__subclasses__(C)
[<class '__main__.D'>]

I think it is quite similar in spirit to your Register, isn't it ?

> Corrections encouraged, pedantry especially.

your sentence

> metaclasses just allow you to tweak a class definition on the fly.  Mainly 
> adding or removing attributes just before the class becomes 'real'. 

is correct when you redefine the __new__ method of the metaclass. However,
in your examples, you redefine the __init__ method. This means that the
class has been *already* created, i.e. it is 'real', and it is modified
*after* creation. This is just nitpicking, but since you asked for pedantry...

Have fun,

                                   Michele



From nhodgson at bigpond.net.au  Mon Apr  7 08:28:58 2003
From: nhodgson at bigpond.net.au (Neil Hodgson)
Date: Mon, 07 Apr 2003 12:28:58 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org> <yfsr88fkppr.fsf@black132.ex.ac.uk>
Message-ID: <eaeka.1333$1h6.8933@news-server.bigpond.net.au>

Alexander Schmolck:

> He certainly knows about Emacs (have a look at the index of his book), but
I
> guess obtaining patents and attracting attention as some sort of visionary
is
> unnecessarily complicated by pointing out the prior art yourself. And
although
> the most crucial (and brilliant) idea, incremental search, is nicked (and
> uncredited on his webpage),

   Priority for incremental search is uncertain - Jef Raskin has been
pushing LEAP since at least the mid 80s with the Swyftcard and Canon Cat
products. The earliest publication reference found in a quick google is
November 1985. When did EMACS get incremental search?
http://archaic-apples.shauny.de/files/cat/canon.html

   Neil





From BPettersen at NAREX.com  Sun Apr 20 01:38:48 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sat, 19 Apr 2003 23:38:48 -0600
Subject: Is this a super bug?
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE00C@admin56.narex.com>

> From: Bjorn Pettersen 

Hate to reply to myself, but...

If this isn't the __getattribute__ bug I read about, it needs to be
documented that super only supports _direct_ access to attributes, but
not properties (although the Python version does), of a class in the
__mro__, but will not access any attribute the language defines as
"hooks" for various functionality indirectly... (i.e. all the __xx__
methods -- I could understand __getattribute__ would be a problem, but
even __getattr__ should have been possible to implement).

Sort of a special purpose proxy for the superclass... not what I
expected from the name. And there doesn't seem there is a way to
implement it in Python either. If I understand correctly, these
attributes are looked for directly in the instance's class, skipping
lookup in the instance itself, which is why the super instance's
__getattr__ or __getattribute__ aren't called. Question: are __getattr__
and __getattribute__ ever called on the metaclass? Why not/why not in
this case?

-- bjorn

> > From: Terry Reedy [mailto:tjreedy at udel.edu] 
> > 
> > "Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message
> >
> > >>> class ff(str):
> > ...    def foo(self):
> [...]
> > ...      print 'super(ff, self)[0]', super(ff, self)[0]
> > ...
> > >>> f = ff('asdf')
> > >>> f.foo()
> > super(ff, self)[0]
> > Traceback (most recent call last):
> >   File "<stdin>", line 1, in ?
> >   File "<stdin>", line 5, in foo
> > TypeError: unsubscriptable object
> > ----------
> > I suspect that printing type(super(ff,self)) might answer your
> > question.
> 
> I'm afraid not :-)  I allready did that, and from what I can 
> tell, this
> is the sequence that "should" happen:
> 
> 1  super(ff, f)[0]  # f == self
> 2  superinst(type:<class ff>, obj:f)[0]
> 3  si.__getitem__(0)
> 4    si.__getattr__('__getitem__')
> 5    .. return str.__dict__['__getitem__'].__get__(f)
> 6    boundmethod(str.__getitem__, f)
> 7  bmeth(0)
> 8   'a'
> 
> well, the transition from 3 to 4 doesn't happen which surprised me a
> little. What surprised me more was that I couldn't get 
> __getattribute__
> to be called either (neither in the class or its metaclass... 
> I've only
> got 2.3a2, and I read a note about this not working before a4. Is this
> what is happening? If not, the only other reason I could see would be
> from the refMan 5.3.2, "The primary must evaluate to an object of a
> sequence or mapping type."  If that is the case, then I still 
> think it's
> a super bug, i.e. there is no reason super should assume that it's not
> "going to be" a sequence, in fact the direct superclass of ff _is_ a
> sequence. To fix it simply add:
> 
>   def __getitem__(self, v):
>       m = self.__getattr__('__getitem__')
>       return m(v)
> 
> to the definition of super. This will change the exception to
> AttributeError if a superclass isn't a sequence. To keep the current
> semantics, add the lookup to only instances that are sequence proxies
> (example below, refactor between __getattr__ as desired...)
> 
> -- bjorn
> 
> class Super(object):
>     def __init__(self, type, obj=None):
>         self.__type__ = type
>         self.__obj__ = obj
> 
>         if isinstance(self.__obj__, self.__type__):
>             starttype = self.__obj__.__class__
>         else:
>             starttype = self.__obj__
>         mro = iter(starttype.__mro__)
>            
>         for cls in mro:             
>             if cls is self.__type__:
>                 break               
> 
>         for cls in mro:
>             if '__getitem__' in cls.__dict__:
>                 def getitem(v):
>                     m = self.__getattr__('__getitem__')
>                     return m(v)
>                 self.__possibleGetitem = getitem
>                 break
> 
>     def __possibleGetitem(self, v):
>         raise TypeError("unindexable object")
> 
>     def __getitem__(self, v):
>         return self.__possibleGetitem(v)
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 




From km-news1 at onlinehome.de  Fri Apr 25 15:37:42 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Fri, 25 Apr 2003 21:37:42 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <Kq9qa.9083$h%2.957726@read1.cgocable.net> <mailman.1051273778.5621.python-list@python.org> <fzo6emyy.fsf@python.net> <mailman.1051277378.13076.python-list@python.org>
Message-ID: <oprn6944nxmu3s0q@news.cis.dfn.de>

> And what about effbot's Tkinter 3000?
How active is this project? Maybe the release date could be the year 3000 
;-)

-- 
Gru? - regards Klaus :-)



From gregory_r_warnes at groton.pfizer.com  Fri Apr 18 12:38:41 2003
From: gregory_r_warnes at groton.pfizer.com (Warnes, Gregory R)
Date: Fri, 18 Apr 2003 12:38:41 -0400
Subject: PCA principal component analysis
Message-ID: <D7A3CFD7825BD6119B880002A58F06C202F2C847@groexmb02.pfizer.com>

- As others have noted, the correct URL for R is http://www.r-project.org.

- The 'Rpy' (http://rpy.sourceforge.net) interface between R and Python is
more reliable and better supported than the 'RSPython' interface.     I use
rpy to provide a SOAP interface to R
(http://software.biostat.washington.edu/statsoft/snake/RSOAP) and on top of
that I've build an interface between R and Zope
(http://software.biostat.washington.edu/RSessionDA).

- Finally, there are several different functions available for PCA.  I use
the 'prcomp' function from the 'mva' package, which is included in the R
distribution.

> -----Original Message-----
> From: Suchandra Thapa [mailto:ssthapa at classes.cs.uchicago.edu]
> Sent: Thursday, April 10, 2003 7:30 PM
> To: python-list at python.org
> Subject: Re: PCA principal component analysis
> 
> 
> sebastien <s.thuriez at laposte.net> wrote:
> > Is there any PCA analysis tools for python ?
> > If it does, do you have any idea on how well it would scale ?
> > 
> > I have already seen PyClimate (but it is not available for Windows
> > which will be one of the target). Is there some LAPACK like 
> packages ?
> 
> Have you considered using R (www.r-project.com) and RS/Python
> (http://www.omegahat.org/RSPython/).  R is a statistical 
> programming language
> with libraries to handle most common statistical methods.  
> Using RS/Python
> you can create R objects, populate them, and pass them into R 
> functions within
> python or vice versa.
> 
> -- 
> --------------------------------------------------------------
> --------------
> Suchandra Thapa                       
> s-thapa-11 at NOSPAMalumni.uchicago.edu  
> --------------------------------------------------------------
> --------------
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 


LEGAL NOTICE
Unless expressly stated otherwise, this message is confidential and may be privileged. It is intended for the addressee(s) only. Access to this E-mail by anyone else is unauthorized. If you are not an addressee, any disclosure or copying of the contents of this E-mail or any action taken (or not taken) in reliance on it is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately.




From sean at activeprime.com  Wed Apr 30 11:07:15 2003
From: sean at activeprime.com (Sean)
Date: 30 Apr 2003 08:07:15 -0700
Subject: Determining Unicode encoding.
References: <ef7a350b.0304291242.42c9b024@posting.google.com>
Message-ID: <ef7a350b.0304300707.7c43ad4c@posting.google.com>

Thanks for the amazingly helpful responses everyone.  Perhaps a more
detailed explanation of what I'm doing might be in order.  My program
is reading data from a commercial customer database through their API
(that I don't have access to).  Most of the data I get back are python
string objects, but I'll occasionally come across a Unicode object.  I
need to write this data out to text (.csv format) so that it can later
be read back in and reinserted into the database.  In order to write
to disk, I need to encode the Unicode object using some kind of
encoding.  What I am worried about is when I write to the file,
someone's accented 'e' or random non-Latin character will get changed
into something else to fit the standard character set.  Making the
problem worse is that the company that makes the database software I'm
accessing has spotty language support at best, and even worse
technical support.

>From what I'm hearing here, it sounds like I can just use UTF-8 and
not have to worry about it.



From BPettersen at NAREX.com  Sat Apr 12 16:00:20 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sat, 12 Apr 2003 14:00:20 -0600
Subject: win32all odbc/MS SQLServer/bigint
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEB0@admin56.narex.com>

> From: M.-A. Lemburg [mailto:mal at lemburg.com] 
> 
> Bjorn Pettersen wrote:
[...]

> The win32all odbc module fetches BIGINTs as floats. That's
> where you probably get your corruption from...

That's what I thought from reading the code, but I didn't want to imply
it publically in case I was wrong. I'm sure there are areas where loss
of precision is acceptable, but in business applications there is a
rather strong reaction when this happens, especially without
documentation, or warning, and the fact that the Python type in the
result set is a long not a float...

> > I looked at the mxODBC documentation, and it says it's doing 
> > a conversion [automatic] to string for bigint data... (again 
> > too slow). I'm pretty flummoxed, the ODBC C calls for 
> > bigint/__int64 are pretty straight forward(?)
> 
> They are ? __int64 is not exactly portable across platforms...

AIX, IRIX, SGI, Linux, SUN, HP, z/OS, OS/390, and Win* all support a 64
bit integer type (I did the research at my last company, and only for
the platforms we supported... -- I'm sure there are exceptions, but ODBC
isn't a lowest common denominator specification by design).

> BTW, conversion from strings to Python longs is done in C,
> so it's not as slow as you may think.

I haven't looked at the Py_LongFromLONGLONG(sp?) algorithm, so you may
be right. It was too slow when I wrote our odbc wrapper library in
C++...

> > As an added bonus, someone had the pleasure of showing me 
> > that a simple select/fetchall (i.e. all C code for Python), 
> > was over three times faster in C#..., and the win32all odbc 
> > module has crashed python somewhere deep down in sql three 
> > times in the last two hours (we're expecting this app to 
> > run for ~4 days...). Looks like it's back to C++ (*sigh*).
> 
> Have you tried mxODBC on this ? We have many customers running
> mxODBC under Zope on heavily loaded sites for weeks... so far
> without any complaints about its stability.

I will definitely try it later today (your reputation precedes you, but
upper management is cheap, especially when something doesn't come from
MS <groan>). My other options are wrapping our db lib as a Python
extension (mostly done, but not DB API compliant), or recoding in C++
(should be fast, I've got a working prototype :-)

Btw., are any of your customers running with 3Tb of collections
("someone didn't pay their bill") data, with throughput requirements and
hard deadlines ("if you're not done by noon you must pay us back
$50K"/hr) <wink>? (sorry if I come across too harsh, not my intention,
but after going through the joy of formally proving my program could do
no harm to either the data or the throughput of the server [oh, and
being called at 3am on Sunday], my sense of humor seems to have gone
into hiding -- maybe next week I'll have a beer and laugh about it :-)

-- bjorn

ps: In case you were wondering, we're not doing this on live production
data because we don't know what we're doing...




From roy at panix.com  Tue Apr 15 07:56:21 2003
From: roy at panix.com (Roy Smith)
Date: Tue, 15 Apr 2003 07:56:21 -0400
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org>
Message-ID: <roy-31F2E6.07562115042003@reader1.panix.com>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> wrote:
> If strings were to
> use a preallocation strategy--similar to what lists do--MANY string
> appends could avoid a memory copy on the original string.

I certainly agree that mutable strings would be useful in some 
situations.  For example, in the quadratic run-time program I mentioned 
in a parallel thread, I could easily see doing:

s = ""
s.reserve (len (line))
for word in line.split():
   blah blah blah
   s += stuff

and being very happy that I was now getting linear run time with only a 
trivial change in logic.

The problem is, strings are currently documented as being immutable.  
Making them mutable would break existing code.  I can, for example, find 
the length of a string and use that length someplace later in my 
program, knowing that it hasn't changed.  Not to mention using strings 
as dictionary keys.



From abpillai at lycos.com  Thu Apr 24 09:06:49 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 24 Apr 2003 06:06:49 -0700
Subject: sys.argv problem in Win 2k
References: <bd993a2f.0304240121.7c795cf0@posting.google.com> <mailman.1051177117.30168.python-list@python.org>
Message-ID: <bd993a2f.0304240506.5b206f3d@posting.google.com>

Sure...

#example+
#cmdtest.py
import sys

print sys.argv
for x in sys.argv:
    print x

D:\OpensourceApps>cmdtest.py passing arguments
['D:\\OpensourceApps\\cmdtest.py']
D:\OpensourceApps\cmdtest.py
#example-

 As you see the arguments I pass are not being printed!
This is the case with all my scripts now. They were working
well when I was on a WinNT workstation :-?

Regards

ABP

Gerhard Haering <gh at ghaering.de> wrote in message news:<mailman.1051177117.30168.python-list at python.org>...
> Anand B Pillai wrote:
> > Hi pythonistas
> > 
> >        All my python scripts on Win 2k does not take command line arguments.
> > i.e, len(sys.argv) is always 1. Because of this I cannot use any scripts which
> > takes cmd line arguments. [...]
> 
> #v+
> D:\tmp>copy con: test.py
> import sys
> print sys.argv
> ^Z
>          1 file(s) copied.
> 
> D:\tmp>c:\Python22\python.exe test.py bla blubs
> ['test.py', 'bla', 'blubs']
> 
> D:\tmp>test.py bla blubs
> ['D:\\tmp\\test.py', 'bla', 'blubs']
> 
> D:\tmp>
> #v-
> 
> Works fine here. Can you show exactly what does not work with a test case?
> 
> -- Gerhard



From jjl at pobox.com  Sat Apr  5 07:48:31 2003
From: jjl at pobox.com (John J Lee)
Date: Sat, 5 Apr 2003 13:48:31 +0100 (BST)
Subject: cgi.py -- anybody using it?
In-Reply-To: <1049529516.6022.11.camel@lothlorien>
Message-ID: <Pine.LNX.4.44.0304051328020.365-100000@alice>

On 5 Apr 2003, Ian Bicking wrote:

> On Fri, 2003-04-04 at 17:58, John J Lee wrote:
> > I just tried testing some HTML form upload code using cgi.py, and
> > discovered that it wants lower-case HTTP headers :-/
>
> Have you actually seen this behavior, or are you getting this from
> reading the code?

No, I've seen it, but by using cgi.FieldStorage directly in unit tests
rather than actually sending HTTP requests.


> I don't even think the code handles headers -- the
> CGI process turns headers into environmental variables, and case isn't
> preserved in that process.

Ah.  That'll be it.  That sort of thing did pass through my mind, but for
some reason I didn't follow it up.


> In some places you may read code that uses the mimetools module to parse
> headers.  It has a dictionary interface, but keys are case-insensitive,
> so if it uses m["content-type"] it's still work with Content-Type or
> somesuch.

Well, though I had forgotten about mimetools, presumably that isn't the
issue, assuming it's the CGI process' responsibility to lower-case
everything as you tell me.  Might be an easy way to fix my unit test,
though.


> And yes, the cgi module is in very active use, not just for cgi
> programs, but in many systems that handle the web (since it has
> functions for query string parsing and many other things).

Oh.  There should be a pointer to the cgi docs from the urllib docs, then.
I once wrote a urldecode function to match urllib.urlencode -- hadn't
noticed cgi.parse_qsl.  I'll file a bug about *that*.


John





From xtian at toysinabag.com  Mon Apr  7 18:33:36 2003
From: xtian at toysinabag.com (xtian)
Date: 7 Apr 2003 15:33:36 -0700
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com>
Message-ID: <d111eafa.0304071433.17b1af0c@posting.google.com>

danra at walla.co.il (Danra) wrote in message news:<fd1c664e.0304071056.31a0d377 at posting.google.com>...
> Hi,
> 
> Here's my experience with 'super':
> 
> >>> class b:
> 	pass
> 

Super only works on new-style classes.
Because you don't subclass anything, b is an old-style class. To make
b a new-style class, b should subclass object. You can also create a
new-style class by subclassing a builtin type, like list or str.

> >>> class c(b):
> 	pass
> 
> >>> super(c)
> Traceback (most recent call last):
>   File "<pyshell#6>", line 1, in ?
>     super(c)
> TypeError: super() argument 1 must be type, not classobj

A new-style class is a type (ie, type(newstyleclass) == <type 'type'>)
while an old-style class is not (type(b) == <type 'class'>). The error
message is a bit cryptic.

Hope that clarifies things - once the type-class unification is
complete, this distinction will go away (I think).

cheers,
xtian



From ckea25d02 at sneakemail.com  Tue Apr  8 20:38:05 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 09 Apr 2003 12:38:05 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1049789354.28270.python-list@python.org>
References: <mailman.1049789354.28270.python-list@python.org>
Message-ID: <b6vpfa$9imjc$1@ID-169208.news.dfncis.de>

Christian Tanzer wrote:
> If you want to implement something comparable
> to Emacs' power, you'd better look at the Elisp Reference Manual for
> quite a while.

That's one way of looking at it. Arguably, however, it
might be better *not* to study the Elisp Reference Manual
too closely, so as not to pollute your mind with all
those crufty hacks and jeopardise your chances of
coming up with a better overall design. A sort of
conceptual clean-room approach.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From martin at v.loewis.de  Tue Apr 22 18:17:40 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 23 Apr 2003 00:17:40 +0200
Subject: pygettext but no msgfmt
References: <b83snm$1ms$1@news.tu-darmstadt.de> <b844b0$92e$1@zeppelin.rz.uni-potsdam.de> <b84aum$62g$1@news.tu-darmstadt.de>
Message-ID: <m37k9mtbwb.fsf@mira.informatik.hu-berlin.de>

Mirko Koenig <koenig at v-i-t.de> writes:

> Or can you help me?

Install msgfmt.

If that is still unacceptable, download the Python source code
from www.python.org.

Regards,
Martin



From peter at engcorp.com  Wed Apr  9 08:40:37 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 09 Apr 2003 08:40:37 -0400
Subject: sorting strings written to files
References: <2745895.1049870828@dbforums.com> <b70igq$9n0ff$2@ID-99293.news.dfncis.de> <3E93D198.1030801@mxm.dk>
Message-ID: <3E9414C5.7CCD3BC5@engcorp.com>

Max M wrote:
> 
> This could give problems:
> 
> l.sort()
> print l
>  >>> ['a\ta', 'a\tad', 'a\tb', 'a\tbe', 'a\tc', 'a\tcf', 'aa\ta',
>   'ab\tb', 'ac\tc', 'ad\tad', 'ae\tbe', 'af\tcf']
> 
> You would need to split them on the '\t' first.
> 
> # decorate sort undecorate
> l2 = [i.split('\t')[:2] + [i] for i in l]
> l2.sort()
> sorted = [i[-1] for i in l2]
> print sorted
> 
>  >>> ['a\ta', 'a\tad', 'a\tb', 'a\tbe', 'a\tc', 'a\tcf', 'aa\ta',
> 'ab\tb', 'ac\tc', 'ad\tad', 'ae\tbe', 'af\tcf']

Uh.... is there supposed to be some difference between your two
output lists?  

Given the problem as described, William's solution was quite adequate.

-Peter



From djc at object-craft.com.au  Wed Apr 30 07:48:19 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 30 Apr 2003 21:48:19 +1000
Subject: Web programming and a different "type" problem
In-Reply-To: <1051673492.698.214.camel@lothlorien>
References: <mailman.1051532982.25352.python-list@python.org>
	<UXara.22428$K35.745468@news2.tin.it>
	<vaqvrb3ipqblf2@corp.supernews.com>
	<slrnbarhfb.67t.jon+usenet@snowy.squish.net>
	<b8lj2q$l14$1@newshost.accu.uu.nl>
	<m3fzo1v561.fsf@ferret.object-craft.com.au>
	<mailman.1051654786.4587.python-list@python.org>
	<m365owlu6v.fsf@ferret.object-craft.com.au>
	<1051673492.698.214.camel@lothlorien>
Message-ID: <m3vfwwryt8.fsf@ferret.object-craft.com.au>

>>>>> "Ian" == Ian Bicking <ianb at colorstudy.com> writes:

Ian> On Tue, 2003-04-29 at 19:12, Dave Cole wrote:
>> So what Albatross does is allow (require) you to specify when a
>> field should return a list.  If you build a form which contains
>> multiple inputs with the same name ("radio" is an obvious
>> exception) then it will raise an exception if you do not use the
>> "list" attribute on all of the instances of the field in the form.

Ian> How do you specify that you expect a list?  Does it happen at
Ian> runtime, when you fetch the value from the request with an
Ian> I'm-expecting-a-list method?  Or is there some more declarative
Ian> schema involved?

You tell the template interpreter that you expect a list of values to
be returned for a particular input field.  This information is
included in the form as a hidden field.  When the browser request form
that form is submitted the values from the request are handled
according to the information from the hidden field.

If you do not specify that a field should return a list of values and
your template includes multiple fields with the same name then the
template interpreter will raise an exception.

>> In our experience you can get subtle and annoying bugs in your
>> application as a result of incorrect handling of fields which
>> sometimes return multiple values.

Ian> Indeed, but I've usually found that the bugs were from form
Ian> generation, not from form processing.

That is true, The bug is in the form generation.  Where it causes
problems though is in the processing of requests from that form.
Albatross forces you to declare (in the form) that you expect multiple
values in response to a form.

Ian> It might be better if the error is signaled earlier and in a more
Ian> pleasant way, but IMHO throwing away duplicate values when they
Ian> aren't expected just covers up bugs.

Albatross reports the problem immediately as an exception.  Since it
is could be a programming error to build forms with duplicate input
field names, the exception is the right thing to do.  The way to avoid
the exception is to tell the template interpreter (via the "list"
attribute in the <al-input> tag) that you really did want multiple
values for the field.

Ian> But I don't think that's what you're doing in this case (though
Ian> it sounds like jonpy does this, for instance)...?

I haven't looked at jonpy yet.

>> When the </al-form> tag is processed, the record of fields is
>> encoded into a hidden field called __albform__.  The field value is
>> MD5 signed by combining the encoded value with a server side secret
>> and the sign is combined with the __albform__ value.

Ian> Again out of curiosity -- what exactly goes into the __albform__
Ian> value?  I.e., what information are you saving there?

The important thing to realise here is that Albatross does not force
you to do much of anything in any single way.  If you use an execution
context with inherits from the NameRecorderMixin then you get the
following (from the next release of Albatross - FILE is a new type).

The actual source from Albatross is included with commentary.

class NameRecorderMixin:

    # These are the three different types that we expect an input
    # field to have.  For each field in a form we record the name and
    # type of that field.

    NORMAL, LIST, FILE = range(3)

    def __init__(self):

        # The NameRecorderMixin is used in execution contexts.  This
        # means that this constructor is called when the application
        # object receives a new browser request.  The execution
        # context is destroyed once request processing is complete.

        self.__elem_names = {}

    def form_open(self):

        # This is called by the template interpreter when an <al-form>
        # open tag is processed.  We just reset the field type
        # information.

        self.__elem_names = {}
        self.__need_multipart_enc = 0

    def form_close(self):

        # This is called by the template interpreter when an
        # </al-form> tag is processed, but before the </form> tag is
        # written.  We create an extra hidden field in the form which
        # contains the field type information.

        # First we pickle the field type information.

        text = cPickle.dumps(self.__elem_names, 1)

        # Then we ask the application to sign the pickle.  Currently
        # Albatross provides the PickleSignMixin which MD5 signs a
        # secret and the pickle and returns the MD5 sign prepended to
        # the pickle (without the secret of course).

        text = self.app.pickle_sign(text)

        # If zlib is available we then compress and bas64 encode the
        # signed pickle.

        if have_zlib:
            text = zlib.compress(text)
        text = base64.encodestring(text)

        # Finally the hidden field is written to the output.

        self.write_content('<input type="hidden" name="__albform__" value="')
        self.write_content(text)
        self.write_content('">\n')

        # We reset the form record and tell the template interpreter
        # whether or not any type="file" input fields were included in
        # the form.
        #
        # This is done because the template interpreter generated the
        # content of the <form> tag inside a "content trap".  The
        # execution context maintains a stack of content traps.  All
        # content written between the call to push_content_trap() and
        # pop_content_trap() is returned as a string to the caller of
        # pop_content_trap() as a string.  This allows the <al-form>
        # tag to delay the output of the <form> tag until after all of
        # the form content has been written.  At the end of the form
        # the <al-form> tag automatically includes the
        # enctype="multipart/form-data" attribute if any file input
        # fields were present.

        self.__elem_names = {}
        return self.__need_multipart_enc

    def input_add(self, itype, name, unused_value = None, return_list = 0):

        # This is called by the template interpreter for each
        # <al-input> tag encountered.  The type of each field is
        # recorded and checked for consistency.

        if itype == 'file':
            self.__need_multipart_enc = 1
            self.__elem_names[name] = self.FILE
        elif self.__elem_names.has_key(name):
            prev_multiples = self.__elem_names[name]
            implicit_multi = itype in ('radio', 'submit', 'image')
            if not return_list and not implicit_multi:
                raise ApplicationError(
                    'al-input "%s" not defined as "list"' % name)
            if return_list and implicit_multi:
                raise ApplicationError(
                    'al-input "%s" should not be defined as "list"' % name)
            if prev_multiples == 0 and return_list != 0:
                raise ApplicationError(
                    'al-input "%s" initially not defined as "list"' % name)
            if return_list == 0 and prev_multiples != 0:
                raise ApplicationError(
                    'al-input "%s" initially defined as "list"' % name)
        else:
            if return_list:
                self.__elem_names[name] = self.LIST
            else:
                self.__elem_names[name] = self.NORMAL

    def merge_request(self):

        # This is called when the application wants to merge the
        # browser request into the execution context namespace.  The
        # Request object has already been attached to the execution
        # context in the request member.
        #
        # Try to get the hidden field out of the request.  If it is
        # not present then assume the request was GET rather than
        # POST.  Fall back to default behaviour and just merge all
        # request values into namespace.

        if not self.request.has_field('__albform__'):
            for name in self.request.field_names():
                value = self.request.field_value(name)
                self.set_value(name, value)
            return

        # Found hidden field - reverse the decoding process.

        text = self.request.field_value('__albform__')
        text = base64.decodestring(text)
        if have_zlib:
            text = zlib.decompress(text)

        # Ask the application to check and remove the sign on the
        # pickle.  If the sign does not match we get back an empty
        # string.  At this point we ignore the request (refuse to
        # merge it).

        text = self.app.pickle_unsign(text)
        if not text:
            return

        # All is good - merge the raw browser request into the
        # execution context according to the type information we
        # recorded.

        elem_names = cPickle.loads(text)
        for name, mode in elem_names.items():
            if mode == self.FILE:
                value = self.request.field_file(name)
            elif self.request.has_field(name):
                value = self.request.field_value(name)
            else:
                x_name = '%s.x' % name
                y_name = '%s.y' % name
                if self.request.has_field(x_name) \
                   and self.request.has_field(y_name):
                    value = Point(int(self.request.field_value(x_name)),
                                  int(self.request.field_value(y_name)))
                else:
                    value = None
            if mode == self.LIST:
                if not value:
                    value = []
                elif type(value) is not type([]):
                    value = [value]
            self.set_value(name, value)

Having explained the above, it is not set in stone.  You can implement
your own scheme and have your execution context class inherit from
that mixin class rather than the NameRecorderMixin.

There is an alternate mixin provided called the StubRecorderMixin.  It
is shown in full below.

class StubRecorderMixin:
    def form_open(self):
        pass

    def form_close(self):
        pass

    def input_add(self, itype, name, value = None, return_list = 0):
        pass

    def merge_request(self):
        for name in self.request.field_names():
            value = self.request.field_value(name)
            self.set_value(name, value)

We do have plans for other recording schemes.

- Dave

-- 
http://www.object-craft.com.au





From exarkun at intarweb.us  Tue Apr  8 15:48:29 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Tue, 8 Apr 2003 15:48:29 -0400
Subject: Tiobe index algorithm now published
In-Reply-To: <b6v8id$qli$1@reader1.tiscali.nl>
References: <b6v8id$qli$1@reader1.tiscali.nl>
Message-ID: <20030408194829.GA4589@meson.dyndns.org>

On Tue, Apr 08, 2003 at 09:37:31PM +0200, GerritM wrote:
> We have had discussions about the "TIOBE Programming Community Index" at
> http://www.tiobe.com/tpci.htm
> One of our consistent critics was that the algorithm was not published nor
> verifiable. Tiobe has now added the (simple) algorithm on this page. They
> also explain which "programming languages" are monitored.

  Ooh, a Google search.  How scientific. ;)

  Jp

-- 
Somewhere, something incredible is waiting to be known.
                -- Carl Sagan
-- 
 up 19 days, 15:01, 4 users, load average: 0.10, 0.12, 0.23




From aahz at pythoncraft.com  Sun Apr 20 17:34:20 2003
From: aahz at pythoncraft.com (Aahz)
Date: 20 Apr 2003 17:34:20 -0400
Subject: Good Python generators example?
References: <paDoa.1611$hT2.1228371@news2.news.adelphia.net>
Message-ID: <b7v3os$6k0$1@panix3.panix.com>

In article <paDoa.1611$hT2.1228371 at news2.news.adelphia.net>,
Robert Oschler <no_replies at fake_email_address.invalid> wrote:
>
>Can someone point me to a nice concise Python 2.2 compatible example of the
>use of generators?

ARK's example was a good example of the algorithmic use of generators;
here's a short example of what I'd call workday use of generators:

import re
def grep(regex, seq):
    '''
    regex can be either string or compiled regex; ``re`` doesn't care.  
    seq can be an iterator -- there's just no good abbreviation other
    than iter()
    '''

    regex = re.compile(regex)
    for line in seq:
        if regex.search(line):
            yield line

The nice thing about this is that it pipelines *very* well with other
generators.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From aleax at aleax.it  Thu Apr 17 12:49:58 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 16:49:58 GMT
Subject: [Que]Do any SGML lib good for Python?
References: <2778697.1050595278@dbforums.com>
Message-ID: <WWAna.28468$LB6.652269@news1.tin.it>

dolephi9080 wrote:

> 
> I search  the SGML Lib for python, but it's hard to find. even the
> python contain a sgml parser itself. But that's a limite parser.
> Does some one know it? pls tell me. Thx

Perhaps:

http://www.garshol.priv.no/download/software/pysp/

???  Doesn't seem to be very actively maintained, but perhaps
it doesn't need to be... I wouldn't know, it's ages since I last 
had to parse very general SGML input.  Worst case, you could
re-wrap James Clark's SP, e.g. with Boost Python...


Alex




From martin at v.loewis.de  Mon Apr 14 16:07:20 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 14 Apr 2003 22:07:20 +0200
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <3e9af967$0$49112$e4fe514c@news.xs4all.nl> <ptnosvmm.fsf@morpheus.demon.co.uk>
Message-ID: <m3d6jovo5j.fsf@mira.informatik.hu-berlin.de>

Paul Moore <gustav at morpheus.demon.co.uk> writes:

> PS Is there any way of forcing the Windows 2000/XP console to display
>    Unicode in any form (UTF-8, maybe) rather than having to deal with
>    code pages???

Yes, through WriteConsoleW.

Regards,
Martin



From a.schmolck at gmx.net  Sun Apr 13 19:22:18 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 14 Apr 2003 00:22:18 +0100
Subject: isearch, LEAP and patents  was: Re: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org> <yfsr88fkppr.fsf@black132.ex.ac.uk> <eaeka.1333$1h6.8933@news-server.bigpond.net.au> <yfs1y0el9xc.fsf_-_@black132.ex.ac.uk> <e0caaad7.0304100951.57cb4572@posting.google.com>
Message-ID: <yfsadeudlud.fsf@black132.ex.ac.uk>

usenet at rob.dqd.com (Rob Mayoff) writes:

> The LEAP patent (US patent 5,019,806) does not claim general
> incremental search.  
> It claims an incremental search mechanism where one key enables a forward
> search mode while the key is pressed, and a second key enables a backward
> search mode while the second key is pressed. In other words, it claims
> quasi-modal incremental search. Emacs doesn't have that.

Have you got a publically available reference at hand by any chance? I'd be
really curious to learn how the patent claim is justified.


'as



From cben at techunix.technion.ac.il  Thu Apr 17 18:12:41 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Fri, 18 Apr 2003 01:12:41 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <b7i2jt$170ph$1@ID-169208.news.dfncis.de>
Message-ID: <20030417221241.0346222093@techunix.technion.ac.il>

In comp.lang.python, you wrote:
>Robin Munn wrote:
>> Paul Foley <see at below.invalid> wrote:
>>>
>>>What if another thread (B) does the same thing, and changes sys.stdout
>>>while this thread (A) is still inside the "try" block?  And if thread
>>>A runs its "finally" clause and resets sys.stdout while thread B is
>>>still in its "try" block, they screw each other.
>> 
>> This is the classic problem that arises when two threads need to use any
>> shared resource, isn't it? 
>
>No, actually it's the classic problem that arises when
>two threads spuriously share something (a global variable)
>that shouldn't have been shared in the first place.
>
True.  Blocking here on locks is completely unnecessary.

>This is *not* an argument for dynamic scoping. It's an
>argument for explicitly passing things the state that
>they need!
>
No, this is just an argument for *passing* things the state that they need.
Whether it is done explicitly or implicitly is quite irrelevant to the obvious
requirement that things should work accross threads.  It's still an
interesting question, or we wouldn't write so much <wink> but it doesn't
follow directly.

Of course, currently Python allows to fulfil this only by passing things
explicitly.  So there is no choice.  But there could be.

I still don't see why is dynamic binding inherently buggy.  I'm talking about
dynamic binding done right, so that a new thread gets a copy of all dynamic
variables from the parent thread; afterwards both threads can alter them
independently.  (This sounds like a very uneffecient implementation; it can be
optimized in various ways to copy only those variables that are needed).

Note that the semantics is perfectly correct - every called function gets all
dynamic variables as implicit parameters, just like you would get with
explicit parameters.  The only difference is whether you have to spell this
out in your code.  So the question is just: which leads to less programmer
mistakes?

Truly, some confusion can result when one controls the same thing both with a
dynamicaly bound implicit parameter and an explicit parameter and when you
don't know which parameter is controlled in which way.  However, it is easy to
use it cleanly.  If you decide once that rebinding sys.stdout is the *right*
way to pass an output parameter, and use this consistently in your program,
there is nothing evil to it!  These two approaches are completely equivallent:

def foo():
    bar()
    with output to file("f2"): baz()   # neutrally bogus syntax :-)
    
def foo(f1):
    bar(f1)
    baz(file("f2"))

If inheriting the parameter (like in the `bar` call) is the typical situation,
the dynamic approach seems better; if overriding it (like in the `baz` call)
is mroe typical, explicit passing seems better.  The main criteria for
deciding on each parameter how it should be passed are:

- Whether you work with only one value of it most of the time.
- Whether one wants it by default to be inherited by called functions.

For output, the answers are usually yes, so UNIX's designers decided it should
be passed implicitly - and got it right IMHO, at least at the process level. 
Perhaps inside a single program the balance is different but I think the
implicit passing would be nice inside a program too.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From no_spam_artyprog at wanadoo.fr  Sat Apr 12 10:33:13 2003
From: no_spam_artyprog at wanadoo.fr (Salvatore)
Date: Sat, 12 Apr 2003 16:33:13 +0200
Subject: distutils module
Message-ID: <b79863$dil$1@news-reader10.wanadoo.fr>

Hello,

I've just installed Debian Knoppix,and i was surprised
to see that distutils module could not be found.

Does anybody experienced the same problem ?

Regards

Salvatore




From khowe at perfnet.ca  Thu Apr 24 16:52:35 2003
From: khowe at perfnet.ca (Kevin Howe)
Date: Thu, 24 Apr 2003 16:52:35 -0400
Subject: php equivalents?
References: <OHSpa.8996$h%2.915957@read1.cgocable.net> <b890vo$rpj$1@news1.wdf.sap-ag.de>
Message-ID: <e9Ypa.3964$b67.980821@read2.cgocable.net>

Thanks for the suggestions Daniel. I've since discovered an equivalent to
PHP's register_shutdown_function() as follows:

def myFunc ():
  """ perform cleanup just before the script exits """

import atexit
atexit.register(myFunc)

The var_export equivalent however, is much more tricky. I'm going to post a
more specific thread concerning it's difficulties.

- Kevin






From benoit.besse at club-internet.fr  Wed Apr 16 19:34:49 2003
From: benoit.besse at club-internet.fr (Benoit BESSE)
Date: Thu, 17 Apr 2003 01:34:49 +0200
Subject: Réception et décodage d'une trame UDP
References: <3e9c9b1e$0$1509$7a628cd7@news.club-internet.fr><3E9CA7C2.2C9730EA@engcorp.com> <mailman.1050456911.19513.python-list@python.org>
Message-ID: <3e9de899$0$1516$7a628cd7@news.club-internet.fr>

Ok for the language and still all my excuses.
It's exactly what I want to do.
To decode UDP and to insert it in a base and the whole in Python.

"Gerhard H?ring" <gh at ghaering.de> a ?crit dans le message de news:
mailman.1050456911.19513.python-list at python.org...
* Peter Hansen <peter at engcorp.com> [2003-04-15 20:45 -0400]:
> Benoit BESSE wrote:
> >
> > Bonsoir tout le monde,
> >
> > qui a d?j? ?crit un module qui ?coute, re?oit, d?code et insert dans une
BD
> > une trame UDP.
> > Je voudrais d?coder la trame UDP ainsi et l'ins?rer dans une table :
> >     - Port Source
> >     - Port Destination
> >     - Longueur
> >     - Somme de contr?le
> >     - Donn?es
> >
> > Merci par avance.
>
> Hmm... almost looks like another "Twisted has that" but I'm
> not sure what "BD" is so I can't be sure...  ;-)

BD = base donn?es = database.

Of course, twisted has that (twisted.enterprise.adbapi,
twisted.enterprise.row) ;-)

The original poster might also want to have a look at a packet sniffer
like tcpdump - looks like tcpdump can be talked to do pretty much all of
that, except the database interface.

Btw. the usual language on this group is English, but there *is* a
french Python newsgroup: fr.comp.lang.python.

Gerhard
--
mail:   gh at ghaering.de
web:    http://ghaering.de/





From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr  3 12:29:58 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 3 Apr 2003 17:29:58 GMT
Subject: PEP 313 - Roman numeral literals
References: <mailman.1049296805.12278.python-list@python.org> <b6fvj9$518n3$1@ID-169208.news.dfncis.de>
Message-ID: <slrnb8orsl.174.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <b6fvj9$518n3$1 at ID-169208.news.dfncis.de>,
Greg Ewing (using news.cis.dfn.de) wrote:
> sismex01 at hebmex.com wrote:
>> No, the correct writing of 39 in roman numerals is "XXXIX",
>> the proposed rule for only 3 of X, I or C in a single number
>> is premature, or incomplete.
>> 
>> It should read "3 consecutive X, I or C digits".
> 
> You shouldn't need to explicitly state any such rule
> if the rest of the rules are formulated properly.
> 
> The best way would be to specify a grammar in some
> form, e.g. a regular expression.

Right, like this:

   M{0,3}(D?C{0,3}|C[DM])(L?X{0,3}|X[LC])(V?I{0,3}|I[VX])

But since the use of subtraction was apparently a late addition
to the system, we should allow the purists to do without it:

   M{0,3}((D?C{0,3}|C[DM])(L?X{0,3}|X[LC])(V?I{0,3}|I[VX])|D?C{0,4}L?X{0,4}V?I{0,4})

The problem of confusion with existing identifiers can be dealt
with once Python supports Unicode source files, when we can use
this simple expression:

   \u216F{0,3}((\u216E?\u216D{0,3}|\u216D[\u216E\u216F])(\u216C?\u2169{0,3}|\u2169[\u216C\u216D])(\u2164?\u2160{0,3}|\u2160[\u2164\u2169])|\u216E?\u216D{0,4}\u216C?\u2169{0,4}\u2164?(\u2160{0,4})

Of course we would also need to deal with overbars for 5000,
10000, and so on.



From hwlgw at hotmail.com  Sun Apr 13 06:24:27 2003
From: hwlgw at hotmail.com (Will Stuyvesant)
Date: 13 Apr 2003 03:24:27 -0700
Subject: win32 COM test question
References: <mailman.1050209239.15961.python-list@python.org>
Message-ID: <cb035744.0304130224.6323b4a0@posting.google.com>

[Roman Yakovenko]
> http://www.hps1.demon.co.uk/users/andy/pyvb/spamsrv.htm
> 
> And next time try to search for the answer with google ( python com
> visual basic )

That URL leads to a nice "tutorial" and "idiots guide" about using COM
and Python together in the real world, with **examples**.  It is
great, more of that!  And I do not mean the COM subject, I mean the
"tutorials" and the "examples".  With many opensource projects, it is
hard to find people to write good tutorial and examples, and Python is
not an exception.  I am always very pleased to see stuff like that. 
Unfortunately my english is not as good as I would like so writing
them myself is um, ..., well I did some.  Hrm.

As a side note, quoted from above URL:

"Distribution is a problem - not everyone will want to know they have
python installed, and a on-step installation that puts it there
suitably hidden will be needed in time"

...!!!!  This has also been my experience.  But it is very funny to
see somebody else with it too.  This is also a serious matter.  Maybe
it is a good idea for the technical people in the python-dev core to
think again about integrating a C compiler with python?  It has been
my understanding, from reading python-dev and c.l.py occasionally,
that a real Python compiler is not feasible, but maybe packaging
Python with the native C compiler for the platform you choose to
download Python for?  So the "users" will only see the .exe file you
produce on Windows, and they do not have to have Python installed. 
And *after* they have been using your applications for a while you can
show them the light, unveiling that you did most of it with Python. 
That is the only way I have ever convinced people of taking Python
seriously.  And still, then their reaction was like "Oh my god, did we
use open source software?  What about *security*!  All those weird
hackers out there!  These geeky people with their stinking clothes! 
We want nothing to do with those!"

You know managers...

-- 
The light at the end of the tunnel may be an oncoming dragon.



From peter at engcorp.com  Wed Apr 16 12:55:55 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 12:55:55 -0400
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com> <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk>
Message-ID: <3E9D8B1B.24050D34@engcorp.com>

Michael Hudson wrote:
> 
> Peter Hansen <peter at engcorp.com> writes:
> > The last time I used Google (of which I'm well aware of course), it
> > told me that Knuth was in fact merely paraphrasing someone else
> > (google can tell you who it was, if you wish ;-),
> 
> Yes, you can find lots of people claiming it was Hoare that said this.
> However, their plurality doesn't necessarily make them right.

True... but neither does the fact that Knuth (also?) said it make
them wrong.  On the other hand, this is the first time I've seen
a claim, found via google or not :-), that Hoare did not in fact 
say that.  Is your view authoritative, or a supposition?

-Peter



From borco at go.ro  Mon Apr 28 14:47:24 2003
From: borco at go.ro (Ionutz Borcoman)
Date: 28 Apr 2003 11:47:24 -0700
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <80dac0f4.0304281047.9143cd0@posting.google.com>

Hi, 

Why don't you try "How to Think Like a Computer Scientist: Learning with Python"

http://www.ibiblio.org/obp/thinkCSpy/

It's free and it's easy reading for people with no programming experience.

Ionutz



From bbfoto2002 at netscape.net  Wed Apr  2 10:15:16 2003
From: bbfoto2002 at netscape.net (BBFoto)
Date: Wed, 02 Apr 2003 15:15:16 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org>
Message-ID: <3e8afeb9.1125578@news-server>

execfile('c:/program files/pysol/pysol-4.80/data/PySol_22.pyc')

invalid syntax 

???

Thanks

BB



On Wed, 2 Apr 2003 13:02:52 +0400, "Raseliarison nirinA"
<nirina at mail.blueline.mg> wrote:

>> >"BBFoto" <bbfoto2002 at netscape.net> wrote:
>> >
>> >> I run the Python executable and type in "PySol" at the prompt on the
>> >> black screen and get "name not defined."
>> >
>> >What happens if you type:
>> >import PySol
>
>> "module not found"
>
>I suppose 'PySol.py' is the full name of your game and it is located at
>'c:/your/game/directory', so you can try:
>
>import os
>os.chdir('c:/your/game/directory')
>execfile('PySol.py')
>
>or in one line:
>
>execfile('c:/your/game/directory/PySol.py')
>
>nirinA
>
>
>
>
>




From python at dylanreinhardt.com  Wed Apr  2 17:39:25 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 02 Apr 2003 14:39:25 -0800
Subject: Zope style (was: tuning our site (was: help! advocacy
	resources needed fast))
In-Reply-To: <v8mjffpuph1oc2@corp.supernews.com>
References: <mailman.1046819836.25704.python-list@python.org>
	 <mailman.1047414442.16521.python-list@python.org>
	 <v81d26sb3vf170@corp.supernews.com> <irdbl-1ds.ln1@news.lairds.org>
	 <v8mjffpuph1oc2@corp.supernews.com>
Message-ID: <1049323165.6723.111.camel@ida.dylanreinhardt.com>

On Wed, 2003-04-02 at 12:54, Cameron Laird wrote: 
> I find
> it really interesting when one technology is unequivocally su-
> perior to another.  I think we have that with DTML vs. Python
> scripts.

[ jumping in mid-stream ] 

DTML and Python Scripts are different tools, each best-suited to a
particular thing.  Determining which is better obscures a more important
point: they're each good tools for a particular set of problems.

DTML's problem space is quite a bit smaller, but that doesn't make it a
bad tool... just one that should be applied more selectively than it
typically is. :-)

> There are a couple of pertinent facts.  First, much of the
> passion "Web designers" are supposed to have for "templates"
> comes from the supposition that they want stuff to look like 
> HTML, so they can use their "HTML editors".  Is *anyone* doing
> that with Zope, though?  

It's not really about the visual editors.  DTML is largely incompatible
with HTML editors... or so I hear. :-)

The purpose is more basic than that: one is a presentation tool, one is
a logic tool.

Templating is appropriate when 80-90% of what you want to do is simply
write out text.  With templating, there is no need to *say* that text
should be written out... everything you type is returned unless you
specify otherwise.

When you're performing logic, the reverse is true... most of what you
type *produces* your content and nothing is returned aside from that
which is *explicitly* returned.

Templating provides the least cluttered, most concise way of expressing
*integrating* static and dynamic content.  Python represents the most
concise way of expressing logical operations.  Each problem will consist
of some part logic and some part presentation.  Most of my Zope projects
use a blend of templating and scripting.

> Finally, while your dtml-let is an effective style, 

I almost never use dtml-let.  In my book, it's one of the first signs
that you're trying to do something DTML isn't well-suited for.  Another
is use of nested dtml-ifs more than a couple layers deep.

> projects always, always, 
> always seem to need more parametrization; managers say, "I want
> it just like that, except it needs to {apply to a different
> continent,show the current contents of the order,show the 
> expanded display when the user has selected that, ...}."

If that requirement is thought of as parameterization, DTML is a less
worthy tool than Python scripts, no doubt.  But if that same set of
requirements is thought of as "skinning" DTML is a great tool.

DTML's usefulness depends, to some degree, on your use of Acquisition. 
Your overall site design strategy might have a lot to do with why one
seems more powerful than the other for any given task.

Dylan





From abpillai at lycos.com  Sat Apr 26 07:24:02 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 26 Apr 2003 04:24:02 -0700
Subject: sys.argv problem in Win 2k
References: <bd993a2f.0304240121.7c795cf0@posting.google.com> <mailman.1051177117.30168.python-list@python.org> <bd993a2f.0304240506.5b206f3d@posting.google.com> <Xns936790B52180Eduncanrcpcouk@127.0.0.1> <bd993a2f.0304242310.586305df@posting.google.com> <Xns9368583183AC8duncanrcpcouk@127.0.0.1>
Message-ID: <bd993a2f.0304260324.4e9a4f4e@posting.google.com>

Well I was hoping that would fix it!
I tried

Python.File=D:\Python22\python.exe "%1" %*
and
Python.File=D:\Python22\python.exe "%*"
and another 2 combinations,

But still I cannot pass the cmd line arguments!
Let me try and figure out myself :-O

ABP
Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in message news:<Xns9368583183AC8duncanrcpcouk at 127.0.0.1>...
> I wrote:
> >> D:\temp>ftype Python.File
> >> Python.File=D:\Python22\python.exe "%1" %*
> 
> abpillai at lycos.com (Anand B Pillai) wrote in
> news:bd993a2f.0304242310.586305df at posting.google.com: 
> 
> > F:\users\anand\LocalWorkspaces\ResModWS>ftype Python.File
> > Python.File=D:\Python22\python.exe "%1"
> > 
> 
> Spot the difference.
> 
> The "%1" passes the name of the script to python.exe, the %* passes all the 
> other arguments. Without the %* windows will simply drop the other 
> arguments on the floor.



From jepler at unpythonic.net  Thu Apr 17 08:41:30 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Thu, 17 Apr 2003 07:41:30 -0500
Subject: "pythonic"
In-Reply-To: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
Message-ID: <20030417124128.GC20849@unpythonic.net>

On Thu, Apr 17, 2003 at 08:00:27AM +0000, Afanasiy wrote:
> Too many times now I have seen the term "pythonic" used to vaguely
> describe something which is never explicitly defined, as if it cannot be.

I've given up worrying about that (see my e-mail address)

Jeff




From dgallion1 at yahoo.com  Sat Apr 19 15:43:07 2003
From: dgallion1 at yahoo.com (darrell)
Date: Sat, 19 Apr 2003 19:43:07 GMT
Subject: Parent references: hot or not?
References: <mailman.1050762692.5175.python-list@python.org>
Message-ID: <fFhoa.6843$_N.1979@twister.nyroc.rr.com>

Check the weakref module.
The child can keep a weakref to the parent. This won't set up a cycle.

--Darrell


Dylan Reinhardt wrote:

> Hi all,
> 
> I'm working on a project where I have multiple instances of a container
> class that each hold a large number of another class.
> 
> I need the "contained" object to be able to make use of methods and
> attributes in its container. Inheritance won't do the trick since the
> things I need to refer to will be pretty dynamic.
> 
> A very minimal example of what I'm doing:
> 




From mis6 at pitt.edu  Wed Apr 30 08:44:24 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 30 Apr 2003 05:44:24 -0700
Subject: Problem subclassing tuple
References: <F7DB8D13DB61D511B6FF00B0D0F062330489D915@mail.hebmex.com> <mailman.1051645702.13393.python-list@python.org>
Message-ID: <2259b0e2.0304300444.3494163d@posting.google.com>

Skip Montanaro <skip at pobox.com> wrote in message news:<mailman.1051645702.13393.python-list at python.org>...
> >> You might want to make it "cooperative", that is, instead of using
>     >> "tuple.__new__", use "super(Holder,cls).__new__"
>  
>     >> This delegates the search for the superclass to the function super();
>     >> and that gives you the ability to have multiple inheritance, since
>     >> super() will "do the right thing" when getting your current class'
>     >> superclass; so every class which defines a __new__ method has it
>     >> called.
> 
> I don't believe it matters here.  As I recall, builtins don't support
> multiple inheritance very well (if at all).  The most you can do is add
> mixin classes.
> 
> Skip

Indeed. According to Guido (see
http://www.python.org/2.2.2/descrintro.html#subclassing):

 You can use multiple inheritance, but you can't multiply inherit from
different built-in types (for example, you can't create a type that 
inherits from both the built-indict and list types). This is a
permanent restriction; it would require too many changes to Python's
object
implementation to lift it. However, you can create mix-in classes by 
inheriting from "object".

Here is an exapmple for the OP's edification:

class Holder(tuple):
    def __new__(cls, *args, **kargs):
        return super(Holder,cls).__new__(cls, args)

class Pretty(object):
    def __str__(self):
        return '*** %s ***' % super(Pretty,self).__str__()


class PrettyHolder(Pretty,Holder): pass

t=PrettyHolder(1,2)

print t  #=> *** (1, 2) ***



From dan at isvara.net  Tue Apr  8 11:39:00 2003
From: dan at isvara.net (Dan Ellis)
Date: 8 Apr 2003 08:39:00 -0700
Subject: Stackless example
Message-ID: <a3a44f02.0304080739.252d4c8@posting.google.com>

Can someone provide -- or point to -- a simple example of using
tasklets and channels in Stackless Python?

Thanks,
Dan.



From dextech at terrigal.net.au  Fri Apr  4 23:40:40 2003
From: dextech at terrigal.net.au (Paul Dexter)
Date: Sat, 5 Apr 2003 14:40:40 +1000
Subject: Python commands
References: <3e8bb9c5$1_1@news.iprimus.com.au> <Xns9352D1FE6C1E0cliechtigmxnet@62.2.16.82>
Message-ID: <3e8e5d85$1_1@news.iprimus.com.au>

    Hi all,
                Thanks for all the advice and links to the libraries.
    I sometimes build hardware to connect to the computer via the serial or
parrallel ports, so a peripheral driver is not usually available.
    I now have a lot of reading to do, to get Python to transfer data thro'
the port to my device.
    Things aren't as straight forward as the DOS days with all the
multitasking multithreading features we seem to demand.         LOL

    cheers,
                Paul






From imbosol-1049688353 at aerojockey.com  Mon Apr  7 01:23:08 2003
From: imbosol-1049688353 at aerojockey.com (Carl Banks)
Date: Mon, 07 Apr 2003 05:23:08 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net> <b6qs04$7eqi1$1@ID-169208.news.dfncis.de>
Message-ID: <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>

"Greg Ewing (using news.cis.dfn.de)" wrote:
> Carl Banks wrote:
>> As an example, take the section on buffers.  First function is
>> "current-buffer".  
>> Next three functions are "save-current-buffer", "with-current-buffer",
>> and "with-temp-buffer".  These are actually not functions, but special
>> forms and macros.
> 
> In a truly Pythonic design, I think I would do away with
> the notion of a current-buffer as a dynamically scoped
> global variable. Instead, the buffer to be operated on
> would be passed as a parameter to anything which does
> things to a buffer (or they would be methods of Buffer).
> 
> You may need to have a current-buffer and associated
> functions for elisp compatibility, but they should be
> implemented on top of a more Pythonic layer.


Three objections:

First, current-buffer is a function, not a dynamically-scoped
variable.

Second, given that the keyboard focus can only be active in one buffer
at a time, the idea of a current buffer is absolutely necessary.  You
could argue that you need a current window, not a current buffer, but
you need a current something.  When you type "a", some function is
going to have to call "current-buffer".

Third, there are a *ton* of functions in Emacs that operate on a
certain point in a certain buffer (delete-char, kill-region, etc.).
These kinds of functions typically get bound to keystrokes and other
sorts of commands, which a user expects to work in the current buffer.
If all these functions had to be passed the current buffer as an
argument, it would just complicate things.

The user-interaction functions can, in turn, call general functions
that operate on a given buffer.

The idea of a "Pythonic" layer is a good one, but let's not get rid of
uncomplicated functions like delete-char just because it violates some
notion of what's Pythonic.



-- 
CARL BANKS



From hal_wine at yahoo.com  Mon Apr  7 16:05:50 2003
From: hal_wine at yahoo.com (Hal Wine)
Date: Mon, 7 Apr 2003 13:05:50 -0700 (PDT)
Subject: win98 results (was Re: Cygwin Python regression testers wanted)
In-Reply-To: <20030407135551.GG2536@tishler.net>
Message-ID: <20030407200550.42132.qmail@web14103.mail.yahoo.com>

plain old 98!
--- Jason Tishler <jason at tishler.net> wrote:
> Hal,
> 
> On Fri, Apr 04, 2003 at 04:08:35PM -0800, Hal Wine
> wrote:
> > Jason Tishler wrote:
> > >3. post any failures and/or hangs other than
> test_commands
> > 
> > Final command:
> [snip]
> 
> Is the above for 98 or 98SE?
> 
> Thanks,
> Jason
> 
> -- 
> PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc
> or key servers
> Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD
> E4CC ECF4 8EF6


__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com




From mcclaine at comcast.net  Thu Apr 17 01:00:00 2003
From: mcclaine at comcast.net (Em)
Date: Thu, 17 Apr 2003 01:00:00 -0400
Subject: Python in HTML files
In-Reply-To: <mailman.1050550654.3607.python-list@python.org>
References: <mailman.1050550654.3607.python-list@python.org>
Message-ID: <_2ednU2VIv5QqQOjXTWcqw@comcast.com>

I'm trying to get Python commands to run from within HTML files (a la 
PHP) or at least simulate that behavior.

 From what I've found so far it looks like I need to have a lot more 
permissions than I do. Basically I have access to a /public_html 
directory with a CGI-BIN and that's it. I can't copy any files outside 
of my /public_html directory.

So my question is, do you know of any technologies like this that are 
not Servlets and do not copy files outside a locked-down user's profile?




From simon at klickit.com  Fri Apr 18 14:10:22 2003
From: simon at klickit.com (Simon Withers)
Date: 18 Apr 2003 11:10:22 -0700
Subject: Python COM Object
Message-ID: <78088784.0304181010.91e3641@posting.google.com>

I am experimenting with the python com server found in programming
python second edition (comserver.py) - it looks to be essentially the
same as all the other python com server examples that I have seen.

When I use the com server from within a python script ie:
    from win32com.client import Dispatch
    server = Dispatch('PythonServers.MyServer')

it works fine, however when I use the com server from somewhere else
say within jscript ie:
    var server = new ActiveXObject('PythonServers.MyServer');
I get an 'Automation server can't create object' error.

I have registered the server using:
    from win32com.server.register import UseCommandLine
    UseCommandLine(pyclass)

and a search of my registry finds what look to be appropriate values.

I am at a loss as to why I cannot use the python com server in
non-python contexts.

Simon



From jtiller at sjm.com  Fri Apr 11 20:27:35 2003
From: jtiller at sjm.com (Tiller, Jason)
Date: Fri, 11 Apr 2003 17:27:35 -0700
Subject: split encloser
Message-ID: <ADBFFED9CB40D5118C5A0008C7864BA10185AF25@USSVML03>

Hi, "Aussie", :)

Now, to be totally honest, I kicked myself after Robin pointed out that you
can get a CSV add-on module for Python.  DOH!  ::sigh::  Of course something
like this is available - this is hardly the first-ever application of CSV
(or TSV or whatever)...

I apologize if the commented regex wasn't detailed enough to completely
explain its function.  If you have any  questions, don't hesitate to ask.  I
can tell that the base regex I described in my previous message doesn't meet
all the possible situations that a more sophisticated CSV module might.

If you're just interested in understanding regex'es for potential
application in your own scripts, the regex chapter in "Programming Perl" is
/very/ good, IMHO.  There's always Jeffrey Freidl's "Mastering Regular
Expressions" - all of the solid concepts are there, regardless of language.
I apologize that I can't really offer insight into Python-based regex
processing - I just haven't studied it well, and the Programming Python
(1.3) book that I have devotes only five or so pages to the entire topic
("Programming Perl"'s regex chapter is 70 pages, and "Mastering..." is 308
pages).  I would hope that the more modern "Programming Python" does a
better job...

However, for solving your particular problem, I encourage you to take a look
at the CSV module that Robin mentioned in his post.

Good luck!

---Jason

P.S. - I found documentation for the 'findall' method inside 'pre': "pydoc
pre" from the prompt displayed it.   The 'pre' module looks like Emacs ugly
"infinite backslash" notation (\\\\ just for a literal backslash!).
However, the 'sre' module uses the single-backslash notation, which is a lot
simpler.  I found documentation on python.org here:
http://www.python.org/doc/current/lib/node99.html that discusses the
'finditer' method.  Funny that I couldn't find it via PyDoc...

-----Original Message-----
From: aussie2010 at yahoo.com [mailto:aussie2010 at yahoo.com] 
Sent: Thursday, April 10, 2003 5:54 PM
To: python-list at python.org
Subject: Re: split encloser


Jason Tiller <jtiller at sjm.com> wrote in message
news:<mailman.1049417184.1536.python-list at python.org>...

> 
> Good luck!  I hope this pattern at least gives you a starting point
> for implementing the split() in Python.
> 
> ---Jason
> Sonos Handbell Ensemble

Jason, thanks a lot for the example and the explanation. I fired up
Perl and it worked like a charm. Now to try and understand it!  I
looked at stuff about Python regular expressions and there is
something called finditer and findall that might be useful for parsing
a string, but can't seem to find any documentation on them.
-- 
http://mail.python.org/mailman/listinfo/python-list


***************************** 
This communication may contain information that is proprietary, privileged,
confidential or legally exempt from disclosure.  If you are not a named
addressee, you are notified that you are not authorized to read, print,
retain, copy or disseminate this communication without the consent of the
sender and that doing so may be unlawful. If you have received this
communication in error, please notify the sender via return e-mail and
delete it from your computer. Thank you. St. Jude Medical, Inc. 
*****************************




From tim.one at comcast.net  Mon Apr 14 17:11:46 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 14 Apr 2003 17:11:46 -0400
Subject: Complexity of standard Python data structures
In-Reply-To: <slrnb9m2qj.9r2.marcus@infa.abo.fi>
Message-ID: <BIEJKCLHCIOIHAGOKOLHGEHHFFAA.tim.one@comcast.net>

[Tim]
> Note that it's not true that Python's list is really a vector/array --
> thelanguage doesn't define its implementation, and doesn't want to.

[Marcus Alanen]
> Hmm.  So the Python language is more about specifying _what_ will
> happen, instead of _how_, with no concern for actual speed or memory?

The Python language reference manual, and the Python standard library
reference manual, are like that, and that's par for the course for language
specs:  they specify syntax ("how is it spelled?") and semantics ("what does
it mean?").  Everything else falls under pragmatics, often called "quality
of implementation" issues.  The STL is close to unique, outside the
real-time world, in making promises related to performance.  For the other
side, I spent the first 15 years of my career working largely on optimizing
Fortran compilers and runtime libraries:  no user base was more concerned
about speed than Fortran's, yet the Fortran standard didn't breathe a word
about how fast or slow any construct or standard function might be, could
be, or should be.  Similarly, the IEEE-754 standard for floating-point
arithmetic doesn't say anthing about speed either.  Standards bodies are
generally loathe to constrain implementations.

Of course the Python language itself is distinct from (albeit related to)
its implementations.  People working on implementations are typically quite
concerned about speed and memory.  The pragmatic tradeoffs can (and do) vary
across implementations, though (same as for C, C++, Fortran, Java, etc).





From skip at pobox.com  Mon Apr 14 14:29:03 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 13:29:03 -0500
Subject: Complexity of standard Python data structures (Re: Speed of string
        += string)
In-Reply-To: <BIEJKCLHCIOIHAGOKOLHGEFPFFAA.tim.one@comcast.net>
References: <slrnb9lsbe.4pa.marcus@infa.abo.fi>
        <BIEJKCLHCIOIHAGOKOLHGEFPFFAA.tim.one@comcast.net>
Message-ID: <16026.65007.680812.181463@montanaro.dyndns.org>

    Marcus> If somebody would say "python's list is a really a
    Marcus> vector/array", that's more or less the only documentation one
    Marcus> will ever need,
    ...

    Tim> I believe they would be to you, but I also believe you're in a
    Tim> minority there.

As long as we are talking about inferring significant semantics from
individual words or short phrases, both "vector" and "array" imply (to me)
numerically indexed containers of objects of a single type.  That doesn't
properly describe Python's lists which aren't restricted to containing
objects of a single type.

Skip




From exarkun at intarweb.us  Wed Apr 30 11:42:48 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 30 Apr 2003 11:42:48 -0400
Subject: plugin system
In-Reply-To: <20030430153624.GA22939@meson.dyndns.org>
References: <WHIra.27759$su3.3307778@twister.nyc.rr.com> <20030430153624.GA22939@meson.dyndns.org>
Message-ID: <20030430154248.GA26204@meson.dyndns.org>

On Wed, Apr 30, 2003 at 11:36:24AM -0400, Jp Calderone wrote:
> 
>   http://twistedmatrix.com/documents/TwistedDocs/current/api/

  Ahh frames how I hate thee:

  http://twistedmatrix.com/documents/TwistedDocs/current/api/public/twisted.python.plugin.html

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 41 days, 11:04, 5 users, load average: 0.14, 0.11, 0.03




From abpillai at lycos.com  Thu Apr  3 01:06:55 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 2 Apr 2003 22:06:55 -0800
Subject: Text Editor with wxPython
References: <3e7c561a.8718311@news.utanet.at> <af8b8v0lol2vqvqr09ag5aalbiid70huja@4ax.com> <bd993a2f.0303310456.2f201c43@posting.google.com> <9oai8vshpth0hvll13kbv85q0kefrpti0s@4ax.com>
Message-ID: <bd993a2f.0304022206.13b307a6@posting.google.com>

All the best. As for being a "newbie", dont worry about that.
I was a newbie 4 months back. Python is one of those languages
suitable to fast learning. ASPN python recipes are a sure-fire
way to help you get your python skills correct. As regards
Scintilla, they have implemented a small editor 'styce' using
scintilla. Dont miss that.

Regards

Anand Pillai

Franz Steinh?usler <f.steinhaeusler at amatic.com> wrote in message news:<9oai8vshpth0hvll13kbv85q0kefrpti0s at 4ax.com>...
> On 31 Mar 2003 04:56:56 -0800, abpillai at lycos.com (Anand B Pillai)
> wrote:
> 
> >I am thinking of doing something similar using wxWindows
> >(C++ port). I have not started doing anything on it, but
> >you can view my post on wxWindows mailing list.
> >[...]
> 
> Hello Anand,
> 
> thank you for your answer.
> 
> It is only an idea, I'm still a "Python newbie",
> but I will see.
> It seems, Scintilla and using wxStyledTextCtrl will be the best
> approach.
> 
> best regards,



From johnson at equadriga.com  Sat Apr 19 05:45:25 2003
From: johnson at equadriga.com (Elester Johnson)
Date: Sat, 19 Apr 2003 15:15:25 +0530
Subject: Python and SSL sockets
In-Reply-To: <3EA116D9.3020603@ghaering.de>
Message-ID: <LIELKILHEABECBGHPOOJGEEJCIAA.johnson@vdc.equadriga.com>


I'm Sorry Gerhard.......

Kind Rgds,
	Elester Johnson

-----Original Message-----
From: python-list-admin at python.org
[mailto:python-list-admin at python.org]On Behalf Of Gerhard Haring
Sent: Saturday, April 19, 2003 2:59 PM
To: python-list at python.org
Subject: Re: Python and SSL sockets


Elester Johnson wrote:
> Hi Friends,

People posting 9 KB of HTML (and in your case, even imgages, too) for a 
two-line question to a mailing list are definitely NOT my friends.

Please stop sending that junk NOW.

> How do I transfer an HTML file from a windows machine to a linux 
> machine (server), using SSL sockets.

I'd suggest to try uploading the file via HTTPS. Because Pre-2.3 Windows 
versions of Python aren't built with SSL support, you'll need my 
_socket.pyd replacement:

http://pypgsql.sourceforge.net/misc/python-ssl.html

-- Gerhard









From mchermside at ingdirect.com  Wed Apr 30 15:53:25 2003
From: mchermside at ingdirect.com (Chermside, Michael)
Date: Wed, 30 Apr 2003 15:53:25 -0400
Subject: os.path.walk() lacks 'depth first' option
Message-ID: <7F171EB5E155544CAC4035F0182093F0421213@INGDEXCHSANC1.ingdirect.com>

Tim wrote:
> Thanks for the nudge!  If you hadn't reminded us, I bet this would have been
> forgotten.  (I would have replied earlier, except my head got so heavy it
> took this look to peel my lips off the floor.)

Thanks. Someday when I'm NOT starting a new job and also in my first
year of parenthood, I'll try to contribute some work to Python. In
the meantime, little nudges and answering the occasional question on
c.l.p is the best I can contribute. But I find that I am learning a
great deal from reading python-dev -- some about python, some about 
how to successfully run a large project like this one, and sometimes
just a look at a really beautiful bit of design or code.

-- Michael Chermside


This email may contain confidential or privileged information. If you believe you have received the message in error, please notify the sender and delete the message without copying or disclosing it.





From john_bradbury at ___cableinet.co.uk  Thu Apr  3 14:49:11 2003
From: john_bradbury at ___cableinet.co.uk (John Bradbury)
Date: Thu, 3 Apr 2003 19:49:11 +0000 (UTC)
Subject: web server in python
References: <b6gnau$d7q$1@news.freedom2surf.net>
Message-ID: <b6i37m$m1f$1@sparta.btinternet.com>

Buy "Python Web Programming" by Steve Holden Pub. New Riders.  It has
examples & is one of the best books on computing that I have come across.

John Bradbury

"Tom Saunders" <toms at f2s.com> wrote in message
news:b6gnau$d7q$1 at news.freedom2surf.net...
> i am curerntly trying to make simple dynamic site in python which i intend
> to expand on. i can make the html using python but i have a few questions.
> using the BaseHTTPServer class, is it then possible to write all the code
> for a site inside one python program, which then serves based on requests
> from a website?
> ive been having trouble finding information about this and i would be
> greatfull if someone could give me an example of how to start.
> Hope you can help
>
> Tom
>
>





From missive at frontiernet.net  Fri Apr 18 09:31:32 2003
From: missive at frontiernet.net (Lee Harr)
Date: Fri, 18 Apr 2003 13:31:32 GMT
Subject: Zope REQUEST question
References: <OG2dnWc4TKANfAOjXTWcrg@august.net>
Message-ID: <slrnb9vhjh.76.missive@localhost.my.domain>

In article <OG2dnWc4TKANfAOjXTWcrg at august.net>, Terry Hancock wrote:
> Okay, this is a little tricky to explain.  This is technically a Zope 
> question, but it has to do with programming products in Python.
> 
> I'm doing a Forbidden Thing with a Zope product, which is overloading its 
> __getattr__ function.  It's a necessary Evil, though. ;-)
> 
> And actually, it works really well, as far as it goes.  It *is* a nightmare 
> to debug, which is why I've moved nearly all of the code out of this part 
> so it can be tested independently.
> 
> Anyway, it turns out, I need to be able to access the REQUEST object
> from there (I'm trying to implement an HTTP_REFERER block).
> 
> So far, I've tried:
> 
> getattr(self, 'REQUEST')
> 
> and
> 
> ObjectManager.__getattr__(self, 'REQUEST')
> 
> which should do the same thing, except the ObjectManager.__getattr__ 
> is much safer (it's not recursive).  I'm not sure why I thought that REQUEST
> would be an attribute of the object at this point, but I'm pretty sure I 
> read this somewhere and didn't just think it up out of thin air. However, 
> in this case at least, it's wrong -- these both give me attribute errors 
> (at least my object returns "not found" which is what happens when you
> raise an attribute error from inside of __getattr__ -- other errors usually
> give you the traceback you'd normally expect).
> 
> Does anyone have a better idea for how to get the REQUEST?
> 


1)  You really, really want to join the zope at zope.org mailing list. There are
     some people there who can answer this question in a second. If you do
     not like the high traffic, you can read the list through a mail-news
     gateway like gmane.

2)  How about self.REQUEST  ?




From aleax at aleax.it  Tue Apr  1 04:29:02 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 09:29:02 GMT
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <uznnb43a8.fsf@boost-consulting.com> <b6ajbj$3d15q$1@ID-169208.news.dfncis.de> <uadfb3vxc.fsf@boost-consulting.com>
Message-ID: <yZcia.53526$i26.1636429@news2.tin.it>

David Abrahams wrote:

> "Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> writes:
> 
>> greg at cosc.canterbury.ac.nzDavid Abrahams wrote:
>>> In my application it was easy to predict where the bulk of
>>> the time would be spent, and had I written the core engine in C++
>>> early on I might have saved myself a redesign effort.
>>
>> Or you could have designed that part of the system with
>> a possible C++ interface in mind, if it turned out to
>> be necessary later.
> 
> I thought I was doing that at the time.  The problem is that the
> edges of subsystems define themselves differently when there's a
> language boundary there from the beginning.

Yes, I agree and I think that's a part of a wider issue -- e.g.
to ensure your code is portable, it's best to develop in on TWO
platforms at once -- to ensure you can switch RDBMS engines later,
it's best to develop for two or three not very compatible engines
right from the start -- etc, etc.

The applicable italian proverb is, "practice is worth more than
grammar" ("val piu` la pratica che la grammatica").  Theory is
fine, but human beings tend to work better when SOME practical
counterpart is also present.  That's also why prototyping is SUCH
a good idea...


Alex




From tug at wilson.co.uk  Tue Apr 29 12:19:13 2003
From: tug at wilson.co.uk (John Wilson)
Date: Tue, 29 Apr 2003 17:19:13 +0100
Subject: Problem subclassing tuple
References: <F7DB8D13DB61D511B6FF00B0D0F062330489D361@mail.hebmex.com>
Message-ID: <092201c30e6b$13933590$640a0b0a@handel>

----- Original Message ----- 
[snip]
>
> Howdy.
>
> First of all... What are you trying to do? :-)

I am trying to add some methods to a tuple. Obviously the example I posted
has been considerably simplified to just demonstrate the problem I'm hitting
;-))

>
> Second.  If that's the full definition and intention
> of "Holder", then you might want to know (or are
> overlooking) the fact that "values", inside __init__,
> is already a tuple.

Yes, but I want an object which has all the characteristics of a tuple with
additional methods.

I could use delegation to do this but subclassing appeared to be an easier
way.

I'm really quite interested in why this doesn't work. I *thought* I'd
grasped the concept of the new style classes but this seems to demonstrate
that I haven't!

Thanks for the response.

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From imbosol-1049990205 at aerojockey.com  Thu Apr 10 13:27:33 2003
From: imbosol-1049990205 at aerojockey.com (Carl Banks)
Date: Thu, 10 Apr 2003 17:27:33 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk>
Message-ID: <9Qhla.7802$d8.3212@nwrdny02.gnilink.net>

Alexander Schmolck wrote:
> Jeff Epler <jepler at unpythonic.net> writes:
> 
>> Ignoring the fact that "globals" are module-level, you can obtain
>> "stacking" of global variables with only a single level of name lookup,
>> including automatic removal on return, in Python.  All that you need is a
>> convenent way to write complex things in the 'body' argument.  Untested
>> code, of course
>> 
>> _let_undef = object()
>> def let(body, **kw):
>>       restore = {}
>>       g = globals()
>>       for k, v in kw.items():
>>               restore[k] = g.get(k, _let_undef)
>>               g[k] = v
>>       try:
>>               return body()
>>       finally:
>>               for k, v in restore.items():
>>                       if v is _let_undef:
>>                               del g[k]
>>                       else:
>>                               g[k] = v
>>
> 
> Sure, but this is a but a clumsy, almost unusable imitation of
> scheme's (non-standard, but common) fluid-let macro. So what are you
> doing for multithreaded code (which scheme's faked dynamic extent
> via fluid-let doesn't cover easily but CL's special variables do)?
>
> Maybe I'm missing something, but non-dynamically scoped globals just
> seem like a bad idea to me.

What you're missing is that Globals are Evil(tm), and furthermore,
that if your program depends on changing a global in a dynamic scope,
it badly needs to be redesigned.  (To wit, that functions that refer
to the global you want to change need to be changed to accept it as an
argument.)

I think dynamic scpoing is a wholly bad idea, and simplifying a very
misguided programming practice doesn't make it any less so.


-- 
CARL BANKS



From fredrik at pythonware.com  Wed Apr 23 15:21:35 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 23 Apr 2003 21:21:35 +0200
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com>
Message-ID: <b86p2m$h47$1@main.gmane.org>

Aahz wrote:

> There were several times when I wrote what I thought were
> correct tests for the BCD module, and my code was right and
> my test was wrong.

here's the effbot "don't think if you don't have to" approach
to successful testing:

1. create a test script.
2. use 'doctest' to verify that your test script works as
   expected.  add functions with docstrings to the test
   script, and add code to the docstrings that exercises
   parts of your code.
3. add a few simple tests with expected results
4. write code
5. when all the simple tests pass, add more sophisticated
   tests, but don't add any expected output just yet.
6. run the test again, to see what results your code is producing.
   verify the results manually, in some suitable way.
   - for each result that turns out to be correct, copy the
      result into the test suite.
   - if you get a bogus result, go back to item 4.
7. if you feel that you need more tests, go to item 5.
8. if you need to add more features, go to item 3.
9. ????
X. profit!

</F>







From mis6 at pitt.edu  Fri Apr 25 14:00:47 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 25 Apr 2003 11:00:47 -0700
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk>
Message-ID: <2259b0e2.0304251000.20c5904a@posting.google.com>

Michael Hudson <mwh at python.net> wrote in message news:<7h3bryulp3e.fsf at pc150.maths.bris.ac.uk>...
> mis6 at pitt.edu (Michele Simionato) writes:
> 
> > The meta-type conflict is rather annoying, indeed. According to the
> > metaclass book, the language should generate the correct metaclass
> > that solves the conflict automatically, without burden for the
> > programmer. Python is not that magic (yet) therefore you have to
> > solve the conflict yourself.
> 
> I guess a potential problem is that if you do
> 
> class C(object):
>     __metaclass__ = A
> 
> class D(object):
>     __metaclass__ = B
> 
> class E(C, D):
>     pass
> 
> class F(C, D):
>     pass
> 
> you'd like both E and F to have the *same* (rather than merely
> equivalent) metaclasses.  This would seem to be tricky to arrange.
> 
> Cheers,
> M.

Yes, but I didn't posted  my full solution ;)

I take a list of all generated metaclasses and I reuse them:

  def withmemory(f):
      """This closure remembers all f invocations"""
      argskw=[]; result=[]
      def _(*args,**kw): 
          akw=args,kw
          try: # returns a previously stored result
              i=argskw.index(akw)
          except ValueError: # there is no previously stored result
              res=f(*args,**kw)  # returns the new result
              argskw.append(akw) # update argskw
              result.append(res) # update result
              return res
          else:
              return result[i]  
      _.argskw=argskw #makes the argskw list accessible outside
      _.result=result #makes the result list accessible outside
      return _

 
  _generatemetaclass=withmemory(_generatemetaclass) 

Now, when _generatemetaclass is invoked with the same bases it returns the
same metaclass:
   
>>> class A(type): pass
...
>>> class B(type): pass
...
>>> class C: __metaclass__=A
...
>>> class D: __metaclass__=B
...
>>> E=child(C,D,name='E')
>>> type(E)
<class 'oopp._AB'>
>>> F=child(C,D,name='F')
>>> type(F)
<class 'oopp._AB'>
>>> type(E) is type(F)
True

It works enough for me ;)

                                    Michele



From a.schmolck at gmx.net  Mon Apr  7 10:31:48 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 07 Apr 2003 15:31:48 +0100
Subject: Handling slices
References: <mailman.1049676910.30971.python-list@python.org> <yfsistrkofa.fsf@black132.ex.ac.uk> <mailman.1049693951.4240.python-list@python.org>
Message-ID: <yfssmsujs4r.fsf@black132.ex.ac.uk>

Jp Calderone <exarkun at intarweb.us> writes:

> On Mon, Apr 07, 2003 at 03:54:17AM +0100, Alexander Schmolck wrote:
> > [snip]
> > 
> > Here is a simple example:
> > 
> > class SliceMaker:
> >     """Convinience class to make slices."""
> >     def __getitem__(self, a): return a
> > 
> 
>   It may be worth noting that this is essentially the slice() builtin.

Not quite. The key is "Convinience" (even if misspelled :) :

 SliceMaker()[3:10:-1, NewAxis, ::-2, 0]

vs:

 (slice(3, 10, -1), None, slice(None, None, -2), 0)

I find the first much easier to parse and write correctly (the false analogy
between slice and range doesn't help).

alex



From andrew-pythonlist at puzzling.org  Wed Apr 30 09:53:13 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Wed, 30 Apr 2003 23:53:13 +1000
Subject: looping through a file
In-Reply-To: <uWOra.24347$3M4.683692@news1.tin.it>
References: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au> <uWOra.24347$3M4.683692@news1.tin.it>
Message-ID: <20030430135313.GA18706@frobozz.local>

On Wed, Apr 30, 2003 at 12:01:30PM +0000, Alex Martelli wrote:
[...]
> I think this will most likely solve your practical problems.  However,
> supposing for the sake of argument that InFile could be SO large that
> reading it all at once with readlines would be a problem (say, tens
> of megabytes), you could solve that too, for example, using again
> modern Python (2.2 or better in this case):
> 
> InFile = open('users.dat', 'r')
> 
> for line in InFile:
>     file_username = line.strip()
>     file_password = InFile.next().strip()
>     if file_username==users_username and file_password==users_password:
>         authorized = 1
>         break
> else:
>     authorized = 0
> 
> InFile.close()

I thought there might be something in 2.3's itertools module to help with
this -- I was hoping you'd be able to write something like:

    InFile = open('users.dat', 'r')
    
    for username, password in multistep(InFile, 2):
        if username == users_username and password == users_password:
            authorized = 1
            break
    else:
        authorized = 0
        
    
    InFile.close()

Where "multistep" would probably have a better name, and be defined like:

    def multistep(iterable, count):
        # XXX: What should happen when len(iterable) % count != 0 ?
        iterator = iter(iterable)
        while 1:
            values = []
            for i in range(count):
                values.append(iterator.next())
            yield values

I couldn't see anything in itertools to save me writing that out.  But I
just figured out how to do it with itertools -- izip can do it:

    def multistep(iterable, count):
        return itertools.izip(*((iter(iterable),)*count))

I guess you could write this directly:

    iterator = iter(InFile)
    for username, password in izip(iterator, iterator):
        ...

So it's probably not worth defining a seperate function...

-Andrew.





From mis6 at pitt.edu  Mon Apr 21 11:10:25 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 21 Apr 2003 08:10:25 -0700
Subject: Is this a super bug?
References: <mailman.1050811779.16487.python-list@python.org>
Message-ID: <2259b0e2.0304210710.7ab29457@posting.google.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1050811779.16487.python-list at python.org>...

> <snip> this is the sequence that "should" happen:
> 
> 1  super(ff, f)[0]  # f == self
> 2  superinst(type:<class ff>, obj:f)[0]
> 3  si.  getitem  (0)
> 4    si.  getattr  ('  getitem  ')
> 5    .. return str.  dict  ['  getitem  '].  get  (f)
> 6    boundmethod(str.  getitem  , f)
> 7  bmeth(0)
> 8   'a'

I don't understand what you are trying to say here ... 

> well, the transition from 3 to 4 doesn't happen which surprised me a
> little. What surprised me more was that I couldn't get   getattribute  
> to be called either (neither in the class or its metaclass... I've only
> got 2.3a2, and I read a note about this not working before a4. Is this
> what is happening? If not, the only other reason I could see would be
> from the refMan 5.3.2, "The primary must evaluate to an object of a
> sequence or mapping type."  If that is the case, then I still think it's
> a super bug, i.e. there is no reason super should assume that it's not
> "going to be" a sequence, in fact the direct superclass of ff  is  a
> sequence. To fix it simply add:
> 
>   def   getitem  (self, v):
>       m = self.  getattr  ('  getitem  ')
>       return m(v)
> 
> to the definition of super.

You may do that, but I don't like it. I like the fact that

super(cls,self)[index]

is *not* equivalent to

super(cls,self).__getitem__(index)

The present behaviour is the natural one, when you realize that super
objects are attribute descriptors. In other words:

super(cls,self) is *not* self

In you case 'self' is a string and has a __getitem__ method; nevertheless,
since super(cls,self) is a descriptor and not a string, I would not expect 
it to have a __getitem__ method, nor the other methods of a string. 

What happens in reality (disclaimer: I don't know nothing about the C 
implementation, I am guessing here) is that super(cls,self) has a 
__getattribute__ method that call the methods of the string object: the 
fact that you can write

super(cls,self).__getitem__(index)

does *not* means that super(cls,self) has a __getitem__ method.

I agree this can be confusing at first. The solution would be to
document ``super`` and the attribute descriptors. 

Maybe one of these days I will collect my notes and fill this gap,
but this would require at least 20-30 pages, so it would take some
time ...


Hope this helps,

--
Michele Simionato - Dept. of Physics and Astronomy
210 Allen Hall Pittsburgh PA 15260 U.S.A.
Phone: 001-412-624-9041 Fax: 001-412-624-9163
Home-page: http://www.phyast.pitt.edu/~micheles/



From mjd at plover.com  Wed Apr 23 14:17:39 2003
From: mjd at plover.com (Mark Jason Dominus)
Date: Wed, 23 Apr 2003 18:17:39 +0000 (UTC)
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <1azna.115402$It5.22896@news2.central.cox.net> <tyfhe8xqhmy.fsf@pcepsft001.cern.ch> <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com>
Message-ID: <b86lc3$tnd$1@plover.com>

In article <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4 at 4ax.com>,
Courageous  <jkraska at san.rr.com> wrote:
>
>>But he regrets the inclusion of a crippled lambda, rather than the
>>inclusion of support for anonymous functions (IIRC).
>
>Anyway, I guess what I'm saying is that I suspect that Guido
>didn't have any good alternatives. 


I don't believe that's correct.  At an early Open Source Conference,
perhaps around 1998 or so, I was present at a conversation between
Guido and Tom Christiansen.  Tom asked Guido why Python didn't have
closures, and Guido's reply (and I think this is an exact quote) was
"Closures aren't important."

I remember it because I was so astonished by the obtuseness of the answer.




From staschuk at telusplanet.net  Wed Apr 30 10:23:25 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 30 Apr 2003 08:23:25 -0600
Subject: looping through a file
In-Reply-To: <20030430135313.GA18706@frobozz.local>; from andrew-pythonlist@puzzling.org on Wed, Apr 30, 2003 at 11:53:13PM +1000
References: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au> <uWOra.24347$3M4.683692@news1.tin.it> <20030430135313.GA18706@frobozz.local>
Message-ID: <20030430082325.G334@tibia.amotlpaa.bogus>

Quoth Andrew Bennetts:
  [...]
> I thought there might be something in 2.3's itertools module to help with
> this -- I was hoping you'd be able to write something like:
> 
>     InFile = open('users.dat', 'r')
>     
>     for username, password in multistep(InFile, 2):

Such a function (with a name something like group) has been
proposed, but I don't think it's in itertools yet.

  [...]
> I couldn't see anything in itertools to save me writing that out.  But I
> just figured out how to do it with itertools -- izip can do it:
> 
>     def multistep(iterable, count):
>         return itertools.izip(*((iter(iterable),)*count))

Heh.  That's a clever implementation.  (Perhaps too clever.)

  [...]
> So it's probably not worth defining a seperate function...

Imho a separate function is desirable for clarity.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From simonb at webone.com.au  Fri Apr 11 19:20:05 2003
From: simonb at webone.com.au (Simon Burton)
Date: Sat, 12 Apr 2003 09:20:05 +1000
Subject: instance creatiion as an "event"
References: <mailman.1050095779.7490.python-list@python.org>
Message-ID: <pan.2003.04.11.23.20.04.65410@webone.com.au>

On Fri, 11 Apr 2003 17:15:32 -0400, Arthur wrote:

> I am trying to re-engineer, as lightly as possible, a Python app I 
> wrote, to be make it more functional when used at the interactive prompt.
> 
> The only thing I seem to be missing is some kind of "hook" to trigger a 
> general update cycle of all existing class instances upon the creation 
> of a new class instance at the interactive prompt
> 

class Member:
	members = []
	def __init__(self):
		for m in Member.members:
			m.notify(self)
		Member.members.append(self)
	def notify(self,other):
		print self,"a new member:",other

#Just an idea,
#
#Simon Burton.






From aleax at aleax.it  Sun Apr 13 05:21:42 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 09:21:42 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <3E98D281.8C0CE87D@easystreet.com>
Message-ID: <G_9ma.9028$T34.282778@news2.tin.it>

achrist at easystreet.com wrote:

> We went around on this one about 4 years ago. Is this a url:
> 
> 
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&threadm=37054490.E79ADCC9%40easystreet.com&rnum=1&prev=/groups%3Fnum%3D100%26hl%3Den%26lr%3D%26ie%3DISO-8859-1%26as_drrb%3Db%26q%3Dgroup%253Acomp.lang.python%2Binsubject%253Astring%2Bauthor%253Aachrist%2540easystreet.com%26btnG%3DGoogle%2BSearch%26as_mind%3D12%26as_minm%3D5%26as_miny%3D1993%26as_maxd%3D12%26as_maxm%3D4%26as_maxy%3D1999

> About 6 messages into the thread, Tim Peters came up with a way that
> was about 5 times faster than string.join().

Not exactly -- the "way" you mention just spreads the costs out,
so they don't SHOW when you're "joining" the pieces (which are
already "joined").  In modern terms:

[alex at lancelot python2.3]$ python -O timeit.py -s'''
import array; SAR=array.array("c")
data=map(str,range(9999))
''' 'for x in data:' ' SAR.fromstring(x)' 'z=SAR.tostring()'
100 loops, best of 3: 2.67e+04 usec per loop

[alex at lancelot python2.3]$ python -O timeit.py -s'''
data=map(str,range(9999))
''' 'z="".join(data)'
100 loops, best of 3: 2.14e+03 usec per loop


Sure, the cost of the .tostring is tiny, but when you have a
lot of little pieces you must loop over them and .fromstring
each -- and THAT is where you're paying the cost.  So .join
is actually an order of magnitude faster...!-)


Alex




From gh at ghaering.de  Fri Apr  4 12:22:05 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: 4 Apr 2003 17:22:05 GMT
Subject: What are the differences between _weak_ references and "normal" references?
References: <169a7283.0304040901.4bc3531b@posting.google.com>
Message-ID: <slrnb8rfhp.1l8.gh@haering.opus-gmbh.net>

sdieselil <sdieselil at yahoo.com> wrote:
> Hi
> 
> I noticed a module weakref in Python standard library. What is weak
> reference and in what way it differs from "normal" reference?

Not that I fully understand it myself, but have you looked into the module
documentation for weakref and/or http://www.python.org/peps/pep-0205.html
yet?

>By the way, how can I implement usual reference in Python (I mean C++
>reference)?

::ASSIGN REFERENCE::

C++:

    Foo foo1("foo1");
    Foo& foo2 = foo1;

Python:

    foo1 = Foo("foo1")
    foo2 = foo1

::COPY VALUE::

C++:

    Foo foo1("foo1")
    Foo foo2 = foo1;    // call copy constructor

Python:

    import copy
    foo1 = Foo("foo1")
    foo2 = copy.copy(foo1)

I'm currently re-learning C++, so there may be bugs in the above C++ :-)

-- Gerhard



From imbosol-1049925772 at aerojockey.com  Wed Apr  9 18:17:53 2003
From: imbosol-1049925772 at aerojockey.com (Carl Banks)
Date: Wed, 09 Apr 2003 22:17:53 GMT
Subject: Macros in Python?
References: <b71qde$q71$07$1@news.t-online.com>
Message-ID: <l_0la.1989$d8.141@nwrdny02.gnilink.net>

Dominic wrote:
> 
> 
> (defmacro with-root-privs (() &body body)
>   (let ((oldidsym (gensym)))
>     `(let ((,oldidsym (geteuid)))
>        (seteuid 0)
>        (unwind-protect
>           (progn , at body)
>         (seteuid ,oldidsym)))))
> 
> Is there an easy way to produce similar code
> in Python which gurantees that certain things
> happen around some code? (Maybe by using
> generators?)
> 
> The normal approach would be to enclose everything
> with try: finally: but this would scatter the
> code all over.


It doesn't have to.  First, define:

    def get_root_privledges():
        euid = os.geteuid()
        os.seteuid(0)
        return euid

    def release_root_privledges(euid):
        os.seteuid(euid)


Then you can do:

    old_euid = get_root_privledges()
    try:
        if x in user:
            active.append(x)
            ...
    finally:
        relase_root_privledges(old_euid)


I hope this isn't what you consider "scattering the code all over."
There is a certain niceness to the with macros (the ones that don't
rebind any symbols, anyways), but try ... finally does the job just
fine.


-- 
CARL BANKS



From m.faassen at vet.uu.nl  Tue Apr  1 18:05:24 2003
From: m.faassen at vet.uu.nl (Martijn Faassen)
Date: 1 Apr 2003 23:05:24 GMT
Subject: Mimic a web surfer ?
References: <3e894b7d$0$29158$626a54ce@news.free.fr>
Message-ID: <b6d5vk$lvo$3@newshost.accu.uu.nl>

Hey,

You may want to take a look at ClientCookie and ClientForm. I've used
those to mimic a web user.

http://wwwsearch.sourceforge.net/ClientCookie/

Regards,

Martijn
-- 
History of the 20th Century: WW1, WW2, WW3?
No, WWW -- Could we be going in the right direction?



From okumee at compuserve.de  Wed Apr 16 08:26:00 2003
From: okumee at compuserve.de (mnietz)
Date: 16 Apr 2003 05:26:00 -0700
Subject: trouble using signal
Message-ID: <6b9f8eda.0304160426.69064546@posting.google.com>

Hi Folks,

I am trying to use signals for a timeout in a http-request as
following:


#------------------------------------------------------------------------------
def requestUrl(host, data):
    """starts a http-request for given data to the given url
       and returns status code. If request will not be answered in
specifiend
       time, status 404 will be returned
    """
    conn = httplib.HTTPConnection(host)
    conn.request("GET", data)
   
    status = 404
    content = None
    
    signal.signal(signal.SIGALRM, requestUrlTimoutHandler)
    signal.alarm(10)
    
    r = conn.getresponse()
    
    signal.alarm(0)
    
    status = r.status
       
    if status == 200:
       content = r.read()
    
    return (status, content)   

#------------------------------------------------------------------------------
def requestUrlTimoutHandler(signum, frame):
    print "the url request timed out, status stays = 404"



but that will bring up following Exception:
....
  File "/usr/local/lib/python2.2/httplib.py", line 760, in getresponse
    response.begin()
  File "/usr/local/lib/python2.2/httplib.py", line 269, in begin
    version, status, reason = self._read_status()
  File "/usr/local/lib/python2.2/httplib.py", line 231, in
_read_status
    line = self.fp.readline()
IOError: [Errno 4] Interrupted system call


Is the reason for this, that r.read is interrupted by the
timeout-handler? But timeout handling should work this way. Or should
I just catch the exception and ignore it?

Thanks for any advice,

Best Regards,
Matthias



From syver-en+usenet at online.no  Tue Apr  1 16:41:52 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 01 Apr 2003 23:41:52 +0200
Subject: ANNOUNCE:  SCons 0.13 (build tool in Python) is now available
References: <mailman.1049219524.21980.python-list@python.org>
Message-ID: <uk7educ7z.fsf@online.no>

I get an error when trying to download the 0.13 version. Sourceforge
complains that the file doesn't exist.
-- 

Vennlig hilsen 

Syver Enstad



From dsoulayrol at free.fr  Tue Apr 22 10:01:47 2003
From: dsoulayrol at free.fr (David)
Date: 22 Apr 2003 07:01:47 -0700
Subject: [Newbie] A problem while trying Xsl transformation with 4Suite
Message-ID: <cbdb2fee.0304220601.58d81d40@posting.google.com>

Hi,

I'm quite new to Python, and I'd like to use it to apply some XSL
transformations to XML Documents. I first had a look at PyXML, but I
had problems, and I read here some messages suggesting the use of
4XSLT.

Then I installed it. I wrote this little program to test it :

-----
from Ft.Xml.InputSource import DefaultFactory
from Ft.Xml.Domlette import ValidatingReader
from Ft.Xml.Xslt.Processor import Processor

lProc = Processor()
lSheet = DefaultFactory.fromUri('file:../../xsl/page.xsl')
lDoc = DefaultFactory.fromUri('file:file.xml')
lProc.appendStylesheet(lSheet)
result = lProc.run(iSrc = lDoc, outputStream = open('out.html', 'w'))
-----

It works very well... but not at all when I have a <!DOCTYPE ... >
declaration in the header of my XML documents. For example, this
document is correctly transformed :

------
<?xml version="1.0" encoding="ISO-8859-15"?>
<article>
  <title>Acc?s non autoris?</title>
  <section>
    <para>
      Vous n'?tes pas le bienvenue sur cette page. Ouste ! Du balai !
...
  </para>
  </section>
</article>
------

... whereas this one :

------
<?xml version="1.0" encoding="ISO-8859-15"?>
<!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
            "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">

 <the rest is the same>
------

... will fail with this error :

Traceback (most recent call last):
  File "tXslt.py", line 13, in ?
    result = lProc.run(iSrc = lDoc, outputStream = open('out.html',
'w'))
  File "/usr/lib/python2.2/site-packages/Ft/Xml/Xslt/Processor.py",
line 150, in run
    raise XsltException(Error.SOURCE_PARSE_ERROR, iSrc.uri or '<Python
string>', e)
Ft.Xml.Xslt.XsltException: Source document (file:error404.xml): cannot
concatenate 'str' and 'NoneType' objects


Can someone please help ?
David.



From mertz at gnosis.cx  Tue Apr  8 14:03:05 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 14:03:05 -0400
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com>
 <mailman.1049768113.16134.python-list@python.org>
 <kauu6b.f2d1.ln@boundary.tundraware.com>
Message-ID: <Z7wk+kKkXwrV092yn@gnosis.cx>         

Tim Daneliuk <tundra at tundraware.com> wrote previously:
|I am unclear, though, on whether a mathematician understands
|1.0 to mean 1.0000000... (a single number) or the _neighborhood_
|around 1 with one digit precision.

It depends what type of mathematician (and what mathematical context).
"Pure" math deals with approximations and confidence levels less often
than does applied math.  So in those areas, mathematicians are likely to
have a first thought of Real or Rational numbers.  In statistics,
however, I assume every mathematician would read 1.0 as both a number
and a precision.

Yours, Lulu...

--
 mertz@  _/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: \_\_\_\_    n o
gnosis  _/_/             Postmodern Enterprises            \_\_
.cx    _/_/                                                 \_\_  d o
      _/_/_/ IN A WORLD W/O WALLS, THERE WOULD BE NO GATES \_\_\_ z e





From donn at u.washington.edu  Fri Apr 18 19:45:37 2003
From: donn at u.washington.edu (Donn Cave)
Date: 18 Apr 2003 23:45:37 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu>
Message-ID: <b7q2n1$24d2$1@nntp6.u.washington.edu>

Quoth David Eppstein <eppstein at ics.uci.edu>:
| In article <C8_na.646$dP1.2863 at newsc.telia.net>,
|  "Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote:
|>> What's wrong with:
|>>
|>> t = 0
|>> for y in x:
|>>      t += y[1]
|>>
|>> The extra variable?  The number of lines?
|> 
|> I am really curious about that to, I would most definitley say that this is
|> the most pythonic way. It is simple and easily readabel by anyone and I am
|> sure it is just as fast any of the other way.
|
| If you want the sum of a list of items, you should write it in a way 
| that looks like "the sum of a list of items", not in a way that looks 
| like "loop over these items, maintain another variable t, perform a 
| sequence of additions".  Why do we have high level languages if not to 
| express our intentions at a higher level and let the language worry 
| about what low-level operations are needed to implement it?

So you see the functional programming answers as "higher level"
than the imperative, procedural approach taken above?  I think
many functional programming enthusiasts would see it that way,
but it's more interesting to hear it from a Python programmer.

I personally don't see it - I mean, I think there are some good
things about the functional approach, but I guess I have higher
standards for higher level.

	Donn Cave, donn at u.washington.edu



From jester at NOSPAM.mcsnospam.vuw.acNOSPAM.nz  Fri Apr 18 03:23:18 2003
From: jester at NOSPAM.mcsnospam.vuw.acNOSPAM.nz (Mike McGavin)
Date: Fri, 18 Apr 2003 19:23:18 +1200
Subject: Zope authentication from python scripts
Message-ID: <HJNna.3123$AB5.402420@news02.tsnz.net>

Hi everyone.

I'm experimenting with Zope but after a few weeks trying on and off, I'm 
still quite confused with how the user authentication works. Sorry if 
this is a straightforward question.

I've been trying to make an experimental website using page templates 
and python scripts where a user can log in by entering a username and 
password via a web form. If possible I'd like to use a standard Zope 
user folder to store the users in, and use Zope as much as possible to 
keep track of who's logged in in what session.

The specific problem I'm having is that I can't figure out how to get 
the code that handles the submitted web form to authenticate and log the 
user in so that Zope recognises an Authenticated User in the session. I 
have discovered that if I call:

  u=container.acl_users.authenticate("myusername","mypassword",request)

from a script, then u will be assigned a User object if the password was 
correct, or None otherwise.

Comments around the source code for the authenticate() method imply that 
it's not meant to be public though, so I'm not sure if it's even correct 
to use it. In any case, it still doesn't seem to get Zope to recognise a 
specific Authenticated User in the session, or do anything besides 
confirm that the password is correct. In particular, running 
request['AUTHENTICATED_USER'] still returns a user object containing an 
Anonymous User, implying that nobody was logged in.


I could kludge my own login system by storing the current user in a 
session variable, but this seems to side-step all of the security that 
Zope has built in. It'll forever be seeing the user as anonymous even if 
my session variable knows who it is.


I'd appreciate any help on this or a pointer to a good reference. I 
haven't been able to figure much out from the online documentation or 
reading the source code.

Thanks.
Mike.




From andrew-pythonlist at puzzling.org  Sat Apr 26 06:36:24 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sat, 26 Apr 2003 20:36:24 +1000
Subject: Forgetting "()" when calling methods
In-Reply-To: <WRsqa.9333$K35.305964@news2.tin.it>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it>
Message-ID: <20030426103624.GC26530@frobozz.local>

On Sat, Apr 26, 2003 at 10:05:42AM +0000, Alex Martelli wrote:
[..snip..]
> I don't know what to DO about it, but some minor tweaks might help.
> For example, I cannot think of any code I've ever written that would
> break if methods (and functions) spouted a __nonzero__ special method
> (or, of course, the C-coded equivalent thereof) raising an exception
> with a clear error message -- this would at least provide prompt

I can!  I'm sure Twisted, and probably other places, use this idiom:

    # Call a handler if one is defined
    func = getattr(self, 'handler_' + name', None)
    if func:
        func(...)

This idiom could be easily adapted to use "if func is not None", but it's
not clear that that's a better way to express it.  Anyway, my point is that
making function's __nonzero__ raise an exception will certainly break
existing code!

The same could probably be said of, e.g. setting class's and type's
__nonzero__ to raise an exception, to avoid "instance = MyClass" bugs too.
I think this would also break things.  "if foo:" is just too commonly relied
upon to mean "is 'foo' *something*, or just an empty placeholder?" in
Python, and so making anything's __nonzero__ raise an exception is probably
a bad idea.

-Andrew.





From adalke at mindspring.com  Wed Apr 16 18:10:16 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 16 Apr 2003 16:10:16 -0600
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <a8b7f07a.0304161158.29c0ef6b@posting.google.com>
Message-ID: <b7kk53$fe7$1@slb2.atl.mindspring.net>

A. Lloyd Flanagan
> This mathematical analysis of an algorithm's behavior (and best-case,
> average, and worst-case performance) can be trivial.  For example, x +
> 1 is O(1).

Or not so trivial.  If x can be an arbitrary size integer, represented
in memory as a chunk of bits, the value 'n' for the number of bits,
and a sign flag, then addition could be O(n) == O(log(x)).  Eg,
suppose    x =  1111111111111111111111111111111111111111
add 1 to get = 10000000000000000000000000000000000000000
so 49 bits were affected.

But for fixed size integers, the above is true.

                    Andrew
                    dalke at dalkescientific.com





From tjg at craigelachie.org  Wed Apr 23 13:42:59 2003
From: tjg at craigelachie.org (Timothy Grant)
Date: Wed, 23 Apr 2003 10:42:59 -0700
Subject: Data-driven testing
In-Reply-To: <b86e9f$c15$1@panix1.panix.com>
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com>
Message-ID: <200304231042.59613.tjg@craigelachie.org>

On Wednesday 23 April 2003 09:16 am, Aahz wrote:
>
> Which brings up the question: do the unit-testers here write tests for
> throwaway scripts?

Absolutely! Well, mostly if I'm going to be writing a throw-away that will be 
manipulating important data. I wrote some code to reformat all the phone 
numbers in a *very* large database, that contained phone numbers and other 
things in the phone number field, and the phone number formats were all 
different, etc.

I made a list of every exception I could find, and wrote tests that each one 
of those conditions was correctly handled, and that the script bailed if it 
came across a condition it couldn't handle.

By the time I ran that thing against our live database, I was extraordinarily 
confident that it would work flawlessly.

-- 
Stand Fast,
    tjg.

Timothy Grant
www.craigelachie.org





From grante at visi.com  Sun Apr 27 23:33:51 2003
From: grante at visi.com (Grant Edwards)
Date: 28 Apr 2003 03:33:51 GMT
Subject: WML examples using mod_python?
References: <slrnbals94.2qp.grante@tuxtop.visi.com> <uLFqa.9353$h%2.1023630@read1.cgocable.net> <slrnbamprf.2th.grante@tuxtop.visi.com> <b8h94q$g7v$0@216.39.172.122>
Message-ID: <slrnbap8ht.14n.grante@tuxtop.visi.com>

In article <b8h94q$g7v$0 at 216.39.172.122>, Bengt Richter wrote:
> On 27 Apr 2003 05:10:45 GMT, Grant Edwards <grante at visi.com> wrote:
>>In article <uLFqa.9353$h%2.1023630 at read1.cgocable.net>, Mark Nenadov wrote:
> [...]
>>> If you are looking for something to make WML-generation
>>> quicker, you might want to check out "pywml"
>>> (http://sourceforge.net/projects/pywml).
>>
>>Thanks!  That's what I was looking for.  I was sure there had
>>to be something like that, but all my searches found were a
>>Zope/WML howto.
>
> I was curious, but I found just an empty project. Maybe there
> used to be a Weapon of Mass Lunacy there, and it had to be
> removed quickly ;-/

If you're talking about http://sourceforge.net/projects/pywml:
there are files in CVS though none have been officially
"released".  If you click on "browse CVS" on the project page,
you should be able to find the source code file.  The project
seems to be pretty dead, so I've no idea what state the code is
in (I haven't tried it, so I can't even say it compiles).

I picked up a remaindered WML book today, and have been playing
with the IMAP module (an IMAP to WML gateway is the project I
have in mind).

-- 
Grant Edwards                   grante             Yow!  HERE!! Put THIS
                                  at               on!! I'm in CHARGE!!
                               visi.com            



From bourassa-m at rmc.ca  Thu Apr 10 14:22:42 2003
From: bourassa-m at rmc.ca (bourassa)
Date: Thu, 10 Apr 2003 14:22:42 -0400
Subject: Python compared toJava
In-Reply-To: <3E95A2A7.238BFA6C@engcorp.com>
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com> <3E95A2A7.238BFA6C@engcorp.com>
Message-ID: <3e95b7d6$1@rmcnt03a.rmc.ca>

Would it have bee so hard to either ignore the post or to give a useful 
url?  Learn some manners before _your_ next post.

Peter Hansen wrote:
> Blaktyger wrote:
> 
>>I recently read that the Python interpreter is like the Java Virtual Machine.
>>Is that true?
> 
> 
> Yes!  In fact, it's exactly the same thing!  We just use a different name
> for it to avoid the wrath of Sun Microsoftsystems, and because we kind of like
> snakes.
> 
> Note: see http://www.catb.org/~esr/faqs/smart-questions.html before posting
> again please. 
> 
> -Peter




From sismex01 at hebmex.com  Tue Apr 22 14:35:32 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Tue, 22 Apr 2003 13:35:32 -0500
Subject: True/False
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F06233046FEA98@mail.hebmex.com>

> From: Ian Bicking [mailto:ianb at colorstudy.com]
> Sent: Tuesday, April 22, 2003 1:30 PM
> 
> [...snip...]
>
> import sys
> if sys.version_info < (2, 3):
>     True, False = 1==1, 0==1
> 
> But I *so* don't want to do this.  Is there any reason why reassigning
> True and False will be a problem in the future?  Should I tell people
> who are putting in version tests that it isn't necessary?
> 
>   Ian
>

No problem, because (1==1) will evaluate to True [or bool(1)]
and (1==0) will evaluate to False [or bool(0)].

hth

-gus




From koenig at v-i-t.de  Sun Apr 20 21:25:34 2003
From: koenig at v-i-t.de (Mirko Koenig)
Date: Mon, 21 Apr 2003 03:25:34 +0200
Subject: importing modules with the same name but from different directories
Message-ID: <b7vhad$1t6$1@news.tu-darmstadt.de>

Hi

The subject is a little bit long but i didn't how how to describe it
shortly.

I hope my english is good enough to explain my problem.

I have a program consisting of many modules. Every modul has its own
directory in a directory called modul, eg.

.../modul/modul1
.../modul/modul2

In every modul directory is a diretory for the language files. These
files are used to display the modules label in different languages.
eg.

.../modules/module1/language/german.py
.../modules/module1/language/english.py
.../modules/module2/language/german.py
.../modules/module2/language/english.py

My problem is:
If is start modul1 from within its modul directory ( .../modules/modul1 )
And modul1 calls modul2, then modul2 doesn't load its own language file.
Instead it loads the langugae file of modul1.

In modul1 i wrote sys.path.append( "./languages" ) then import german.
In modul2 i wrote the same. I know in that way it coulnd't work.

What can i do, that every modul knows its 'homedir', no matter from where
i start it ?

I don't want to set environment variables, because the programm is for
beginners, that are not used to pc's and want an easy going programm and
not to set env vars.

I'm developing under linux if that help solving my problem

thanx for comments

Mirko Koenig



From BPettersen at NAREX.com  Wed Apr 16 06:48:31 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 16 Apr 2003 04:48:31 -0600
Subject: Embedding Python Engine into a windows executable
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDF56@admin56.narex.com>

> From: Karam Chand [mailto:karam_chand02 at yahoo.com] 
> 
> Hello,
> 
> Is is possible to embed Python engine completely into a windows
> executable? 

Yes (all it involves is linking with a third party library and calling
functions in it -- unless I'm misunderstanding your question...) You'll
have to use Boost.Python if you insist on the static requirement below.
Note: you _will_ need to extend Python with modules too, to make your
datatypes available to Python (it's the extension part Boost.Python can
do statically vs the others).

> Basically, I want to give some kind of python scripting to
> allow users extend my applicaion and I dont want to ship python*.dll
> separately.

First reconsider. If you're creating a traditional windows executable,
you allready have a collection of files to distribute. If you're doing
this for the sole purpose of getting a single executable file, I'd
suggest going the other way: develop in Python, extend with
Boost.Python, and use McMillan's Installer or py2exe to create your
executable (I've used Installer, but not py2exe, so someone correct me
if I'm wrong... :-)

If you're trying to create a singular executable to make distribution
and installation easier, you're headed in the wrong direction. I haven't
met many GUI applications that doesn't need to set up its "environment"
before it can start running, most logically in a separate initialization
program, ...

The only valid reason I can come up with is "production environment
policy" (which seems to be quite common although mostly misguided
<wink>). If this is the case, you'll need to recompile Python from
sources yourself as a static library. AFAIK this is never done on
Windows, but is done on various other OSen so I can't imagine it to be
too hard (though terribly boring compared to real coding <wink>).

-- bjorn




From ianb at colorstudy.com  Wed Apr  9 02:21:41 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 09 Apr 2003 01:21:41 -0500
Subject: [ANNOUNCE] KirbyBase 1.0
In-Reply-To: <tdNka.76495$0X.14720956@twister.columbus.rr.com>
References: <tdNka.76495$0X.14720956@twister.columbus.rr.com>
Message-ID: <1049869301.697.136.camel@lothlorien>

On Tue, 2003-04-08 at 23:21, Jamey Cribbs wrote:
> 4. KirbyBase allows you to specify field types, then it  allows you to 
> use Python's regular expression syntax to search by string fields and 
> Python expressions to search by numeric fields.  So, for exampe, to 
> search a table containing info on WWII fighter planes and select US 
> planes that had a maximun speed higher than 400 mph, you would say 
> something like:
> 
>      db.select('plane.tbl', ['country', 'speed'], ['USA', '>400'])
> 
> I like doing this more than building SQL query strings in Python.

You might be interested in the query builder I'm using in SQLObject
(sqlobject.org), in the SQLObject.SQLBuilder module.  It allows you to
write a query like:

  (column.country == "USA") & (column.speed > 400)

Or:

  country_code = "USA"
  required_speed = 400
  (column.country == country_code) & (column.speed > required_speed)

-- 
Ian Bicking  ianb at colorstudy.com  http://colorstudy.com
4869 N. Talman Ave., Chicago, IL 60625  /  773-275-7241
"There is no flag large enough to cover the shame of 
 killing innocent people" -- Howard Zinn





From ark at research.att.com  Sun Apr 20 16:44:02 2003
From: ark at research.att.com (Andrew Koenig)
Date: Sun, 20 Apr 2003 20:44:02 GMT
Subject: Good Python generators example?
References: <paDoa.1611$hT2.1228371@news2.news.adelphia.net>
Message-ID: <yu99y9243nn1.fsf@europa.research.att.com>

Robert> Can someone point me to a nice concise Python 2.2 compatible
Robert> example of the use of generators?

Here's a program that solves the "8 queens" problem--that is, the
problem of how to place 8 queens on a chessboard so that none of
them attacks any of the others.  It prints all the solutions.

from __future__ import generators

class Queens(object):
    def __init__(self, n):
        assert n >= 0
        self.N = n
        self.board = []

    def generate(self):
        if len(self.board) == self.N:
            yield self.board[:]
        else:
            for i in range(self.N):
                if self.safe(i):
                    self.board.append(i)
                    for i in self.generate():
                        yield i
                    self.board.pop()

    def safe(self, n):
        "Is is safe to append a queen to the board in row n?"
        m = len(self.board)
        for i in range(m):
            b = self.board[i]
            if n == b or b+(m-i) == n or b-(m-i) == n:
                return False
        return True
    
for i in Queens(8).generate():
   print i


-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From cben at techunix.technion.ac.il  Thu Apr  3 12:51:43 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Thu, 3 Apr 2003 20:51:43 +0300 (IDT)
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <x7y92uhzpw.fsf@guru.mired.org>
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <Pine.GSO.4.51.0304032019150.1807@techunix.technion.ac.il>

Mike Meyer wrote on 2003-04-01:

> PEP: 313

PEP: CCCXIII you mean?  I think this would make the PEP index much
more convenient to navigate :-).

> [snip]

> Compatability Issues
>
>     No new keywords are introduced by this proposal.  Programs that
>     use variable names that are all upper case and contain only the
>     characters M, D, C, L, X, V and I will be affected by the new
>     literals.  These programs will now have syntax errors when those
>     variables are assigned, and either syntax errors or subtle bugs
>     when those variables are referenced in expressions.  Since such
>     variable names violate PEP 8 [3], the code is already broken, it
>     just wasn't generating exceptions. This proposal corrects that
>     oversight in the language.
>
While I sympatise with the "broken" sentiment, that's clearly not
Python's way to handle such changes.  The right thing is obviously::

    Python MMM (#I Apr 1 2003 12:00-48)
    [GCC III.V (Red Hat Linux VI^H^HEMACS;)]
    >>> III = 'example'
    >>> III
    'example'
    >>> # ---vvvvvvvv-----------------------
    >>>
    >>> from __past__ import roman_numerals
    >>>
    >>> # ---^^^^^^^^----------------------
    >>> III
    3
    >>> III = 'bad'
    SyntaxError: can't assign to literal

Yes, I know __past__ is reserved for the past in Python's development,
not the surrounding universe's past; importing from it would encourage
keeping obsolete baggage, so it will never be implemented; that means
you can't do::

    from __future__ import __past__

to test the above code in current or soon-to-appear Pythons...
Perhaps something like::

    from __never__ import __past__

would be nice to have for such situations...  I wonder if pymacs
together with emacs' `M-x do`_ would allow time-travel access from
Python, that would at least relieve some burden from the devolopers, I
think.

.. M-x do:
   http://savannah.gnu.org/cgi-bin/viewcvs/emacs/emacs/etc/
   future-bug?rev=HEAD&content-type=text/vnd.viewcvs-markup

Another approach altogether would be to give actual meaning to the
`__date__` module attribute: python would run the program according to
the current Python documentation at this date.  That should fix
compatibility issues once and for all.  Of course, this depends on
your system's ``/dev/time`` being seekable in both directions...

was-doing-homeworkds-on-April-fools-day-ly y'rs
  Beni Cherniavsky <cben at tx.technion.ac.il>

If I don't code for myself, who will?
And if I don't GPL it, what am I?




From aleax at aleax.it  Thu Apr 24 16:58:36 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 20:58:36 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org>
Message-ID: <0eYpa.1909$3M4.45981@news1.tin.it>

Bjorn Pettersen wrote:
   ...
> Then define a hierarchy:
> 
>  class Counted(object):
> __metaclass__ = metaIC
> 
>  class Point(object):
> __metaclass__ = metaMetaBunch
> x = 0
> y = 0
> 
>  class CountedPoint(Counted, Point):
>      def __init__(self):
>          super(CountedPoint.self).__init__(self)

hmmm. I think you want a comma there, not a dot, in the super args.
And you shouldn't pass self as an argument here, either.  Just:

           super(CountedPoint, self).__init__()

But really, this whole __init__ method has no reason for being,
as it's just delegating to exactly the same method that would be
used if it didn't exist!!!  So, snip it...

   ...
> not exactly clear _what_ the conflict is, i.e. it's not that the meta
> classes must have a common base (both isinstance(x, type), it's not that
> type.__new__ should perhaps be super(x, type).__new__, it _is_ that the
> metaclasses aren't related by inheritance. Trying to fix this by, e.g.

"Related by inheritance" may be a bit too weak.  More specifically:
when class A(B, C, D): ... the constraint is that type(A) is a 
subclass of each of type(B) and type(C).  E.g., try this:

class meta1(type): pass
class C1: __metaclass__ = meta1

class meta2(type): pass
class C2: __metaclass__ = meta2

try:
    print 'C1, C2',
    class C3(C1, C2): pass
except TypeError:
    print 'no'
else:
    print 'yes'

class meta3(meta2, meta1): pass
class C3(C1, C2): __metaclass__ = meta3
print "with metaclass MI, though, it's fine!"


But in some cases, yes, weaker inheritance constraints between
meta-classes can suffice.



> making metaIC a subclass of metaMetaBunch (and changing the __new__ call
> appropriately), gives you:
> 
> Traceback (most recent call last):
>   File "M:\python\cpoint.py", line 38, in ?
>     class CountedPoint(Counted, Point):
>   File "M:\python\cpoint.py", line 25, in __new__
>     return metaMetaBunch.__new__(cls, name, bases, dict)
>   File "M:\python\cpoint.py", line 20, in __new__
>     return type.__new__(cls, classname, bases, newdict)
> TypeError: multiple bases have instance lay-out conflict
> 
> which means you're stuck (?)

Well no -- you MIGHT be stuck if you ever needed two different AND
incompatible __new__ in the same class, but here metaIC may just as
well use __init__ instead (as it ain't messing with e.g. __slots__:-)
so everything works if you recode the whole mess as, e.g.:


class metaMetaBunch(type):
    def __new__(cls, classname, bases, classdict):
        def __init__(self, **kw):
            for k in self.__dflts__: setattr(self, k, self.__dflts__[k])
            for k in kw: setattr(self, k, kw[k])

        newdict = {'__slots__':[], '__dflts__':{}, '__init__':__init__}

        for k in classdict:
            if k.startswith('__'):
                newdict[k] = classdict[k]
            else:
                newdict['__slots__'].append(k)
                newdict['__dflts__'][k] = classdict[k]

        return type.__new__(cls, classname, bases, newdict)

class metaIC(metaMetaBunch):
    def __init__(cls, name, bases, dict):
        cls.InstCount = 0
    def __call__(cls, *args, **kwds):
        cls.InstCount += 1
        return type.__call__(cls, *args, **kwds)

class Counted(object):
        __metaclass__ = metaIC
        
class Point(object):
        __metaclass__ = metaMetaBunch
        x = 0
        y = 0
        
class CountedPoint(Counted, Point): pass
               
cp = CountedPoint()
print cp        


But given the dependencies I don't think this particular mixin
mesh makes much sense.  I'd rather merge two metaclasses into a
third one (solving any __new__-related conflicts of course;-)
rather than arbitrarily deriving one from the other (and it IS
arbitrary -- now you're forcing ANYTHING that's counted to be
SOME kind of metabunch...).


Alex




From exarkun at intarweb.us  Sat Apr  5 13:42:35 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sat, 5 Apr 2003 13:42:35 -0500
Subject: Q: The roles of Python
In-Reply-To: <5LEja.3680$945.10664@tor-nn1.netcom.ca>
References: <5LEja.3680$945.10664@tor-nn1.netcom.ca>
Message-ID: <20030405184235.GB23069@meson.dyndns.org>

On Sat, Apr 05, 2003 at 01:08:24PM -0800, Carl Waldbieser wrote:
> It seems obvious to me how Python can be extremely useful on the server side
> of a distributed application, but I was wondering if it has been widely used
> on the client side of such applications or in desktop applications in
> general.  Are there some good articles on the issues associated with using
> Python in this way?  If you wrote a desktop application in Python, are there
> any examples of how you would distribute Python with your application?  How
> transparent would (or could?) the process be to the end user?
> 

  This question is asked here frequently.  It is answered in the archives
and the FAQ.

> I think Python offers a lot of advantages over the other two big choices in
> the Windows scripting world (JScript and VBScript), notably the fact that it
> seems a lot easier to organize Python scripts into discrete modules.  It has
> been my experience that in big applications, the other scripting engines I
> have used tend to have problems with colliding names, and information tends
> to be passed between scripts using global variables.  The net result is that
> they tend to get very messy and hard to debug very quickly.

  Luckily, Python isn't a "scripting engine".  It's a programming language. 
It scales to large projects easily.

> 
> I think that Python scripts could be a lot easier to manage than other
> scripting enviornments, but I have not heard much about Python being used
> much behind the scenes in desktop scenarios.

  I think you're right.

> 
> Also, I would be very interested in hearing about client-side and desktop
> applications where the application was written primarily in Python rather
> than simply being the scripting engine.
> 

  I don't know what kind of applications you're thinking of, but I've
developed a couple "client-side" apps in Python that were also scriptable in
Python.  It's all pretty straightforward.  The one thing Python can't give
you (yet) that some of those "scripting engines" might offer is something
resembling a secure, restricted execution environment.

> Thanks for any insights and information in advance!
> 

  Hope this helps,

  Jp

-- 
A sad spectacle.  If they be inhabited, what a scope for misery 
and folly.  If they be not inhabited, what a waste of space.
                -- Thomas Carlyle, looking at the stars
-- 
 up 16 days, 14:00, 7 users, load average: 1.25, 1.31, 1.27




From anton at vredegoor.doge.nl  Sun Apr 20 03:23:12 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Sun, 20 Apr 2003 09:23:12 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com>
Message-ID: <b7ti62$ibg$1@news.hccnet.nl>

Courageous <jkraska at san.rr.com> wrote:

>grokked what you did. Probably Python's not the language for a
>construct where the author expresses surprise that seasoned
>programming veterans "get" his work, you know?

It's one thing to follow the adagium to keep Python as simple as
possible. It's quite another thing (in fact the opposite thing, seen
from some perspective) to throw out anything because it's too
complicated. These kinds of religious aberrations are common these
days, unfortunately. I'd be interested in an explanation of what
continuations are and what their advantages would be. It wouldn't
surprise me in the least if unexperienced programmers would have no
difficulties with them while programmers with more cognitive ballast
would have trouble integrating the new stuff into their mental
framework. 

Anton



From do-not-spam-ben.hutchings at businesswebsoftware.com  Tue Apr  1 10:18:21 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 1 Apr 2003 15:18:21 GMT
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <roy-26F31A.18263529032003@reader2.panix.com> <x78yuwloxj.fsf@guru.mired.org>
Message-ID: <slrnb8jbdu.1fs.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <x78yuwloxj.fsf at guru.mired.org>, Mike Meyer wrote:
> Roy Smith <roy at panix.com> writes:
> 
>> I just make a trivial change to a single source file in a project I'm 
>> working on and times how long it took to run make.  It took 1 minute, 41 
>> seconds.  The vast majority of that time was make itself, recursing up 
>> and down the directory tree to discover that nothing else had changed.  
>> Had the change been to a core header file, I might have been waiting for 
>> 15 minutes while make recompiled the whole world.  With Python, I would 
>> have been able to test my change in a matter of seconds.
> 
> It's fairly well know that recursive make files are slow.

They are also incapable of expressing all the dependencies you might want.

> There are tools built to solve this problem (e.g. jam), and technics for
> using include's in make files so the entire tree is processed by one
> make. Either one results in significantly faster builds.

Yes.  I came up with an include-based makefile system that would often
require make to read rules from a hundred or so files (it would read some
common files several times over with different substitutions) and to
examine a thousand or so source and dependency files.  It only took about
a second to start up, and that was on an Sun Ultra 2 (I think), which is
not a speed demon.

However, C++ compilation can be very slow because the compiler often needs
to see so many related class definitions to compile each source file.  You
can avoid a lot of that with the pimpl idiom, though.

> Of course, that's still not faster than having an interpreter. One of
> the things that attracted me to python was that it came with an
> interpreter, as opposed to having add it to Perl or Java.

It gives a huge productivity boost.

> Batteries included, indeed.



From TOttinge at indy.rr.com  Wed Apr 16 22:24:22 2003
From: TOttinge at indy.rr.com (Tim Ottinger)
Date: Thu, 17 Apr 2003 02:24:22 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <bDSma.22753$T34.687158@news2.tin.it> <slrnb9o8s8.iea.news@datatailors.com> <mailman.1050427771.31620.python-list@python.org>
Message-ID: <3E9E0FBA.6050503@indy.rr.com>

Donnal Walter wrote:
> I am a physican first and programmer second. I've never taken a course 
> in computer science, and I do not know the meaning of O(N). OTOH, just 
> from following comp.lang.python I think I have a pretty good idea why
> 
>     s += str(i)
> 
> would be a bad idea. Nevertheless, I certainly would not be opposed to 
> deepening my understanding. (PythonIAN is great, BTW.)

I came in late. If we're talking about schools, I'm not sure whether 
we're talking about colleges and adult ed, or high school, or elementary 
school. Are we talking about teaching people who know nothing about
programming?

If a 3rd grader wrote the s += str(i) or s = s+str(i) I would be pretty
impressed. I don't expect them to understand O() notation. I wouldn't
be to thrilled to see it in professional code. I don't expect that it
would make it through the PEPs. But it would be expressing the basic
idea, and then later it could be explained why it is so awfully slow
on the big loops.

Maybe not too bad for 5th or 6th grade, since they're far from writing
any professional code. You tolerate more with rank beginners, and you
correct it in time.

tim




From mwh at python.net  Thu Apr  3 10:35:35 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 3 Apr 2003 15:35:35 GMT
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com>
Message-ID: <7h3znn7k308.fsf@pc150.maths.bris.ac.uk>

drew at astro.pas.rochester.edu (Drew Moore) writes:

> Thanks for responding.. Maybe posting on April 1st was a mistake,
> people might have thought I was joking around!!

I certainly did.

Cheers,
M.

-- 
  same software, different verbosity settings (this one goes to
  eleven)                             -- the effbot on the martellibot



From mwm at mired.org  Wed Apr  2 12:45:29 2003
From: mwm at mired.org (Mike Meyer)
Date: Wed, 02 Apr 2003 17:45:29 GMT
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <just-0075D9.21243801042003@news1.news.xs4all.nl>
Message-ID: <x7of3oiyiu.fsf@guru.mired.org>

Just <just at xs4all.nl> writes:

> In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer <mwm at mired.org> 
> wrote:
> 
> >     Roman numeral literals will consist of the characters M, D, C, L,
> >     X, V and I, and only those characters.  They must be in upper
> >     case, and represent an integer with the following rules:
> 
> I *strongly* object to the upper case requirement.

But without that requirement, you get syntax errors on otherwise
correct programs.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From cben at techunix.technion.ac.il  Wed Apr  9 14:00:57 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Wed, 9 Apr 2003 21:00:57 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <b6vq8r$9qev7$1@ID-169208.news.dfncis.de>
References: <m2u1dcw22s.fsf@localhost.localdomain>       
 <3E9040CF.2040606@removethis.free.fr>        <m2k7e7addl.fsf@localhost.localdomain>
        <b6pgcm$1t3$1@panix3.panix.com>        <m27ka7aae0.fsf@localhost.localdomain>
        <6ee58e07.0304061722.2472b3d0@posting.google.com>       
 <3E9175A5.1060607@kyborg.dk>        <yfsof3ijmua.fsf@black132.ex.ac.uk>   
     <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org>
 <b6vq8r$9qev7$1@ID-169208.news.dfncis.de>
Message-ID: <Pine.GSO.4.51.0304092053190.1222@techunix.technion.ac.il>

Greg Ewing (using news.cis.dfn.de) wrote on 2003-04-09:

> Skip Montanaro wrote:
> > Look again.  In Elisp, when foo() looks up the value of case-fold-search it
> > will find the value defined in the let statement inside bar() because it's
> > lexically scoped  (looks up the call chain for values).
>
> Your terminology is confused -- that's *dynamic* scoping.
>
> > If this was Python, the reference to case-fold-search in foo()
> > would be resolved at the global scope because Python is statically
> > scoped (looks in locals, globals, then builtins only
>
> *This* is called lexical scoping (since the scope is determined
> by where the code is written). (Although calling it "static"
> is correct too, since it's not dynamic.)
>
To clarify the issue completely, the emacs lisp manual contains a
must-read section:

http://www.gnu.org/manual/elisp-manual-21-2.8/html_node/elisp_147.html#SEC147

Note the difference between scope and extent and how the confusing "X
scoping" terminology maps to it.

It also describes some implementation details - not that I believe we
want dynamic scope...  Note that the listed motivation is mainly
implemention simplicity and effeciency.  I read somewhere that this is
no longer true but for a long time dynamic scope was considered faster
even for compiled code because so much effort of LISP hackers went
into optimizing it in clever ways :-)...

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From skip at pobox.com  Thu Apr 17 20:25:25 2003
From: skip at pobox.com (Skip Montanaro)
Date: Thu, 17 Apr 2003 19:25:25 -0500
Subject: Python "meetup" in the Chicago area - May 8 7pm
Message-ID: <16031.17909.876817.764198@montanaro.dyndns.org>

Chicago (and other) area Pythonistas might be interested in getting together
via meetup.com.  The next Chicago meetup is tentatively scheduled for May
8th at 7pm.  You can sign up and express your preference for location here:

    http://python.meetup.com/

Skip




From duncan at NOSPAMrcp.co.uk  Wed Apr 30 04:34:09 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 30 Apr 2003 08:34:09 +0000 (UTC)
Subject: Pyrex Hard Case
References: <3EAD2D06.7010505@elo.utfsm.cl>
Message-ID: <Xns936D611B4747Cduncanrcpcouk@127.0.0.1>

Rodrigo Benenson <rodrigob at elo.utfsm.cl> wrote in
news:3EAD2D06.7010505 at elo.utfsm.cl: 

> I had allready done that. This command create an
> 
> extern MACRO_NAME(EXPECTED_ARGUMENTS)
> 
> Then macro name is remplaced by something like
> 
> extern double * C2F(stack).STK + x - 1 (EXPECTED_ARGUMENTS);
> 
> and the C compile is missed, this is the line I have comment via hand
> edit. 
> 
> rodrigob.
> 

I think you may need to cut this down to a minimal example that shows your 
problem and post it. Here is a minimal example that shows that how to use 
macros from Pyrex:

---- myhdr.h ----
#include <stdio.h>
#define PRINT(e) printf("%s", e)
---- test.pyx ----
cdef extern from "myhdr.h":
    void PRINT(char *EXPECTED_ARGUMENT)

def test(char *x):
    PRINT(x)
---- end ----

This compiles and from Python you can:

    import test
    test.test("Hello world")

which has the expected effect (printing the string).

N.B. You have to declare the variable types and return type when declaring 
the macro within Pyrex, otherwise it will try to pass Python objects as 
parameters and expect an object as a result.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From python at dylanreinhardt.com  Fri Apr 25 15:24:25 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 25 Apr 2003 12:24:25 -0700
Subject: Zope-Dev: Curious Problem
In-Reply-To: <45jiavkpsr5kl41u0ajs2jud1o0qd7rm0d@4ax.com>
References: <45jiavkpsr5kl41u0ajs2jud1o0qd7rm0d@4ax.com>
Message-ID: <1051298664.6009.26.camel@ida.dylanreinhardt.com>

[ already answered on the Zope list ]


On Fri, 2003-04-25 at 07:58,  wrote:
> Hi,
> 
> I have a python external method which reads from an external database,
> unpickles a class object, and concatenates some of the object's
> properties into a string which is returned by the external method.
> The class definition is included in the module.
> 
> When I run it from the linux prompt in the Extensions directory it
> behaves exactly as it should. The class is successfully unpickled, the
> properties are read and the string is returned.
> 
> However, when I set it up as an external method and call it I get the
> infamous:
> 
> Error Type: SystemError
> Error Value: Failed to import class MyObject from module __main__
> 
> >From my perspective Zope should be 'blissfully unaware' of the
> existence of the class or the instantiation of the object.  Since the
> object is only accessed in the external method, the module contains
> the class definition, and only a string is returned, why is this
> happening?
> 
> What am I missing here?
> 
> Platform: Red Hat 7.3, Python 2.1.3, Zope 2.51
> 
> Curious,
> 
> Sean
> 
> P.S. Also posted to zope-dev but I wanted the 'python' perspective as
> well





From andy at reportlab.com  Thu Apr 17 16:38:50 2003
From: andy at reportlab.com (Andy Robinson)
Date: Thu, 17 Apr 2003 21:38:50 +0100
Subject: Pythonic way of web-programming
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <B8Pma.20554$T34.656139@news2.tin.it> <mailman.1050471334.22000.python-list@python.org>
Message-ID: <bh3u9vgjof7ke1r7fvlr97ee8c38cl4cfv@4ax.com>

On 16 Apr 2003 00:35:05 -0500, Ian Bicking <ianb at colorstudy.com>
wrote:

>On Tue, 2003-04-15 at 03:11, Alex Martelli wrote:
>> However -- look at the XML situation for contrast.  There IS just such
>> a standard (and parts of it ARE in the standard libraries, while others,
>> PyXML stuff, integrate as seamlessly as they can with those via some
>> import tricks).  AND YET -- you guys have developed and maintain PyRXP,
>> which uses substantially different interfaces -- *AND WELL IT DOES*,
>> because thanks to the different interfaces (trees of bare tuples, or
>> lists as the case may be, rather than intricate objects as in DOM) it
>> can save a LOT of memory, and generally zip along much, much faster.
>> 
>> Aren't there analogies here...?  

I m having a little trouble following this as my newsreader does not
show my own post and the first day of replies (hardly a surprise
with BlueYonder), but I will try to infer what went between.

I did gripe a lot about Python's XML tools up to about 2 years ago,
but the main driver was the general lack of how-to documentation.
This problem has been very well addressed in PyXML and in
several good books, and Python does have a clear standard now.

We went our own way for what I hope are very clearly stated
reasons on the pyRXP page.  Also, RXP is GPL, which killed
any plans to try and get it into the Python distro - I tried
to get that changed but could not.  But I do have a "roadmap" to 
bring that back into the fold one day.
1. The "tuple tree" structure and a standard getattr-based DOM-like
wrapper for it (whih I know Alex has been playing with) could be in
the Python XML package or even the standard library
2. A modest change to pyexpat would let it be generated at C-like
speeds without pyRXP
3. If you can bear GPL - or pay us :-) - you can use pyRXP for
validation.  But you won't really need it to get the speed of our
approach.


Finding the time to execute the plan is of course much harder - it
might happen if some huge corporate swallows up ReportLab and 
pays us to write cool code, but we now have a "solved problem"
so it's hard to really get wound up for it.  

I'll join in with the rest when i find the intervening
2 days somewhere :-)

- Andy



From mis6 at pitt.edu  Mon Apr 28 17:36:26 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 14:36:26 -0700
Subject: a new pythonista?
References: <pan.2003.04.26.04.22.16.562041@mn.rr.com> <7xvqa.4236$b67.1131392@read2.cgocable.net> <slrnbaqtfe.ao.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <2259b0e2.0304281336.413e9b55@posting.google.com>

Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> wrote in message news:<slrnbaqtfe.ao.do-not-spam-ben.hutchings at tin.bwsint.com>...
> I wrote my first programs when I was 6, in BASIC.  If I remember correctly,
> they converted temperatures in degrees celsius to degrees fahrenheit, and
> vice versa.  If Python had existed then, I might have been able to write
> something interesting and wouldn't have got stuck in that stupid language
> for so long.

By a strange coincidence, my first program in Python converted
fahrenheit in celsius; but even last year I was somewhat older than six ;)
Still, it is strange to think that I would not have discovered Python
if Pittsburgh had a better whether ...

                                Michele



From patrickdlogan at attbi.com  Wed Apr 23 01:42:26 2003
From: patrickdlogan at attbi.com (Patrick Logan)
Date: 22 Apr 2003 22:42:26 -0700
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050569128.32513.python-list@python.org>
Message-ID: <27cc4d4a.0304222142.199ba7c9@posting.google.com>

Christian Tismer <tismer at tismer.com> wrote:

> Let me simply end this pamphlete with some simple sentences:
> Stackless Python is more capable of tasklets switching than any
> other light-weight threading software package.
> If anyone disagrees, please give me a runnable counter-example.

This would match the results seen in other implements of lightweight
threads, e.g. in Erlang, PLT Scheme, Gambit Scheme, etc.

So I can believe a similar implementation of Python would achieve
similar results.

-Patrick



From theller at python.net  Thu Apr 17 13:47:19 2003
From: theller at python.net (Thomas Heller)
Date: 17 Apr 2003 19:47:19 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <d6jmqr2q.fsf@python.net> <Xns936072232AA84duncanrcpcouk@127.0.0.1> <he8xgr3g.fsf@python.net> <Xns9360A249561CBduncanrcpcouk@127.0.0.1> <Xns9360A663C7BA5duncanrcpcouk@127.0.0.1> <znmpf64d.fsf@python.net>
Message-ID: <d6jlf23c.fsf@python.net>

Thomas Heller <theller at python.net> writes:

> Duncan Booth <duncan at NOSPAMrcp.co.uk> writes:
> 
> > > The difference code could actually be made much faster quite
> > > easily.  If I added a search for the first difference, then I
> > > could do the ellipsis and truncating first and only pass what is
> > > left to difflib.
> > 
> > Old version took 0.18 seconds on Python 2.3, gave up waiting on
> > Python 2.2.  This version takes 0.18 on Python 2.2, 0.08 on Python
> > 2.3.
> 
> Cool! I can now use 10 million character strings, and the test only
> takes 1.5s in 2.2, and 0.7s in CVS python.
> 

I forgot to say, that your patch is even cooler.  Without it,
failUnlessEqual(a, b) tries to print the 10 million character strings
*completely*.

Would you care to upload your patch to SF, so it doesn't get lost?

Shouldn't there by a python function which limits the output of
'print obj' to *some* amount?

Thomas



From uwe.schmitt at procoders.net  Tue Apr 22 10:19:49 2003
From: uwe.schmitt at procoders.net (Uwe Schmitt)
Date: 22 Apr 2003 14:19:49 GMT
Subject: zope TAL
Message-ID: <b83j25$3namt$1@hades.rz.uni-saarland.de>

Hi,

are there any good resources concerning TAL ?
I just found a reference on zope.org but nothing more.

Greetings, Uwe.

-- 
Dr. rer. nat. Uwe Schmitt   http://www.procoders.net                      
schmitt at procoders.net      "A service to open source is a service to mankind."



From point14 at magma.ca  Sun Apr 27 02:18:38 2003
From: point14 at magma.ca (Martin d'Anjou)
Date: Sun, 27 Apr 2003 06:18:38 GMT
Subject: Looping around a simple list of tuples
References: <3EAB7213.4030209@magma.ca> <mailman.1051419821.24369.python-list@python.org>
Message-ID: <3EAB87F7.9090306@magma.ca>

Andrew Bennetts wrote:
> Which could be spelt out as:
> 
>     for el in a:
>         r, l = el
>         print "R "+`r`
>         print "L "+`l`
> 
> Does this make sense?

Yes, now I got it. el is a tuple. My second loop did not make sense. Thanks!

Martin




From aleax at aleax.it  Sat Apr 19 15:31:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 19:31:52 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org>
Message-ID: <Iuhoa.58588$T34.1568946@news2.tin.it>

Steven Taschuk wrote:

> Quoth Alex Martelli:
>   [...]
>> "sum([y[1] for y in x])" would be very simple and direct, IMHO; the
>> "same" thing expressed with "reduce(operator.add,[y[1] for y in x])"
>> may only feel "simple and direct" to PhD's in maths -- it feels
>> _abstruse_ to most people, in my opinion.
> 
> What if operator.add were changed to take any number of arguments?
> Then you could write
>     operator.add(*[y[1] for y in x])
> Clear?  Abstruse?

That would be quite acceptable.  After all, it would then take
nothing more than a simple:

  from operator import add as sum

to get my favourite syntax sugar for the operation:-).

What would you have operator.add return when called with no
arguments, though?  Either 0 or 0.0 or '' or ... have problems.
I think it should be callable with ONE OR MORE arguments,
not with ANY number.


> Btw, I don't know about most people, but I don't have a PhD in
> math, and reduce(operator.add, L) seems simple and clear to me.
> (I do agree that there is a point beyond which abstraction brings
> diminishing returns; but reduce is well inside that limit, imho.)

Guess we'll have to agree to disagree.  I came to Python with
lots of experience disposing me to favour reduce (no PhD either,
I'm an engineer -- but, lots of APL, functional languages, etc).

Four years later, having coded a lot of Python, taught it widely,
written a lot about it, and so on, I've changed my mind: I now
think that reduce is more trouble than it's worth and Python
would be better off without it, if it was being designed from
scratch today -- it would not substantially reduce (:-) Python's 
power and WOULD substantially ease the teaching/&c task.  That's
not a strong-enough argument to REMOVE a builtin, of course, and
thus that's definitely NOT what I'm arguing for.  But I do suggest
avoiding reduce in most cases -- that's all.


Alex




From tanzer at swing.co.at  Mon Apr 14 13:01:41 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Mon, 14 Apr 2003 19:01:41 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "Mon, 14 Apr 2003 16:50:40 GMT."
             <slrnb9l507.1q1.rmunn@localhost.localdomain>
Message-ID: <E1957Kv-0003KM-00@tswings.swing.cluster>

Robin Munn <rmunn at pobox.com> wrote:

> Paul Foley <see at below.invalid> wrote:
> > I.e., write broken and fragile code just to avoid a perfectly good
> > construct that you happen to dislike for religious reasons.
> > Brilliant!
>
> How is try ... finally "broken and fragile code"? How is this:
>
>     try:
>         old_stdout = sys.stdout
>         new_stdout = cStringIO.StringIO()
>         sys.stdout = new_stdout
>         function_whose_output_I_want_to_capture()
>         do_something_with(new_stdout)
>     finally:
>         sys.stdout = old_stdout
>
> in any way inferior to this (making up a semi-Pythonic syntax):
>
>     with sys.stdout = cStringIO.StringIO():
>         function_whose_output_I_want_to_capture()
>         do_something_with(sys.stdout)
>
> Yes, sample #2 is shorter. But it would require adding a new keyword to
> the language, which is a pretty heavy price to pay. And sample #2
> doesn't need a temporary variable -- but sample #1's temporary variable
> is a function-local variable, so speed and namespace-clutter costs are
> minimal.
>
> What I really want to know is why you would consider sample #1 "broken
> and fragile code". I can't think of any circumstances where sample #1
> would fail to achieve its desired effect. Can you?

Disclaimer: I'm not arguing for adding dynamic scopes or variable
declarations to Python <wink>.

I see three problems with try/finally for temporary rebindings of
multiple variables:

- It is fragile for multiple rebindings unless you save all the old
  values before the try.

- It is easy to forget/inadvertantly delete some restore operations in
  the finally clause.

- It destroys locality of reference for the human reader. Your toy
  example needs eight try/finally lines vs. three lines for the
  alternative solution. For multiple rebindings, the ratio gets worse.

Still-greatly-preferring-Python-over-Lisp-ly yr's,

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From tismer at tismer.com  Sat Apr 19 00:03:40 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sat, 19 Apr 2003 06:03:40 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de>
Message-ID: <3EA0CA9C.2010105@tismer.com>

Martin v. L?wis wrote:
> Steven Taschuk <staschuk at telusplanet.net> writes:
> 
> 
>>>At the same time, I dismissed all of my Stackless 1.0 code, which was
>>>continuation-based, an absolute no-no in Guido's eyes. [...]
>>
>>Is there a short way to say why Guido didn't want continuations?
> 
> 
> He didn't like the implementation strategy; I believe he was neutral
> about the feature per se.

No, he was not. He believes continuations are an insane feature,
and he hates them. Said that several times in private emails.

ciao - chris

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From delgado at okeeffe-pc3.la.asu.edu  Mon Apr 14 13:16:20 2003
From: delgado at okeeffe-pc3.la.asu.edu (Olaf Delgado Friedrichs)
Date: 14 Apr 2003 17:16:20 GMT
Subject: Pyrex: wrapping #defines with a type name argument
References: <b76ivu$eri$1@bob.news.rcn.net> <b7dfi7$dguav$1@ID-169208.news.dfncis.de>
Message-ID: <slrnb9lufd.eqb.delgado@okeeffe-pc3.la.asu.edu>

Greg Ewing wrote:
>Edward C. Jones wrote:

[...]

>>     #define GET_ELEM(type,x,n) (*(type*)((char*)(x) + n))
>> 
>> What is a good way to wrap this in Pyrex (if any)?

[...]

>(2) Don't use the macro at all, and do the casting
>     yourself in Pyrex each time you need it.

Just an odd thought: couldn't one feed a Pyrex file with C macros to
the C preprocessor?

Olaf



From aleax at aleax.it  Sat Apr 26 12:42:33 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 16:42:33 GMT
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de>
Message-ID: <ZFyqa.7824$3M4.220335@news1.tin.it>

Klaus Meyer wrote:

>> I don't know what to DO about it, but some minor tweaks might help.
> 
> I also have made this mistake sometimes.
> 
> On reason, i think, is the mix of function-call and values, example:
> 
> import time
> print time.time()
> print time.timezone
> 
> You have to remember the doc of the module carefully to make no mistake.

Well, in this case it's no big deal, actually -- 

>>> print time.time
<built-in function time>

i.e. if you forget parentheses you get an immediate reminder of that.


> In this case time() must be a function, because it has to read the actual
> timer value, but must timezone be a value?

There is NO case in which something "HAS" to be a value -- it's just
handier and faster and neater when it CAN be.

> My question about style:
> Should a Module/Class etc. return values, if possible, or better use
> function to return values?

If arguments are needed, then of course functions / methods.

If arguments *aren't* needed, then a class instance gives you more
freedom of choice, as you can define *properties* -- user code
syntactically accesses a property without any overt function call,
but "under the covers" Python calls a method you have designated
to be the "getter" when buildint the property.  The ability to start
with a simple attribute access -- and turn it into a getter-call at
any time that may be needed, without any need for changes in code
using it, is really crucial, and a good motivation to NEVER design
"argumentless getter methods that just access what's really and
conceptually a simple stored attribute value" into your classes.


> If PyChecker could check such cases this would be very helpful.

When you're actually DOING something with the function or method object,
even if what you're doing is not *calling* them, it's basically out of the
question -- apparently not even testing truth value can be removed, due
to the "x=getattr(someobject, attrname, None)" usage followed by testing
just for truth value ("if x:") rather than for None-hood ("if x is None:").

A checker that gives lots of false positives soon becomes worse than
useless, makes you waste more time than it gains you, and withers.

So, I think your best bet is to apply the metaclass I already posted,
selectively, to modules you're writing, and hopefully quite transiently
as you proceed to learn that "to call callables, you need to CALL them"
and not just MENTION them.


Alex




From rjgruet at yahoo.com  Fri Apr 11 16:43:37 2003
From: rjgruet at yahoo.com (Richard Gruet)
Date: Fri, 11 Apr 2003 22:43:37 +0200
Subject: ANN: Python Quick Reference 2.2 available.
Message-ID: <b779f9$inh$1@aphrodite.grec.isp.9tel.net>

Hi all

Finally I found enough courage to update the Quick Reference  for Python 2.2, shortly before the 2.3 is released ;-)
It is available in HTML, zipped HTML and pdf at http://rgruet.free.fr/ , http://www.brunningonline.net/simon/python/PQR.html, and probably soon at http://www.vex.net/parnassus/.

Basically, it is a long single HTML page (about 40 printed pages) that does... what one usually expect from a quick reference for the Python language, i.e. summarize a maximum of things in a (relatively) short space.

Please report to me (rjgruet.yahoo.com) any error, innaccuracies, and comments/suggestions.

Cheers

Richard Gruet.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030411/9f452e60/attachment.html>

From aleax at aleax.it  Wed Apr 16 04:18:17 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 08:18:17 GMT
Subject: Pythonic way of web-programming
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <B8Pma.20554$T34.656139@news2.tin.it> <mailman.1050471334.22000.python-list@python.org>
Message-ID: <dl8na.20637$LB6.515421@news1.tin.it>

Ian Bicking wrote:

> On Tue, 2003-04-15 at 03:11, Alex Martelli wrote:
>> However -- look at the XML situation for contrast.  There IS just such
>> a standard (and parts of it ARE in the standard libraries, while others,
>> PyXML stuff, integrate as seamlessly as they can with those via some
>> import tricks).  AND YET -- you guys have developed and maintain PyRXP,
>> which uses substantially different interfaces -- *AND WELL IT DOES*,
>> because thanks to the different interfaces (trees of bare tuples, or
>> lists as the case may be, rather than intricate objects as in DOM) it
>> can save a LOT of memory, and generally zip along much, much faster.
>> 
>> Aren't there analogies here...?
> 
> There may be compelling differences in some areas.  For instance, a
> template that is DOM-like (i.e., externally controlled) vs. a template
> that is self-evaluating, like Cheetah.  But among self-evaluating
> templates, the interface seems like it should be possible to unify.

Possible, yes.  Feasible, I dunno.  Compare Cheetah and preppy, which
both process templates to produce Python modules.  preppy does it in
a very simple, rather Pythonic way -- inserting "directives" enclosed
in {{...}} that are Python expressions or a few Python statements
modified to have explicit endif/endfor/&c termination rather than
relying on indentation.  Any Python programmer should have no problem
understanding preppy in 10 minutes and then happily using e.g.
    {{myobject.func()}}
or
    {{myobject['index']}}
or
    {{myobject.attr}}
&c &c in a fully appropriate manner -- since by definition a Python
programmer already understands perfectly the difference between
calling a function (and KNOWS you need empty parentheses when you
call a function without arguments -- something that keeps tripping
beginners up, btw), indexing a container, accessing an attribute.

But suppose your design intention is different -- you want your
templates to be maintained by NON-programmers, people which DO
have problems understanding the differences between the above
distinct concepts!  In Cheetah, the above constructs would be
    $myobject.func
    $myobject.index
    $myobject.attr
i.e., totally polymorphic.  The Cheetah designer, on the basis of
a LOT of experience trying to have non-programmers maintain
templates, believe this extreme polymorphism makes things VERY
much easier for non-programmers (perhaps particularly ones that
have some previous exposure to languageoids such as Javascript
and VBScript that may also try to hide such differences...).

So who's right?  Quite possibly both -=- a programmer will find it
easier and more productive to use and maintain preppy templates,
closer to Python, and a non-programmer Cheetah templates, which
requires him or her to learn fewer concepts (and do NOT require
empty parentheses for argumentless function calls, in particular).


> There are deep differences and shallow ones.  It would be really nice to
> get rid of the shallow ones.

Personally I agree that the difference between {{x.y()}} and $x.y is
quite shallow.  But I don't know if it's best gotten rid of -- I _have_
seen nonprogrammers (and even programmers with e.g. a Pascal background)
wrestle with forgetting the () over and over, yet as a programmer I FAR
prefer the clarity of the distinctions drawn by preppy (and Python)
for programming use...


Alex




From piet at cs.uu.nl  Wed Apr 30 10:37:41 2003
From: piet at cs.uu.nl (Piet van Oostrum)
Date: 30 Apr 2003 16:37:41 +0200
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3ea7b12a@news.012.net.il> <7h3k7dkkwko.fsf@pc150.maths.bris.ac.uk> <Rpu197AqP9p+Ew5w@jessikat.demon.co.uk> <87vfx28vz8.fsf@pobox.com>
Message-ID: <wzu1cgnj9m.fsf@nono.cs.uu.nl>

>>>>> jjl at pobox.com (John J. Lee) (JJL) wrote:

JJL> The worst thing about Mac vs. PC keyboards is that Mac keyboards have
JJL> the home-key bumps on d and k, whereas PC keyboards have them on f and
JJL> j.  This really messes up my touch-typing.  I have no doubt I'd get
JJL> used to it eventually, though.

My iBook has them on F and J. 
-- 
Piet van Oostrum <piet at cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.van.Oostrum at hccnet.nl



From danb_83 at yahoo.com  Fri Apr 11 17:49:39 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 11 Apr 2003 14:49:39 -0700
Subject: Loss of precision / data corruption with Tkinter.DoubleVar ?!
References: <f70e3538.0304111007.b0a72c3@posting.google.com>
Message-ID: <ad052e5c.0304111349.1a007b28@posting.google.com>

fnord at u.washington.edu (Lonnie Princehouse) wrote in message news:<f70e3538.0304111007.b0a72c3 at posting.google.com>...
> Greetings, 
> 
> I stumbled across this curiosity yesterday:
> 
> >>> x = 0.33333333333333098
> >>> x
>  0.33333333333333098     
> >>> import Tkinter
> >>> y = Tkinter.DoubleVar(Tkinter.Tk())
> >>> y.set(x)
> >>> y.get()
> 0.33333333333300003  
> 
> What the....?
> 
> Sure, it's only on the order of 10e-13 delta, 
> but why is there any difference at all??

It looks like Tkinter.DoubleVar stores numbers in decimal floating
point (with 12 digits of precision) instead of binary like it ought
to.  Notice that you do get the correct result when x is (as close as
possible to) a power of 0.1.



From peter at engcorp.com  Wed Apr  9 14:36:50 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 09 Apr 2003 14:36:50 -0400
Subject: wxPython - Can you explain why I crash
References: <33803989.0304090652.174ac9de@posting.google.com>
Message-ID: <3E946842.5D897F5@engcorp.com>

Miki Tebeka wrote:
> 
> I've written this little client. It used to work fine but now no
> matter which button I press the application crash.
> I'm running Python2.2 and wxPython 2.4.0.7
> Any ideas?

[snip 480 lines of code]

I have an idea: spend a little of your own time cutting down the
program to a *much* smaller example which still crashes.  Then,
and only then, post it to this newsgroup and ask for others' help.

-Peter



From peter at engcorp.com  Tue Apr 15 20:42:01 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 15 Apr 2003 20:42:01 -0400
Subject: Complexity of standard Python data structures
References: <mailman.1050354988.8671.python-list@python.org> <tdnk7dwlfda.fsf@shell01.TheWorld.com> <A6Uma.99232$It5.2059@news2.central.cox.net> <b7i6gl$172lm$1@ID-169208.news.dfncis.de>
Message-ID: <3E9CA6D9.36C55DF3@engcorp.com>

"Greg Ewing (using news.cis.dfn.de)" wrote:
> 
> He forgot to mention that you need the new experimental
> quantum de-interference nails being prototyped by the PSU,
> which are capable of penetrating normal matter without
> otherwise disturbing it. Unfortunately, it's very difficult
> to obtain advance samples of these, and even mentioning
> their existence in a public newsgroup tends to result in
> the PSU cutting off your

Oh, I'm sorry, but this is "abuse".  Capons are in thread 12A, 
just back a few messages...

-Peter



From piet at cs.uu.nl  Tue Apr  8 07:07:57 2003
From: piet at cs.uu.nl (Piet van Oostrum)
Date: 08 Apr 2003 13:07:57 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <6ee58e07.0304061144.35c5ab92@posting.google.com> <m2k7e78j5s.fsf@localhost.localdomain> <mailman.1049662570.29853.python-list@python.org> <yfsr88fkppr.fsf@black132.ex.ac.uk> <mailman.1049687898.18892.python-list@python.org> <b6tcb5$8p1s0$1@ID-169208.news.dfncis.de> <mailman.1049771354.18382.python-list@python.org>
Message-ID: <wzznn1s0vm.fsf@nono.cs.uu.nl>

>>>>> Ian Bicking <ianb at colorstudy.com> (IB) wrote:

IB> On Mon, 2003-04-07 at 21:41, Greg Ewing (using news.cis.dfn.de) wrote:
>> Ian Bicking wrote:
>> > Given a novel form of input device, perhaps there's something to what
>> > he's trying to do.  I'm skeptical, though.
>> 
>> Foot pedals. Hold down the Search pedal and type away.
>> 
>> Maybe one of those car-racing-game accessories could
>> be used to prototype this...?

IB> There are foot pedals that map to things like Shift, for use by people
IB> with RSI.  Little expensive ($50?  Maybe $100...), but easy to get
IB> started with otherwise.

Of course `foot pedal' is doublespeak. `pedal' implies `foot'.
That said, an organ pedal would give you a tremendous amount of control :=)
-- 
Piet van Oostrum <piet at cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.van.Oostrum at hccnet.nl



From bvdpoel at kootenay.com  Tue Apr 22 19:50:04 2003
From: bvdpoel at kootenay.com (Bob van der Poel)
Date: 22 Apr 2003 16:50:04 -0700
Subject: Locale settings
References: <3ea437a1_1@dns.sd54.bc.ca> <b81k5m$822$06$1@news.t-online.com>
Message-ID: <fc3b2087.0304221550.5ebdf985@posting.google.com>

"Martin v. L?wis" <martin at v.loewis.de> wrote in message news:<b81k5m$822$06$1 at news.t-online.com>...
> Bob van der Poel wrote:
> 
> > Now, for the most part this works. However, one user of my program has 
> > just upgraded to Linux Mandrake 9.1 and (I think) he has set his 
> > language to "de". In any event, the getdefaultlocal() seems to return 
> > 'de' which isn't a valid python encoding (??) and the above code then 
> > converts to his 'ascii'.
> 
> Using getdefaultlocale() is broken, and cannot be fixed. On Unix, use
> 
> locale.nl_langinfo(locale.CODESET)
> 
> instead; if supported, this will tell you the user's preferred encoding.
> Be aware that
> a) neither nl_langinfo nor CODESET may be supported by the system,
> b) they may return a codeset for which a Python codec is not available,
> c) you may have to invoke setlocale(locale.LC_CTYPE, "") before
>     nl_langinfo returns a reasonable response (e.g. on Solaris)
> 
> Python 2.3 has a function locale.getpreferredencoding which does all
> that. I recommend you copy its source code into your application, and
> use that instead.

Thanks for the reply. I guess I'll be downloading 2.3 earlier than I'd
planned on. I really do like to get past the beta versions :) But, if
I can snarf the local stuff out of it and have it work with my 2.2
code, that'd be fine. Guess I  can use the included version for awhile
and then, once 2.3 become common, just tell users to use that python
version.

The funny think about all this is that my user didn't have any
problems when he was using Mandrake Linux 9.0. Just the upgrade to 9.1
which caused the problem. Which is why I was thinking that he'd not
set the locale stuff up properly at his end.

Thanks.



From ianb at colorstudy.com  Tue Apr 29 18:19:37 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 29 Apr 2003 17:19:37 -0500
Subject: Web programming and a different "type" problem
In-Reply-To: <m3fzo1v561.fsf@ferret.object-craft.com.au>
References: <mailman.1051532982.25352.python-list@python.org>
	 <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com>
	 <slrnbarhfb.67t.jon+usenet@snowy.squish.net>
	 <b8lj2q$l14$1@newshost.accu.uu.nl>
	 <m3fzo1v561.fsf@ferret.object-craft.com.au>
Message-ID: <1051654776.675.3767.camel@lothlorien>

On Tue, 2003-04-29 at 07:50, Dave Cole wrote:
> Albatross solves this yet another way.  As input tags are processed by
> the template interpreter the form tag accumulates a summary of all the
> input fields.  When the form is closed the summary is then pickled,
> MD5 signed with a secret, compressed, and base64 encoded in a hidden
> field.

You mean, it takes the expected names?  Like a form:

Object 1: <input type=checkbox name="id*" value=1>
...
Object 4: <input type=checkbox name="id*" value=4>

You create a hash of (in this small form) just "id*", and so confirm the
user isn't sending extra fields, or changing field names?

Seems a little odd... or are you doing something else?  There's always
the issue of the security of hidden fields, but that's kind of a
different issue.

  Ian






From aussie2010 at yahoo.com  Sat Apr 12 07:25:00 2003
From: aussie2010 at yahoo.com (Chris)
Date: 12 Apr 2003 04:25:00 -0700
Subject: split encloser
References: <435d40e0.0304031530.1c197742@posting.google.com> <c76ff6fc.0304040038.4ca3d62f@posting.google.com>
Message-ID: <435d40e0.0304120325.ec8b99a@posting.google.com>

sjmachin at lexicon.net (John Machin) wrote in message 
> 
> Aussie cobber,
> 
> REs are about as useful as a hip pocket on a singlet when it comes to
> data that includes the "encloser" as well. For example in a database,
> yer old Auntie Mabel's address_line_1 might be
> 
>    "Emoh Ruo", 123 Smith St
> 
> but would appear as 
> 
>    """Emoh Ruo"", 123 Smith St"
> 
> in an exported file (or it should).
> 
> Get onta Dave Cole's you-beaut CSV module at:
> 
> http://www.object-craft.com.au/projects/csv/
> 
> Cheerio,
> John

John,
Thanks for the info. Very slick. Nice that it handles embedded
carriage returns, that's something that Outlook outputs in it's CSV if
you have them in your notes/comments field,and the other Personal
Information Managers don't seem to be able to handle when you try and
import that CSV file. They could use that module.

I tried to do the full install, but it didn't work for some reason
with OpenWatcom (the only C compiler I have installed):
-----------------
D:\Downloads\Python\csv\csv-1.0>d:\python22\python setup.py install
running install
running build
running build_ext
building 'csv' extension
d:\watcom\BINNT\cl.exe /c /nologo /Ox /MD /W3 /GX
-Id:\python22\include /Tccsv.c
 /Fobuild\temp.win32-2.2\Release\csv.obj
csv.c
d:\python22\include\pyport.h(480): Error! E1091: "LONG_BIT definition
appears wr
ong for platform (bad gcc/glibc config?)."
csv.c(106): Warning! W202: Symbol 'parse_prepend_char' has been
defined, but not
 referenced
error: command 'cl.exe' failed with exit status 255

D:\Downloads\Python\csv\csv-1.0>
---------------------------------------

I did get the Win32 binary to work (although I had to try a couple of
times to find out what directory to put it in since they don't tell
help us newbies out by listing it).



From ark at research.att.com  Fri Apr  4 16:02:46 2003
From: ark at research.att.com (Andrew Koenig)
Date: Fri, 4 Apr 2003 21:02:46 GMT
Subject: output buffering--am I missing something obvious?
Message-ID: <yu99istu9ds9.fsf@europa.research.att.com>

Can someone please explain the following behavior?  Under Python 2.2.2
and Solaris 2.8:

    >>> import sys
    >>> sys.stdout.write("foo")
    >>> sys.stdout.flush()
    >>> sys.stdout.write("bar\n")
    bar

I would have expected the call to sys.stdout.write to put "foo" into
the output buffer, and perhaps to have written it to the standard
output stream.  If the call to sys.stdout.write didn't write it, I
would have expected the call to sys.stdout.flush to have done so.

What I don't understand at all is how the second call to
sys.stdout.write should be able to write its output on the
standard output stream without the output from the first call
ever appearing.

What am I missing?

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From sdhyok at yahoo.com  Tue Apr  1 22:46:36 2003
From: sdhyok at yahoo.com (sdhyok)
Date: 1 Apr 2003 19:46:36 -0800
Subject: [Q]How to append paths in a platform-independent way?
Message-ID: <420ced91.0304011946.56a22d2a@posting.google.com>

How to do the following jobs in a platform-independent way?

path = /home/sdhyok
file = data.txt

1. Append the file name to the path => /home/sdhyok/data.txt
2. Extract only path from /home/sdhyok/data.txt => /home/sdhyok

Daehyok



From peter at engcorp.com  Wed Apr  2 16:27:14 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 16:27:14 -0500
Subject: Shadowing builtins (was Re: a pyrex-inspired for-i-from-1-to-n...)
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4>
	        <b68jti$2mj66$1@ID-169208.news.dfncis.de>
	        <RbCia.6782$9I.555309@news20.bellglobal.com>
	        <mailman.1049298845.7655.python-list@python.org>
	        <3E8B4094.F61515B3@engcorp.com>
	        <ZpHia.23$Sd.4@fed1read01>
	        <3E8B48E0.C786D192@engcorp.com> <mailman.1049316186.31356.python-list@python.org>
Message-ID: <3E8B55B2.D8D7D2F2@engcorp.com>

Skip Montanaro wrote:
> 
>     >> range = my_custom_range_function
>     >>
>     >> would be OK. While,
>     >>
>     >> some_random_module.range = my_custom_range_function
>     >>
>     >> would not.
> 
>     Peter> Hmm.... that is exactly what is needed to make this effective in
>     Peter> unit testing.  One imports the module under test, then fakes the
>     Peter> open method in that other module:
>     ...
>     Peter> I'm not sure what other approach could be as simple, elegant, and
>     Peter> effective and I worry about losing it.  Surely there are many
>     Peter> other people doing similar things who would be hurt by this.
> 
> How about calling new.module() to create a new module object from the one
> you want to test which has suitably munged globals?

Interesting idea... but I'm a little unclear on what approach we would
be taking to get all the names/references from the old module into the
new one.  Pretend I'm really sleepy and having trouble following you here
(I don't have to pretend... it's the truth).  

import new
from mock_files import mockOpen
import module_under_test as mut

# create substitute module
mungedMut = new.module('dummy')

# populate with stuff from old module
for key,item in mut.___dict__.items():
    mungedMut.__dict__[key] = item

# install mock open again
mungedMut.open = mockOpen

What I don't understand is the difference between the final line above
and what I was originally trying to do, except that this is a newly
created module instead of one that was imported.

Okay, I guess I must have missed a key point and I'm way off the 
track here.  Please clarify.

-Peter



From exarkun at intarweb.us  Mon Apr 14 14:19:31 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 14 Apr 2003 14:19:31 -0400
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <3e9afaa2$0$49106$e4fe514c@news.xs4all.nl>
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl> <3E9AF8BD.BC918E5E@engcorp.com> <3e9afaa2$0$49106$e4fe514c@news.xs4all.nl>
Message-ID: <20030414181931.GB25237@meson.dyndns.org>

On Mon, Apr 14, 2003 at 08:14:45PM +0200, Irmen de Jong wrote:
> Peter Hansen wrote:
> >Irmen de Jong wrote:
> >
> >>[offtopic]
> >>
> >>Erik Max Francis wrote:
> >>
> >>
> >>>>>>Course(lowEarthOrbit, lowMarsOrbit).transfers()
> >>>
> >>>[<ExtractionTransfer @ 0x82893ac (out of `low orbit around `Earth'')>,
> >>><HohmannTransfer @ 0x812cef4 (Hohmann from `Earth' to `Mars')>,
> >>><InsertionTransfer @ 0x82adb64 (into `low orbit around `Mars'')>]
> >>>
> >>
> >>Is this actual code? What does this do? I'm interested.
> >
> >
> >Given that Python is an "agile language", it seems very likely that
> >it must be an implementation of Robert Zubrin's XP-style "Mars Direct" 
> >program for a manned mission to Mars.  
> >
> ><0.9 wink but I wish it were lower-ly y'rs> 
> >
> >-Peter
> 
> I *was* serious...
> If it is real code I suspect it somehow determines a path for a
> free moving object in space to transfer from one orbit to another
> orbit around different bodies in the solar system.
> At least that is what a quick google for "Hohmann orbit transfer"
> tells me. And I'm still interested ;-)
> 

  Indeed, I would guess it is part of an autopilot program.

  Cool stuff.

  Jp

-- 
http://catandgirl.com/view.cgi?44
-- 
 up 25 days, 15:02, 4 users, load average: 1.19, 1.11, 1.25




From blunck at gst.com  Thu Apr 17 21:13:59 2003
From: blunck at gst.com (Christopher Blunck)
Date: 17 Apr 2003 18:13:59 -0700
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com> <mailman.1050606271.32628.python-list@python.org>
Message-ID: <1147e466.0304171713.1b325644@posting.google.com>

Steven Taschuk <staschuk at telusplanet.net> wrote in message news:<mailman.1050606271.32628.python-list at python.org>...
> Quoth Christopher Blunck:

> There is another point:  Although there are indeed lots of
> programmers who do not deserve this trust and will betray it every
> time, the feeling among Pythonistas is, I think, that (a) this is
> not sufficient justification for hamstringing the competent, and
> (b) no language feature will make an appreciable dent in the
> incompetent's ability to screw everything up.
> 
> Dealing with incompetent programmers is an organizational problem,
> not a language design problem.


I completely agree with you!

Incompetence is a human problem, requiring a human solution
(termination being a good candidate).  Management is difficult because
it often involves confrontation, which usually ends in emotional
turmoil and discord on the team.  As a result, some managers may feel
more comfortable trying to lasso an incompetent programmer (or team)
by creating coding standards, forcing the use of one particular
language, or by some form of a community review (code review).  In
doing so, the manager hopes that these "process improvements" will fix
the human problem.  Frequently it does not, and (as you so elegantly
pointed out) what ends up happening is hamstringing of the stars on
team.  That's when it's time to look for a new job.  In the end
everyone loses:  the team looses star players and the star players
have to look for a new job, which usually is not a fun process.


-c



From gh at ghaering.de  Thu Apr 17 15:48:04 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Thu, 17 Apr 2003 21:48:04 +0200
Subject: Downloading TV listings with urllib
In-Reply-To: <E196EHo-0003Gb-00@swan.mail.pas.earthlink.net>
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4>
	<mailman.1050602375.2763.python-list@python.org>
	<E196EHo-0003Gb-00@swan.mail.pas.earthlink.net>
Message-ID: <3E9F04F4.3070300@ghaering.de>

Josh wrote:
> On 17 Apr 2003, you wrote in comp.lang.python:
>>Josh wrote: [...]
>>>        h.putheader('Accept-Encoding:',
>>>        'gzip,deflate,compress;q=0.9') [...]
>>>when I use err.read() to read the object returned I get everything in
>>>Hex. 
>>
>>My guess is that you get what you're asking for - gzip-compressed
>>data. 
> 
>  Well I don't want gzipped data. I want HTML so I can parse it and capture 
> the listings table into a list or something. Any ideas?

Don't tell the webserver that you'll accept gzip-compressed data?

I. e. remove the line I quoted above.

-- Gerhard




From mads at troest.NEVERMORE.dk  Sun Apr 13 15:24:24 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Sun, 13 Apr 2003 21:24:24 +0200
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com>
Message-ID: <18lb3o2cs1se2$.isoixa5igmj6$.dlg@40tude.net>

On 13 Apr 2003 11:58:36 -0700, srijit at yahoo.com wrote:

> Or may be it is possible to emulated Protected Methods in Python.

It can be "emulated" by prepending the method/attribute name with 2 
underscores. This will make Python mangle the name so noone accidentally 
can access it. Evil minds can still do whatever they want, but then again - 
it's their own responsibility if they break things then. ;)

I also feel it's weird that Python doesn't have a truly enforced 
encapsulation scheme, but one can live with it.

HTH,
/\/\\ads Orbesen Troest



From artyprog at wanadoo.fr  Thu Apr 17 06:22:34 2003
From: artyprog at wanadoo.fr (Salvatore)
Date: Thu, 17 Apr 2003 12:22:34 +0200
Subject: variable in Python
Message-ID: <b7lvbh$an3$1@news-reader13.wanadoo.fr>

Hello,

I have a little doubt.
when you declare x = 3
does the variable x 'contains' the value 3 (like in Pascal)
or does it contains an address which points to
a memory place containing the value 3 ?

Regards

Salvatore





From FBatista at uniFON.com.ar  Tue Apr 29 12:18:38 2003
From: FBatista at uniFON.com.ar (Batista, Facundo)
Date: Tue, 29 Apr 2003 13:18:38 -0300
Subject: import problem
Message-ID: <A128D751272CD411BC9200508BC2194D01885457@escpl.tcp.com.ar>

#- i am having problems importing a module in my python script.  i have 
#- the following two import statements in the script:
#- 
#- from WVWService.util import XMLUtil
#- from WVWService.test import HTTPClient
#- 
#- the first import works just fine.  but, the second one i get 
#- the error 
#- "ImportError: No module named test".  i have checked for 
#- mispellings;  
#- i have checked to see if the directory test exists; i have 
#- checked the 
#- system path by printing out sys.path from w/i the script to 
#- see if the 
#- directory path is in sys.path 
#- (/home/vjb/workspace/wvwBase3/src/wvw).  

Check if you have a __init__.py file in the directory (you must).

Facundo.





From mike at bindkey.com  Fri Apr  4 17:20:42 2003
From: mike at bindkey.com (Mike Rovner)
Date: Fri, 4 Apr 2003 14:20:42 -0800
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com>
Message-ID: <b6l0eq$tf0$1@main.gmane.org>

FWIW, I decided for myself:

- wrap small C libs with Pyrex,
- wrap big C libs with SWIG,
- wrap small C++ libs with boost.python,
- wrap big C++ with (I still have to do that ;)

Mike

PS.
For me small is tens, big is hundreds and more

"sdieselil" <sdieselil at yahoo.com> wrote in message
news:169a7283.0304040821.9ad0976 at posting.google.com...
> Hi
>
> I want to write C extension for Python. I know that there are 2 most
> popular libraris for this purpose: SWIG and Boost.Python. What are the
> differences between them? Which one is better for little extensions
> and which is better for large-scale extensions?
> --
> http://mail.python.org/mailman/listinfo/python-list
>







From staschuk at telusplanet.net  Mon Apr 21 13:19:53 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 21 Apr 2003 11:19:53 -0600
Subject: Explanation of ttx.py program
In-Reply-To: <3EA3C11B.8020609@antares.enst-bretagne.fr>; from Anish.Mehta@enst-bretagne.fr on Mon, Apr 21, 2003 at 11:59:55AM +0200
References: <3EA3C11B.8020609@antares.enst-bretagne.fr>
Message-ID: <20030421111953.C258@tibia.amotlpaa.bogus>

Quoth Mehta, Anish:
> I am interesting in making some changes in ttx.py program written in 
> python. I am attaching the file with email. Can someone explain what 
> each function is for here in this piece of code.
> 
> I will be waiting for reponse.

Tell you what: study the program yourself and come back with a
specific question when you run into trouble.

Many of us are willing to help if you try and have trouble, but
very few of us are willing to do your work for you.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From mike at bindkey.com  Mon Apr  7 17:10:30 2003
From: mike at bindkey.com (Mike Rovner)
Date: Mon, 7 Apr 2003 14:10:30 -0700
Subject: Pyrex, enumeration constants, and wrapping #define's
References: <b6soip$okk$1@bob.news.rcn.net>
Message-ID: <b6spf7$rk2$1@main.gmane.org>

"Edward C. Jones" <edcjones at erols.com> wrote in message
news:b6soip$okk$1 at bob.news.rcn.net...
> Harbison and Steele ( 4th edition, Section 5.5) say that some C
> compilers will accept an enum where some constant expression is written
> in terms of a previously defined enumeration constant:
>
>      enum xx {
>          a = 2,
>          b = a + 3
>      }
>
> Does Pyrex accept the following?
>
>      enum
>          a = 2
>          b = a + 3
>

No ;)
Use that:
cdef enum xx:    a = 2    b = a+3AFAIR, Pyrex take expressions in enums
since v0.4

> If not, how should I wrap this?
>
>      #define IPL_DEPTH_SIGN 0x80000000
>      #define IPL_DEPTH_8S  (IPL_DEPTH_SIGN| 8)


IPL_DEPTH_SIGN=0x80000000
IPL_DEPTH_8S=(IPL_DEPTH_SIGN| 8)

Mike







From mwh at python.net  Wed Apr 16 12:10:21 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 16:10:21 GMT
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com>
Message-ID: <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk>

Peter Hansen <peter at engcorp.com> writes:

> Michael Hudson wrote:
> > 
> > Peter Hansen <peter at engcorp.com> writes:
> > > One of the most important things to learn as a programmer is this,
> > > as stated by various people but most popularly attributed to Knuth:
> > > "Premature optimization is the root of all evil in programming."
> > 
> > And Knuth really did say it!  In "Structured Programming with goto
> > Statements", on page 268 of the rather naff scan[*] I have of it, we
> > find:
> > 
> >  "We /should/ forget about small efficiencies, say about 97% of the time:
> >  premature optimization is the root of all evil."
> > 
> > So even you have managed to misquote him, unless he (or someone else)
> > said it differently somewhere else.
> > 
> > [*] you can get your own naff scan easily enough with google; I can't
> >     remember where mine came from.
> 
> The last time I used Google (of which I'm well aware of course), it
> told me that Knuth was in fact merely paraphrasing someone else 
> (google can tell you who it was, if you wish ;-),

Yes, you can find lots of people claiming it was Hoare that said this.
However, their plurality doesn't necessarily make them right.

> and since I was not directly quoting him, but merely paraphrasing
> "various people" myself, I made no claims that he said precisely
> that.

I wasn't meaning to accuse you of inaccuracy, but there is *so* much
vagueness about this particular quote that I like to inject a little
precision every now and again.

> It is, after all, the sentiment that matters, more than the specific
> words used to convey them.

Indeed!

Cheers,
M.

-- 
  If design space weren't so vast, and the good solutions so small a
  portion of it, programming would be a lot easier.
                                            -- maney, comp.lang.python



From aleax at aleax.it  Sat Apr 12 12:31:34 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 16:31:34 GMT
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com> <b7726n$8pq$0@216.39.172.122>
Message-ID: <GbXla.6790$T34.183733@news2.tin.it>

Bengt Richter wrote:
   ...
>>The problem: I need to design a "decision criterion" to classify
>>"observations".  For each observation I measure the values of a
>>number N of features, x1, x2, ... xN; the desired design criterion
>>is a set of weights w1, w2, ... wN such that for any observation I
>>will then just compute a weighted sum
>>  S = w1*x1 + w2*x2 + ... + wN*xN
>>and classify the observation as Black if S<=1, White if S>1.  To
   ...
> Well, that was my reaction, FWIW. I'm sure I'm not the only one curious
> as to what you are measuring and classifying ;-)

Fair enough -- I owe it to the group to clarify that (AND collective
thanks for the huge numbers of answers, both here and privately --
"logistic regression" is the name I wanted but couldn't remember,
but many other suggestions such as "perceptron" provide fertile
criteris for other studies).

What I'm doing is, now that I (theoretically) have SOME spare time
again (no book projects actively proceeding), going back to my real
passion -- research on the theoretical fundamentals of the game of
contract bridge.  In bridge, every one of the four players is dealt
13 cards out of a 52-cards deck; a phase of "bidding" follows, where 
pairs of players playing as partners try to determine what "contract"
to bid for (if any).  A key issue during the bidding phase is to
evaluate hands' "strength" -- trick-taking potential.  As I mention
in my "short bio" whenever space allows, out of all I've accomplished
in my life, the single achievement of which I'm proudest is the pair of
articles published in the Jan-Fen 2000 issues of "The Bridge World",
under the title of "How Shape Influences Strength", demonstrating how
to use a computer to practically apply an evaluation method described
as an ideal, abstract approach in the 1920's -- and how some results
of such application prove assertions about hand strengths that were
also first made in print in the 1920's and disprove some others.  (If
anybody manages to look at those articles -- no need to tell me that
the variance-evaluation I use in them is wrong -- far too pessimistic,
so I could have gotten away with samples about 1/10th as large to
reach the thousandth-of-a-trick accuracy I required... a note pointing
that out was already published in "American Statistician", providing
ample opportunities for crow-eating for both me, and the Bridge World
editor, whose Ph.D in maths should have helped him see that...;-).

Note that I'm NOT trying to get a computer to play bridge -- I could
care less about that, and several researchers and commercial groups
are dealing with the practical issues around that goal -- MY goal is
to gain "a measure of light" as to how bridge *actually works*, maybe
to help people actually evaluate their hands better at the table, and/or
design better systems and approaches for the task; computers are tools
for the purpose, from my POV.


But anyway, back to my request.  By far the most popular practical
method for hand evaluation is known as "point-counting" -- a weighted
sum of the numbers of aces, kings, etc, in a hand.  And by far the
most popular "point count" (set of weights) in use is one first
published in 1915 (allegedly derived from one used in Italian Tarocco
games as early as 1320, though I think the connection's a bit weak)
which became overwhelmingly popular in the 1940's: 4-3-2-1-0 (count
"4 points" for each Ace, "3 points" for each King, etc, down to "0
points" [i.e., just ignore them] for each Ten-spot).  Almost every
theoretician knows/believes this specific point-count approach is way 
off the mark (with the signal exception of Dr. Jean Rene Vernes, of
France, maybe THE major theoretician of the application of statistics
to contract bridge theory, who claims 4-3-2-1-0 is "good enough" and
tries to support that claim with data which, I believe, show that
things are, in fact, otherwise).  So, I want to see what I can find
out about the issue, mostly by working with a million-deals library
of deals played at "double dummy" (i.e. ignoring the fundamental fact
that the play of the hand at bridge is a game of incomplete
information -- corrections for THAT, if any, will need to come
later) published by Dr. Ginsberg, another major theoretician in the
field (and more than a theoretician -- he ALSO runs a firm which
publishes and sells one of the best computer-player programs, as well
as pursuing his academic research in the same field).

Out of all decisions in bidding, the single most frequent one is
whether to contract for the "game contract" of "three notrumps"
(undertaking for the partnership to make at least 9 of the 13 tricks,
without any trump-suit).  So, I focused on that, as well as on "flat"
hands -- deals for which all suits have between 2 and 4 cards in each
partner's hand (no especially short nor especially long suit) -- for
which the 3NT target is particularly relevant and point-count more
important than for "shapely" hands (hands with certain suits being
particularly long or short, for which many other considerations may
apply).  I extracted the 80,000+ deals from Ginsberg's library that
meet these constraints and classified partners' hands by the counts of
Aces, Kings, ... down to Tens (2,000+ different combinations of
counts), with the "ground truth" information being whether that given
pairs of hands (played at NT at double dummy) makes 9+ tricks, or 8
tricks or less.  With the classic 4-3-2-1-0 "point-count", I have seen
that the correct threshold for 3NT is the sum of 25+ "points" in the
hands of the two partners, which is indeed the threshold currently
taught (for some reason in the '40s and '50s a threshold of 26 was
taught instead, but experience over the decades made the threshold of
25 emerge as more optimal).

It's an error to bid 3NT when the two combined hands make 8 or fewer
tricks; it's also an error NOT to bit 3NT when the two combined hands
make 9 or more tricks.  In one case, a small gain turns into a small
loss; in the other, an opportunity to turn a small gain into a large
gain is missed.  It so happens that (at least in some cases and forms
of bridge competition) the cost of errors of each type is quite close
so for simplicity I can actually start with the assumption that each
kind of error has the same cost.

So, what I propose is to find the best assignment of weights ("point
count") and threshold for the decision of whether to bid 3NT or not --
for starters.  Many variations are clearly possible (e.g. use weights
that best predict the number of tricks, rather than just the fact
of >=9 vs <=8 tricks, etc, etc) and I plan to explore many of them...
but the key intermediate goal is "finding good/optimal assignments
of weights" ("point-counts" -- and later, if and when I've settled
on what's a good point-count, I may then explore variations such as
pluses or minuses for groupings and positioning of honors, etc etc)
and threshold for the 3NT bid to be profitable.


Again thanks to all who helped!

Alex




From aahz at pythoncraft.com  Wed Apr  2 00:54:34 2003
From: aahz at pythoncraft.com (Aahz)
Date: 2 Apr 2003 00:54:34 -0500
Subject: ZODB: Quixote vs. Twisted
References: <pan.2003.04.01.13.58.42.127815.759@thomas-guettler.de> <b6d5s3$lvo$2@newshost.accu.uu.nl> <pan.2003.04.02.07.45.41.185943.810@thomas-guettler.de> <b6dtjn$unc$1@newshost.accu.uu.nl>
Message-ID: <b6dtuq$q31$1@panix3.panix.com>

In article <b6dtjn$unc$1 at newshost.accu.uu.nl>,
Martijn Faassen <m.faassen at vet.uu.nl> wrote:
>
>-- 
>History of the 20th Century: WW1, WW2, WW3?
>No, WWW -- Could we be going in the right direction?

You know, we're three years into the 21st century.  I'm surprised that
the PSU hasn
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From koenig at v-i-t.de  Tue Apr 22 17:07:37 2003
From: koenig at v-i-t.de (Mirko Koenig)
Date: Tue, 22 Apr 2003 23:07:37 +0200
Subject: pygettext but no msgfmt
References: <b83snm$1ms$1@news.tu-darmstadt.de> <b844b0$92e$1@zeppelin.rz.uni-potsdam.de>
Message-ID: <b84aum$62g$1@news.tu-darmstadt.de>

On Tue, 22 Apr 2003 21:14:40 +0200, Martin v. L?wis wrote:

>> I tried to use gettext for internationalizing but i can't find the
>> msgfmt tool.
>>
>> I have debian woody with python2.1 and python 2.2 installed. In the
>> debian archive the file msgfmt.py isn't listed either.
>>
>> How can i convert the po files then?
> 
> I recommend to install the gettext package, and use msgfmt(1).
> 
> If you want to use msgfmt.py instead, you'll find it in
> /usr/share/doc/python2.1/examples/Tools/i18n/msgfmt.py.

There's no such file. There's even no such directory.
debian woody even doesn't contain msgfmt.py

i found no package including this file.

Or can you help me?

Mirko Koenig



From phil at river-bank.demon.co.uk  Wed Apr 16 18:35:02 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Wed, 16 Apr 2003 23:35:02 +0100
Subject: PyQT and Eric
In-Reply-To: <b7k998$2c9e$1@news.imp.ch>
References: <b7is7g$i1d$1@rex.ip-plus.net> <mailman.1050479366.18545.python-list@python.org> <b7k998$2c9e$1@news.imp.ch>
Message-ID: <200304162335.02066.phil@riverbankcomputing.co.uk>

On Wednesday 16 April 2003 8:00 pm, Daniele Trovato wrote:
> @Phil   I hope the answers helpful enough !
>
> > What version of PyQt?
>
> 3.5-54
>
> > python -c "from qt import qVersion, PYQT_VERSION"
>
> Traceback (most recent call last)
> File "<string>", line 1, in ?
> File "/usr/lib/python2.2/site-packages/qt.py", line 39, in ?
> import libqtc
> ImportError: /usr/lib/python2.2/site-packages/libqtcmodule.so: undefined
> symbol: _ZNK16QAssistantClient9classNameEv

The problem is your Qt installation. It has been built and installed to a 
separate location (rather then left in situ) which means that the library 
containing QAssistantClient has been stripped of too many symbols. I think 
the problem has been fixed in Qt v3.1.2.

I can only suggest raising it as a bug with your distro provider, or 
rebuilding Qt.

Phil




From news at nospam.prana.eml.cc  Wed Apr 16 12:46:03 2003
From: news at nospam.prana.eml.cc (Mahesh Padmanabhan)
Date: Wed, 16 Apr 2003 16:46:03 GMT
Subject: How many of you are Extreme Programmers?
In-Reply-To: <mailman.1050504987.840.python-list@python.org>
References: <F7DB8D13DB61D511B6FF00B0D0F06233045E4632@mail.hebmex.com> <mailman.1050504987.840.python-list@python.org>
Message-ID: <fNfna.27443$4P1.2478698@newsread2.prod.itd.earthlink.net>

Christopher Blunck wrote:
> On Wed, Apr 16, 2003 at 09:36:14AM -0500, sismex01 at hebmex.com wrote:
> 
>>Many of us have used XP techniques without actually knowing
>>that our methodology was labelled as "XP", it comes natural
>>in the scope of our tools and goals.
> 
> 
> Absolutely!  Which is why I believe that many of us in the Python
> community follow many of the principals of "XP" without knowing it.
> 
> 
> 
>>I don't know; "OOP", "XP", "Structured Programming", etc.. are
>>all monickers, buzzwords, fads, which come and go; they come
>>into existance when some mass-market-oriented technical book
>>writer says "Hey, this programmer is successful, let's see
>>what he's doing"; once done, the writer creates some catchy
>>name for this guy's techniques, and sells his book.
>>
>>I really dislike market speak, as you can tell. :-)
> 
> 
> Me too.  What's discouraging is that it often takes books, buzzwords,
> and fads to get the attention of higher-ups.  Developing software
> is not all that difficult - it's always so confusing why people make
> software development more complex than it actually is.
> 
> 
> -c
> 

Yeah, XP is a buzzword too. I programmed a lot of C++ code ( I wish it 
were python) and when I came to know about XP, I realized that all the 
software engineering that I put into the C++ code could be classified as 
XP (including the most controversial of all - pair programming !)




From aleax at aleax.it  Sat Apr 12 18:35:35 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 22:35:35 GMT
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net> <3E96D270.B8E99B27@engcorp.com> <b76ult$plu$1@slb5.atl.mindspring.net> <52e5ab5f.0304121323.1d1e376b@posting.google.com>
Message-ID: <Xw0ma.5903$LB6.168491@news1.tin.it>

Michael Robin wrote:
   ...
> class C:
>     x = 42
>     def foo(): self.x += 1

No doubt you intend foo to have a self argument, right?

> other way around?) - but in this case I think we'd vote for the extra
> typing for the class accessor, just as we don't mind "self.".

It's not a matter of extra typing, but of changing semantics!!!

>>> class C:
...   x = 42
...   def foo(self): self.x = self.x + 1
...
>>> class D:
...   x = 42
...   def foo(self): self.x = D.x + 1
...
>>> class subC(C): x = 23
...
>>> class subD(D): x = 23
...
>>> aa = subC(); aa.foo(); print aa.x
24
>>> aa = subD(); aa.foo(); print aa.x
43
>>>

See the difference?  By switching to D.x instead of self.x, you
inhibit the data override in class subD from working as intended --
you cannot any more make a subclass have a different default /
starting value for the x attribute of its instances that is
different from that in the base class.

You need to know which semantics you desire, but in most cases
allowing data overriding adds flexibility at no cost -- thus it's
what you should be doing by default.  Use self.attribute, NOT
someclass.attribute, even in methods written inside someclass,
in order to allow data overriding to work.


Furthermore, would you be happy calling aa.__class__.foo(), if
aa.foo didn't automatically go get the attribute from the class
when it doesn't find it in the instance?  I doubt so.  Automatic
delegation of attribute access from instance to class is THE
prime mechanism of Python's O-O, after all.  So what horrid kludges
would you suggest, to ensure that accessing self.x delegates to
the class (when instance self has no specific attribute x) just
in some cases but not in others...?

MUCH better as we have it today -- simpler, more practical,
more flexible, more regular and general -- ALWAYS delegate
the lookup when the attribute isn't found right in the instance.
How could piling exceptions on top of this simple rule make
things any better?!


Alex




From skip at pobox.com  Wed Apr 23 09:27:27 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 23 Apr 2003 08:27:27 -0500
Subject: True/False
Message-ID: <16038.38079.397544.701140@montanaro.dyndns.org>

    Python 2.3a2+ (#2, Mar 21 2003, 22:13:05) 
    [GCC 3.2.3 20030316 (Debian prerelease)] on linux2
    Type "help", "copyright", "credits" or "license" for more information.
    >>> try:
    ...     None
    ... except NameError:
    ...     None = 0
    ... 
    <stdin>:4: SyntaxWarning: assignment to None

What does this have to do with conditional assignment to True and False?
Sure, in 2.3 you get a SyntaxWarning when assigning to None.  So don't do
that.  None has been around for a good long time (probably since the Big
Bang or close enough to it that our crude atomic clocks can't measure
precisely when it was added), while True and False are recent inventions.
Assigning to True or False today doesn't produce a SyntaxWarning, and given
that it's rather common to define them in programs, it's likely that it
won't for a long time.  On the other hand, Guido would like to eventually
make None a keyword, so a SyntaxWarning now makes sense.

Skip




From amk at amk.ca  Mon Apr 14 20:28:23 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Mon, 14 Apr 2003 19:28:23 -0500
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it> <pan.2003.04.14.18.33.53.118906@effectif.com> <_gGma.14411$LB6.389752@news1.tin.it>
Message-ID: <C3udnSC5Koq6zwajXTWcrg@speakeasy.net>

On Mon, 14 Apr 2003 22:05:46 GMT, 
	Alex Martelli <aleax at aleax.it> wrote:
> Yes, I would agree they should be.  I don't know how (and indeed, if)
> one can submit patches for AMK's "what's new" documents, though -- I
> don't normally think of them as part of "Python's docs" 

They're in Doc/whatsnew/ .  People can submit patches or, for small things,
make changes and check them in; I'll rewrite later if I feel like it.

I've been only half-following this thread.  In the end, is there any
consensus on __slots__? Should the "What's New" document only mention them
as a way to save memory, or should their use be discouraged?  (I have no
reservations about describing a new feature as not particularly useful, or
about suggesting that a feature's use be avoided.)

--amk                                                             (www.amk.ca)
"Eureka" is Greek for "This bath is too hot".
      -- The Doctor, in "The Talons of Weng-Chiang"



From danb_83 at yahoo.com  Tue Apr  1 14:01:33 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 1 Apr 2003 11:01:33 -0800
Subject: Questions on linked lists
References: <kl8ia.27315$OV.127379@rwcrnsc54> <5vonl-vm3.ln1@beastie.ix.netcom.com> <slrnb8ja7a.421.ramen@lackingtalent.com>
Message-ID: <ad052e5c.0304011101.485fb2ae@posting.google.com>

Dave Benjamin <ramen at lackingtalent.com> wrote in message news:<slrnb8ja7a.421.ramen at lackingtalent.com>...
> In article <5vonl-vm3.ln1 at beastie.ix.netcom.com>, Dennis Lee Bieber wrote:
> > 
> >         I suspect that, in the case of Python, it would be reinventing the 
> > wheel. Python has lists, and list operations, as a native data type; 
> > one doesn't have to program list manipulation from scratch. What Python 
> > does not have is user control of dynamic allocation and access to 
> > addresses.
> 
> That said, what would be an example of a legitimate need for a custom linked
> list data structure in Python? I can't say I've come across one yet. Do
> linked lists outperform native Python lists at certain operations?

Linked lists are much faster when you need to insert data in the middle of the list.



From achrist at easystreet.com  Tue Apr 29 14:42:22 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Tue, 29 Apr 2003 11:42:22 -0700
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org> 
	     <1bf4033f.0304290627.40f7d831@posting.google.com> 
	     <uptn56xxi.fsf@fitlinxx.com> <mailman.1051638646.15981.python-list@python.org>
Message-ID: <3EAEC78E.9D51D668@easystreet.com>

When attempting to introduce a new programming language into a 
corporate shop, you've got to look at it from the manager's
side.  

1. There is a tremendous potential downside to having any language
in your shop if you have fewer than 3 fluent developers.

2. If you introduce a new language, you hope that it will succeed and
be used in your shop for 5-10 years.

3. Multiply (1) times (2) and you see that the decision to introduce
a new language is well into a seven-figure $US investment.

4. Seven-figure $US is well beyond the spending authority of most
managers who know anything much about programming.  In many companies
the IS director can't spend that much without approval of the CEO and/or
board of directors.

5. A relatively low-level manager might get his company into such an
investment without explicit approval by slipping in a new language,
but the manager will not be very comfortable doing so if there is any
significant downside potential.

6. The way to counter all of this in favor of Python is to make the
case that fewer lines of code result in much lower maintenance cost,
so that those who look beyond the short-term see benefits and not
costs.  One thing that might be interesting to try would be to take
a VB application from 1995 or so (16-bit Win 3.1) and see how easy it
is to get it running under the current version of VB, ie VB.Net.  Then
take a python version 1.4 app and bring it up to verson 2.3.  I believe
that python will look pretty good in this comparison. 


Al



From tim_one at email.msn.com  Fri Apr 25 01:37:40 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Fri, 25 Apr 2003 01:37:40 -0400
Subject: Dict handling of floats
In-Reply-To: <1051243370.926621@yasure>
Message-ID: <LNBBLJKPBEHFEDALKOLCCEBHEIAB.tim_one@email.msn.com>

[Donn Cave, on float displays]
> ...
> You're looking at the repr() output, which for reasons we won't
> go into here renders the value to its ultimate precision.  If
> you run the same numbers through str(), they'll look different.
>
> Maybe the big question here is whether your notion of "right"
> has any computational significance, that is, does it matter to
> your application which of these renditions more accurately
> describes the value.  Because in that respect, repr is right
> and str is almost right.

Nope, it's very rare for either to tell you the whole truth.  repr()
produces the minimal number of digits so that non-equal floats never have
equal reprs, but that's usually a fraction of the number of decimal digits
that would be needed to display the true value stored in the machine.  For
example, as explained in the oft-referenced Tutorial Appendix, in

>>> 0.1
0.10000000000000001
>>>

the 17-digit repr is still just an approximation to the value in the
machine, which latter is exactly

0.1000000000000000055511151231257827021181583404541015625

repr() rounds that to 17 significant digits, because 17 is the smallest
number of digits you can round back to in all cases without producing
identical strings for some non-equal floats.  It's also the smallest number
you can round back to and always get back the same float you started with if
you read the string in again.  Telling the whole truth wasn't a goal here;
telling enough of the truth to avoid gross surprises was.  That "0.1" isn't
exactly one-tenth is a gross surprise inherited from the hardware, of
course.





From sdhyok at yahoo.com  Mon Apr 14 15:31:59 2003
From: sdhyok at yahoo.com (sdhyok)
Date: 14 Apr 2003 12:31:59 -0700
Subject: Standard way to express a date in a real number?
Message-ID: <420ced91.0304141131.73a6206e@posting.google.com>

I am looking for a way to express a date in a real number
in Python.

For instance, in MatLab,
n = datenum(2001,12,19,18,0,0) returns n = 731204.75.

This kind of expression turned out to be very helpful
when handling or storing dates.

Is there any funtion for the purpose in Python (e.g. in datetime module)?

Daehyok Shin



From scott_list at mischko.com  Tue Apr 22 13:03:53 2003
From: scott_list at mischko.com (Scott Chapman)
Date: Tue, 22 Apr 2003 10:03:53 -0700
Subject: Pythonwin don't show the complete list in interactiv window!?
In-Reply-To: <oprn0862owmu3s0q@news.cis.dfn.de>
References: <oprnzwcechmu3s0q@news.cis.dfn.de> <d17pa.6563$8K2.53295@news-server.bigpond.net.au> <oprn0862owmu3s0q@news.cis.dfn.de>
Message-ID: <200304221003.53526.scott_list@mischko.com>

On Tuesday 22 April 2003 06:31, Klaus Meyer wrote:
> >> The interactive windows shows only a part of the list (last value 771).
> >> (Win2000)
> >
> > It is a Scintilla bug or limitation. Newer versions of Scintilla don't
> > stop at 771 but there are still limits.
>
> Thanks.
>
> For a Python beginner it's a little frustrating to search bugs in my own
> code but could not find and then later find out, that it was a tools bug.
> Because Idle also has bugs (for example, after a little working time won't
> save my source file so i lost the new source-version!) i switched to
> ConTEXT (http://www.fixedsys.com/context/).

If you want a really slick text editor for Winodows, please have a look at 
Ultraedit.  It is really excellent.  http://www.ultraedit.com/

Scott




From ianb at colorstudy.com  Thu Apr 17 15:42:36 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 17 Apr 2003 14:42:36 -0500
Subject: HTML form pre-populating functionality or hints on building
	myown
In-Reply-To: <Hhwna.114818$It5.51036@news2.central.cox.net>
References: <200304161939.13004.scott_list@mischko.com>
	 <mailman.1050560546.21370.python-list@python.org>
	 <Hhwna.114818$It5.51036@news2.central.cox.net>
Message-ID: <1050608556.720.298.camel@lothlorien>

On Thu, 2003-04-17 at 06:32, Steve Holden wrote:
> I'd be happy to :-). See
> 
>     http://wwwsearch.sourceforge.net/ClientForm/
> 
> It's a very interesting module, and appears (from my own limited testing) to
> work well.
> 
> [...]
> 
> >From the documentation: """ClientForm is a Python module for handling HTML
> forms on the client side, useful for parsing HTML forms, filling them in and
> returning the completed forms to the server. It developed from a port of
> Gisle Aas' Perl module HTML::Form, from the libwww-perl library, but the
> interface is not the same"""

I think what the OP was talking about was different, meant to be used on
the server side.  Like you'd have a pseudo-template something like:

template = """
<form action="" method="POST">
<input type=text name=firstname value="">
<input type=text name=lastname value="">
<input type=submit value=register>
</form>
"""

Note that it doesn't look like a template at all, not even the most
minimal markup.  Then someone comes along and submits the form, but
forgets the firstname (though they remember the lastname).  You do
something like:

errors = False
if not form["firstname"]:
    write("Please enter your first name<br>")
    errors = True
if not form["lastname"]:
    write("Please enter your last name<br>")
    errors = True
if errors:
    write(subform(form, template))

And that subform function goes in, finds the <input...> tags, and fills
them in from the form dictionary, so that the user is presented with
their input and a chance to make corrections.

  Ian






From pinard at iro.umontreal.ca  Sat Apr 12 08:51:52 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 12 Apr 2003 08:51:52 -0400
Subject: Is there any way to make repr(aString) use double quotes?
In-Reply-To: <698f09f8.0304120127.52417f53@posting.google.com>
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
	<mailman.1049836294.24621.python-list@python.org>
	<698f09f8.0304120127.52417f53@posting.google.com>
Message-ID: <oqhe93swt3.fsf@titan.progiciels-bpi.ca>

[Jeremy Fincher]

> Francois Pinard <pinard at iro.umontreal.ca> wrote in message news:<mailman.1049836294.24621.python-list at python.org>...
> > # Python delimits a string it by single quotes preferably, unless
> > # single quotes appear within the string while double quotes do
> > # not, in which case it uses double quotes for string delimiters.
> > # Checking the string contents, the C code stops at the first NUL.
> > # We prefix the string with a single quote and a NUL, this forces
> > # double quotes as delimiters for the whole prefixed string.  Then,
> > # we get rid of the representation of the single quote and the NUL.
> > 
> > write('"' + repr("'\0" + value)[6:])

> Ah, perfect!  Thanks :)

You are welcome, of course.  But this may not be perfect! :-) We are relying
here on an undocumented behaviour of Python, and there is no reasonable[1]
guarantee that some later version of Python will not break the above code.
So in practice, it is only usable as a short or medium term kludge, that you
should be prepared to _happily_ revise if the need ever arises.

--------------------
[1] What means "reasonable"?  It is a difficult concept.  To illustrate it a
tiny bit, consider the timely finalisation of Python objects.  This is a
somewhat undocumented behaviour for which we have reasonable guarantee that
it will never change (in the official, standard Python).

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From ark at research.att.com  Sun Apr 20 14:37:45 2003
From: ark at research.att.com (Andrew Koenig)
Date: Sun, 20 Apr 2003 18:37:45 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050857614.6299.python-list@python.org>
Message-ID: <yu99adel3thi.fsf@europa.research.att.com>

Tim> Continuations are a different *foundation* for implementing
Tim> control-flow features, more than they're a feature in their own
Tim> right.  While they're an elegant foundation, they're not the
Tim> foundation Guido (or 99.9+% of the world's other language
Tim> designers) chose at the start, and it remains unlikely that any
Tim> Python Guido works on will adopt them.

By way of further context for continuations, it might be worth
mentioning that Scheme (an MIT dialect of Lisp) is based on
continuations in a fundamental way.  Also, the Bell Labs/Princeton
implementation of ML (``Standard ML of New Jersey'') is based internally
on continuations, although this fact is pretty much invisible to
user programs.

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From achim.domma at syynx.de  Mon Apr 14 02:51:34 2003
From: achim.domma at syynx.de (Achim Domma)
Date: Mon, 14 Apr 2003 08:51:34 +0200
Subject: Inserting Python commands into HTML documents
References: <zpqdne0Kztcb0AejXTWcow@comcast.com>
Message-ID: <b7dlgh$ir2$04$1@news.t-online.com>

> Is there a way for it to act more like PHP, as in run Python commands
> from inside the (index|contacts|about|etc).html or at least simulate
> that behavior? (This is on a linux server and I am not to use PHP, FYI.)

That page might help you:
http://www.python.org/cgi-bin/moinmoin/WebProgramming

regards,
Achim






From graham at effectif.com  Mon Apr 14 15:07:37 2003
From: graham at effectif.com (Graham Ashton)
Date: Mon, 14 Apr 2003 20:07:37 +0100
Subject: Reading/writing files
References: <pan.2003.04.14.17.32.01.209427@soul.net>
Message-ID: <pan.2003.04.14.19.07.36.357067@effectif.com>

On Mon, 14 Apr 2003 18:32:02 +0100, lost wrote:

> Hi there,
> I'm a bit lost in terms of reading and writing files. I've been reading
> the python docs that came with my distibution, and they aren't too
> clear. I need something like this
> 
> for line in a file
> 	print line
> print done

You're very nearly there. I made a file called foo.txt with these lines
in:

shoe
fruit
wheel barrow

Then I wrote a script to print out each line:

    for line in file('foo.txt'):
        print line
    print 'done'

If you run it you'll find you get a blank line between each line in your
input file. That's because print automatically appends a new line
character to the end of each line it prints, and the file already contains
new line characters. To get round that, you could import the sys module
and use sys.stdout.write(line) inplace of the print statement.
Alternatively you could use the rstrip() method on a string (see below).

When you open a file for writing, you need to say that you want to write
to it. Add a second option to the write() method like this:

    output_file = file('bar.txt', 'w')

Improving on the example above, I wrote this:

    input_file = file('foo.txt')
    output_file = file('bar.txt', 'w')
 
    line_num = 0
    for line in input_file:
        line_num += 1
        output_file.write('%d: %s' % (line_num, line))
 
    input_file.close()
    output_file.close()
 
    print 'done'

Note that this example stores each file object in a variable and calls
close() on them both when it's done with them. If you don't close a file
yourself it'll get closed by Python, but it'll happen at a time of
Python's choosing, not yours. It's generally thought to be better practice
to close them yourself.

> Also, do i need to take
> into account the \n at the end of a line? for example, if I have
> hello:there
> and
> hello:there\n
> how could i ignore the \n?

If you have 'hello:there\n' in a variable called line, you can get rid of
the whitespace off the end by calling this:

    line = line.rstrip()

Note that you have to reassign the result of rstrip() to the line
variable, as strings are immutable. There are plenty of other methods that
you can call on strings; read the docs for the string module for more.

Good luck...

--
Graham Ashton



From 'tert at pncen.hf'.decode  Fri Apr  4 00:04:45 2003
From: 'tert at pncen.hf'.decode (Greg Krohn ('rot-13'))
Date: Thu, 3 Apr 2003 23:04:45 -0600
Subject: about NLTK in Python
References: <97e89849.0304032019.28a612ad@posting.google.com>
Message-ID: <b6j3pr01cf5@enews4.newsguy.com>

"abdi" <a_abdi406 at yahoo.com> wrote in message
news:97e89849.0304032019.28a612ad at posting.google.com...
> hi,
>
> I have question about the modules, I downloaded NLTK module for win32
> and I followed the instruction for installation from the web, having
> dowloaded I the wizardinstallation guided me to complete the
> installation and it installed the same directory as Python22,  I
> opened the python interactive window to check whether the installation
> was successfull I called the module as follows:
>
> >>> from nltk.token import *
> Traceback (most recent call last):
>   File "<interactive input>", line 1, in ?
> ImportError: No module named nltk.token
> >>>
>
> I consulted the FAQ before posting the question, so what is going on ?
>
> folder structure is as follows:
>
> C:\\Python22\\nltk1.0
>
>
>
> thanks in advance


Is C:\\Python22\\nltk1.0 in your PYTHONPATH?





From harri.pasanen at trema.com  Thu Apr 10 11:00:47 2003
From: harri.pasanen at trema.com (Harri Pasanen)
Date: 10 Apr 2003 08:00:47 -0700
Subject: python 2.3a2 Win32 expected failures in regression tests?
References: <mailman.1049910795.4707.python-list@python.org>
Message-ID: <e2d925a8.0304100700.3a51653a@posting.google.com>

Tim Peters <tim.one at comcast.net> wrote in message news:<mailman.1049910795.4707.python-list at python.org>...
> [Harri Pasanen]
> > I've built python 2.3a2 myself on Win32, using Visual C++ 6.0.
> 
> Which flavor of Windows ("Win32" is ambiguous -- they're not identical)?
> 2.3a2 passed all tests on Win98SE and Win2K before it was released.
> 
> > When running the regression tests, the following 3 fail:
> >
> > test_tarfile, test_timeout, test_zipimport
> >

Ok, found out that my testtar.tar file was corrupted (Python was
imported to local CVS, and .tar was not declared to be binary file ->
checkout to W2K got it all mangled).

On the test_timeout, it is always the testRecvTimeout that fails, but
that is likely an artifact of me running it in vmWare linux hosting
w2k.

Harri



From spam at mattbergin.co.uk  Sun Apr 13 12:45:02 2003
From: spam at mattbergin.co.uk (Matt Bergin)
Date: Sun, 13 Apr 2003 17:45:02 +0100
Subject: Reference Tracking
In-Reply-To: <KUYla.7409$T34.201025@news2.tin.it>
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it>
Message-ID: <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk>

Alex Martelli wrote:
> That might cause you problems at termination time, I think.  Also,
> is there any risk that your objects could end up as parts of cycles
> of mutual references?  If so, defining a __del__ makes it impossible
> for the garbage collector to ever reclaim them.

I don't think there's any risk of reference cycles. I should have been 
more clear about what the objects are. I'm trying to write a class to 
wrap MySQLdb and hand out database cursors to worker threads.

> You can try to avoid the periodic scanning overhead by using a
> tricky __del__, or a non-tricky __del__ or weak-reference callback
> for some object that's "packed together with" the object you give
> out (keeping the latter alive in a per-class container anyway).  A
> radically different and far cleaner approach would be to specify
> a way (e.g. a method) for the using-code to say "I'm finished with
> this object, reuse it freely from now on" -- no black magic at all
> (of course, perhaps marginally smaller convenience for the using
> code, but it's not all that bad to ensure a finalization method
> gets explicitly called, IMHO).

I'm just worried that if an exception occurs in a thread then the cursor 
will never be returned, using the 'no black magic' method which appeals 
to me the most. Could I prevent this by careful use of try: finally?

-- 
Matt Bergin




From kristofer at hotpop.com  Fri Apr 25 01:02:50 2003
From: kristofer at hotpop.com (Kristofer Wouk)
Date: Fri, 25 Apr 2003 05:02:50 GMT
Subject: variable scoping problem
In-Reply-To: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com>
References: <p60hav0mnenbsio44uoo3pti7b78rn5ij0@4ax.com>
Message-ID: <_j3qa.1978$001.1638401@monger.newsread.com>

Tung Wai Yip wrote:
> Consider this
> 
> 
>>>>x=1
>>>>def foo():
> 
> ...   print x
> ...
> 
>>>>def bar():
> 
> ...   #print x   <- referenced before assignment error
> ...   x=2
> ...   print x
> ...
> 
>>>>foo()
> 
> 1
> 
>>>>bar()
> 
> 2
> 
>>>>print x
> 
> 1
> 
> 
> What rules in Python explain the scoping of x? Is there a name for the
> scope of the first x? 'module' scope? Why is bar() not able to access
> x but foo() can?
> 
> Wai Yip Tung
> 
OK, I'm a newbie but let me give this a try. I think since foo() has no 
internal variable x it is forced to use the global(?, or whatever) x. 
However, since bar() actually assigns something to x, it just creates a 
new variable in bar()'s scope. Am I correct? I'd like to know myself...

Kris




From sbaptista at yahoo.com  Wed Apr 23 14:32:51 2003
From: sbaptista at yahoo.com (Stan Baptista)
Date: Wed, 23 Apr 2003 11:32:51 -0700 (PDT)
Subject: Off topic: What usenet news reader to use?
In-Reply-To: <Xns93665EDC8E0E0duncanrcpcouk@127.0.0.1>
Message-ID: <20030423183251.96440.qmail@web11604.mail.yahoo.com>

> "Wai-Yip Tung \(wtung\)" <wtung at cisco.com> wrote
> > I haven't used Usenet news for a long time. Now
> that I have subscribed
> > to python-list I find that email does not scale
> well to the high volumne

I also want to switch over to a newsgroup for the same
reason so I downloaded xnews and tried to access
these:

comp.lang.python
and
comp.lang.python.announce

I get error: "10110 unknown error while looking up
comp.lang.python[.announce]"

What's up?

Thanks,
Stan


__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com




From imbosol-1050454138 at aerojockey.com  Tue Apr 15 21:47:23 2003
From: imbosol-1050454138 at aerojockey.com (Carl Banks)
Date: Wed, 16 Apr 2003 01:47:23 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com>   <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <yfsr883b96o.fsf@black132.ex.ac.uk>
Message-ID: <LC2na.19696$Gb1.9950@nwrdny01.gnilink.net>

Alexander Schmolck wrote:
> 
> 
> Carl Banks <imbosol-1050288268 at aerojockey.com> writes:
> 
> Firstly:
> 
>> > Have you had traumatic childhood experiences with early perl or lisp
>> > dialects?  :)
>> 
>> Yes
> 
> OK, let's try this. Just consider for a moment the following possibility:
> 
>  THERE IS NOT A *SINGLE* DRAWBACK TO DYNAMICALLY SCOPED *GLOBALS*.

I don't think you're giving me enough credit here.

Fully aware that lexically-scoped local variables will still exist
alongside dynamically scoped variables, and need not be confused with
each other, I simply do not want to see dyanamic scoping at all.

The disadvantages of dynamic scoping are there, completely
irrespective of lexical variables.  Labeling it a "global" doesn't
make these problems go away (as if there were a even useful
distinction between dynamic globals and locals).  The mere possibility
of temporarily rebinding a variable is a wart.

And, frankly, the legitimate uses of temporarily rebound globals are
very rare, otherwise you're doing something wrong.  The right way to
pass parameters is the argument list.

Yes!  I fully admit that there are practical reasons to not pass
parameters as arguments, and occasionally a dynamically scoped
variable can do exactly what you want.  But it comes with a completely
unexcusable cost, namely: an inferior, bug-prone scoping system; the
tendency for people to lazily use dynamically scoped variables when
they should be passing them as parameters; add complexity and
confusion to the language.

Also, dynamic scoping violates a chunk of the Python Zen: Explicit is
better than implicit.  (You should pass parameters explicitly.)
Simple is better than complex.  (One scoping strategy is simpler than
two.)  Special cases aren't special enough to break the rules.  (Valid
uses for dyanmic scoping are special cases and should not break the
rule of lexical scoping.)

All this, when you can get the same effect in a thread-safe way with
try ... finally.


Forget it.  The benefit is miniscule, work-around-able, and the
drawbacks are enormous.  Dyanmic scoping is just a bad idea that needs
to die.


-- 
CARL BANKS



From dave at pythonapocrypha.com  Wed Apr  9 01:36:59 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Tue, 8 Apr 2003 22:36:59 -0700 (PDT)
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <3e936947$1@nntp0.pdx.net>
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org>
 <BAB8B22D.2022%dogen@mindspring.com> <mailman.1049839153.2128.python-list@python.org>
 <3e936947$1@nntp0.pdx.net>
Message-ID: <Pine.LNX.4.51.0304082233300.1595@ns0.tbye.com>

On Tue, 8 Apr 2003, Scott David Daniels wrote:

> Dave Brueck wrote:
> > ...
> > def IsPal(s):
> >   r = list(s)
> >   r.reverse()
> >   r = ''.join(r)
> >   ...
>
> This bit should go significantly faster with array:
>
>      from array import array
>
>      def IsPal(s):
>         t = array('c', s)
>         t.reverse()
>         r = t.tostring()

Not necessarily - it has a lot to do with the size of the input string
's'. For short strings (under about 30 chars) the array version is
actually _slower_ than the list one, so you'd definitely want to avoid it
if you're checking for palindromes in single words or small sets of words.
:)

-Dave




From jepler at unpythonic.net  Sat Apr  5 12:32:15 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Sat, 5 Apr 2003 11:32:15 -0600
Subject: Novelwriting 0.3: rule-based random document generation
Message-ID: <20030405173215.GB13698@unpythonic.net>

http://unpythonic.net/~jepler/novelwriting/index.html

Description
-----------

Novelwriting[1] generates random text based on a user-defined grammar.
It is similar to the Dada Engine[2], which is used by the Postmondernism
Generator[3].  The name was chosen based on the Monty Python sketch,
"Thomas Hardy / Novel Writing"[4].  It requires Python 2.2.


Example
-------
Start: folk-saying;
folk-saying: person " once said that " folk-saying ".\n";
person: "My " relative | "The President" | "Cardinal Fang";
relative: "mother" | "aunt" | "grandmother" ;
folk-saying: "a rolling stone gathers no moss"
    | "a stitch in time saves nine"
    | "no-one expects the Spanish Inquisition";
;;


Advantages compared to the Dada Engine
--------------------------------------
The Dada Engine has only a limited degree of programmability.  In
Novelwriting, you can use the full power of Python at any point in a rule.
For example:
	Start: product-claim
	product-name: "Novelwriting" | "Dada Engine";
	product-claim:
		@set(this-product, product-name)
		" is twice as good as "
		@alternative(product-name, @get(this-product));
	;;
	from novelwriting.calls import get, set, alternative
(The facility provided by @get/@set is available in the Dada Engine, 
but @alternative is impossible to write)


Limitations compared to the Dada Engine
---------------------------------------
Novelwriting does not have a formatting library like the Dada Engine.
Novelwriting does not have a nice syntax for text substitution rules or
maps.  Novelwriting does not have parameterized rules.


[1] http://unpythonic.net/~jepler/novelwriting/index.html
[2] http://dev.null.org/dadaengine/
[3] http://www.csse.monash.edu.au/cgi-bin/postmodern
[4] http://www.stone-dead.asn.au/albums-cds/sketches/matching-tie-and-handkerchief/thomas-hardy-novel-writing.html




From glyn at millingtons.org  Tue Apr  8 02:59:55 2003
From: glyn at millingtons.org (Glyn Millington)
Date: Tue, 08 Apr 2003 07:59:55 +0100
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <m27ka7aae0.fsf@localhost.localdomain>
 <b6pri6$drf$1@panix2.panix.com>
 <mailman.1049656989.5896.python-list@python.org>
 <b6qhl6$sht$1@panix2.panix.com>
Message-ID: <878yul4gpg.fsf@millingtons.org>

aahz at pythoncraft.com (Aahz) writes:
> Using emacs with its default keybindings hurts my fingers.  If I were to
> seriously use emacs, I'd need a hefty .emacs, which would make me unable
> to be productive with it in an emergency.  As for using emacs in vi
> mode, why not just stick with vi?

Because the advantages of using (X)Emacs are not restricted to its
superior key-bindings ? ;-)




Glyn



From aleax at aleax.it  Sun Apr 13 15:58:55 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 19:58:55 GMT
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk> <m38yuepdva.fsf@mira.informatik.hu-berlin.de> <LtFzMJA4ZZm+EwDh@jessikat.fsnet.co.uk>
Message-ID: <3kjma.11793$T34.382463@news2.tin.it>

<posted & mailed>

Robin Becker wrote:
   ...
> resolve and break any loop. As you say tuples (and the None object)
> aren't weakrefable (for some reason).

Micro-optimization...?

> The trivial silly solution is to attempt to hide the tupleness inside
> another class, but that won't work as the wrapper dies immediately.
> 
> The tuples are coming from pyRXP's tuple tree representation.

The simplest solution is to have pyRXP return a tree of *LISTS*,
instead of a tree of *TUPLES*, as it's able to do since about 6
months ago thanks to the tiny changes you and I made to it.  As
I showed back then, there's no measurable cost connected to that,
and the advantages are many.


Alex




From mal at lemburg.com  Tue Apr 29 16:44:43 2003
From: mal at lemburg.com (M.-A. Lemburg)
Date: Tue, 29 Apr 2003 22:44:43 +0200
Subject: mxODBC puzzle
In-Reply-To: <3EAED98B.2090009@emseal.com>
References: <3EAED98B.2090009@emseal.com>
Message-ID: <3EAEE43B.1040307@lemburg.com>

Bill Witherspoon wrote:
> I've been fooling around with mxODBC/mxDateTime for a couple of days 
> now, trying to effect an Update query. Can anyone spot the error I'm 
> making with the code below:
> 
> 
> sql = "Update tblOEOrders set tblOEOrders.ScheduleDate = #" + 
> newdate.strftime("%m/%d/%Y") + "# where tblOEOrders.OrderID = " + 
> str(orderid) + ";"
> 
> myconn = mx.ODBC.Windows.Connect('production')
> mycursor = myconn.cursor()
> numlines_affected = mycursor.execute(qrystring)

This should read:

mycursor.execute(...)
numlines = mycursor.rowcount

> The DB is Access97 via DSN (which is *not* read only), and query seems 
> to return OK (1 record affected), but the record never gets updated. 
> Using the same code I can do Selects without difficulty.
> Do I have to do a fetchall() equivalent to make the update query run, or 
> is something else wrong?

mxODBC defaults to transactional operation. To see the updates
in the database you have to commit your work, i.e. call

myconn.commit()

HTH,
-- 
Marc-Andre Lemburg
eGenix.com

Professional Python Software directly from the Source  (#1, Apr 29 2003)
 >>> Python/Zope Products & Consulting ...         http://www.egenix.com/
 >>> mxODBC, mxDateTime, mxTextTools ...        http://python.egenix.com/
________________________________________________________________________
EuroPython 2003, Charleroi, Belgium:                        56 days left





From puccio_13 at yahoo.it  Wed Apr 16 08:15:45 2003
From: puccio_13 at yahoo.it (Alessio Pace)
Date: Wed, 16 Apr 2003 14:15:45 +0200
Subject: reading Unix mailbox
Message-ID: <hZbna.17434$DN.480144@tornado.fastwebnet.it>

Hi all,
I am using Python2.3 alpha and I thought there was something updated to
iterate over the messages of a mailbox in mbox format, but the next()
method still return a deprecated RFC822  message object reference, am I
wrong? Is there a way to read mailboxes in mbox format but iterating over
MIMEMessages?
Thanks to all who can help me

-- 
bye
Alessio Pace



From wilfred_s at webtv.net  Tue Apr 15 15:37:39 2003
From: wilfred_s at webtv.net (W S)
Date: Tue, 15 Apr 2003 15:37:39 -0400 (EDT)
Subject: O'Reilly Python Books
Message-ID: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>

Hello,

I find the material, as presented, in the book Programming Python
frustrating and extremely difficult to understand.

Are the nutshell and cookbook O'Reilly Python books easier to read?

I'm fairly new to programming but trying to learn. I have read Learning
Python, Core Python, Python Bible and a bunch of perl books

Thanks,
Wilfred.




From tebeka at cs.bgu.ac.il  Wed Apr 30 08:36:09 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 30 Apr 2003 05:36:09 -0700
Subject: looping through a file
References: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au>
Message-ID: <33803989.0304300436.49efc6ab@posting.google.com>

Hello Psybar,

> hi all - ive just started with python (after having done java and c)
You've save the best for last :-)

> but it appears that python doesn't have a do... while...
> until loop.
See http://www.python.org/doc/current/tut/node6.html#SECTION006200000000000000000
and 
http://www.python.org/doc/current/tut/node5.html#SECTION005200000000000000000

> 
> #!/usr/local/bin/python
> 
> import cgi, string
> 
> form = cgi.FieldStorage()
> 
> if form.has_key('usernameField') and form.has_key('passwordField'):
>          users_username = string.strip(form['usernameField'].value)
>          users_password = string.strip(form['passwordField'].value)
authorised = 0
lines = open('users.dat').readlines()
for i in range(0, len(lines), 2):
    if (users_username == strip(lines[i]) and (users_password ==
strip(lines[i + 1]):
    authorised = 1
    break

HTH.
Miki



From ckea25d02 at sneakemail.com  Mon Apr 14 22:34:50 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 15 Apr 2003 14:34:50 +1200
Subject: Pyrex: wrapping #defines with a type name argument
In-Reply-To: <slrnb9lufd.eqb.delgado@okeeffe-pc3.la.asu.edu>
References: <b76ivu$eri$1@bob.news.rcn.net> <b7dfi7$dguav$1@ID-169208.news.dfncis.de> <slrnb9lufd.eqb.delgado@okeeffe-pc3.la.asu.edu>
Message-ID: <b7fqfr$geod$2@ID-169208.news.dfncis.de>

Olaf Delgado Friedrichs wrote:
> Just an odd thought: couldn't one feed a Pyrex file with C macros to
> the C preprocessor?

I suppose you could... You'd void your warranty,
though...:-)

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From rmunn at pobox.com  Mon Apr 14 11:15:47 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 14 Apr 2003 15:15:47 GMT
Subject: Python FAQ (Was: Re: Confusion: assignments create object references.)
References: <mailman.1050080058.23175.python-list@python.org> <3E970241.483698C7@engcorp.com>
Message-ID: <slrnb9kveb.1q1.rmunn@localhost.localdomain>

Peter Hansen <peter at engcorp.com> wrote:
[snip a very good explanation of names and references]
> (A lengthy explanation for what others have explained better and more
> succinctly in the past.  Search the archives for them, and check the 
> FAQ, if you are still confused.)

Actually, I've re-written the same explanation multiple times in the
past myself. I usually look in the FAQ or in Google archives and can't
find anything that really jumps out at me as being exactly what is
needed. This is clearly a frequently-asked question, but can you find it
in the FAQ? I couldn't.

In fact, the Python FAQ needs some trimming, IMHO. With the "Programming
in Python" section containing over 100 questions, few are going to have
the patience to sit down and read the entire list to see if their
question is on the list. Thus we get the same questions over and over in
the newsgroup. The FAQ is not serving its purpose.

I am not suggesting we get rid of the information currently contained
in the FAQ, but we should probably refactor the FAQ into two levels: one
for the truly frequently-asked questions, and one for "Here are more
questions that crop up from time to time; look through this for useful
nuggets of information." I think a shorter FAQ would be read a lot more
often.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From pobrien at orbtech.com  Mon Apr  7 09:53:58 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 07 Apr 2003 08:53:58 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net> <b6qs04$7eqi1$1@ID-169208.news.dfncis.de> <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net> <mailman.1049693590.23386.python-list@python.org>
Message-ID: <m28yum76rt.fsf@localhost.localdomain>

Ian Bicking <ianb at colorstudy.com> writes:

> On Mon, 2003-04-07 at 00:23, Carl Banks wrote:
> > Third, there are a *ton* of functions in Emacs that operate on a
> > certain point in a certain buffer (delete-char, kill-region, etc.).
> > These kinds of functions typically get bound to keystrokes and other
> > sorts of commands, which a user expects to work in the current buffer.
> > If all these functions had to be passed the current buffer as an
> > argument, it would just complicate things.
> 
> Maybe instead they should be methods of the buffer, and a keystroke
> automatically calls a method on the current buffer.

That should be easy enough.

> Managing so many methods from diverse sources might be difficult.  I
> want generic methods, I keep thinking of places where they'd be mighty
> useful... like Mixins but better.

This I don't understand.  Could you elaborate?

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From tim.golden at viacom-outdoor.co.uk  Wed Apr 16 03:31:13 2003
From: tim.golden at viacom-outdoor.co.uk (Tim Golden)
Date: 16 Apr 2003 00:31:13 -0700
Subject: Object Craft MSSQL for Python 2.2 on Win32
Message-ID: <8360efcd.0304152331.71384be4@posting.google.com>

Does anyone have a version of Dave Cole's mssql module compiled
against the Python 2.2 libraries on Win32? I don't have access to a C
compiler here, and I really want to migrate to Python 2.2, not least
because several modules and snippets coming out now are Py22-specific.
I could go to the source, obviously, but I thought I'd ask around
first.

<plug for MSSQL module - reads a bit like an advert, but so what?> 

As a database developer maintaining apps on half a dozen MS Sql Server
databases here, this module is a godsend. None of our apps is written
in Python, but I use it in all sorts of ways: occasional housekeeping;
generating pages for the intranet (MSSQL & Andy Dustman's HyperText);
an invaluable sql-to-excel routine (MSSQL, win32com and wxPython); a
series of display boards for the sales teams (PyGame, MSSQL and Pyro);
PDF documentation for the database (MSSQL and ReportLab); and... well,
you get the idea.

I've been using the MSSQL module for two years and it's never broken
on me, not once. Which is a testament both to the stability and ease
of extension of Python and the skills of the Object Craft people, to
whom I raise my hat in thanks.

</plug>

Tim Golden



From aahz at pythoncraft.com  Thu Apr 17 13:21:13 2003
From: aahz at pythoncraft.com (Aahz)
Date: 17 Apr 2003 13:21:13 -0400
Subject: fixedpoint.py vs Python 2.1
References: <$uJOMKAPtqn+Ew9c@jessikat.demon.co.uk>
Message-ID: <b7mnq9$gn7$1@panix2.panix.com>

In article <$uJOMKAPtqn+Ew9c at jessikat.demon.co.uk>,
Robin Becker  <robin at jessikat.fsnet.co.uk> wrote:
>
>I'm having a bit of difficulty with the fixedpoint module in Python 2.1.

Use the version of FixedPoint.py from my home page.  You might ask the
SF folks to link to it.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From theller at python.net  Wed Apr 30 07:35:36 2003
From: theller at python.net (Thomas Heller)
Date: 30 Apr 2003 13:35:36 +0200
Subject: Python code -> to C code
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com> <b8lpof$q0f$1@pegasus.csx.cam.ac.uk> <Xns936C8AB67B739duncanrcpcouk@127.0.0.1> <vatlfeb16jlled@corp.supernews.com> <Xns936D5DA488705duncanrcpcouk@127.0.0.1> <bryo1gj6.fsf@python.net> <Xns936D7A5A717Aduncanrcpcouk@127.0.0.1>
Message-ID: <3ck01alz.fsf@python.net>

Duncan Booth <duncan at NOSPAMrcp.co.uk> writes:

> Thomas Heller <theller at python.net> wrote in news:bryo1gj6.fsf at python.net:
> 
> > Normally this is used to implement "callback functions" in Python, to
> > call functions like Windows' EnumWindows(), or to implement COM object
> > method tables.
> > 
> > Does this count as embedding? It seems so, although I have never seen it
> > this way.
> 
> I think it counts as half of embedding. It lets you call Python from C, 
> provided the C was called originally from Python. I don't see that it helps 
> at all with the situation in my original posting which was calling Python 
> from an otherwise purely C program.
> 
> I'm impressed by the example wrapping qsort.

;-)

> I can't think offhand of a way to achieve this easily in Pyrex. That
> is, Pyrex could easily wrap qsort to call a specific Python function
> for the comparison, but there isn't anywhere in the qsort callback to
> store state, and I can't see how you manage to convert a Python
> function into a C level function without some additional state.

Right. Actually ctypes is creating C functions at runtime.  If you're
interested, the source is here:

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/ctypes/ctypes/source/callbacks.c?rev=HEAD&content-type=text/vnd.viewcvs-markup

There is additional state. On Windows, the function AllocCallback() does
the trick.  It uses CallbackTemplate(), which is a small inline
assembler function template, which is copied into malloc'ed memory, and
then some dwords are filled out with the required state. This includes a
pointer to the Python callable.

For other systems, libffi does the trick.

Thomas



From no_replies at fake_email_address.invalid  Sun Apr 20 16:12:05 2003
From: no_replies at fake_email_address.invalid (Robert Oschler)
Date: Sun, 20 Apr 2003 20:12:05 GMT
Subject: Good Python generators example?
Message-ID: <paDoa.1611$hT2.1228371@news2.news.adelphia.net>

Can someone point me to a nice concise Python 2.2 compatible example of the
use of generators?

thx

--

Robert Oschler
Android Technologies, Inc.
http://www.androidtechnologies.com
The home of PowerSell! (tm)
The FrontPage AddIn for Amazon Associates
- "Power Tools for Amazon Associates" (sm)





From mal at lemburg.com  Sat Apr 12 03:35:50 2003
From: mal at lemburg.com (M.-A. Lemburg)
Date: Sat, 12 Apr 2003 09:35:50 +0200
Subject: win32all odbc/MS SQLServer/bigint
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEAF@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEAF@admin56.narex.com>
Message-ID: <3E97C1D6.5040709@lemburg.com>

Bjorn Pettersen wrote:
> We were using the win32all odbc module, trying to fix some records in
> our database. Our account IDs are of SQL type bigint, i.e. __int64 if
> you're on MS. After seeing incorrect results (and spending some time
> looking at our own code), it turns out that the odbc module silently
> corrupts data in paramters passed to cursor.execute(). Has anyone seen
> this/got a workaround for it (we're currently cast'ing to varchar on
> selects, and doing string substitution elsewhere, which is killing
> performance, and we can't get it to work in subselects...)?

The win32all odbc module fetches BIGINTs as floats. That's
where you probably get your corruption from...

> I looked at
> the mxODBC documentation, and it says it's doing a conversion
> [automatic] to string for bigint data... (again too slow). I'm pretty
> flummoxed, the ODBC C calls for bigint/__int64 are pretty straight
> forward(?)

They are ? __int64 is not exactly portable across platforms...

BTW, conversion from strings to Python longs is done in C,
so it's not as slow as you may think.

> As an added bonus, someone had the pleasure of showing me that a simple
> select/fetchall (i.e. all C code for Python), was over three times
> faster in C#..., and the win32all odbc module has crashed python
> somewhere deep down in sql three times in the last two hours (we're
> expecting this app to run for ~4 days...). Looks like it's back to C++
> (*sigh*).

Have you tried mxODBC on this ? We have many customers running
mxODBC under Zope on heavily loaded sites for weeks... so far
without any complaints about its stability.

-- 
Marc-Andre Lemburg
eGenix.com

Professional Python Software directly from the Source  (#1, Apr 12 2003)
 >>> Python/Zope Products & Consulting ...         http://www.egenix.com/
 >>> mxODBC, mxDateTime, mxTextTools ...        http://python.egenix.com/
________________________________________________________________________
EuroPython 2003, Charleroi, Belgium:                        73 days left





From aahz at pythoncraft.com  Fri Apr 18 00:21:30 2003
From: aahz at pythoncraft.com (Aahz)
Date: 18 Apr 2003 00:21:30 -0400
Subject: Memory leak while looping
References: <1020cbf8.0304162100.6fc49e6f@posting.google.com> <1020cbf8.0304171900.110a1675@posting.google.com>
Message-ID: <b7nuga$4lq$1@panix1.panix.com>

In article <1020cbf8.0304171900.110a1675 at posting.google.com>,
Roger Hancock <roger_hancock at hotmail.com> wrote:
>
>Thanks for the help.  The xrange call solved my problem.  However, I'm
>still a little puzzled by the behavior I'm seeing with the range call.
> With my now better understanding of the range call, I would expect to
>see my function take a big chunk of memory until the function exited. 
>Instead, the function continues to grab more and more memory.  First,
>it takes most of my physical memory (256 MB) and then proceeds to eat
>up my swap memory.  Eventually my machine reboots.  Pretty impressive,
>huh?  Just imagine the damage I could do with 4 lines of code.

Please post the precise code you see doing this.  I'd suspect you've
probably got some kind of recursive function that's keeping the memory
alive.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From ASBUILT at easynet.be  Tue Apr 15 12:10:17 2003
From: ASBUILT at easynet.be (Asbuilt Easynet)
Date: Tue, 15 Apr 2003 18:10:17 +0200
Subject: file read : feature or bug ?
Message-ID: <3e9c2eea$0$6528$afc38c87@sisyphus.news.be.easynet.net>

Hello,

I have a problem to use the read function that seems read more than
requested
as you can see in this sample demo.

Thanks

Jef

# This is a program to read a file in a reverse order by 3 characters block
# the demo.log file is 2 characters per line in windows so ended with CRLF
# 01
# 02
# 03
# 04
#
# we can see it like 01__02__03__04 where __ is CRLF
#
# PROBLEM :
# the program normaly should return
#   _04
#   03_
#   2__
#   __0
#   11
#
# BUT it returns
#   _04
#   03_
#   2_0     <-- From where is coming the 0 ?
#   _02     <-- From where is coming the 2 ?
#   11
#  (this is an extract from rgrep.py)

import string

l_bufsize = 3
l_file    = open("demo.log")
l_file.seek(0, 2)
l_pos      = l_file.tell()

while l_pos > 0:
  l_size = min(l_pos, l_bufsize)
  l_pos  = l_pos - l_size
  l_file.seek(l_pos)
  l_buffer = l_file.read(l_size)
  print "l_buffer = " + l_buffer + "[" + str(len(l_buffer)) +"]"

  l_lines = string.split(l_buffer, "x")
  print "Split = " + str(l_lines)
  del l_buffer





From juneaftn at REMOVETHIShanmail.net  Wed Apr 16 10:50:11 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Wed, 16 Apr 2003 23:50:11 +0900
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com>
Message-ID: <b7jqi7$q1t$1@news.hananet.net>

"Christopher Blunck" <blunck at gst.com> wrote in message
news:1147e466.0304160630.5b510724 at posting.google.com...
> Was just reading a thread about how python implements protected and
> private methods.  I found JP's response quite interesting (this is
> something I've heard numerous friends of mine say of the language):
>
> [In Python] everything is permissible, but not
> everything is opportune.  And that is very much Python's philosophy:
> rather than focusing on trying to make some things impossible (and
> generally failing -- a simple cast in most implementations of C++ lets
> you blast away any "protected" that a silly library designer may have
> tried to impose on you;-), Python empowers and trusts the programmer.
>
>
> The natural response a non-Python programmer has to this statement is,
> "what?!  you __trust__ the programmer?!  that doesn't work in most
> environments."  Despite the ignorance this response demonstrates, I
> none-the-less thought about the statement and the answer I came up
> with was "we write lots of really thorough tests to demonstrate
> functionality."
>
> That got me to thinking:  continuous testing is a cornerstone of the
> XP methodology.  <generalization>And Python programmers typically
> write lots of tests</generalization> (at least more than their Java
> counterparts from what I've seen).  That being said, can that
> generalization be extended to "Python programmers subscribe to the XP
> methodology"?
>
> So how many of you guys use XP processes?

I use. I was in a company where we developed a proteomics research software
with python and extreme programming. My best and happiest time of
development.

Looking at python through the java window is more painful and less useful
than through the smalltalk window. There is a lot to learn from smalltalk
culture; XP was born from the smalltalk land.

June.




From oren-py-l at hishome.net  Thu Apr  3 10:43:46 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Thu, 3 Apr 2003 10:43:46 -0500
Subject: Python commands
In-Reply-To: <3e8bb9c5$1_1@news.iprimus.com.au>
References: <3e8bb9c5$1_1@news.iprimus.com.au>
Message-ID: <20030403154346.GA35581@hishome.net>

On Thu, Apr 03, 2003 at 02:37:26PM +1000, Paul Dexter wrote:
>     Hi all,
>                 I have just discovered Python and I was wondering if anyone
> could help me find an equivalent command in Python for the "OUT" and "INP"
> commands in Qbasic. I wish to convert a Qbasic program to Python.
>     The following snippet is an example of the task I wish to perform;
> 
> base0 = &H378:
> 
> interface:
> OUT base0, &HA0
> OUT base0, &H80
> OUT base0 + 2, &H8
> 
> V = (INP(base0 + 1) AND &H10) * &H8
> 
>     The base address is 378 (lpt) and the "interface" subroutine outputs
> data to the port (378) and "V" is the input data statement I have used. I
> also use the "GETKEY", "LOCATE" and "DRAW" commands in Qbasic
> 
>     Does anyone know how to perform the same tasks/commands in Python or are
> they even available. I have come across other Basic computer language
> variations that do not support the commands I require.
> 
>     Thankyou for any help,

Direct access to I/O ports is usually reserved for device drivers. 

You can try WinIO from http://www.internals.com - it's a library that 
installs a driver to let you poke your hardware directly from user mode.
You can call this library from Python using calldll, ctypes, SWIG and 
similar tools.

    Oren




From just.say.no.to.spam at cox.net  Fri Apr 25 10:58:27 2003
From: just.say.no.to.spam at cox.net (stuffduff)
Date: Fri, 25 Apr 2003 14:58:27 GMT
Subject: Zope-Dev: Curious Problem
Message-ID: <45jiavkpsr5kl41u0ajs2jud1o0qd7rm0d@4ax.com>

Hi,

I have a python external method which reads from an external database,
unpickles a class object, and concatenates some of the object's
properties into a string which is returned by the external method.
The class definition is included in the module.

When I run it from the linux prompt in the Extensions directory it
behaves exactly as it should. The class is successfully unpickled, the
properties are read and the string is returned.

However, when I set it up as an external method and call it I get the
infamous:

Error Type: SystemError
Error Value: Failed to import class MyObject from module __main__

>From my perspective Zope should be 'blissfully unaware' of the
existence of the class or the instantiation of the object.  Since the
object is only accessed in the external method, the module contains
the class definition, and only a string is returned, why is this
happening?

What am I missing here?

Platform: Red Hat 7.3, Python 2.1.3, Zope 2.51

Curious,

Sean

P.S. Also posted to zope-dev but I wanted the 'python' perspective as
well



From andy at wild-flower.co.uk  Tue Apr 22 14:03:24 2003
From: andy at wild-flower.co.uk (Andy Jewell)
Date: Tue, 22 Apr 2003 19:03:24 +0100
Subject: Regarding Dictionaries in python
In-Reply-To: <3EA455DB.9000300@antares.enst-bretagne.fr>
References: <3EA455DB.9000300@antares.enst-bretagne.fr>
Message-ID: <200304221903.24570.andy@wild-flower.co.uk>

On Monday 21 Apr 2003 9:34 pm, Mehta, Anish wrote:
> Hello !
>
> I am having a small query regarding dictionaries in python. From
> Orelly's Programming Python
>
>  >>>D= {'a' : 1, 'b' : 2, 'c' : 3, 'd' : 4}
>  >>>D.items()
>
> [ ('b', 2), ('c', 3), ('d', 4), ('a', 1)]
>
>
> why it is not showing the items in order in which i have inserted.
> Please tell me why it is happing like this.
>
> Thanks in advance.
>
> Regards...


Dictionaries are not ordered entities - check the Tutorial.  They use a 
hashing algorithm to retrieve items, and as items are added, they are stored 
in a 'bucket' that is unique to that item.  The order of the buckets cannot 
be influenced (it's coded into Python).  For details on how hashing 
algorithms work, scan google...

Numeric items are always added in ascending order (as far as my little 
experiments showed anyway):  


>>> d={}
>>> d[1]="one"
>>> d
{1: 'one'}
>>> d[2]="two"
>>> d
{1: 'one', 2: 'two'}
>>> d[3]="three"
>>> d
{1: 'one', 2: 'two', 3: 'three'}
>>> d[4]="four"
>>> d
{1: 'one', 2: 'two', 3: 'three', 4: 'four'}

String values are added in a manner appropriate for sequences:

>>> for l in "abcdef":
             d[l]=l
             print d

	
{'a': 'a'}
{'a': 'a', 'b': 'b'}
{'a': 'a', 'c': 'c', 'b': 'b'}
{'a': 'a', 'c': 'c', 'b': 'b', 'd': 'd'}
{'a': 'a', 'c': 'c', 'b': 'b', 'e': 'e', 'd': 'd'}
{'a': 'a', 'c': 'c', 'b': 'b', 'e': 'e', 'd': 'd', 'f': 'f'}
{'a': 'a', 'c': 'c', 'b': 'b', 'e': 'e', 'd': 'd', 'g': 'g', 'f': 'f'}

I'd imagine tuples are handled similarly, but I haven't tried them...

Hope that helps.

-andyj




From dpugh at uwc.ac.za  Tue Apr 22 05:08:54 2003
From: dpugh at uwc.ac.za (David Pugh)
Date: Tue, 22 Apr 2003 11:08:54 +0200
Subject: Problems compiling _tkinter
Message-ID: <0AD1BE8F-74A2-11D7-9102-0050E4CEF567@uwc.ac.za>

Dear Python people,

    I get the following errors when I try to make python2.2.2:

building '_tkinter' extension
gcc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -fPIC -DWITH_APPINIT=1 
-I/usr/X11R6/include -I. -I/usr/local/Python-2.2.2/./Include 
-I/usr/local/include -I/usr/local/Python-2.2.2/Include 
-I/usr/local/Python-2.2.2 -c /usr/local/Python-2.2.2/Modules/_tkinter.c 
-o build/temp.linux-i686-2.2/_tkinter.o
/usr/local/Python-2.2.2/Modules/_tkinter.c: In function `Merge':
/usr/local/Python-2.2.2/Modules/_tkinter.c:351: warning: passing arg 2 
of `Tcl_Merge' from incompatible pointer type
/usr/local/Python-2.2.2/Modules/_tkinter.c: In function `Split':
/usr/local/Python-2.2.2/Modules/_tkinter.c:383: warning: passing arg 4 
of `Tcl_SplitList' from incompatible pointer type
/usr/local/Python-2.2.2/Modules/_tkinter.c: In function `Tkapp_Call':
/usr/local/Python-2.2.2/Modules/_tkinter.c:627: warning: initialization 
discards qualifiers from pointer target type
/usr/local/Python-2.2.2/Modules/_tkinter.c: In function `SetVar':
/usr/local/Python-2.2.2/Modules/_tkinter.c:935: warning: assignment 
discards qualifiers from pointer target type
/usr/local/Python-2.2.2/Modules/_tkinter.c:947: warning: assignment 
discards qualifiers from pointer target type
/usr/local/Python-2.2.2/Modules/_tkinter.c: In function `GetVar':
/usr/local/Python-2.2.2/Modules/_tkinter.c:988: warning: assignment 
discards qualifiers from pointer target type
/usr/local/Python-2.2.2/Modules/_tkinter.c:991: warning: assignment 
discards qualifiers from pointer target type
/usr/local/Python-2.2.2/Modules/_tkinter.c: In function 
`Tkapp_SplitList':
/usr/local/Python-2.2.2/Modules/_tkinter.c:1197: warning: passing arg 4 
of `Tcl_SplitList' from incompatible pointer type
/usr/local/Python-2.2.2/Modules/_tkinter.c: In function 
`Tkapp_CreateCommand':
/usr/local/Python-2.2.2/Modules/_tkinter.c:1357: warning: passing arg 3 
of `Tcl_CreateCommand' from incompatible pointer type
gcc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -fPIC -DWITH_APPINIT=1 
-I/usr/X11R6/include -I. -I/usr/local/Python-2.2.2/./Include 
-I/usr/local/include -I/usr/local/Python-2.2.2/Include 
-I/usr/local/Python-2.2.2 -c 
/usr/local/Python-2.2.2/Modules/tkappinit.c -o 
build/temp.linux-i686-2.2/tkappinit.o
gcc -shared build/temp.linux-i686-2.2/_tkinter.o 
build/temp.linux-i686-2.2/tkappinit.o -L/usr/X11R6/lib -L/usr/local/lib 
-ltk8.4 -ltcl8.4 -lX11 -o build/lib.linux-i686-2.2/_tkinter.so
WARNING: removing "_tkinter" since importing it failed

The only other errors in the make are some warnings about tempnam_r in 
posixmodule.c.

Any ideas?

David
--------------------------------
Dr DJR Pugh
Department of Biotechnology
University of the Western Cape
Private Bag X17
Bellville  7535
South Africa
tel: 2721 959 2211
fax: 2721 959 3505
email: dpugh at uwc.ac.za





From dave at boost-consulting.com  Mon Apr 21 18:18:49 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Mon, 21 Apr 2003 18:18:49 -0400
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com>
 <a3b5b608.0304080302.c26ad68@posting.google.com>
 <af6b5bdd.0304181001.558da6f@posting.google.com>
 <84of317cd6.fsf@boost-consulting.com> <b7vlfb$27c$0@216.39.172.122>
 <84of30m4cp.fsf@boost-consulting.com>
 <p0p8avsd7nlqu52p4rnjobuemctd3foil8@4ax.com>
Message-ID: <848yu3o5o6.fsf@boost-consulting.com>

Mark E. <snowball3 at bigfoot.com> writes:

> David Abrahams <dave at boost-consulting.com> wrote:
>
>>    xml.sometag(attr1="value1", attr2="value2", ...)[ "contents" ]
>
> Of course there are always corner cases to deal with. For example, what's a good way
> to handle XHTML's 'class' attribute?
>
> xml.sometag(class="css_class_name")["contents"]
> won't work.
>
> One could map 'foo_' to 'foo' to handle conflicting names used in xml/html, or allow
> a dictionary to serve the same purpose:
>
> xml.sometag({"attr1" : "value1", "class" : "value2"})["contents"]

Sure, you have lots of options. Also:

    xml.sometag(class_("value2"), attr1 = "value1")["contents"]

    xml.sometag({'class':"value2", 'while':"value3"}, attr1 = "value1")["contents"]

    xml.sometag(attr1 = "value1", _ = {'class':"value2", 'while':"value3"})["contents"]

it's all in the fun of designing domain-specific sublanguages.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From pinard at iro.umontreal.ca  Fri Apr 18 08:01:46 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 18 Apr 2003 08:01:46 -0400
Subject: Do any SGML lib good for Python?
In-Reply-To: <m3ptnk9p41.fsf@mira.informatik.hu-berlin.de>
References: <2778697.1050595278@dbforums.com>
	<m3of35ni8d.fsf@mira.informatik.hu-berlin.de>
	<2781052.1050630152@dbforums.com>
	<m3ptnk9p41.fsf@mira.informatik.hu-berlin.de>
Message-ID: <oq3ckgauad.fsf@titan.progiciels-bpi.ca>

[Martin v. L?wis]

> dolephi9080 <member27604 at dbforums.com> writes:

> > What is your mean parsing the output of nsgmls?

> Invoke nsgmls, and see for yourself.

To that task, reading `nsgmls' documentation also helped me! :-)

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From aleax at aleax.it  Thu Apr 24 17:14:24 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 24 Apr 2003 21:14:24 GMT
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com>
Message-ID: <QsYpa.3418$K35.76394@news2.tin.it>

Peter Hansen wrote:
   ...
> I could have said that, but I know Aahz knows when something is really
> throwaway, and there are such things, so I thought I'd try answering

Well, then Aahz is wiser and more perceptive than me, because _I_
don't know -- I've noticed that often what I *believed* I would throw
away, I end up reusing instead.  SOME things i eventually do throw
away (or lose track of...:-), but I can't predict which ones in advance.

> the question directly for a change.  ;-)  And I did emphasize that this
> was for _literally_ throwaway scripts, which do exist.  If nothing else,
> you could call it a spike solution which is expected to be used, once
> only, then tossed.  If it turns out it's got to be used again, then it
> should be re-written anyway so it's not a serious problem that it lacks
> tests.

Why "should" something be rewritten rather than tweaked if it already
does 90% of what you now need?  Pah.  Of COURSE I'm just gonna tweak
it, because I'm in a hurry (I always am) and that will save me lots of
time.  So -- it IS a problem that it lacks tests.

> throwaway than the ones I "know" are not.  If I'm wrong... no big deal:
> I rewrite as a real script with proper tests.  Since the script was
> only twenty or thirty lines of code (any more and it could not possibly
> be throwaway, right?) it isn't really a big deal.

Wrong.  "Throwaway" means "I [think I] am going to need to run this
only once".  Whether it's 30 or 60 lines has nothing to do with it.


Alex





From dl_un-spam_rubin at yahoo.com  Wed Apr  9 15:43:39 2003
From: dl_un-spam_rubin at yahoo.com (Daniel)
Date: Wed, 9 Apr 2003 12:43:39 -0700
Subject: How to free a temporary clob in DCOracle2?
Message-ID: <b71t5c$d3f$1@news.Stanford.EDU>

How do I free the temporary clob when I'm done with it in DCOracle2? I get
an error when I try to close the temporary clob:

>>> clob = connection.LobLocator('SQLT_CLOB')
>>> clob.write('<f>foo bar</f>')
30
>>> d='insert into foo values (XMLTYPE.createXML(:m))'
>>> connection.execute(d, clob)
4
>>> clob.close()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
dco2.DatabaseError: (22289, 'ORA-22289: cannot perform %s operation on an
unopened file or LOB')


Thanks,
Daniel





From ckea25d02 at sneakemail.com  Tue Apr 29 22:01:16 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 30 Apr 2003 14:01:16 +1200
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
In-Reply-To: <slrnbajbre.e5n.ramen@lackingtalent.com>
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <slrnbajav2.e5n.ramen@lackingtalent.com> <pan.2003.04.25.17.49.27.407432@jerf.org> <slrnbajbre.e5n.ramen@lackingtalent.com>
Message-ID: <b8n9v6$bog4h$1@ID-169208.news.dfncis.de>

Dave Benjamin wrote:
> In article <pan.2003.04.25.17.49.27.407432 at jerf.org>, Jeremy Bowers wrote:
> 
>>Implementation of the brain object left as an exercise for the reader. I
>>know it can be done, as I am in possession of such an object right now.
> 
> It'd make for one hell of a base class.

And just think what you could achieve with multiple
inheritance! It could give the phrase "two heads are
better than one" a whole new meaning.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ipellew at pipemedia.co.uk  Wed Apr  2 10:15:28 2003
From: ipellew at pipemedia.co.uk (Ian Pellew)
Date: 2 Apr 2003 07:15:28 -0800
Subject: Compiled Python on Solaris8  at 2.2.2 - Problem with (_)socket
Message-ID: <30875970.0304020715.73954878@posting.google.com>

Hi;

I have compiled python 2.2.2 on Solaris :-
SunOS sopix 5.8 Generic_108528-12 sun4u sparc SUNW,UltraAX-i2

I have a problem with python trying to import "_socket"
Here is my problem:-
./python                                                      
Python 2.2.2 (#2, Apr  1 2003, 10:10:26) 
[GCC 3.1] on sunos5
Type "help", "copyright", "credits" or "license" for more information.
>>> import os as o
>>> import cgi as i
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/home/Ipellew/tmp/Python2.2/Python-2.2.2/Lib/cgi.py", line 39, in ?
    import urllib
  File "/home/Ipellew/tmp/Python2.2/Python-2.2.2/Lib/urllib.py", line 26, in ?
    import socket
  File "/home/Ipellew/tmp/Python2.2/Python-2.2.2/Lib/socket.py", line 41, in ?
    from _socket import *
ImportError: No module named _socket
>>> import socket
>>> from socket import *
>>>

My "socket" works OK, but not "_socket"

Can someone please advise what to do.

Regards
Ian



From abpillai at lycos.com  Mon Apr  7 02:14:32 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 6 Apr 2003 23:14:32 -0700
Subject: Emacs python-mode problem
References: <m2of3ktq67.fsf@db.hit.edu.cn>
Message-ID: <bd993a2f.0304062214.2027bed8@posting.google.com>

Chunyu <dddkk at sina.com> wrote in message news:<m2of3ktq67.fsf at db.hit.edu.cn>...
> While debugging a C/C++ program with GDB in emacs (version 21.3.50),
> there is an arrow symbol `=>' in the source window to indicate the
> current line normally. But After I load the python-mode.el file from
> my ~/.emacs.el, the arrow symbol disappear while debugging with GDB in
> emacs. (The python-mode.el is from
> http://www.python.org/emacs/python-mode, version 4.6.)
> 
> I load the python in my ~/.emacs.el by
> (require 'python-mode)
> (setq auto-mode-alist
>       (cons '("\\.py$" . python-mode) auto-mode-alist))
> (setq interpreter-mode-alist
>       (cons '("python" . python-mode) interpreter-mode-alist))
>

     In my .emacs file I see the following code.
(setq auto-mode-alist
       (append '(("\\.C$"  . c-mode)
                ("\\.cc$" . c++-mode)
                 ...
                 ...
                 ("\\.py$" . python-mode)) auto-mode-alist)

Perhaps you can try that form.

Btw you dont need to 'require python-mode or for that matter any
of the mode-specific functions before your definition of auto-mode-alist.
The 'require' call can come later, no problem.
 
> Now, I have to comment these lines, and load python-mode.el by hand
> when I want to use python-mode. How can I solve it? Is there another
> better python-mode code? Thanks for your replay.

Thanks

Anand



From aleax at aleax.it  Thu Apr 17 05:51:56 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 09:51:56 GMT
Subject: Web Hosting
References: <2rjna.171$hT2.182473@news2.news.adelphia.net> <6Bjna.111823$It5.95226@news2.central.cox.net> <e5os9vg7aj3gfek0abt05s1bjh9pfvfvb3@4ax.com>
Message-ID: <0Puna.38530$T34.1097507@news2.tin.it>

Afanasiy wrote:
   ...
> I was under the impression mod_python was not safe for shared
> environments. If this is true, I don't think you will find a

I'm not aware of such concerns; could you please provide an
URL discussing the issue, or what gave you this impression?

> reasonably price host which is also a good host.
> 
> If it is not true, I wonder why pair.com will not run mod_python.

No idea, but inertia is often a helpful explanation.

> CGI, on the other hand, is CGI. :p

Sure, but that's often not too bad anyway (particularly if you
can run Webware in an external process and just use the CGI
adapter to connect to Webware...).


Alex




From staschuk at telusplanet.net  Thu Apr  3 23:28:25 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 3 Apr 2003 21:28:25 -0700
Subject: Finding field widths of floats for neat printing
In-Reply-To: <2830c89c.0304030645.51ad149c@posting.google.com>; from andrew.gregory@npl.co.uk on Thu, Apr 03, 2003 at 06:45:25AM -0800
References: <2830c89c.0304030645.51ad149c@posting.google.com>
Message-ID: <20030403212825.B2938@tibia.amotlpaa.bogus>

Quoth Andrew Gregory:
> I want to write lists of floating point numbers to text files in neat
> columns with decimal points aligned, right-hand edges straight (rather
> than ragged), and without unnecessary trailing zeros, e.g.
> 1.1    -34.67     0.006  
> 2.0     -1.05     0.010  
> etc.
  [...]

How about this?

    def strparts(value):
        s = ('%f' % value).rstrip('0').split('.')
        if len(s) < 2:
            s.append('')
        return s

    def formatfor(values):
        leftparts, rightparts = zip(*map(strparts, values))
        leftwidth = max(map(len, leftparts))
        precision = max(map(len, rightparts))
        width = leftwidth + precision
        if precision:
            width += 1 # decimal point
        return '%%%d.%df' % (width, precision)

    table = [[1.1, -34.67, 0.006], [2, -1.05, 0.01]]
    formats = map(formatfor, zip(*table))
    for row in table:
        for fmt, val in zip(formats, row):
            print fmt % val,
        print

Output:

    1.1 -34.67 0.006
    2.0  -1.05 0.010

-- 
Steven Taschuk                                7\ 7'Z {&~         .
staschuk at telusplanet.net                        Y r          --/hG-
                                            (__/ )_             1^1`




From Greg.Lindstrom at acxiom.com  Tue Apr 15 13:04:55 2003
From: Greg.Lindstrom at acxiom.com (Lindstrom Greg - glinds)
Date: Tue, 15 Apr 2003 12:04:55 -0500
Subject: Printing Columns
Message-ID:     <ABEE81BE08ADD311B6A1009027DE908D16A949D6@conmsx03.corp.acxiom.net>

Hello-
I have been working on a little project and have come across an interesting
(to me) problem.  I have a list of stuff that I would like to print to
stdout in columns.  Now, it's no problem to print it out as follows:

myList = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H']

	A	B	C
	D	E	F
	G	H

but I would like

	A	D	G
	B	E	H
	C	F

problem is, the code (all 12 lines) is a bit convoluted and would not be
easy for the next person using it to figure out (and in my mind,
maintainability is priority one; that's why I use Python).  I would like to
know if there is a "pythonic" way to do this. So, here are the specs....

Given a list of length N with 0<=N<=infinity and a number of columns, C (C>0
and may be less than N), how do we print out the elements of the list in the
above (column major?) form?

Thanks for your thoughts....I'll post mine in a day or two.
--greg





**********************************************************************
The information contained in this communication is
confidential, is intended only for the use of the recipient
named above, and may be legally privileged.
If the reader of this message is not the intended
recipient, you are hereby notified that any dissemination, 
distribution, or copying of this communication is strictly
prohibited.
If you have received this communication in error,
please re-send this communication to the sender and
delete the original message or any copy of it from your
computer system. Thank You.





From peter at engcorp.com  Wed Apr  2 22:19:12 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 22:19:12 -0500
Subject: ANNOUNCE: pyGDChart2 alpha
References: <mailman.1049336408.10652.python-list@python.org>
Message-ID: <3E8BA830.277CCDA@engcorp.com>

Aldo Cortesi wrote:
> 
> pyGDChart2 itself can be found at:
> 
>         http://www.nullcube.com/software/pygdchart2.html

This link is on that page, but is missing:

http://www.nullcube.com/software/software/pygdchart2.html

-Peter



From artyprog at wanadoo.fr  Thu Apr 17 08:36:54 2003
From: artyprog at wanadoo.fr (Salvatore)
Date: Thu, 17 Apr 2003 14:36:54 +0200
Subject: variable in Python
In-Reply-To: <pBwna.114820$It5.55830@news2.central.cox.net>
References: <b7m1g4$qqb$2@news-reader11.wanadoo.fr> <pBwna.114820$It5.55830@news2.central.cox.net>
Message-ID: <b7m77d$blq$1@news-reader13.wanadoo.fr>

Steve Holden wrote:
> "Salvatore" <artyprog at wanadoo.fr> wrote ...
> 
>>Hello,
>>
>>I have a little doubt concerning variables.
>>When one declare x = 3
>>does x contains the value 3 or
>>does x acts like pointer in C ?
>>
> 
> 
> Salvatore:
> 
> All Python variables are references to the data values that have been
> assigned to them - as are list and dictionary elements. This can lead to
> surprises for beginners, who don't understand why
> 
> 
>>>>lst1 = [1, 2, 3]
>>>>lst2 = lst1
>>>>lst2[2] = "bam!"
> 
> 
> leads to the lst1 variable showing the value [1, 2, "bam!"]. The answer, of
> course, is that the third assignment changes the last element of the list
> that is referred to by both variables, lst1 and lst2.
> 
> Many Python authors prefer to talk about "binding" nowadays, since
> "assignment" implies to users of other languages that a value is being
> placed into the variable. In Python, however, this isn't the case. Data
> values are created in memory allocated from a heap, and this memory is only
> reclaimed when no references to the data vablue remain. Once the data value
> is created, variables reference it.
> 
> In  the CPython implementation each reference to a value increments its
> reference count, and deletion of a reference decrements that count. Certain
> "cyclic" structures, being self-referential, would never be reclaimed, and
> so recent versions of CPython have a garbage collection algorithm that can
> detect otherwise-unreferenced cyclic structures.
> 
> The JPython implementation, which compiles Python into Java bytecodes, does
> not use reference counting at all and relies on Java's garbage colection
> scheme to reclaim unused memory.
> 
> regards
> --
> Steve Holden                                  http://www.holdenweb.com/
> How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
> Python Web Programming                 http://pydish.holdenweb.com/pwp/
> 
> 
> 
> 
That's what i call 'a real response' ;-)

Thank you Steve




From mwatts3 at stny.rr.com  Wed Apr  2 16:15:40 2003
From: mwatts3 at stny.rr.com (Mark Watts)
Date: Wed, 02 Apr 2003 21:15:40 GMT
Subject: Python, XML and ECMAScript for XML?
Message-ID: <0qIia.2653$ii.1781@twister.nyroc.rr.com>

I am fairly new to python so please forgive me if this has already been
asked and answered even though the topic is pretty new.  Appropriate
googling has been done....

This article discusses some languages changes made by BEA to ECMAScript
(JavaScript) that permit easier XML manipulation.  I am NOT an XML
expert but reading the article, it sure seems that the changes might
actually make XML easier to handle.  These changes have been submitted
to ECMA for integration into the standard and ECMA appears to have
agreed to add the changes although we won't know until the new spec
comes out Q1 2004.

http://dev2dev.bea.com/articles/JSchneider_XML.jsp

The ECMA announcement can be found at 

http://www.ecma-international.org/news/ECMA%20E4X%20Final%20Final.pdf


My question:  Has anyone looked at doing something similar in Python or
has it already been done.


--
Thanks
- Mark



From ckea25d02 at sneakemail.com  Mon Apr 14 01:02:08 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 14 Apr 2003 17:02:08 +1200
Subject: Adding item in front of a list
In-Reply-To: <CE0ma.5922$LB6.169206@news1.tin.it>
References: <mailman.1050038237.25551.python-list@python.org> <ZUXla.7029$T34.191442@news2.tin.it> <mailman.1050176305.12881.python-list@python.org> <CE0ma.5922$LB6.169206@news1.tin.it>
Message-ID: <b7deoe$devs8$1@ID-169208.news.dfncis.de>

Alex Martelli wrote:
> I *hate* the few places in the Nutshell in which
> I couldn't explain a function's behavior by means of Pythonic "optional
> arguments", but rather had to resort to brackets meaning "optional"

Maybe Python should in general allow optional arguments to
be on the left instead of the right:

   def myrange(start = 0, stop):
     ...

and then range() and insert() wouldn't be anomalous!

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From xxdigitalhellxx at hotmail.com  Wed Apr  2 09:41:33 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Wed, 02 Apr 2003 08:41:33 -0600
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com>
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <b6esse$pge$1@heckle.cc.uregina.ca>

Ben Hutchings wrote:
> In article <x7y92uhzpw.fsf at guru.mired.org>, Mike Meyer wrote:
> <snip>
> 
>>    3.  At most three Is, Xs and Cs may appear in any given literal.
> 
> <snip>
> 
> OK cleverclogs, how do we write 39?

Although I honestly don't know how, I'm sure it's possible. Here's how I 
would do it, although it's probably not the right way. My roman numerals 
are a bit rusty. :)

XIL

?

-- 
Ryan




From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Sat Apr  5 14:53:23 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Sat, 05 Apr 2003 21:53:23 +0200
Subject: Distributed environment
In-Reply-To: <3e8d1edc$0$235@hades.is.co.za>
References: <3e8d1edc$0$235@hades.is.co.za>
Message-ID: <3e8f3430$0$49112$e4fe514c@news.xs4all.nl>

Gerrit van Dyk wrote:
> Hi all,
> 
> I am in the process of designing a system that will need to use a 
> distributed environment. 

More choice is better, I think, so I'd like to point you to Pyro.
http://pyro.sourceforge.net

Pyro has a topic-based publish-subscribe Event Service that might
serve your needs. If not, you could roll your own server 'hub'
quite easily.


--Irmen de Jong




From imbosol-1050288268 at aerojockey.com  Mon Apr 14 01:52:31 2003
From: imbosol-1050288268 at aerojockey.com (Carl Banks)
Date: Mon, 14 Apr 2003 05:52:31 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com>   <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk>
Message-ID: <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net>

Alexander Schmolck wrote:
> Carl Banks <imbosol-1050078176 at aerojockey.com> writes:
>> Alexander Schmolck wrote:
>> > Carl Banks <imbosol-1049990205 at aerojockey.com> writes:
>> >
>> > You did notice, of course, that you don't only need to change that particular
>> > function foo, but also everything (directly or indirectly) calling foo? As
>> > well as everything that is called (directly or indirectly) by foo and that
>> > uses this global state information? Including 3rd party code and builtin
>> > constructs of the language?
>> 
>> Yes.  Since the idea here is to rewrite the Emacs in Python, you have
>> to rewrite all of this anyways.
> 
> I'm simply interested in the question as to whether globals of the
> type C, Java and most languages, but specifically python, have them
> are a good idea or whether dynamically scoped variables are
> superior. So I'd prefer not get too sidetracked by Emacs and its
> uses of dynamic scope (some bad, some good, IMO).  The above
> describes by no means a problem limited to refactoring code in
> dynamically scoped legacy languages but anyway, since you bring it
> up: there is rewriting and rewriting -- not that anyone who's got
> the necessary ressources would be stupid enough to rewrite emacs
> anyway[1] -- I'd personally rather rewrite leveraging more rather
> than fewer powerful abstractions.

Ok, let's leave Emacs out of this.


>> >> I think dynamic scpoing is a wholly bad idea, and simplifying a very
>> >> misguided programming practice doesn't make it any less so.
>> > 
>> > How much time did you actually spend thinking about global variables and
>> > dynamic scope or what I wrote before you posted?
>> > 
>> > Why are global variables evil?
>> [snip reasons globals are evil]
>> 
>> I didn't say globals are evil. I said Globals are Evil(tm). Big difference.
>> (You should learn the meaning of (tm); it means the writer fully realizes
>> that it's only a general truth, even though others consider it an absolute
>> truth.)
> 
> I can't see were I attribute the claim that "globals are evil" to
> you -- *I* happen to think globals of the broken kind python suffers
> from are evil (plain and simple, no (tm)) and dynamic scope would
> provide a much better solution.  I'm always willing to hear why I'm
> wrong.
> 
>> I do not oppose global variables in general.
> 
> That was never in doubt. The question is not so much whether or not
> one should have implicitly shared variables (in python and similar
> languages), but how they should work and how they should be
> used. Your answer to my question on your opinion on why python
> provides global variables seems to be (correct me if I'm wrong):
> 
> I) to provide *one*[2] implicitly shared state that should only be
> altered in an irrecoverable fashion.
> 
> My answer would be:
> 
> II) to provide a poor substitute for implicitly shared state that
> can be shared at different granularities and so that the same name
> can simultanously be exposed with different values to different
> values to different parts of the program -- all with automatic
> stack-popping style recovery of former values on natural or
> unnatural block exit.

Let's leave Python (and C and Java) globals out of this.


>> I do oppose global variables where there are situations where it feels
>> necessary to change it and restore it.  That tells you right away that
>> your functions should not use a global, but should have it passed in.
>> Designing such a function to refer to the global anyways is Bad
>> Programming(tm).
> 
> Repetition doesn't make this claim doesn't make it more or less
> true. Why not instead back it up in some way or attack a provided
> example where I feel that you *do* want to modify and restore state?

I didn't think I needed to back it up nor attack an example.  We all
know Globals are Evil(tm), and passing a parameter into a function
with a global is therefore Bad Programming(tm).  This *is* usually
true.

It thought it was obvious.

But see below.


> You say:
> 
>> These issues should be handled by global variables if they indeed are
>> global in nature.  The stuff you mentioned usually is.
> 
> and then:
> 
>> I have not said lexical scoping handles the issues you have referred
>> to better than dynamic scoping.  I have said that the issues you refer
>> are only issues in the presense of Bad Programming(tm).  Anyone using
>> globals discriminately will not have such issues to deal with.
> 
> I infer this to mean that it's OK to have logging and to use global
> state to control logging, as long as I don't want one part of the
> program to be treated differently. So to come back to an example
> from my previous post, a particularly critical routine that should
> be logged at a higher level without affecting the rest of the
> logging operations, would be Bad Programming(tm)?  Same for output
> precision, assertions and other debugging options (or localized
> fine-tuning of e.g. floating point behavior or error handling etc.)?

Ok, let me back off a bit.  That you might want to change a program's
state in a dynamic scope is reasonable.

Typically, however, that kind of thing should be handled by passing
objects around as arguments.  For example, consider the use of current
buffer in Emacs.  (We are still leaving Emacs issues out of it; I'm
just using it as an example of bad programming.)  Most functions in
Emacs operate on the current buffer, a piece of global state.  In
order to operate on a non-current buffer, you have to temporarily
change the current buffer.  That is bad programming; functions should
take the buffer to operate on as an argument.  Some people might whine
that the absense of dynamic scope would make life difficult, but, in
fact, it is their own incompetence and/or laziness that makes life
difficult.

However, as you say, there are certain side concerns, like logging,
where it is at best inconvenient, and at worst counterproductive, to
pass a value around as an argument, so it gets made a global.  It's
very common to do this, of course.

However, I think changing a global value within a dynamic scope, in
situations where it wouldn't be better to pass around an argument, is
a very uncommon concern.  And frankly, I think it's too much of a
special case to introduce a confusing and separate scoping rule that
has all kinds of nasty and big-ridden problems.

So if you need that kind of thing, please don't destroy the language
by adding dynamic scoping.  Just live with a try ... finally.

Or get off your lazy ass and refactor your code like you should.


>> I have not said lexical scoping handles the issues you have referred
>> to better than dynamic scoping.  I have said that the issues you refer
>> are only issues in the presense of Bad Programming(tm).  Anyone using
>> globals discriminately will not have such issues to deal with.
> 
> A bold claim. I bet a few people in the Lisp community would disagree that
> their use of dynamic variables reflects their limited coding and design
> abilities.

Retracted.


> Do you ever use try-finally?
 
Yes, but not to change any program-visible globals.

(Yes, I know open affects process file decriptor tables, but all my
code sees is a file object which is a local variable, that I pass as
an argument to any function I call.  Almost everything I use try
... finally for is stuff that the OS forces to be part of the
process's global state.)


>> But to answer your question: yes, dyanmic scoping is so awful it
>> doesn't justify this particular benefit, especially when the benefit
>> only arises in Bad Programming(tm).
> 
> Indeed, wickedness should be its own reward -- if only we could
> agree that Bad Programmers(tm) are the sole beneficiaries.
>
>> In fact, dynamic scoping is so awful is doesn't justify any benefits.
> 
> Have you had traumatic childhood experiences with early perl or lisp
> dialects?  :)

Yes


[snip stuff about many people agreeing with me and it doesn't mean anything]
> Also note that, as pointed out in Beni Cherniavsky's informative
> posting, python already shows various signs of dynamic infiltration
> such as exceptions.

I don't think exceptions matter one bloody bit when it comes to
scoping issues.


> [I assume the following was written under the impression that I advocated
> global scope for all variables, but I'll reply anyway just in case]
> 
>> I don't wish to go into detail here.  The main problems are: It forces
>> you to use defensive naming schemes, even for your local variables.
> 
> Not if you have lexical variables as default.

Ok.  Let's not forget that, lexical locals or not, this is still a
problem for dynamic globals.

With all globals in dynamic scope, software packages that declare
globals must take steps to avoid name clashes.  Or, you can make a
pitiful attempt to work around this such as the absolutely horrid
abomination known as the Lisp Package system.


>> You cannot determine the variables in effect just by looking at a function
>> in context. You cannot do lexical closures. It's less efficient.
> 
> All equally true of globals as they are in python. BTW, not that it matters,
> but dynamically scoped variables may well be more efficient than passing
> parameters (the reason that it doesn't matter is of course that such design
> decisions shouldn't be influenced by minor performance characteristics).

The first is kind of true in Python; but not generally true of
lexically-scoped variables.  Even in Python, when you see a global,
you know at least what module it belongs to.  When you see
dynamically-scoped variable, you can't tell what it is or who set it.
Python, as it is, is still much better in this regard.



Closing remarks:

You seem to acknowledge the existence of the well-known problems with
dynamic scoping (because you agree lexical scoping is better for
locals).  But, you seem to have ignored the fact that these problems
don't go away just because the variable in question is global.  You
say globals should be dynamically scoped because they can do this
neat, funky trick, but never consider the excess baggage the dynamic
scoping brings.

Making all globals dynamically scoped is a mistake because of the
problems dynamic scoping causes, which, I say, are far more severe
than the slight inconvenience of having to explicitly save and restore
a value in the few special cases where that would be useful (and not
bad programming).

I say dynamic scoping shouldn't be in the language, period, because,
despite your good intentions, someone will put it to evil use.  Ok,
you can say that about any language feature, but this is an evil I
cannot bear.  Having such a dispicable wart in the language is simply
not worth saving you the inconvenience of using a try ... finally to
change your log level.


-- 
CARL BANKS



From skip at pobox.com  Wed Apr 16 11:06:43 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 16 Apr 2003 10:06:43 -0500
Subject: reading Unix mailbox
In-Reply-To: <K1ena.17601$DN.484833@tornado.fastwebnet.it>
References: <hZbna.17434$DN.480144@tornado.fastwebnet.it>
        <mailman.1050499476.11070.python-list@python.org>
        <K1ena.17601$DN.484833@tornado.fastwebnet.it>
Message-ID: <16029.29059.271690.152161@montanaro.dyndns.org>

    >> from spambayes import mboxutils
    ...

    Alessio> But that requires to have spambayes, doesn't it?

Yes, but it's available from SourceForge cvs and runs with 2.2.x.

Skip




From squan at web.de  Thu Apr 24 10:31:57 2003
From: squan at web.de (Stefan Quandt)
Date: 24 Apr 2003 07:31:57 -0700
Subject: Problem with PyQt on SuSE 8.2
References: <b842hl$dsa$1@online.de> <b844nl$fst$1@online.de>
Message-ID: <ae5c48d7.0304240631.465b9b0e@posting.google.com>

Birgit Lachner <schulz.online at gmx.de> wrote in message news:<b844nl$fst$1 at online.de>...
> > we try ro run several Python programs which are using PyQt on SuSE 8.2.
> > Every time the same error message occurs:
> >  :
> > We tried the original PyQt packages which comes with SuSE 8.2. We also
> > have compiled PyQt from the source RPMs, but the same error occurs.
> > 
> Okay ... as we compiled the original riverbank-sources the programms are
> running now.
> 
> Bye ... Birgit & Bernd Lachner

I try to compile the PyQt-x11-gpl-3.5 sources from riverbank but build.py gives me
many X11 related linker errors from when it tries to link a (generated) qttest.cpp
(also on SuSE 8.2).

How did you do this??



From jdhunter at ace.bsd.uchicago.edu  Fri Apr 25 08:57:18 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Fri, 25 Apr 2003 07:57:18 -0500
Subject: GUI libs
In-Reply-To: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> (Alessio Pace's
 message of "Fri, 25 Apr 2003 11:36:01 +0200")
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
Message-ID: <m2ptnazqdt.fsf@mother.paradise.lost>

>>>>> "Alessio" == Alessio Pace <puccio_13 at yahoo.it> writes:

    Alessio> Hi, I'd like to make a GUI for a program written in
    Alessio> python, meant to run on mac, win and linux: which is the
    Alessio> preferable solution? I saw around tkinter, wxpython and
    Alessio> pyqt....

pygtk runs on windows, unix and mac os and is an excellent and mature
GUI toolkit.  It also supports glade, a visual GUI rapid application
design tool.  If you've ever run the GIMP, you have an idea of how
sophisticated GTK applications can be.  I also think that pygtk apps
are more visually appealing than Tkinter apps.  Certainly wx is
powerful and makes great looking apps, but make sure it can handle
printing on your platforms of choice before committing -- last time I
checked, printing on some platforms was still a work in progress.

  Homepage: http://www.daa.com.au/~james/pygtk/
  Docs: http://www.gnome.org/~james/pygtk-docs/
  Tutorial: http://www.moeraki.com/pygtktutorial/pygtk2tutorial/index.html

Here's a screenshot of an application I'm working on in pygtk

  http://nitace.bsd.uchicago.edu:8080/files/share/loc3d_jr.png


Cheers,
John Hunter




From rogue9 at ntlworld.com  Wed Apr 23 19:22:18 2003
From: rogue9 at ntlworld.com (Rogue 9)
Date: Thu, 24 Apr 2003 00:22:18 +0100
Subject: Creating dynamically named lists/dictionaries
Message-ID: <ydFpa.15$IO5.10324@newsfep2-gui.server.ntli.net>

Hi all,
	I'm new to both programming and Python in particular and have run in to a
snag while trying to write my own program.I have looked through books and
the web but haven't as yet found an answer so I though I might try a post
here to see if anyone could point me in the right direction.
	Okay my program is for analysing a Lottery Draw results text file I have
pre-formatted so that each line contains 7 comma separated values (Draw
number and six values representing the numbers drawn in numerically
ascending order.
	I've sussed out how to open the file read in the lines and I have split
the strings first by the '\n' newline symbol and then I have split it
again by ','.I've also turned the values back into integers as they were
strings.Now what I've ended up with is a list containing one draw result
for each week (768 at the last count)as individual items in that list e.g
list =[[767,11,22,33,44,45,46]......[1,20,33,41,45,46,47]]
	What I really want is a list or dictionary for each individual weeks
result but I can't seem to work out how to dynamically create a list or
dict with a unique name for each week.The results file will grow and I
don't fancy creating 768+ lines of code just to instantiate a list or
dict.At first I tried looping a variable equal to the length of the file
as in:
	x=len(filelength)
	'draw'+str(x)=[drawresult]
I thought this would dynamically make a whole bunch of lists such as
draw1,draw2 etc. but all I got was errors.
	I hope I have explained my 'problem' sufficiently for someone to help or
to at least point me in the right direction to work towards a solution.
	Yours hopefully,
		Rogue9



From cbuffer at removethis.rosecott.ukfsn.org  Sun Apr 27 18:04:47 2003
From: cbuffer at removethis.rosecott.ukfsn.org (Ken Parkes)
Date: Sun, 27 Apr 2003 23:04:47 +0100
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com>
Message-ID: <1051481042.66194.0@doris.uk.clara.net>

On Sun, 27 Apr 2003 09:57:18 +0000, Steven Hodgen wrote:

> Hello,
> 
> I'm currently teaching a high-school introduction to programming course.
> I'm currently teaching in Java, but am seriously considering changing to
> Python. 


Have a look at "Learn to Program Using Python"  by Alan Gauld 
pub, Addison-Wesley.
I'm about half-way through the book, with a background of modest late
seventies Basic.  No problems so far but of course it depends on where you
hope to get to.

Ken.



From upclick at upclick.com.br  Wed Apr  2 03:42:51 2003
From: upclick at upclick.com.br (Sites para empresas de RH)
Date: Wed, 2 Apr 2003 05:42:51 -0300
Subject: Boletim UP Click - Resultado de Busca - Seu site
Message-ID: <41297-2200343284251290@vaio>

	
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030402/bdaf09f2/attachment.html>

From mcfletch at rogers.com  Sun Apr 13 21:05:39 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 13 Apr 2003 21:05:39 -0400
Subject: properties & exceptions
In-Reply-To: <3e9a0144$1@news.microsoft.com>
References: <3e9a0144$1@news.microsoft.com>
Message-ID: <3E9A0963.1040300@rogers.com>

Any chance you don't have a __metaclass__ = type declaration?  You don't 
subclass object in the example you cite, and without one of those 
properties just don't work.

class FOO(object):

HTH,
Mike


Saveen Reddy [MSFT] wrote:
...

>class FOO :
>  
>
...

> bar1 = property ( _getbar1 )
> bar2 = property ( _getbar2 )
>  
>
...

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From gh at ghaering.de  Tue Apr 15 04:36:07 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Tue, 15 Apr 2003 10:36:07 +0200
Subject: vim configuration help
In-Reply-To: <b7gc3g$lb54$1@ID-141451.news.dfncis.de>
References: <P4Oma.78974$rd4.2975295@twister.austin.rr.com> <b7gc3g$lb54$1@ID-141451.news.dfncis.de>
Message-ID: <3E9BC477.4090408@ghaering.de>

Karl Scalet wrote:
> Hi Randall,
> 
> Randall Smith wrote:
> 
>> I'm learning Python and I've been looking for a good IDE to use with 
>> it.    I haven't found an IDE that has text editing capabilities close 
>> to Vim/Emacs. [...]
>> I have no idea how to configure Vim.  I'm a total newbie.  Can someone 
>> please help me get started.  I'm looking to configure Vim for Python. 
>> Autoindenting and 4 space tabs are high priority now.
> 
> looking in my vimrc file:
> 
> set autoindent
> set tabstop=4
> set shiftwidth=4
> set expandtab
> 
> could give you a start

No, no, no! That's evil!

Tabs are 8 characters wide. Period. So you want something like this instead:

" Tabulator policy:
" - Show existing tabs as 8 spaces
" - Tab key/autoindent/reindent insert 4 spaces
set ts=8 sw=4 smarttab
set expandtab

The full glory of my .vimrc is attached ;-)

-- Gerhard
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: .vimrc
URL: <http://mail.python.org/pipermail/python-list/attachments/20030415/ee04fe87/attachment.ksh>

From gh at ghaering.de  Wed Apr 16 10:03:30 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 16 Apr 2003 16:03:30 +0200
Subject: Python for VB6 programmers (was Re: Complexity of standard Python
 data structures)
In-Reply-To: <Ajdna.110595$It5.16632@news2.central.cox.net>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net>
Message-ID: <3E9D62B2.2010702@ghaering.de>

Steve Holden wrote:
> "Aahz" <aahz at pythoncraft.com> wrote:.
>>Right now, I'm trying to get some momentum moving on a book titled
>>_Python for VB6 Programmers_.  Anyone interested should subscribe to
>>marketing-python.
>>
>>http://wingide.com/mailman/listinfo/marketing-python
> 
> You'd do better to call it "Python for VB programmers", expecially since the
> change to VB6 is so radical that migrating to a new and more sensible
> language might be an attractive option to some!

You're confusing VB6 with VB.NET (VB7).

-- Gerhard





From no_replies at fake_email_address.invalid  Sun Apr 20 18:16:35 2003
From: no_replies at fake_email_address.invalid (Robert Oschler)
Date: Sun, 20 Apr 2003 22:16:35 GMT
Subject: Good Python generators example?
References: <paDoa.1611$hT2.1228371@news2.news.adelphia.net> <yu99y9243nn1.fsf@europa.research.att.com>
Message-ID: <7%Eoa.1624$hT2.1255148@news2.news.adelphia.net>

"Andrew Koenig" <ark at research.att.com> wrote in message
news:yu99y9243nn1.fsf at europa.research.att.com...
> Robert> Can someone point me to a nice concise Python 2.2 compatible
> Robert> example of the use of generators?
>
> Here's a program that solves the "8 queens" problem--that is, the
> problem of how to place 8 queens on a chessboard so that none of
> them attacks any of the others.  It prints all the solutions.
>
> from __future__ import generators
>
> class Queens(object):
>     def __init__(self, n):
>         assert n >= 0
>         self.N = n
>         self.board = []
>
>     def generate(self):
>         if len(self.board) == self.N:
>             yield self.board[:]
>         else:
>             for i in range(self.N):
>                 if self.safe(i):
>                     self.board.append(i)
>                     for i in self.generate():
>                         yield i
>                     self.board.pop()
>
>     def safe(self, n):
>         "Is is safe to append a queen to the board in row n?"
>         m = len(self.board)
>         for i in range(m):
>             b = self.board[i]
>             if n == b or b+(m-i) == n or b-(m-i) == n:
>                 return False
>         return True
>
> for i in Queens(8).generate():
>    print i
>
>
> --
> Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark


Andrew,

Thanks!  Interesting choice of examples, last time I solved that problem was
with Prolog.  With a stack of Python generators you can pretty well emulate
Prolog backtracking, except for the automatic unbinding of variables.

thx
--

Robert Oschler
Android Technologies, Inc.
http://www.androidtechnologies.com
The home of PowerSell! (tm)
The FrontPage AddIn for Amazon Associates
- "Power Tools for Amazon Associates" (sm)






From b_orcutt at pobox.com  Mon Apr  7 19:00:10 2003
From: b_orcutt at pobox.com (Bill Orcutt)
Date: 7 Apr 2003 16:00:10 -0700
Subject: Embedded MacPython in a Multi-Threaded App
Message-ID: <4ecaae91.0304071500.2f66da7a@posting.google.com>

Hi-

I'm writing an C application for OS 9 and I need/want to call some
python
code from inside of an MP Task Thread. Possible? I've had a look at
the official
Python thread documentation and googled and gleaned as much as I can
from
the web, but I still have yet find something that works.

In my main thread I intialize python and threads--

<<

    PyMac_Initialize();
    PyEval_InitThreads();
    
    // save a pointer to the main PyThreadState object
    mainThreadState = PyThreadState_Get();
    // release the lock
    PyEval_ReleaseLock();

>>

and then in my worker thread call--

<<

    PyEval_AcquireLock();
    // get a reference to the PyInterpreterState
    mainInterpreterState = mainThreadState->interp;
    // create a thread state object for this thread
    myThreadState = PyThreadState_New(mainInterpreterState);
    // free the lock
    PyEval_ReleaseLock();

    // grab the global interpreter lock
    PyEval_AcquireLock();
    // swap in my thread state
    PyThreadState_Swap(myThreadState);
    
    // execute some python code
        PyRun_SimpleString("print 'howdy-doo'"); //python here

    // clear the thread state
    PyThreadState_Swap(NULL);
    // release our hold on the global interpreter
    PyEval_ReleaseLock();

    // grab the lock
    PyEval_AcquireLock();
    // swap my thread state out of the interpreter
    PyThreadState_Swap(NULL);
    // clear out any cruft from thread state object
    PyThreadState_Clear(myThreadState);
    // delete my thread state object
    PyThreadState_Delete(myThreadState);
    // release the lock
    PyEval_ReleaseLock();

>>

after the worker thread exits I finalize and clean up in the main
thread.

I've tried every permutation on this I could come up with, including a
<<Py_BEGIN_ALLOW_THREADS>> block and bracketing the python call with
<<MPEnterCriticalRegion>>. Also have had a look at pysvr.c,
pythonrun.c & ceval.c.

Clearly I'm flying blind. Any assistance or tips would be greatly
appreciated.

thanks

Bill



From rmunn at pobox.com  Mon Apr 14 11:03:07 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 14 Apr 2003 15:03:07 GMT
Subject: Calling Python Code Snippets within Python Application
References: <236da4ed.0304140143.25abc577@posting.google.com>
Message-ID: <slrnb9kumi.1q1.rmunn@localhost.localdomain>

Marc Floessel <gildemere at hotmail.com> wrote:
> This is very likely a newbie question, but I'd still appreciate a
> pointer. I need to call Python Code snippets (saved in a file/database
> etc) within my main Python Application. I would like to do the
> following things:
> 
> - Code Snippet should not have access to or knowledge of Main
> Application variables and methods - ie. each code snippet has its own
> private scope
> - main application exposes a number of specific methods and variables
> ("environment") to the snippet that it can use
> - a way for the code snippet to return some variables (ie. a success
> bool) - possibly using (re-importing back from the snippet) the
> variables exposed by the main application
> 
> In which direction should I look?

Look at the eval() or execfile() built-ins:

    http://www.python.org/doc/current/lib/built-in-funcs.html

Note that eval() evaluates an expression -- if you want to use eval() to
run chunks of Python code, you might have to wrap those chunks in a
function and eval() the result of calling that function. execfile() will
work if the code you want is in a file; if it's in a database, you might
have to write it to a temporary file and then execfile() that temporary
file.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From exarkun at intarweb.us  Wed Apr 30 17:03:26 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 30 Apr 2003 17:03:26 -0400
Subject: network programming problem??
In-Reply-To: <14bd1399.0304301236.7fe19be5@posting.google.com>
References: <14bd1399.0304301236.7fe19be5@posting.google.com>
Message-ID: <20030430210326.GA29023@meson.dyndns.org>

On Wed, Apr 30, 2003 at 01:36:01PM -0700, kkennedy wrote:
> I have written a small script to retrieve the python.org web page 7
> times. I did this to try and provide data for a percieved problem in
> another Python program. I ran this script on Windows XP and SuSE Linux
> 8.2 both running on the same machine (Athlon XP 2000+, 512MB RAM,
> D-Link DFE-530TX+ NIC).  Here is the script:

  DNS is a likely culprit.  Try using an IP instead of a hostname.  Or, try
using strace and noting the functions that take an overly large amount of
time.

  Jp


-- 
http://catandgirl.com/view.cgi?44
-- 
 up 41 days, 17:04, 5 users, load average: 0.97, 1.00, 0.66




From daniel.dittmar at sap.com  Thu Apr  3 10:23:26 2003
From: daniel.dittmar at sap.com (Daniel Dittmar)
Date: Thu, 3 Apr 2003 17:23:26 +0200
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com>
Message-ID: <b6hjle$b4m$1@news1.wdf.sap-ag.de>

Drew Moore wrote:
> some say in how this takes place. I'm at a loss to explain why
> the "most trivial case of augmented assignment" was denied this power.

Performance comes to mind. Assignment is a frequent statement and having to
search for an __assign__ method will slow things down a lot.

And if an object has an __assign__ method, how do you perform a real rebind?
Do we add a 'becomes' operator (similar to the ==/is operators)?

Daniel






From matt at teaphoo.fsnet.co.uk  Fri Apr 25 18:42:24 2003
From: matt at teaphoo.fsnet.co.uk (Matt Russell)
Date: 25 Apr 2003 15:42:24 -0700
Subject: Pickling objects...
References: <mailman.1051291900.11523.python-list@python.org>
Message-ID: <459f9b68.0304251442.6885dc2d@posting.google.com>

Hi,

you should not override the class's namespace.
i.e do not define class docContainer twice.

In the ocde you posted , the script that sets state defines 'class
docContainer',
and in the code that reads it you've overriden the class to be a blank
class of the same name.
if you need that class in the script that reads state, simply import
it from the script that sets state.

i.e

from script1 import docContainer

packedObj = open('packed.file', 'r')
doc = cPickle.load(packedObj)

Cheers,
Matt

"Bo M. Maryniuck" <b.maryniuk at forbis.lt> wrote in message news:<mailman.1051291900.11523.python-list at python.org>...
> Hello.
> 
> Well, I would like to anybody help me to serialize an instance 
> from 4DOM HtmlLib. Somebody know where (and HOW) to push 
> __setstate__ / __getstate__ methods? 8-| I tried a lot of different
> code, but _always_ I got the same traceback (below).
> 
> Yes, I know, I need __setstate__ stuff, but here I dumb where to
> push these... Or it is impossible to serialize it?.. Simple classes serialises
> without the problem. But this.......
> 
> This pack's OK:
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
> #!/usr/bin/python
> from xml.dom.ext.reader import HtmlLib as HL
> import cPickle
> 
> class docContainer:
>     def __init__(self, doc):
>         self.doc = doc
> 
>     def __getstate__(self):
>         return self.__dict__.copy()
> 
>     def __setstate__(self, dict):
>         self.__dict__.update(dict)
> 
> reader = HL.Reader()
> doc = docContainer(reader.fromString(open('foo.html').read(), None, None))
> 
> packIt = open('packed', 'w')
> p = cPickle.Pickler(packIt, 1)
> p.dump(doc)
> packIt.close()
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
> 
> This should read, but raises an exception (below):
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
> #!/usr/bin/python
> import cPickle
> 
> class docContainer:
>     pass
> 
> packedObj = open('packed.file', 'r')
> doc = cPickle.load(packedObj)
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
> 
> ...and the traceback:
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
> Traceback (most recent call last):
>   File "./unpkl.py", line 10, in ?
>     doc = cPickle.load(packedObj)
> TypeError: ('__init__() takes exactly 1 argument (2 given)', <class xml.dom.html.HTMLDocument.HTMLDocument at 0x81dd47c>, (None,))
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
> 
> Can anybody push me where to find and WHAT to subclass, if I need it...?
> Thank you.



From pobrien at orbtech.com  Sun Apr  6 01:51:39 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 00:51:39 -0600
Subject: Could Emacs be rewritten in Python?
Message-ID: <m2u1dcw22s.fsf@localhost.localdomain>

If you were crazy enough to think that you could create a program
along the lines of Emacs, but written in Python, how would you go
about it?  How would you design the domain model for files, buffers,
windows, and frames?  How would you allow the same level of
customizability?  How would you map functions (or methods or whatever)
to keys?  Any thoughts?

Here is the background.  I'm working on a file editor (called
PyAlaMode) using wxPython and Scintilla.  It's an extension of the
PyCrust code base.  I've got a basic version working, and it supports
editing multiple files and all the usual, basic features.  But I want
it to be as customizable and extensible as Emacs.  In fact, I want to
model it after Emacs, so that elisp code could be rewritten in Python
and work with PyAlaMode.  To do that, I've got to support the same
primitives as Emacs; expose files, buffers, regions, and windows the
same way Emacs does; handle the Emacs keybindings and ability to remap
keys, the same way Emacs does, etc.

Sound like fun?

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From amk at amk.ca  Sun Apr 20 09:59:20 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Sun, 20 Apr 2003 08:59:20 -0500
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl>
Message-ID: <ruadnZIgGvclOj-jXTWcpA@speakeasy.net>

On Sun, 20 Apr 2003 09:23:12 +0200, 
	Anton Vredegoor <anton at vredegoor.doge.nl> wrote:
> days, unfortunately. I'd be interested in an explanation of what
> continuations are and what their advantages would be. 

See an old posting by Tim Peters, archived at  
http://mail.python.org/pipermail/python-dev/1999-July/000467.html

--amk



From skip at pobox.com  Wed Apr 30 10:00:18 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 30 Apr 2003 09:00:18 -0500
Subject: Can urllib be used with nonblocking sockets?
In-Reply-To: <877k9c2nq0.fsf@r2d2.rfc1149.org>
References: <877k9c2nq0.fsf@r2d2.rfc1149.org>
Message-ID: <16047.55026.945080.733649@montanaro.dyndns.org>

    Arne> I have written a little python program that uses non-blocking
    Arne> sockets with select and I now want to urrlib in my program and I
    Arne> cannot find a method in urllib that would allow me to tell urllib
    Arne> that it should use non blocking sockets.

In 2.3 you can set the default timeout for sockets before opening the URL:

    dto = socket.getdefaulttimeout()
    socket.setdefaulttimeout(mytimeout)
    try:
        try:
            f = urllib2.urlopen("http://www.python.org/")
            x = f.read()
        except socket.error, msg:
            print "timeout"
    finally:
        # restore default
        socket.setdefaulttimeout(dto)

Skip




From tug at wilson.co.uk  Wed Apr 30 13:14:57 2003
From: tug at wilson.co.uk (John Wilson)
Date: Wed, 30 Apr 2003 18:14:57 +0100
Subject: Problem subclassing tuple
References: <1051720502.3eaffb367a905@mcherm.com>
Message-ID: <40a701c30f3c$06b119b0$640a0b0a@handel>

----- Original Message ----- 
From: "Michael Chermside" <mcherm at mcherm.com>
[snip]
> For more about it, including some discussion of why two stage
> initialization is so good, see 
> http://cocoa.mamasam.com/MACOSXDEV/2003/03/1/58013.php (and I thought
> I'd heard a discussion of the subject in Python recently, but couldn't
> find a link).

Thanks - the discussion looks very interesting.

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From fidtz at clara.co.uk  Fri Apr 25 16:09:28 2003
From: fidtz at clara.co.uk (Fidtz)
Date: Fri, 25 Apr 2003 21:09:28 +0100
Subject: Reading a non-standard floating point format
References: <mailman.1051286382.24604.python-list@python.org> <vaip4ts74ajqd0@corp.supernews.com>
Message-ID: <Xns9327331C4783Bfidtzxyzabc@195.8.68.209>

claird at lairds.com (Cameron Laird) wrote in
news:vaip4ts74ajqd0 at corp.supernews.com: 

> In article <mailman.1051286382.24604.python-list at python.org>,
> Tim Peters  <tim.one at comcast.net> wrote:
>>[Fidtz]
>                .
>                .
>                .
>>> While I think I can see the way to do this with various bit-grubbing
>>> techniques, the potential for error is massive, esp considering my
>>> non comp-sci background.
>>
>>Post a link to a definition of this format, and I bet someone will
>>help. 
>>
>>> Does anyone know of a library in python or even (fairly) easily
>>> wrappable C that might let me specify the format in a more high
>>> level way?
>>
>>Precise English will translate into Python easily enough <wink>.
>>
>>
> 
> It's true that getting floating-point formats correct
> to the far-right bit is, in general, tedious; Fidtz
> shows good instincts in his caution about that prospect.
> 
> The bad news is that, no, there are no widely-applicable
> parametrizations of the range of interesting floating-
> point formats.  This remains an area of "craftwork".
> Learn to use a hand-chisel, or find someone who will do
> so for you.

Thanks for the info, I was worried that might be the situation!

The closest thing I have to a definition from the AlphaBasic programming
manual: The reason for this is that floating point numbers occupy six
bytes of storage. Of the 48 bits in use for each 6-byte variable, the
high order bit is the sign of the mantissa. The next 8 bits represent
the signed exponent in excess-128 notation, giving a range of
approximately 2.9*10^-39 through 1.7*10^38. The remaining 39 bits
contain the mantissa, which is normalized with an implied high-order bit
of one. This gives an effective 40-bit mantissa which results in an
accuracy of 11 significant digits. 

I had a go this afternoon and managed to read the sign and something
approximating the exponent, though I will have to test the edge
conditions carefully :) 

Dom  



From roy at panix.com  Tue Apr  8 13:26:15 2003
From: roy at panix.com (Roy Smith)
Date: 8 Apr 2003 13:26:15 -0400
Subject: exceptions == errors?
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com> <16018.5597.933725.584314@montanaro.dyndns.org> <mailman.1049817314.14810.python-list@python.org>
Message-ID: <b6v0nn$1h8$1@panix2.panix.com>

Jim Meyer  <jmeyer at pdi.com> wrote:
>It's always been my impression that exceptions impose considerable
>performance overhead and therefore should be used to handle only the
>truly exceptional cases; for any predictable case, one should perform
>validation and/or error-handling outside the exception mechanism. 
>
>[BTW, I picked up this habit after reading "Effective C++" by Scott
>Meyers, a book I cannot recommend highly enough. =]

"Effective C++" is indeed an excellent book for improving your C++
code.  But why would you think it would teach you anything about
improving your Python code?  Most of the book is very C++ specific.



From fedor at mailandnews.com  Wed Apr 30 17:11:15 2003
From: fedor at mailandnews.com (fedor)
Date: Wed, 30 Apr 2003 23:11:15 +0200
Subject: list index question
References: <e62b26d3.0304010613.39f98991@posting.google.com> <pan.2003.04.01.15.38.54.181811@mn.rr.com> <pan.2003.04.01.15.42.18.291186@mn.rr.com> <e62b26d3.0304011217.7f41254f@posting.google.com> <slrnb8lf0v.ccp.rmunn@localhost.localdomain>
Message-ID: <b8pe54$6kn$1@news.hccnet.nl>

> > Can I wildcard? I don't know the values in the second column.  I want
> > to find the first entry what has a particular value in the first
> > column.

> The list.index() function won't help you, then: it just does a strict
> equality comparison. You'll have to write a simple loop:
>
>     index = -1
>     for item in sql_result_list:
>         index += 1
>         if item[0] == 100:
>             print "Found it"
>             break
>     if index >= len(sql_result_list):
>         print "Not found"
>
> I know you said you wanted to avoid looping, but this may not be
> possible. Why did you want to avoid looping, anyway?
>
> Another thing you may not have thought of is to loop through your
> results once and build a dict that you could use for lookup:
>
>     index_dict = {}
>     for item in sql_result_list:
>         key = item[0]
>         index_dict[key] = item
>
> Now you can use index_dict to do O(1) lookups based on the first item of
> each tuple. Note that as written, though, only *one* item with the same
> key will be kept. There are other ways around that, but I'm getting a
> bit far off your original question.
>

If you realy want to avoid you can use the less eficient:

>>> list1=[(100,1),(100,10),(101,1),(101,5)]
>>> zip(*list1) #transpose
[(100, 100, 101, 101), (1, 10, 1, 5)]
>>> list(zip(*list1)[0]).index(100) #select the first column and transform
it into a list (because a tuple doesn't have an index function)
0






From h_schneider at marketmix.com  Thu Apr 17 10:56:00 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Thu, 17 Apr 2003 16:56:00 +0200
Subject: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net> <b7mbnb$pp2$06$1@news.t-online.com> <8yu9gpnb.fsf@python.net>
Message-ID: <b7mf9u$2jb$03$1@news.t-online.com>

It's in HKEY_LOCAL_MACHINE.

The registry calls are in another module, which is imported. The registry
calls work, when I move them to the Service framework. So this seems to be
an import problem which only happens on 2.2 ...

> What happens when you run the service as not-py2exe'd service (via
> PythonService.exe)?

It just says
Could not start the service - error 203
Registering the Python Service Manager..

Very strange ..

--Harald



"Thomas Heller" <theller at python.net> schrieb im Newsbeitrag
news:8yu9gpnb.fsf at python.net...
> "Harald Schneider" <h_schneider at marketmix.com> writes:
>
> > No permission probs, logged in as Admin. Platform is Win2K, not updated
from
> > NT.
> > There are no exceptions. Reading a registry key just claims it would not
> > exist. When I run the same code outside the service wrapper, I can read
the
> > key.
> >
>
> Strange.
>
> Do you mean *you* are logged in as Admin, or does the *service* run under
> the Admin account?
>
> Per default a service runs under the system account. Maybe the registry
> keys are really not there then? HKEY_CURRENT_USER, probably?
>
> What happens when you run the service as not-py2exe'd service (via
> PythonService.exe)?
>
> Thomas
>
> PS: As far as i can tell, there are no differences between the 2.2 and 2.1
> release of py2exe. Not yet ;-)





From FBatista at uniFON.com.ar  Fri Apr 11 16:18:33 2003
From: FBatista at uniFON.com.ar (Batista, Facundo)
Date: Fri, 11 Apr 2003 17:18:33 -0300
Subject: ASN.1
Message-ID: <A128D751272CD411BC9200508BC2194D018853E1@escpl.tcp.com.ar>

Hi people!

I'm searching a ASN.1 parser, to load the structure of a file to memory and
then be able to parse the file.

Do you know something like this for python?

Thanks!

	Facundo





. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
ADVERTENCIA  

La informaci?n contenida en este mensaje y cualquier archivo anexo al mismo,
son para uso exclusivo del destinatario y pueden contener informaci?n
confidencial o propietaria, cuya divulgaci?n es sancionada por la ley. 

Si Ud. No es uno de los destinatarios consignados o la persona responsable
de hacer llegar este mensaje a los destinatarios consignados, no est?
autorizado a divulgar, copiar, distribuir o retener informaci?n (o parte de
ella) contenida en este mensaje. Por favor notif?quenos respondiendo al
remitente, borre el mensaje original y borre las copias (impresas o grabadas
en cualquier medio magn?tico) que pueda haber realizado del mismo. 

Todas las opiniones contenidas en este mail son propias del autor del
mensaje y no necesariamente coinciden con las de Telef?nica Comunicaciones
Personales S.A. o alguna empresa asociada. 

Los mensajes electr?nicos pueden ser alterados, motivo por el cual
Telef?nica Comunicaciones Personales S.A. no aceptar? ninguna obligaci?n
cualquiera sea el resultante de este mensaje. 

Muchas Gracias.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030411/2d09c390/attachment.html>

From anton at vredegoor.doge.nl  Mon Apr 21 10:44:29 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 21 Apr 2003 16:44:29 +0200
Subject: __setitem__ puzzle
References: <b80q7c$cmd$1@news.hccnet.nl> <b80rmm$2c3$1@panix1.panix.com>
Message-ID: <b81083$l2l$1@news.hccnet.nl>

On 21 Apr 2003 09:28:54 -0400, aahz at pythoncraft.com (Aahz) wrote:

>While I understand your point about expirimenting, I can't give you any
>useful advice unless I know what your code is intended to do, and I doubt
>many people on c.l.py will do much better.

Reading the subject line might be informative. According to Derrida a
thing that is equal to itself is dead, so how dead will something be
that is exactly as intended?

Anton



From jarausch at skynet.be  Sun Apr 27 15:20:01 2003
From: jarausch at skynet.be (Helmut Jarausch)
Date: Sun, 27 Apr 2003 21:20:01 +0200
Subject: 2.3b1 and Tk8.4.2
Message-ID: <3eac2d62$0$594$ba620e4c@reader1.news.skynet.be>

Hi,

I've tried to build 2.3b1 but it fails to build _tkinter.so with the
error message


-- *** WARNING: renaming "_tkinter" since importing it failed: 
build/lib.linux-i686-2.3/_tkinter.so: undefined symbol: TkSetClassProcs


and indeed, libtk8.4  (from 8.4.2) does not contain this symbol
while libtk8.3 does.

Is there a fix other than keeping Tk8.3 just for Python?

Many thanks for a hint,

Helmut Jarausch

Lehrstuhl fuer Numerische Mathematik
RWTH - Aachen University
D 52056 Aachen, Germany




From aleax at aleax.it  Tue Apr  1 03:56:38 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 08:56:38 GMT
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <bd993a2f.0303310333.5a95ecb3@posting.google.com> <roy-D262FA.09230131032003@reader2.panix.com> <b6aam1$7o6$1@news.island.liu.se> <mailman.1049170087.26854.python-list@python.org>
Message-ID: <avcia.53237$i26.1632089@news2.tin.it>

Steven Taschuk wrote:
   ...
> What do templates have to do with interfaces?

C++ templates, and "interfaces" (entirely abstract classes), are related 
by offering two separate approaches to obtaining polymorphism, with very
different trade-offs. Python polymorphism is very close to the "generic
programming" afforded by C++ templates, since both hinge on signature
based polymorphism.  However, Python's polymorphism is resolved at runtime,
like that based on C++'s virtual-method calls, not at compile-time, like
that afforded by C++'s templates, so the parallels are not exact.


Alex




From dberlin at dberlin.org  Mon Apr 14 13:39:06 2003
From: dberlin at dberlin.org (Daniel Berlin)
Date: Mon, 14 Apr 2003 13:39:06 -0400
Subject: Speed of string += string
In-Reply-To: <5d7j9vch642sjfipee9ul333iafkfoimgb@4ax.com>
Message-ID: <FD63B586-6E9F-11D7-8B3E-000393575BCC@dberlin.org>

On Sunday, April 13, 2003, at 01:26  PM, Courageous wrote:

>
>> But is that /guaranteed/ (as it is with STL), or does it just happen 
>> to be
>> that way with the current implementation (however unlikely it is to
>> change)?
>
> It's amortized. It's a highly tuned hash table. Python doesn't have
> implementation _specs_ (requirements) in the nature of STL.

Right.
This is because nobody could determine what country Tim Peters *really* 
represents, so we couldn't form an ISO committee to spend 5 years doing 
it.
>
> C//
>
> -- 
> http://mail.python.org/mailman/listinfo/python-list





From ark at research.att.com  Mon Apr 28 16:41:41 2003
From: ark at research.att.com (Andrew Koenig)
Date: Mon, 28 Apr 2003 20:41:41 GMT
Subject: Slice confusion : a[n:p] is a list exclude the last  element p (thus spoke Dijkstra)
References: <mailman.1051532982.25352.python-list@python.org> <b8k0fl$dsa$1@news.hananet.net>
Message-ID: <yu99bryqjqwq.fsf@europa.research.att.com>

June> A bit related to this is,
June> "Why Numbering Should Start At Zero" by Edsger Dijkstra at
June> http://www.cs.utexas.edu/users/EWD/ewd08xx/EWD831.PDF

Thanks for the reference.  I hadn't seen this (3-page) paper before; I
think it's elegant.

Incidentally, Dijkstra's argument about ``natural numbers'' translates
directly into an argument about pointers in C, which permits a pointer
to any element of an array, or to a position immediately past the last
element of an array, but does not allow a pointer to a position before
the beginning of an array.

-- 
Andrew Koenig, ark at research.att.com, http://www.research.att.com/info/ark



From pjdurai at hotmail.com  Mon Apr 28 18:59:59 2003
From: pjdurai at hotmail.com (pj)
Date: 28 Apr 2003 22:59:59 GMT
Subject: GUI testing with python  (SendKeys)
References: <b8c6c9$p2m$0@198.102.102.163> <b8ecgs$47i$1@sirius.dur.ac.uk>
Message-ID: <b8kbpf$63a$0@198.102.102.163>

I took a look at a girder.
It seemed far excessive for my requirements.
Frankly I havnt understood what its all about yet.

Just the sendkeys offered by Windows Scripting object isnt enough either. I
need some mouse manipulation too. So I thought perls Win32::GuiTest was a
good starting point.

Is there another easier to read description of girder somewhere on the net ?

thanks
pj


"Christopher Saunter" <christopher.saunter at durham.ac.uk> wrote in message
news:b8ecgs$47i$1 at sirius.dur.ac.uk...
> Hi,
>   You might want to check out Girder - http://www.girder.nl
> It offers more robust and flexible ways of remote controlling Win32 GUI
> stuff than sendkeys.  There is a very noddy Python client for talking to
> Girder at http://www.dur.ac.uk/christopher.saunter/soft.html
>
> HTH
> cds
>
> pj (pjdurai at hotmail.com) wrote:
> : Greetings
>
> : Is there a python module similar to Perl's Win32::GuiTest?
> : Basically something that would implement SendKeys and friends on Win32 ?
>
> : Before I went ahead and rolled my own, I just wanted to check.
> : Web searches didnt turn up anything useful.
>
> : cheers
> : pj
>
>
>





From ben at hal.rescomp.berkeley.edu  Thu Apr 24 19:04:40 2003
From: ben at hal.rescomp.berkeley.edu (Benjamin Hartshorne)
Date: Thu, 24 Apr 2003 23:04:40 +0000 (UTC)
Subject: string to variable name
Message-ID: <b89qi8$2oij$1@agate.berkeley.edu>

<disclaimer -- perl programmer trying to learn python>

I'm trying to figure out how to get from a string to a varible name.  I just
picked up Python last week, so I'm still kinda new at this. 

Suppose I've got a text file:

"""
blah blerg blarf
foo bar baz
spam eggs ham
"""

I have a class:

class myobject:
  def __init__(self):
    self.blerg = "blarf"
    self.bar = "baz"
    self.eggs = "ham"

I have a global:

inst = myobject()

I have a function that I want to change the variables in my class:

def myassigner():
  file = open("mytextfile", 'r')
  for line in file.readline():
    m = re.match("^(\w+)\s+(\w+)\s+(\w+)$", line)
    (newval, var, oldval) = m.groups()
    inst.var = newval

The equivelent (of what I _want_ to do) in perl:

sub myassigner() {
  FILE = open("<mytextfile");
  while (<FILE>){
    /^(\w+)\s+(\w+)\s+(\w+)$/;
    inst.$2 = $1;
  }
}

But, as I'm sure you're all aware, the python does not do what the perl does.
Instead, it writes newval into a variable in inst called 'var'.  

Is there any way to (easily) achieve what I want to do in python?

Thanks,

-ben

p.s.  The inspiration behind asking this question: I want to store all the
values passed in by a CGI script in an instance of a class.

for key in mycgi.keys():
  mystorage.key = mycgi[key].value

Hey! Could I do it by overloading the special function that's called when you
treat something like a dictionary?  i.e. could I make my class respond to 
  key = "bar"
  inst[key] = "baz"
as though it got
  inst.bar = "baz"

How?

-- 


-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Ben Hartshorne



From lx1978 at 263.net  Sun Apr  6 16:16:07 2003
From: lx1978 at 263.net (Xin Liu)
Date: Mon, 7 Apr 2003 04:16:07 +0800
Subject: Is there swap function in Python?
References: <b6q0n6$1c55$1@mail.cn99.com> <3r119vosrdukh6til7bdmgredlhhehqg64@4ax.com>
Message-ID: <b6q1so$1cm2$1@mail.cn99.com>

>
> What do you mean by the swap function? What should it do
> precisely?
>
> -- 
> Ren? Pijlman
It is a function like swap(a, b) . Use to swap the value of a and b.
A example in C++ language maybe like:
template <class T> void swap(T & a, T & b)
{
    T * tmp;
    tmp = a;
    a = b;
    b = tmp;
    return
}





From peter at engcorp.com  Wed Apr 23 22:41:04 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 23 Apr 2003 22:41:04 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com>
Message-ID: <3EA74EC0.4BAB336@engcorp.com>

Aahz wrote:
> 
> Which brings up the question: do the unit-testers here write tests for
> throwaway scripts?

For throwaways, I don't write quite so many tests, and in many cases if 
it's only a "small" throwaway script (e.g. one or two subroutines only, 
plus a little mainline) I will write a test or two for the subroutines 
(because I'm doing TDD, not because I want to have unit tests to make 
sure my script works properly... if you see the distinction) but never
for the mainline.  

Basically, the purpose of the tests is to catch problems you create 
when refactoring, so if you're planning to throw the script away, 
you need the tests only to help you write the code in the first
place.  No, that's two reasons for writing the tests - for refactoring,
and to drive the development.  And to document the requirements and
the API for the code.  There are three reasons... oh, bugger ....

Basically, with a literally throwaway script, you do lose much of
the business value of the tests, so as far as I can see you write 
them only if they truly help you write the code properly, which 
should once you're proficient at TDD and then only sometimes, since
you're a very experienced programmer.  Otherwise, and certainly in some 
cases no matter who you are, they really would be a waste of time.  
IMHO.  Except maybe to help you develop the habits.   _Four_ reasons...

-Peter



From imbosol-1049420093 at aerojockey.com  Thu Apr  3 22:19:11 2003
From: imbosol-1049420093 at aerojockey.com (Carl Banks)
Date: Fri, 04 Apr 2003 03:19:11 GMT
Subject: Is there a python realaudio player? For windows?
References: <b6i6uq$rcs$0@216.39.172.122> <yH4ja.7115$ey1.562521@newsread1.prod.itd.earthlink.net>
Message-ID: <PQ6ja.7085$wM6.2874@nwrdny01.gnilink.net>

Tipton Bandy wrote:
> A *.ram file is a "real audio metafile". It does not contain any
> sound at all.  If you save the file and open it up in a text
> editor, you will find a line that looks like this:
> rtsp://server/path/file.rm
> This is a streaming media file.  I was hoping to be able to copy
> the *.rm file to my hard disk so that I could listen offline, but
> so far I haven't been able to do that.  Google hasn't helped.
> Tipton


A couple ideas:

On Linux, Real Player (at least the version I have) has Enlightened
Sound Daemon support.  If you set Real Player to use ESD, you can use
esdmon to output the sound to a file.  I was able to use this trick to
get a sound clip from "The Price Is Right" (specifically, the music
from the game Cliffhangers) to use in a genetic algorithms
presentation.  However, this didn't work smoothly.

A lot of sound card drivers have an option to record anything played
to a file.

Or you can write your own client for RTSP protocol to download the
stream.  Here is the home page for RTSP:

    http://www.cs.columbia.edu/~hgs/rtsp/

I doubt the format of Real Media sound itself is open, but it might be
possible to more or less throw a bunch of datagrams together and it'll
work as a .rm file.


-- 
CARL BANKS



From ma at Bell.com  Wed Apr 30 00:56:22 2003
From: ma at Bell.com (xam)
Date: Wed, 30 Apr 2003 04:56:22 GMT
Subject: plugin system
Message-ID: <WHIra.27759$su3.3307778@twister.nyc.rr.com>

I am interested in writing a plugin system where the main app is python but
it is capable of accepting helper "plugins" which can be either python or
C++. I guess I want some sort of an interface spec for the plugins that can
be implemented in either language. Which technologies should I research? Are
there other projects that I can refer to as a model -

thanks.





From belred1 at yahoo.com  Thu Apr 24 22:01:51 2003
From: belred1 at yahoo.com (Bryan)
Date: Fri, 25 Apr 2003 02:01:51 GMT
Subject: xmlrpc introspection
References: <bbefdd1c.0304220823.1a4a8b3c@posting.google.com> <af6b5bdd.0304240434.fd7e6ad@posting.google.com>
Message-ID: <jG0qa.76986$gK.180897@rwcrnsc52.ops.asp.att.net>

poiboy,

thanks!!!!  the xml-rpc debugger is awsome.  and the results are exactly
what i was after.  of course the introspection doc was too, but someone
should update the introspection docs with hard examples like these... puts
all the questions to rest.

thanks again.

bryan



"poiboy" <bitbucket at safe-mail.net> wrote in message
news:af6b5bdd.0304240434.fd7e6ad at posting.google.com...
> Quick answer: Register 'system.listMethods', 'system.methodSignature',
> and 'system.methodHelp' in your server and make sure they perform as
> stated in the docs (listMethods returns a list of strings, etc.). Yup
> - just do it all by hand.
>
> The output for these methods are typical <methodResponse>s (snippets
> available at http://www.xmlrpc.org/spec) - which *should* be taken
> care of automatically by the XMLRPC server; lists and/or strings
> converted to something like
>
> <?xml version="1.0"?>
> <methodResponse>
>     <params>
>         <param>
>             <value>
>                 <array>
>                     <data>
>
> <value><string>remoteMethodOne</string></value>
>
> <value><string>remoteMethodTwo</string></value>
>
> <value><string>remoteMethodThree</string></value>
>                     </data>
>                 </array>
>             </value>
>         </param>
>     </params>
> </methodResponse>
>
> ..for 'server.listMethods'. In other words, you *shouldn't* have to
> worry about the raw XML. Verify your server's output democratically
> via the debugger link above
> (http://www.dscpl.com.au/xmlrpc-debugger.php - added to my bookmarks,
> mahalo Graham!) or manually by testing 'server.listMethods' at
> http://xmlrpc.usefulinc.com/demo/server.php.
>
> A headstart on introspection can be found at
> http://xmlrpc-c.sourceforge.net/hacks/xmlrpc_registry.py, and there's
> a detailed account of a comprehensive 'system.describeMethods'
> implementation at
http://xmlrpc-epi.sourceforge.net/specs/rfc.system.describeMethods.php
> which I recommend for practical documentation consideration.
>
> Two unsolicited cents worth: Complimentary RPC test cases are
> invaluable (list all methods as links, click to test individually or
> click "test all"). Had to get that off my chest.
>
> Aloha, the poiboy





From tim_one at email.msn.com  Tue Apr 15 00:02:05 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Tue, 15 Apr 2003 00:02:05 -0400
Subject: Built-ins vs. modules
In-Reply-To: <BAC0CDB6.35077%berkowit@silcom.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCOEAKEHAB.tim_one@email.msn.com>

[Paul Berkowitz]
> I see from the Python Library Reference (for 2.2) that a lot of
> things which I thought required a module to be imported can also be
> accessed without doing so:
>
>     >>> "boo".capitalize()   #built-in
>     'Boo'
>     >>> import string
>     >>> string.capitalize("boo")
>     'Boo'
>
> Why import the string module if you can do the same things anyway?

Nostalgia; inertia; laziness; perverseness; whimsy; Luddism.

> The page of built in String Methods bears an uncanny resemblance to the
> page on the string module. Is there a historical reason?

Yes:  string methods didn't exist before Python 1.6.  The string module has
always been there.

> Are the built in type methods quite recent?

Python 1.6 was released in 2000.  Whether that's quite recent depends on
whether you reckon in, say, geological time or computer time <wink>.

If you like investigating this kind of thing, see Misc/HISTORY and Misc/NEWS
in your Python distribution (the latter gets prepended to the former as the
centuries fly by).





From skip at pobox.com  Tue Apr 29 17:33:23 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 29 Apr 2003 16:33:23 -0500
Subject: Finding unused code
In-Reply-To: <pan.2003.04.29.21.07.45.44830@effectif.com>
References: <pan.2003.04.29.21.07.45.44830@effectif.com>
Message-ID: <16046.61347.502523.859663@montanaro.dyndns.org>

    Graham> I'd like to write a script to analyse the source code and tell
    Graham> me which bits haven't been used, but I'm at a bit of a loss for
    Graham> where to start.

Check out pychecker.  That may get you 90% of the way there.  If not, it
probably provides a good framework to add the necessary logic.

    http://pychecker.sf.net/

Skip






From dale at riverhall.NOTHANKS.co.uk  Wed Apr 16 10:04:47 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 16 Apr 2003 15:04:47 +0100
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com> <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com>
Message-ID: <cnoq9vci702imrunkenpulnefautq0n5ua@4ax.com>

Dale Strickland-Clark <dale at riverhall.NOTHANKS.co.uk> wrote:

>danra at walla.co.il (Danra) wrote:
>
>>def f(fileobj):
>>    if not hasattr(fileobj,'write'):
>>        raise TypeError, "Given arg must have the 'write' attribute."
>>    .
>>    .
>>    fileobj.write('...')
>
>You want assert, you do:
>
>def f(fileobj):
>    assert hasattr(fileobj, 'write'), 'Bang! Sort your code out.'

Obviously, there's a 'not' missing from the above :-(.
-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From howard at eegsoftware.com  Sat Apr 26 11:59:00 2003
From: howard at eegsoftware.com (Howard Lightstone)
Date: Sat, 26 Apr 2003 15:59:00 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <14bd1399.0304252032.71752cb5@posting.google.com> <3EAA13D3.F7B4B9A6@easystreet.com>
Message-ID: <Xns93695AE339FBBhowardeegsoftwarecom@66.75.162.201>

achrist at easystreet.com wrote in news:3EAA13D3.F7B4B9A6 at easystreet.com:

> kkennedy wrote:
>> 
>> RTF format (MS open format???) is a moving target.  RTF files created
>> by MS Word 97 will look odd or worse in MS Word 95, and be completely
>> unreadable in WordPad. Either find out what version of RTF WordPad
>> uses and use that, or require your users to have a specific version
>> of MS Word (or the free download - MS Word viewer).
>> 
> 
> That's one reason why I asked if there were any solutions already
> available.  If someone had already identified the forms of rtf 
> markup that worked well in all those versions, I would not want 
> to rediscover the same set of tires.  
[snip]

About 8 years ago I did sort of the opposite.  We wanted a way to keep 
documentation, specifications, and code aligned so I dreamed up a method of 
using Word as a (semi-)literate editor with different (defined by me) 
styles for each kind of thing.  Of course, then I had to write code to 
extract out the various headers, code segments, build instructions , etc. 
which had to feed to the many existing compilers, linkers, and such.  I 
made them keep the code in RTF so I could *just* write an ASCII parser for 
it.    

Right.

Every version of Word and many different combinations of operating systems 
kept me constantly updating the parser.  The RTF "rules" (all three 
versions I downloaded from Microsoft) do NOT contain all the different 
possibilities that Word expects/generates nor do they explain just how you 
need to order various pieces.  We resolved most of our issues by smacking 
the knuckles of anyone who tried a "new" way of formatting and by making 
them responsible for proper code extraction (although they did a lot of 
whining at me).

If you have simple boilerplate requirements, try making a simple file with 
one example of each formatting kind and then dumping the RTF.   You will 
then see.....   (I used a package for parsing RTF which I HIGHLY recommend
   http://www.dubois.ws/software/RTF 
that happens to have the RTF spec with it).



From rs55862 at NOSPAMyahoo.com  Wed Apr  9 07:17:58 2003
From: rs55862 at NOSPAMyahoo.com (Rami Saarinen)
Date: Wed, 09 Apr 2003 14:17:58 +0300
Subject: Making python module using Borland C++ 5
In-Reply-To: <mailman.1049880855.22786.python-list@python.org>
References: <b6u7fp$d2c$1@news.cc.tut.fi> <b6ujd4$pe5$1@news.hccnet.nl> <b70jiq$men$1@news.cc.tut.fi> <mailman.1049880855.22786.python-list@python.org>
Message-ID: <b70vh5$13a$1@news.cc.tut.fi>

>>> Please have a look at:
>>>
>>> http://www.cyberus.ca/~g_will/pyExtenDL.shtml
>>>
>> I finally managed to build the module using bcpp. Thank you Anton and 
>> Anders. It was all about including the EXPORT initpyclips=_initpyclips 
>> in a def file (see the link above).
> 
> 
> Why don't you just use disutils? It will work without any further setup, 
> then. All you need to read is the part of the Python documentation about 
> using non-MS compilers on Windows.

(I am involved in a research project and currently just trying to find 
out what can and cannot be done.)

The main reson is that I am going to use the module (if possible) within 
a custom python system (see www.reachin.se for more, if interested) that 
probably does not include distutils. I will have a look later on, but 
currently I was able to compile and link the module using Borland C++ 
builder 5 so that it does work on normal Python installation. However 
there seems to be some other things to consider when using the custom 
system.

I may later try to find out if there is distutils on the custom system, 
altough I am pretty sure that it does not include python in any other 
way than embedded in the rest of the system. Currently the problem lies 
elsewhere.

I have nothing against Distutils, I'd be glad to use it if possible. :)

--
Rami Saarinen




From rob at zoism.org  Wed Apr  2 23:09:53 2003
From: rob at zoism.org (Rob Brown-Bayliss)
Date: 03 Apr 2003 16:09:53 +1200
Subject: plugin system?
In-Reply-To: <20030403035224.GA8901@unpythonic.net>
References: <mailman.1048836529.7090.python-list@python.org>
	 <KVVga.31670$i26.818965@news2.tin.it> <1049072379.21707.34.camel@musicbox>
	 <20030331014306.GB21551@unpythonic.net> <1049340834.8760.22.camel@musicbox>
	 <20030403035224.GA8901@unpythonic.net>
Message-ID: <1049343012.8760.34.camel@musicbox>

> In my view, the thing returned should act as a factory (including the
> possibility of just being a class), so if you call it more than once you
> should get two independent instances of the plugin.

Yes, my example each module was imported once and once only, which
suited the app it was built for... basically only loading parts that the
users had used rather than the whole app...


-- 

*  Rob Brown-Bayliss
*  =================	
*      zoism.org






From cben at techunix.technion.ac.il  Fri Apr 11 09:43:00 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Fri, 11 Apr 2003 16:43:00 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1050048561.24768.python-list@python.org>
References: <mailman.1050048561.24768.python-list@python.org>
Message-ID: <Pine.GSO.4.51.0304111549170.15707@techunix.technion.ac.il>

Christian Tanzer wrote on 2003-04-11:

> Carl Banks <imbosol-1049990205 at aerojockey.com> wrote:
>
> > > Maybe I'm missing something, but non-dynamically scoped globals just
> > > seem like a bad idea to me.
> >
Agreed.  If you have any globals at all that someone might ever want
to check, making them dynamically scoped is the right thing.  Simple
single-valued non-scoped globals are the true evil.

> > What you're missing is that Globals are Evil(tm), and furthermore,
> > that if your program depends on changing a global in a dynamic scope,
> > it badly needs to be redesigned. (To wit, that functions that refer
> > to the global you want to change need to be changed to accept it as an
> > argument.)
> >
> > I think dynamic scpoing is a wholly bad idea, and simplifying a very
> > misguided programming practice doesn't make it any less so.
>
That's a good idea, in theory :-).  In practice, the problem is that
you don't always know before-hand what will need parametrizing.  Can
you list down in your interface spec all the parameters your code
might depend on?  What if deep down it contains some sanity check that
can print an error message to stdout?  Or writes a log file to the
current directory?  Or launches another program to your job that
depends on some environment variable?  It's hard to eliminate all
these dependencies and passing them around is too insane; so you can
either make them hardcoded, modifiable global (almost equally evil) or
dynamic (managable)...  If you forgot to add these minor parameters to
your function's interface, you can't fix it in a compatible way,
except by dynamic scoping.  It allows you to "tunnel"  a parameter
through code that is unaware of it.  It lifts the burden of specially
handling it all the way up.  It sounds like ... exceptions!
Exceptions are dynamically scoped, and this solves *so many* issues.
It's the single most important idea of exceptions vs. traditional
error return technics.

So I argue that dynamic scoping done right, unlike simple globals, is
not always evil.  It's completely equivallent to passing all your
environment, except that you don't have to write that explicitly,
which can make all the difference.  Frequently, what you really want
is to parametrize a *proccess*, not just a specific call.

For example, in Unix, I/O redirection, the current working directory
and most other process parameters are dynamically scoped: it's
implemnented by copying from each process to its children but to you,
it's just magically inherited.  Actually the copying semantics are the
only right way to do dynamic scoping in the face of paralell
processes/threads - it's inherited but later it can be changed
independently.

The parametrize-a-whole-process idea is the most convenient way
invented to date to handle issues like I/O.  Unfortunately, most
languages don't have with-output-to and the like, so inside a each
proccess it works as a single-valued global, with all the
inconvenience.  The fix is obvious - the language should provide
dynamic scoping emulation to complete the illusion <wink>.

Any parameter that changes with infrequenlty, whose use is scattered
throughout the system and that is not interesting to other subsystems
is a good candidate for dynamic scoping.  Consider environmet
variables in Unix: they are probably the biggest evidence for dynamic
scoping working on large scales.  In fact, they work on much larger
scales than any other mechanism I can think of - they allow disperate
programs written independently to communicate, "tunnelling" through
other applications, and with quite good results!  Lexical scoping only
works inside a single program, as far as I can see...

> You are of course right. Surprisingly though, it works very well for
> Emacs. Passing arguments would *not* work in the context of Emacs.
>
That's what remains to be seen :-).  Probably not all aspects of Emacs
are best handled by dynamic scoping but some probably are.


I think that the most Python way to abstract dynamic scoping would be
to disconnect it from actual scoping, leaving only the dynamic binding
part.  "Special variables", like in CL and Scheme's fluid-let won't
fit well because in Python a value can be stored in so many ways - as
a module/class/instance attribute, as a local variable, as part of a
data structure...  So I think we need something analogous Java's
ThreadLocal objects - a single "magic" object that dispatches you to
the binding you should see.

Then (ignoring compatibility issues), the sys module would contain
three Dynamic attributes - std{in,out,err} and one would be able to
write:

with sys.stdout.bound_to(file('quux')):
  do_something_involving_printing_somewhere()

You get the idea :-).

> Practicality-beats-purity-ly yr's,

the-Right-Thing-always-has-negative-overhead-
(if-only-you-find-this-thing)-ly y'rs
  Beni Cherniavsky <cben at tx.technion.ac.il>




From gh at ghaering.de  Mon Apr 21 19:42:21 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 22 Apr 2003 01:42:21 +0200
Subject: *Newbie Question* --> Is it viable to distribute Python apps
 over the Internet?
In-Reply-To: <25a672ee.0304211504.4f3cd067@posting.google.com>
References: <25a672ee.0304210914.1cc2049c@posting.google.com>
	<mailman.1050946774.25112.python-list@python.org>
	<3ea434fd$0$180$a1866201@newsreader.visi.com>
	<25a672ee.0304211504.4f3cd067@posting.google.com>
Message-ID: <3EA481DD.2060702@ghaering.de>

Denis Agar wrote:
> I guess I understand the situation then. Is there any way to use a GUI
> that most people already have such as the ones employed by Visual
> Basic or Visual C++?

VB's GUI code is likely in the VB runtime. With "VC GUI" you likely mean 
the MFC classes, which are usually in the MFC42.DLL. It can be used from 
Python via the win32 extensions, which are a relatively slim download.

But looking at the included examples, I wouldn't want to develop a GUI 
using the win32 extensions, as it's basically hardcore win32 API 
programming.

-- Gerhard





From claird at lairds.com  Fri Apr 25 13:27:29 2003
From: claird at lairds.com (Cameron Laird)
Date: Fri, 25 Apr 2003 17:27:29 -0000
Subject: Divided by a common language (was: OO + Python Developers Wanted)
References: <mailman.1051183297.29110.python-list@python.org> <yyyznmezekm.fsf@adams.patriot.net>
Message-ID: <vais01o44vtc0d@corp.supernews.com>

In article <yyyznmezekm.fsf at adams.patriot.net>,
Nick Vargish  <nav at adams.patriot.net> wrote:
>"Neeta Ramchandani" <nramchandani at harveynash.com> writes:
>
>> This is a role for an Investment Bank in the City, for duration of
>> 3-6 months. Rates are OPEN!!
>
>Which "City" would that be?
			.
			.
			.
<URL: http://www.cityoflondon.gov.uk/ >
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From dripton at ripton.net  Wed Apr 30 10:42:03 2003
From: dripton at ripton.net (David Ripton)
Date: Wed, 30 Apr 2003 10:42:03 -0400
Subject: bug in Jython 2.1? VerifyError
In-Reply-To: <3eafb3b0$0$49112$e4fe514c@news.xs4all.nl>
References: <3eafb3b0$0$49112$e4fe514c@news.xs4all.nl>
Message-ID: <3EAFE0BB.1090503@ripton.net>

Irmen de Jong wrote:
> I've run into some trouble using Jython 2.1
> (jdk 1.4, 1.3, windows, linux).
> 
> The following piece of Python code:
> ----
> def method():
>     try:
>         for dummy in [1,2,3]:
>             try:
>                 return "result"
>             except:
>                 pass
>     finally:
>         pass
> 
> print method()
> ----
> 
> fails to compile in Jython, I get this error:
> 
> java.lang.VerifyError: java.lang.VerifyError: (class: test$py, method: 
> method$1 signature: 
> (Lorg/python/core/PyFrame;)Lorg/python/core/PyObject;) Register 3 
> contains wrong type
> 
> 
> That piece of code above seems to be the minimal piece of code to trigger
> the problem. Something to do with nested try blocks and a loop?
> 
> Can someone confirm this or help me? Thanks!

I can reproduce the problem, with Jython 2.1 on JDK 1.4.2 beta on Linux.

2.1 is pretty old -- you might want to try building Jython from CVS and 
see if this has been fixed before reporting it.

If you don't get a better answer here, maybe try the jython-users or 
jython-dev list.

-- 
David Ripton    dripton at ripton.net





From andrew.markebo at telia.com  Sat Apr 26 14:22:16 2003
From: andrew.markebo at telia.com (Andrew Markebo)
Date: Sat, 26 Apr 2003 18:22:16 GMT
Subject: zipfile with umlaut in filename
References: <3ea9a4bf$1_1@news.vo.lu>
Message-ID: <m3ist1ruee.fsf@telia.com>

/ Patrick Useldinger <uselpa at myrealbox.com> wrote:
| Hello everybody,
| I have a problem when creating an archive using zipfile. If the name
| of the file to be added contains a special character like an umlaut,
| this character gets replaced by another character.
| What can I do against this?

If my memory doesn't fail me, basically the zip-standard only supports
ASCII, meanwhile winzip walks around this, because thats what the user
want, so they did an own addition to the standard.

A quick note maybe is, which format is the string sent to the
zip-functions? I mean, does the zipfile classes handle wide
(unicode/utf-8) strings and so on, these can't be used in the
zip-file.. yet..?

    /Andy

-- 
 Your fault -- core dumped



From aahz at pythoncraft.com  Wed Apr  2 13:12:58 2003
From: aahz at pythoncraft.com (Aahz)
Date: 2 Apr 2003 13:12:58 -0500
Subject: Pycon DC was terrific
References: <5b4785ee.0303310948.26afe16b@posting.google.com> <slrnb8lgns.ccp.rmunn@localhost.localdomain> <7h3d6k4ls53.fsf@pc150.maths.bris.ac.uk>
Message-ID: <b6f97a$1f5$1@panix1.panix.com>

In article <7h3d6k4ls53.fsf at pc150.maths.bris.ac.uk>,
Michael Hudson  <mwh at python.net> wrote:
>Robin Munn <rmunn at pobox.com> writes:
>> Wade Leftwich <wade at lightlink.com> wrote:
>>>
>>> Many thanks to the PSF, the sponsors, the organizers, Guido, and
>>> conference chairman / master of ceremonies Steve Holden.
>> 
>> Hear, hear! *applauds*
>
>(with-competive-streak :transatlantic
>  EuroPython will be better, dammit!
>)

Not if you insist on using Lisp.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From hust_zxq524 at 263.net  Thu Apr  3 22:11:56 2003
From: hust_zxq524 at 263.net (Frank Zheng)
Date: Fri, 4 Apr 2003 11:11:56 +0800
Subject: mp3 or wave  ---> midi , in python , how?
Message-ID: <b6isqq$23hs$1@mail.cn99.com>

Hi, All:

   Who can tell me how to do this in python ,are there any modules existing
can help me?


Thanks in advance

Frank








From shalehperry at attbi.com  Tue Apr  8 00:35:59 2003
From: shalehperry at attbi.com (Sean 'Shaleh' Perry)
Date: Mon, 7 Apr 2003 21:35:59 -0700
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <b6qhl6$sht$1@panix2.panix.com>
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049656989.5896.python-list@python.org> <b6qhl6$sht$1@panix2.panix.com>
Message-ID: <200304072135.59864.shalehperry@attbi.com>

> Using emacs with its default keybindings hurts my fingers.  If I were to
> seriously use emacs, I'd need a hefty .emacs, which would make me unable
> to be productive with it in an emergency.  As for using emacs in vi
> mode, why not just stick with vi?

took me about 3 months to fully break my vi fingers.




From patrick at fraley.de  Thu Apr 17 15:26:36 2003
From: patrick at fraley.de (Patrick W. Fraley)
Date: Thu, 17 Apr 2003 21:26:36 +0200
Subject: Zope Sequences and Python Scripting
References: <pan.2003.04.17.12.34.15.777976@fraley.de> <GN6dnaRxi5queAOjXTWcpw@august.net>
Message-ID: <pan.2003.04.17.19.26.22.654130@fraley.de>

On Thu, 17 Apr 2003 10:34:32 +0000, Terry Hancock wrote:

> Patrick W. Fraley wrote:
>> I have a python script which returns a [] containing {}.  Now I can go
>> thru the returned array (I know this is perl lingo, sorry that is my
>> background) by using the <dtml-in> method, but I can not
>> access the values of the dictionary contained within the sequence item.
>> 
>> And here is what I am trying to do with the returned array:
>> 
>> <dtml-in filter_row>
>> <dtml-var filter_id><br>
>> <dtml-var bez_ref><br>
>> <dtml-var col_span><br>
>> <hr>
>> </dtml-in>
> 
> <dtml-in filter_row prefix="sq">
>    <dtml-var expr="sq_item['filter_id']"><br>
>    <dtml-var expr="sq_item['bez_ref']"><br>
>    <dtml-var expr="sq_item['col_span']"><br>
>    <hr>
> </dtml-in>
> 
> or
> 
> <dtml-in filter_row>
>   <dtml-with seq-item>
>        <dtml-var filter_id><br>
>        <dtml-var bez_ref><br>
>        <dtml-var col_span><br>
>        <hr>
>   </dtml-with>
> </dtml-in>
> 
> (sorry, they're untested -- you may have to fix)
> 
> The point is, you will loop through the *sequence* of *dictionaries*. Then 
> you act on each dictionary -- referencing its elements.  You can use python 
> syntax to retrieve dictionary elements from the seq-item (prefix is a DTML 
> kludge to make the syntax more python compatible), or "dtml-with" to make 
> the dictionary the current namespace.  There are probably other ways to do 
> it as well.
> 
> Cheers,
> Terry

Well unfortunately the first one did the trick.  Unfortunately, because
this causes some other problems (this was just a short example).  I need
those variables in the current namespace.   But now I do have a basic idea
of what needs to be accomplished and I think I can take it from here.

Also after reading up a little more on Zope doc's I get the feeling that
I should be doing this with page templates instead. Any comments,
suggestions?

Thank you for your response,
Patrick





From ckea25d02 at sneakemail.com  Wed Apr  2 18:03:50 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 03 Apr 2003 11:03:50 +1200
Subject: When Python outruns C++ ...
In-Reply-To: <b6fmcq$dh0$1@panix2.panix.com>
References: <mailman.1048973984.17441.python-list@python.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com> <wBIia.2000$y3.310439@news010.worldonline.dk> <b6fmcq$dh0$1@panix2.panix.com>
Message-ID: <b6fpou$4t3ka$1@ID-169208.news.dfncis.de>

Roy Smith wrote:
> In that
> way lies bit-twiddling and assembler language and soldering irons.

I'm looking forward to the day when a board full of
user-programmable FPGAs becomes standard equipment.
That should make for some interesting optimisation
problems... :-)

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From phil at river-bank.demon.co.uk  Wed Apr 16 03:48:05 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Wed, 16 Apr 2003 08:48:05 +0100
Subject: PyQT and Eric
In-Reply-To: <b7is7g$i1d$1@rex.ip-plus.net>
References: <b7is7g$i1d$1@rex.ip-plus.net>
Message-ID: <200304160848.06522.phil@riverbankcomputing.co.uk>

On Wednesday 16 April 2003 7:13 am, Daniele Trovato wrote:
> Hi
>
> I tried to install Eric, but it won't work and I don't no why.
> I'm using SuSE 8.2 and I installed Python, PyQT and PyQT Develop,
> QScintilla, SIP ..... packages from the SuSE distribution.
> My problem is, when I'm trying to install Eric with the following command
> "python install.py", the shell returns me an error: "Sorry, Please install
> PyQT"  but I have PyQT already installed.
> Does anybody now where the problem is or how i can solve the problem ?

Can you run any of the PyQt examples?

What's the output of running the following (there shouldn't be any)?

python -c "from qt import qVersion, PYQT_VERSION"

The above is what the eric installer is running to see if PyQt is installed.

What version of PyQt?

Phil




From mis6 at pitt.edu  Mon Apr 21 17:22:47 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 21 Apr 2003 14:22:47 -0700
Subject: Is this a super bug?
References: <mailman.1050944075.8903.python-list@python.org>
Message-ID: <2259b0e2.0304211322.749bb6ce@posting.google.com>

> "Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message 
> news:<mailman.1050811779.16487.python-list at python.org>...
> 
> > <snip> this is the sequence that "should" happen:
> > 
> > 1  super(ff, f)[0]  # f == self
> > 2  superinst(type:<class ff>, obj:f)[0]
> > 3  si.__getitem__(0)
> > 4    si.__getattr__('  getitem  ')
> > 5    .. return str.__dict__['  getitem  '].__get__(f)
> > 6    boundmethod(str.__getitem__, f)
> > 7  bmeth(0)
> > 8   'a'
> 
> I don't understand what you are trying to say here ... 

> Special method lookup goes through the class, and since the class is
> Super, not obj.__class__, they can't be trapped through __getattr__.

Okay.


> (And for some reason they can't be trapped in a metaclass either...?)

?? Still unclear...

> Since Python has special attribute lookup rules for some attributes, the
> normal extension would be to define super as an object with data
> attributes from foo, and method attributes from Foobase. But that is not
> even correct, it has to have the method attributes from Foo, except when
> calling foo.m will call Foobase::m, but if Foobase::m calls m2, it will
> start lookup in Foo.
>
> One property this has is that given an empty subclass:

>  class Foo : Foobase: pass

> and a program context C with a hole, C[*] (i.e. C gives the value of the
> program once you put a value in the hole), then:

>   C[self] == C[super]

> proving properties in this half-baked scheme is much harder.

Not clear either ...

<snip>

> recall that 
>  obj.meth.__get__(None, obj) is an unbound method

You mean   obj.meth.__get__(None,ObjClass)

> while
> m = obj.meth.__get__(ObjClass, obj) is a bound method

you mean  m = obj.meth.__get__(obj,ObjClass)

> when saying obj.meth, you're doing the latter which means:
>  1. if the instance contains the bound method, return it.
>  2. if the class contains an unbound method with this name
>    bind it to the object (create a bound method object
>    and return it.

Correct.

> And then the special case semantics:
>  obj[x]
>
> calls ObjClass.__getitem__(self, x) directly, not going through the
> steps above (thus my confustion that I can't trap it in the metaclass).

It seems working as a charm:

>>> class C(object):
...     def __getitem__(self,a):
...             return a
...
>>> c=C()
>>> c[1]
1
>>> c.__getitem__.__get__(c,C)
<bound method C.__getitem__ of <__main__.C object at 0x402f6c0c>>
>>> c.__getitem__.__get__(c,C)(1)
1

Here obj[x] does not seems a special case ...

> Thus, if I ask for __getitem__ by name, we follow the general rules, if
> the __getitem__ semantics is invoked indirectly through subscripting, we
> follow the special rules.

The only special behavior I see (see my other posting) is that when __getitem__
is not defined c[x] gives a TypeError and not an AttributeError.



From donn at drizzle.com  Wed Apr  9 16:27:51 2003
From: donn at drizzle.com (Donn Cave)
Date: Wed, 09 Apr 2003 20:27:51 -0000
Subject: signals and system calls
References: <mailman.1049908218.16228.python-list@python.org>
Message-ID: <1049920070.202131@yasure>

Quoth Michael Stenner <mstenner at phy.duke.edu>:

| I have a long-running python program.  I want it to terminate nicely
| (ie not lose data) upon certain common signals.  The problem comes
| with system calls.  If a system call is interruped by a signal, it
| ends up raising an exception.  Now, the way you address this in C is
| to set SA_RESTART via sigaction, but that doesn't appear the be
| available in python.  [ I trust there is a good reason, btw, and am
| curious about it, but am trying to keep this email goal-oriented :) ]

Actually I think that's the key issue.  I don't know why, either,
but I don't think there's any satisfactory substitute, so I guess
in your shoes I'd be tempted to do this in C.  That could get hairy,
since signal handling is integrated a little more intimately with
the interpreter than your average external interface, but I'm not
looking at the source as I write this, maybe it doesn't need to be
all that bad.

	Donn Cave, donn at drizzle.com



From aahz at pythoncraft.com  Sat Apr 26 14:59:44 2003
From: aahz at pythoncraft.com (Aahz)
Date: 26 Apr 2003 14:59:44 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <mailman.1051376980.15852.python-list@python.org>
Message-ID: <b8ekv0$o3e$1@panix2.panix.com>

In article <mailman.1051376980.15852.python-list at python.org>,
Steven Taschuk  <staschuk at telusplanet.net> wrote:
>
>(Aside: A lot of questions of the form "why doesn't this work?"
>get answered here by programmers who are experienced enough that
>the problem is obvious, and who consequently might have trouble
>explaining how the querent could have figured it out themselves.
>It would be nice to be able to point people at a good on-line
>debugging tutorial for beginners; anybody know of such a thing?)

Doing a Google for "debugging tutorial" pulls this as the top link:
http://heather.cs.ucdavis.edu/~matloff/debug.html
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From sholden at holdenweb.com  Wed Apr 16 15:08:48 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Wed, 16 Apr 2003 19:08:48 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org> <slrnb9qlrf.7eb.rmunn@localhost.localdomain>
Message-ID: <4Thna.110739$It5.98016@news2.central.cox.net>

"Robin Munn" <rmunn at pobox.com> wrote ...
> Gerhard H?ring <gh at ghaering.de> wrote:
> > * Dan Jacobson <jidanni at dman.ddts.net> [2003-04-15 13:59 +0800]:
> >> I need to use one liners every day, especially in Makefiles.  I thus
> >> need to know how to make any kind of python program into a big long one
> >> liner. However, right out of the starting gates I run into
> >> $ python -c 'print 2;for i in (1,4):print i'
> >>   File "<string>", line 1
> >>     print 2;for i in (1,4):print i
> >>               ^
> >> SyntaxError: invalid syntax
> >
> > Try:
> >
> > python -c """
> > print 2
> > for i in (1, 4):
> >     print i
> > """
> >
> > It seems to work under a bash script at least. I once did something with
> ><<EOF but have since forgotten the little knowledge I had a about shell
> > scripting ;-)
>
> If you're in a shell script, you could do:
>
> ----- Cut here -----
> # ... do some shell script stuff ...
> rm temporary_script.py  # Make sure there isn't a symlink lying around
> cat >temporary_script.py <<EOF
> # Start of Python code
> import math
> print math.pi
> for i in range(5):
>     print i
> print "All done"
> EOF
> python temporary_script.py
> rm temporary_script.py
> # ... more shell script stuff ...
> ----- Cut here -----
>
> Notice the first "rm temporary_script.py" to make sure someone hasn't
> "carelessly" left a symlink lying around pointing to a file like
> /etc/passwd that you wouldn't want overwritten. There is still a race
> condition here -- if someone manages to create a symlink between the rm
> and the cat, you could be in trouble -- but it would be tricky to
> exploit, and you could do away with the problem entirely by using random
> (and therefore unpredictable) filenames for your temporary files.
>
> This way you clean up the directory after you're done, and you don't
> need to have a ton of small scripts lying around; you can consolidate
> them all into one big file. This is the approach used by the GNU
> autoconf suite -- take a look at the source of configure sometime.
>

In shell scripting, it's conventional to use the shell variable $$ in
filenames that must be more-or-less unique. Since $$ yields the (?) process
number, no other process is likely to collide with the filename. Your
file-creation dance would then be a little less likely to interfere with
other similar tasks.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From ianb at colorstudy.com  Tue Apr 22 14:29:54 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 22 Apr 2003 13:29:54 -0500
Subject: True/False
Message-ID: <1051036194.697.177.camel@lothlorien>

I've been using True and False for a while now, and I put this at the
top of my files:

True, False = 1==1, 0==1

But people seem concerned about this with Python 2.3.  I don't think
they should be concerned, while I'm reassigning True and False, I'm not
redefining them.  I could add something like:

import sys
if sys.version_info < (2, 3):
    True, False = 1==1, 0==1

But I *so* don't want to do this.  Is there any reason why reassigning
True and False will be a problem in the future?  Should I tell people
who are putting in version tests that it isn't necessary?

  Ian






From revanna at mn.rr.com  Tue Apr 22 22:37:46 2003
From: revanna at mn.rr.com (Anna)
Date: Wed, 23 Apr 2003 02:37:46 GMT
Subject: A few errors in "Python in a Nutshell" - Numeric related
References: <b6frd5$uq$1@peabody.colorado.edu>
Message-ID: <pan.2003.04.03.00.14.15.232608@mn.rr.com>

On Wed, 02 Apr 2003 16:23:16 +0000, Fernando Perez wrote:

> Hi,
> 
> I recently got Python in a Nutshell, and as I expected, I've found it to
> be extremely informative and well written.  I haven't read it in detail,
> but I'm glad I got it.
> 
> However, I'd like to point out a few problems I found in Chapter 15,
> Numeric processing.  I did not read it in excruciating detail, so please
> don't consider this an exhaustive bug hunt of that chapter.  And by the
> way, that chapter is a fabulous reference to NumPy, far better for most
> things than the standard manual which is organized in a 'cute' (read
> random) way and for some things is nearly useless.  Ch. 15 actually
> explains many things which the NumPy docs gloss over, and the function
> reference is *alphabetized*!! (how hard would it be to do that in the
> normal manual??)
> 
> Anyway, my few nits for future reference:

<snip>

> Anyway, these are just some small nits which readers which use Numeric may
> want to be aware of.  If there is an errata page for the book, it may be
> worth adding them to it.

Alex Martelli is at Python-UK. 

You might wish to submit this to the online errata form. Here's the url for doing so. 

http://www.oreilly.com/cgi-bin/errata.form/pythonian

cordially,
Anna



From tjreedy at udel.edu  Thu Apr 10 18:53:11 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Thu, 10 Apr 2003 18:53:11 -0400
Subject: Can I define nested classes in Python?
References: <169a7283.0304101432.13f5786c@posting.google.com>
Message-ID: <1JGdnc952fcaaAijXTWcrg@comcast.com>

"sdieselil" <sdieselil at yahoo.com> wrote in message
news:169a7283.0304101432.13f5786c at posting.google.com...
> Can I define nested classes in Python

Try it and see what happens.

tjr





From mike at bindkey.com  Tue Apr 22 21:22:32 2003
From: mike at bindkey.com (Mike Rovner)
Date: Tue, 22 Apr 2003 18:22:32 -0700
Subject: SGMLParser questions
References: <igmbavg5ughls6ksiden2rgnael1s0ub8q@4ax.com>
Message-ID: <b84puk$dev$1@main.gmane.org>

"Tung Wai Yip" <tungwaiyip at yahoo.com> wrote in message

>   attributes.get['src']
>
> but this doesn't work for the list. Is there some short cut to
> retrieve value in a list of (name, value) pairs?

dict(attributes).get['src']








From gh at ghaering.de  Fri Apr  4 17:03:25 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: Sat, 5 Apr 2003 00:03:25 +0200
Subject: how to test when correct output is "random"?
In-Reply-To: <20030404215312.GE7895@unpythonic.net>
References: <20030404215312.GE7895@unpythonic.net>
Message-ID: <20030404220324.GA1195@mephisto.ghaering.test>

* Jeff Epler <jepler at unpythonic.net> [2003-04-04 15:53 -0600]:
> My current project, Novelwriting[1] is a knock-off of something called
> the Dada Engine[2].  Essentially, you write a file that describes the
> structure of a document, and Novelwriting outputs a document that
> matches the structure.
> 
> I'd like to use this project as an opportunity to teach myself something
> about testing methodology, but since the output is based on a random
> number generator, how can the normal "unit test" or "doctest" tools work
> for me? [...]

As you don't want to test the quality of the PRNG, how about using a
custom PRNG that is *not random*, but *predictable* in the unit test?

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/




From sholden at holdenweb.com  Thu Apr 17 07:53:57 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 11:53:57 GMT
Subject: variable in Python
References: <b7m1g4$qqb$2@news-reader11.wanadoo.fr>
Message-ID: <pBwna.114820$It5.55830@news2.central.cox.net>

"Salvatore" <artyprog at wanadoo.fr> wrote ...
> Hello,
>
> I have a little doubt concerning variables.
> When one declare x = 3
> does x contains the value 3 or
> does x acts like pointer in C ?
>

Salvatore:

All Python variables are references to the data values that have been
assigned to them - as are list and dictionary elements. This can lead to
surprises for beginners, who don't understand why

>>> lst1 = [1, 2, 3]
>>> lst2 = lst1
>>> lst2[2] = "bam!"

leads to the lst1 variable showing the value [1, 2, "bam!"]. The answer, of
course, is that the third assignment changes the last element of the list
that is referred to by both variables, lst1 and lst2.

Many Python authors prefer to talk about "binding" nowadays, since
"assignment" implies to users of other languages that a value is being
placed into the variable. In Python, however, this isn't the case. Data
values are created in memory allocated from a heap, and this memory is only
reclaimed when no references to the data vablue remain. Once the data value
is created, variables reference it.

In  the CPython implementation each reference to a value increments its
reference count, and deletion of a reference decrements that count. Certain
"cyclic" structures, being self-referential, would never be reclaimed, and
so recent versions of CPython have a garbage collection algorithm that can
detect otherwise-unreferenced cyclic structures.

The JPython implementation, which compiles Python into Java bytecodes, does
not use reference counting at all and relies on Java's garbage colection
scheme to reclaim unused memory.

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From dale at riverhall.NOTHANKS.co.uk  Wed Apr 16 13:40:34 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Wed, 16 Apr 2003 18:40:34 +0100
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com> <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com> <cnoq9vci702imrunkenpulnefautq0n5ua@4ax.com> <Plena.22932$LB6.544795@news1.tin.it> <3cvq9vc9smcp7fbott8ivb1ujv0bgsk855@4ax.com> <O3gna.23369$LB6.556843@news1.tin.it>
Message-ID: <7c5r9vskojh1tnic92qr65fs092d443vr6@4ax.com>

Alex Martelli <aleax at aleax.it> wrote:

>Dale Strickland-Clark wrote:
>   ...
>>>def f(fileobj):
>>>    write = fileobj.write
>>>    write('')
>>>    # rest of function follows -- just call write('...') when needed
>> 
>> Way too fussy and obscure for my liking. Also, the assert is debugging
>> code and nicely goes away in production.
>
>...while the bound-methods stay in local variables and accelerate operations
>even "in production" -- oh woe!  No accounting for tastes, I guess;-)
>
>
>Alex

It might grow on me. ;-)

-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From rainer.klueting at epost.de  Tue Apr  8 04:09:08 2003
From: rainer.klueting at epost.de (Rainer Klueting)
Date: Tue, 08 Apr 2003 10:09:08 +0200
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com> <urnka.4658$Du.16154@newsc.telia.net>
Message-ID: <i6059vgtvavjp99ilkfeo8rtc003mha57l@4ax.com>

Hi Tim,

wow, so easy...!

Yes it works indeed. Thank you!

I'm just learning Python; I dropped onto it about a week ago and had
the strong impression that it might be what I was looking for. :-)

I had already seen that there is another GUI toolkit called wxWindow,
and I even had downloaded it. But my impression was, that Tkinter is
kind of a GUI-standard for Python, and I tried to stick to the
standard.

Is that true? Do you have any information about how widespread the use
of wxWindows is? I don't want to steer into a dead-end-road.

Thanks again!

Rainer Klueting, Stuttgart, Germany




From duncan at NOSPAMrcp.co.uk  Mon Apr 28 05:32:41 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Mon, 28 Apr 2003 09:32:41 +0000 (UTC)
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com> <7ULqa.9804$3M4.264757@news1.tin.it> <g3qdncyq1Zx2FzGjXTWcrg@comcast.com> <mailman.1051497463.22158.python-list@python.org>
Message-ID: <Xns936B6B217EB86duncanrcpcouk@127.0.0.1>

Lulu of the Lotus-Eaters <mertz at gnosis.cx> wrote in 
news:mailman.1051497463.22158.python-list at python.org:

> And to be even trickier, answer the question for the following lists:
> 
>     [u'x', 'x', chr(255)].sort()
>     ['x', chr(255), u'x'].sort()
>     [1j, '1', 1].sort()
>     [1, 1j, '1'].sort()
> 
> 
> Then try to explain the answer to ANYONE you know.

I must admit I was surprised that 3 of these failed to sort, since I'd 
guessed that you would have selected 2 that worked and 2 that failed. So my 
question is, did one of the last two actually work on your system?

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From duncan at NOSPAMrcp.co.uk  Tue Apr 22 06:42:52 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 22 Apr 2003 10:42:52 +0000 (UTC)
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <d6jmqr2q.fsf@python.net> <Xns936072232AA84duncanrcpcouk@127.0.0.1> <he8xgr3g.fsf@python.net> <Xns9360A249561CBduncanrcpcouk@127.0.0.1> <Xns9360A663C7BA5duncanrcpcouk@127.0.0.1> <znmpf64d.fsf@python.net> <d6jlf23c.fsf@python.net>
Message-ID: <Xns93657709D4D28duncanrcpcouk@127.0.0.1>

Thomas Heller <theller at python.net> wrote in news:d6jlf23c.fsf at python.net:

> Would you care to upload your patch to SF, so it doesn't get lost?
> 
Ok, uploaded as patch #725569.

I modified it a bit further before uploading it. There is no longer any 
dependency on an undocumented difflib function, I extracted just the 
relevant code which makes my patch smaller and faster as well as cleaner.
I also timed the code for finding the first difference, and increased the 
slice size to 1000 characters as that seems to provide about the best 
compromise performance both for short and long strings.

> Shouldn't there by a python function which limits the output of
> 'print obj' to *some* amount?

Perhaps an option on pprint? pprint.PrettyPrinter already has arguments 
depth (how far to drill down) and width (the screen width), someone could 
add a height option to try to constrain the output to a certain number of 
lines.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From duncan at NOSPAMrcp.co.uk  Fri Apr 25 03:47:19 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Fri, 25 Apr 2003 07:47:19 +0000 (UTC)
Subject: sys.argv problem in Win 2k
References: <bd993a2f.0304240121.7c795cf0@posting.google.com> <mailman.1051177117.30168.python-list@python.org> <bd993a2f.0304240506.5b206f3d@posting.google.com> <Xns936790B52180Eduncanrcpcouk@127.0.0.1> <bd993a2f.0304242310.586305df@posting.google.com>
Message-ID: <Xns9368583183AC8duncanrcpcouk@127.0.0.1>

I wrote:
>> D:\temp>ftype Python.File
>> Python.File=D:\Python22\python.exe "%1" %*

abpillai at lycos.com (Anand B Pillai) wrote in
news:bd993a2f.0304242310.586305df at posting.google.com: 

> F:\users\anand\LocalWorkspaces\ResModWS>ftype Python.File
> Python.File=D:\Python22\python.exe "%1"
> 

Spot the difference.

The "%1" passes the name of the script to python.exe, the %* passes all the 
other arguments. Without the %* windows will simply drop the other 
arguments on the floor. 

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From tanzer at swing.co.at  Wed Apr  9 03:00:24 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Wed, 09 Apr 2003 09:00:24 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "Wed, 09 Apr 2003 12:38:05 +1200."
             <b6vpfa$9imjc$1@ID-169208.news.dfncis.de>
Message-ID: <E1939ZI-0002dE-00@tswings.swing.cluster>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> wrote:

> Christian Tanzer wrote:
> > If you want to implement something comparable
> > to Emacs' power, you'd better look at the Elisp Reference Manual for
> > quite a while.
>
> That's one way of looking at it. Arguably, however, it
> might be better *not* to study the Elisp Reference Manual
> too closely, so as not to pollute your mind with all
> those crufty hacks and jeopardise your chances of
> coming up with a better overall design. A sort of
> conceptual clean-room approach.

Well, I didn't argue for cloning the cruft but for gathering
information about the problem space. For all the crufty hacks, Emacs
is an excellent editor. Ignoring the past is rarely a good guide to
improve the state of the art.

IMHO it is not possible to do top-down clean-room design for such a
piece of software. Given enough sweat, it should certainly be possible
to come up with a substantially better overall design.

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From peter at engcorp.com  Sun Apr 20 11:58:35 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 20 Apr 2003 11:58:35 -0400
Subject: Toronto-area Python user group: meeting Tuesday April 22nd
Message-ID: <mailman.1050926645.15848.clpa-moderators@python.org>

We apologize for the late notice (again!):  PyGTA, the user group 
for Python, Zope and their ilk in the Greater Toronto Area and environs, 
will be meeting at the usual location (519 Church St. Community Centre, 
room 23), time (8pm to 10pm) and date (fourth Tuesday of the month, or 
April 22nd this time).

Nobody has (yet?) committed to doing a presentation this month, but 
Peter Hansen, Director of Software Engineering at Kaval Wireless 
in Markham hopes to present a brief introduction to "Using Python for 
Industrial Control".

Please come and support the group with your participation.  The usual 
post-meeting meeting will be convened at Just Desserts on Yonge.

Meeting information is always available on our web site at
http://web.engcorp.com/pygta/wiki/NextMeeting .

If you didn't receive this message by email, you are not yet on our 
announcement mailing list.  See the web page to subscribe.

-Peter Hansen, Ian Garmaise, meeting organizers




From skip at pobox.com  Mon Apr  7 15:14:59 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 7 Apr 2003 14:14:59 -0500
Subject: socket question(2)
In-Reply-To: <20030407184913.69195.qmail@web20707.mail.yahoo.com>
References: <20030407184913.69195.qmail@web20707.mail.yahoo.com>
Message-ID: <16017.52787.696861.182511@montanaro.dyndns.org>

    Rick> eventhough i have installed openssl in my solaris 2.8
    Rick> box, i still get these errors. ..???
    ...
    Rick> ImportError: ld.so.1: python: fatal: libssl.so.0.9.6:
    Rick> open failed: No such file or directory

Did your openssl install create a file named libssl.so.0.9.6?  If not,
that's the reason.  If so, make sure it's in a location which ld.so will
look.  If it's not, try setting LD_LIBRARY_PATH before running the Python
interpreter.

Skip





From mis6 at pitt.edu  Mon Apr 21 11:52:16 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 21 Apr 2003 08:52:16 -0700
Subject: Is this a super bug?
References: <mailman.1050817233.1530.python-list@python.org>
Message-ID: <2259b0e2.0304210752.561a4cc3@posting.google.com>

Oops!!

I forgot a couple of dirs:

>>> dir(super(ff,f))

and 

>>> dir(Super(ff,f))

give the outputs I wrote before ;)



From ianb at colorstudy.com  Tue Apr 22 20:00:08 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 22 Apr 2003 19:00:08 -0500
Subject: True/False
In-Reply-To: <3EA5B173.9420324C@alcyone.com>
References: <mailman.1051036237.30103.python-list@python.org>
	 <3EA5B173.9420324C@alcyone.com>
Message-ID: <1051056008.697.250.camel@lothlorien>

On Tue, 2003-04-22 at 16:17, Erik Max Francis wrote:
> If done, then a blanket statement like yours above would start
> generating warnings eventually; they'd be harmless warnings, but they'd
> be unnerving and annoying.
> 
> > import sys
> > if sys.version_info < (2, 3):
> >     True, False = 1==1, 0==1
> 
> This would be okay (although the untyped version of the True and False
> builtins were introduced in 2.2, not 2.3).  The solution I've adopted
> is:
> 
> 	try:
> 	    True
> 	except NameError:
> 	    True, False = 1, 0

Not safe at all:

Python 2.3a2+ (#2, Mar 21 2003, 22:13:05) 
[GCC 3.2.3 20030316 (Debian prerelease)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> try:
...     None
... except NameError:
...     None = 0
... 
<stdin>:4: SyntaxWarning: assignment to None

Syntax-related errors/warnings aren't runtime.  Though strangely:

>>> if 0:
...     None = 0
... 
>>> if False:
...     None = 0
... 
<stdin>:2: SyntaxWarning: assignment to None


Maybe there's an optimization going on in this case, though (and to do
the optimization properly in both cases you'd have to catch assignment
to False too!)

  Ian






From jkraska at san.rr.com  Thu Apr 17 23:11:25 2003
From: jkraska at san.rr.com (Courageous)
Date: Fri, 18 Apr 2003 03:11:25 GMT
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <87k7dt5m3j.fsf@interim.henrik-motakef.de> <1azna.115402$It5.22896@news2.central.cox.net> <tyfhe8xqhmy.fsf@pcepsft001.cern.ch>
Message-ID: <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com>

>But he regrets the inclusion of a crippled lambda, rather than the
>inclusion of support for anonymous functions (IIRC).

I was toying around in my own python parser about a year ago,
attempting to implement anonymous functions there. It was a
struggle, and in the end, I felt basically *forced* down the
same path the Python team went, until I included a set of tokens
for explicit scope (hey! this is just a toy environment, don't
look scandalized!).

Anyway, I guess what I'm saying is that I suspect that Guido
didn't have any good alternatives. If someone has some, that
fit well with our beloved block delimited language, SHOW ME
DA MONEY! :-)

C//




From aleax at aleax.it  Wed Apr 23 10:50:21 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 14:50:21 GMT
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz>
Message-ID: <NKxpa.85054$T34.2112670@news2.tin.it>

Frantisek Fuka wrote:

> I want a simple test to see if str contains any occurences of substr. I
> am currently using this:
> 
> import string
> if string.find(str,substr) + 1:
>      ...do something...
> 
> But it somehow looks "ugly" to me. Is there more "standard" way to test
> this?

With the close-to-beta-release Python 2.3, sure:

[alex at lancelot src]$ python2.3
Python 2.3a2+ (#11, Mar 28 2003, 12:17:11)
[GCC 3.2 (Mandrake Linux 9.0 3.2-1mdk)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> 'foo' in 'barfooble'
True
>>> 'gzap' in 'barfooble'
False
>>>

and more generally, "needle in haystack".

However, such new functionality doesn't get back-ported to previous
releases of Python, such as 2.2.*.  In all Python releases from 1,6
included to 2.3 excluded, "needle.find(haystack) > 0" is the idiom.
Module string is needed if you want to use antediluvian versions of
Python, such as 1.5.2, where strings didn't have methods.


Alex




From aleax at aleax.it  Tue Apr 15 08:29:49 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 12:29:49 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <7h3vfxhpl23.fsf@pc150.maths.bris.ac.uk> <yfsu1d1cisj.fsf@black132.ex.ac.uk> <7h3adettapc.fsf@pc150.maths.bris.ac.uk> <yfsptnpce02.fsf@black132.ex.ac.uk> <7h3k7dwrt2h.fsf@pc150.maths.bris.ac.uk>
Message-ID: <1XSma.22915$T34.689468@news2.tin.it>

Michael Hudson wrote:

> Alexander Schmolck <a.schmolck at gmx.net> writes:
> 
>> > Requiring declarations of all variables would be a HUGE change to
>> > Python, IMHO.
>> 
>> Sure and I'm in no way arguing for such a change (and neither Paul
>> Foley who, unlike me, considers implicit declaration a major
>> flaw). I just don't understand why it would "unfeasible
>> syntactically" -- *syntactically* it seems quite unproblematic to
>> me.
> 
> Well, maybe I should have been more clear than just "syntactically" --
> but you do understand what I mean, no?  I don't think you can
> distingusish establishment and mutation of bindings in Python without
> significant damage to the "feel" of the language.

I think you COULD choose to add to Python two special assignment
delimiters -- say := and ~= for definiteness -- with semantics as
follows:

    <name> := <expression>
    
        if <name> was previously unbound in this scope, bind it;
        if <name> was already bound, raise an exception.


    <name> ~= <expression>

        if <name> was already bound in this scope, re-bind it;
        if <name> was previously unbound, raise an exception.

while leaving good old = to mean exactly what it means today.

Would such additions "significantly damage the feel of the
language"?  I don't know.  I doubt they'd add any important
expressiveness, or see any significant use; therefore, in
practice, I think they'd remain obscure, rarely-used "dusty
corners of the language" and thereby indirectly do some
damage (complexity not pulling its weight).  But I don't see
any sense in which these difficulties could possibly be
described as "syntactical", even lato sensu.


Alex





From mis6 at pitt.edu  Wed Apr 30 10:25:00 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 30 Apr 2003 07:25:00 -0700
Subject: Question about accessing class-attributes.
References: <mailman.1051662005.2247.python-list@python.org>
Message-ID: <2259b0e2.0304300625.4e0ebace@posting.google.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1051662005.2247.python-list at python.org>...
> SF 729913. I have not been able to find an exact description of the
> lookup rules for new-style classes anywhere.
> 
> I'm also missing an exact description of what super() returns. It looks
> like it's an object that when you call method 'm', invokes the proper
> super class' method 'm', and if that 'm' invokes any otherm method,
> 'm2', the original objects definiton 'm2' is used. So it looks like it
> should be substitutable, or at least a first class object, but as you
> said: "of course not". Am I missing some documentation, or is it
> unreasonable of me to think it should work this way (what way should it
> work... exactly)?
> <snip> The
> thinking was that if Super's metaclass could intercept special methods
> through it's   getattr  , the object returned from super() could look
> more like a regular object.  That Super.  getattr   doesn't check for a
>   getattr   in the   mro   chain is a problem, but a separate one (sorry
> for the confustion).
> 
> -- bjorn

super seems to be quite subtle, indeed. Unfortunately, the only documentation
is Guido's essay, which does not describe the internal details of the
implementation, it shows only a Python implementation, as you know already.

I think I understand your point, now. It is like in the following example:

class B(object):
    def __getattr__(self,somename):
        return somename

class C(B):
    pass

b=B()
c=C()

print b.name # works
print c.name # works
print super(C,c).__getattr__('name') # "means" B.__getattr__(c,'name'); works
print super(C,c).name # does not work

"Means" does not mean it is *literally* equivalent. If it was so,
this would work:

class M(type):
    def __getattr__(cls,somename):
        return somename
    
class B(object):
    __metaclass__=M
    def __getattr__(self,somename):
        return somename

class C(B):
    pass

b=B()
c=C()

print B.name # works
print C.name # works
print b.name # works
print c.name # works
print super(C,c).__getattr__('name') # works
print super(C,c).name # does not work, it is not equivalent to B.name !

super(C,c) is a proxy to the methods of class B, passing to them
the instance c, but it does not see if the lookup on B has been
altered by a custom metaclass overriding __getattr__

Not sure if this is a bug or a design choice, anyway. You should post to
python-dev. I am more bothered by more trivial issues with super, like the
following one:

class C(object):
    pass

C.s=super(C)

help(C)

Traceback (most recent call last):
  File "<stdin>", line 6, in ?
  File "/usr/local/lib/python2.3/site.py", line 293, in __call__
    return pydoc.help(*args, **kwds)
  File "/usr/local/lib/python2.3/pydoc.py", line 1539, in __call__
    self.help(request)
  File "/usr/local/lib/python2.3/pydoc.py", line 1575, in help
    else: doc(request, 'Help on %s:')
  File "/usr/local/lib/python2.3/pydoc.py", line 1368, in doc
    pager(title % desc + '\n\n' + text.document(object, name))
  File "/usr/local/lib/python2.3/pydoc.py", line 279, in document
    if inspect.isclass(object): return self.docclass(*args)
  File "/usr/local/lib/python2.3/pydoc.py", line 1122, in docclass
    lambda t: t[1] == 'method')
  File "/usr/local/lib/python2.3/pydoc.py", line 1057, in spill
    name, mod, object))
  File "/usr/local/lib/python2.3/pydoc.py", line 280, in document
    if inspect.isroutine(object): return self.docroutine(*args)
  File "/usr/local/lib/python2.3/pydoc.py", line 1145, in docroutine
    realname = object.__name__
AttributeError: 'super' object has no attribute '__name__'

Here is the simple fix. I have submitted the bug report already.

class Super(super):
    def __init__(self,cls,obj=None):
        super(Super,self).__init__(cls,obj)
        self.__name__='super(%s)' % cls.__name__

class C(object):
    pass

C.s=Super(C)

help(C)

--------------------------------------------------------------------------
Help on class C in module __main__:

class C(__builtin__.object)
 |  Methods defined here:
 |  
 |  s = super(C)(...)
 |  
 |  ----------------------------------------------------------------------
 |  Data and other attributes defined here:
 |  
 |  __dict__ = <dictproxy object>
 |  
 |  __weakref__ = <attribute '__weakref__' of 'C' objects>


Notice the advantage with the present implementation of super as a
class: you can subclass it and correct its flaws ;)

HTH,

                     Michele



From nanthan79 at lycos.com  Mon Apr  7 02:33:49 2003
From: nanthan79 at lycos.com (Pathmenanthan Ramakrishna)
Date: Mon, 07 Apr 2003 06:33:49  0000
Subject: NEED HELP!!!EDITOR
Message-ID: <FKLGAICCOMIIPAAA@mailcity.com>

hi,

im Nanthan, a postgraduate student from Malaysia.im very new to 
python language but im already getting use of it.Currently,i 
have been working in windows, but my development actually is on 
Linux, so can u suggest me the best editor(i mean use friendly 
and graphical,i dont want command line editor)for python?

is cooledit the right choice.

i hope u'll reply soon

thanks

Nanthan.R



____________________________________________________________
Get 25MB, POP3, Spam Filtering with LYCOS MAIL PLUS for $19.95/year.
http://login.mail.lycos.com/brandPage.shtml?pageId=plus&ref=lmtplus




From ariza at flexatone.com  Fri Apr  4 15:19:10 2003
From: ariza at flexatone.com (christopher ariza)
Date: Fri, 4 Apr 2003 15:19:10 -0500
Subject: writing a midi file w/ midi.py / python
In-Reply-To: <mailman.1049485689.18855.python-list@python.org>
References: <9383481e.0304012011.78f05548@posting.google.com>
 <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com>
 <mailman.1049388384.29614.python-list@python.org>
 <9383481e.0304031917.1d0100bf@posting.google.com>
 <mailman.1049485689.18855.python-list@python.org>
Message-ID: <Pine.OSX.4.50.0304041508520.1664-100000@pythagoras-v.local>

does anyone have some sample code of how to write a basic midi file with
python?

i dont need to deal with real-time communication, i just want to
take a few note-events (as a python data structure) and spit out a midi file.

i have been looking at Will Ware's midi.py:

http://www.python.or.kr/pykug/midi_2epy

but cant get it to do what i want. i cant find any more information about
this module, examples of it in use, or a way to contact the author. anyone
every use midi.py?

if anyone has some tips, many thanks in advance.

________________________________________
christopher ariza \ http://flexatone.net





From alloydflanagan at attbi.com  Tue Apr  8 10:14:55 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 8 Apr 2003 07:14:55 -0700
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org> <b6fj57$626$1@news-reader11.wanadoo.fr> <3E8B53BA.DBB0C84C@engcorp.com>
Message-ID: <a8b7f07a.0304080614.251fce66@posting.google.com>

Peter Hansen <peter at engcorp.com> wrote in message news:<3E8B53BA.DBB0C84C at engcorp.com>...
...
> 
> Generally this doesn't work.  You have to import
> a module (or use other techniques) for it to be compiled.
> Python does not store a .pyc or .pyo file for the main 
> module, perhaps surprisingly.
> 
> -Peter

Yes, that's surprising.  Does anyone happen to know why this is the
case?  It doesn't make sense to me.



From gh at ghaering.de  Tue Apr 15 17:42:00 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: Tue, 15 Apr 2003 23:42:00 +0200
Subject: MySQLdb: getting inserted data
In-Reply-To: <b7hq7q$v3e$1@news.tu-darmstadt.de>
References: <b7hq7q$v3e$1@news.tu-darmstadt.de>
Message-ID: <20030415214200.GA11833@mephisto.ghaering.test>

* Mirko Koenig <koenig at v-i-t.de> [2003-04-15 22:31 +0200]:
> Hi
> 
> I want to insert a new empty row into my table and get the value of the
> primary key, so that i can identify it. [...]

Use the .lastrowid property of the cursor object. It's an optional
DB-API extension that MySQLdb implements. E. g.:

# cx is the connection object
cu = cx.cursor()
name = "Alice"
cu.execute("INSERT INTO PERSON(PERS_NAME) VALUES (%s)", (name,))

title = "Macbeth"
# BK_OWNER is the foreign key that refers to the PERSON table:
cu.execute("INSERT INTO BOOKS(BK_TITLE, BK_OWNER) VALUES (%s, %i)",
        (title, cu.lastrowid))
cx.commit()     # These two INSERTs are atomic

HTH,

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/




From tweedgeezer at hotmail.com  Fri Apr 11 17:38:46 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 11 Apr 2003 14:38:46 -0700
Subject: module for doing unix "find"-like things?
References: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com>
Message-ID: <698f09f8.0304111338.70b20757@posting.google.com>

Mark Harrison <mh at pixar.com> wrote in message news:<7LDla.1483$Pq2.135830754 at newssvr14.news.prodigy.com>...
> Is there such a thing?  I didn't see such a thing, but
> "find" is not the easist thing to search for.
> 
> Specifically, I would like to replace an external call to
>     for i in `find $top -type l -maxdepth 1`; ...
> 
> Many TIA,
> Mark

You're looking for os.path.walk and fnmatch/glob.

Jeremy



From donn at drizzle.com  Wed Apr 23 01:21:50 2003
From: donn at drizzle.com (Donn Cave)
Date: Wed, 23 Apr 2003 05:21:50 -0000
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <f6a31cb8.0304220810.547f5acd@posting.google.com> <oRKdnYDxiYyb5TijXTWcpA@comcast.com>
Message-ID: <1051075309.87191@yasure>

Quoth "Terry Reedy" <tjreedy at udel.edu>:
| "Ted Drain" <teddrain at earthlink.net> wrote in message
| news:f6a31cb8.0304220810.547f5acd at posting.google.com...
|> My $.02:
|>
|> We use Python as a scripting interface to an underlying C++ system.
|> The users write Python code to do what they want.  We ended up writing
|> __repr__ and __str__ in our classes to be the same thing because our
|> users were very annoyed when they typed:
|>
|> >>> s
|> >>> print s
|>
|> and got different results.
|
| If you think of repr(o) as the codewriter view of an object o and
| str(o) as the (passive) user view of o, then printing repr(o) in
| interactive (codewriter) mode makes perfect sense.

Coincidentally, today I was looking into a little Mailman problem.
Its subclass of Message defines __repr__ to return __str__().
Unfortunately, __str__ is a fairly involved function that not only
returns information that's completely inappropriate for situations
like the interpreter, it's also prone to failure with Unicode troubles.

So here's this object.  You can't repr() it, including implicitly
from the interpreter prompt - you get a traceback instead.  Of course
that also goes for the list object that contains this object.  To
me, this is not right, and maybe it ought to be another point to
consider if we're reconsidering the documentation for __repr__.
Wish I could specify this better, but I believe there's a minimalist
way to think about implementing __repr__ that will tend to make it
less interesting but more reliable.

	Donn Cave, donn at drizzle.com



From trooper at ttnet.net.tr  Tue Apr 29 11:44:41 2003
From: trooper at ttnet.net.tr (=?iso-8859-9?Q?Tongu=E7?= Yumruk)
Date: Tue, 29 Apr 2003 18:44:41 +0300
Subject: Pipe Problem
Message-ID: <20030429154441.GA2679@Serafettin.fazlamesai.net>

Hi,

I don't know if it's the proper place to ask but I couldn't find any
better place.

Here is the problem:

whichpipe = os.popen("which xloadimage")
billGates = whichpipe.readlines()
whichpipe.close()

It works when I try it on python interpreter, itworks when I execute it
in my script... But it never works when I call it from lineakd.

I have one of those fancy "Easy Access" keyboards, and I am using LinEAK
(http://lineak.sourceforge.net) to use the extra buttons on my keyboard.
I have a little python script that changes the wallpaper randomly...
When I call it from lineakd, this happens:

Traceback (most recent call last):
  File "/home/tonguc/Python/randombg", line 37, in ?
    whichpipe.close()
IOError: [Errno 10] No child processes

When I don't close the pipe it all works perfect, but I think it's not
good coding when you use a pipe then assign the same name to another
pipe without closing it.

I'm using Debian Woody r1, Python 2.1.3

Thanks...

--
Love Respect Linux
################################################################################
: The following (relative to AutoSplit 1.03) attempts to please everyone
: and perhaps pleases no one:
 
I think that's way cool.
             -- Larry Wall in <199709292015.NAA09627 at wall.org>
################################################################################
Tongu? Yumruk
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030429/e6da7b3e/attachment.sig>

From jon+usenet at unequivocal.co.uk  Wed Apr 30 17:08:49 2003
From: jon+usenet at unequivocal.co.uk (Jon Ribbens)
Date: 30 Apr 2003 21:08:49 GMT
Subject: Web programming and a different "type" problem
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com> <slrnbarhfb.67t.jon+usenet@snowy.squish.net> <b8lj2q$l14$1@newshost.accu.uu.nl>
Message-ID: <slrnbb0er1.fb6.jon+usenet@snowy.squish.net>

In article <b8lj2q$l14$1 at newshost.accu.uu.nl>, Martijn Faassen wrote:
> Doesn't jonpy allow the Evil User to manipulate the type of your Python
> variables too? The Evil User could get rid of the '*' in the form and
> submit, right? Though I grant it looks like the code will always fail
> in that case, which is good. It's a neat idea.

Well, you answer your own question. The "*" is kept in the 'params'
mapping so when the code looks for "foo*" it wouldn't find it, which
is a situation it has to cope with anyway when a genuine user simply
selects none of the options.



From jepler at unpythonic.net  Tue Apr  8 22:51:29 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Tue, 8 Apr 2003 21:51:29 -0500
Subject: Problem with Tkinter
In-Reply-To: <3E9361FD.6000504@teknovusDOTcom>
References: <3E9361FD.6000504@teknovusDOTcom>
Message-ID: <20030409025128.GA1488@unpythonic.net>

On Tue, Apr 08, 2003 at 11:55:30PM +0000, news.easynews.com wrote:
> I just managed to get Python configured to run with TK. I am running 
> Python 2.2.2 with TCL/TK 8.3.5 on Redhat linux 8.0.
> 
> When I try to run the recommended TK test, I get the following error:

Works fine here (redhat 9, etc)

What does
	>>> unicode("\347", "iso-8859-1")
give you?  It should give you
	u'\xe7'
but I'm guessing it'll give you the same LookupError message.  This
indicates that something about unicode codecs is incorrectly installed, not
a problem with Tkinter (necessarily)

    if TclVersion >= 8.1:
        try:
            text = text + unicode("\nThis should be a cedilla: \347",
                                  "iso-8859-1")
        except NameError:
            pass # no unicode support

I've never *heard* of a python misinstalled in this way, though...

Jeff




From sholden at holdenweb.com  Mon Apr  7 21:15:45 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 08 Apr 2003 01:15:45 GMT
Subject: How to protect python code ?
References: <mailman.1049758213.26256.python-list@python.org> <slrnb945tu.p4.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <5ppka.23067$It5.15751@news2.central.cox.net>

"Ben Finney" <bignose-hates-spam at and-zip-does-too.com.au> wrote in message
news:slrnb945tu.p4.bignose-hates-spam at rose.localdomain.fake...
> On Mon, 07 Apr 2003 16:33:44 -0700, Brian Quinlan wrote:
> > >> What kind of business *pays* to become dependent on
> > >> software they don't control?!
> >
> >  I'm sure that the original poster was being facetious but...power
> >  customers, telecom customers, computer customers.
>
> I didn't read facetiousness; rather, I read in the original question a
> sentiment perhaps better expressed as:
>
>   "If a business *pays* to become dependent on software they don't
>   control, what can one conclude about that business?"
>
> or even:
>
>   "How could it possibly be of benefit to a business to *pay* to become
>   dependent on software they don't control?"
>
> Thus, a rhetorical question perhaps.  The "What kind of business" was
> not a query for what *industry* the business is in, but rather in a
> spirit of "what kind of idiot".
>

Of course, the businesses that make a good living out of proprietary
software encourage all sorts of FUD about open source. Consequently, large
organizations implement rules forbidding the use of open source software
products because "they aren't supported" when in fact the support for some
open source products is actually superior to that of most proprietary
products.

The fact of the matter is that most companies would prefer to pay for a
black-box proprietary product because ultimately there's assumed to be a
responsible (and accountable) organization behind it. If they bothered to
read the licenses, of course, they'd realize that the producer's liability
is usually limited to the purchase price of the software.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From just at xs4all.nl  Tue Apr 29 15:53:27 2003
From: just at xs4all.nl (Just)
Date: Tue, 29 Apr 2003 21:53:27 +0200
Subject: Python 2.3b1: RuntimeError using rexec
References: <b8mh4l$2oo0$1@nntp6.u.washington.edu>
Message-ID: <just-53D0EC.21532729042003@news1.news.xs4all.nl>

In article <b8mh4l$2oo0$1 at nntp6.u.washington.edu>,
 "Russell E. Owen" <no at spam.invalid> wrote:

> I stumbled across this when testing some existing code. This code uses 
> rexec to convert a string representation of a dictionary to a dictionary 
> as part of reading in font preferences.
> 
> Any suggestion on how to do that without rexec would also be 
> appreciated. (I was quite disappointed that "dict" didn't do it!)

For this I often use

def safeEval(data):
   """A safe replacement for eval."""
   return eval(data, {"__builtins__": {}}, {})

Exactly _how_ safe it is I don't know, but it's safe enough for my 
purposes...

Just



From piet at cs.uu.nl  Fri Apr 25 18:42:30 2003
From: piet at cs.uu.nl (Piet van Oostrum)
Date: 26 Apr 2003 00:42:30 +0200
Subject: Twisted 1.0.4
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local> <341410000.1051007646@leeloo.intern.geht.de> <20030422121836.GA8567@frobozz.local> <1051031635.18957.24.camel@software1.logiplex.internal> <mailman.1051064256.4034.python-list@python.org> <3EA60532.6ADC21C@alcyone.com>
Message-ID: <wzu1cm8ai1.fsf@nono.cs.uu.nl>

>>>>> Erik Max Francis <max at alcyone.com> (EMF) wrote:

EMF> Andrew Bennetts wrote:
>> Using gcc doesn't require you to link against anything, let alone
>> LGPLed
>> libraries -- although of course that is the most common situation;
>> most
>> systems using gcc will also be using glibc.  But the libraries linked
>> against
>> are a seperate issue to the point I was making about gcc; using gcc
>> doesn't
>> necessitate using LGPLed libraries, ...

EMF> While that's in some sense technically true, any use of gcc to build an
EMF> application -- that is, actually create an application rather than
EMF> simply compile to object code via the -c option -- _will_ silently link
EMF> in those pesky LGPL libraries like libc or libg++.

That depends on the platform. For example Mingw, which also is gcc, links
to the Microsoft libraries rather than glibc, AFAIK.
-- 
Piet van Oostrum <piet at cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.van.Oostrum at hccnet.nl



From synthespian at uol.com.br  Thu Apr 17 23:57:53 2003
From: synthespian at uol.com.br (synthespian)
Date: Fri, 18 Apr 2003 00:57:53 -0300
Subject: Why functional Python matters
In-Reply-To: <yfs7k9vcu2p.fsf@black132.ex.ac.uk>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk>
Message-ID: <b7nt4s$2o4qe$1@ID-78052.news.dfncis.de>

Alexander Schmolck wrote:
> Dave Benjamin <ramen at lackingtalent.com> writes:
> 
> 
>>The next thing I read will probably not surprise you if you have ever done a
>>Google search on functional programming. It was David Mertz's "Charming
>>Python" series on IBM's DeveloperWorks. His papers were not easy to digest
>>in one sitting, but I made a great effort to understand them and apply the
>>techniques he described. At this point, I should mention that I was still
>>programming only in PHP!
> 
> 
> I think your life might be much happier if you just went to
> http://www.drscheme.org/, dowloaded Dr Scheme and used *that* for
> webdevelopment. Then you can use a real functional programming language and
> read books by people who actually have a real clue about functional
> programming (there is lots of *excellent* material available on-line and for
> free, like SICP and HTDP, the latter uses Dr Scheme -- also have a look at
> http://www.cs.brown.edu/courses/cs173/2001/Lectures/). Why settle for some
> second-rate immitation? [1]
> 
> (Dr Scheme *does* have web libraries, I've never used them but I'd suspect
> they shouldn't be too bad, because they seem to have hit on trick of using
> webdevelopment as application domain to bring things like continuations to the
> unwashed masses (viz. their sophomores) :).
> 
  Hahaha...

  The OP might also be interested, then, in the Clean language. This is 
haskell-like, but with modifications in the type system (combines static 
/and/ dynamic typing and has destructive updates [1]), to improve 
performance (it compiles). Some issues remain in what regards exception 
handling, it seems (for pretty much the same reasons that they remain in 
haskell, probably. I'm no expert on FP, so I might be wrong).
  It was developed by serious people in the FP research community, but 
it is intended to be used in the "real" world. They have an IDE writen 
in Clean, I/O libraries, and even a graphics Game library! So, they mean 
what they say.
  Licensed under the LGPL. IDE for win32 only (for now).


home page:
http://www.cs.kun.nl/~clean/

some libraries: 
http://www.cs.kun.nl/~clean/About_Clean/page_modified/page_modified.html

foldoc (kind of useless, too outdated)
http://wombat.doc.ic.ac.uk/foldoc/foldoc.cgi?Clean

Regs

Henry

[1] "Clean is the only functional language in the world which offers 
/uniqueness typing/. This type system makes it possible in a pure 
functional language to incorporate destructive updates of arbitrary data 
structures (including arrays) and to make direct interfaces to the 
outside imperative world. The type system makes it possible to develop 
efficient applications."





From mcfletch at rogers.com  Sun Apr  6 14:41:51 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 06 Apr 2003 14:41:51 -0400
Subject: properties and get/set methods
In-Reply-To: <lA6dnc_HzLQazg2jXTWcqg@comcast.com>
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org> <lA6dnc_HzLQazg2jXTWcqg@comcast.com>
Message-ID: <3E9074EF.1070504@rogers.com>

Terry Reedy wrote:

>"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
>news:mailman.1049604615.22077.python-list at python.org...
>  
>
>>In these cases, you're using properties to simplify the modelling of
>>domain objects, or to provide some *generic* functionality over and
>>above a regular value get/set.  You don't want to have to specify
>>    
>>
>that
>  
>
>>functionality for every property, you just want to say "this is one
>>    
>>
>of
>  
>
>>_those_ properties".  The code bloat (not to mention mindless
>>duplication of code) when you get up to 15 or 20 properties for an
>>object would be ridiculous.
>>    
>>
>
>Perhaps what you want are __get/setattr__ methods that check if the
>attribute is one of 'those-properties' (in a set/dict lookup) and if
>so, does the generically different thing.  My impression is that
>properties are meant for doing attribute specific customization
>without loading up/slowing down __get/setattr__ with a long list of
>if/elif blocks.
>
>Terry
>  
>

The annoyance there is that you then have to have every object aware of 
every property-type you want to support.  You either wind up with a 
"super property-holding class" from which every object must be derived, 
and which is aware of all of the property types, or you do holder-class 
mix-ins, or you have some sort of plug-in system to lookup the 
appropriate function based on the property type... sounds a lot like 
what you get with having the property's type (i.e. class) store the 
appropriate method to begin with.

When you get dozens of property-types, and individual properties need to 
store data about themselves (e.g. their constraints, their data-types) 
which requires another out-of-band storage mechanism if you're not 
storing them in the properties themselves, the use of an external 
__getattribute__/__setattr__ mechanism starts to fall down pretty fast. 
Shrug :) .

However, it works very well for such things as the original poster was 
describing (i.e. a single (or small number of) generic thing(s) to be 
done for a large fraction of properties of an object).  You can see a 
nice example in the Zope3 persistence module.

Enjoy all,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From non at non.com  Wed Apr  2 16:40:33 2003
From: non at non.com (Iwan van der Kleyn)
Date: Wed, 02 Apr 2003 21:40:33 +0000
Subject: DVD on Mdk 9.1: VideoLan works, but what about Xine?
References: <3e8b3dbc$0$2228$e4fe514c@dreader6.news.xs4all.nl>
Message-ID: <3e8b3fad$0$2241$e4fe514c@dreader6.news.xs4all.nl>

Sorry, posted to wrong group :-(
Please ignore....

Iwan



From dkuhlman at rexx.com  Tue Apr 22 20:17:02 2003
From: dkuhlman at rexx.com (Dave Kuhlman)
Date: Tue, 22 Apr 2003 17:17:02 -0700
Subject: Python support for Amazon Web services
Message-ID: <b84m1p$65u9i$1@ID-139865.news.dfncis.de>

I've developed a small amount of Python support for the
XML-over-HTTP mode of Amazon Web services.  Amazon describes this
as their REST mode.

The support I've developed consists mostly consists of Python code
that helps to parse and process the Amazon WS XML documents.

Also, to produce this support I used generateDS.py (see:
http://www.rexx.com/~dkuhlman/#generateDS). Therefore, I needed
XML Schema documents.  So, being a Calvinist [1], I decided
implement generateXsd.py, which extracts an XML Schema from an XML
instance document.  It's included in the distribution.

There is a document describing this support at:

    http://www.rexx.com/~dkuhlman/amazon_ws_support.html

And the distribution file is at:

    http://www.rexx.com/~dkuhlman/amazon_ws_support-1.0.tar.gz

The Amazon WS developer's kit is available at:

    http://www.amazon.com/webservices

  - Dave

[1] Calvin and Hobbes, the comic strip --  Once, when Calvin's Mom
tried to force him to clean up his room, Calvin decided to invent a
robot to do the work.  After much struggle, Hobbes asked him:
"Wouldn't it be less work to just clean your room yourself?" Calvin
replied, without looking up: "It's only work if they make you do
it."

-- 
Dave Kuhlman
dkuhlman at rexx.com
http://www.rexx.com/~dkuhlman



From tim.one at comcast.net  Fri Apr  4 13:20:56 2003
From: tim.one at comcast.net (Tim Peters)
Date: Fri, 04 Apr 2003 13:20:56 -0500
Subject: My Object pool isnt working
In-Reply-To: <292c8da4.0304040145.480bcacc@posting.google.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCGEJDECAB.tim.one@comcast.net>

[sam]
> I need to pool some database connections so i thought I'd write a
> object pool. I thought I'd use a design I've used in Java before -
> having a locked and unlocked hashtable and keeping references of the
> connections in two states (i.e. allocated and unallocated). The
> problem is that the total number of connections always seems to rise.
> Also the total number of connections is less than the sum of the
> lengths of the hashtables (dictionaries) which shouldn't happen. I
> believe the problems are connected. As I said I've used this design
> before in Java and its been fine. The only difference I can think of
> (barring shoddy implementation, which I can't rule out) is that
> Hashtables in Java are thread safe, but I'm not sure if the built in
> dict type in python is.

It is.

> I've decided to use jonpy's dbpooling module
> (http://sourceforge.net/projects/jonpy) instead of writing my own now,
> but I'd still like to know whats going on. I have tested on python
> 1.52 and python 2.2 (both on win2k) and get the same result. All
> thoughts appriciated

"both on win2k" was the crucial clue:  you're indexing a dict by timestamp,
but time.time() on all flavors of Windows updates only about 18.2 times per
second:  it's extremely common on Windows for successive calls to
time.time() to return the same value.

>>> import time
>>> for i in range(10):
...     print time.time()
...
1049479681.84
1049479681.84
1049479681.84
1049479681.84
1049479681.84
1049479681.84
1049479681.84
1049479681.84
1049479681.84
1049479681.84
>>>

> Regards
>
> Sam Owen
>
> <code>
> from Threading import *

This is no module named Threading in Python (although there is a module
named "threading").  Is this the actual code you ran?  Do you have your own
module named Threading.py?

> import time
> import random
> import sys
>
> class Object:
>     def sayHello(self, caller):
>         print "hello", self, caller
>         time.sleep(int(random.random() * 5))

Note that it's not *necessary* to do the int() business here; Python's
time.sleep() can sleep for fractional seconds.

> class ObjectPool:
>     LOCKED= {}
>     UNLOCKED= {}
>     sema_connect = Semaphore(1)
>     sema_return = Semaphore(1)
>     expiretime = 60 #30 mins

The comment here doesn't make sense, of course.

>     totalConns=0
>     numAllocated = 0
>     numReturned = 0

The intent of all these assignments is unclear.  Do you intend that all
instances of ObjectPool share these values?  That would be surprising.  If
it's your intent that each instance of ObjectPool have its own set of these
variables, then these assignments should be in an ObjectPool __init__
method, a la

        def __init__(self):
            self.LOCKED = {}
            # etc

You create only one ObjectPool instance in this code, so the difference
isn't apparent.

>     def getConnection(self):
>         try:
>             self.sema_connect.acquire()
>             if len(self.UNLOCKED):

More idiomatic would be

              if self.UNLOCKED:

However, the "if" statement here doesn't appear to serve a purpose; if
UNLOCKED is empty, the following "for" won't enter the body (.keys() will
return an empty list).

>                 for ts in self.UNLOCKED.keys() :
>                     if ts + self.expiretime < time.time():
>                         print "deleting ..."
>                         self.totalConns = self.totalConns-1
>                         del self.UNLOCKED[ts]
>                     else:
>                         print "allocating ..."
>                         self.numAllocated = self.numAllocated+1
>                         o = self.UNLOCKED[ts]
>                         self.LOCKED[o]=1
>                         del self.UNLOCKED[ts]
>                         return o
>             print "creating ..."
>             self.numAllocated = self.numAllocated+1
>             self.totalConns = self.totalConns+1
>             o = Object()
>             self.LOCKED[o] = 1
>             return o
>         finally:
>             self.sema_connect.release()
>
>     def returnConnection(self, con):
>         try:
>             self.sema_return.acquire()
>             if self.LOCKED.has_key(con):
>                 print "returning"
>                 self.numReturned = self.numReturned +1
>                 self.UNLOCKED[time.time()]=con

This is where you're getting killed on Windows:  whenever time.time()
repeats a value t previously seen, you're simply losing the connection
object that *used* to be stored at UNLOCKED[t].  time.clock() has much finer
resolution on Windows than time.time() (>1MHz vs 18.2Hz), so you could use
time.clock() instead.  If you feel you need to do a portable timestamp-based
gimmick, you'll need to implement an object that passes out unique
timestamps.  Alternatively, you could use a dict mapping a timestamp to a
list of connection objects.  Simple hack:  replace the last line above with:

    t = time.time()
    # Break ties (2.2 syntax used here).
    while t in self.UNLOCKED:
        t += random.random() / 10.0  # no more than 0.1 second away
    self.UNLOCKED[t] = con

> ...





From ianb at colorstudy.com  Tue Apr 22 14:05:55 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 22 Apr 2003 13:05:55 -0500
Subject: Should I prefer an external database
In-Reply-To: <e9570f37.0304220712.37a3109d@posting.google.com>
References: <u3ckbymri.fsf@mail.afm.dtu.dk>
	 <mailman.1051001675.31208.python-list@python.org>
	 <e9570f37.0304220712.37a3109d@posting.google.com>
Message-ID: <1051034755.703.159.camel@lothlorien>

On Tue, 2003-04-22 at 10:12, Graham Fawcett wrote:
> > Now, that all said, you might want an external database with an
> > object-relational mapper.  Using SQLObject (plug: http://sqlobject.org),
> > you'd do it like:
> > 
> > from SQLObject import *
> > 
> >  [snipped code]
> 
> 
> Oh, that's very interesting, Ian. I wasn't aware of the project. 
> 
> Do you have, or are you working toward, support for transactions?

I exposes the database's transactional support.  So on Postgres, yes, on
MySQL, no (technically MySQL "maybe", but practically "no").

> I wonder, too, if you support Postgresql as a backend and if so, is
> there any plan to use its inheritance features for mapping Python
> subclasses. Just curious...

Yes, Postgresql is supported, but the inheritance features aren't.  You
could probably *use* such tables easily enough, but some of the more
advanced features might not be available, and you couldn't automatically
create such tables.  But it's not out of the question that something
like this could be supported... though SQLObject is currently fairly
lowest-common-denominator.

  Ian






From irmen at -NOSPAM-REMOVETHIS-xs4all.nl  Fri Apr 11 17:04:01 2003
From: irmen at -NOSPAM-REMOVETHIS-xs4all.nl (Irmen de Jong)
Date: Fri, 11 Apr 2003 23:04:01 +0200
Subject: distributed computing implementations
In-Reply-To: <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <d116fbae.0304100704.68ef4833@posting.google.com> <c48$3e960655$516049d2$3498@nf1.news-service.com> <3e96efac$0$49107$e4fe514c@news.xs4all.nl> <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk> <3e96f481$0$49099$e4fe514c@news.xs4all.nl> <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk> <3e970abc$0$49101$e4fe514c@news.xs4all.nl> <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>
Message-ID: <3e972e43$0$49109$e4fe514c@news.xs4all.nl>

Robin Becker wrote:

> sigh :(
> for real distributed computing I wouldn't want to hold the sockets open.

Sorry if I missed that requirement. I was just responding to Duncan's
CORBA timings...

> I just tried the original server.py/time-client.py pair again this time
> on a freeBSD system
> 
> freeBSD-->localhost 500 secs
> freeBSD-->192.168.0.3 50 secs (a local win32 machine).
> 
> 
> I guess I'm just not doing what I think I'm doing.

I really can't explain these numbers either? Localhost is 10 times slower
than over the network???

Unless your client & server both struggle for CPU cycles on the same machine 
at the same time...

--Irmen




From andrew-pythonlist at puzzling.org  Fri Apr 11 10:32:27 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sat, 12 Apr 2003 00:32:27 +1000
Subject: looking for "optimal weighting" algorithm
In-Reply-To: <b73ql8011te@enews4.newsguy.com>
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <20030411143227.GC19915@frobozz.local>

On Thu, Apr 10, 2003 at 03:13:11PM +0200, Alex Martelli wrote:
> I _know_ there's a decent algorithm to solve the following problem,
> but I don't recall its name and thus can't rapidly google for the 
> details... can somebody help?

[..snip..]

Perhaps you are after the ID3 or C5 (or is it C5.0?) algorithm?

Some googling finds this, which explains the ID3 algorithm:
    http://yoda.cis.temple.edu:8080/UGAIWWW/lectures95/learn/C45/

I'm sure I remember reading the exact same page during my AI course at Uni.
:)

-Andrew.





From ckea25d02 at sneakemail.com  Tue Apr 15 20:05:42 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 12:05:42 +1200
Subject: instance creatiion as an "event"
In-Reply-To: <P4Tma.22981$T34.691702@news2.tin.it>
References: <mailman.1050095779.7490.python-list@python.org> <pan.2003.04.11.23.20.04.65410@webone.com.au> <mailman.1050117319.5102.python-list@python.org> <b7gt8r$em0$1@panix1.panix.com> <P4Tma.22981$T34.691702@news2.tin.it>
Message-ID: <b7i63s$1869h$1@ID-169208.news.dfncis.de>

Alex Martelli wrote:
> If Member.members is coded
> as above, then you have circular reference loops involving the
> class and each instance so just about everything becomes sort of
> "immortal" -- with weak references in Member.members, no circular
> reference loops, no problems (or, am I missing something...?).

Circular references aren't a disaster these days, with
the cyclic garbage collector.

But if you want to be sure things are reclaimed as
soon as possible, weak refs are still a good idea.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From staschuk at telusplanet.net  Sat Apr  5 13:52:02 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 5 Apr 2003 11:52:02 -0700
Subject: how to test when correct output is "random"?
In-Reply-To: <20030404215312.GE7895@unpythonic.net>; from jepler@unpythonic.net on Fri, Apr 04, 2003 at 03:53:13PM -0600
References: <20030404215312.GE7895@unpythonic.net>
Message-ID: <20030405115202.C371@tibia.amotlpaa.bogus>

Quoth Jeff Epler:
  [...]
> But other parts are based on randomness.  "Alternation rules" are the
> foundation on which this is built.  An alternation rule might read
> like this:
>     S: "a" | "b" | "c";
> Then, str(S) would yield one of the alternatives.  Over many runs, it
> should yield each choice about an equal number of times.  So how do
> I test this?  Run a large number of trials and do a statistical test?
> This doesn't seem to fit in with the philosophy of software testing,
> which stresses exact repeatability.

I've encountered this problem too, trying to write unit tests for
genetic algorithms.  I haven't found an entirely satisfactory
solution, but here's my thoughts on it so far.

1. Separate the deterministic behaviour from the random behaviour.

I really like this approach, where possible.

For example, above you could have
	>>> possibilitiesFor('S') # deterministic
	['a', 'b', 'c']
	>>> expand('S') # random
	'c'
	>>> expand('S')
	'c'
	>>> expand('S')
	'a'
Testing possibilitiesFor is straightforward.  And in this case,
expand is extremely simple:
	def expand(symbol):
		return random.choice(possibilitiesFor(symbol))
This is, imho, too simple to break; no tests required.  If you
*did* write a test for it, you'd really be testing random.choice,
which is not a worthwhile activity.

2. Mock the pseudo-random number generator.

I see that others have suggested this, both in the form of
supplying your own PRNG, and in the form of calling
random.setstate.  I don't much like this approach; in any but the
simplest cases (which generally don't need to be tested), it
couples the unit test to the exact use of random numbers in the
code under test.  That is, such tests test not only that the code
is functionally correct, but also that it uses the random numbers
in a specific way.  But why would you want to test that?

Consider two implementations of this function:
	def permute(seq):
		"""Returns a copy of seq, randomly permuted."""
		# ...
One implementation builds up the result by randomly selecting,
without replacement, elements from seq, until seq is exhausted.
The other implementation picks a single random integer in
[0..2**len(seq)) and applies the permutation corresponding to that
number (in some ordering of permutations).  These implementations
are functionally equivalent, but they make very different use of
the PRNG.  If I mock the PRNG, changing from one implementation to
the other will probably break my unit tests.

3. Do statistical tests on the output.

If you're worried that your function's output is not "random
enough", then do statistical tests.  I don't think there's any way
around this, short of proving your algorithm correct; the concern
in such cases must be that the function makes such complicated use
of the PRNG that it is difficult to be sure its output has a
desirable distribution.  (If it makes simple use of the PRNG,
testing the output distribution is testing the underlying PRNG,
which is, as noted above, undesirable.)

The annoying thing about statistical tests is that they can
generate false negatives.  But if you know more about statistics
than I do, perhaps you can code to make this possibility extremely
unlikely.

4. Do sanity checks on the output.

Run the random function n times and do sanity checks on the
output.  For the permute function, for example, you could verify
that its output actually is a permutation of its input, say, by
sorting both sequences and comparing them.  For your Dadaist
program, you could verify that the output actually matches the
production.

(Under this approach, it is desirable to record the state of the
PRNG before running the test, and to report that state as part of
the assertion-failed message.  Thus the troubleshooter can
duplicate failures deterministically.)

Such tests are not strictly speaking reliable, in the sense that
they can generate false positives.  But over time, with many
successful runs, they increase your confidence in the correctness
of the program, which is the point of testing in the first place.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From sam at insidegabriel.com  Fri Apr 25 08:51:57 2003
From: sam at insidegabriel.com (Sam)
Date: 25 Apr 2003 05:51:57 -0700
Subject: boost.python - Virtual Functions with Default Implementations
References: <3c539fdc.0304241407.54dac3dc@posting.google.com> <ubryvguik.fsf@boost-consulting.com>
Message-ID: <3c539fdc.0304250451.5deafcc3@posting.google.com>

Thanks for the help.  I'll make a note of the c++-sig site for the future.
-Sam



From pan-newsreader at thomas-guettler.de  Wed Apr  9 08:42:37 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 09 Apr 2003 14:42:37 +0200
Subject: Adding item in front of a list
Message-ID: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de>

Hi!

I think it would be nice if
 l=[2, 3, 4]
 l.insert(1)

would insert in front of the list, as opposed to append() which adds new
elements at the end.

l.insert(1) looks better than l.insert(-1,1)

 thomas


-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From mlh at furu.idi.ntnu.no  Sun Apr 13 15:29:17 2003
From: mlh at furu.idi.ntnu.no (Magnus Lie Hetland)
Date: Sun, 13 Apr 2003 19:29:17 +0000 (UTC)
Subject: PEP305 csv package: from csv import csv?
References: <mailman.1049814434.25284.python-list@python.org>        <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>  <mailman.1050010580.22697.python-list@python.org>  <slrnb9jdvr.se1.mlh@furu.idi.ntnu.no>
Message-ID: <slrnb9jekd.48.mlh@furu.idi.ntnu.no>

In article <slrnb9jdvr.se1.mlh at furu.idi.ntnu.no>, Magnus Lie Hetland wrote:
>In article <mailman.1050010580.22697.python-list at python.org>, Skip
>Montanaro wrote:
[snip]
>When will it be available in the main Python cvs? (Just curious...)

Sorry -- was a bit impatient there. Just re-updated my cvs distro, and
there it was. Hooray again! :)

-- 
Magnus Lie Hetland               "Nothing shocks me. I'm a scientist." 
http://hetland.org                                   -- Indiana Jones



From lekishvili at python.qartu.com  Fri Apr  4 17:39:19 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 4 Apr 2003 14:39:19 -0800
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049477200.971.python-list@python.org>
Message-ID: <a3b5b608.0304041439.392bdb4b@posting.google.com>

What about Webware's PSP? What is the feeling?



From mwh at python.net  Thu Apr 10 13:17:04 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 10 Apr 2003 17:17:04 GMT
Subject: SYSV IPC module maintainer?
References: <mailman.1049992636.32037.python-list@python.org>
Message-ID: <7h3istmz2zz.fsf@pc150.maths.bris.ac.uk>

Jonathan Corbet <corbet-python at lwn.net> writes:

> I'm using the Python module for SYSV shared memory and semaphores; it's
> part of a twisted bit of code that lets me run Oasis banner ads on
> LWN.net.  The module is great, but, as it turns out, it's also responsible
> for a memory leak I've been chasing for quite some time.  So I tried
> dropping a note to the author (Vladimir Marangozov) with a patch, but it
> came bouncing right back to me.
> 

He posted to python-dev just today:

http://mail.python.org/pipermail/python-dev/2003-April/034465.html

You could try that email address.

Cheers,
M.

-- 
  Usenet is like a herd of performing elephants with diarrhea --
  massive, difficult to redirect, awe-inspiring, entertaining, and
  a source of mind-boggling amounts of excrement when you least 
  expect it.                                             -- spaf (1992)



From titanrebel at comcast.net  Wed Apr 30 22:26:03 2003
From: titanrebel at comcast.net (kkennedy)
Date: 30 Apr 2003 19:26:03 -0700
Subject: network programming problem??
References: <14bd1399.0304301236.7fe19be5@posting.google.com>
Message-ID: <14bd1399.0304301826.48cedeec@posting.google.com>

You guys are genius'.  DNS was the culprit.  I'm definately not a
network guru!  I wonder why it doesn't effect Windows XP the same way?
 Oh well, I'm just happy to have it fixed in Linux.

Thanks a lot!



From tim at bladerman.com  Wed Apr 16 03:36:25 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Wed, 16 Apr 2003 07:36:25 GMT
Subject: flickering in wxPython with bliting and memmoryDC
Message-ID: <ZJ7na.241$dP1.739@newsc.telia.net>

Hi
I have a wxPanel that flickers a whole lot, first I didn't give it much
consideration since I though the flickering was a common problem and would
stop as soon as I introduced a double buffer. But now, when I have tried
that, and it still flickers just as much I feel stuck.

I guess I have overlooked something and would much appreciate if someone
would have a quick look at my paint method (wich gets called everytime a
refresh is issued). Speed is not an issue, just the flickering .


def onPaintEvent(self, event):
#self.output is the wxPanel where I want my drawings
    dc = wxPaintDC(self.output)
    mdc = wxMemoryDC()
    x,y = self.output.GetSizeTuple()

#I read that these thre lines are needed
    bitmap = wxEmptyBitmap( x,y, -1 )
    mdc.SelectObject( bitmap )
    mdc.BeginDrawing()

#This loop draws my objects onto a dc
#In this case the memoryDC I would presume?
    for drawObj in self.contents:
        drawObj.draw(mdc)

#Also needed
    mdc.EndDrawing()

#Here I just blit the mdc-img over to the dc in one move
#How can it flicker then?
    dc.BeginDrawing()
    dc.Blit( 0,0,x,y, mdc, 0,0)
    dc.EndDrawing()


The panel is not large or anything and refreshes arent done at a particular
high speed, but it flickers everytime. Just like if the dc gets cleared
before the blitting but I can't se that I do that.

I have been looking at this for hours and would appreciate any pointer.

Thanks

Tim





From imbosol-1050452170 at aerojockey.com  Tue Apr 15 20:45:33 2003
From: imbosol-1050452170 at aerojockey.com (Carl Banks)
Date: Wed, 16 Apr 2003 00:45:33 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org>
Message-ID: <NI1na.16223$883.16217@nwrdny02.gnilink.net>

Dan Jacobson wrote:
> I need to use one liners every day, especially in Makefiles.  I thus
> need to know how to make any kind of python program into a big long one
> liner. However, right out of the starting gates I run into
> $ python -c 'print 2;for i in (1,4):print i'
>  File "<string>", line 1
>    print 2;for i in (1,4):print i
>              ^
> SyntaxError: invalid syntax
> 
> Help. BTW tell me what to do in all the other cases when ";" can't be
> a replacement for newline.  I.e. does python's nifty indentation rules
> spell the death of one-liners?

That can be a problem.  I can understand why it might be a pain to
litter your directory with little Python scripts.

You can work around most little problems if there is only one loop.
Stuff that comes before the loop would have to be folded into the loop
somehow.  See below.


> BTW, how do I import a module from the command line when using -c? Is
> there an option switch for importing modules?
> 
> This makes no output and returns no error value to the shell:
> $ python -c 'import math' -c 'for i in range(200,500,50): print
>  i,360/2/math.pi*math.asin(i/1238.4)'


The most common thing you want to put before a loop is import.  The
workaround is to import the module inside the loop with __import__:

python -c 'for i in range(200,500,50): print i,360/2/__import__("math").pi\
*__import__("math").asin(i/1238.4)'


Another thing you might want to do before the loop is set a variable,
like this:

    a=1; for i in range(20): a = complicated_function(a)

You can set the global inside the loop with the following trick:

python -c 'for i in range(20): globals("a").setdefault(1); \
a = complicated_function(a)'


Sometimes you want to print something before the first loop.  You can
work around this using the short-circuiting "and" operator to test
whether you're on the first iteration, along with the import trick
above and sys.stdout, like this:

python -c 'for i in range(20): i==0 and __import__("sys").stdout.\
write("Hello.\n"); do_whatever()'


You can simulate a nested loop with a list comprehension.  For
example, if you wanted to do something like this in one line:

    for i in range(10):
        for j in range(10):
            do_whatever(i,j)

You can do this:

python -c 'for (i,j) in [ (i,j) for i in range(10) for j in range(10) ]:\
do_whatever(i,j)'


What you can do in one line of Python code is surprising, if you know
all kinds of little workarounds like this.

The best thing to do might be to find a preprocessor that'll accept
Python code with braces, and feed it to Python in indented form.  Or
just litter your directories with scripts.



-- 
CARL BANKS



From mwh at python.net  Wed Apr 16 09:16:21 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 13:16:21 GMT
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com>
Message-ID: <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk>

Peter Hansen <peter at engcorp.com> writes:

> One of the most important things to learn as a programmer is this,
> as stated by various people but most popularly attributed to Knuth:
> "Premature optimization is the root of all evil in programming."

And Knuth really did say it!  In "Structured Programming with goto
Statements", on page 268 of the rather naff scan[*] I have of it, we
find:

 "We /should/ forget about small efficiencies, say about 97% of the time: 
 premature optimization is the root of all evil."

So even you have managed to misquote him, unless he (or someone else)
said it differently somewhere else.

Cheers,
M.

[*] you can get your own naff scan easily enough with google; I can't 
    remember where mine came from.

-- 
  I'm not particularly fond of singing GSTQ because she stands for
  some things I don't, but it's not really worth letting politics
  getting in the way of a good bawling.     -- Dan Sheppard, ucam.chat



From andymac at bullseye.apana.org.au  Mon Apr 21 01:33:58 2003
From: andymac at bullseye.apana.org.au (Andrew MacIntyre)
Date: Mon, 21 Apr 2003 16:33:58 +1100 (edt)
Subject: Zope and FreeBSD
In-Reply-To: <e786d63d.0304201806.4c5293a6@posting.google.com>
Message-ID: <Pine.OS2.4.44.0304211628130.25380-100000@tenring.andymac.org>

On 20 Apr 2003, Wenshan Du wrote:

> I've built a website with Zope 2.5.1 and FreeBSD 4.3, it's better than
> other solutions.
>
> The server got so slow recently that I have to reboot every 3 days,
> and Zope stop sometimes.
>
> I don't know why, is zope not stable on FreeBSD? or FreeBSD got virus?

FreeBSD 4.3 is pretty old now, and there have been quite a few security
fixes, so I'd upgrade to 4.8 first off.

You don't say anything about the size of the hardware so its hard to tell
whether you've got a scale problem, although I'd suggest not.

You might want to check the memory usage of your Zope process(es), in case
you've cornered yourself with some code that suffers gross memory
consumption (thus using swap space, slowing things down).

I vaguely recall interned strings being an issue, because they don't get
released when you expect.

--
Andrew I MacIntyre                     "These thoughts are mine alone..."
E-mail: andymac at bullseye.apana.org.au  | Snail: PO Box 370
        andymac at pcug.org.au            |        Belconnen  ACT  2616
Web:    http://www.andymac.org/        |        Australia





From darin at REMOVEblueyonder.co.uk  Tue Apr 22 14:47:34 2003
From: darin at REMOVEblueyonder.co.uk (Darin Willits)
Date: Tue, 22 Apr 2003 18:47:34 GMT
Subject: Embedding Python in an MFC application
Message-ID: <a7gpa.456$OK2.4518100@news-text.cableinet.net>

Hi,

I know that the first response to this will be "Why are you embedding
when you could just extend?" but please believe that I have my reasons
and for now this is what I have to do.  However I am something of a
newbie with this stuff so bear with me.

Now for the question:  I have created a simple MFC application (with
AppWizzard) which mearly calls Py_Initialize() and Py_Finalize() in the
OnInit function.  When I build and run this app it starts the python
interpreter fine, and shuts it down without a noticable problem. 
However when I close the application I get a huge memory leak.  Now for
the wieird part.  If I switch the above application from linking to MFC
as a shared dll, to statically linking to MFC the memory dump is gone. 
Why is this?  This is driving me nuts as I really need to use the MFC
as a shared dll.

Please tell me I am missing something simple.  I am learning this stuff
as I go and this one has me stumped.

Thanks,

Darin



From andrew at acooke.org  Thu Apr 24 05:02:09 2003
From: andrew at acooke.org (andrew cooke)
Date: Thu, 24 Apr 2003 05:02:09 -0400 (CLT)
Subject: Why functional Python matters
In-Reply-To: <Xns93675F5FC17Cduncanrcpcouk@127.0.0.1>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> 
     <1azna.115402$It5.22896@news2.central.cox.net> 
     <tyfhe8xqhmy.fsf@pcepsft001.cern.ch> 
     <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com> 
     <b86lc3$tnd$1@plover.com> 
     <ccc7084.0304231430.6b81effb@posting.google.com> 
     <mailman.1051139802.16916.python-list@python.org> 
     <Xns93675F5FC17Cduncanrcpcouk@127.0.0.1>
Message-ID: <33040.127.0.0.1.1051174929.squirrel@127.0.0.1>

so it does!  i have no idea how that happened, but thanks very much!

andrew

Duncan Booth said:

> "andrew cooke" <andrew at acooke.org> wrote in
> news:mailman.1051139802.16916.python-list at python.org:
>
>> a related question - why does queue have to be passed to helper() in the
>> code below?  i tried to define it inside walktree (but outside helper)
>> since it's global to the function, but python objects saying that the
>> local variable "queue" in helper is accessed before it's defined.  this
>> seems to be an example of where closures would work perfectly...
>
> Well, I don't know what you did but the problem was your's not Python's. I
> tried your code and it ran, I then added queue = [basepath] immediately
> before the def helper, and removed the parameter queue and the [basepath]
> from the call. I made no other changes. It still runs.
>
> Perhaps you had some other version of the code where you tried assigning
> to
> queue inside the helper function? That would give the problem you
> describe,
> but simply mutating it as the code you posted does is fine.
>
> --
> Duncan Booth                                             duncan at rcp.co.uk
> int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
> "\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?
> --
> http://mail.python.org/mailman/listinfo/python-list
>
>


-- 
http://www.acooke.org/andrew




From aleax at aleax.it  Wed Apr 16 04:26:45 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 08:26:45 GMT
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net> <420ced91.0304152121.2af84a33@posting.google.com>
Message-ID: <9t8na.27736$T34.825154@news2.tin.it>

sdhyok wrote:

> Tyler Eaves <tyler at cg1.org> wrote in message
> news:<bEGma.462$8c4.82 at tornadotest1.news.pas.earthlink.net>...
>> Nicola Mingotti wrote:
>> 
>> > import time
>> > t = time.time()
>> > # t is the number you asked for
>> > 
>> > If you want to know more about this look at
>> > the "time" module .
>> > 
>> > 
>> > bye.
>> 
>> The problem is that you can only represent dates from 1970 -
>> 2030-someting that way.
> 
> So, this is not the solution I am seeking.
> Better module to handle wider range of time?

In Python 2.3, module datetime may serve your needs:

>>> datetime.date(2303,4,16).toordinal()
840893
>>> datetime.date(1303,4,16).toordinal()
475651
>>>

Of course, dates before the Gregorian calendar's establishment
(which happened in different years in different nations...!)
will give somewhat arbitrary results, but apart from that you're
all right.  The result is an integer, but it's not hard to turn
it into a float if that's what you desire for some reason or other.


Alex




From andersjm at inbound.dk  Thu Apr  3 14:35:35 2003
From: andersjm at inbound.dk (Anders J. Munch)
Date: Thu, 3 Apr 2003 21:35:35 +0200
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com>
Message-ID: <T20ja.2512$y3.403898@news010.worldonline.dk>

"Drew Moore" <drew at astro.pas.rochester.edu> wrote:
[...]
> voltage += 5  # (raise the voltage by 5 volts)
> voltage *= 2  # (double the current voltage)
>
> but when I do
>
> voltage = 3  # (set the voltage to 3 volts)
>
> my voltage object is clobbered with an integer object.
> I don't want to require the user to type any more than
> name = value # no name.val = newval, no name.setval(value)

How about
  name(value)

- Anders






From just at xs4all.nl  Wed Apr  9 13:33:50 2003
From: just at xs4all.nl (Just)
Date: Wed, 09 Apr 2003 19:33:50 +0200
Subject: python 2.3a2 Win32 expected failures in regression tests?
References: <e2d925a8.0304090914.2fd8c154@posting.google.com>
Message-ID: <just-93CA3F.19335009042003@news1.news.xs4all.nl>

In article <e2d925a8.0304090914.2fd8c154 at posting.google.com>,
 harri.pasanen at trema.com (Harri Pasanen) wrote:

> I've built python 2.3a2 myself on Win32, using Visual C++ 6.0.
> 
> When running the regression tests, the following 3 fail:
> 
> test_tarfile, test_timeout, test_zipimport
> 
> (test_zip passes)
> All in all, I have 201 tests that pass, but those 3 fail.
> 
> I wonder if this is normal state of the affairs for Python 2.3a2 on
> win32, or something in my build that does not click.

test_zipimport started failing when DST became active... This has been 
fixed in CVS. Don't know about the others.

Just



From do-not-spam-ben.hutchings at businesswebsoftware.com  Wed Apr  9 09:38:06 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 9 Apr 2003 13:38:06 GMT
Subject: detecting modifications to an instance? (for debugging)
References: <ef08387c.0304082116.e658c31@posting.google.com>
Message-ID: <slrnb988hu.t8.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <ef08387c.0304082116.e658c31 at posting.google.com>,
Fortepianissimo wrote:
> I have a big chunk of Python code and I'd like to detect any run-time
> modification to instances of a particular class.
> 
> I'd like to know how to do it with a minimal addition of code, i.e.,
> it's much preferrable to make the code report any such instances
> instead of using a third-party debugging tool for it.
> 
> Note I just want to detect the modification, not to prevent them
> (which I think is impossible in Python).

It is possible to prevent it.

> Any suggestion is extremely welcome!

Here's how to detect it:

    def __setattr__(self, name, value):
        self.__dict__[name] = value
        print ('in %s instance 0x%x, set %s=%s'
               % (self.__class__.__name__, id(self), name, repr(value)))

    def __delattr__(self, name):
        del self.__dict__[name]
        print ('in %s instance 0x%x, deleted %s'
               % (self.__class__.__name__, id(self), name)

This assumes your class doesn't use __slots__ and doesn't have a
super-class that defines __setattr__ or __delattr__.  If either of
those is true then you should call the super-class implementation
instead of manipulating self.__dict__.

If you want to prevent modifications, you can raise an exception in
those methods, but be aware that these methods will be called even
when you set attributes in __init__, so you would need to manipulate
self.__dict__ there.



From dgemic at nospam.net  Thu Apr 17 07:36:08 2003
From: dgemic at nospam.net (Drazen Gemic)
Date: Thu, 17 Apr 2003 13:36:08 +0200
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net> <Wrena.33096$T34.876856@news2.tin.it> <b7kqr8$jsq$1@panix2.panix.com> <6e07b825.0304162200.34884d36@posting.google.com>
Message-ID: <pan.2003.04.17.13.36.08.2131.744@nospam.net>

> People have said it in the past, the Python community needs to pick ONE
> of the Python wxWindow GUI builders currently under development and
> really promote the hell out of it.

I don't think it will ever happen. You people with microsoft background
do not understand some things. In Python, PHP, Perl, Linux comunities
there are no 'winners'. People just use software they like the most.
Some use this GUI builder, other use that, other do not use GUI builder
at all, without bothering each other.
People do not want to be "guided" by some kind of superior authority
that decides what is the best for everybody.

                                             DG



From sdhyok at yahoo.com  Wed Apr 16 01:21:58 2003
From: sdhyok at yahoo.com (sdhyok)
Date: 15 Apr 2003 22:21:58 -0700
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net>
Message-ID: <420ced91.0304152121.2af84a33@posting.google.com>

Tyler Eaves <tyler at cg1.org> wrote in message news:<bEGma.462$8c4.82 at tornadotest1.news.pas.earthlink.net>...
> Nicola Mingotti wrote:
> 
> > import time
> > t = time.time()
> > # t is the number you asked for
> > 
> > If you want to know more about this look at
> > the "time" module .
> > 
> > 
> > bye.
> 
> The problem is that you can only represent dates from 1970 - 2030-someting
> that way.

So, this is not the solution I am seeking.
Better module to handle wider range of time?



From aleax at aleax.it  Wed Apr 16 11:15:02 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 15:15:02 GMT
Subject: Python for VB6 programmers (was Re: Complexity of standard Python data structures)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi> <b7ih78$fb2$1@panix1.panix.com> <Ajdna.110595$It5.16632@news2.central.cox.net>
Message-ID: <Wrena.33096$T34.876856@news2.tin.it>

Steve Holden wrote:
   ...
> You'd do better to call it "Python for VB programmers", expecially since
> the change to VB6 is so radical that migrating to a new and more sensible
> language might be an attractive option to some!

Changes from VB5 to VB6 weren't all that radical.  It was the change
from VB6 to VB.NET, aka VB7, that was the shocker (dozens of backwards
incompatible changes -- mostly taking VB's semantics closer to Python's,
btw, and therefore actually for the better, but still...).  I thought
the window of opportunity -- millions of VB6 programmers reluctant to
move to dotNET -- had more or less closed by now, but Aahz apparently
has identified that it's still open.  If that perception is correct,
then the book would indeed have quite a target-market, potentially.

Presumably, putting "VB6" in the title is meant to convey "you don't
need to understand dotNET to profit from this book".  I don't know if
there are substantial slices of programmers still stuck on VB5, VB4, ...,
and so on backwards, that might be scared off by the "6" in the title.


Alex




From mcfletch at rogers.com  Tue Apr  1 00:49:46 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Tue, 01 Apr 2003 00:49:46 -0500
Subject: SimpleParse and LookAhead
In-Reply-To: <9ee55987.0303312042.47b3283e@posting.google.com>
References: <9ee55987.0303312042.47b3283e@posting.google.com>
Message-ID: <3E89287A.60706@rogers.com>

Hi Yvan,

Here's the code I assume you intended to send as the problem-description:
8<______________
declaration = r'''line := ?([a-z], root)
root := char*
<char> := [a-zA-Z0-9-]
'''

testData = '''root - Dir
'''

from simpleparse import generator

from mx.TextTools import TextTools

parser = generator.buildParser(declaration).parserbyname('line')

taglist = TextTools.tag(testData, parser)
print taglist

for tag, beg, end, parts in taglist[1]:
   print testData[beg:end]
8<______________

Now, the problem that you are running into is that you are declaring
that the entire group "([a-z], root)" is a look ahead production.  Look
ahead productions return results exactly equal to the result that would
be returned by the regular group.  Since the [a-z] is consuming the
first character, the production "root" returns values from after that
first character to the end of the match.  There should be no consumption
of characters (i.e. the next position is position 0).

Here are the alternatives, (you haven't really explained what you're
trying to do with the look ahead, so can't come up with anything more
helpful):

   declaration = r'''line := ?([a-z], root)
   root := char*
   <char> := [a-zA-Z0-9-]
   '''
gives:
   (1, [('root', 1, 4, [])], 0)

that is, it matches, consuming zero characters.  The look ahead has
ignored the results of its first sub element and added its second
sub-element ("root")'s results to the result list.

   declaration = r'''line := ?[a-z], root
   root := char*
   <char> := [a-zA-Z0-9-]
   '''
gives:
   (1, [('root', 0, 4, [])], 4)

This version simply says that the first character of the production
"root" must be a lowercase letter. It will match the first four
characters of the test data, declaring that to be a regular match of "root".

Hope that helps,
Mike

yvan wrote:

>Hi, 
>
>I must be missing something here. I was expecting the following script
>to get me: 'root', it outputs 'oot'
>Reading the docs, my understanding what that if '?' is a prefix, then
>the parser object returns to the previous location. Is that right? If
>not, how do i get that behaviour?
>
>
>
>declaration = r'''line := ?([a-z], root)
>root:= char
>'''
>
>testdata = '''root - Dir
>'''
>from simpleparse import generator
>
>from mx.TextTools import TextTools
>
>parser = generator.buildParser(declaration).parserbyname('line')
>
>taglist = TextTools.tag(testdata, parser)
>
>for tag, beg, end, parts in taglist[1]:
>
>    print testdata[beg:end]
>
>Any help appreciated,
>
>-Yvan
>  
>

-- 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From anton at vredegoor.doge.nl  Tue Apr 29 08:32:25 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Tue, 29 Apr 2003 14:32:25 +0200
Subject: Python-2.3b1.exe install error on W98se
References: <mailman.1051538624.9122.python-list@python.org> <b8k9rp$9fn$1@news.hccnet.nl> <p0jsavcfgt346tad93i4hnnsdr90cdmfji@4ax.com>
Message-ID: <b8lrjp$b1u$1@news.hccnet.nl>

Richie Hindle <richie at entrian.com> wrote:

>> Please run this program and show me the output and I will be
>> able to complete Python-2.3b1.exe myself. 
>> [...]
>
>Here you go:

Thanks. I also got this by email, thanks to this sender too! Now I can
make all seven separate downloads of Python-2.3b1.exe result in
identical output by running the following script on them:

def main():
    data = list(file("Python-2.3b1.exe",'rb').read())
    patch = iter(list([chr(int(i)) for i in\
                       file("numbers.txt").read().split()]))
    for i in xrange(15800,16200):   data[i] = patch.next()
    file("python23b1.exe",'wb').write(''.join(data))

if __name__=='__main__':
    main()

However, agreeing with each other, the executables don't yet agree
with the Python-2.3b1.exe at python.org:

import md5

def test():
     fn5  ='Python23b1.exe'
     data = file(fn5,'rb').read()
     d = md5.new(data).hexdigest()
     print d

if __name__=='__main__':
    test()

output is:

8ecce4129dd8ca5a53aa9fa3b6b4b1b9

I guess I could split Python23b1.exe in halves and ask questions here
about the respective expected md5 hexdigests of the halves and use a
kind of bisect algorithm to locate the erroneous parts :-)

Anton

ps1. my modem is a simulation of a modem, a so called "software modem"
or winmodem, on an audio modem riser card. If I knew more details I
could write my own modem in Python. The strange thing is that it is a
very good simulation of a modem, I (almost?) never have problems with
downloading files, it seems to have problems *only* with
Python-2.3b1.exe 

ps2. I successfully downloaded Python-2.3b1.tgz and compiled it under
cygwin, so the question is, although still interesting as an exercise,
becoming a little academic: Scite, my main Python editor, can run
cygwin Python just as well as windows Python.



From piotrus at sympatico.ca  Sat Apr 19 23:25:26 2003
From: piotrus at sympatico.ca (Peter)
Date: Sat, 19 Apr 2003 23:25:26 -0400
Subject: red hat 8.0 and python
Message-ID: <3EA21326.7070100@sympatico.ca>

I've just installed Red hat 8.0, and although python 2.2.1 comes with 
it, i cannot use anything which uses the Tkinter library (this includes 
IDLE). I a few other packages work, and the interpreter runs fine. I 
suppose that this might be due to the setup, but I cannot figure it out. 
  Could it be that Tkinter module does not come with the python 
distribution which came with Red Hat 8.0? I read that it should, but I 
searched my whole computer for anything with ?tk? in the name with no luck.

I downloaded the rpm's for the python 2.2.2 (from www.python.org) but 
simple "rpm -i" or "rpm -U" installing gives me errors; conflicts with 
the currently installed version. I am not quite sure whether I should 
try uninstalling what is there now because I've read that the a few 
programs need it to run.

Any suggestions? Also if anyone knows of a webpage listing the 
step-by-step instructions on how to install/uninstall/upgrade the python 
distribution, please reply. I have read the info provided on 
www.python.org, but it is not detailed enough for me.

Thanks
Peter




From geoff at gerrietts.net  Tue Apr  1 17:12:52 2003
From: geoff at gerrietts.net (Geoff Gerrietts)
Date: Tue, 1 Apr 2003 14:12:52 -0800
Subject: How to obtain python file path
In-Reply-To: <b6d260$lg$1@news.eusc.inter.net>
References: <20030401163212631-0500@news.covad.net> <slrnb8jusc.5u1.ramen@lackingtalent.com> <20030401165546102-0500@news.covad.net> <b6d260$lg$1@news.eusc.inter.net>
Message-ID: <20030401221252.GF25283@isis.gerrietts.net>

Dan Grassi wrote:
>Not quite, that only provides the path to the initial .py file.
>
>With a directory structure:
>fa
>    a.py
>    fb
>        b.py
>
>cd .../fa
>execute a.py
>where a adds fb to sys.path
>a imports b.py
>
>I need b.py to report it's path which would be .../fa/fb

If you import fb, fb.__path__[0] will be a relative path to the
directory that produced fb.

If you import fb.b, fb.b.__file__ will be the name of the file that
was loaded to create module b (usu. a *.pyc, *.py, or *.pyo file).

You can use the sys.argv[0] and os.path tricks to find your answers
from there, I think?

Luck,
--G.

-- 
Geoff Gerrietts             "Punctuality is the virtue of the bored." 
<geoff at gerrietts net>                               --Evelyn Waugh




From spam at magnetic-ink.dk  Wed Apr 23 12:11:31 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Wed, 23 Apr 2003 16:11:31 +0000 (UTC)
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz>
Message-ID: <96bef3c9-7dd4-49e5-a9ea-13c3f71bb715@news.szn.dk>

Frantisek Fuka wrote:

> I want a simple test to see if str contains any occurences of substr.
> I am currently using this:
> 
> import string
> if string.find(str,substr) + 1:
>      ...do something...

If you're using a newer version of Python (2.2+?), you should be able
to do something like this:

#v+

>>> contains = lambda haystack, needle: haystack.find(needle) >= 0
>>> contains('foo bar fly', 'foo')
1
>>> contains('foo bar fly', 'xyz')
0
>>> 

#v-


  // Klaus

-- 
 ><> 	http://pnx.dk/



From claird at lairds.com  Mon Apr 28 15:35:05 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 28 Apr 2003 19:35:05 -0000
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <3EAC2962.2050609@soraia.com> <b8hqpk$rn3$1@panix2.panix.com>
Message-ID: <var0j9cvmnab9c@corp.supernews.com>

In article <b8hqpk$rn3$1 at panix2.panix.com>, Aahz <aahz at pythoncraft.com> wrote:
>In article <3EAC2962.2050609 at soraia.com>,
>Joe Francia  <usenet at soraia.com> wrote:
>>
>>The Deitels have a textbook called _Python How to Program_ that is aimed 
>>at beginning programmers.  I'm not sure how good/bad it is.  Three out 
>>of four reviewers on Amazon seem to like it (the fourth guy gave it one 
>>star based on its lack of Perl coverage).
>
>If you Google through the history of c.l.py, you'll find many people
>giving it a thumbs-down.
			.
			.
			.
Including at least a couple of its technical reviewers,
that is, people with at least an affective stake in its
prosperity.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From mstenner at phy.duke.edu  Wed Apr  9 15:01:07 2003
From: mstenner at phy.duke.edu (Michael Stenner)
Date: Wed, 9 Apr 2003 15:01:07 -0400
Subject: signals and system calls
In-Reply-To: <3E94690B.586BC8C8@engcorp.com>; from peter@engcorp.com on Wed, Apr 09, 2003 at 02:40:11PM -0400
References: <mailman.1049908218.16228.python-list@python.org> <3E94690B.586BC8C8@engcorp.com>
Message-ID: <20030409150106.B30646@phy.duke.edu>

On Wed, Apr 09, 2003 at 02:40:11PM -0400, Peter Hansen wrote:
> Michael Stenner wrote:
> > 
> > Just to be clear, I'm not trying to "use" signals here.  They exist,
> > they WILL be sent (when a system shuts down, for example), and I'm
> > just trying to tolerate them in a way that doesn't lose data.
> > 
> > So, the bottom line is:  what do people recommend?
> 
> Don't.
> 
> That is, don't try to be incredibly tolerant even in the face of
> system shutdown and signals that kill your program.
> 
> Instead, be tolerant of partially written data structures, half-completed
> operations, and so forth.
> 
> There is *no* way you can protect against a power outage, and no
> way you can protect against something like "kill -9" anyway, so
> you are trying to achieve an unattainable level of perfection if
> you insist that under other circumstances you *must not* ever
> lose data.
> 
> At least with this view, you'll find it much easier to deal with
> the particular nastiness of signals...  just my view.

Actually, this is really excellent advice.  I have only one minor
quibble: as stated, your argument comes off as a bit extreme.  It
seems akin to those argements that often lead security conversations
down that road to "why bother locking the case when a band of ogres
wielding crowbars and soldering irons can ....".

Just because you can't protect against a power outage or kill -9
(which are freak events in a stable and well-managed system) doesn't
mean you shouldn't _try_ to protect against _normal_ events such as
SIGTERM.  I take your advice more as "don't put all your eggs in one
basket", which is probably how you meant it anyway :)

I will take the latter part of your advice ("be tolerant of partially
written data structures...") but would still prefer to deal with the
signals if I can.

				Thanks,
				 Michael

-- 
  Michael Stenner                       Office Phone: 919-660-2513
  Duke University, Dept. of Physics       mstenner at phy.duke.edu
  Box 90305, Durham N.C. 27708-0305




From max at cNOvSisiPonAtecMh.com  Tue Apr 15 14:15:05 2003
From: max at cNOvSisiPonAtecMh.com (Max Khesin)
Date: Tue, 15 Apr 2003 18:15:05 GMT
Subject: super-problem
Message-ID: <J_Xma.4475$Up3.2766033@twister.nyc.rr.com>

I was playing with super() and here is my session - the suggested usage of
super appears to produce an error. What am I missing? thanks -
max.


PythonWin 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32.
>>> print super.__doc__
super(type) -> unbound super object
super(type, obj) -> bound super object; requires isinstance(obj, type)
super(type, type2) -> bound super object; requires issubclass(type2, type)
Typical use to call a cooperative superclass method:
class C(B):
    def meth(self, arg):
        super(C, self).meth(arg)
>>> class B:
...  def meth(self, arg):
...   print ('B::meth, arg = %s'%arg)
...
>>> class C(B):
...  def meth(self, arg):
...   super(C, self).meth(arg)
...
>>> c = C()
>>> c.meth('Ho')
Traceback (most recent call last):
  File "<interactive input>", line 1, in ?
  File "<interactive input>", line 3, in meth
TypeError: super() argument 1 must be type, not class

--
========================================
Max Khesin, software developer -
max at cNvOiSsPiAoMntech.com
[check out our image compression software at www.cvisiontech.com, JBIG2-PDF
compression @
www.cvisiontech.com/cvistapdf.html]






From glenfant at NOSPAM.bigfoot.com  Wed Apr 16 08:14:52 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Wed, 16 Apr 2003 14:14:52 +0200
Subject: Problem Cron-Job using Python-Script
References: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de>
Message-ID: <b7jgni$1q8r$1@biggoron.nerim.net>

"Andreas Penzel" <mail at andreas-penzel.de> a ?crit dans le message de news:
b7jeem$1l5i0$1 at ID-69142.news.dfncis.de...
> Hello!
>
> I have a Python-Script to create Log-Files und work with them.
> Very well.
> But if I start the Script with Cron it works not correct.
> I often use os.system() to send commands to the shell including paths from
> the os.
> Could this be the problem?
> And if, how to solve?
>
> Thanks, Andreas
>


Append this to your crontab line :

> /var/log/mycronjob.log  2>&1

Change if needed /var/log above against what's suitable to your system and
available for write to the user that runs the cron job.

Then watch this file to see the console output from your script and
tracebacks (if any).

HTH

--Gilles
>




From staschuk at telusplanet.net  Sun Apr 20 02:31:02 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 20 Apr 2003 00:31:02 -0600
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <KIioa.58844$T34.1577116@news2.tin.it>; from aleax@aleax.it on Sat, Apr 19, 2003 at 08:55:06PM +0000
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it>
Message-ID: <20030420003102.A8441@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
  [...]
> So, I tried implementing sum in the simplest way, as a
> separate patch to operator.c (leaving add alone) -- THAT
> gives me 124 microseconds, a 50% speedup with respect to
> the simple for loop and clearly sufficient to justify
> sum's existence (together with its simplicity, clarity,
> and likely frequency of use).  [...]

+1.  Optimization of the most common case.  Excellent.

I'm not sure such a sum function belongs in the operator module,
though.  At the moment I think it might fit best as a built-in; it
goes nicely with min and max.

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From matthew at newsgroups.com  Tue Apr  8 02:38:23 2003
From: matthew at newsgroups.com (matthew)
Date: Tue, 08 Apr 2003 18:38:23 +1200
Subject: passing no arguments to a callback function question
Message-ID: <b6tqig$93k$1@lust.ihug.co.nz>

Hi,

I'm pretty new to python and most of my stumblings get me somewhere but 
I'm stumped on this. Given the following code/exception how do I avoid 
the exception generated when notify_listeners is called?


     def __init__(self, file_to_process):
       	NC_Scanner.__init__(self, file_to_process, self.lexicon, 
self.event_types)
         self.add_listener('any', print_event, 'any')
         self.add_listener('position', print_params, args=None)


     def event_end(self, scanner):
         self.notify_listeners('any')
         self.notify_listeners(self.state_name) # initiates callbacks 
for   the particular event scanned in Plex
         self.begin('')

     def print_params():
     	print 'PARAMS: ', params

gives:-

   File "D:\Python22\Lib\site-packages\TeePee\notifier.py", line 39, in 
__call__
     return self.fn(*(self.args + args), **d)
TypeError: print_params() takes no arguments (1 given)

I hope this isn't too garbled. Thanks for the help. matthew.




From srijit at yahoo.com  Sat Apr  5 08:31:33 2003
From: srijit at yahoo.com (srijit at yahoo.com)
Date: 5 Apr 2003 05:31:33 -0800
Subject: Stackless python
Message-ID: <221d8dbe.0304050531.107bc603@posting.google.com>

I am interested to try Stackless Python. I have downloaded python22.dll.zip. 
How do I install this zip file? I already have standard python 2.2.2 installed.

I also do not see any documentation like tutorials, example programs, demos etc.

Is it 100% compatible with standrad Python 2.2.2?

With best regards,
Srijit



From jrfsousa.removeFF2A45B2this at esoterica.pt  Tue Apr 15 09:09:13 2003
From: jrfsousa.removeFF2A45B2this at esoterica.pt (=?iso-8859-1?Q?Jos=E9_Rui_Faustino_de_Sousa?=)
Date: Tue, 15 Apr 2003 14:09:13 +0100
Subject: book on Python for Fortran programmers
References: <3064b51d.0304150458.595afc30@posting.google.com>
Message-ID: <b7h09k$q6u1$1@ID-134488.news.dfncis.de>

beliavsky at aol.com wrote:
> started in Python?
>

There is a welt of information on Python's website:

http://www.python.org/

More directly you can try:

http://www.python.org/doc/Newbies.html

http://www.python.org/doc/Intros.html

http://www.python.org/cgi-bin/moinmoin/PythonBooks

And there is always:

news://comp.lang.python

It is pretty high traffic newsgroup (maybe too much...)

And do not forget to have fun learning Python ;-)

Best regards
Jos? Rui


--
========================================================================
Hands that help are better than lips that pray. - Robert Ingersoll.
========================================================================
iam://Jose Rui Faustino de Sousa http://homepage.esoterica.pt/~jrfsousa/
mailto://jrfsousa.removeFF2A45B2this at esoterica.pt phone://+351-239444940
address://rua Carlos A. Pinto de Abreu 30C, 1 3040-245 Coimbra Portugal
========================================================================




From adalke at mindspring.com  Sun Apr 27 15:52:36 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Sun, 27 Apr 2003 13:52:36 -0600
Subject: Another itertool function?
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no>
Message-ID: <b8hc64$6n2$1@slb9.atl.mindspring.net>

Magnus Lie Hetland:
> def weave(*iterables):
>     iterables = map(iter, iterables)
>     while iterables:
>         for it in iterables[:]:
>             try:
>                 yield it.next()
>             except StopIteration:
>                 iterables.remove(it)

Being nit-picky here.  I don't like the 'remove'.  That requires
a test for equivalence in the iterators, when there's no real
need to make that assertion.  What about the following
untested code?

def weave(*iterables):
    iterables = map(iter, iterables)
    i = 0
    while iterables:
      i = i % len(iterables)
      it = iterables[i]
      try:
        yield it.next()
      except StopIteration:
        del iterables[i]
      else:
        i = i + 1

As to inclusion in itertools, that's a different question,
to which I have no comment.

                    Andrew
                    dalke at dalkescientific.com





From dharms98 at yahoo.com  Mon Apr 28 23:17:43 2003
From: dharms98 at yahoo.com (Dan Harms)
Date: 28 Apr 2003 20:17:43 -0700
Subject: error in
References: <mailman.1050622890.4267.python-list@python.org> <1147e466.0304271906.113457ad@posting.google.com>
Message-ID: <6703f28d.0304281917.602c680@posting.google.com>

blunck at gst.com (Christopher Blunck) wrote in message news:<1147e466.0304271906.113457ad at posting.google.com>...
> Mod_Python executes your modules using restricted execution mode. 

I don't think it does. According to this, this is more likely a Python 2.2.1 bug:

http://www.modpython.org/pipermail/mod_python/2002-October/002607.html

D.

> Under restricted execution mode, a module cannot execute arbitrary
> code (which your handler module qualifies as).  PyXML tries to load
> encodings (usually UTF-8) when processing your xml message.  The
> loading of the encodings is quantified as executing arbitrary code. 
> It is promptly stopped with the message you displayed.
> 
> I'm looking into ways to either Bastion PyXML or to get mod_python to
> support PyXML.  Will post more when I learn more.
> 
> -c
> 
> Vince Buonassisi <vjb at vuit.com> wrote in message news:<mailman.1050622890.4267.python-list at python.org>...
> > hi,
> > 
> > i am getting the following error when i load my test script in my web  
> > server:
> > 
> > Mod_python error: "PythonHandler reqhandler"
> > 
> > Traceback (most recent call last):
> > 
> >    File "/usr/lib/python2.2/site-packages/mod_python/apache.py", line  
> > 181, in Dispatch
> >      module = import_module(module_name, _req)
> > 
> >    File "/usr/lib/python2.2/site-packages/mod_python/apache.py", line  
> > 335, in import_module
> >      module = imp.load_module(mname, f, p, d)
> > 
> >    File "/usr/local/apache/htdocs/reqhandler.py", line 9, in ?
> >      from xml.dom.ext.reader.Sax2 import FromXmlStream, FromXml
> > 
> >    File  
> > "/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/__init__.py",  
> > line 21, in ?
> >      encoder = codecs.lookup("utf-8")[0] # encode,decode,reader,writer
> > 
> >    File  
> > "/usr/src/build/143041-i386/install/usr/lib/python2.2/encodings/ 
> > __init__.py", line 51, in search_function
> > 
> > RuntimeError: cannot unmarshal code objects in restricted execution mode
> > 
> > the script i am running is very simple.  i just send a text message to  
> > the browser.  i am using the following versions of software:
> > 
> > apache 1.3.26
> > python 2.2.1
> > mod python 2.7.8
> > PyXML 0.7.1
> > linux redhat 8
> > 
> > any help would be appreciated.
> > 
> > thanks
> > 
> > vince
> > --



From sholden at holdenweb.com  Mon Apr 21 09:36:05 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Mon, 21 Apr 2003 13:36:05 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it>
Message-ID: <9tSoa.142846$It5.87082@news2.central.cox.net>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:KIioa.58844$T34.1577116 at news2.tin.it...
> Alex Martelli wrote:
>
> > Steven Taschuk wrote:
> >
> >> Quoth Alex Martelli:
> >>   [...]
[and ... again: sum() experiments ...]
>
> So, I tried implementing sum in the simplest way, as a
> separate patch to operator.c (leaving add alone) -- THAT
> gives me 124 microseconds, a 50% speedup with respect to
> the simple for loop and clearly sufficient to justify
> sum's existence (together with its simplicity, clarity,
> and likely frequency of use).  As I was at it, I also
> ensured that operator.sum(x) would immediately delegate
> to ''.join(x) if x's first item is a string -- this way
> it gets the same top performance as ''.join(x), avoiding
> the trap into which the "for xx in x: t+=xx" falls here
> (for x=map(str,range(999)), sum(x) clocks in at 79
> microseconds, vs 78 for ''.join(x) and 1260 for the
> simple loop with += ...).
>
> Interesting... I have my doubts on whether Guido would
> accept such a patch just a week before a beta, but maybe
> I should try submitting it anyway...
>

Only one way to resolve those doubts.

The use of sum() to concatenate a list (please, not a tuple!) of strings in
such a deterministic way is a masterstroke. It would surely deserve to
become immediately idiomatic.

looking-forward-to-the-python-dev-summari-ly y'rs -  steve
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From max at alcyone.com  Mon Apr 14 20:27:59 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 17:27:59 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl> <3E9AF8BD.BC918E5E@engcorp.com>
Message-ID: <3E9B520F.7D6E0360@alcyone.com>

Peter Hansen wrote:

> Given that Python is an "agile language", it seems very likely that
> it must be an implementation of Robert Zubrin's XP-style "Mars Direct"
> program for a manned mission to Mars.

Actually, it can manage any kind of trip:

>>> info(stationaryEarthOrbit, highTritonOrbit)
Source: stationary orbit around `Earth'
Destination: medium orbit around `Triton'
Opportunities:
 Opportunity: Earth/Neptune
  Angle: 1.97496428786 rad (113.157118384 deg)
  Period: 31756371.7964 s
Transfers:
 Transfer: out of `stationary orbit around `Earth''
  Duration: 0.0 s
  Burns: 1273.39353752 m/s
 Transfer: Hohmann from `Earth' to `Neptune'
  Duration: 965260524.876 s
  Burns: 11651.5022539, 4053.91436026 m/s
 Transfer: into `Triton'
  Duration: 0.0 s
  Burns: 1817.74724404 m/s
 Transfer: into `medium orbit around `Triton''
  Duration: 0.0 s
  Burns: 322.586540999 m/s
Course: from `stationary orbit around `Earth'' to `medium orbit around
`Triton''
Duration: 965260524.876 s
Deltavee: 19119.1439368 m/s

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Society attacks early when the individual is helpless.
\__/ B.F. Skinner
    CSBuddy / http://www.alcyone.com/pyos/csbuddy/
 A Counter-Strike server log file monitor in Python.



From tlavoie at acm.org  Tue Apr 15 10:24:34 2003
From: tlavoie at acm.org (Tim Lavoie)
Date: 15 Apr 2003 09:24:34 -0500
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1049901964.20840.clpa-moderators@python.org> <b72fo5$bq6$1@panix1.panix.com> <87ptnuxryk.fsf@theasylum.dyndns.org> <b7g2fk$g1u$1@panix3.panix.com>
Message-ID: <87smsj4z4t.fsf@theasylum.dyndns.org>

>>>>> "aahz" == aahz  <aahz at pythoncraft.com> writes:

    aahz> In article <87ptnuxryk.fsf at theasylum.dyndns.org>,
    aahz> Tim Lavoie  <tlavoie at acm.org> wrote:
    >>>>>>> "aahz" == aahz <aahz at pythoncraft.com> writes:
    >>
    aahz> So how does Capon differ from SCons?
    >>  Well, the name has funnier connotations, for one.  :)

    aahz> You deserve to be roasted for that comment.

Well, sometimes you stick your neck out... 
"Capon: Batteries included, nuts aren't."   :)

-- 
"I went to the museum where they had all the heads and arms from the
statues that are in all the other museums."
                -- Steven Wright



From peter at engcorp.com  Tue Apr  1 15:06:10 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 15:06:10 -0500
Subject: else: w/o except: - why not?
References: <mailman.1049156943.12744.python-list@python.org> <v8j9pb3v2cnra6@corp.supernews.com> <698f09f8.0304011201.1a4939a1@posting.google.com>
Message-ID: <3E89F132.9C2AD2B4@engcorp.com>

Jeremy Fincher wrote:
> 
> claird at lairds.com (Cameron Laird) wrote in message news:<v8j9pb3v2cnra6 at corp.supernews.com>...
> > And, for you, does
> >   try:
> >       f1()
> >   except:
> >       f2()
> >       raise
> > become *not* bad practice?
> 
> Wouldn't that be better written like thiS?
> 
> try:
>   f1()
> finally:
>   f2()

Doesn't finally *always* execute, while the example above
would execute f2() only if f1() raised an exception?

-Peter



From simmonds at rlmsystems.com.au  Mon Apr 28 23:27:47 2003
From: simmonds at rlmsystems.com.au (Scott Simmonds)
Date: 28 Apr 2003 20:27:47 -0700
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com>
Message-ID: <84a129e2.0304281927.a79e1ce@posting.google.com>

An option to consider would be to generate an XML file from Python,
then run Jade (the DSSSL processor) on the XML output with RTF as the
output format.
There are a few options for the choice of XML schema, Docbook
certainly has stylesheets that support RTF, and I think TEI do too
(DITA as well ?), or you could roll your own if you want to do battle
with DSSSL :)
This is also a crossplatform approach as jade will run on Unix and
Windows.

Cheers,
Scott.


achrist at easystreet.com wrote in message news:<3EA97D61.3CC9FE74 at easystreet.com>...
> Any suggestions of the shortest path from a Python program to RTF
> output?  RTF seems to be the best (reasonable combination of 
> features and compatability) common denominator for exporting
> data so that (windows) users can bring it into their own word
> processing program, whatever that may be.  (Any problem with
> that?)
> 
> Requirements are about like this:
> 
> 1. Don't need an edit control, just want to write a file, but if
> I have to open an edit window, I can.
> 
> 2. Don't want to do too much fancy.  Text, headings, bulleted
> lists,  maybe some variations in font size.
> 
> 3. Development platform is WinNT, target platform is any reasonably
> recent Windows machine (eg either anything newer than Win95 release 1
> or Win98 or better). 
> 
> TIA,
> 
> 
> Al



From SocLcrBoy2 at netscape.net  Sat Apr 19 17:57:50 2003
From: SocLcrBoy2 at netscape.net (SocLcrBoy2 at netscape.net)
Date: Sat, 19 Apr 2003 17:57:50 -0400
Subject: Tkinter help
Message-ID: <48BEA671.1EA9CE07.0CCD53D2@netscape.net>

im having trouble learning Tkinter, can anyone point me too a good guide, book, etc. 

__________________________________________________________________
Try AOL and get 1045 hours FREE for 45 days!
http://free.aol.com/tryaolfree/index.adp?375380

Get AOL Instant Messenger 5.1 for FREE! Download Now!
http://aim.aol.com/aimnew/Aim/register.adp?promo=380455




From eppstein at ics.uci.edu  Tue Apr 15 21:23:47 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Tue, 15 Apr 2003 18:23:47 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <slrnb9p3vo.vg8.ramen@lackingtalent.com> <mailman.1050454017.965.python-list@python.org>
Message-ID: <eppstein-8427F2.18234715042003@news.service.uci.edu>

On Tue, Apr 15, 2003 at 11:04:06PM -0000, Dave Benjamin wrote:
> > I think that list comprehensions bridge the gap somewhat between FP and OO
> > methods, and this is one thing I really like them for. For instance:
> > 
> >   [obj.method(x) for x in data]
> > 
> > looks nicer to me than:
> > 
> >   map(lambda x: obj.method(x), data)

Sure, but wouldn't it be even better to use
    map(obj.method, data)
?

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From cben at techunix.technion.ac.il  Sun Apr 13 16:50:15 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Sun, 13 Apr 2003 23:50:15 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <oqhe947i87.fsf@titan.progiciels-bpi.ca>
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr>
 <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com>
 <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com>
 <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk>
 <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org>
 <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <Pine.GSO.4.51.0304092053190.1222@techunix.technion.ac.il>
 <oqhe947i87.fsf@titan.progiciels-bpi.ca>
Message-ID: <Pine.GSO.4.51.0304132051160.26349@techunix.technion.ac.il>

Fran?ois Pinard wrote on 2003-04-11:

> [Beni Cherniavsky]
>
> > To clarify the issue completely, the emacs lisp manual contains a
> > must-read section:
>
> > http://www.gnu.org/manual/elisp-manual-21-2.8/html_node/elisp_147.html#SEC147
>
> > [...] It also describes some implementation details - not that I believe
> > we want dynamic scope... [...]
>
> Oh, I would not say that the above text is clarifying the issue completely.
> It rather gives a flying glimpse over the problem, within the context of a
> discussion about a project considering a rewrite of Emacs in Python.
>
You're right; I drop the "completely" ;-).  I only meant the classic
lexical/dynamic scopings, not all the variants emacs has...

> For example, the shallow binding implementation is not really acceptable for
> buffer local variables, even if most non buffer local variables use it.  Not
> to speak about the complexity associated with various defaulting or
> initialising mechanism for variables in buffers.
>
Hmm, buffer-local things are an interesting idea in emacs.  It
approximately amounts to running a separate emacs in every buffer.

> Bindings are also different, despite simultaneous, for functions and for
> non-functions.

You mean it's a "LISP2" while Python is "LISP1"?  I don't see any
architectural importance to it, it's just a matter of local convenince
w.r.t. naming things.

> Over these two, properties allow one to simulate many concurrent
> flavours of bindings to symbols.

You mean Python's descriptors/properties, or LISP's properties
attached to symbols?

> Evaluation contexts may also be fully switched through usage of
> alternate obstacks.
>
I don't really follow you here.  Care to explain?

> Now, key bindings are another kind of binding which does not work exactly
> like variables, yet the project quoted above might be wise to consider as
> well.  For example, key bindings may be associated with limited text areas
> within buffers, ...



> ... and while the cursor moves between windows, or over text
> within windows, key bindings may change quite often.  Emacs uses heaps (some
> call them priority stacks) for tracking these changes.
>
Really?  I always thought it simply searches all active keymaps.
Searching one is quick (it's a trie) and how many are there at once?
You only do keymap lookup when the user presses something so this
doesn't seem to be speed-critical.  I must miss something if emacs
bothers to optimize it - but what?

> So, all in all, in real practice, we have a flurry of different
> types and qualities of bindings with Emacs, and a variety of
> implementation techniques for them.

I think implementation techniques are secondary; first we should
choose what behavior we want.  Would be nice to get a list of all
kings of bindings/lookups.  XEmacs did some interesting generalization
of some of them - "specifiers":

http://xemacs.org/Documentation/21.5/html/lispref_45.html#SEC624

Looks to me like they are onto something but didn't carry it far
enough :-).

One type that comes to mind that is not currently good enough in emacs
is combining and switching groups of settings under user control.
Any big enough mode seems to implement something in this direction -
e.g. cc-mode's "styles", bunches of variables saved in lists given
names switchable by the user.  That kind of thing should be availiable
in any place around, perhaps as part of the custom system - but there
should be good support for having different "themes" active in
different buffers/modes/something at the same time.

What this all reminds me is the cascading in CSS.  People keep
reinventing the idea of lookup in several places in order of priority
all the time: X resources, CSS, emacs keymaps, multiple config files
(every self-respecting unix program supports at least 2: site-wide and
per-user), emacs buffer-local variables defaulting to the global
binding...

> It would not be wise to underestimate the problem, or to quickly
> dismiss this complexity as being gratuitous.
>
I completely agree.  I didn't mean to understimate it.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From opengeometry at yahoo.ca  Wed Apr  9 03:33:32 2003
From: opengeometry at yahoo.ca (William Park)
Date: 9 Apr 2003 07:33:32 GMT
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <b70icc$9n0ff$1@ID-99293.news.dfncis.de>

Jim Shady <iamshady at rediffmail.com> wrote:
> Hello,
> 
> I am a beginner at Python and I just joined this group.
> 
> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string. For
> example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?

out = []
for i in 'abcd/df/a/iiwk/abcdefghijkl/b/c'.split('/'):
    out.append( (len(i), i) )
print max(out)[1]

-- 
William Park, Open Geometry Consulting, <opengeometry at yahoo.ca>
Linux solution for data management and processing. 



From aahz at pythoncraft.com  Mon Apr 28 09:55:09 2003
From: aahz at pythoncraft.com (Aahz)
Date: 28 Apr 2003 09:55:09 -0400
Subject: Coding style and Python versions (was Re: Another itertool function?)
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no> <b8hc64$6n2$1@slb9.atl.mindspring.net>
Message-ID: <b8jbrt$41$1@panix1.panix.com>

In article <b8hc64$6n2$1 at slb9.atl.mindspring.net>,
Andrew Dalke <adalke at mindspring.com> wrote:
>
>        i = i + 1

I'm a little curious why you didn't write this as 

    i += 1

Any reason (i.e. still need to support 1.5.2) or just habit?  If the
latter, why haven't you changed your habit?
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From alloydflanagan at attbi.com  Wed Apr 16 17:30:11 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 16 Apr 2003 14:30:11 -0700
Subject: Typing arguments in python
References: <fd1c664e.0304160524.295bbdc1@posting.google.com> <e9oq9vc2lfmb50ga5aipr0unembkhfd9da@4ax.com>
Message-ID: <a8b7f07a.0304161330.1979b409@posting.google.com>

Dale Strickland-Clark <dale at riverhall.NOTHANKS.co.uk> wrote in message news:<e9oq9vc2lfmb50ga5aipr0unembkhfd9da at 4ax.com>...
> danra at walla.co.il (Danra) wrote:
> 
> >def f(fileobj):
> >    if not hasattr(fileobj,'write'):
> >        raise TypeError, "Given arg must have the 'write' attribute."
> >    .
> >    .
> >    fileobj.write('...')
> 
> You want assert, you do:
> 
> def f(fileobj):
>     assert hasattr(fileobj, 'write'), 'Bang! Sort your code out.'

Actually, I'm going to disagree.  Some people don't recommend using
assert for checking parameters, for a couple of reasons.  First, it's
probably a good idea to continue to check parameters of your functions
even in production use.  That's really a pretty basic defensive
programming practice.

The second reason is more subtle.  An assert is technically a
statement that a certain condition can't happen under any
circumstances.  It's used to verify that the code in the function has
the expected effect on the state of the program.  But being called
with an incorrect parameter is NOT a flaw in the code that contains
the assert -- it's a problem with the caller.  Basically, instead of
saying, "hey, buddy, that parameter is wrong" you're now saying
"whoops, there's a bug in my code somewhere".  For a bad parameter you
should raise ValueError or something similar, not AssertionError.

Just my two cents as usual...



From ramen at lackingtalent.com  Tue Apr 15 19:27:29 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 15 Apr 2003 23:27:29 -0000
Subject: AOP in Python (was: Post-OOP: not just an "unfortunate acronym"?)
References: <pan.2003.04.15.21.27.33.796814@mn.rr.com>
Message-ID: <slrnb9p5bk.vg8.ramen@lackingtalent.com>

In article <pan.2003.04.15.21.27.33.796814 at mn.rr.com>, Anna wrote:
> It's time to look at what might await us in the post-OOP
> era (apart from an unfortunate acronym).

Does anyone have an example of this newfangled "Aspect Oriented" paradigm's
techniques implemented in Python? The only thing I've seen in any language
(and I haven't looked at AspectJ or anything) was a casual reference in a C#
text that seemed something like function-call pre/post event-handlers. What
might Python stand to gain from this methodology?

Thanks,
Dave



From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Fri Apr 11 12:39:08 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Fri, 11 Apr 2003 18:39:08 +0200
Subject: distributed computing implementations
In-Reply-To: <c48$3e960655$516049d2$3498@nf1.news-service.com>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de> <d116fbae.0304100704.68ef4833@posting.google.com> <c48$3e960655$516049d2$3498@nf1.news-service.com>
Message-ID: <3e96efac$0$49107$e4fe514c@news.xs4all.nl>

Duncan Grisby wrote:

> Interesting article. Unfortunately, the first two timings of the CORBA
> client aren't timing what Mike thinks they are. The first time, for
> "connecting to server" doesn't actually connect to the server at all.
> It just creates an object reference for it.
> 
> The second time, for "send string" _is_ timing sending a string, but
> that is also when the TCP connection to the server is made. A
> significant portion of the time is spent setting up the connection,
> not transferring the string. If I modify the client to do the first
> call twice, I get
> 
> Connecting to server
> Time to connect to server, 0.000386
> 
> Sending a long string to the server
> Time to send a string of 21000 chars, 0.002099
> 
> Sending a long string to the server
> Time to send a string of 21000 chars, 0.001034
> 
> Recieving a long stirng from the server
> Time to receive a string of 22000 chars, 0.001088
> 
> Sending lots of ints to the server
> Time to send 5000 ints, 0.921309 (0.000184 per call)
> 


To add some more to the mix, I benchmarked Pyro (3.2) :

Connecting to server
Time to connect to server, 0.026004

Sending a long string to the server
Time to send a string of 21000 chars, 0.003536

Recieving a long stirng from the server
Time to receive a string of 22000 chars, 0.003387

Sending lots of ints to the server
Time to send 5000 ints, 10.057741 (0.002012 per call)

I find this very fast for a pure Python solution....

I also measured the message sizes with tcpdump as mentioned in the article:

Actual message size sending 1,000 characters: 1390
Actual message size sending 100 integers: 27698 (with CORBA amongst the 
smallest)


> Lies, damn lies, and statistics...

Amen.

--Irmen de Jong




From sholden at holdenweb.com  Sun Apr  6 11:06:14 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sun, 06 Apr 2003 15:06:14 GMT
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com> <3E8F8C10.A7F921CF@engcorp.com> <m6009vst7snfmo645pd9h8clbgio2dp0lp@4ax.com>
Message-ID: <GnXja.18450$It5.16446@news2.central.cox.net>

"Simon Faulkner" <news at titanic.co.uk> wrote in message
news:m6009vst7snfmo645pd9h8clbgio2dp0lp at 4ax.com...
> Thanks for all the help on this chaps,
>
> The file has changed and I am now importing 25,000 15 digit numbers!
>
> I am compiling a report as I go on each number, wether is is duped
> within this batch and then looking it up in a MySQL database.
>
> It seems to slow down as it goes and I guess that this is the overhead
> of checking each number against the dictionary?
>

Nope. Dictionary access time grows very slowly as the size of the dictionary
increases. The code that supports the dictionary structure goes to great
trouble to modify things as growth occurs to keep access efficient.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From mal at lemburg.com  Mon Apr 14 10:53:01 2003
From: mal at lemburg.com (M.-A. Lemburg)
Date: Mon, 14 Apr 2003 16:53:01 +0200
Subject: win32all odbc/MS SQLServer/bigint
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEB1@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEB1@admin56.narex.com>
Message-ID: <3E9ACB4D.7040401@lemburg.com>

Bjorn Pettersen wrote:
>>From: M.-A. Lemburg [mailto:mal at lemburg.com] 
>>
>>Sure, but __int64 is only available in VC C++ AFAIK. Many
>>compilers have a "long long" type which could be used, but
>>then again, how do you know whether the ODBC driver was
>>compiled with the same C type and layout as the application
>>using it ?
> 
> I haven't found a compiler without a long long type, but then I haven't
> checked e.g. the Palm or Cray's, so I'm assuming they're out there :-)
> This isn't a concern for the win32all odbc module of course, but all
> cross platform projects I've seen so far has a #define LONGLONG to
> either long long or __int64 somewhere (with an appropriate comment
> <wink>).

Sure, but the ODBC standard docs only refer to __int64 and
as a result ODBC drivers and managers on other platforms
have to use some other notion of "give me an integer
with at least 64 bits".

Here's an example of what unixODBC does:

#if (ODBCVER >= 0x0300)
#if (SIZEOF_LONG == 8)
#  define ODBCINT64         long
#  define UODBCINT64    unsigned long
#else
# ifdef HAVE_LONG_LONG
#  define ODBCINT64         long long
#  define UODBCINT64    unsigned long long
# else
/*
  * may fail in some cases, but what else can we do ?
  */
struct __bigint_struct
{
     int             hiword;
     unsigned int    loword;
};
struct __bigint_struct_u
{
     unsigned int    hiword;
     unsigned int    loword;
};
#  define ODBCINT64         struct __bigint_struct
#  define UODBCINT64    struct __bigint_struct_u
# endif
#endif
#ifdef ODBCINT64
typedef ODBCINT64       SQLBIGINT;
#endif
#ifdef UODBCINT64
typedef UODBCINT64      SQLUBIGINT;
#endif
#endif

This is what the SAP DB ODBC driver uses:

/* the ODBC C types for SQL_C_SBIGINT and SQL_C_UBIGINT */
#if (ODBCVER >= 0x0300)
#if (_MSC_VER >= 900)
#define ODBCINT64       __int64
#endif

#if defined(_WIN32) || defined(_WIN64)
#else
#   if defined (BIT64)
#      define   ODBCINT64   long      /* integer (8 byte, signed)   */
#   else
#      define   ODBCINT64   long long     /* integer (8 byte, signed)   */
#   endif
#endif /* UNIX(tm) tested on AIX,DEC,LINUX,HP,SNI and SUN ;-) from sapdb_types.h */

/* If using other compilers, define ODBCINT64 to the
         approriate 64 bit integer type */
#ifdef ODBCINT64
typedef ODBCINT64       SQLBIGINT;
typedef unsigned ODBCINT64      SQLUBIGINT;
#endif
#endif  /* ODBCVER >= 0x0300 */

The introduction of Unicode has similar
problems. Unicode on Windows is actually UCS2 (two bytes)
or UTF-16 (depending on the app's view). Now, on other
platforms such as Linux wchar_t has 4 bytes. As a result,
some ODBC drivers use 4 bytes per character on Linux while
other use 2 bytes.

> If you call
> 
>   SQLColAttribute(hstmt, n, SQL_DESC_TYPE, NULL, 0, NULL,
> (SQLPOINTER)&sqlType);
> 
> and sqlType == SQL_BIGINT, that would be a good indication for result
> sets, and if

Right. That's what mxODBC uses.

>   SQLBindParameter(hstmt, num, SQL_PARAM_INPUT,
>                    SQL_C_SBIGINT, SQL_BIGINT, 
>                    sizeof(__int64), 0, 
>                    (void*)val, sizeof(__int64), &m_nullInd);
> 
> returns either SQL_SUCCESS or SQL_SUCCESS_WITH_INFO that ought to be
> good enough for parameters?

Yes, but only if __int64 is available. I suppose that mxODBC
could use __int64 on Windows without any problems...

>>It's one of those MS things again... the half cooked, "works
>>for me" kind of attitude.
> 
> 
> And here I thought it was the "everyone else did it that way, so we'll
> do it different just because we can" attitude <wink>.

The "everyone else" part doesn't fit: MS defines what the ODBC
standard is, not everybody else :-/

-- 
Marc-Andre Lemburg
eGenix.com

Professional Python Software directly from the Source  (#1, Apr 14 2003)
 >>> Python/Zope Products & Consulting ...         http://www.egenix.com/
 >>> mxODBC, mxDateTime, mxTextTools ...        http://python.egenix.com/
________________________________________________________________________
EuroPython 2003, Charleroi, Belgium:                        71 days left





From zyziu at interia.pl  Wed Apr 16 17:13:54 2003
From: zyziu at interia.pl (zyziu)
Date: Wed, 16 Apr 2003 23:13:54 +0200
Subject: scrollbar
Message-ID: <b7kh3g$51u$1@nemesis.news.tpi.pl>

Hello. I made lanChat in Python. I have problem with Text widget. I want to
text from Entry goes to Text widget and scrollbar goes down, because I can't
see what write another user. Now I must do it manually. How do it to
scrollbar goes down automatically? Thanks for answer. Zyziu from Poland.





From aleaxit at yahoo.com  Fri Apr 11 12:34:46 2003
From: aleaxit at yahoo.com (Alex Martelli)
Date: Fri, 11 Apr 2003 18:34:46 +0200
Subject: looking for "optimal weighting" algorithm
In-Reply-To: <20030411143227.GC19915@frobozz.local>
References: <b73ql8011te@enews4.newsguy.com> <20030411143227.GC19915@frobozz.local>
Message-ID: <200304111834.46888.aleaxit@yahoo.com>

On Friday 11 April 2003 04:32 pm, Andrew Bennetts wrote:
> On Thu, Apr 10, 2003 at 03:13:11PM +0200, Alex Martelli wrote:
> > I _know_ there's a decent algorithm to solve the following problem,
> > but I don't recall its name and thus can't rapidly google for the
> > details... can somebody help?
>
> [..snip..]
>
> Perhaps you are after the ID3 or C5 (or is it C5.0?) algorithm?
>
> Some googling finds this, which explains the ID3 algorithm:
>     http://yoda.cis.temple.edu:8080/UGAIWWW/lectures95/learn/C45/

Thanks!  Actually what i had in mind (and was unable to remember
the name of) was "logistic regression", but it's nice to see several
cognate approaches pointed out -- tx!


Alex





From mhammond at skippinet.com.au  Tue Apr 29 18:55:47 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Tue, 29 Apr 2003 22:55:47 GMT
Subject: Installer Problem WinNT to Win 98
In-Reply-To: <3EADE94B.379C63A0@easystreet.com>
References: <3EADD5C2.ABC48D5E@easystreet.com> <Xns936C27AC56317JanC@213.224.134.81> <3EADE94B.379C63A0@easystreet.com>
Message-ID: <TpDra.7013$lD4.74951@news-server.bigpond.net.au>

achrist at easystreet.com wrote:

> JanC wrote:
> 
>>achrist at easystreet.com schreef:
>>
>>Did you use the Unicode version of wxPython?  
> 
> 
> Thanks, but no.  I've got wxPythonWIN32-2.4.0.7-Py22.exe, not
> wxPythonWIN32-2.4.0.7u-Py22.exe.  
> 
> 
> I've just spent some time looking inside the various pyd files that
> the McMillan installer did package with my app to see what else I might
> need.  I found:
> 
> In almost all of them:  MSVCRT.DLL.  Is there any way that a pure
> vanilla Win 98 machine (very little installed software, nostly used
> to run what came with the computer) might be missing this file?  Is it
> legit for me to redistribute it with a python app (I do have an MSVC
> license)?  

If I understand things, that is possible (but unlikely)

> 
> Also:
> 	
> 	ADVAPI32.dll
> 	Shell32.DLL
> 	OLEAUT32.DLL
> 	Kernel32.DLL
> 	User32.DLL

That is impossible.

You really need a way to see the exception.  This could be as simple as 
hacking your app to open a file and set sys.stderr and sys.stdout to it. 
  Rebuild the package, and try again on the same machine, this time 
looking at the file.

Or simply install Python on that machine and run from source code, just 
to determine the error.

Mark.




From reply.in at the.newsgroup  Wed Apr  2 12:41:02 2003
From: reply.in at the.newsgroup (Rene Pijlman)
Date: Wed, 02 Apr 2003 19:41:02 +0200
Subject: Socket timeouts for HTTP connections (httplib)?
References: <9df3d134.0304020550.7e32bf6b@posting.google.com>
Message-ID: <s18m8v4tkt4jgp02ntu85qe886g86j93oe@4ax.com>

nick:
>There were rumblings about Guido implementing this at a socket level
>in June 2002, but I've not been able to find anything since. 

It's in 2.3:
http://www.python.org/2.3/highlights.html

But I don't know if and how this ripples up to higher level
APIs.

-- 
Ren? Pijlman



From jane.doe at acme.com  Fri Apr  4 14:32:20 2003
From: jane.doe at acme.com (Jane Doe)
Date: Fri, 04 Apr 2003 21:32:20 +0200
Subject: [Newbie] Opinions on GUI and distribution
Message-ID: <17nr8vgmif2df7dr8dd5saut781b1fr044@4ax.com>

Hi,

	I'd like to check if Python is a viable alternative to
VisualBasic to develop stand-alone apps for Windows, and I did read
the archives and stuff on the Web about those issues but... I'd like
some feedback from people experienced with this kind of developement.

1. How easy is it to build and deploy Python apps on hosts that either
have nothing pre-installed (ei. Python itself, the GUI framework, and
my scripts), or need some updates (ie. my scripts in a newer version)?

2. Are the GUI frameworks currently available (win32all as part of
ActivePython, wxWindows, PyGTK, PythonQT, etc.) really good enough to
build professional applications?

3. What about performance as compared to VB, VC++ or PowerBasic?

Thx much for any infos
JD.



From ramen at lackingtalent.com  Thu Apr 24 03:04:41 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Thu, 24 Apr 2003 07:04:41 -0000
Subject: Why functional Python matters
References: <mailman.1051154381.32098.python-list@python.org>
Message-ID: <slrnbaf35b.ebi.ramen@lackingtalent.com>

In article <mailman.1051154381.32098.python-list at python.org>, Tim Peters wrote:
> (Python statements are not Python expressions, so no
> matter how perverse you try to be, you can't write a Python lambda that
> contains a Python statement -- what you can do inside a Python lambda is use
> a crippled subset of the language).

This is what I was talking about in my original post. I don't believe that
the Python lambda offers a crippled subset of the language any more than I
believe that a constant offers a crippled subset of the functionality of a
variable. Lambdas are expressions, for use with other expressions. Defs
are statements, for use with other statements. I don't find this crippled or
redundant.

Perhaps this is an over-generalization, but I think that nestability
encourages readability. The nice thing about lambdas in Python are that they
enforce this property.

> That said, Guido regrets lambdas anyway <wink>.

He may one day live to regret regretting lambdas. =) In the meantime, all
I'm saying is that I really like Python the way it is. Why ruin a good thing?

One last thing and I'm off of my soapbox. ;) I've been thinking more about
the parallel between list comprehensions and relational calculus vs.
map/filter functions and relational algebra. I mentioned this briefly before
but I want to bring it up again because nobody commented. I'm totally open
to criticism. Nonetheless, did anybody notice that filter+lambda can express
a "select" operation in relational algebra using almost identical syntax:

r.a.: select [age > 20] (people)
python: filter(lambda x: x.age > 20, people)

Now, most of the time, perhaps we use a form of relational calculus (SQL) to
do this type of thing anyway. But relational algebra is still around. It's
still a useful concept. Why?

 1. It's very regular
 2. It's easy to parse
 3. It's very portable across languages
 
I can port my map/filter code to PHP, for instance, or to LISP, or even to
ActionScript (with appropriate language hacks). List comprehensions, on the
other hand, are proprietary Python (or Haskell) extensions from a
meta-linguistic perspective. I love them to death, still, but just think
about it.

Okay, I'm off to meta-land. 'Night, y'all. =)
Peac,
Dave



From phil at river-bank.demon.co.uk  Tue Apr  1 13:50:57 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Tue, 1 Apr 2003 18:50:57 +0000
Subject: PyQt QTableItem inheritance...
In-Reply-To: <x77kaejf9b.fsf@guru.mired.org>
References: <x77kaejf9b.fsf@guru.mired.org>
Message-ID: <200304011850.57083.phil@river-bank.demon.co.uk>

On Tuesday 01 April 2003 5:31 pm, Mike Meyer wrote:
> I've already tried posting this to the pyKDE mail list, and gotten no
> response.
>
> I'm trying to create an item in a QTable that uses the
> QDoubleValidator on the input edit field. The only way I can see to do
> this is subclass QTableItem, and override createEditor with a method
> that sets the validator, like so:
>
> class Float_Field(QTableItem):
>
>     def __init__(self, table, data):
>         print "Creating float field."
>         QTableItem.__init__(self, table, QTableItem.OnTyping, str(data))
>
>     def createEditor(self):
>         print "createEditor"
>         editor = QLineEdit(self.text(), self.table().viewport())
>         self.connect(editor, SIGNAL("textChanged(QString &)"),
>                      self.table(), SLOT("doValueChanged"))
>         editor.setValidator(QDoubleValidator(self))
>         return editor
>
>
>
> While I get the output from __init__, I never see the output from
> createEditor. Any help would be appreciated.

A small but complete example would help.

Phil




From usenet_spam at janc.invalid  Tue Apr 29 13:44:15 2003
From: usenet_spam at janc.invalid (JanC)
Date: Tue, 29 Apr 2003 17:44:15 GMT
Subject: Installer Problem WinNT to Win 98
References: <3EADD5C2.ABC48D5E@easystreet.com> <Xns936C27AC56317JanC@213.224.134.81> <3EADE94B.379C63A0@easystreet.com>
Message-ID: <Xns936CC8BDB1288JanC@213.224.134.81>

achrist at easystreet.com schreef:

> In almost all of them:  MSVCRT.DLL.  Is there any way that a pure
> vanilla Win 98 machine (very little installed software, nostly used
> to run what came with the computer) might be missing this file?

My Win98 installation is 3 years old, so not really "pure vanilla" 
anymore...  :-)

>      ADVAPI32.dll
>      Shell32.DLL
>      OLEAUT32.DLL
>      Kernel32.DLL
>      User32.DLL

AFAIK these are all standard Windows system DLLs, I think Windows won't 
boot without them.

-- 
JanC

"Be strict when sending and tolerant when receiving."
RFC 1958 - Architectural Principles of the Internet - section 3.9



From exarkun at intarweb.us  Fri Apr 11 12:31:27 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 11 Apr 2003 12:31:27 -0400
Subject: a module who knows its caller
In-Reply-To: <2259b0e2.0304110812.4eea37de@posting.google.com>
References: <2259b0e2.0304110812.4eea37de@posting.google.com>
Message-ID: <20030411163127.GA31416@meson.dyndns.org>

On Fri, Apr 11, 2003 at 09:12:35AM -0700, Michele Simionato wrote:
> I have a module (say m.py) that can be called by different scripts (say a.py 
> and b.py). I want m.py to be able to tell if it has been called by a.py or
> by b.py.

  Why?

> 
> What I do now is to pass the global name __file__ to a function m.beencalledby:
> something like
> 
> # a.py
> import m
> m.beencalledby(__file__)  # __file__ is a.py
> 
> 
> # b.py
> import m
> m.beencalledby(__file__)  # __file__ is b.py
> 
> # m.py
> 
> def beencalledby(fname): print 'I have been called by',fname
> 
> it works, but I wonder if there is a more elegant solution.
> 

  The inspect module can certainly do this with "magically", but I would
recommend against this kind of thing for anything other than getting extra
information for debugging (and even then, I would recommend pdb first).

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 22 days, 13:01, 5 users, load average: 0.99, 0.98, 0.99




From pwatson at redlinec.com  Fri Apr 11 14:24:02 2003
From: pwatson at redlinec.com (Paul Watson)
Date: Fri, 11 Apr 2003 13:24:02 -0500
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <v9e21mhemm8c2@corp.supernews.com>

Were you required to learn how to refine crude oil into petrol before you
were taught how to drive a car?

Surely you must have studied the basic physics of fluid mechanics before you
could be successful in the WC.

We need to teach students correct design priciples to build something
greater than what already exists.  We will never get very far if we require
everyone to start with a quark or atom.  Yes, of course we need some people
who design silicon and create microcode.  They will learn the low-level
details what they need to know as they need it.  Knowing great design and
organization principles will enable them to make the most of it.


"Harald Massa" <cpl.19.ghum at spamgourmet.com> wrote in message
news:Xns935A6D886AC5cpl19ghumspamgourmet at 62.153.159.134...
> > [cut]
> > I think python is a very good language for education. You could get in
> > contact with local schools.
>
> Thomas, I am not so sure about this.
>
> Yes, I love Python, and it is the most elegant programming language I
> ever met.
>
> Python gives you objectoriented, functionorientet and imperative
> programming paradigm in one fitting package.
>
> Python does not bother beginners with a million typecasts. Very great.
>
> For education there are small drawbacks:
>
> - 50% of all exercises "to teach programming" are already solved in the
> standard-library or are a method of a buildin object / are allready build
> in the core language
>
> - 45% of these exercises are solved within "the python cookbook",
> available also online
>
> - if you learn programming with Python and later have to use Java or
> Visual Basic, it will be the most frustrating experience for the young
> fellows.
>
>
> Anyway, I regret not having met Python earlier in my programming career.
>





From ianb at colorstudy.com  Sun Apr  6 23:58:06 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 06 Apr 2003 22:58:06 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <yfsr88fkppr.fsf@black132.ex.ac.uk>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	 <6ee58e07.0304061144.35c5ab92@posting.google.com>
	 <m2k7e78j5s.fsf@localhost.localdomain>
	 <mailman.1049662570.29853.python-list@python.org>
	 <yfsr88fkppr.fsf@black132.ex.ac.uk>
Message-ID: <1049687886.660.208.camel@lothlorien>

On Sun, 2003-04-06 at 21:26, Alexander Schmolck wrote:
> He certainly knows about Emacs (have a look at the index of his book), but I
> guess obtaining patents and attracting attention as some sort of visionary is
> unnecessarily complicated by pointing out the prior art yourself. 

I'm not sure if that's better or worse than ignorance...

> And although
> the most crucial (and brilliant) idea, incremental search, is nicked (and
> uncredited on his webpage), the quasi-modal character (i.e. that you have to
> keep the search key pressed) of LEAP (tm, patented and what have you) seems
> like an actual difference -- you don't have to press return/C-g/movement
> commands to get out of it, which I would guess (I haven't tried it -- no mac)
> makes it more natural for moving in text (it also places the cursor at the
> beginning of the match, I think, which seems similarly useful for moving
> around).

I did try THE out, though only a little.  I seriously doubt that the use
of shift would become natural, even after more heavy use -- it binds
your hand in a certain position, which makes natural typing difficult. 
You can't even switch shift keys.  But it might be reasonable if you
were holding down a key with your thumb... I think he mentions that,
though most keyboards don't have an available thumb-accessible key, so
he settles for shift.

By his technique for analyzing simplicity of keyboard shortcuts, the
shift technique isn't really better than C-S.  They both involve three
key hits (Shift-Space-period vs. Control-S ... Enter).  And the C-S form
doesn't involve hitting many keys simultaneously.  The quasi-modal stuff
is an idea, but not a successful one.  Maybe it's meant to handle the
multi-key sequences that Emacs has.  Doing LEAP AGAIN is also
particularly painful.  Common commands have to be easy to type.

Given a novel form of input device, perhaps there's something to what
he's trying to do.  I'm skeptical, though.

> > Now, I'm not saying Emacs couldn't be improved.  But it's a far, far
> > better starting point that "modern" text editors.  I'm thoroughly
> > convinced that modality is the superior interface for most situations,
> > and I think GUIs are leaning more in that direction as time goes on.  I
> > think this is why vi remains popular -- even though I personally don't
> > like it -- because it's modal.  At least anything that's keyboard
> > friendly should be modal, and a text editor that isn't keyboard friendly
> > is clearly stupid.
> 
> Huh?

Which part?  Obviously text editors need to be keyboard friendly.  I
believe to be friendly you have to be modal, because manipulating focus
with a keyboard is awkward and error-prone.  There might be better ways
to display your context, but displaying context isn't the same as a
non-modal interface.

> Squeak also happens to sport an UI element whose usefulness seems incredibly
> obvious to me but which I haven't seen anywhere else (exchanging contents of
> current and prior selection).

That does sound like a good idea.  On the topic of funny little UI
features, Oberon had neat scrollbars, kind of like Xaw scrollbars, only
when you did a left click, the line parallel was placed at the top of
the screen.  Right click and it was placed at the bottom.  It was great
for reading text, because you'd know exactly where to pick up on the
next page (since you were actually indicating exactly how the next page
should display).  While you can't do it in the scrollbars, I guess
that's easy enough to do as a keyboard shortcut in Emacs... I'll have to
add that shortcut and try to start using it...

Anyway, I'm sure there's a bunch of interesting UI ideas out there, most
in now-dead projects.  There must be a graveyard of user interfaces out
there somewhere... 

  Ian






From roy at panix.com  Sat Apr  5 08:40:13 2003
From: roy at panix.com (Roy Smith)
Date: Sat, 05 Apr 2003 08:40:13 -0500
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com> <b6l0eq$tf0$1@main.gmane.org> <mailman.1049539029.20263.python-list@python.org>
Message-ID: <roy-C473C1.08401305042003@reader2.panix.com>

> MR> - wrap small C libs with Pyrex,
> MR> - wrap big C libs with SWIG,
> MR> - wrap small C++ libs with boost.python,
> MR> - wrap big C++ with (I still have to do that ;)
> 
> pyste = (python + gcc-xml + boost.python)

Plus I saw a recent reference to something called SIP?

There was a recent item on slashdot lamenting how much effort is wasted 
in the free software community on reinventing the wheel, and how much 
better off we would be if there was less NIH going on and more people 
cooperating on fewer projects.  Do we really need a half-dozen ways to 
make Python talk to C/C++?

Other obvious examples include the plethora of linux distros, Gnome vs. 
KDE, Open/Free/Net-BSD.



From mrsinger at usa.net  Tue Apr  1 15:17:06 2003
From: mrsinger at usa.net (Matt Singer)
Date: 1 Apr 2003 12:17:06 -0800
Subject: list index question
References: <e62b26d3.0304010613.39f98991@posting.google.com> <pan.2003.04.01.15.38.54.181811@mn.rr.com> <pan.2003.04.01.15.42.18.291186@mn.rr.com>
Message-ID: <e62b26d3.0304011217.7f41254f@posting.google.com>

Can I wildcard? I don't know the values in the second column.  I want
to find the first entry what has a particular value in the first
column.


"Anna" <revanna at mn.rr.com> wrote in message news:<pan.2003.04.01.15.42.18.291186 at mn.rr.com>...
> On Tue, 01 Apr 2003 09:38:54 +0000, Anna wrote:
> 
> > 
> >>>> list1=[(100,0), (10,5), (100,10), (100,0)] print list1
> 
> Have I mentioned that I *hate* it when my mailreader does this to a simple
> copy/paste from IDLE... 
> 
> That print statement should, of course, have been on a separate line...
> 
> Ugh.
> Anna



From wester at ilt.fhg.de  Fri Apr  4 06:34:37 2003
From: wester at ilt.fhg.de (Rolf Wester)
Date: Fri, 04 Apr 2003 13:34:37 +0200
Subject: Severe c-module problem
In-Reply-To: <b6jo6v$715$1@nets3.rz.RWTH-Aachen.DE>
References: <b6jo6v$715$1@nets3.rz.RWTH-Aachen.DE>
Message-ID: <b6jql1$937$1@nets3.rz.RWTH-Aachen.DE>

Thanks a lot. I copied the code without reading the documentation (and 
without thinking I gues). That really helped.

Rolf




From rudy.schockaert at pandora.be  Sun Apr 20 11:03:34 2003
From: rudy.schockaert at pandora.be (Rudy Schockaert)
Date: Sun, 20 Apr 2003 15:03:34 GMT
Subject: telnet
References: <hryoa.600205$sj7.23974187@Flipper>
Message-ID: <aFyoa.53976$t_2.5005@afrodite.telenet-ops.be>

> how can I assign a port with code like this?
>
> import telnetlib
>
> HOST = "localhost"
>
> tn = telnetlib.Telnet(HOST)
>
> print tn.read_all()
>

import telnetlib
HOST = "localhost"
PORT = 25
tn = telnetlib.Telnet(HOST,PORT)





From martin at v.loewis.de  Sun Apr 27 13:37:44 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 27 Apr 2003 19:37:44 +0200
Subject: Python 2.3b1 + Tkinter on Solaris 2.6
References: <mailman.1051453836.28998.python-list@python.org>
Message-ID: <m3fzo3lu3b.fsf@mira.informatik.hu-berlin.de>

Jeff Bauer <jbauer at rubic.com> writes:

> Thanks for your response.  I removed every occurrence of
> the tcl/tk libraries from my system, then rebuilt tcl/tk
> from source.  I then rebuilt Python2.3b1 from scratch, but
> the same error occurs.  :-(

So what compiler version are you using? Where do you get libgcc from?
Does it have __eprintf? Is that linked into libtcl? Into _tkinter?
Into python? etc.etc.

Martin




From andersjm at dancontrol.dk  Tue Apr  8 11:08:16 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Tue, 8 Apr 2003 17:08:16 +0200
Subject: Making python module using Borland C++ 5
References: <b6u7fp$d2c$1@news.cc.tut.fi> <3e92b527$0$32016$edfadb0f@dread12.news.tele.dk> <b6uhsn$kj9$1@news.cc.tut.fi>
Message-ID: <3e92e5c0$0$10001$edfadb0f@dread11.news.tele.dk>

"Rami Saarinen" <rs55862 at NOSPAMyahoo.com> wrote:
> However, the thing is that eventually I'd be using a custom python 2.1 
> that has its' own python21.lib and includes. I am pretty sure that there 
>   is no distutils for that installation so I need to learn to do it by 
> hand.

Perhaps you can just statically link the whole thing, using the
embedding procedure, and calling initpyclips directly right after
Py_Initialize.

- Anders





From peter at engcorp.com  Tue Apr  8 19:29:57 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 08 Apr 2003 19:29:57 -0400
Subject: How to split string in equally long parts
References: <b6vd0e$i1b$01$1@news.t-online.com>
Message-ID: <3E935B75.64968B9@engcorp.com>

Heiko Henkelmann wrote:
> 
> I need to split a string into an array of equally long substrings, like
> e.g.:
> 
> '1234567890'   -> ['12', '34','56','78','90']
> 
> What is the most efficient way to do this?

Define "efficient" ...

And see also http://groups.google.ca/groups?th=7848e47f9918f3f7

-Peter



From mcfletch at rogers.com  Sun Apr  6 11:56:49 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 06 Apr 2003 11:56:49 -0400
Subject: properties and get/set methods
In-Reply-To: <20030406062457.GB25203@meson.dyndns.org>
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <3E8FB1A9.20507@rogers.com> <20030406062457.GB25203@meson.dyndns.org>
Message-ID: <3E904E41.8080408@rogers.com>

Jp Calderone wrote:

>On Sat, Apr 05, 2003 at 11:48:41PM -0500, Mike C. Fletcher wrote:
>  
>
>>Dan Bishop wrote:
>>    
>>
... heavily edited, there's more byplay here...

>>>Why not just use two different getvar and setvar functions?
>>>      
>>>
... heavily edited, there's more byplay here...

>  Indeed not.  So move things around a little:
>
... neat embed-with-namespaces hack snipped ...

>  Logically, the two layers still exist (good thing), but the first is
>resolved entirely at class-creation time, nicely negating the efficiency
>concerns (good thing).
>
Very nice.  Don't think I'll actually adopt it directly (I prefer a 
classical OO derived-class approach), but definitely a neat approach.  
In the end, it would push all the data (e.g. type declarations, watching 
mechanisms) down into the functions, which should make for fast 
operation, but would likely make the normal introspection mechanisms 
take a fit (well, actually just "do nothing") when trying to describe 
the properties.  I'd looked at define-time-binding the methods of the 
property objects, but it wasn't a noticable win compared to the work the 
methods themselves were doing.

Interesting though, especially for the original poster, where there's 
only a single attribute type ("generic"), so introspection won't matter 
much.

Enjoy,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From peter at engcorp.com  Thu Apr 10 08:50:19 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 10 Apr 2003 08:50:19 -0400
Subject: sorting strings written to files
References: <2745895.1049870828@dbforums.com> <2749980.1049939226@dbforums.com>
Message-ID: <3E95688B.444729B3@engcorp.com>

onefatman wrote:
> 
> They're strings separated by \n so list.sort doesn't work on them. 

??  Since when can't you sort a string with a \n at the end?



From max at alcyone.com  Tue Apr 15 00:36:27 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 21:36:27 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl> <3E9AF8BD.BC918E5E@engcorp.com> <3E9B520F.7D6E0360@alcyone.com> <b7fltn$b7v$1@slb6.atl.mindspring.net> <3E9B6821.AACF1ADD@alcyone.com> <b7g0ps$hi7$1@slb3.atl.mindspring.net>
Message-ID: <3E9B8C4B.25A1AF0F@alcyone.com>

Andrew Dalke wrote:

> I hadn't heard of a bielliptic transfer before, but then again, nearly
> everything I know about this topic comes from reading Heinlein.  ;)
> (Okay, and reading magazines like Nature and various articles
> about space flight.)

I wouldn't worry about it too much, bielliptic transfers are an obscure
element of orbital dynamics which are very rarely used.

> Still, a gravitational assist can get things out faster and with less
> fuel
> than either of those two.  But they are non-trivial to compute, and
> Heinlein sure hadn't heard of them, given especially the constraint
> of having to work it all out on paper and sliderule.

Direct propulsive capability (which will be much more easily available
in the era when the game opens) will be a much better way to get from
place to place, either because you want to make a transfer off nominal
time (i.e., I want to take a Hohmann-like transfer from Earth to Mars
but I don't want to wait for the right opportunity), or just to use a
dead-run thrust from point A to point B.  When you have advanced fusion
or even antimatter drives, your exhaust speed is high enough that you
get some serious bang for your buck (momentum divided by reaction mass,
or thrust divided by fuel consumption rate).

Gravitational assists are certainly a possibility, but they sort of get
in the way of the approximations of the model (namely that the objects
in question are far too small to "accidentally" bump into unless you are
deliberately making a transfer away from or toward one of them). 
Furthermore, favorable arrangements for gravitational slingshots are
much rarer than Hohmann transfer opportunities (to make a Hohmann
transfer, you have to wait for the right alignment between your source
and destination points -- expressed as a relative central angle between
them).  So for the purposes of the model I've decided to ignore them. 
Another possibility, straddling the difference between letting gravity
doing the work and direct propulsive capability, would be to allow
solar/magnetic sails; these take a considerably long time, but
essentially require zero deltavee (you're letting the Sun's photon
flux/solar wind and gravity do the work for you).  Again, if you really
want to go from here to there quickly, it's a lousy solution, but if
you're just ferrying bulk materials as part of a trading network it
doesn't matter how long at takes as long as the "train" keeps going.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ They have rights who dare defend them.
\__/ Roger Baldwin
    EmPy / http://www.alcyone.com/pyos/empy/
 A templating system for Python.



From luc at nospam.com  Sat Apr 12 09:00:09 2003
From: luc at nospam.com (luc wastiaux)
Date: Sat, 12 Apr 2003 13:00:09 +0000 (UTC)
Subject: readline clear buffer
References: <slrnb9edf0.tds.luc@grizzly.dont-panic.info> <FWRla.4654$T34.116245@news2.tin.it>
Message-ID: <slrnb9g3en.uvs.luc@grizzly.dont-panic.info>

In article <FWRla.4654$T34.116245 at news2.tin.it>, Alex Martelli wrote:
> luc wastiaux wrote:
> 
>> Hello,
>> when using readline, I can retrieve the contents of the line buffer using
>> get_line_buffer()
>> 
>> but is there a way to clear this buffer ? none of the provided functions
>> seem to allow that: http://www.python.org/doc/lib/module-readline.html
> 
> Right: it seems that Python's readline.c does not expose interfaces to
> rl_delete_text, rl_kill_text, or other GNU Readline functions that would
> allow such behavior.  I suspect the simplest solution would be to patch
> readline.c to add such interfaces (also patch the docs accordingly), and
> submit the patch to sourceforge, where I think it would be welcome (and
> still in time to get into Python 2.3, for a modest addition, I think).

Hello,
thank you for your advice, I patched readline.c and added a delete_text 
method, I will test it and if it behaves correctly I will submit the patch 
to sourceforge.

-- 
luc wastiaux
 $ finger luc at info.4002.org



From dino_levy at mindlessSPAM.com  Mon Apr 21 16:56:49 2003
From: dino_levy at mindlessSPAM.com (Dino Levy)
Date: Mon, 21 Apr 2003 22:56:49 +0200
Subject: Mail and text extraction
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com> <3EA34DAE.EC67C1E3@engcorp.com> <pan.2003.04.21.13.28.15.303279.2294@mindlessSPAM.com> <3EA3DB6B.DAC620BD@engcorp.com>
Message-ID: <pan.2003.04.21.22.56.47.757048.1444@mindlessSPAM.com>

On Mon, 21 Apr 2003 13:52:11 +0200, Peter Hansen wrote:
> Oh, it's quite achievable that way.  It's just the wrong way to do it.
> 
> Mixing in the various operations like that is a form of optimization,
> and since nothing's actually working yet it's clearly premature, and
> therefore not good. :-)

Hehe, I certainly agree with this one... But I like to do such things,
like polishing the code while it does not yet exist, or make hundreds of
new options DURING the programming. That's why many of my programs never
see the daylight:)

> MIME and POP3 are orthogonal.  MIME is an encoding, POP3 is a transport
> protocol (or something).  You can have one, or the other, or both, or
> neither...
> 
> By the way, the "quite achievable" part above simply means you could
> easily build this capability into custom software.  The "wrong way" just
> means that would likely be a large waste of time, given that the
> individual pieces already exist in standard modules, quite well tested
> and debugged.  Just combine them in appropriate ways and keep life
> simple.
> 
> -Peter

Sooorry, Mea culpa. I overlooked your sentence and wrote it
automatically. I was actually thinking about IMAP. I don't want IMAP, I
want POP3. Sorry.
I'm going to experiment with this downloading the message first and then
parsing it method... 
Thanks...
Dino



From bdesth at removethis.free.fr  Fri Apr 18 19:30:05 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Fri, 18 Apr 2003 23:30:05 +0000
Subject: Why functional Python matters - A kind of conclusion
References: <mailman.1050527910.25653.python-list@python.org>
Message-ID: <3EA08A7D.3010206@removethis.free.fr>

Bjorn Pettersen wrote:
>>From: laotseu [mailto:bdesth at removethis.free.fr] 
>>
>>Paul Foley wrote:
>>
>>>On Tue, 15 Apr 2003 23:49:37 +0000, laotseu  wrote:
>>>
>>>
>>>
>>>>Even for OO programmers, functionnal features in Python are IMHO a
>>>>great plus, and BTW functionnal and OO paradigm does not have to
>>>>conflict (that would be functionnal vs imperative and object vs
>>>>procedural). CLOS is one of the great system objects out here, and
>>>>it's been implemented on top of a functionnal language,
>>>
>>>Oh yes?  Which functional language would that be?
>>>
>>
>>CLOS means Common Lisp Object System.
> 
> 
> "Houston: we have a claim that CL is a functional language, please
> proceed with crosspost to c.l.sml, c.l.haskell, ..." <grin>.
> 

Ok folks

Due to my deep ignorance, I came to believe that Lisp was the first 
programming language based on the lambda calculus theory, the first 
language to promote functions as first class objects, and the first 
language to promote side-effect free programming style...

It seems that good old time is gone, and that even lispers themselves - 
BTW, is Paul Graham still a Lisper ? -  claims that Lisp is not a 
functional language. So I just can agree with them and <troll>proudly 
announce that Lisp is now a procedural, pascal-style language</troll>. 
oops, sorry, that was not that either. Er, a general multi-paradigm 
language.

So :

for x in range(100):
   print "I shall not say that Lisp is a functional language anymore"


Laotseu




From rmunn at pobox.com  Wed Apr  2 17:44:17 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 22:44:17 GMT
Subject: Python enthusiasts in Central Arkansas?
References: <mailman.1049316006.24627.python-list@python.org>
Message-ID: <slrnb8m4u8.dih.rmunn@localhost.localdomain>

(no name) <yuba at cyberback.com> wrote:
> I am attempting to find enough Python enthusiasts in Central Arkansas to form
> a user's group (we have folks running an open source, as well as a perl
> group).  If you're interested, drop me a note off-list.
> 
> Thanks!
> --greg

I'm not in Central Arkansas; but I replied to your message to put a
subject on it, so that the people you want reading it will be more
likely to see it. :-)

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From skip at pobox.com  Sun Apr 13 18:12:40 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sun, 13 Apr 2003 17:12:40 -0500
Subject: PEP305 csv package: from csv import csv?
In-Reply-To: <slrnb9jdvr.se1.mlh@furu.idi.ntnu.no>
References: <mailman.1049814434.25284.python-list@python.org>
        <slrnb99d6s.2ve.mlh@furu.idi.ntnu.no>
        <mailman.1050010580.22697.python-list@python.org>
        <slrnb9jdvr.se1.mlh@furu.idi.ntnu.no>
Message-ID: <16025.57560.690450.379931@montanaro.dyndns.org>

    Magnus> When will it be available in the main Python cvs? (Just curious...)

Checked in a couple days ago.

Skip




From martin at v.loewis.de  Sun Apr 13 16:35:51 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 13 Apr 2003 22:35:51 +0200
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk> <3e970abc$0$49101$e4fe514c@news.xs4all.nl> <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk> <1e740$3e99a2f6$516049d2$15395@nf1.news-service.com> <2+aM4KAFkam+EwV+@jessikat.fsnet.co.uk>
Message-ID: <m37k9y2l08.fsf@mira.informatik.hu-berlin.de>

Robin Becker <robin at jessikat.fsnet.co.uk> writes:

> Well after thinking about it I guess you're right. I guess I'm harking
> back to times when only small numbers of file handles could be held and
> thinking that sockets might be subject to the same kind of limits.

You are right that file descriptors are a limited resource, and any
carefully designed distributed-computing library needs to take this
into account. *However*, if taken into account, you can make good use
of this limited resource by nearly exhausting it:

Allow your library to consume a certain number of file descriptors,
just about as much so that enough are left for local file IO. Then the
library should implement some collection mechanism for unused socket
connections, and shut down anything that has not been used for a
while. The protocol needs to allow shut-downs from either side, as
both partners may experience file descriptor exhaustion.

In CORBA, you need to keep open all connections from which you expect
a response. So if your allocated socket descriptor pool is exhausted,
you first close those connections that have no outstanding requests.
If that is still insufficient, you also close the socket on which you
have been waiting for a response longest, and tell the application
that this connection has timed out.

Regards,
Martin



From revanna at mn.rr.com  Tue Apr  1 11:47:17 2003
From: revanna at mn.rr.com (Anna)
Date: Tue, 01 Apr 2003 10:47:17 -0600
Subject: else: w/o except: - why not?
References: <mailman.1049211247.3694.python-list@python.org>
Message-ID: <pan.2003.04.01.16.47.16.340267@mn.rr.com>

On Tue, 01 Apr 2003 07:33:57 +0000, Michael Chermside wrote:

> I can't speak for Tim, but my OWN opinion is precisely in alignment with
> what you have written.
> 
> I think that a "bare except" is useful in only two ways. The first is like
> a "finally" clause that's skipped if everything works. In other words, the
> following bits of code is (almost[1]) equivalent to your second example:
> 
>     everythingWorked = False
>     try:
>         f1()
>         everythingWorked = True
>     finally:
>         if not everythingWorked:
>             f2()
> 
> This doesn't come up much, but it's occasionally useful, and your version
> (using bare except with a raise) is a little prettier.

I thought that finally clauses are *always* run, whether an exception was
raised or not. 

Looks at code closely... So, in this case the finally is run (in terms of
evaluating the "if" statement), but function f2 may not happen depending
on the results of the "if". So the finally isn't really skipped, it just
doesn't do much if everything worked properly in the try clause. Right?

I'm not being picky - just seeing if I understand this correctly. Thanks.

Anna





From khowe at perfnet.ca  Fri Apr 25 11:13:21 2003
From: khowe at perfnet.ca (Kevin Howe)
Date: Fri, 25 Apr 2003 11:13:21 -0400
Subject: pythong referencing/aliasing
Message-ID: <ylcqa.9116$h%2.963225@read1.cgocable.net>

How does Python handles variable references/aliases?

Example:

a = 1
b = (REF)a
b = 2
print a

In this case, since "b" is just a reference to "a", changes to "b"
automatically change "a".

Does python have this ability?







From rstephens at vectron.com  Mon Apr 21 09:44:55 2003
From: rstephens at vectron.com (Ron Stephens)
Date: 21 Apr 2003 06:44:55 -0700
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org> <QxAoa.540832$3D1.296696@sccrnsc01>
Message-ID: <a5415304.0304210544.1342b436@posting.google.com>

Boa Constructor is, in my opinion, a great project. It will give
Python an open source Delphi- or Visual Studio- like IDE with visual
gui-creator.

PythonCard , in my opinion, goes beyond this in giving us an extremely
easy to learn and use version of the same thing. Boa Constructor, like
VS and Delphi, is quite complex. Learning Visual Studio or Boa
Constuctor is a majopr undertaking all by itself. PythonCard, on the
other hand, is extremely user friendly. I think PythonCard is
revolutionary in empowering programmers who are not prefessional
computer scientists.

Ron Stephens
But Boa's great too.
"Howard Dunlavy" <home at dunlavy.net> wrote in message news:<QxAoa.540832$3D1.296696 at sccrnsc01>...
> "Tim Butterfield" <timbutterfield at mindspring.com> wrote
> 
> >>>>>>
> What I envision is an IDE for python similar to VB and Delphi
> (PyDE?) that allows for those property editors and resource files to
> be integrated into an editor/debugger and builder.  Of course, being
> able to add custom widgets and tie in data support to those widgets
> would be needed.  Is such a thing desired?
> <<<<<<
> 
> There is project that I have come across called Boa Constructor which lists
> as its goal the creation of the equivalent of Delphi for Python.  It can be
> found at Source Forge also.
> 
> -Howard



From mertz at gnosis.cx  Tue Apr  1 00:48:07 2003
From: mertz at gnosis.cx (David Mertz)
Date: Tue, 01 Apr 2003 00:48:07 -0500
Subject: Histogram (was: When Python outruns C++ )
Message-ID: <XgSi+kKkXUFb092yn@gnosis.cx>                       

Julian Tibble <chasm at rift.sytes.net> wrote previously:
...histogram program...
|file of size ~80MB, the C program outperformed the python one
|$ (time ./words < biginput) > /dev/null            #  0m8.476s
|$ (time python words.py < biginput) > /dev/null    # 1m59.747s

This much difference is surprising to me.  I figured there had to
be something wrong with Tibble's Python version; but it seems
like it's not that bad.  The fastest Python histogram program I
know is below (the example came up before :-))... but in a quick
test against a smaller 5MB file, the below takes over half the
time of Tibble's (FWIW, the version I present in my book,
<http://gnosis.cx/TPiP/chap2.txt>, is between the two in
speed--but mine has a few more options).  I tested under 2.3a2,
FWIW.

------------------------------------------------------------------------
#!/usr/local/bin/python
# $Id: wordfreq.python,v 1.9 2001/05/11 17:44:00 doug Exp $
# http://www.bagley.org/~doug/shootout/
#
# adapted from Bill Lear's original python word frequency counter
#
# Joel Rosdahl suggested using translate table to speed up
# word splitting.  That change alone sped this program up by
# at least a factor of 3.
#
# with further speedups from Mark Baker

import sys

def main():
    count = {}
    i_r = map(chr, range(256))

    trans = [' '] * 256
    o_a, o_z = ord('a'), (ord('z')+1)
    trans[ord('A'):(ord('Z')+1)] = i_r[o_a:o_z]
    trans[o_a:o_z] = i_r[o_a:o_z]
    trans = ''.join(trans)

    rl = sys.stdin.readlines

    lines = rl(4095)
    while lines:
        for line in lines:
            for word in line.translate(trans).split():
                try:
                    count[word] += 1
                except KeyError:
                    count[word] = 1
        lines = rl(4095)

    l = zip(count.values(), count.keys())
    l.sort()
    l.reverse()

    print '\n'.join(["%7s\t%s" % (count, word) for (count, word) in l])

import time
start = time.clock()
main()
sys.stderr.write('%4.2f seconds\n' % (time.clock()-start))


--
 mertz@   _/_/_/_/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY:_/_/_/_/ v i
gnosis  _/_/                    Postmodern Enterprises         _/_/  s r
.cx    _/_/  MAKERS OF CHAOS....                              _/_/   i u
      _/_/_/_/_/ LOOK FOR IT IN A NEIGHBORHOOD NEAR YOU_/_/_/_/_/    g s





From km-news1 at onlinehome.de  Sat Apr 26 15:03:16 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Sat, 26 Apr 2003 21:03:16 +0200
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de> <ZFyqa.7824$3M4.220335@news1.tin.it>
Message-ID: <oprn827qp6mu3s0q@news.cis.dfn.de>

On Sat, 26 Apr 2003 16:42:33 GMT, Alex Martelli <aleax at aleax.it> wrote:

> as you proceed to learn that "to call callables, you need to CALL them"

You misunderstand me (and the OP, i think). It's not a problem of 
understanding, it's a simple type error or a "false remember" (or a 
"Fl?chtigkeitsfehler" in German, i don't know a translation).
Programmers sometimes make simple mistakes ;-)


from the OP:

if object.isGreen:
        do something...
?
PyChecker could (optional?) warn in such cases if isGreen is a function.

-- 
Gru? - regards Klaus :-)



From phil at river-bank.demon.co.uk  Tue Apr 15 09:49:23 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Tue, 15 Apr 2003 14:49:23 +0100
Subject: Implementing Metatypes in C
In-Reply-To: <lk2o9v8bamjhgdmas6ce2g2qvuthril237@4ax.com>
References: <mailman.1050275480.8590.python-list@python.org> <lk2o9v8bamjhgdmas6ce2g2qvuthril237@4ax.com>
Message-ID: <200304151449.23490.phil@riverbankcomputing.co.uk>

On Tuesday 15 April 2003 2:34 pm, Greg Chapman wrote:
> On Mon, 14 Apr 2003 00:10:10 +0100, Phil Thompson
> <phil at river-bank.demon.co.uk>
>
> wrote:
> >1. Is my understanding correct, or is there a more correct way to do what
> > I want that avoids the problem?
> >
> >2. If my solution is correct, can the structure defined in typeobject.c be
> >published in a header file as the comment says? Please?
>
> I believe your solution may break in 2.3a2 if subtypes declare tp_members,
> since the PyMemberDefs for these are kept in the variable part of the type
> structure which follows the things declared in etype.  At any rate, check
> out this patch (accepted after the release of 2.3a2):
>
>    http://www.python.org/sf/696193
>
> which should allow you to cleanly add extra data slots to metatypes.

I have a temporary hack in place based on the current CVS version (which 
includes the above patch) which I can remove once 2.3b1 is released.

Phil




From gh at ghaering.de  Tue Apr 29 08:02:23 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 29 Apr 2003 14:02:23 +0200
Subject: SV: Function to convert fetchone() result from a list to a
	dictionary
In-Reply-To: <BHEFKIOIMEDGILGDACKDKEBEDFAA.carsten@gehling.dk>
References: <BHEFKIOIMEDGILGDACKDKEBEDFAA.carsten@gehling.dk>
Message-ID: <3EAE69CF.5010303@ghaering.de>

Carsten Gehling wrote:
>>Fra: python-list-admin at python.org
>>[mailto:python-list-admin at python.org]Pa vegne af Neil Padgen
>>Sendt: 29. april 2003 11:16
> 
> 
>>Take a look at the cursorclass argument:
> 
> 
> <snip>
> 
>>>>>c = db.cursor(cursorclass=MySQLdb.cursors.DictCursor)
> 
> 
> <snip>
> 
> Hmm... Nice. However, is this MySQLdb.cursors.DictCursor available in other
> DB modules 

No.

> - say, if I someday changed to (shudder) Postgres or Oracle?

I don't use (shudder) MySQL :-P

pyPgSQL has it's own tuple type that will also allow you to access 
columsn via name:

for row in cursor.fetchall():
	print row.FIELD_1, row[2], row["FIELD_2"] # 3 syntaxes possible

For my Oracle uses, I adapted pyPgSQL's PgResultSet class and built a 
generic DB-API wrapper around it (it wraps connections and cursors, such 
that all cursor.fetchXX now return PgResultSets).

AFAIK both MySQLdb and psycopg support the .dictfechXXX() methods.

My favourite so far is db_row (http://opensource.theopalgroup.com/) as 
it can be used with any database and seems to be pretty darn fast. Our 
plans are to adapt this or develop something similar for pyPgSQL 3 and 
PySQLite 0.5.

-- Gerhard




From tweedgeezer at hotmail.com  Tue Apr  1 15:07:57 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 1 Apr 2003 12:07:57 -0800
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift> <mailman.1049181423.2644.python-list@python.org>
Message-ID: <698f09f8.0304011207.5552813c@posting.google.com>

Jp Calderone <exarkun at intarweb.us> wrote in message news:<mailman.1049181423.2644.python-list at python.org>...
>   It looks like this could easily be improved with the application of a
> slightly smarter searching algorithm, such as Rabin-Karp search or
> Boyer-Moore find.

These searches depend on a fixed needle to search for in the haystack.
 They can't be applied to a regular expression search.

Jeremy



From mail at andreas-penzel.de  Wed Apr 23 05:19:48 2003
From: mail at andreas-penzel.de (Andreas Penzel)
Date: Wed, 23 Apr 2003 11:19:48 +0200
Subject: Problem Cron-Job using Python-Script
References: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de>
Message-ID: <b85lrk$62480$1@ID-69142.news.dfncis.de>

"Andreas Penzel" <mail at andreas-penzel.de> schrieb im Newsbeitrag
news:b7jeem$1l5i0$1 at ID-69142.news.dfncis.de...
> Hello!
>
> I have a Python-Script to create Log-Files und work with them.
> Very well.
> But if I start the Script with Cron it works not correct.
> I often use os.system() to send commands to the shell including paths from
> the os.
> Could this be the problem?
> And if, how to solve?
>

I think it happens here:

os.system("iptables -L -v > /path/logfile.txt")

The file logfile.txt in /path is after running the script empty.
Is there a other way to become the result from iptables -L -v in a file?

- Andreas







From bobx at linuxmail.org  Wed Apr 16 16:55:26 2003
From: bobx at linuxmail.org (Bob X)
Date: Wed, 16 Apr 2003 20:55:26 GMT
Subject: Web Hosting
Message-ID: <2rjna.171$hT2.182473@news2.news.adelphia.net>

Can you name a good (reasonably priced) Web Hosting company that supports
Python (either CGI or mod_python)?

Bob





From aahz at pythoncraft.com  Wed Apr 30 09:45:56 2003
From: aahz at pythoncraft.com (Aahz)
Date: 30 Apr 2003 09:45:56 -0400
Subject: importing modules with the same name but from different directories
References: <mailman.1050892833.31640.python-list@python.org> <b80pdp$81h$1@news.tu-darmstadt.de> <T8Qra.32135$WQ4.29932@news2.central.cox.net>
Message-ID: <b8ok2k$p8d$1@panix1.panix.com>

In article <T8Qra.32135$WQ4.29932 at news2.central.cox.net>,
Steve Holden <sholden at holdenweb.com> wrote:
>
>What you definitely *can't* do is this:
>
>    m = "modulename_or_path"
>    import m as something
>
>To emulate this non-available solution and actually execute an import you
>would need to use this:
>
>    m = "modulename_or_path"
>    exec "import %s as something" % m
>
>I'd advise against any solution that involves exec'ing or eval'ing unless
>there's really no alternative, which is not the case here. What you would
>need for this is the __import__() function, built in to the language and
>just waiting for you to take advantage of it.

Use the ``imp`` module.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From aleax at aleax.it  Wed Apr 16 05:08:26 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 09:08:26 GMT
Subject: Protected Methods and Python
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <18lb3o2cs1se2$.isoixa5igmj6$.dlg@40tude.net> <12aeh974i1kdh.dfbwdn5e18px.dlg@40tude.net> <mailman.1050275130.5456.python-list@python.org> <jg4na.146561$S14.33908@news1.central.cox.net>
Message-ID: <e49na.20877$LB6.518824@news1.tin.it>

Venkatesh Prasad Ranganath wrote:
   ...
> All said, how do you ensure "design by interface" in Python?   If every
> feature of a class is open to the user then he/she
> will use them as he/she sees fit.  However, the writer of the class could
> have coded some methods for ancillary purposes and
> later decides to do away with them in the next version of the class.  How
> should the user deal with this situation?

Quite simple, really: a name that starts with a leading underscore
indicates that the name is there "for ancillary purposes" as you put
it -- it's not in the "published public interface" of the module,
class, etc; a name _without_ such a leading underscore is part of
the public interface, and won't be "done away with" in the next
version of the package, module, class, and so forth.

Thus, the user should deal with this situation by using only names
exposed in the public interface of the package, module, class, etc --
names that don't start with a leading underscore -- unless of course
said user *WANTS* to tie himself or herself to the specific version
of the reusable component.

> In my opinion, protected and
> private as in C++ or Java empowers the library developer to expose just
> the interface that he/she
> intends the user to see.  Nothing more and nothing less.  (It's an issue
> dealing with abstraction.)  

Not quite true in C++, since casts allow the user to work around this
(in most implementations) -- just as Python allows working around
exactly the same attempts by the library designer to impose his or her
views of appropriate abstraction by using marked-as-internal names.

In Python and Java the user can also work around such limitations by
introspection, unless the user's code is treated as untrusted and
is run inside a sandbox that tries to limit the damage the user can
do (that was the design intent of standard library modules rexec and
Bastion in Python -- enforcing those restrictions that are normally
just respected by convention, by treating user code as untrusted and
potentially intended to be actively harmful -- they're currently broken,
but equivalent alternatives such as package Sandbox are being explored).

But issues of security are really very different from issues of
abstraction, IMHO, even though Java does its best to conflate them --
if you must think of user code as actively trying to do maximum damage
to the system, you've got a LOT of things to worry about.  For normal
use, having the library designer consider the future library user as
a nasty adversary to be foiled at every turn is really an unproductive
mindset, even though it's the one that accessibility restrictions suggest.


> As a result, the user can be more care-free
> while using the library code as he cannot screw up the internals if the
> library code doesn't do it on it's
> own.  On the other hand, if all members of a class are visible to the user
> then the user needs to be more diligent while using
> the library.  I prefer the first option.

The user can rather easily "screw up" things if and when he or she
does it DELIBERATELY -- and in fact, even an accident is quite likely.
E.g., consider the hypersimplified code where the library has
something like:

class Base {
  private: virtual int another() { 
    return 100; 
  }
  public: virtual int amethod() {
    return 23+another();
  };
};

Method another is virtual because class Base is actually derived from
others and that method is overridden inside the library here and there,
etc, etc.  But it's private, so the user can't tamper accidentally
with it, right?  Yeah, right...:-).  So, way over there in userland,
the user, not even "seeing" (you do keep saying "visible" while what's
supposed to happen has to do with "accessible" instead...) the private
parts of the class Base it gets from the library, happily codes:

class Derived: public Base {
  public: virtual int another() { return 1000; }
};

int xx(Base* d)
{
  return d->amethod();
}

int main()
{
  Derived dd;

  std::cout << xx(&dd) << '\n';
  return 0;
}

guess what this prints...?


In Python, you can choose between calling the meant-to-be-internal
method _amethod (the user then must KNOW not to use it) or avoiding
accidental clashes by calling it __amethod (so it gets name-mangled
by the Python compiler to help avoid _accidental_ clashes).  Not
foolproof, yet (homonymous classes in different modules are still
*possible*, unless the designer takes care about that...), but quite
a bit better than what pure *accessibility* restrictions achieve in C++
in cases such as this.


> On the issue of working around bugs in a library via runtime protected
> attributes, it would be better if the skillful coder
> fixed the bug ;-) or use another library.  This is just what I would do.

Modifying source code inside a component that is maintained by
another party -- basically "forking" that component -- even where
_possible_ at all because you do get source, is a grievous mistake
(correctly identified as such in the important book "Antipatterns").

By making your own fixes (forking) the component, you make your
situation *UNTENABLE* when new releases of that component come out --
you're contemned to keep maintaining your fixes forever, porting
them to each successive release, rebuilding from source every time,
distributing the .DLL for YOUR fixed component (when you can even
have the LICENSE to do so...) rather than relying on system-provided
versions...

The option to avoid buggy libraries isn't always there - sometimes
one IS condemned to e.g. develop on top of Microsoft MFC, because
that's part of the specs for your component.  Then, you start cursing
a language which has let the designer of those libraries impede you
from access to a zillion things you desperately DO need to access
(or, you start doing heavy-duty casting, but there are still some
limitations that will stump you quite badly, such as the use of
concrete classes rather than factory-idioms where you NEED to be
able to interpose your own subclass of some widget and the deuced
framework insists on generating the concrete baseclass instead...).


Alex




From dooms at info.ucl.ac.be  Wed Apr 23 06:21:22 2003
From: dooms at info.ucl.ac.be (=?iso-8859-1?Q?Gr=E9goire?= Dooms)
Date: Wed, 23 Apr 2003 12:21:22 +0200
Subject: python language: infimum and supremum of integers
Message-ID: <20030423102121.GA837@yhwh>

Hi,
I need a lower and higher integer for initializing lower and higher
bounds in an algorithm.  I know integers have no limits(maximum and
minimum) in python and infimum and supremum would do the job too.

I empirically discovered (Python 2.2.2) that None could be considered an
infimum and any string could be considered a supremum.
e.g "a">int(x) is always true and None<int(x) is always false

But this fact is not stated in the language reference. 
Citing http://www.python.org/doc/current/ref/comparisons.html : 
"The objects need not have the same type. If both are numbers, they are
converted to a common type. Otherwise, objects of different types always
compare unequal, and are ordered consistently but arbitrarily."

Could/should that order be hardcoded in the language ?

-- 
Gr?goire Dooms
Department of Computer Engineering
Universit? Catholique de Louvain.




From max at alcyone.com  Tue Apr  1 20:12:55 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 01 Apr 2003 17:12:55 -0800
Subject: The Whitespace Language
References: <283adf56.0304011143.4f4a6934@posting.google.com> <3E8A12CD.6194A9A8@alcyone.com> <eGWdnSuSpbdBqBejXTWc3w@comcast.com>
Message-ID: <3E8A3917.7FABDFBD@alcyone.com>

Terry Reedy wrote:

> Do you mean this WhiteSpace (XML-centered Java extension)?
> http://www.bluering.nl/whitespace/faq.html

No, this one:

	http://compsoc.dur.ac.uk/whitespace/

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The work of many cannot be done alone.
\__/ (a Bemba proverb)
    7 Sisters Productions / http://www.7sisters.com/
 Web design for the future.



From aleax at aleax.it  Wed Apr 30 13:12:46 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 17:12:46 GMT
Subject: looping through a file
References: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au> <uWOra.24347$3M4.683692@news1.tin.it> <mailman.1051711067.28906.python-list@python.org>
Message-ID: <iuTra.42265$K35.1222818@news2.tin.it>

Andrew Bennetts wrote:
    ...
>> InFile = open('users.dat', 'r')
>> 
>> for line in InFile:
>>     file_username = line.strip()
>>     file_password = InFile.next().strip()
>>     if file_username==users_username and file_password==users_password:
>>         authorized = 1
>>         break
>> else:
>>     authorized = 0
>> 
>> InFile.close()
> 
> I thought there might be something in 2.3's itertools module to help with
> this -- I was hoping you'd be able to write something like:
> 
>     InFile = open('users.dat', 'r')
>     
>     for username, password in multistep(InFile, 2):

Careful!!!  You'd still need to strip them before you compare them,
else the trailing \n or any other whitespace will make the comparisons
fail.  imap(str.strip, multistep(InFile, 2)) might work here (if
multistep existed).

> Where "multistep" would probably have a better name, and be defined like:
> 
>     def multistep(iterable, count):
>         # XXX: What should happen when len(iterable) % count != 0 ?

I think the last trailing part X with 0<=len(X)<count should be dropped.
Otherwise, unpacking assigments such as the above would fail -- and,
"padding with None's" isn't in fashion any more;-).

>         iterator = iter(iterable)
>         while 1:
>             values = []
>             for i in range(count):
>                 values.append(iterator.next())
>             yield values

I think I'd prefer expressing this loop (naming issues apart) as:

   while True: return [iterator.next() for i in range(count)]

> 
> I couldn't see anything in itertools to save me writing that out.  But I
> just figured out how to do it with itertools -- izip can do it:
> 
>     def multistep(iterable, count):
>         return itertools.izip(*((iter(iterable),)*count))
> 
> I guess you could write this directly:
> 
>     iterator = iter(InFile)

Unneeded, btw: iter(inFile) is inFile (I also exploit that in my
less-elegant snippet above where I call inFile.next(), secure that
this will consume a line).

>     for username, password in izip(iterator, iterator):
>         ...
> 
> So it's probably not worth defining a seperate function...

I think it is (if anything can still be added to new module itertools
in 2.3 -- I'm not sure ANY functionality can be added after beta-1), as
it's frequently needed.  But, if it isn't, the "trick" with the
iterator explicitly repeated is good and IMHO worth keeping in mind.

But one does still need the .strip call...;-)


Alex




From srfarley at yahoo.com  Wed Apr 16 07:06:25 2003
From: srfarley at yahoo.com (Steven F.)
Date: 16 Apr 2003 04:06:25 -0700
Subject: Standard way to express a date in a real number?
References: <420ced91.0304141131.73a6206e@posting.google.com> <d64f1eb0a893848f6b69af2538989f9b.124124@mygate.mailgate.org> <bEGma.462$8c4.82@tornadotest1.news.pas.earthlink.net> <420ced91.0304152121.2af84a33@posting.google.com>
Message-ID: <86d3b75b.0304160306.faafadc@posting.google.com>

> So, this is not the solution I am seeking.
> Better module to handle wider range of time?

Perhaps mxDateTime would be helpful:
http://www.egenix.com/files/python/mxDateTime.html

++Steve



From aleax at aleax.it  Sun Apr 13 16:17:21 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 20:17:21 GMT
Subject: How to subclass datetime?
References: <ba1e306f.0304130339.3bece039@posting.google.com> <ba1e306f.0304130850.1db6bf69@posting.google.com>
Message-ID: <lBjma.9598$LB6.256808@news1.tin.it>

Jane Austine wrote:

>> > Running Python 2.3a2 on win XP.
>> > 
>> > I'm interested in the new datetime module and I want to subclass
>> > date/datetime class to modify the behaviour. However, simply
>> > inheritting the classes(like "class mydate(date):" ) didn't work.
>> 
>> Please define "didn't work"...:
>> 
>> >>> import datetime
>> >>> class mydate(datetime.date):
>> ...   pass
>> ...
>> >>> xx=mydate(2003,4,14)
>> >>> print xx
>> 2003-04-14
>> >>>
>> 
>> Seems to be working just fine.  What exactly isn't working for you?

"seems" is the key word here -- behind the covers, weird things are
indeed happening;-).  Sorry for not looking into it a bit deeper in
the first place, but the generic "didn't work" indication wasn't
very encouraging to do so;-).


> For example,
> 
>>>> class mydate(datetime.date):
> def abc(self):
> print "abc"
> 
> 
>>>> z=mydate(2003,4,14)
>>>> z.abc()
> Traceback (most recent call last):
>   File "<pyshell#6>", line 1, in -toplevel-
>     z.abc()
> AttributeError: 'datetime.date' object has no attribute 'abc'
>>>> type(z)
> <type 'datetime.date'>

VERY interesting.  It seems to boil down to function date_new in
module datetimemodule.c -- i.e., in Python terms, date.__new__ --
it ignores the PyTypeObject it's passed, and generates an object
of the SPECIFIC type PyDateTime_DateType no matter what.

I think it can be stated that you've found a bug, one way or
another -- if the design intent is that the type cannot be
subclassed, then the type shouldn't have the Py_TPFLAGS_BASETYPE
flag set (which it does), so the attempt at subclassing it would
immediately give a clear error message; if subclassing is meant
to be allowed, then date.__new__ should work more canonically...

I suggest you submit the bug report to sourceforge -- that's what
alpha releases are for, after all.


Alex





From heiland at ncsa.uiuc.edu  Fri Apr 25 11:37:27 2003
From: heiland at ncsa.uiuc.edu (Randy Heiland)
Date: Fri, 25 Apr 2003 10:37:27 -0500
Subject: install/import modules, /site-packages
References: <3EA94E2D.6010505@ncsa.uiuc.edu> <3EA952D6.70C6A89A@engcorp.com>
Message-ID: <3EA95637.60703@ncsa.uiuc.edu>


Peter Hansen wrote:
> Randy Heiland wrote:
> 
>>I need some basic Python module knowledge.  On my RH Linux installation,
>>I've got the following:
>>
>>/usr/local/lib/python2.2/xml
>>and
>>/usr/local/lib/python2.2/site-packages/_xmlplus
>>
>>and when I >>> import xml.sax, I get:
>>'/usr/local/lib/python2.2/site-packages/_xmlplus/sax/__init__.pyc'
>>
>>How does this module (in site-packages) take precedence over the other?
> 
> 
> "Use the source, Luke."  
> 
> Look in lib/python2.2/xml/__init__.py and all will be revealed.  :-)
> 
> -Peter

Indeed. I had this env set:
PY_SAX_PARSER=xml.sax.drivers.drv_pyexpat

Ugh.  thanks,
-Randy




From martin at v.loewis.de  Mon Apr 14 17:34:13 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 14 Apr 2003 23:34:13 +0200
Subject: Displaying Unicode on the console (Windows)
References: <mailman.1050352043.17028.python-list@python.org>
Message-ID: <m38yucvk4q.fsf@mira.informatik.hu-berlin.de>

sismex01 at hebmex.com writes:

> > Yes, through WriteConsoleW.
> It accepts UTF-8?  Or only 16bit chars?

It accepts WCHAR, which is 16-bit chars.

Regards,
Martin



From tjreedy at udel.edu  Tue Apr 29 12:57:54 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 29 Apr 2003 12:57:54 -0400
Subject: The behavior of List.insert
References: <b8m8dn$b03$1@pegasus.csx.cam.ac.uk>
Message-ID: <TZOdnU3XYIdTMzOjXTWcpQ@comcast.com>

"York" <yorklee70 at yahoo.com> wrote in message
news:b8m8dn$b03$1 at pegasus.csx.cam.ac.uk...
> mylist = [1,2,3,4,5],
>
> mylist.insert(-1,99) will make mylist to be [99,1,2,3,4,5]
>
> just wonder why not let it to be [1,2,3,4,99,5]

Guido decided otherwise for whatever reason.  What you observed is
intended: from LibRef 2.2.6.4 Mutable Sequence Types
"(4)
When a negative index is passed as the first parameter to the insert()
method, the new element is prepended to the sequence. "
so this is not an implementation bug.

Terry J. Reedy







From rupole at hotmail.com  Mon Apr 21 03:39:58 2003
From: rupole at hotmail.com (Roger Upole)
Date: Mon, 21 Apr 2003 00:39:58 -0700
Subject: win32net.NetUseAdd syntax, possible bug
References: <mailman.1050496946.25294.python-list@python.org>
Message-ID: <3ea36c99_4@corp-goliath.newsgroups.com>

Both of the levels of info for the function use 'password' spelled out.
Looks like you want level 2 info, user is not included in level 1. (key is
'username')
You'll probably also have to pass the 'domainname' for level 2 also.
   HTH
      Roger




"Scott Prive" <scottprive at earthlink.net> wrote in message
news:mailman.1050496946.25294.python-list at python.org...
> Hello,
>
> I'm attempting to use win32net.NetUseAdd, had a problem and I think
there's
> a bug in the API, or the docs (or me! :-)
>
> The ActiveState made it clear most of the properties need to be in a
> dictionary. It lacked an example use, but referred me to the MSDN site,
and
> at MSDN could see there the data structure. In theory I should combine the
> two docs and know all the things I can pass into the call via dictionary.
>
> Small problem: the MSDN data structure calls the PW key "password", and
> win32net.NetUseAdd does not understand this. It looks like the function
> needs a key called "passwd", and if not provided it falls upon some
default
> value (or the current credentials), making it appear that your user/pass
> values are not valid on the server when in fact they were.
>
> This is my first post to the list (cc: me, I'm not subscribed). Is this
the
> appropriate place for possible bug reports? How do I determine a module's
> "owner", and should I mail them directly?
>
> The code looks like:
>
>     import win32net
>     data = {}
>     data['remote'] = r'\\mypdc.example.com\myshare'
>     data['local'] = 'Z:'
>     data['password'] = 'foo' #Hmm.. I should call this     data['passwd']
=
> 'foo'
>     data['user'] = 'foo'
>     data['asg_type'] = 0
>     win32net.NetUseAdd(None,1,data)
>
> -Scott
>
>
>
>
>




-----------== Posted via Newsfeed.Com - Uncensored Usenet News ==----------
   http://www.newsfeed.com       The #1 Newsgroup Service in the World!
-----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =-----



From exarkun at intarweb.us  Fri Apr  4 00:26:36 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Fri, 4 Apr 2003 00:26:36 -0500
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304031917.1d0100bf@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <mailman.1049388384.29614.python-list@python.org> <9383481e.0304031917.1d0100bf@posting.google.com>
Message-ID: <20030404052636.GA8878@meson.dyndns.org>

On Thu, Apr 03, 2003 at 07:17:19PM -0800, Drew Moore wrote:
> Jp Calderone <exarkun at intarweb.us> wrote in message news:<mailman.1049388384.29614.python-list at python.org>...
> 
> > [snip - example usage of property()]
> >
> 
> Well, not exactly what I had in mind.. how bout this?

  I know exactly what you meant.  I was proposing an alternative solution
that actually works given the current form of the language, and which
requires no gut wrenching disfigurement of the current language semantics.

> [snip - example of gut wrenching disfigurement of the current language
>  semantics]
> 
> Drew

  Jp

-- 
"There is no reason for any individual to have a computer in their
home."
                -- Ken Olson, President of DEC, World Future Society
                   Convention, 1977
-- 
 up 15 days, 2:00, 2 users, load average: 0.07, 0.04, 0.01




From peter at engcorp.com  Tue Apr  1 15:09:30 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 01 Apr 2003 15:09:30 -0500
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <mailman.1049226964.14755.python-list@python.org>
Message-ID: <3E89F1FA.EDFBE919@engcorp.com>

Dave Brueck wrote:
> 
> On Tue, 1 Apr 2003, Mike Meyer wrote:
> 
> > PEP: 313
> > Title: Adding Roman Numeral Literals to Python
> > Version: $Revision: 1.2 $
> > Last-Modified: $Date: 2003/04/01 17:41:34 $
> > Author: Mike Meyer <mwm at mired.org>
> > Status: Draft
> > Type: Standards Track
> > Content-Type: text/plain
> > Created: 01-Apr-2003
> > Python-Version: 2.4
> > Post-History:
> 
> +1, well done!
> 
> Before we submit this to the BDFL, I propose that we wait a short while to
> give the community an opportunity to debate its merits, followed by a
> vote, just so Guido can get idea of where the community stands. Further, I
> propose that Aahz be the official vote counter. Does anybody have any
> preferences as to the voting method we should use?

Wouldn't "we who are about to die salute you" somehow be appropriate here?

SPQR!



From ckea25d02 at sneakemail.com  Tue Apr 15 18:58:48 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 10:58:48 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1050394215.20671.python-list@python.org>
References: <mailman.1050394215.20671.python-list@python.org>
Message-ID: <b7i26g$17t5m$1@ID-169208.news.dfncis.de>

Christian Tanzer wrote:
> This has nothing to do with touching point but how cursor
> movements behave. For example, searching doesn't work properly in some
> cases unless you let `inhibit-point-motion-hooks` to a true value.

Er... if it has nothing to do with touching point, how
come you have to fiddle with some global that has 'point'
in its name?

> But there isn't that much difference between an attribute of a buffer
> object (in the Python sense) and a buffer-locale variable (in the
> Emacs sense).

The difference is that the buffer (or some other data structure
from which the buffer can be easily reached) would be an explicit
argument.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From hejj at china.com  Wed Apr  9 13:20:11 2003
From: hejj at china.com (hejj)
Date: Thu, 10 Apr 2003 01:20:11 +0800
Subject: I can't execute my IDLE
Message-ID: <b71kp8$1g8r$1@mail.cn99.com>

The command line mode can execute well.but no response when i click the IDLE
icon in the program.what is wrong?

thx!





From stephan.diehl at gmx.net  Thu Apr 10 10:41:15 2003
From: stephan.diehl at gmx.net (Stephan Diehl)
Date: Thu, 10 Apr 2003 16:41:15 +0200
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <b73vov$d92$00$1@news.t-online.com>

You are probably looking for "linear regression".

Stephan


Alex Martelli wrote:

> I _know_ there's a decent algorithm to solve the following problem,
> but I don't recall its name and thus can't rapidly google for the
> details... can somebody help?
> 
> 
> The problem: I need to design a "decision criterion" to classify
> "observations".  For each observation I measure the values of a
> number N of features, x1, x2, ... xN; the desired design criterion
> is a set of weights w1, w2, ... wN such that for any observation I
> will then just compute a weighted sum
>   S = w1*x1 + w2*x2 + ... + wN*xN
> and classify the observation as Black if S<=1, White if S>1.  To
> train my classifier I have a large corpus of observations already
> made and annotated with "ground-truth" data about whether each
> given observation should have been classified as B or W, and an
> error-cost value for each kind of classification (Ebw is the cost
> of erroneously classifying a feature as W when it should be B,
> Ewb is that of classifying it as B when it should be W).  So,
> what's the algorithm to estimate the weights given the corpus of
> observation and ground-truth data, and the error-costs?
> 
> Responses welcome either here or directly to my mailbox -- if I
> do get responses to my mailbox, I'll eventually summarize here --
> as a bonus, I promise to make available a Python implementation of
> whatever algorithm I end up with, if one's not already around...;-).
> 
> 
> TIA,
> 
> Alex




From bobx at linuxmail.org  Wed Apr 23 16:50:27 2003
From: bobx at linuxmail.org (Bob X)
Date: Wed, 23 Apr 2003 20:50:27 GMT
Subject: Curses on Win32
References: <b86r8p$27n$1@gateway.northgrum.com> <3ea6f33d$0$185$a1866201@newsreader.visi.com>
Message-ID: <n0Dpa.3882$hT2.2046982@news2.news.adelphia.net>

"Grant Edwards" <grante at visi.com> wrote in message
news:3ea6f33d$0$185$a1866201 at newsreader.visi.com...
> In article <b86r8p$27n$1 at gateway.northgrum.com>, Dang Griffith wrote:
>
> I was going to say I agree 100% with the subject line, but on
> second thought Win32 is it's own curse.  It doesn't need them
> added on... ;)
>
LOL

I was going to say the same thing!





From hupp at cs.wisc.edu  Fri Apr 11 14:54:44 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Fri, 11 Apr 2003 13:54:44 -0500
Subject: pyorbit1.99.4  and Python2.3 problem
In-Reply-To: <wFmdnSwh7MtWKwujXTWcpg@comcast.com>
References: <slrnb9cuvj.49s.stehule@kix.fsv.cvut.cz> <wFmdnSwh7MtWKwujXTWcpg@comcast.com>
Message-ID: <20030411185444.GB23331@upl.cs.wisc.edu>

On Fri, Apr 11, 2003 at 08:43:07AM -0400, Terry Reedy wrote:
> 
> "Pavel Stehule" <stehule at kix.fsv.cvut.cz> wrote in message
> news:slrnb9cuvj.49s.stehule at kix.fsv.cvut.cz...
> > I tried pyorbit with new version  of Python and when I try run test
> I get
> > error:
> > [root at stehule pyorbit-1.99.4]# cd tests/
> > [root at stehule tests]# python everything_server.py
> > TypeError: MRO conflict among bases Object, B1, BaseServer, B2
> > TypeError: MRO conflict among bases Object, B1, BaseServer, B2
> >
> > ** (process:13354): WARNING **: could not generate stub for base
> > IDL:orbit/test/C1:1.0
> 
> MRO stands for 'method resolution order'  2.3 has improved the
> definition and calculation of mro.  I believe the error message says
> that there is a conflict or ambiguity in one of your object
> inheritance trees (that was apparently not caught before).  The What's
> New file should say something about this and maybe give a pointer.
> 
> > Is working orbit with 2.3 Python?
> 
> I believe that changing the order of your base classes may solve the
> problem.  I hope someone else knows more about debugging this.


everything_server.py is the server side of the pyorbit tests so there
is definitely something broken.  I'm seeing the same thing on
debian/unstable with pyorbit 1.99.4, python 2.3a2+ (packaged as
2.2.101-1exp1?).

I filed a bug with debian, see 
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=188610 


-Adam




From BPettersen at NAREX.com  Mon Apr 21 00:33:48 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sun, 20 Apr 2003 22:33:48 -0600
Subject: Teaching : Python, Scheme, Java...
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE015@admin56.narex.com>

[I sent this while my PC was crashing, and it shows as sent in my
mailbox, but I can't see it on e.g. Google, so my apologies if this is a
resend...]

> From: Alex Martelli [mailto:aleax at aleax.it] 
> 
> On Friday 18 April 2003 04:52 pm, Bjorn Pettersen wrote:
> >
> > [..tail call to 'assembly' translation of zero argument
functions...'
> 
> You're not dealing with the question of the stack frame.

Yes, I was hoping to gloss over that -- talking about tail call
optimizations in general, just using exmples from a specific and common
implmentation area as far as the tail-call optimization was concerned,
but since you asked... <wink>. The concept of a _stack_ frame is rather
curious if you went directly from theory to implementation without
knowing anything about C (Algol, et.al.). The fact that certain
languages don't even have such a basic concept as a continuation
(arguably as simple as goto [call/cc being a particularly
incomprehensible introduction to continuations :-]), shouldn't mean that
we must use these crippled languages "lowest-upper-bound" implementation
technique for normal languages <wink>.

Theoretically, the straight-forward way of executing f(args), i.e.
without any arbitrary limitations imposed by e.g. implementation
decisions like datastructures (but assuming you can call functions only
with values [not variables/terms/expressions...], so I don't have to
create an entire grammar...), can go something like: First, let's define
an Environment to be a mapping from visible variables to their values at
a given point during program execution (i.e. lexical scoping, nested
calls), "executing a function invocation, f(actual_argument), happens by
executing the code of the function [details about how to get the code
depends on the definition of executing a function _definiton_], in an
Environment (Env' [i.e. Env-prime]) which we create from the current
Environment (Env) extended with a mapping from f.formal_argument to
actual_argument [multiple arguments adds nothing of theoretical
interest, just complexity, so is frequently omitted, sometimes assuming
multiple arguments are passed as a single tuple etc. etc.], mappings for
defined variables, DV(f) = vars declared in function (should be the set
Env setminus FV(f), the free variables, i.e. variables defined lexically
outside the function). And f.name->f.code [the name the function when
defined as opposed to called for recursion]".

Using formal notation, with the term currently executing [f(actual)] at
the bottom, and the steps defining the meaning of saying f(actual) in
this language listed in order from top to bottom over the line, (where +
is sequential update, <B> is bottom [a upside down T] indicating
undefined, and execution of a program [happening in the current
Environment, Env] results in a value [potentially <B>] and a new
Environment, Env'' (because we defined it as a mapping with "time"
they're immutable and we must create new objects -- and since this is
theory not implementation we mostly care about what is easier to reason
about :-):

   fn = Env[f]
   Env' = Env + (fn.formal -> actual) + (DV(fn) -> <B>) +
{fn.name->fn.code}
   Env' |- fn.code -> v, Env''
   -----------------------------------
   Env |- f(actual) -> v, Env''

Note the CPython implementation of locals seem to calculate DV(fn), but
you could get the same effect (semantically) by extending the
Environment on a local assignment. The only difference is that with the
above you can say "UnboundLocalError", with the other only "NameError"
:-)

a Python version of the above might be (Env implemented by a dict):

  def executeFunctionCall(Env, f, actual):
      fn = Env[f] # funciton object, as opposed to function name
      formals = fn.formals # list of formal arguments
      argmapping = dict(zip(formals, actual))
      fnLocals = dict(zip(fn.vars, None))
      Env' = copy.copy(Env).update(argmapping).update(fnLocals)
      v, Env'' = executeCode(Env', fn.code)
      return v, Env''

and assuming there exists a function "fn, Env' =
executeFunctionDefinition(Env, name, formals, body)" which creates the
function object fn with attributes formals, vars, and code computed by
executing the definition as opposed to the code (just like Python). It
also places fn in Env' with key "name", and through successive
modifications of the Environment passed to and returned from the various
executeTermX, we find this object in the Environment passed to us.

[which is how you normally create an interpreter for a language, well,
at least it used to be before all these fancy things that compiled
language1 to a language2 requiring two interpreters instead of one --
strange :-)]

I haven't defined State (i.e. memory layout of the program at any point
in time), but doing so changes nothing substantial in the definition
above, but does add another complexity. (Defining it is easy, carrying
it through all the formulas is tedious :-) E.g. define Environment to be
the mapping of visible variables to locations in the Program State, S
(a.k.a memory) at a point in program execution, and State as a mapping
from locations to values at a point in program execution, you can get C
style variables by defining the semantics to be:

   loc = Env[x]
   S' = S + loc->v
   ---------------------------------------
   Env * S |- x := v -> v * Env * S'

and Python style semantics by keeping the old definiton of Environment
and using the following formula, which defines reference counting to be
part of the semantics of the language as opposed to a property of the
implementation. (Similarly to define the semantics of C you need to
define two types of state/memory stack and heap mappings with
appropriate properties):

   curval = Env[x]
   S' = S + {loc(curval)->decref(curval)}
   Env' = Env + {x->v}
   incref(v)
   ---------------------------------
   Env * S |- x := v -> v * Env' * S'

The semantics of a language is defined by the combination of the
operational formulas for all "parts" (terms of the grammar) of the
language -- or more practically, given the interpreter, I, defined by
OS(L) [the Operational Semantics of the language L], the semantics of a
program, P is v, defined by v = I(P). Type systems, optimizations, other
program transformations can be defined in the same way, and given a
formal semantics you can prove different properties of the language...
[and creating formal definitions of assembly language, you can prove
correctness of e.g. peep-hole optimizers and even compilers...]

Worth mentioning: The Environment is defined by the language (just a
regular object/value containing the current values of all visible
variables [thinking of possible implementations would be premature]).
There is nothing magical about this object/value, i.e. why wouldn't you
be able to assign it to a variable, pass/return it to/from functions,
etc. Of course if you think of semantics (i.e. the meaning of the
program, what happens when it executes) in terms of what any specific
language implementation (or CPU architecture) is limited to you're
putting the cart before the horse. You're mapping an algorithm
[semantics] onto a language [implementation], the other way would be
like implementing assembly in C (possible, but pointless).

Note, C does have an Environment, but it is extrinsic to the language,
mostly only visible through a debugger...

Indeed, there is nothing preventing you from defining the "PC" (current
Program Counter) inside the language so you can operate on it too. Why?
Well, if you combine the PC with the current Environment (PC, Env), you
just created a continuation (defining the PC as some kind of function
would be a "good idea" :-). 

[[If you tilt your head and think of PC as indicating the next term,
which if you tilt further can be defined as a function taking the value
of executing the previous term as an argument and the rest of the
program as the body, lambda x: rest_of_program(x), as an example if the
program has executed to "a = [PC] f(b)",

  # opAssign defined as a "fundamental" operation, e.g. inlined..
  PC = lambda x, k: k(opAssign(a, x))

x being the result of f(b), and k being the rest_of_program (passed as a
an input argument) you're at CPS - Continuation Passing Style/only
forward program and data flow (just imagine how simple program and data
analysis ought to become, not to mention garbage collection <wink>, as a
bonus, call/cc looks a tad more reasonable]].

Going from mappings and set operations, to requiring that the data
layout is physically contigous seems like a rabid overspecification in
general, and rather unintuitive for this algorithm (the argument seems
to be (a) to implement local "scope" we need a stack, (b) we have a
stack, so (c) a stack is necessary for implementing local "scope"
[http://www.wikipedia.org/wiki/Affirming_the_consequent])? And why would
anyone assume it's ok to just remove a random set of values (i.e. in C I
can store a local int variable in a global and it will survive, but an
int* doesn't if what it is pointing to is created by this function,
directly or indirectly, as a local..?) -- in particular, why would you
use time of creation (instead of when you're done with on object
[defined however you want either directly or e.g. reachability,
reference count, life-time analysis]) as a criterion for deletion? It
may have been convenient when we were using assembly, but only a
convenience. Considering that a function can easily store the
Environment object, which, to belabor the point, contains the
variable->value mappings of all nested function calls up to the current
function call (i.e. what C has on the stack), there doesn't seem to be a
correct stack based implementation that doesn't involve a lot of
copying... Can it be that it is the wrong datastructure for locals?

> [.. what about implementation issues: calling conventions, memory
layout,
>     memory (de)alloc, how to optimize tail-call in this scheme..]

You need to think more about the inherent beauty of simple, logically
defined, systems of formulaes with a theory that has direct implications
for what you're doing. Pure (sometimes advanced) science, logic, and
math that can be directly applied -- pretty special. Why waste your time
on unsatisfying problems like how to limit expressiveness purely for the
sake of the convenience of an application developer <chuckle>. Seriously
though, C semantics is barely more than a formalization of the most
effective way to map the terms to a specific implementation algorithm
that places great arbitrary limits on the semantics purely because
otherwise the abstract data-layout would be different from the most
effective hardware layout for assembly language calling conventions... 

Note: I'm saying converting C to assembly should be simple relatively
because (a) they're very similar, (b) possible higher order concepts are
specifically excluded. I'm explicitly not stating either of (1) C isn't
an important language [just not terribly interesting in its semantic
theory], (2) C optimization techniques aren't advanced, i.e. C is very
fast [less time translating gives you more time optimizing...], (3) it's
hard to find a more efficient mapping of algorithm to machine run-time
than the implementation of the algorithm written using C limitations by
an expert programmer. (Which is why I'm using Python, how long something
takes me is much more important than how long it takes the computer.)

You seem to think that compilers are somehow special because they have
requirements (i.e. must compile to byte-code, or MIPS, etc.). If you
think of a language, L, as the requirements document for a regular
application trnalating a text file with a certain formal structure, to a
simpler language defined by another requirements document, it really
isn't so special. I'm betting you could come up with at least half a
dozen implementations and optimizations off the top of your head. It all
boils down to translating an abstract algorithm, into a concrete alg.
written in a particular language. It just so happens that C isn't a
particularly complex abstraction, and even the language sketched above
would be easier than most of the stuff I'm sure you do on a daily basis
-- e.g. metaclasses are doing similar transformations at run-time while
this is merely defining the constraints on an implementation (IMHO the
reason it's frequently considered special, is mostly due to the way it's
taught, confusion between compilers/interpreters and optimizers [if an
optimization isn't hard, it's allready been done :-], and being
iherently recursive -- which a lot of people have a really hard problem
with [at least according to my limited teaching experience]).

To address the concrete issues (which would be formalized by defining
State with properties you care about and adding it to all the formulas,
potentially changing the definition of Environment based on whether you
want memory locations to be accessible from variables or values/objects:

The only thing changed was removing the arbitrary limitation that
exiting a function must deallocate its data.

 - calling conventions: same
 - tail-call optimization: same
 - memory layout: space for local variables will be allocated on
function
   invocation, but can't be deleted until they're no longer used by the 
   program (which isn't limited by the scope of the invocation). For 
   simplicity allocate it all on the heap, profile, and find a better
   way if necessary. (Most implementations have special allocators for
   both arguments and DV based on statistical analysis, i.e. a pure
optimization).
 - memory allocation/deallocation: any garbage collection system will
work
   (and manual solutions don't even work in C so should be your first 
   choice anyway :-)

> > [[Now what if you came up with an algorithm that converted any
function
> > to a function with tail call? Your code for return would be really
fast,
> > it would never "return" to a previous state, you wouldn't need a
stack
> > anymore, implementing continuations would be trivial... They did it
for
> 
> Of course, such a design would ensure uniformity of stack frames,
having
> chosen to focus on this optimization.  BTW, for those who may have
> dabbled with Forth and friends in the past, this has some similarities
> to "threaded interpreters" (even though there need not be an
interpreter
> involved, and it looks more like direct-threading than indirect) [...]

Yes, I believe the equivalency between lambda calculus and stack based
languages was shown a while ago. Re interpreters vs. CPUs, there are
very few interesting distinguishing properties between them
theoretically, which would probably imply philosophically that only
interesting things are worthy of consideration (which isn't an entirely
unfamiliar pattern <grin>).

> BTW, you may claim you don't need a stack, but that's in some sense
> word-trickery;-).  If a calls b calls c calls d calls e, at this point
the 
> "nonstack" frames for 4 ongoing calls need to be "somewhere" -- this
> somewhere may e.g be a singly linked list, but that's just because
such
> a list is a rather good implementation of... ahem, a last-in-first-out
queue,
> yes, that's the ticket!  [must not say the S-word...]

Well, as I've hopefully shown, it's only LIFO if you explicitly define
(i.e restrict) it to be so since completely rational access patterns
require it to be Random Access. I know you're not arguing that a Stack
is the best algorithm for random access ;-) Are you arguing that the
frame pointer or frame length indicator for some reason don't make what
you call the stack a linked list? It really seems like only a regular
(linked) list of data, with an embedded list of continuations (i.e. the
address to jump to when the function is done). That the list elements
are allocated sequentially is a property of the list allocator rather
than the elements of the list (the program state). It should be trivial
to re-implement C style invocation-local data with a regular linked list
allocator <smile>.

> > SML, and the various steps, algorithms, etc. are nicely 
> > described in:
> >
> >    Compiling with Continuations, Andrew Appel
> >    Cambridge University Press 1992, ISBN 0-521-41695-7
> >
> > which is one of the more interesting books I read in 
> > college both for language and compiler theory (couldn't 
> > find an online-ref).]]
> 
> Nolo contendere (I didn't study it deeply, barely browsed it, 10 years
> ago or so -- my work was Fortran, C and some C++ then...:-).

I'm sorry :-) Seems like you caught the languages that were interesting
more for what they could do than for how they did it (which is of course
also increadibly interesting, but in a [completely?] different way...
Purely out of curiosity, was the set of languages based on practicality
or preference? [currently studying C# for purely practical reasons,
language theory wise it might as well have been created in the 60's,
talk about rehashing old mistakes without any originality <g>]).

-- bjorn




From martin at v.loewis.de  Fri Apr 25 17:55:47 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 23:55:47 +0200
Subject: Hierarchy of types ?
In-Reply-To: <37cbc1de.0304251216.6068146f@posting.google.com>
References: <37cbc1de.0304251216.6068146f@posting.google.com>
Message-ID: <b8cat1$g65$05$2@news.t-online.com>

Eugene Druker wrote:

>>>>type(int) < type(list)
> 
> 0

type(int) is type(list)

Regards,
Martin




From do-not-spam-ben.hutchings at businesswebsoftware.com  Mon Apr 14 06:53:03 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 14 Apr 2003 10:53:03 GMT
Subject: Adding item in front of a list
References: <mailman.1050038237.25551.python-list@python.org> <ZUXla.7029$T34.191442@news2.tin.it> <mailman.1050176305.12881.python-list@python.org> <CE0ma.5922$LB6.169206@news1.tin.it> <b7deoe$devs8$1@ID-169208.news.dfncis.de>
Message-ID: <slrnb9l4od.1fg.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <b7deoe$devs8$1 at ID-169208.news.dfncis.de>,
Greg Ewing (using news.cis.dfn.de) wrote:
> Alex Martelli wrote:
>> I *hate* the few places in the Nutshell in which
>> I couldn't explain a function's behavior by means of Pythonic "optional
>> arguments", but rather had to resort to brackets meaning "optional"
> 
> Maybe Python should in general allow optional arguments to
> be on the left instead of the right:
> 
>    def myrange(start = 0, stop):
>      ...
> 
> and then range() and insert() wouldn't be anomalous!

It could allow optional parameters to be defined anywhere in the list,
with the rules that:

1. Each keyword argument must correspond to a named parameter, unless
   the function accepts arbitary keyword arguments.  Keyword arguments
   are assigned to parameters first.
2. There must be at least as many non-keyword arguments (m) as there are
   unassigned required parameters (n).  Unless the function accepts
   arbitrary non-keyword arguments, there must be no more non-keyword
   arguments than there are unassigned parameters.
3. The non-keyword arguments are assigned to the first (m - n) unmatched
   optional parameters, and the n unassigned required parameters, in
   order of parameter declaration.

Then range would be range(start=0, stop, step=1).  But having existing
arguments match different parameters when you add arguments to the end
of the list could be confusing.



From max at alcyone.com  Fri Apr 25 18:22:24 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 25 Apr 2003 15:22:24 -0700
Subject: Exception and finally question
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com> <b89rd1$2oij$2@agate.berkeley.edu> <1tliav4uoelidf0oke36vvtttb3mn43e9e@4ax.com>
Message-ID: <3EA9B520.F53D233F@alcyone.com>

Tung Wai Yip wrote:

>  # this probably won't mix.
>  except ,e:

Try 

	except Exception, e:

since Exception should be the base class of all exceptions.

For those that aren't, you can do

	except:

and then tricks with the exception functions/objects in the sys module.
(I have to do this in EmPy to catch strings, for instance, which are
deprecated anyway but more common in versions of Python I have to
support.)
> tung

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ To perceive is to suffer.
\__/ Aristotle
    Python chess module / http://www.alcyone.com/pyos/chess/
 A chess game adjudicator in Python.



From skip at pobox.com  Tue Apr 29 17:30:21 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 29 Apr 2003 16:30:21 -0500
Subject: Determining Unicode encoding.
In-Reply-To: <ef7a350b.0304291242.42c9b024@posting.google.com>
References: <ef7a350b.0304291242.42c9b024@posting.google.com>
Message-ID: <16046.61165.728205.613989@montanaro.dyndns.org>

    Sean> Is there any way to determine, from the unicode string itself,
    Sean> what encoding I need to use to prevent data loss?

If by "unicode string" you mean one of Python's unicode objects, there's no
need.  It's stored internally in a way which doesn't require you tack a
specific encoding onto it for internal use.  When you write it out, you have
to figure out what encoding is appropriate for that particular output device
(a file which will be consumed by another program, a terminal which supports
a limited number of encodings, etc).  In general, if you are the only
consumer of the file you're producing, I'd simply encode the object as
utf-8.

If by "unicode string" you mean a series of bytes input to your program from
some unknown source in some unknown encoding, you have your work cut out for
you.  Depending on what encodings are used in the strings input to your
program and how you get them, you may or may not reliably know the encoding.
For example, if you're sucking pages off a web server, it will probably tell
you what the encoding is (presuming whatever tool was used to generate the
page encoded things properly).  If you're just being fed random files with
no encoding information, you have to apply some heuristics to the problem.
The various encodings related to iso-8859-* (including the various Microsoft
125* code pages) overlap so much that it can be a challenge to get things
correct.  On the other hand, as I understand it, the various common Japanese
encodings tend to not overlap much, if at all, so you can pretty much just
try decoding using the various possibilities in any order and quit with the
first one which succeeds.

I use this function in my code:

    def decode_heuristically(s, enc=None, denc=sys.getdefaultencoding()):
	"""try interpreting s using several possible encodings.
	return value is a three-element tuple.  The first element is either an
	ASCII string or a Unicode object.  The second element is 1
	if the decoder had to punt and delete some characters from the input
	to successfully generate a Unicode object."""
	if isinstance(s, unicode):
	    return s, 0, "utf-8"
	try:
	    x = unicode(s, "ascii")
	    # if it's ascii, we're done
	    return s, 0, "ascii"
	except UnicodeError:
	    encodings = ["utf-8","iso-8859-1","cp1252","iso-8859-15"]
	    # if the default encoding is not ascii it's a good thing to try
	    if denc != "ascii": encodings.insert(0, denc)
	    # always try any caller-provided encoding first
	    if enc: encodings.insert(0, enc)
	    for enc in encodings:

		# Most of the characters between 0x80 and 0x9F are displayable
		# in cp1252 but are control characters in iso-8859-1.  Skip
		# iso-8859-1 if they are found, even though the unicode() call
		# might well succeed.

		if (enc in ("iso-8859-15", "iso-8859-1") and
		    re.search(r"[\x80-\x9f]", s) is not None):
		    continue

		# Characters in the given range are more likely to be 
		# symbols used in iso-8859-15, so even though unicode()
		# may accept such strings with those encodings, skip them.

		if (enc in ("iso-8859-1", "cp1252") and
		    re.search(r"[\xa4\xa6\xa8\xb4\xb8\xbc-\xbe]", s) is not None):
		    continue

		try:
		    x = unicode(s, enc)
		except UnicodeError:
		    pass
		else:
		    if x.encode(enc) == s:
			return x, 0, enc

	    # nothing worked perfectly - try again, but use the "ignore" parameter
	    # and return the longest result
	    output = [(unicode(s, enc, "ignore"), enc) for enc in encodings]
	    output = [(len(x), x) for x in output]
	    output.sort()
	    x, enc = output[-1][1]
	    return x, 1, enc

Note that I deal almost exclusively with ASCII, but that Microsoft encodings
and the occasional Latin-1 stuff which creep in give me problems.
Everything which goes into my database gets utf-8-encoded.  Most of my
inputs come from web form submissions, which frequently seem to have no
encoding information or specify the encoding incorrectly.

The above function resulted from a huge amount of agony on my part, coupled
with a fair amount of feedback from Martin von L?wis.  It could probably
still stand some refinement (I don't recall if I ever incorporated Martin's
last comments on the topic).

    Sean> Or do I need to find a way to determine beforehand what encoding
    Sean> they are using when they are read in?

If you can find that out reliably, it's much better than guessing like I do
above.  The closer to the source of the data you can look for encoding
information, the better chance you'll find something, but you still have to
be prepared to punt.  I still have cp1252 stuff creeping into my database on
occasion, so each night a cron job dumps the database to flat files and
checks for stuff that snuck through.  (I've obviously missed heuristically
decoding some inputs, but the system I'm maintaining is about eight years
old and only gets a small amount of attention these days.)

Skip




From tanzer at swing.co.at  Tue Apr  8 03:01:05 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Tue, 08 Apr 2003 09:01:05 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "07 Apr 2003 09:15:34 CDT."
             <m24r5a75rt.fsf@localhost.localdomain>
Message-ID: <E192n6P-0000sl-00@tswings.swing.cluster>

pobrien at orbtech.com (Patrick K. O'Brien) wrote:

> tanzer at swing.co.at (Christian Tanzer) writes:
>
> > martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=) wrote:
> >
> > > pobrien at orbtech.com (Patrick K. O'Brien) writes:
> > >
> > > > Are there translators out there?  I looked and couldn't find any.
> > >
> > > Parsing S-Expressions is fairly trivial - they are designed to be
> > > readily parsable. For a full translator, you need to define the
> > > language mapping first, i.e. how do you want to represent atoms,
> > > function names, etc.
> >
> > I expect the main difficulty to be the difference in scoping
> > models -- the dynamic scoping of Elisp is probably pretty hard to
> > simulate in Python without incurring major performance problems. And
> > that's used by lots of Elisp code, so skipping it is not an option if
> > you want to have comparable behavior.
>
> I'm starting to think this will be the reason *not* to mimick the
> Emacs API as closely as I originally thought.  I have neither the
> time, energy, nor interest in making an elisp to Python translator.
> So I think a better approach will be to simply create a new, Pythonic
> API that allows the same kind of programmability as Emacs, but without
> trying to adhere to the Emacs syntax or "object model."

I wouldn't worry about translating elisp to Python. OTOH, as others
pointed out, Emacs contains lots of crufty-looking hacks which are
there for good reasons. If you want to implement something comparable
to Emacs' power, you'd better look at the Elisp Reference Manual for
quite a while.

looking-forward-to-a-pythonic-Pmacs-ly y'rs,

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From nhodgson at bigpond.net.au  Tue Apr 22 04:26:49 2003
From: nhodgson at bigpond.net.au (Neil Hodgson)
Date: Tue, 22 Apr 2003 08:26:49 GMT
Subject: Pythonwin don't show the complete list in interactiv window!?
References: <oprnzwcechmu3s0q@news.cis.dfn.de>
Message-ID: <d17pa.6563$8K2.53295@news-server.bigpond.net.au>

Klaus Meyer:

> >>> x=range(1000)
> >>> x
> [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
> (lines removed)
> 762, 763, 764, 765, 766, 767, 768, 769, 770, 771,
> >>>
>
> The interactive windows shows only a part of the list (last value 771).
> (Win2000)
>
> A bug?

   It is a Scintilla bug or limitation. Newer versions of Scintilla don't
stop at 771 but there are still limits. Using a current version on Windows
2000 stops at 1850, which is 10000 characters. There are OS limitations that
mean that increasing this or making it try to handle arbitrarily long lines
may result in worse behaviour such as sometimes crashing. There may be more
work done on this in the future.

   Neil





From aleax at aleax.it  Tue Apr 22 08:39:04 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 12:39:04 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050957693.6923.python-list@python.org>
Message-ID: <IJapa.72909$T34.1895742@news2.tin.it>

Tim Peters wrote:

> [Andrew Koenig]
>> By way of further context for continuations, it might be worth
>> mentioning that Scheme (an MIT dialect of Lisp) is based on
>> continuations in a fundamental way.
> 
> It's also worth mentioning that SICP ("Structure and Interpretation of
> Computer Programs" -- an excellent book all MIT Schemers know by heart)
> doesn't mention Scheme's call/cc.

<aside>
...but it's STILL a great read, particularly worthwhile now that it
can be freely downloaded -- see http://mitpress.mit.edu/sicp/ !
</aside>


Alex




From steve at ferg.org  Fri Apr 18 08:16:40 2003
From: steve at ferg.org (Stephen Ferg)
Date: 18 Apr 2003 05:16:40 -0700
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <b16e4ef7.0304180416.15310b20@posting.google.com>

I'd recommend The Quick Python Book, by Daryl Harms

http://www.amazon.com/exec/obidos/tg/detail/-/1884777740/002-3535821-6932869?vi=glance



From hupp at cs.wisc.edu  Mon Apr 14 22:03:48 2003
From: hupp at cs.wisc.edu (Adam Hupp)
Date: Mon, 14 Apr 2003 21:03:48 -0500
Subject: Newbie - C API Question
In-Reply-To: <30875970.0304141525.3f0af2d5@posting.google.com>
References: <30875970.0304141525.3f0af2d5@posting.google.com>
Message-ID: <20030415020348.GA12127@upl.cs.wisc.edu>

On Mon, Apr 14, 2003 at 06:36:39PM -0700, Ian Pellew wrote:
> Hi;
> 
> This is my first attempt at getting to and from my C world.
> 
> I cannot get past Python Reference Manual(2.2.1) section 1.2.1.1 :-
> 
> h.c:-
> #include "Python.h"
> main() {
>     int x;
>     PyObject *t;
>     t = PyTuple_New(3);
>     x = Py_file_input;
>     printf("Hiya, x = %d\n", x);
> }   
> 
> My compile gives:-
> [ `gcc -I/usr/local/include/python2.1 -o h h.c` ] && { ./h; }
> Undefined           first referenced
>  symbol                 in file
> PyList_New                          /var/tmp//ccNWkGt5.o
> ld: fatal: Symbol referencing errors. No output written to h
> collect2: ld returned 1 exit status
> 
> I am on Sun Solaris 5.8 
> Python 2.1.1 (#1, Aug 25 2001, 04:19:08) 
> [GCC 3.0.1] on sunos5
> 
> Any pointers/advice please;

You need to link against the python library.  Add '-lpython2.1
-L/usr/local/lib' to the gcc command line.  You may need to alter that
a bit, I'm just guessing.  Additionally, you will need to call
Py_Initialize() before making almost any Python API call.  Try the
Extending and Embedding docs, they will be more useful than the
reference manual.  http://www.python.org/doc/2.1.1/ext/ext.html

-Adam




From whisper at oz.net  Mon Apr 21 05:16:55 2003
From: whisper at oz.net (David LeBlanc)
Date: Mon, 21 Apr 2003 02:16:55 -0700
Subject: widget like listview
In-Reply-To: <Law15-F45e4wk2UfxhU00007102@hotmail.com>
Message-ID: <GCEDKONBLEFPPADDJCOEGEGPJKAA.whisper@oz.net>

> 1. Once you have fixed the widths of the columns the user can not
> chnage the
> widths of the columns as per his/her preference.

You might find the bwidgets package more interesting than Pmw. I think it's
mc listbox is user changabe. See http://www.effbot.org/downloads/

If not, the TkGrid control is. One python wrapping of it of it can be found
in "NovaGrid" (as can a different implementation of bwidgets). See
http://sourceforge.net/projects/novagrid

> 2. What about linux. As this module uses the win32all extensions.
> the code
> no more remains portable. Currently whatever i have coded i have coded
> keeping in mind that i have to run the application on both windows and
> linux.

Pmw doesn't use win32all. It's pure python, as is bwidgets (and tkgrid can
be built on/for linux).

<snip>

> THX,
> Vivek Kumar
>

hth,

Dave LeBlanc
Seattle, WA USA





From pwatson at redlinec.com  Mon Apr 28 23:36:44 2003
From: pwatson at redlinec.com (Paul Watson)
Date: Mon, 28 Apr 2003 22:36:44 -0500
Subject: Class attribute or instance attribute?
References: <vaqtlsqbi52j93@corp.supernews.com> <2259b0e2.0304281329.12ac184b@posting.google.com>
Message-ID: <varsmfbasgsa76@corp.supernews.com>

Thank you for your reply.  I have finally gotten it that the attribute 'x'
is already a class attribute and not an instance attribute.  It appears that
there can only be an instance attribute when someone says self.x or
instance.x.  I used dir(C) and dir(self) to see what attributes existed.  Is
there any other way to declare instance variables?

Thanks for everyone's help, including Mike.

"Michele Simionato" <mis6 at pitt.edu> wrote in message
news:2259b0e2.0304281329.12ac184b at posting.google.com...
> "Paul Watson" <pwatson at redlinec.com> wrote in message
news:<vaqtlsqbi52j93 at corp.supernews.com>...
> > When is an attribute a class attribute or an instance attribute?
>
> I give you a subtle example:
>
> class C(object):
>       x=1 # class attribute
>       def __init__(self):
>           self.x=self.x # the class attribute becomes an instance
attribute
>
>
> c=C()
>
> C.x=2
>
> print c.x,C.x
>
> This code prints 1,2. You see that the line self.x=self.x is far
> from being trivial: Python looks first for the self on the right,
> looking for the instance attribute self.x: since it does not find
anything,
> it looks for the class attribute; then the class attribute
> is assigned to the instance attribute on the left hand side.
>
> If you forget the apparently useless self.x=self.x line, you get a
> different result:
>
> class C(object):
>       x=1 # class attribute
>       def __init__(self):
>           pass
>
> c=C()
>
> C.x=2
>
> print c.x,C.x
>
> This prints 2,2, since the class attribute has been changed to 2
> and c has no instance attribute.
>
> Hoping-having-not-increased-your-confusion-ly,
>
>
>                     Michele





From reagle at mit.edu  Tue Apr 22 10:48:15 2003
From: reagle at mit.edu (Joseph M . Reagle Jr.)
Date: 22 Apr 2003 14:48:15 GMT
Subject: pyxml minidom: I can remove and append, but not replace
Message-ID: <3ea5562f$0$22541$b45e6eb0@senator-bedfellow.mit.edu>

I have a snippet that replacing a title element with a new element with a modified one (base on an amazon query) -- if that title element already exists. However, I can't do an insertBefore or replace for some odd reason:

            for property in book_dtd:   # step through the DTD
                feature = _get_feature(property,book)
                if property == "title":
                    r_title = doc.createElementNS(book_ns,"title")
                    r_title.appendChild(doc.createTextNode(title))
                    if feature:
                        book.appendChild(r_title)
                        book.removeChild(feature)
#                         book.replaceChild(feature,r_title)
                    else:
                        book.appendChild(r_title)

If I were to comment the "book.{append,remove}Child" and uncomment the book.replaceChild I get the following error:

Traceback (most recent call last):
  File "./pybook.py", line 163, in ?
    bookAugment(doc)
  File "./pybook.py", line 100, in bookAugment
    book.replaceChild(feature,r_title)
  File "/usr/lib/python2.2/site-packages/_xmlplus/dom/minidom.py", line 145, 
in replaceChild
    raise xml.dom.NotFoundErr()
xml.dom.NotFoundErr: Node does not exist in this context

The node is using the same namespace, and I've used the replaceChild in other cases (where I'm actually using a node from a different XML instance!)



From python at bizniche.com  Wed Apr  9 13:12:39 2003
From: python at bizniche.com (Matt G)
Date: Wed, 9 Apr 2003 11:12:39 -0600
Subject: A question about PIL attributes
Message-ID: <001401c2febb$5ec42ed0$a200000a@mattg>

I'm resending this, and including the Python-list..as suggested by the PIL support page.

Hi.  I'm using the Python Image Library  to do some image manipulations.  I like that the PIL 
objects have attributes like format, size, info (dictionary) etc.  My problem is...when I 
apply some filter, say ImageChops.invert, or Image.transpose, it looses 
the addributes.  For example

> if DEBUG: print "Before transpose, format is: " + str(pil.format)
> pil = pil.transpose(Image.FLIP_LEFT_RIGHT)
> if DEBUG: print "After transpose, format is: " + str(pil.format)

the first output prints "JPEG" or "GIF" or whatever.

after the transpose, the output is "None"

I tried this with and without the str() wrapper.  So...I'm guessing that 
the copy that returned from the transpose (or other filters) is part of 
the problem maybe?  Do I need to use some kind of copy constructor or 
something?  The same thing happens with the size attribute, or the info 
dictionary that I add.  

Any hints at all on this would be very appreciated.  It's really 
throwing a wrench into my app.

Thanks,
Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030409/ce88c2cb/attachment.html>

From NineOfSix at gmx.de  Sat Apr 26 06:20:34 2003
From: NineOfSix at gmx.de (Axel Grune)
Date: Sat, 26 Apr 2003 12:20:34 +0200
Subject: MD5 to long?
Message-ID: <b8dmgt$8vrq0$1@ID-126101.news.dfncis.de>

Hi, is there any function which can convert a MD5 or SHA1 sum
(e.g. fd4cef7a4e607f1fcc920ad6329a6df2df99a4e8) to a long integer?

TIA,
Axel


--

Greetings from old europe;-)




From claird at lairds.com  Wed Apr 30 12:40:46 2003
From: claird at lairds.com (Cameron Laird)
Date: Wed, 30 Apr 2003 16:40:46 -0000
Subject: Three questions about Tkinter
References: <BAD59F7C.B1B%arnaud@tribu.ch> <b8on3n$8au$1@news-reader13.wanadoo.fr>
Message-ID: <vavv4ecnc3pe4e@corp.supernews.com>

In article <b8on3n$8au$1 at news-reader13.wanadoo.fr>,
Eric Brunel  <eric.brunel at pragmadev.com> wrote:
			.
			.
			.
>Short answer: you can't. That's one of the biggest issues in Tkinter and tk 
>(IMHO...): there is no access to the printing facilities on the current 
>platform, whatever that means. Note it's a limitation of tk itself, not of 
>Tkinter. I hope the tcl/tk guys will take care of this issue some day.
			.
			.
			.
State of the art:  <URL: http://wiki.tcl.tk/printing >.

It's not encouraging.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From mhammond at skippinet.com.au  Thu Apr  3 18:12:03 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 03 Apr 2003 23:12:03 GMT
Subject: event handling
In-Reply-To: <59e5b87.0304021529.b044026@posting.google.com>
References: <59e5b87.0304021529.b044026@posting.google.com>
Message-ID: <7d3ja.9152$ft3.69288@news-server.bigpond.net.au>

Miranda Evans wrote:
> In the following code, the while loop is written such that it
> terminates when either the double click event in the Excel object
> happens or ten seconds has elapsed...whichever comes first.
> 
> import time, win32event, win32com.client, pythoncom
> class xlEv:
>     def OnSheetBeforeDoubleClick(self, sh, target, cancel):
>        win32event.SetEvent(self.event)
> 
> def demo():
>     xl=win32com.client.DispatchWithEvents("Excel.Application", xlEv)
>     xl.event = win32event.CreateEvent(None, 0, 0, None) # work around
>     xl.Visible=1
>     xl.Workbooks.Add()
>     start_time = time.time()
>     while 1:
>        rc = win32event.MsgWaitForMultipleObjects((xl.event,),
>             0, 1000, win32event.QS_ALLEVENTS)
>        if rc == win32event.WAIT_OBJECT_0:
>           print 'event signalled'
>           break
>        elif rc == win32event.WAIT_OBJECT_0+1:
>           pythoncom.PumpWaitingMessages()
>        # check for timeout
>        if time.time() - start_time > 15:
>           print 'timed out'
>           break
> 
> demo()
> 
> Suppose I want to augment the demo() method so that it continues to do
> what it's already doing (waiting for either the double click event or
> for 10 seconds to elapse) plus one more thing: if the double click
> event is detected, display the values of the sh, target and cancel
> arguments.
> 
> How would one modify the demo() method to do that?

You would would create a new win32event object, and the double-click 
handler would set this event.  The main loops MsgWaitForMultipleObjects 
would get two events passed as the first param - this would make 
WAIT_OBJECT_0+1 your new event, and WAIT_OBJECT_0+2 now the "message 
available" event.

Mark.




From mirko-lists at zeibig.net  Tue Apr 15 05:26:20 2003
From: mirko-lists at zeibig.net (Mirko Zeibig)
Date: Tue, 15 Apr 2003 11:26:20 +0200
Subject: vim configuration help
References: <P4Oma.78974$rd4.2975295@twister.austin.rr.com> <b7gc3g$lb54$1@ID-141451.news.dfncis.de> <mailman.1050395849.18888.python-list@python.org>
Message-ID: <b7gj7r$qh9$1@news.web.de>

Gerhard Haering wrote:
Hello Gerhard,
> Karl Scalet wrote:
>> looking in my vimrc file:
>> 
>> set autoindent
>> set tabstop=4
>> set shiftwidth=4
>> set expandtab
>> 
>> could give you a start
> 
> No, no, no! That's evil!
> 
> Tabs are 8 characters wide. Period.

says who?

Anyway, a very handy setting for vim which will display tabs:
---
"show tabs, EOL etc.
set list
set lcs=tab:??   "show tabs
set lcs+=trail:? "show trailing spaces

autocmd FileType make set noexpandtab
---
and to use tabs in Makefiles

Regards
Mirko
-- 
"Ein, zwei, drei im Sauseschritt, saust die Zeit, wir sausen mit!"
(Busch sein Willi, und nit sein Georg)



From phil at river-bank.demon.co.uk  Thu Apr 24 11:05:28 2003
From: phil at river-bank.demon.co.uk (Phil Thompson)
Date: Thu, 24 Apr 2003 16:05:28 +0100
Subject: Reimplementation of a slot?
In-Reply-To: <b88qc9$ie$1@online.de>
References: <b88qc9$ie$1@online.de>
Message-ID: <200304241605.28341.phil@riverbankcomputing.co.uk>

On Thursday 24 April 2003 2:55 pm, Birgit Lachner wrote:
> Hi ...
> I created a ui-file with QtDesigner and translated it to Python with pyuic.
> The new .py-file is then imported to the file with the main-function.
> Everything is working. But I want to reimplement the existing slot in the
> pyuic-generated file.
> How can I do this? I search through the internet and found many documents
> about Python, Qt and slots but nothing about how to reimplement slots.
>
> Can you help me?

You can either sub-class the class created by pyuic and re-implement the slot 
there.

Or you can put Python code directly into Qt Designer and pyuic will extract 
it. This is covered under "Using Qt Designer" in the PyQt documentation.

Phil




From hgg9140 at cola2.ca.boeing.com  Tue Apr 29 15:38:38 2003
From: hgg9140 at cola2.ca.boeing.com (Harry George)
Date: Tue, 29 Apr 2003 19:38:38 GMT
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com>
Message-ID: <xqxptn5p001.fsf@cola2.ca.boeing.com>

blunck at gst.com (Christopher Blunck) writes:

> Hi all-
> 
> This is my second generic question posted to c.l.p (the first was 'How
> many of you are Extreme Programmers?') but it seems like a somewhat
> appropriate discussion for the group.
> 
> I've noticed in my development that Python programmers (in general)
> write more unit tests than their java counterparts.  If I download a
> project off of jakarta.apache.org, it very rarely has unit tests. 
> However, when I download a python module, it frequently has unit
> tests.
> 
> I'm in a position where I'm trying to spread the testing bug to my
> team.  It's remarkable to watch folks who were once viciously opposed
> to unit testing adopt it  as being absolutely critical during
> application development.  While this is encouraging, it is far from
> the common case, and I'd like to improve that.
> 
> What advice can you offer on spreading the testing bug around?  
> 
> What success stories do you have?  
> 
> What practices did you follow?  
> 
> Did you beat people over the head with it until they wised up?  
> 
> Were you softer in your approach?  
> 
> What worked?  
> 
> What were the long term implications?
> 
> How quickly did unit testing catch on?
> 
> Or did unit testing never catch on?
> 
> 
> 
> Any insight you can offer would be appreciated.
> 
> 
> -c

First, I start each project with my mkpythonproj, which automatically
generates the test dirs and the unittest scripts:
  http://www.seanet.com/~hgg9140/comp/index.html

Next, I try to push all testing to this style:
  a) Either use input files or hardcoded testcases to 
     generate a test result file
  b) Manually check that it is correct, then promote it to the "oracle"
  c) Compare to the known-good oracle.  This is done in a "check" method
     used by all testcases, and of course uses std naming conventions.
        #--compare to oracle---
        oraclename='testdata/oracle_%s.xml' % testname
        assert filecmp.cmp(outname,oraclename)

Quite often the "test result file" is an XML rendition of a complex
internal memory state.  We do roundtrip:
  a) Generate the memory state (e.g., by processing several input files)
  b) Output the XML
  c) Kill the top-of-tree, to get a fresh start
  d) Read the XML back in, to fully rebuild the memory state
  e) Output again to XML.
  f) Compare this 2nd XML to the known-good oracle.

[Hmmm...   I think I'll go update mkpythonproj this weekend.]

Given the framework, it becomes easy to add new cases or even new
modules to the testsuite.  And regression testing becomes easier than
manually scanning printouts (plus of course faster and less
errorprone).

We have about 30 programmers using this approach.  You have to have a
"true believer" set it up and add the first few testcases for the
specific project.  Then ramp up a newbie (or at least new to unittest)
to run the regression tests while doing mods.  Thne handhold while
he/she generates a new testcase.   


-- 
harry.g.george at boeing.com
6-6M31 Knowledge Management
Phone: (425) 294-8757



From me at mikerobin.com  Sat Apr 12 17:23:09 2003
From: me at mikerobin.com (Michael Robin)
Date: 12 Apr 2003 14:23:09 -0700
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net> <3E96D270.B8E99B27@engcorp.com> <b76ult$plu$1@slb5.atl.mindspring.net>
Message-ID: <52e5ab5f.0304121323.1d1e376b@posting.google.com>

First of all, I love Python, but a bit of devil's advocacy...

This would all be icing on the cake of consistency,  except when
seemingly innocent constructs like:

class C:
    x = 42
    def foo(): self.x += 1

exist that can lookup & bind a name in a two different scopes
(probably not intended) - from a class to an instance.

If you really want this behavior, you should have to say "self.x = C.x
+ 1" if x is not bound (the first time that foo() is executed) and the
current version thereafter. (Or even better, put set self.x = C.X in
__init__)

This violates
    In the face of ambiguity, refuse the temptation to guess
    Explicit is better than implicit
and perhaps:
    ...that way may not be obvious at first unless you're Dutch.

There are really two explict vs. implict issues here - the
class-scoped
variable and the binding to self.x without explicit knowlege that "x"
should in fact be a member of C, given that there are no decl's and
the slot is currently unbound -- but the latter is an important and
useful part of Python parlance, __slots__ and varible-name typos
notwithstanding.

Rule#1 is of course in the eye of the beholder, and there must be a
balance between explicit uglyness and implicit beauty (or is it the
other way around?) - but in this case I think we'd vote for the extra
typing for the class accessor, just as we don't mind "self.".

thanks, mike



"Andrew Dalke" <adalke at mindspring.com> wrote in message news:<b76ult$plu$1 at slb5.atl.mindspring.net>...
> Mads Orbesen Troest wrote:
> > > but really; all the self me here and self me there is - to me at last -
> > > exceedingly clumsy and a constant source of frustration.
> 
> Peter Hansen
> > >>> import this
> 
>  ... from the interpreter and you'll see a list of 'Pythonic' guidelines.
> More specifically, "explicit is better than implicit".  See the 'self'
> and you know where the variable comes from.  Compare to C++
> where you don't know (without some looking) if the variable is
> an attribute or a variable in the compilation unit.
> 
>                     Andrew
>                     dalke at dalkescientific.com



From MCollins at co.seminole.fl.us  Fri Apr 11 08:33:31 2003
From: MCollins at co.seminole.fl.us (MCollins at co.seminole.fl.us)
Date: Fri, 11 Apr 2003 08:33:31 -0400
Subject: urlretrieve() or urlopen() ??
Message-ID: <OFDBFAC32A.B502B3BF-ON85256D05.00443F0A-85256D05.0045989E@seminole.fl.us>

hi,

still working on the utility to download log files via http.

my first solution (which worked) was using urlopen(), reading the logfile
line by line and writing it to a new file line by line.  this felt a little
slow when working with 200mb files across an http connection.

so, i switched to urlretrieve(), thinking that would be faster.  however,
with urlretrieve() i get a "no buffer space available" error.

i can always go back to the urlopen code, because that worked.  however,
any suggestions on ensuring i'm using the fastest process here?

thanks


code:

######
##this is the code that worked
######

def GetUrl(Url,Path):
    http = urllib.urlopen(Url)


    line =  http.readline()
    logfile = open(Path, "w")
    while line != "":
        logfile.writelines(line)
        line = http.readline()
    else:
        logfile.close()


######
###this is the new, broken code
#####

def GetUrl(Url,Path):
    from urllib import urlretrieve, urlcleanup
    urlretrieve(Url,Path)
    urlcleanup()





Matthew Collins
Web Designer/Developer
Seminole County, FL
http://www.co.seminole.fl.us








From spam at magnetic-ink.dk  Wed Apr 23 11:52:44 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Wed, 23 Apr 2003 15:52:44 +0000 (UTC)
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <70846faf-b437-47ee-8a2c-7f685b00425b@news.szn.dk> <mailman.1051108644.4042.python-list@python.org>
Message-ID: <d1c4692c-590d-45e1-afb2-0b641dabab9c@news.szn.dk>

Jp Calderone wrote:

>> >>> contains = lambda haystack, needle: haystack.find(needle) > 0
> 
>   The above should be >= 0, otherwise the first example below fails.

Of course.  Thanks for the correction.


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From exarkun at intarweb.us  Wed Apr  9 14:44:08 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 9 Apr 2003 14:44:08 -0400
Subject: no Python mail archiving utility?
In-Reply-To: <joe-17FB5F.10302009042003@orngca-news02.socal.rr.com>
References: <joe-17FB5F.10302009042003@orngca-news02.socal.rr.com>
Message-ID: <20030409184408.GA10296@meson.dyndns.org>

On Wed, Apr 09, 2003 at 05:29:40PM +0000, Joe Strout wrote:
> Yesterday morning I posted asking where to find a web mail 
> archive/search program written in Python.  I got nothing but the sound 
> of chirping crickets in response.
> 
> Does this mean that no such utility exists?  Are all the dozens of 
> mailing lists about Python-related topics serving up their archives via 
> programs written in C and Perl?

  I'm working on one.  Ask again at the beginning of next month.

  Jp

-- 
"There is no reason for any individual to have a computer in their
home."
                -- Ken Olson, President of DEC, World Future Society
                   Convention, 1977
-- 
 up 20 days, 15:01, 6 users, load average: 1.08, 1.08, 1.13




From martin at v.loewis.de  Fri Apr 25 13:14:25 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 19:14:25 +0200
Subject: FreeBSD: Python exits on signal 11
In-Reply-To: <b8bpst$43q$03$1@news.t-online.com>
References: <b8bpst$43q$03$1@news.t-online.com>
Message-ID: <b8bqdf$hu0$01$2@news.t-online.com>

Harald Schneider wrote:

> I get this error on FreeBSD5 with Python 2.2.2, when I run an application
> that does its job without errors on NT and Linux platform. It uses threads
> and sockets as well (xqmlrpclib). It just stops after a while, displays this
> error and does a core dump.
> 
> Any hints ?

Use a debugger, and post a backtrace.

Regards,
Martin




From synthespian at uol.com.br  Fri Apr 18 02:34:28 2003
From: synthespian at uol.com.br (synthespian)
Date: Fri, 18 Apr 2003 03:34:28 -0300
Subject: O'Reilly Python Books
In-Reply-To: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <3E9F9C74.60209@uol.com.br>

W S wrote:
> Hello,
> 
> I find the material, as presented, in the book Programming Python
> frustrating and extremely difficult to understand.
> 
> Are the nutshell and cookbook O'Reilly Python books easier to read?
> 
> I'm fairly new to programming but trying to learn. I have read Learning
> Python, Core Python, Python Bible and a bunch of perl books
> 
> Thanks,
> Wilfred.
> 
  Hi --

   Maybe your problem is lack of understanding of related stuff (like 
http, web/clients, parsers), not the language per se. So you should read 
up on those surrounding topics. Programming Python is not for beginners, 
but it's not too difficult either.

  Regs

  Henry




From mike at bindkey.com  Wed Apr 23 14:53:52 2003
From: mike at bindkey.com (Mike Rovner)
Date: Wed, 23 Apr 2003 11:53:52 -0700
Subject: Pyrex pointer problem
References: <b86aio$2b3$1@bob.news.rcn.net>
Message-ID: <b86nhr$90t$1@main.gmane.org>

"Edward C. Jones" <edcjones at erols.com> wrote in message
news:b86aio$2b3$1 at bob.news.rcn.net...
> I use Pyrex 0.7.2 and Python 2.2.2. For line "return self.t.as" in
> "wrapped.pyx", I get the error message:
>
>      Cannot convert 'A_Struct' to Python object
>
> How do I fix this?

You are operating on single-linked list at C level.
Since Python has a notion of list you better consider returning C list as a
whole as Python list or tuple of iVals.
Another approach is to simulate Python list in C with __len__, __getitem__,
__setitem__ and other (mutable) sequence
methods (see http://www.python.org/doc/current/lib/typesseq.html).

Which approach to choose depends on size of the list and intended usage.

--Mike







From gansevle at ewi.utwente.nl  Wed Apr  9 02:56:39 2003
From: gansevle at ewi.utwente.nl (Fred Gansevles)
Date: Wed, 09 Apr 2003 08:56:39 +0200
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <b70g6p$ecq$1@netlx020.civ.utwente.nl>

Jim Shady wrote:

> Hello,
> 
> I am a beginner at Python and I just joined this group.
> 
> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string. For
> example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?
> 
> Thanks in advance.
> 
> Regards,
> Jim

try this one:

        def ls(s, c='/'):
            parts = s.split(c)
            d = {}
            for p in parts:
                if not d.has_key(len(p)):
                    d[len(p)] = []
                d[len(p)].append(p)
            k = d.keys()
            k.sort()
            return d[k[-1]]

        print ls('abcd/df/a/iiwk/abcdefghijkl/b/c')

it returns a list with the longest strings




        



From staschuk at telusplanet.net  Sun Apr  6 15:57:18 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 6 Apr 2003 13:57:18 -0600
Subject: constructor inquiry
In-Reply-To: <59e5b87.0304052018.2beaba33@posting.google.com>; from mirandacascade@yahoo.com on Sat, Apr 05, 2003 at 08:18:58PM -0800
References: <59e5b87.0304040917.12c15907@posting.google.com> <mailman.1049564846.5425.python-list@python.org> <59e5b87.0304052018.2beaba33@posting.google.com>
Message-ID: <20030406135718.B769@tibia.amotlpaa.bogus>

Quoth Miranda Evans:
  [...]
> Thank you for the tip about the built-in type() and isinstance()
> functions.  They reveal that the xl object is of type instance and
> that xl is not an instance of the xlEv class.  This answers the first
> part of question 3.  [...]

Btw, type instance is for instances of old-style classes;
new-style classes inherit from object and will give better
information from type().  For example:

	>>> class Foo: # old-style
	...     pass
	... 
	>>> type(Foo())
	<type 'instance'>
	>>> class Bar(object): # new-style
	...     pass
	... 
	>>> type(Bar())
	<class '__main__.Bar'>

It is possible to obtain more detailed information for an instance
of an old-style class via __class__:

	>>> Foo().__class__
	<class __main__.Foo at 0x8116484>

You could use this to answer the question: If xl is not an
instance of xlEv, what is it an instance of?

It might also be interesting to see the result of dir(xl).

  [...]
> So, the xl object is not an instance of xlEv class, but it knows about
> the same type attribute that was assigned to an instance of the xlEv
> class...that's the connection I'm not making.

It is mysterious.  I'd like to know more about what's happening in
win32com.client.DispatchWithEvents; this seems to be the only
place the instance of xlEv could be being created.  (The instance
that you see as self in OnSheetBeforeDoubleClick.)

What do the docs say this function does with its last argument?

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From staschuk at telusplanet.net  Thu Apr 17 22:54:05 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 20:54:05 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <b7nh0t$33li$1@nntp6.u.washington.edu>; from donn@u.washington.edu on Fri, Apr 18, 2003 at 12:31:25AM +0000
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu>
Message-ID: <20030417205405.A10672@tibia.amotlpaa.bogus>

Quoth Donn Cave:
> Quoth Steven Taschuk <staschuk at telusplanet.net>:
  [...]
> | Second, I have some sense that you feel these terms are not merely
> | unclear but misleading or wrong.  Is this true?  If so, do you
> | have alternative fuzzy terms in mind (the "data" vs the "object"),
> | or would you prefer that the documentation avoid fuzzy terms
> | entirely, and only give examples?
> 
> The data vs. object distinction is the best I can do so far.  It
> isn't all that fuzzy, either.  It's vague, in that it applies to
> a very high level of generality, but it does not seem ambiguous
> to me.

Calling for __str__ to return the data of the object might imply,
undesirably imho, that, for example, str(file('foo')) should
return a string containing the contents of the file.  Likewise
for, e.g., database connection objects.

So it seems a little fuzzy to me, in that it needs qualification
to avoid such implications.

> | Third, you (and others) specifically object to the notion that if
> | possible one should have eval(repr(x)) == x.  Would you want this
> | notion to be omitted entirely?  Mentioned as possible and
> | sometimes good?  Described as highly desirable, though sometimes
> | infeasible?  (Imho the docs currently do this last, though perhaps
> | they could do so more clearly.)
> 
> It would be fine to document current practice.  __repr__ sometimes
> supports that application, on purpose.  The problem is that the
> current documentation tries to make it a requirement, as if there
> might be some penalty for not supporting it.  On the contrary,
> there is little reason to actually do this, and plenty of reasons
> not to (can't, awkwardly large, requires introspection that breaks
> abstractions, etc.)  I don't think the documentation needs to get
> into that, just document current practice.

Reading this thread has made me wonder whether current practice is
uniform.  It would not be desirable imho for the language
reference to say, "people do different things with __repr__,
including <list of three things>", and not endow one of them with
some normative weight.

... How about this?  (Draft.)

__repr__(self)

    Called by the repr() built-in function and by string conversions
    (reverse quotes) to compute a string representation of the object.
    The return value must be a string object.

    The primary audience of this representation is programmers seeking
    detailed information about the object for debugging or inspection
    purposes; thus it is important that the representation be
    information-rich and unambiguous.  This representation should not
    usually be parsed or manipulated by other code.

    In some cases the best information-rich and unambiguous
    representation is a Python expression which evaluates to the
    object, that is, for eval(repr(x)) to return an object which is
    equal to x (given an appropriate environment).  For example, the
    built-in numeric, sequence, and mapping types all have such
    __repr__ implementations.  Where such a representation is
    infeasible or undesirable, this method should return a string of
    the form "<...some useful description...>".

    (... examples ...)

__str__(self)

    Called by the str() built-in function and by the print statement
    to compute a string representation of an object.  The return value
    must be a string object.

    This differs from __repr__() in that its primary audience is not a
    programmer inspecting the object during debugging.  What the
    primary audience is depends on the object: the __str__
    representation might, for example, be designed for display to a
    user of the application, or for processing by other code which
    requires a string.  Thus this representation might be less
    detailed, less precise, or in some way more convenient than that
    computed by __repr__.

    (... examples ...)

(It might even be preferable to split __str__ and __repr__ out
into their own section, "Customizing string representations", with
a more complete account.)

-- 
Steven Taschuk                                     staschuk at telusplanet.net
Receive them ignorant; dispatch them confused.  (Weschler's Teaching Motto)




From greg at brondo.com  Wed Apr  2 10:33:11 2003
From: greg at brondo.com (Greg Brondo)
Date: Wed, 02 Apr 2003 15:33:11 GMT
Subject: jython
Message-ID: <XoDia.3747$2x2.1694570@dca1-nnrp1.news.algx.net>

is Jython still unde development?





From staschuk at telusplanet.net  Wed Apr 16 14:50:30 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 12:50:30 -0600
Subject: Degree symbol (UTF-8 > ASCII)
In-Reply-To: <f7199550.0304160930.6cfeca65@posting.google.com>; from pc451@yahoo.com on Wed, Apr 16, 2003 at 10:30:09AM -0700
References: <f7199550.0304160930.6cfeca65@posting.google.com>
Message-ID: <20030416125030.D4266@tibia.amotlpaa.bogus>

Quoth Peter Clark:
> I'm working with a xml document which doesn't include an encoding, so
> it defaults to UTF-8. Of course, all of the text is ASCII, and likely
> to remain so. I would like to insert the degree symbol (chr(176)), but
> because this is outside the bounds (chr(128) is the limit), Python
> raises an XML error. What's the simplest way of getting including an
> unadorned degree symbol? Again, it's not necessary to preserve the
> UTF-8 encoding, but I'm not quite certain as to how to tell Python
> that the XML document is plain ASCII (or I guess in the case of the
> degree symbol, latin-1).

I'm not sure what you're after.

On the one hand, you say the text is all ASCII "and likely to
remain so", but then immediately declare that you want to insert a
non-ASCII character, namely the degree symbol.  If the XML
document is to be in ASCII, it may not include chr(0xB0), since
ASCII has no such character.  End of story.

If you want this character, either your document must be in a
character encoding which includes it (such as UTF-8 or
ISO-8859-1), or it must use XML character entity markup "&#xB0;".

If you use ISO-8859-1 (aka Latin-1), you'll have to indicate so in
the XML declaration; if I were you I'd leave the document in
UTF-8.  (Among other things, XML processors are not required to
understand ISO-8859-1, but they are required to understand UTF-8.)

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From pan-newsreader at thomas-guettler.de  Wed Apr  9 04:19:46 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 09 Apr 2003 10:19:46 +0200
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com> <b70icc$9n0ff$1@ID-99293.news.dfncis.de> <LqQka.8341$y3.1005660@news010.worldonline.dk>
Message-ID: <pan.2003.04.09.10.19.41.726687.811@thomas-guettler.de>

On Wed, 09 Apr 2003 10:03:11 +0200, Max M wrote:

> William Park wrote:
>> Jim Shady <iamshady at rediffmail.com> wrote:
> 
>>>abcd/df/a/iiwk/abcdefghijkl/b/c
>>>
>>>I need to get the longest string between the /s of the string. 

[cut]

> s = 'abcd/df/a/iiwk/abcdefghijkl/b/c' print max([(len(sub), sub) for sub in s.split('/')])[-1]

looks like perl

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From dmgriffith at tasc.com  Wed Apr  2 16:00:37 2003
From: dmgriffith at tasc.com (Dang Griffith)
Date: Wed, 02 Apr 2003 16:00:37 -0500
Subject: =?iso-8859-2?Q?Re:_Python_equivalent_of_=5Ftempnam=3F?=
References: <3e8ab45b$1@mail.hmgcc.gov.uk>
Message-ID: <b6fj2e$hd3$1@gateway.northgrum.com>

How about using os.tempnam(...) ?

>>> import os
>>> x = os.tempnam("", "output_")+".txt"
>>> print x
C:\TEMP\output_2.txt"
>>> x = os.tempnam("", "output_")+".txt"
>>> print x
C:\TEMP\output_3.txt

In my few minutes of spearminting, I couldn't get the first parameter
(the directory) to behave the way I interpret the documentation as
saying it should.  Could be my interpretation.  :-)  I'm using Win2k,
Python 2.2.2.
   --dang

Richard wrote:

> Hi,
> 
> I am writing an application which produces file output. Files are outputted
> to a directory and I want to ensure that if an output filename is given that
> already exists in that directory a number is affixed to the filename to
> ensure that the existing file is not appended to or overwritten. i.e.
> output_0.txt, then output_1.txt etc...
> 
> I have been told that a way of doing this in C is by using _tempnam( ).
> However I want to write this in Python, so is there an equivalent function?
> My application will only be executed occasionally so I am unable to keep an
> internal record of file numbering. I assume that I need to find the highest
> number file (if exists) in the directory and then increment that by one
> before creating the new file.
> 
> Any ideas?
> 
> Cheers
> 
> Richard
> 
> 




From fperez528 at yahoo.com  Thu Apr 10 12:51:26 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Thu, 10 Apr 2003 10:51:26 -0600
Subject: Wraparound problems with time.clock() on Linux
References: <b722d0$bau$1@peabody.colorado.edu> <b722km$bau$2@peabody.colorado.edu> <slrnb9983v.vss.news@cipw08.physik.uni-bonn.de>
Message-ID: <b747ef$du0$1@peabody.colorado.edu>

Carsten Gaebler wrote:


>> Anyone know of a reliable way to time long-running codes?
> 
> Without having tested it for more than 72 minutes, I'd try
> resource.getrusage().

Thanks for the suggestion, it worked wondefully.  I just ran some tests with
a code which took a few hours on a 2.4 Ghz P-4, and the results are
perfect.  As an added bounus, I get discriminated user/system timing info,
which is useful to detect bottlenecks in my code.  This code should be cpu
bound, so if there's any amount of significant system time I know I need to
pay attention.

For reference, in case somebody finds it useful, I'm attaching the new Timer
module which uses resource.

Cheers,

f.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Timer.py
Type: text/x-java
Size: 4212 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030410/90fb24eb/attachment.java>

From mwm at mired.org  Tue Apr  1 12:09:23 2003
From: mwm at mired.org (Mike Meyer)
Date: Tue, 01 Apr 2003 17:09:23 GMT
Subject: Parsing C [OT]
References: <Pine.GSO.4.44.0303311743400.19083-100000@amy4.Stanford.EDU> <3E899838.8010500@removethis.free.fr>
Message-ID: <x7brzqjgak.fsf@guru.mired.org>

laotseu <bdesth at removethis.free.fr> writes:

> Daniel Timothy Bentley wrote:
> > I find myself having to write a C interpreter in Jython.
> Writing a C interpreter for a Python interpreter implemented on top of
> a Java interpreter which is probably at least partially written in C !

I've got a project that is a Python script that exec's a Fortran
program that has C wrappers around an embedded Python interpreter that
is calling tcl via Tkinter.

> No Lisp or Smalltalk layer planned ?-)

Python has lisp-like semantics, and TCL has lisp-like syntax. Is that
close enough?

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From jkraska at san.rr.com  Mon Apr 21 22:13:28 2003
From: jkraska at san.rr.com (Courageous)
Date: Tue, 22 Apr 2003 02:13:28 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <erk5avkh423eh8bc13ctvtbjqt9f6kt4kv@4ax.com> <84smscm5df.fsf@boost-consulting.com> <pf58avcm8u8r84183tia6alfo8gctc4fdu@4ax.com> <84el3vptmf.fsf@boost-consulting.com>
Message-ID: <6999avcasif2smdq6q117vrbcouh2id7gd@4ax.com>

>> One *can* on some systems, although this isn't portable, and not
>> supported on many, and guaranteed on none. Hence in the part you
>> deleted: "although certain patterns of setjmp/longjmp aren't
>> always safe." :)
>
>References, please?

--Joe Kraska, 2003.

Did it, long time ago. It was either on a HPUX or Solaris machine.
IIRC. Perhaps I'm no longer current, though. :)

C//




From andrew-pythonlist at puzzling.org  Fri Apr 25 03:34:44 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Fri, 25 Apr 2003 17:34:44 +1000
Subject: Date validation
In-Reply-To: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au>
References: <1051255370.27775.10.camel@pirate.bridge.anchor.net.au>
Message-ID: <20030425073444.GF15695@frobozz.local>

On Fri, Apr 25, 2003 at 05:22:51PM +1000, James Gregory wrote:
> Hi all,
> 
> I'm writing a killer web app in python and I want to validate all the
> input it gets. Amongst said input are dates represented as text strings
> of the format 'YYYY-MM-DD' - how can I check whether or not such strings
> represent a valid date? I had thought that
> 
>     time.strptime('2003-4-31', '%Y-%m-%d')
> 
> would throw an exception (since gnome-something tells me that april only
> has 30 days), but it just gives me back
> 
>     (2003, 4, 31, 0, 0, 0, 3, 121, 0)
> 
> indeed I'm even able to use mktime on that without an exception.
> 
> The docs for time.strptime inform me that it's libc's fault, which it
> probably is, but I'm still stuck for a method to validate this data.

In Python 2.3, strptime has been implemented independently of the platform
libc, and it does raise an exception:

    >>> import time
    >>> time.strptime('2003-4-31', '%Y-%m-%d')
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
      File "/usr/lib/python2.3/_strptime.py", line 496, in strptime
        julian = datetime_date(year, month, day).toordinal() - \
    ValueError: day is out of range for month
    >>> 

You could try stealing Python 2.3's _strptime.py, or use the recipie at:
    
    http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/56036

They appear to be the same code anyway :)

-Andrew.





From exarkun at intarweb.us  Mon Apr 21 22:24:04 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Mon, 21 Apr 2003 22:24:04 -0400
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
In-Reply-To: <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com>
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com>
Message-ID: <20030422022404.GA23037@meson.dyndns.org>

On Mon, Apr 21, 2003 at 08:49:30PM +0000, Dennis Reinhardt wrote:
>
> On Windows, which is where Visual Basic runs after all,  a complete Python
> hello world is 459 KB.   3-4 MB is *not* required.  See the online builder
> for constructing Python hello-world class executables at
> http://www.dair.com/autorun/arpyexe.cgi
> 
> A serious app (SpamAI below) is built differently and has much more code
> than hello world.  Current size is 797 KB.  800 KB *is* likely.
> 
> The python DLL is bundled as are all required PY files in both of the
> executables above.

  Is the app that generates these available (commercially or for free)?

  Jp

-- 
A sad spectacle.  If they be inhabited, what a scope for misery 
and folly.  If they be not inhabited, what a waste of space.
                -- Thomas Carlyle, looking at the stars
-- 
 up 32 days, 23:02, 2 users, load average: 1.49, 1.35, 1.35




From cwhitley at earthlink.net  Wed Apr 30 23:03:59 2003
From: cwhitley at earthlink.net (Cecil H. Whitley)
Date: Thu, 01 May 2003 03:03:59 GMT
Subject: syslogd (sorta)
Message-ID: <z80sa.47803$ey1.4455255@newsread1.prod.itd.earthlink.net>

I am trying to write a syslogd server in python.  How can I "end" the main
thread and have a daemonic thread continue?  I am using windows nt and I
have tried

def syslogd():
# Lots of code initiating a udp socket on 514 and doing a
socket.recvfrom(maxpkt) and then kicking off another thread to process the
data.

main program

thread = threading.Thread(target=syslogd, name=syslogd, args=[], kwargs={})
thread.isDaemon(True)
thread.start()

That's it.  Main exits and activity stops.  If I add

junk = raw_input("Press return to exit")

Everything works just fine, up until I hit return of course.  So where
should I be looking in order to write a "clean" background process?

Thanks,

Cecil Whitley





From sholden at holdenweb.com  Sun Apr  6 11:11:14 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sun, 06 Apr 2003 15:11:14 GMT
Subject: properties and get/set methods
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org> <v908s7echa191b@news.supernews.com>
Message-ID: <msXja.18451$It5.97@news2.central.cox.net>

"John Roth" <johnroth at ameritech.net> wrote ...
> "Mike C. Fletcher" <mcfletch at rogers.com> wrote ...
> > Dan Bishop wrote:
> > >Timothy Grant <tjg at craigelachie.org> wrote ...
> > >
[...]
> >
> > In these cases, you're using properties to simplify the modelling of
> > domain objects, or to provide some *generic* functionality over and
> > above a regular value get/set.  You don't want to have to specify that
> > functionality for every property, you just want to say "this is one of
> > _those_ properties".  The code bloat (not to mention mindless
> > duplication of code) when you get up to 15 or 20 properties for an
> > object would be ridiculous.
>
> Well, I suppose since this is Sunday, it's appropriate to bring
> religion into it.
>

That's a rather exclusive point of view, given that Saturday is the Jewish
sabbath and (I believe) Muslims worship most devoutly on Friday.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From mcherm at mcherm.com  Fri Apr  4 15:24:56 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Fri,  4 Apr 2003 12:24:56 -0800
Subject: My Object pool isnt working
Message-ID: <1049487896.3e8dea182b229@mcherm.com>

> Whew! I always thought that I must use locks to access any object in 
> Python before I modify it from a thread.
> Is it possible to append a list or a dictionary from a thread without 
> locking it? 

Yep. If you write two threads that both run around adding to a list
then the list will wind up with all items, but their order isn't
known (partial ordering of each thread's items IS preserved).

> What does thread-safe mean here?

Basically, it means that each individual bytecode of python (not each
statement) is operated in series, not parallel. Erm... to put it
better, even if you have multiple processors, your threads act as if
they were running on just _one_ cpu and timeslicing, never interrupting
the execution of a single bytecode. Except for threads that are
blocked waiting on I/O, or which are performing C code that is
"independent" of python (doesn't need to hold the GIL). This effect
is achieved because there's a single (global) interpreter lock (GIL)
which must be held for a thread to [start? complete?] a bytecode.

Somebody please correct me if I've gotten some details wrong in the
above explanation.

> For example, what would happen if I walk through a list in a for loop 
> while another thread modifies the list?

Don't do that... you might accidentally trigger the time machine and
risk altering the future of the universe. No, seriously, Python won't
crash, but your program might not do what you expect... it's not usually
recomended. Someone with deeper knowledge can probably explain EXACTLY
what will happen, but I suggest using a queue (see the Queue module) 
instead.

-- Michael Chermside





From roy at panix.com  Wed Apr  2 16:57:46 2003
From: roy at panix.com (Roy Smith)
Date: 2 Apr 2003 16:57:46 -0500
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <b6ckjr$903$1@panix2.panix.com> <slrnb8jkg8.5u1.ramen@lackingtalent.com> <wBIia.2000$y3.310439@news010.worldonline.dk>
Message-ID: <b6fmcq$dh0$1@panix2.panix.com>

I wrote:
>>> the more you can push down into the libraries and interpreter core,
>>> the better off you are most of the time.

"Dave Benjamin" <ramen at lackingtalent.com> wrote:
>> This has been my experience as well. It is a very gratifying discovery,
>> IMHO, because it encourages a less-is-more style of programming where you
>> say what you mean, and little else.

Anders J. Munch <andersjm at inbound.dk> wrote:
> The downside is that sometimes code is contorted to fit whatever has a
> fast high-level implementation in the current interpreter, instead of
> adapting to the nature of the problem.
> [...]
> That coding style tends to adapt to performance accidents of the
> implementation, happens in all interpreted languages.

People have been commiting crimes against good design in the name of
efficiency since punchcard days.  When I first learned C, it was on a
pdp-11, where the compiler could put three variables per function in
fast registers.  My major goal in refactoring (although we didn't call
it that in those days) was to design functions that made extensive use
of no more than three integers or pointers.

Just a couple of weeks ago, I let the efficiency bug bite me and spent
two days totally tearing apart a C++ class to make it use a different
set of STL containers.  The savings?  I turned a rarely-used linear
operation (where N <= 24 most of the time) into a constant-time
operation.  Totally meaningless to overall performance, but once the
bug bit, it was hard to stop myself :-)

The dark side of the force is seductive.  It tries to make us
efficient, even at the cost of clarity and maintainability.  In that
way lies bit-twiddling and assembler language and soldering irons.  We
must be diligent in our fight against it.



From marcoxa at cs.nyu.edu  Thu Apr 17 12:18:33 2003
From: marcoxa at cs.nyu.edu (Marco Antoniotti)
Date: Thu, 17 Apr 2003 12:18:33 -0400
Subject: Why functional Python matters
In-Reply-To: <slrnb9on0q.sho.ramen@lackingtalent.com>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <87k7dt5m3j.fsf@interim.henrik-motakef.de> <1azna.115402$It5.22896@news2.central.cox.net>
Message-ID: <ttAna.2$L4.96@typhoon.nyu.edu>


Steve Holden wrote:

> "Henrik Motakef"  wrote ...
>
> >laotseu  writes:
> >
> >
> >>Lispers friends, let's have you're opinion on that point : is Common
> >>Lisp a functionnal language ?
> >
> >Common Lisp is whatever you want it to be: You can write
> >functional-style programs in it if you want, but is certainly not a
> >"pure" functional language. It supports several other paradigms
> >equally well, and even more important, it will let you integrate
> >yet-unkown paradigms when you think that they are useful.
> >
> >There are really ideas to remove the functional stuff from Python? I
> >wonder why people frequently think that making things impossible will
> >improve programming languages.
> >
> >just-because-it's-flexible-doesn't-mean-it's-perl-ly y'rs
>
>
>
> I can see that, but wouldn't you agree that removing redundant 
> features also
> removes cruft from the source, and eases the maintenance burden, in turn
> reducing the likelihood of implementation bugs? While Guido is known to
> regret the inclusion of lambda in Python, I don't think he would want to
> take it out on purely ideological grounds. Indeed, the suggestion was 
> not so
> much that it would be taken out of the existing language, rather that it
> would not be implemented in some "futurePython", normally called 
> Python 3 or
> Python 3000.
>
> It would be nice to have One True Python, 


Hav I missed something?  I thought that that is exactly what you have now :)

Cheers

--
Marco Antoniotti







From hbl at st-andrews.ac.uk  Thu Apr 10 05:24:34 2003
From: hbl at st-andrews.ac.uk (Hamish Lawson)
Date: Thu, 10 Apr 2003 10:24:34 +0100
Subject: [Python-Dev] PEP305 csv package: from csv import csv?
In-Reply-To: <16020.9071.801846.936864@montanaro.dyndns.org>
References: <5.2.0.9.0.20030409143148.01d0d620@spey.st-andrews.ac.uk>
 <5.2.0.9.0.20030409143148.01d0d620@spey.st-andrews.ac.uk>
Message-ID: <5.2.0.9.0.20030410101950.01c30a48@spey.st-andrews.ac.uk>

Skip Montanaro has informed me that he has forwarded my message to the csv 
mailing list:

     http://manatee.mojam.com/mailman/listinfo/csv


Hamish Lawson





From mertz at gnosis.cx  Tue Apr  8 02:48:58 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 02:48:58 -0400
Subject: How to protect python code ? (fwd)
References: <3E8BE455.4000904@novagrid.com>
  <mailman.1049389032.4904.python-list@python.org>
  <u8bka.651$001.591065@monger.newsread.com>
  <mailman.1049722991.18682.python-list@python.org>
  <v938h26mdl9s8f@corp.supernews.com> <3E921B51.5030401@hotmail.com>
Message-ID: <aDnk+kKkX0yB092yn@gnosis.cx>         

msw <markscottwright at hotmail.com> wrote previously:
|Maybe, but since I'm the only one here who's read the requirements
|document, I submit that I'm the only one qualified to suggest a software
|design, no matter how limited my intelligence.

I'm sure I speak for all the other posters in saying that it is not that
we think you are wrong in your ability to read the requirements
document--it is the requirements document that is wrong in the
presuppositions it makes.

I've read enough of those things in my time to know that a requirments
document never says what it means, nor ever means what it says.  You
might not get fired for writing a program that meets spec, but you sure
as heck won't get promoted for doing so.

In any case, a number posters have provided you with way more than the
amount of information needed for you to develop "protected code."
You're right that no one is volunteering to do your (paid?) job for
you... but it's not exactly *difficult* to modify the behavior of
__import__() a little bit.

Yours, Lulu...

--
 mertz@  _/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: \_\_\_\_    n o
gnosis  _/_/             Postmodern Enterprises            \_\_
.cx    _/_/                                                 \_\_  d o
      _/_/_/ IN A WORLD W/O WALLS, THERE WOULD BE NO GATES \_\_\_ z e





From rganesan at myrealbox.com  Tue Apr 15 04:23:52 2003
From: rganesan at myrealbox.com (Ganesan R)
Date: 15 Apr 2003 13:53:52 +0530
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it> <ouadesfkn2.fsf@andlx-anamika.cisco.com> <7mOma.20244$T34.649927@news2.tin.it>
Message-ID: <ouvfxgdv8n.fsf@andlx-anamika.cisco.com>

>>>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:
>> 
>> Why am I getting a different list for the last call?

> Because you're asking about a distinct (even though equal) value,
> AKA object.  Python MAY but doesn't HAVE TO reuse the same object
> when you use several literals indicating equal immutables -- in
> practice, currently, it does for small integers, but (in Python
> 2.2) not for such strings as you used here.  Python can always
> change this in future releases, since this is strictly an issue of
> implementation -- optimizing one way or another, no more than that.

<andlx-anamika 1> ~ $ python
Python 2.2.1 (#1, Jul 29 2002, 23:15:49) 
[GCC 2.95.4 20011002 (Debian prerelease)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import gc
>>> a = 'hello world'
>>> b = 'hello world'
>>> c = a 
>>> gc.get_referrers(b)
[{'a': 'hello world', 'c': 'hello world', 'b': 'hello world', 'gc': <module 'gc' (built-in)>, '__builtins__': <module '__builtin__' (built-in)>, '__name__': '__main__', '__doc__': None}]

The reason I was confused was because of the output from gc.get_referrers(b). 
I assumed that a and b held a reference to the same object. I should've
checked the ids of a and b. 

I now figured out that a is holding a reference to b because of some
python implementation detail. In fact all module variables seem to hold a
reference to other variables.  

Thanks for the clarifications. 

Ganesan

-- 
Ganesan R




From staschuk at telusplanet.net  Sun Apr 20 13:28:54 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 20 Apr 2003 11:28:54 -0600
Subject: Pythonic way to sum n-th list element?
In-Reply-To: <Nesoa.37952$LB6.862279@news1.tin.it>; from aleax@aleax.it on Sun, Apr 20, 2003 at 07:45:49AM +0000
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <mailman.1050821359.13091.python-list@python.org> <Nesoa.37952$LB6.862279@news1.tin.it>
Message-ID: <20030420112854.A808@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
  [...]
> I think the familiarity with SPECIFIC bulk-functions, big-sigma
> first and foremost, is quite good.  A higher order function that
> abstracts them all, and more, is quite a different issue IMHO --
> it does require a penchant for abstract thinking, like so much
> of functional programming does.  We claim (an old slogan but
> still good) that "Python fits your brain" -- if we want to be
> sincere when saying this to "ordinary programmers off the street",
> then we should take care not too place an excessive emphasis on 
> the generality of functional programming to the detriment of
> clarity and simplicity for ordinary operations such as "sum up
> these numbers".

I entirely agree.  Your sum function optimizes both the
performance and the clarity of the common case.

What I was getting at is that imho an abstraction is easy to
understand if you are already familiar with several concrete
examples of it, and hard to understand if you are not.  So if the
"ordinary programmer off the street" is familiar with only big
sigma (if that) then reduce will indeed seem a strange abstract
thing, and very unclear.  But if they're familiar with a bunch of
similar notations, such as big pi and set operators, reduce will
seem more natural and clear.

This is an attempt to explain why I think reduce is clear, but
your teaching experience says it is not.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"What I find most baffling about that song is that it was not a hit."
                                          -- Tony Dylan Davis (CKUA)




From FBatista at uniFON.com.ar  Fri Apr 11 12:53:16 2003
From: FBatista at uniFON.com.ar (Batista, Facundo)
Date: Fri, 11 Apr 2003 13:53:16 -0300
Subject: a module who knows its caller
Message-ID: <A128D751272CD411BC9200508BC2194D018853DA@escpl.tcp.com.ar>

I already made this routine

def getfather():	
	stack = inspect.stack()
	grandfather = stack[2][3]
	return grandfather

So, you use it in your function like this

	...
	myfather = getfather()
	...

Remember that the father of your function is the grandfather of getfather().

Any question, just ask!

Facundo


#- -----Mensaje original-----
#- De: mis6 at pitt.edu [mailto:mis6 at pitt.edu]
#- Enviado el: Viernes 11 de Abril de 2003 1:13 PM
#- Para: python-list at python.org
#- Asunto: a module who knows its caller
#- 
#- 
#- I have a module (say m.py) that can be called by different 
#- scripts (say a.py 
#- and b.py). I want m.py to be able to tell if it has been 
#- called by a.py or
#- by b.py.
#- 
#- What I do now is to pass the global name __file__ to a 
#- function m.beencalledby:
#- something like
#- 
#- # a.py
#- import m
#- m.beencalledby(__file__)  # __file__ is a.py
#- 
#- 
#- # b.py
#- import m
#- m.beencalledby(__file__)  # __file__ is b.py
#- 
#- # m.py
#- 
#- def beencalledby(fname): print 'I have been called by',fname
#- 
#- it works, but I wonder if there is a more elegant solution.
#- 
#- TIA,
#- 
#-                                    Michele
#- -- 
#- http://mail.python.org/mailman/listinfo/python-list
#- 




From warlock at eskimo.com  Wed Apr 30 14:46:16 2003
From: warlock at eskimo.com (Jim Richardson)
Date: Wed, 30 Apr 2003 11:46:16 -0700
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org> <1bf4033f.0304290627.40f7d831@posting.google.com> <uptn56xxi.fsf@fitlinxx.com> <3EAFE9D4.10805@removethis.free.fr>
Message-ID: <o8i5o-26o.ln1@grendel.myth>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 30 Apr 2003 15:20:52 +0000,
 laotseu <bdesth at removethis.free.fr> wrote:
> David Bolen wrote:
> 
> 
> (snip)
> 
>> 
>> In terms of Python IDEs, most are code/debug only, no GUI.  In the
>> wxPython world, BOA Constructor is getting there but still too much of
>> a moving target to my mind, and you have to like the Delphi model.
>> PythonCard may be promising but is also early on.  wxDesigner is nice,
>> but not integrated into a more general IDE.
> 
> There is also wxGlade, a glade-like RAD tool for wxWindows/wxPython.
> 
> Laotseu
> 

Don't forget PyQT and qt-designer. 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+sBn4d90bcYOAWPYRApFDAKCoSkPrbHKmkGBCT8ZT3Zz5wuwUPgCgsb/U
wGpG/ZEoDyFl/rPoTJUupko=
=DKxV
-----END PGP SIGNATURE-----

-- 
Jim Richardson         http://www.eskimo.com/~warlock

Linux, because eventually, you grow up enough to be trusted with a fork()



From pan-newsreader at thomas-guettler.de  Wed Apr  2 00:37:36 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Wed, 02 Apr 2003 07:37:36 +0200
Subject: Problem with Idle
References: <435d40e0.0304011438.5b7b0595@posting.google.com>
Message-ID: <pan.2003.04.02.07.37.36.373585.810@thomas-guettler.de>

On Wed, 02 Apr 2003 00:38:14 +0200, Chris wrote:

> I have a module and a main program that imports it. When I make changes
> to the module in one IDLE window, and run the program in the main
> window, the changes are not picked up (I did do ctrl-s in the module
> window to save it). But if I go to the MS-DOS prompt and run the main
> program it does run the module with the latest changes to it??????

I don' use IDLE, but maybe it helps if you use the reload function.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From magnus.therning at wanadoo.nl  Sun Apr 20 12:18:01 2003
From: magnus.therning at wanadoo.nl (Magnus Therning)
Date: Sun, 20 Apr 2003 16:18:01 GMT
Subject: printing email parts of type message/external-body??
Message-ID: <ZKzoa.602320$sj7.24103068@Flipper>

Hi,

I am having problems with the following piece of code, would appreciate
any insight that can be offered!

There should be some headers that describes the external body (mime type
and so on), but I haven't found a way to add those.


---- test.py ------------
from email.Message import Message

meb_message = Message()
meb_message.add_header('Content-Type', 'message/external-body',
    access_type='local-file')
meb_message.set_payload('name="test"\n')
print meb_message
-------------------------
magnus at die:[test].$ python test.py

Traceback (most recent call last):
 File "test.py", line 8, in ?
   print meb_message
 File "/usr/lib/python2.2/email/Message.py", line 97, in __str__
   return self.as_string(unixfrom=True)
 File "/usr/lib/python2.2/email/Message.py", line 107, in as_string
   g.flatten(self, unixfrom=unixfrom)
 File "/usr/lib/python2.2/email/Generator.py", line 100, in flatten
   self._write(msg)
 File "/usr/lib/python2.2/email/Generator.py", line 128, in _write
   self._dispatch(msg)
 File "/usr/lib/python2.2/email/Generator.py", line 154, in _dispatch
   meth(msg)
 File "/usr/lib/python2.2/email/Generator.py", line 302, in _handle_message
   g.flatten(msg.get_payload(0), unixfrom=False)
 File "/usr/lib/python2.2/email/Message.py", line 179, in get_payload
   raise TypeError, i
TypeError: 0
-------------------------

/M

-- 
Magnus Therning                    (OpenPGP: 0x6A83A7DF)

No matter what they do, companies are sure to annoy someone.
     -- Deborah Branscum



From paul at prescod.net  Mon Apr  7 18:33:18 2003
From: paul at prescod.net (Paul Prescod)
Date: Mon, 07 Apr 2003 15:33:18 -0700
Subject: Problem setting COM property using win32com
In-Reply-To: <E1B4F80EB080194BBAD3092392F511B3DADEF2@XCH-NW-13.nw.nos.boeing.com>
References: <E1B4F80EB080194BBAD3092392F511B3DADEF2@XCH-NW-13.nw.nos.boeing.com>
Message-ID: <3E91FCAE.2080900@prescod.net>

Unless its proprietary, please send me the makepy generated class 
off-list. I need to knwo what's on line 1217 and it would help to know 
the surrounding context, inheritance structures etc.

  Paul Prescod





From martin at v.loewis.de  Tue Apr 15 15:21:58 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 15 Apr 2003 21:21:58 +0200
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <m3y92cvolo.fsf@mira.informatik.hu-berlin.de> <k7dwsp9l.fsf@morpheus.demon.co.uk>
Message-ID: <m3znmrd0rt.fsf@mira.informatik.hu-berlin.de>

Paul Moore <gustav at morpheus.demon.co.uk> writes:

> > This only works if the console uses a Unicode font (Lucida Sans
> > Unicode).
> 
> How do I get this? I have Lucida Sans Unicode installed, but in the
> console properties screen, the only font I can choose is Lucida
> Console...

Oops, that's the one I meant. For bitmap fonts, arbitrary Unicode
characters are not supported.

Regards,
Martin



From imcmeans at telus.net  Sat Apr 19 21:15:56 2003
From: imcmeans at telus.net (Ian McMeans)
Date: 19 Apr 2003 18:15:56 -0700
Subject: When not to use an RE -- an example
References: <c76ff6fc.0304191459.2adf73a@posting.google.com>
Message-ID: <7f9e1817.0304191715.4a9cd49c@posting.google.com>

How about using a dictionary? I wish we had dictionary comprehensions
:(

>>> user_input = " ZZZZZZZ\n\n"
>>> d = dict( [(x, None) for x in list(user_input.strip())] )
>>> len(d.keys())
1

I don't really understand your regular expression...
"^(?:(.)(?=\1))+\1\Z"
Why the lookahead assertion following the first capture group,
followed by the capture group outside the parantheses? It seems crazy
to me.

".{2,}" is repeated characters, ".{2,}\s*$" is repeated characters
followed by optional whitespace, with nothing following it.

sjmachin at lexicon.net (John Machin) wrote in message news:<c76ff6fc.0304191459.2adf73a at posting.google.com>...
> I needed a check for strings consisting of repeated characters -- like
> when users type "ZZZZZZZ" instead of "UNKNOWN" into a database field.
> After implementing the obvious overlapping-substring comparison, I got
> to thinking how this could be done with REs. The following resulted:
> 
> import re
> repeats1 = re.compile(r"^(?:(.)(?=\1))+\1\Z", re.DOTALL).match
> def repeats2(s):
>    return len(s) > 1 and s[1:] == s[:-1]
> for testvalue, expected in zip(
>    ['','x','xx','xxx','xxxxxx','xy','xxy','xyy','\n\n\n','aaa\n'],
>    [0,  0,  1,   1,    1,       0,   0,    0,    1,       0     ]):
>    print repr(testvalue), not not repeats1(testvalue),
> repeats2(testvalue), expected
> 
> Pergly/phugly, eh? Note the effort required to ensure the newline
> cases worked.



From jkominek at cs.cmu.edu  Wed Apr 23 16:46:50 2003
From: jkominek at cs.cmu.edu (John Kominek)
Date: Wed, 23 Apr 2003 16:46:50 -0400
Subject: .pyc write location
Message-ID: <3EA6FBBA.50701@cs.cmu.edu>

Hi,

I have a question about the location of compiled objects. By default, 
when you compile a source file to a pyc (or pyo) file, the output file 
is in the same directory as your source. I've always found this behavior 
untidy. Maybe it's just me, but I like to build object files in a 
directory separate from my source code.

There compile() funtion in py_compile allows for a non-standard location 
to be specified,

    def compile(file, cfile=None, dfile=None, doraise=False):
    ...
        if cfile is None:
            cfile = file + (__debug__ and 'c' or 'o')
        fc = open(cfile, 'wb')

 But can this be controlled from the command line when invoking the 
interpreter? If there's previous discussion on this topic, or a FAQ 
entry, I'd appreciate any pointers.

john




From frobozz_electric at hotmail.com  Sat Apr 26 16:22:19 2003
From: frobozz_electric at hotmail.com (Sean Ross)
Date: Sat, 26 Apr 2003 16:22:19 -0400
Subject: Some Python Code
References: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com> <mailman.1051378901.13527.python-list@python.org>
Message-ID: <2UBqa.5766$2g5.860168@news20.bellglobal.com>

Following Andrew's suggestion, I've created a page to stick the code on:

http://chat.carleton.ca/~sross/

Thanks to Andrew, and to the others who've pointed out uselesspython as
another possible repository.

Sean





From aahz at pythoncraft.com  Wed Apr  2 11:40:49 2003
From: aahz at pythoncraft.com (Aahz)
Date: 2 Apr 2003 11:40:49 -0500
Subject: Can double-linked lists be implemented in python?
References: <169a7283.0304020432.5a824d1@posting.google.com> <slrnb8lbo0.ccp.rmunn@localhost.localdomain> <3E8AFFB7.9516A49F@engcorp.com>
Message-ID: <b6f3qh$cq$1@panix3.panix.com>

In article <3E8AFFB7.9516A49F at engcorp.com>,
Peter Hansen  <peter at engcorp.com> wrote:
>Robin Munn wrote:
>> sdieselil <sdieselil at yahoo.com> wrote:
>>>
>>> How can I implement double-linked lists and other complex tree-like
>>> structures in Python?
>> 
>> Do you *need* doubly-linked lists? Is there something that Python's
>> builtin list type can't do for you? Consider whether it's worth taking
>> the time to code up a data structure, with all the attendant edge cases,
>> when you've already got such nice useful builtin types...
>
>I can think of at least one example where one would *need* doubly-linked
>lists... a homework assignment.  ;-)

You have a nasty, paranoid mind.  I like that.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From staschuk at telusplanet.net  Sun Apr 20 05:30:38 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 20 Apr 2003 03:30:38 -0600
Subject: Are reference cycles bad form?
In-Reply-To: <5sl3avsia0v41d361c6se63at9hd59po2b@4ax.com>; from mail@manuelmgarcia.com on Sat, Apr 19, 2003 at 11:12:42PM +0000
References: <3o11avop6vkh3ake8dgbtst93t1or771ct@4ax.com> <mailman.1050772931.15672.python-list@python.org> <5sl3avsia0v41d361c6se63at9hd59po2b@4ax.com>
Message-ID: <20030420033038.A8610@tibia.amotlpaa.bogus>

Quoth Manuel M Garcia:
> On Sat, 19 Apr 2003 11:16:10 -0600, Steven Taschuk
> <staschuk at telusplanet.net> wrote:
> 
> >Imho this is an engineering question, not a style question.  [...]
>
> In my current case, it feels more like a style question, instead of an
> engineering question.  [...performance and memory use are irrelevant...]

Ah.  So you're optimizing for clarity (rather than, say,
performance); that is indeed a matter of style.

Imho structures with reference cycles are not intrinsically more
or less clear than structures without, so as a general principle
they are neither good form nor bad form.  (In a particular case,
of course, their presence might enable a clearer algorithm or
simplify other aspects of the data structure; in such cases they
are good form, though only indirectly.)

In other words, don't worry about using reference cycles qua using
reference cycles; only worry about them insofar as they affect the
clarity of other things.

  [...]
> Happily, Python code is easy to refactor, so I am more likely in
> Python to experiment with different data structures and algorithms,
> and pick the one that feels 'cleanest'.

Just so!

-- 
Steven Taschuk                  staschuk at telusplanet.net
"Telekinesis would be worth patenting."  -- James Gleick




From missedby2inches at gmx.net  Fri Apr 25 09:00:10 2003
From: missedby2inches at gmx.net (Two Inches)
Date: Fri, 25 Apr 2003 15:00:10 +0200
Subject: embedding python2.2 in C++/Qt
In-Reply-To: <mailman.1051270058.28503.python-list@python.org>
References: <mailman.1051270058.28503.python-list@python.org>
Message-ID: <b8bbgq$bgn$1@mamenchi.zrz.TU-Berlin.DE>

jdominion at gmx.de wrote:
> hi!
> 
> Who has up-to-date experience in embedding python into C++?  All examples 
> found in the internet or in the (not recommendable) book of Himstedt/M?tzel
> are for python version 1.5.
> 
> I include the Python.h and cannot perform the command Py_Initialize(). what
> is the problem?

I've done this and it works for me. If you can include a more detailed 
error description and maybe a stripped-down version of your code we 
might be able to give better answers.




From shagshag13 at yahoo.fr  Sat Apr 12 10:31:24 2003
From: shagshag13 at yahoo.fr (shagshag13)
Date: Sat, 12 Apr 2003 16:31:24 +0200
Subject: am i missing something ?
Message-ID: <3e98226d$0$26858$626a54ce@news.free.fr>

hello,

i would like to strip only non white space from a string, i have found strip
in documentation as :

strip([chars])
Return a copy of the string with leading and trailing characters removed. If
chars is omitted or None, whitespace characters are removed. If given and
not None, chars must be a string; the characters in the string will be
stripped from the both ends of the string this method is called on.

but i can't get it work :

>>> s = 'this is an example \n'
>>> s.strip("\n")
Traceback (most recent call last):
  File "<pyshell#184>", line 1, in ?
    s.strip("\n")
TypeError: strip() takes no arguments (1 given)

i think i'm missing something...





From tanzer at swing.co.at  Tue Apr 15 03:05:58 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Tue, 15 Apr 2003 09:05:58 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "Mon, 14 Apr 2003 18:09:25 GMT."
             <pPCma.12520$883.10698@nwrdny02.gnilink.net>
Message-ID: <E195KVy-0004Rf-00@tswings.swing.cluster>

Carl Banks <imbosol-1050339953 at aerojockey.com> wrote:

> Christian Tanzer wrote:
> > All these functions (except the one you're writing) might have been
> > implemented years before the text properties and
> > `inhibit-point-motion-hooks` were introduced.
>
> See, that's really the problem.  It's obvious why history made Emacs
> what it is, but it doesn't change the fact that, as it stands now,
> Emacs is a terrible software design.

This kind of hubris is normally associated with Perl programmers, I
thought. Calling Emacs a terrible software design is similar to saying
that tea clippers had a terrible design because they didn't use an
engine...

Emacs is certainly a very successful application. Complex as it is, it
is very easy to program even for somebody like me who hates lots of
silly parentheses.

> Now, Emacs has a variable inhibit-point-motion-hooks.  If Emacs were
> currently designed well, this variable would not be needed, because
> functions would be able to operation without touching the point.

Not true. This has nothing to do with touching point but how cursor
movements behave. For example, searching doesn't work properly in some
cases unless you let `inhibit-point-motion-hooks` to a true value.

> Now, Emacs has a current bufferm, and a current point in a buffer,
> which all text operations take place it.  Right now, you have to use
> macros such as save-excursion and with-current-buffer, because right
> now, text manipulation functions implicitly use the current buffer and
> current point.  If Emacs were currently designed well, functions would
> accept a buffer to work on as an argument, and there would be no need
> to change the buffer temporarily.
>
> And so on.
>
> (As for things like buffer-local variables, I don't think it's a big
> deal to make them attributes of a buffer class.  Frame local
> variables, frame objects.  I think it's quite short-sighted to try to
> force buffer-local variables to be regular variables, and not what
> they obviously should be, attributes of the buffer, just because Emacs
> does it that way.)

Only partly true. Sure, a object-oriented design would make for a
better architecture and a better API. A pythonic solution certainly
would have buffer/frame/... objects with their respective attributes.
But there isn't that much difference between an attribute of a buffer
object (in the Python sense) and a buffer-locale variable (in the
Emacs sense).

You might be able to save some macros, but you would probably still
have lots of context which needs temporary changes for some functions
to work.

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From exarkun at intarweb.us  Sun Apr  6 14:35:58 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 6 Apr 2003 14:35:58 -0400
Subject: Install bug : incorrect permissions
In-Reply-To: <20030406142002.ABD3D40E30@kraid.nerim.net>
References: <20030406142002.ABD3D40E30@kraid.nerim.net>
Message-ID: <20030406183558.GA25907@meson.dyndns.org>

On Sun, Apr 06, 2003 at 04:16:40PM +0200, Yann E. MORIN wrote:
> Hi all!
> 
> (sorry, I'm not subscribed)
> 
> I've found that python 2.2.2 installs its lib-dynload files with what I think
> are incorrect permissions : 0750. That prevent any user, safe root and the
> group, to use those files, like the expat module.
> 
> Changing manualy to 0755 solves this.
> 
> I've built on LFS-4.0 with BLFS cvs from 20030405. lib-dynload is
> /usr/lib/python2.2/lib-dynload
> 

  Python 2.2.2 installed all the lib-dynload/ files with permission 755 on
my system.  Perhaps you have a umask issue?

  Jp

-- 
http://catandgirl.com/view.cgi?90
-- 
 up 17 days, 15:01, 2 users, load average: 0.00, 0.00, 0.00




From peter at engcorp.com  Fri Apr 25 11:23:02 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 11:23:02 -0400
Subject: install/import modules, /site-packages
References: <3EA94E2D.6010505@ncsa.uiuc.edu>
Message-ID: <3EA952D6.70C6A89A@engcorp.com>

Randy Heiland wrote:
> 
> I need some basic Python module knowledge.  On my RH Linux installation,
> I've got the following:
> 
> /usr/local/lib/python2.2/xml
> and
> /usr/local/lib/python2.2/site-packages/_xmlplus
> 
> and when I >>> import xml.sax, I get:
> '/usr/local/lib/python2.2/site-packages/_xmlplus/sax/__init__.pyc'
> 
> How does this module (in site-packages) take precedence over the other?

"Use the source, Luke."  

Look in lib/python2.2/xml/__init__.py and all will be revealed.  :-)

-Peter



From leomasi at libero.it  Wed Apr  2 15:32:32 2003
From: leomasi at libero.it (QWE)
Date: Wed, 02 Apr 2003 20:32:32 GMT
Subject: underline a letter in a button to use "ALT+key" shortcut with Tkinter
References: <b6ebu8$bp2$1@lacerta.tiscalinet.it>
Message-ID: <haim8vcm6b0li0rt121stjkd0gcbr0frge@4ax.com>

On Wed, 2 Apr 2003 11:53:35 +0200, "Federico"
<maschio_77 at hotmail.com> wrote:

>With Tkinter how can I underline a letter in the text in a button and set an
>"ALT+Key" shortcut for that button? is there some library I can use?
>Thanks
>
from Tkinter import *

def click():
    label.configure(text="Hello World")

root=Tk()
button=Button(root,text="Click Me",command=click, underline=6)
button.pack()
label=Label(root)
label.pack()
root.bind('<Alt_L><m>', lambda e:click())
root.mainloop()




From andy at reportlab.com  Mon Apr 14 20:55:26 2003
From: andy at reportlab.com (Andy Robinson)
Date: Tue, 15 Apr 2003 01:55:26 +0100
Subject: OT: Python application deployment?
References: <mailman.1050282088.6885.python-list@python.org>
Message-ID: <n8lm9vk42rtme6pclslos94kaekqvk27mc@4ax.com>

>My two main options appeared to be InnoSetup and NSIS. InnoSetup was eliminated early by reading the FAQ, esp.
>
>http://www.jrsoftware.org/isfaq.php#conditional
>

Hmm.  Both those products can be seriously misread...

The link you point to says that the "Inno Setup Extensions"
package can do what you want, and it can.  Another tip:
what we do, and what the author of pyant.sourceforge.net
does, is to have a Python script which assembles various bits
and pieces and generates/tweaks the Inno ISS file.

OTOH I just glanced at the NSIS page and saw this
"A lovely coding experience with elements of PHP and assembly
(includes 20 general purpose variables, a stack, real flow control,
etc)".  Which would make me run a mile at first glance...

- Andy Robinson





From chrisDOTplachta at teknovusDOTcom  Tue Apr  8 19:55:30 2003
From: chrisDOTplachta at teknovusDOTcom (news.easynews.com)
Date: Tue, 08 Apr 2003 23:55:30 GMT
Subject: Problem with Tkinter
Message-ID: <3E9361FD.6000504@teknovusDOTcom>

I just managed to get Python configured to run with TK. I am running 
Python 2.2.2 with TCL/TK 8.3.5 on Redhat linux 8.0.

When I try to run the recommended TK test, I get the following error:


[chrisp at tampa /home/chrisp]% python
Python 2.2.2 (#1, Apr  8 2003, 15:21:03)
[GCC 3.2 20020903 (Red Hat Linux 8.0 3.2-7)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
 >>> import Tkinter
 >>> Tkinter._test()
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "/usr/local/src/Python-2.2.2/Lib/lib-tk/Tkinter.py", line 3099, 
in _test
     "iso-8859-1")
LookupError: no codec search functions registered: can't find encoding
 >>>

Any ideas or suggestions would be greatly appreciated.

Chris Plachta



Here is my PYTHONPATH variable setting:


 > echo $PYTHONPATH
/usr/local/src/Python-2.2.2/Lib:\
/usr/local/src/Python-2.2.2/Lib/lib-tk:\
/usr/local/src/Python-2.2.2/Lib/compiler:\
/usr/local/src/Python-2.2.2/Lib/curses:\
/usr/local/src/Python-2.2.2/Lib/distutils:\
/usr/local/src/Python-2.2.2/Lib/email:\
/usr/local/src/Python-2.2.2/Lib/encodings:\
/usr/local/src/Python-2.2.2/Lib/hotshot:\
/usr/local/src/Python-2.2.2/Lib/lib-old:\
/usr/local/src/Python-2.2.2/Lib/lib-tk:\
/usr/local/src/Python-2.2.2/Lib/plat-linux2




From grante at visi.com  Sun Apr 27 01:10:45 2003
From: grante at visi.com (Grant Edwards)
Date: 27 Apr 2003 05:10:45 GMT
Subject: WML examples using mod_python?
References: <slrnbals94.2qp.grante@tuxtop.visi.com> <uLFqa.9353$h%2.1023630@read1.cgocable.net>
Message-ID: <slrnbamprf.2th.grante@tuxtop.visi.com>

In article <uLFqa.9353$h%2.1023630 at read1.cgocable.net>, Mark Nenadov wrote:

>> Can anybody point me towards any examples of using mod_python
>> for WML?

> I haven't worked with WML in Python, though I have in other
> languages (PHP, etc). 
> 
> All you will really need to do is assemble your WML decks
> (which are just markup) and then output them along with the
> correct Content-type line. Once you have your WML generated,
> it is just a matter of outputing it. You shouldn't have to do
> anything special with Mod_Python

I've just started reading up on WML (got a couple books on the
way), and was just wondering if there were any real-world
examples in Python that I could look at.

> If you are looking for something to make WML-generation
> quicker, you might want to check out "pywml"
> (http://sourceforge.net/projects/pywml).

Thanks!  That's what I was looking for.  I was sure there had
to be something like that, but all my searches found were a
Zope/WML howto.

-- 
Grant Edwards                   grante             Yow!  Hold the MAYO & pass
                                  at               the COSMIC AWARENESS...
                               visi.com            



From colin.blackburn at durham.ac.uk  Thu Apr  3 09:12:19 2003
From: colin.blackburn at durham.ac.uk (Colin Blackburn)
Date: Thu, 3 Apr 2003 15:12:19 +0100
Subject: The Whitespace Language
References: <283adf56.0304011143.4f4a6934@posting.google.com>
Message-ID: <MPG.18f6519e22cf51679898b3@localhost>

In article <283adf56.0304011143.4f4a6934 at posting.google.com>, 
kdahlhaus at yahoo.com says...
> I'm getting so much entertrainment out of the the joke on the jokers
> that created the Whitespace Language page -since we use indentation to
> delimit block strucutre, white-space has been part of Python's syntax
> since the beginning.  Ha!

Which they fully acknowledge. Where's the joke on them?

"Didn't you know that python already has whitespace as syntax?"

"Yes, we did know. I quite like python, in fact. You might even have 
noticed that the language I chose to implement it in has whitespace as 
syntax. We never said "every" modern programming language ignored 
whitespace, just "most" of them. C, C++, C#, Java, Perl, Scheme all 
swallow whitespace, except perhaps as token delimiters." 

Colin




From johan30 at easynet.be  Wed Apr  2 05:24:52 2003
From: johan30 at easynet.be (Johan Boeckx)
Date: Wed, 2 Apr 2003 12:24:52 +0200
Subject: problem installing tkinter
Message-ID: <3e8aba97$0$6528$afc38c87@sisyphus.news.be.easynet.net>

Hi,

I am trying to install tkinter with python 2.1.3. (from source) I followed
the instructions and changed the /Modules/Setup file for the tkinter part.
The  problem seems to be the
# *** Always uncomment this; X11 libraries to link with:
        -lX11

Because I receive during the make the follwing error :
gcc  -Xlinker -export-dynamic -o python \
                Modules/python.o \




            libpython2.1.a -lpthread -ldl  -lutil -L/usr/lib -ltk8.4 -ltcl8.
4 -lX11  -lm
/usr/bin/ld: cannot find -lX11
collect2: ld returned 1 exit status
make: *** [python] Error 1


Somebody who can help me ?
Greetz,
Johan





From gradha at titanium.sabren.com  Mon Apr 21 17:30:12 2003
From: gradha at titanium.sabren.com (Grzegorz Adam Hankiewicz)
Date: Mon, 21 Apr 2003 23:30:12 +0200
Subject: python dynamic scoping question
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DE01C@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE01C@admin56.narex.com>
Message-ID: <20030421213012.GA2272@pedos.es>

On 2003-04-21, Bjorn Pettersen <BPettersen at NAREX.com> wrote:
> You're correct! God, it's been a long time since I looked at
> it. What you're describing above (and described just as mystically
> in the TCL manuals) is simply call by reference (in the C++
> sense of the term).

Not really, because the called function doesn't require parameters if
you always want a specific variable to be set, like Julia asked for:

$ tclsh
% proc g {} { uplevel set x 5; return }
% set x
can't read "x": no such variable
% g
% set x
5

This can't be done in C portably, and is certainly not call by
reference. And please, do not send me copies of your public ansers.

-- 
 Please don't send me private copies of your public answers. Thanks.




From dmgriffith at tasc.com  Wed Apr 23 15:57:42 2003
From: dmgriffith at tasc.com (Dang Griffith)
Date: Wed, 23 Apr 2003 15:57:42 -0400
Subject: Curses on Win32
Message-ID: <b86r8p$27n$1@gateway.northgrum.com>

   I'm running 2.3a2 and decided to spearmint with curses, only to find
it's not available for import on Win32.  Is/was curses available in 2.2?
 Is there an alternative text user interface library available?
   Thanks,
      --dang



From sdieselil at yahoo.com  Thu Apr 10 18:32:01 2003
From: sdieselil at yahoo.com (sdieselil)
Date: 10 Apr 2003 15:32:01 -0700
Subject: Can I define nested classes in Python?
Message-ID: <169a7283.0304101432.13f5786c@posting.google.com>

Hi

Can I define nested classes in Python and what access do they have to outer
classes?



From aleax at aleax.it  Mon Apr 28 04:41:50 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 08:41:50 GMT
Subject: dynamic function names
References: <20030428102400.21c5074a.bram@nospam.sara.nl>
Message-ID: <iP5ra.14084$3M4.382689@news1.tin.it>

<posted & mailed>

Bram Stolk wrote:

> Hello,
> 
> Is it possible in Python to define functions with a dynamic name?

Yes, but not with the def statement.


> something like:
> 
> def make_func(funcname) :
>   def funcname :

Syntax error -- you're missing parentheses here before the colon.

>     print "This func is named", funcname
> 
> make_func(foo)
> make_func(bar)
> 
> The code above fails due to foo being unknown,
> however, when calling as make_func("foo"), the "foo" string
> cannot be used as a funcname.

Actually, if you fix the syntax error it will work, but not do
what you want:

>>> def makefunc(funcname):
...   def funcname(): print "this function is named", funcname
...   return funcname
...
>>> f = makefunc('foo')
>>> f()
this function is named <function funcname at 0x402c09cc>
>>>

See?  "def funcname():" *RE-BINDS* identifier funcname to
refer to the function object.

So, what you need is to call function 'function' in module
'new'.  You can first build a temporary function with a def
to give you the right code, globals, defaults and closure,
and then pass each of these to new.function together with
the funcname you want to use:

>>> import new
>>> def makefunc(funcname):
...   def temp(): print "this function is named", funcname
...   f = new.function(temp.func_code, globals(), funcname,
...           temp.func_defaults, temp.func_closure)
...   return f
...
>>> f = makefunc('foo')
>>> f()
this function is named foo
>>> f.func_name
'foo'
>>>


Alex




From Frank at no.spam  Thu Apr  3 04:37:55 2003
From: Frank at no.spam (Frank)
Date: Thu, 03 Apr 2003 04:37:55 -0500
Subject: Persistent server with GUI interactive functions ?
References: <mgqn8v01utuhhioq5vq32rm2r2oadtrm8i@4ax.com> <pan.2003.04.03.11.09.29.238820.809@thomas-guettler.de>
Message-ID: <qmvn8v0a4puh88i50totedt0a8b0j3ub7d@4ax.com>

On Thu, 03 Apr 2003 11:09:29 +0200, Thomas Guettler
<pan-newsreader at thomas-guettler.de> wrote:

>On Thu, 03 Apr 2003 10:29:28 +0200, Frank wrote:
>
>> I need to write a GUI app that will serve as a peer to peer
>> client/server.  At the highest level it would be similar to a chat
>> client, which has a server component (waiting for any incoming
>> connections) as well as the client part (initiating a connection,
>> executing user commands).
>
>How are your client and server connected? 

The client and server are the same program (peer to peer).  Think of
any of the chat or video-conference apps around.  They usually CAN
connect to a central server to post their existence, but do not need
to.  Any client-server can connect to or accept connections from any
other client-server, as long as the address is known
>
>Can both create tcp sockets to each other? 
>You can't do this if a client is behind a masquarding (NAT)
>router. If you can use TCP/IP, I think Pyro is the best choice. I have
>not used it yet.
>Can the client make a tcp connection to the server? Some networks only
>allow http through a proxy. If some clients are behind a http proxy
>I would recommend xml-rpc.
> 
Each instance can start a connection and accept one from someone else,
using TCP.  I'm not worried about the NAT/firewall issue, I have that
under control.
>
>> Each of those function appears to involve entering an "endless loop" as
>> the last logical step to initiate processing (i.e. "mainloop()" for the
>> user interface and "serve_forever()" for the server).  I haven't yet
>> grasped the logical structure that will allow me to have two
>> non-returning calls and maintain control of the program.
>
>I would do it with threads. Maybe there it is possible to do it without 
>threads, too.
>
>In the subject you used the word "Persistent". I would use pickles for
>development. If you have too much data you can switch to ZODB later.
>
I meant "Persistent" in the sense of the server function:  it will
wait for a connection, do its thing, and then wait for the next
connection, not terminating until a user command is given.
> thomas




From ianb at colorstudy.com  Tue Apr 22 15:11:22 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 22 Apr 2003 14:11:22 -0500
Subject: True/False
In-Reply-To: <16037.36122.922342.405340@montanaro.dyndns.org>
References: <1051036194.697.177.camel@lothlorien>
	 <16037.36122.922342.405340@montanaro.dyndns.org>
Message-ID: <1051038682.702.191.camel@lothlorien>

On Tue, 2003-04-22 at 13:42, Skip Montanaro wrote:
>     Ian> True, False = 1==1, 0==1
> 
> Why not
> 
>     try:
>         True
>     except NameError:
>         True, False = 1==1, 0==1
> 
> ?  Should work anywhere without complaints.

Yes, that's better than a version check.  I hate how much of this cruft
I'm getting at the top of my files, though... :(  I like Python for its
lack of boiler-plate code, but this feels like boiler-plate.

>     Ian> Is there any reason why reassigning True and False will be a
>     Ian> problem in the future?
> 
> The builtins may get "tightened up" in the future.  This particular idiom
> probably won't be affected, but you may not be able to (easily) shadow a
> builtin in another module's global namespace, e.g.:
> 
>     import a
>     a.True, a.False = 1==1, 0==1
> 
> might not be possible.  The idea here is to allow the byte compiler to know
> all the possible builtin shadowing that goes on in a module by examining
> just the module itself.  This, in turn, would allow access to builtins (and
> globals to some extent I think) to be optimized.

This was the sort of issue I was thinking about.  But another change is:

Python 2.3a2+ (#2, Mar 21 2003, 22:13:05) 
[GCC 3.2.3 20030316 (Debian prerelease)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> None = 10
<stdin>:1: SyntaxWarning: assignment to None

Will this happen to True/False?  Probably not, because that would
*really* mess up backwards compatibility (since try:except: wouldn't
even work).

  Ian






From msg_2222 at yahoo.com  Mon Apr  7 14:49:13 2003
From: msg_2222 at yahoo.com (Rick Y)
Date: Mon, 7 Apr 2003 11:49:13 -0700 (PDT)
Subject: socket question(2)
Message-ID: <20030407184913.69195.qmail@web20707.mail.yahoo.com>

eventhough i have installed openssl in my solaris 2.8
box, i still get these errors. ..???

./viewcvs-install
Traceback (most recent call last):
  File "./viewcvs-install", line 35, in ?
    import compat
  File "./lib/compat.py", line 20, in ?
    import urllib
  File "/usr/local/lib/python2.1/urllib.py", line 26,
in ?
    import socket
  File "/usr/local/lib/python2.1/socket.py", line 41,
in ?
    from _socket import *
ImportError: ld.so.1: python: fatal: libssl.so.0.9.6:
open failed: No such file or directory



__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com




From mfranklin at tcsi.com  Tue Apr  1 14:54:53 2003
From: mfranklin at tcsi.com (Mark Franklin)
Date: Tue, 1 Apr 2003 11:54:53 -0800
Subject: Implementing a "classmethod" in an extension type
Message-ID: <51CE884FE0BAEE47A12058CC902BDEE7188E99@xmail2ca.tcsi.com>

Thanks Thomas,

The key was the DoClassMethods() call at the end of your code. Once I'd copied that over, it worked perfectly:)


void DoClassMethods(PyTypeObject *type)
{
	PyObject *func;
	PyObject *meth;
	PyMethodDef *ml = type->tp_methods;

	for (; ml->ml_name; ++ml) {
		if ((ml->ml_flags & METH_CLASS) == 0)
			continue;
		ml->ml_flags &= ~METH_CLASS;
		func = PyCFunction_New(ml, NULL);
		if (!func)
			return;
		meth = PyObject_CallFunctionObjArgs(
			(PyObject *)&PyClassMethod_Type,
			func, NULL);
		if (!meth)
			return;
		if (-1 == PyDict_SetItemString(type->tp_dict,
					       ml->ml_name,
					       meth))
			return;
	}
}


Mark


> -----Original Message-----
> From: Thomas Heller [mailto:theller at python.net]
> Sent: 01 April 2003 16:10
> To: python-list at python.org
> Subject: Re: Implementing a "classmethod" in an extension type
> 
> 
> "Mark Franklin" <mfranklin at tcsi.com> writes:
> 
> > Hi,
> > 
> > I've got an built-in extension type, implemented in C. I 
> want to be able to implement the equivalent of a 
> "classmethod" in that type. 
> > 
> > I thought I might be able to do something like this, which 
> fairly closely mimics what you do in Python:
> > 
> > 	PyTypeObject myType = {
> > 		// initialised in the usual way. It's 
> tp_methods field includes a method definition for "MyMethod".......
> > 	};
> > 
> > 	PyObject *method = PyObject_GetAttrString(myType, "MyMethod");
> > 	// method is a PyMethodDescrObject
> > 	PyObject *classMethod = PyClassMethod_New(method);
> > 	PyObject_SetAttrString(myType, "MyMethod", classMethod);
> > 	Py_DECREF(method);
> > 
> > 	PyType_Ready(myType);
> > 
> > 
> > However, when I do this I get a TypeError exception from 
> the PyObject_SetAttrString call: "can't set attributes of 
> built-in/extension type 'MyType' ".
> > 
> > So, how do I get the class method into the type object?
> > 
> > 
> > Thanks,
> > Mark
> 
> In Python 2.3, there is a METH_CLASS flag. Unfortunately not in 2.2.
> Here are the relevant snippets that I use - it is from
> http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/ctypes/ctypes/s
> ource/_ctypes.c
> and it compiles with Python 2.2 as well as 2.3.
> 
> HTH,
> 
> Thomas
> 
> 
> #include "Python.h"
> 
> #ifndef METH_CLASS
> #define METH_CLASS 0x0010 /* from Python 2.3 */
> #define NO_METH_CLASS
> #endif
> 
> ...
> 
> /*
>  * If this is a real C classmethod (Python 2.3 and later),
>  * it has METH_O style. So the type is in the first argument,
>  * and the arg in the second.
>  *
>  * We fake this in Python 2.2, where METH_CLASS is not present,
>  * and 'args' is a tuple containing the type as first member.
>  */
> static PyObject *
> CString_from_param(PyObject *cls, PyObject *args)
> {
> 	PyObject *value;
> 
> #ifdef NO_METH_CLASS
> 	PyObject *ignore;
> 
> 	if (!PyArg_ParseTuple(args, "OO", &ignore, &value))
> 		return NULL;
> #else
> 	value = args;
> #endif
> ...
> }
> 
> 
> static PyMethodDef CString_methods[] = {
> #ifdef NO_METH_CLASS /* Python 2.2 */
> 	{ "from_param", CString_from_param, METH_VARARGS | METH_CLASS,
> #else
> 	{ "from_param", CString_from_param, METH_O | METH_CLASS,
> #endif
> 	  from_param_doc },
> 	{ NULL, NULL },
> };
> 
> 
> ...
> 
> 	if (PyType_Ready(&CString_Type) < 0)
> 		return;
> #ifdef NO_METH_CLASS
> 	DoClassMethods(&CString_Type);
> #endif
> 	PyModule_AddObject(m, "c_string", (PyObject *)&CString_Type);
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 




From ckea25d02 at sneakemail.com  Tue Apr  8 01:34:33 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 17:34:33 +1200
Subject: Cross-reference 'import' in a class hierarchy
In-Reply-To: <pan.2003.04.07.23.02.00.101458@jerf.org>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <pan.2003.04.07.23.02.00.101458@jerf.org>
Message-ID: <b6tmff$8tkev$1@ID-169208.news.dfncis.de>

Jeremy Bowers wrote:
> IMHO, this is the largest wart by far in Python when it comes to writing
> "real" programs that really shouldn't have every file dumped into the same
> directory.

It's actually not so bad in the case of an *application*,
because if you structure your application like this:

     myapp.py
     MyAppPkg
        __init__.py
        mymodule1.py
        mymodule2.py
        MySubPkg
          __init__.py
          mymodule3.py

etc., then when you do

   python myapp.py

the directory containing myapp.py effectively becomes a
source of top-level packages for the duration, so you can
refer to MyAppPkg.mymodule1 etc.

It's more of a problem when, e.g. you want to embed a
library as a subpackage of another library, and the
embedded library itself consists of subpackages which
would like to refer to each other without having to
care where they reside in the global picture.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From max at alcyone.com  Mon Apr 14 22:02:09 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 19:02:09 -0700
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl> <3E9AF8BD.BC918E5E@engcorp.com> <3E9B520F.7D6E0360@alcyone.com> <b7fltn$b7v$1@slb6.atl.mindspring.net>
Message-ID: <3E9B6821.AACF1ADD@alcyone.com>

Andrew Dalke wrote:

> You wouldn't really do a Hohmann orbit all the way
> out to Neptune, would you?

No, of course not (unless, say, you were sending bulk materials and were
trying to minimize fuel at any expense -- if you're sending a steady
stream of materials from a mining base on Triton to Earth, it doesn't
matter how fast any particular package gets there, it just matters that
they keep coming).  If you were doing something timely, you'd be more
interested in trading off trip duration for fuel requirements,
particularly in the supposed future where this game would take place,
where high-efficiency drives would be readily available.

There are even in some circumstances more fuel-minimizing transfers than
Hohmann, like bielliptic transfers (which is sort of like a
double-Hohmann transfer to a point way out in space), but they take vast
amounts of time (e.g., a Hohmann-like transfer out to Neptune and then
back simply to go from Earth to Mars).  Under some circumstances this
can take less fuel than a simple Hohmann transfer (although it won't be
energy minimizing).

> Aren't there any
> gravitational assists that could help out?  After all,
> it took Voyager 2 only 12 years or so to get there,
> not 30, though the planets were well aligned at
> the time.

In the envisioned game, one would be able to choose between different
transfer "modes" when selecting a course, so that one could choose
between fuel-minimizing ("Time is no object, use as little fuel as
possible"), time-minimizing ("Fuel is no object, get there as fast as
possible"), or schedule-meeting ("I need to get there by turn 192, use
as much fuel as required to get there but no more"), or budget-meeting
("I need to come back with the same fuel budget so only spend this
much") transfers.  What that means is that the entities in this big
table of transfers will actually be types and coefficients of equations
(polynomials or other similar relations, most likely) that will need to
be solved on demand, abstracting the problem one level.  At present the
calculator is not sophisticated enough to understand this, however.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ It's like being on a diving board that you know is too high.
\__/ Sade Adu
    Crank Dot Net / http://www.crank.net/
 Cranks, crackpots, kooks, & loons on the Net.



From Curzio.Basso at unibas.ch  Thu Apr 10 06:47:02 2003
From: Curzio.Basso at unibas.ch (Curzio Basso)
Date: Thu, 10 Apr 2003 12:47:02 +0200
Subject: PCA principal component analysis
In-Reply-To: <b1ab591.0304100042.1188e4bf@posting.google.com>
References: <b1ab591.0304090601.302cc4b6@posting.google.com>
 <yfs7ka3iom2.fsf@black132.ex.ac.uk>
 <b1ab591.0304100042.1188e4bf@posting.google.com>
Message-ID: <1049971580.21138.11.camel@libya.cs.unibas.ch>

On Thu, 2003-04-10 at 10:42, sebastien wrote:

> The matrix is :
> 
> ((2,1,3),
> (4,2,3),
> (4,1,0),
> (2,3,3),
> (5,1,9))
> 
> 
> The principal component given by IDL is 
> 
> ((0.87, -0.7, 0.69),
> (0.01, -0.64, -0.66),
> (0.49, 0.32, -0.30))
> 
> Runnnig the same matrix with the program that you have kindly provided
> gives :
> 
> ((-0.58, -0.67, 0.45),
> (-0.23, -0.40, -0.88),
> (-0.77, -0.62, -0.074))
> 
> Do you know why is there such a difference ??

Assuming that the data matrix you used holds the vectors in its rows, as
I would assume from your notation, than the principal components,
normalized to one, are the second you wrote, but not with your notation.
That is, the components are not the rows of the matrix, but the columns.
I got the same result with SVD on Matlab. I don't know where the IDL
results come from, but I would take the correctness of the Matlab ones
for granted... could it be the function in IDL is doing something else?

You could also check the difference in the subspaces, and if the
components given by IDL are really orthogonal.

bye

-- 
Curzio Basso <Curzio.Basso at unibas.ch>





From meinmailfilter at gmx.de  Thu Apr 10 09:12:56 2003
From: meinmailfilter at gmx.de (HHaegele)
Date: 10 Apr 2003 06:12:56 -0700
Subject: class too big for one file! How can I split it?
Message-ID: <e956b5f1.0304100512.597654d3@posting.google.com>

Hello,
I have the following structure:

===============
ot_starter.py :
---------------

import OTLib

ot = OTLib.Ot(A, E)

ot.Att()

ot.Btt()

---------------


===============
OTLib.py :
---------------

XX = None

class Ot(a, e):

    def __init__(self, a = None, e = None):
        global XX
        XX = e    
       
    def Att(self):
        global XX
        print XX
        ...
        
---------------

Now, OTLib.py has more then 32000 lines. And that doesn?t work. Python
1.5 can?t execute then. So, how can I add the method Btt()? How can I
make another file, which knows the global XX and who can be handled
from the ot_starter, like it is in one file? Good would also be, if
the Btt() can access XX the same way like Att() can. Is that possible?
Or how must I change the structure?

Thank you 
Heiko



From andersjm at dancontrol.dk  Mon Apr 14 09:27:26 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Mon, 14 Apr 2003 15:27:26 +0200
Subject: more informative stack traces
References: <e0bf36b5.0304132100.4324c3b8@posting.google.com>
Message-ID: <3e9ab723$0$66212$edfadb0f@dread11.news.tele.dk>

"L.C." <deja-google at wotevah.com> wrote:
> I am proposing a different style of stack trace which I found more 
> useful in debugging problems with live code. The main change involves
> showing the actual parameter values in function calls, gdb-style.
> It sounds simple because it really is, but it saves lots of guesswork
> when navigating through the stack to see what caused a problem.
> 
> An implementation is available here: http://www.wotevah.com/code/log.py
> To enable, just import it in your main program.
> 
> It requires syslog (thus Unix-only) because it seemed to be the natural
> mechanism for reporting errors and warnings, although it will also use 
> stderr if you are on a terminal.

May I suggest you refactor your code into two modules: a
platform-independent one that inspects tracebacks and generates the
traceback text in some sort of container or generator.  And a
syslogging module one with the same interface as you have now, which
is a client of the first module.  And perhaps a third module, a client
that uses portable stderr output.

Not only will this give you a fighting chance of convincing people on
syslog-less platforms of the merits of your approach, but it will also
benefit the structure of your code.  Hardwiring the output mechanism
puts obstacles in the way of applications such as CGI scripts and GUI
traceback presenters, where consoles and logfiles may be irrelevant
because user interaction goes through other channels.

- Anders





From skip at pobox.com  Wed Apr  9 08:54:22 2003
From: skip at pobox.com (Skip Montanaro)
Date: Wed, 9 Apr 2003 07:54:22 -0500
Subject: Adding item in front of a list
In-Reply-To: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de>
References: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de>
Message-ID: <16020.6142.803512.179030@montanaro.dyndns.org>

    Thomas> I think it would be nice if
    Thomas>  l=[2, 3, 4]
    Thomas>  l.insert(1)

    Thomas> would insert in front of the list, as opposed to append() which
    Thomas> adds new elements at the end.

What about

    >>> lst = [2,3,4]
    >>> lst.insert(0, 1)
    >>> lst
    [1, 2, 3, 4]

Try:

    >>> help([].insert)
    Help on built-in function insert:

    insert(...)
        L.insert(index, object) -- insert object before index

Skip




From do-not-spam-ben.hutchings at businesswebsoftware.com  Tue Apr  1 10:35:47 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 1 Apr 2003 15:35:47 GMT
Subject: Mimic a web surfer ?
References: <3e894b7d$0$29158$626a54ce@news.free.fr>
Message-ID: <slrnb8jcek.1fs.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <3e894b7d$0$29158$626a54ce at news.free.fr>, shagshag13 wrote:
> hello,
> 
> i would like to develop a script which will mimic a web surfer, that is :
> 
> - follow links as a human and thus always come from previous page (i think
> there is something to do with http 'referer' and cookies)
> - be able to log if necessary
> 
> is there anything already existing ? i'm looking for any tips which will
> help me to start...

The HttpUnit package might be useful, but it's written in Java.

> note : this could be used to have "automatic" access to free webmails not
> providing pop protocol.

That has already been done for some popular webmail providers like Hotmail
(look for "gotmail" or "HotWayDaemon") and Yahoo ("FetchYahoo").



From dickey at saltmine.radix.net  Tue Apr  8 08:05:15 2003
From: dickey at saltmine.radix.net (Thomas Dickey)
Date: 8 Apr 2003 12:05:15 GMT
Subject: curses: xterm vs. linux terminal
References: <3E9065FE.1090108@sympatico.ca> <b6qd2p$ib9$1@news1.radix.net> <3E91E40D.8050409@sympatico.ca>
Message-ID: <b6udtr$das$1@news1.radix.net>

Brian Alexander <brian094 at sympatico.ca> wrote:
> I think you are right. The terminal comes up as gnome-terminal when ps 
> is run. I guess it is supposed to provide xterm-like behaviour.

> I tried running the xterm from gnome-terminal and after it loaded, I ran 
> my box-drawing program. The line characters (ACS_HLINE etc.) still worked.

> What is the difference between these two terminal programs? Does the one 
> just interact with Gnome a little better?

more/less (there are a lot of reasons people would do something like that).

The multi-gnome-terminal (mgt) version incorporates a chunk of code from
XFree86 xterm for line-drawing.  Perhaps sometime it'll become the version
shipped with gnome (seems to have about half as many bugs).

The XFree86 xterm supports ANSI color and VT220 emulation
There's an faq at
	http://invisible-island.net/xterm/xterm.faq.html
	ftp://invisible-island.net/xterm/

-- 
Thomas E. Dickey <dickey at radix.net> <dickey at herndon4.his.com>
http://dickey.his.com
ftp://dickey.his.com



From mis6 at pitt.edu  Wed Apr 30 15:48:58 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 30 Apr 2003 12:48:58 -0700
Subject: Question about accessing class-attributes.
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE169@admin56.narex.com> <mailman.1051689777.10470.python-list@python.org>
Message-ID: <2259b0e2.0304301148.36cdd823@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<mailman.1051689777.10470.python-list at python.org>...
> <snip> Indeed if we were able to
> classify the current behavior as "a bug" in some respect, AND give
> a patch that breaks no currently designed-in behavior, yet "fixes the
> bug" and makes super more useful to you, then that's the kind of
> fix that COULD go in 2.3 beta 2 (new functionality won't, but in
> some sense it might be considered "a bug" if super violates the
> normal expectations of clearly-advanced users such as you).
> <snip>
> Maybe we should break this thread in two -- one about metaclasses
> and one about super -- with suitable subjects... I do not understand
> where the 'bugs' of the two may be related nor actually what they
> really have to do with "accessing class-attributes" except in the
> vaguest, broadest sense.
> <snip>
> Yes, maybe we should separate super-issues and metaclass-issues
> and find a sharp example in each case of where reasonably expected
> behavior from docs, and observed behavior in 2.3 beta 1, differ.  I
> think it does not need to be a "reasonable use case": that would be
> needed if we were arguing for EXTRA functionality of course -- but
> then it would go in 2.4 so there's no hurry, to say the least.  If we're
> arguing there are BUGS, so the fixes could go in 2.3, then just a
> way to show up each bug clearly should suffice (hopefully).
> 
> 
> Alex

I do agree with you. Speaking about wanted functionality (maybe for
2.4)
I would like very much a __superclass__ attribute. Example: now I can
do
(NOTICE: in p23b)

class B(object): pass
class C(B): pass
c=C()
super(C,c).__thisclass__() #=> C
super(C,c).__self_class__() #=> C

I would like to have an additional attribute (not a method)
returning the superclass of C, in this case B:

super(C,c).__superclass__ #=> B

For instance I could do 

b=super(C,c).__superclass__()

returning an instance of B.

I think Bjorn would we happy with this, too.

Certainly the documentation in
http://www.python.org/dev/doc/devel/lib/built-in-funcs.html 

has to be changed. It says:

"""
 super(type[object-or-type])

      Return the superclass of type. If the second argument is omitted
the super object returned is unbound. If the second argument is an
object, isinstance(obj, type) must be true. If the second argument is
a type, issubclass(type2, type)
must be true. """ 


But it DOES not return the superclass! I admit I myself at the
beginning
was confused.


                                                        Michele



From rcs at russellsalsbury.com  Wed Apr 30 14:49:28 2003
From: rcs at russellsalsbury.com (Russ Salsbury)
Date: 30 Apr 2003 11:49:28 -0700
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com>
Message-ID: <3ea624c6.0304301049.52c4027c@posting.google.com>

blunck at gst.com (Christopher Blunck) wrote in message news:<1147e466.0304290638.287687f6 at posting.google.com>...
 
> What advice can you offer on spreading the testing bug around?  

It's very much a cultural issue.  First and foremost you have to have
to use it yourself.  If you are an indifferent user, you can't
convince others to use it.  Be consistent, even for trivial cases.

It's much easier to spread the word if your manager supports you.  A
good manager can encourage others to use unit testing.  S/he can even
make it a requirement of checking in code.  (Unit tests are much more
effective in ensuring quality than are code reviews.)  OTOH, if your
manager is not well respected by your peers, he can actually inhibit
the process.

If you are a manager, find an evangilist in your group.  Get some
successes before you impose the full rigor on everyone.  You don't
need a revolt at the very beginning.  That said, requiring tests for
checkin should be done from the beginning.

Another possible ally is the QA manager, who could enforce the
requirement for tests before checkin.  S/he should also include unit
tests in the regression/acceptance tests.

I found that rigorous unit testing can be difficult to introduce into
a large existing piece of code, esp. if there are multiple developers
of varying degreees of enthusiasm.  In spite of that, I have found
that it is most productive, when inheriting a big chunk of code, to
take the time, even a month or more, to refactor it, write unit tests,
and put in strategic traces and object dumps.  It more than pays off
by the time release comes around.
   
> What practices did you follow?  

I recruited an eager and talented QA tester who wanted to move into
development to set up the JUnit framework and become the first
user/evangelist.

> Did you beat people over the head with it until they wised up?  

A developer may be able to do that, but a manager usually fails at it.
 There are too many ways for the disgruntled to sabotage it, esp. in
larger organizations.
 
> Were you softer in your approach?  

Apparently not soft enough, because I got canned a few months later.
;-)
Actually, it worked in my group, but I was chartered to improve the
development practices of the entire company and when the President was
replaced and the VP of Engineering and the Chief Architect left, the
disgruntled had their way.  Pythonistas do not always triumph. ;-)
 
> What were the long term implications?

The company went bankrupt a few months later. :-)

The long term implications:  no matter what happens you are a better
programmer if you do rigorous automated(!) unit testing.  Your company
produces a better product and has a better chance of long term success
if everyone does.
 
-- Russ



From bignose-hates-spam at and-zip-does-too.com.au  Mon Apr  7 20:23:44 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Tue, 08 Apr 2003 00:23:44 GMT
Subject: How to protect python code ?
References: <mailman.1049758213.26256.python-list@python.org>
Message-ID: <slrnb945tu.p4.bignose-hates-spam@rose.localdomain.fake>

On Mon, 07 Apr 2003 16:33:44 -0700, Brian Quinlan wrote:
> >> What kind of business *pays* to become dependent on
> >> software they don't control?!
>  
>  I'm sure that the original poster was being facetious but...power
>  customers, telecom customers, computer customers.

I didn't read facetiousness; rather, I read in the original question a
sentiment perhaps better expressed as:

  "If a business *pays* to become dependent on software they don't
  control, what can one conclude about that business?"

or even:

  "How could it possibly be of benefit to a business to *pay* to become
  dependent on software they don't control?"

Thus, a rhetorical question perhaps.  The "What kind of business" was
not a query for what *industry* the business is in, but rather in a
spirit of "what kind of idiot".

-- 
 \      "I hope some animal never bores a hole in my head and lays its |
  `\   eggs in my brain, because later you might think you're having a |
_o__)          good idea but it's just eggs hatching."  -- Jack Handey |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From mwm at mired.org  Wed Apr  2 12:34:15 2003
From: mwm at mired.org (Mike Meyer)
Date: Wed, 02 Apr 2003 17:34:15 GMT
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk>
Message-ID: <x7u1dgiz1k.fsf@guru.mired.org>

Alexander Schmolck <a.schmolck at gmx.net> writes:
> "Max Khesin" <max at cNOvSisiPonAtecMh.com> writes:
> > > Yes, but humans aren't good optimizing compilers.
> > So who writes optimizing compilers? :).
> Humans. If they were good optimizing compilers themselves why would they
> bother?
> 
> Whether e.g.  assembler code can theoretically achieve better performance than
> fortran code is irrelevant. What matters is whether it is humanly possible to
> hand-code such assembler code.
> 
> For most cases it isn't

I disagree with that statement. Clearly, *someone* knows all the
coding tricks used by the compiler - because they put it into the
compiler. So they can achieve the same level of performance. Unless
you believe that optmiizing compilers produce the best possible
performance - in which case, why are people still doing research on
them? - then humans can probably improve the code in some way, meaning
it will be faster.

The criticial question isn't whether it's possible for a human to
achieve better performance than compiled code. The critical question
is whether the savings from the faster human-produced code is worth
the cost of producing the code. Since producing the code requires
highly specialized, non-portable - and hence expensive - knowledge and
a lot of time per operation, the answer is often "no".

To drag this thread back to the topic, this same argument is the one
seen here for why Python beats C. If your application is "fast enough"
in Python, then any time spent optimizing it in C represents 0
savings, and is a bad investment.

This is why you should write in Python, profile, tune in Python,
profile again, and then resort to C or Fortran. Make sure your
expensive developer time is going to where it makes a difference.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From andrew-pythonlist at puzzling.org  Sun Apr 27 04:06:22 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 18:06:22 +1000
Subject: Forgetting "()" when calling methods
In-Reply-To: <bOLqa.13827$K35.455183@news2.tin.it>
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <mailman.1051423062.28410.python-list@python.org> <bOLqa.13827$K35.455183@news2.tin.it>
Message-ID: <20030427080621.GN26530@frobozz.local>

On Sun, Apr 27, 2003 at 07:38:47AM +0000, Alex Martelli wrote:
> Andrew Bennetts wrote:
>    ...
> > For that matter, removing __nonzero__ from functions still doesn't help
> > this case:
> > 
> >     sys.exit
> > 
> > So it's of pretty minimal benefit.
> 
> Not really.  Write the following erro.py file:
> 
> import sys
> sys.exit
> 
> and now run:
> 
> [alex at lancelot booksa]$ pychecker erro.py
> Processing erro...
> 
> Warnings...
> 
> erro.py:2: Statement appears to have no effect
> [alex at lancelot booksa]$
> 
> See?  PyChecker CAN and DOES warn about "this case"!  But OF COURSE
> it can't help with the following erro1.py:

Sure, but I said: "removing __nonzero__ from functions still doesn't help
this case".  You can catch that today, with PyChecker, as you've demonstrated.
And that's great -- that's helped me perhaps once or twice, iirc, back when
I used to do some VB in between Python work, and occasionally got confused
for a moment.

It's still not a justification for breaking __nonzero__ on functions.  There
are plenty of cases where "if func:" is perfectly valid, and no amount of
"SINERGY" is going to convince me that making that fail (either in PyChecker
or at runtime) is a good idea :)

-Andrew.





From mcherm at mcherm.com  Fri Apr  4 13:43:56 2003
From: mcherm at mcherm.com (Michael Chermside)
Date: Fri,  4 Apr 2003 10:43:56 -0800
Subject: Interactive loop with InteractiveInterpreter
Message-ID: <1049481836.3e8dd26c822a1@mcherm.com>

> Hmm replying to myself here but I just figgured out a way that maybe is
> good... or bad... ad an extra layer before I call
> 
>     InteractiveInterpreter.runcode()
> 
> If I am about to send a for statement or a while statement to the
> interpreter I skip that line and set up my own loop with the parameters from
> the the loopline and then feed each line to the interpreter several times,
> handling the looping myself.

If ALL you have to handle is the code:

    for i in list:
        print i

Then I think you can probably "set up your own loop". But I hope you
realize that this doesn't generalize very well. If you had this:

    x = 1
    while complexFunc(x):
        x = doFirstThing(x)
        x = doSecondThing(x)

Then I'd be VERY surprised if you could "set up your own loop" --
basically you'd have to write your own python interpreter!

Fortunately, are correct that looking at debugging hooks is the
way to go for what I _think_ you want. (I think you want an interactive
display that stops after each "line" of code, right?) Whether or
not that's done via debuggermodule, I'm not sure, but if you explore
the docs a little and still have questions, I'm sure others here can
answer them.

-- Michael Chermside




From robin at jessikat.fsnet.co.uk  Fri Apr 11 12:47:11 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Fri, 11 Apr 2003 17:47:11 +0100
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
 <d116fbae.0304100704.68ef4833@posting.google.com>
 <c48$3e960655$516049d2$3498@nf1.news-service.com>
 <3e96efac$0$49107$e4fe514c@news.xs4all.nl>
Message-ID: <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk>

In article <3e96efac$0$49107$e4fe514c at news.xs4all.nl>, Irmen de Jong
<irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> writes
>Duncan Grisby wrote:
......
>
>To add some more to the mix, I benchmarked Pyro (3.2) :
>
>Connecting to server
>Time to connect to server, 0.026004
>
>Sending a long string to the server
>Time to send a string of 21000 chars, 0.003536
>
>Recieving a long stirng from the server
>Time to receive a string of 22000 chars, 0.003387
>
>Sending lots of ints to the server
>Time to send 5000 ints, 10.057741 (0.002012 per call)
>
>I find this very fast for a pure Python solution....
>
>I also measured the message sizes with tcpdump as mentioned in the article:
>
>Actual message size sending 1,000 characters: 1390
>Actual message size sending 100 integers: 27698 (with CORBA amongst the 
>smallest)
.....was this on the same machine? I find things seem to be awful off
the machine even if I pass in a dotted numeric address. I'm getting the
idea that I'm doing name lookups even if there's no apparent need. 
>--Irmen de Jong
>

-- 
Robin Becker



From jkraska at san.rr.com  Sat Apr 12 22:28:48 2003
From: jkraska at san.rr.com (Courageous)
Date: Sun, 13 Apr 2003 02:28:48 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <1c8pot8sof9m3$.1a39huke99f0m.dlg@40tude.net> <GC0ma.5919$LB6.169092@news1.tin.it>
Message-ID: <koih9vo3vdvifa8i4bdbo4dp52110i1m42@4ax.com>

>people, this particular trap comes as a surprise (it did to me the
>first time I met it [which wasn't in Python, but in a C++ library

Ah, but I wasn't so intelligent. I knew all about immutability,
encountered it before, and still wrote a bunch of code that did
aggressive string concatentation. Thankfully, Python's profiler is
so easy to use. :-)

C//




From aleax at aleax.it  Thu Apr 17 06:52:54 2003
From: aleax at aleax.it (Alex Martelli)
Date: Thu, 17 Apr 2003 10:52:54 GMT
Subject: Python in HTML files
References: <mailman.1050550654.3607.python-list@python.org> <_2ednU2VIv5QqQOjXTWcqw@comcast.com>
Message-ID: <aIvna.39620$T34.1116151@news2.tin.it>

Em wrote:

> I'm trying to get Python commands to run from within HTML files (a la
> PHP) or at least simulate that behavior.
> 
>  From what I've found so far it looks like I need to have a lot more
> permissions than I do. Basically I have access to a /public_html
> directory with a CGI-BIN and that's it. I can't copy any files outside
> of my /public_html directory.
> 
> So my question is, do you know of any technologies like this that are
> not Servlets and do not copy files outside a locked-down user's profile?

You could have your CGI script execute any templating utility of your
choice (including the toy-level YAPTU I wrote for the Cookbook, THAT
one at least shouldn't be a problem to embed in your scripts or put
side by side with them, since it's so tiny -- 80 lines or thereabouts
once you remove the test & docs).  Would that be OK...?  It wouldn't
give you "request", "response", and "session" objects, though -- just
the embedding.  Your CGI scripts could also use the normal cgi module
just to parse incoming HTML forms, and perhaps Cookie and e.g. shelve
to simulate some minimal session-continuity, if you need that.

But, I'm not aware of any already-integrated framework that puts these 
few pieces together seamlessly and is designed to work under the strict
deployment constraints you specify -- looks like a fun project for
a free weeekend... if you have any such thing;-).


Alex




From paramahlawat at yahoo.com  Fri Apr  4 11:24:51 2003
From: paramahlawat at yahoo.com (param ahlawat)
Date: Fri, 4 Apr 2003 08:24:51 -0800 (PST)
Subject: test_hotshot crashed 
Message-ID: <20030404162451.84639.qmail@web11101.mail.yahoo.com>

Hi!!

I've build python-2.2.2 on NEC SX machine which
doesn't
support shared libraries.
I've also included the "hotshot" module through
Modules/Setup file.
But while running the test suite through "make test" ,
i am getting the test failure as:

 test test_hotshot crashed -- exceptions.SystemError:
_PyImport_FixupExtension: module hotshot not loaded

and actually what is happening is:

"PyDict_GetItemString(modules, name)" is returning
NULL in "/Python/import.c" file.

any pointers to what should i do to correct this ??

regards,
Param.



__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com




From bharaths24 at yahoo.com  Thu Apr 17 09:54:11 2003
From: bharaths24 at yahoo.com (Bharath Sreekanth)
Date: Thu, 17 Apr 2003 06:54:11 -0700 (PDT)
Subject: Audit tools
Message-ID: <20030417135411.61020.qmail@web13801.mail.yahoo.com>

Hello Everyone,I want to do an architectural audit on python. Can you suggest me some tools that are ideal. Also, is it possible to get the design documents for python. Thanks, Bharath


---------------------------------
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030417/aaebf9b4/attachment.html>

From richardd at hmgcc.gov.uk  Tue Apr 15 04:15:14 2003
From: richardd at hmgcc.gov.uk (Richard)
Date: Tue, 15 Apr 2003 09:15:14 +0100
Subject: Decoding / Encoding ASN1 BER?
References: <3e9a5fc6@mail.hmgcc.gov.uk> <3E9A6E11.8B5D9191@engcorp.com>
Message-ID: <3e9bbf94$1@mail.hmgcc.gov.uk>

"Peter Hansen" <peter at engcorp.com> wrote in message
news:3E9A6E11.8B5D9191 at engcorp.com...
> (which might be something you require as part of a package being "most
popular"),

Bingo...... I honestly just wanted some personal opinion on what people out
there are using, that IMO is normally what you expect when you post a
request for a popular solution in a technical newsgroup.

A Google search returns websites containing solutions, but very rarely will
you find a website proclaiming that they are not the best solution and
recommending visitors to go elsewhere, hence the reason I asked in the
newsgroups.





From bokr at oz.net  Thu Apr 17 19:45:41 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 23:45:41 GMT
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr> <mailman.1050609281.18019.python-list@python.org>
Message-ID: <b7neb5$g6a$0@216.39.172.122>

On 17 Apr 2003 14:54:36 -0500, Ian Bicking <ianb at colorstudy.com> wrote:

>On Thu, 2003-04-17 at 02:55, Jean-Paul Roy wrote:
>> - Python is bad : tail recursion is not iterative (also astonished at 
>> that, I can understand with an interpreter, but compiler ?)
>
>Some would argue that tail recursion is not good for teaching and
>understanding, because you loose the call stack.  In the likely event of
>a traceback, it can be very confusing if you don't fully indicate how
>you got there, or if points upon the path were lost in an optimization.
>
Or if the stack trace is so deep the first part scrolls away and you
give up waiting for the last part ;-)

Perphaps a trace output that would transform the traceback of, e.g.,

 >>> def recu(n):
 ...     if n>0: recu(n-1)
 ...     raise Exception, 'test'
 ...
 >>> recu(5)
 Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "<stdin>", line 2, in recu
   File "<stdin>", line 2, in recu
   File "<stdin>", line 2, in recu
   File "<stdin>", line 2, in recu
   File "<stdin>", line 2, in recu
   File "<stdin>", line 3, in recu
 Exception: test

to

 Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "<stdin>", line 2, in recu
 [ ... above 1 line repeated 4 more times ...]
   File "<stdin>", line 3, in recu
 Exception: test

or the like, could be handy?
And IWT short repeating cycles could be handled too.

Regards,
Bengt Richter



From salvatore at wanadoo.fr  Tue Apr  1 16:46:28 2003
From: salvatore at wanadoo.fr (Salvatore)
Date: Tue, 01 Apr 2003 23:46:28 +0200
Subject: Creating instance
Message-ID: <b6d191$j7s$1@news-reader12.wanadoo.fr>

Hello,

I would like to create an instance of class
from the string name of the class

class Foo:
	pass

classname = 'Foo'
instance = create_instance('Foo')


Thank you for your help

Regards

Salvatore







From essai1 at mci.local  Fri Apr 18 04:24:46 2003
From: essai1 at mci.local (News M Claveau /Hamster-P)
Date: Fri, 18 Apr 2003 10:24:46 +0200
Subject: wmi ?
References: <b7khea$kt2$1@news-reader13.wanadoo.fr> <b7liis$3s1$1@slb3.atl.mindspring.net>
Message-ID: <b7s2l3$b33$1@news-reader11.wanadoo.fr>

Merci beaucoup

Thank you very very very very very very very very very very very very very
much





From aleax at aleax.it  Wed Apr 16 16:39:40 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 20:39:40 GMT
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it> <b7kd74$jqm$1@panix2.panix.com>
Message-ID: <gcjna.35728$T34.931514@news2.tin.it>

Aahz wrote:

> In article <iCzma.16405$T34.518484 at news2.tin.it>,
> Alex Martelli  <aleax at aleax.it> wrote:
>>
>>def test():
>>    class Bunch: pass
>>    loc = Bunch()
>>    loc.count = 0
>>    def inc_count():
>>        loc.count += 1
>>    inc_count()
>>    inc_count()
>>    print loc.count
> 
> Note that Guido deeply regrets permitting one-line blocks; this is
> something he may try fixing for Python 3.0.

Forcing linebreak after each & every colon?  Oh well -- he's probably
getting bribes from large-screen and vertical-aspect-ratio monitor
manufacturers, is my guess...;-)


Alex




From norproaj at yahoo.com  Fri Apr 11 12:18:03 2003
From: norproaj at yahoo.com (Dianne Combs)
Date: Fri, 11 Apr 2003 09:18:03 -0700 (PDT)
Subject: Media/The Northeast Progressive Advocate-Journal
Message-ID: <20030411161803.6418.qmail@web20704.mail.yahoo.com>

Dear Online Reader or Editor ... If you could ...
please take a look at our new online treeless edition
newspaper, The Northeast Progressive Advocate-Journal,
which is produced exclusively by our staff to serve
the Northeast area - with additional non-mainline
media perspective. News stories and columns are
researched, thoughtful and blend politics with
society, regional and local topics, etc. - in the
tradition of the premier monthly magazines and the
cosmopolitan newspapers. We are also a non-profit, but
we do accept ads and subscription requests for online
weekly e-mail links to the latest edition - so please
take advantage of these services, provided your
interest justifies it. 
 
Here are the following online links to the newspaper
... www.norprogaj.com ... 
 
Again, if you feel its appropriate, please take
advantage of our ad service - a competitive rate sheet
is available upon request. Online subscription
requests are also currently offered at $5.00 USD a
month to help support our activities as a truly
progressive publication - committed to providing
factual news and stories that just do not appear on
the local or for that matter the national scene. 
 
We truly appreciate your time and let others know
about it ... The Northeast Progressive
Advocate-Journal.
 
... thanks again ... Walter Harry Combs,
Editor-in-Chief
 
Dianne Combs, Features Editor & Ads/Subscriptions
Contact
 
norproaj at hotmail.com 
 
(9a-5p EDT - Phone) 518-793-8501
(24/7 Fax/Phone-Voicemail) 760-462-2751 
 
? 2003 THE NORTHEAST PROGRESSIVE ADVOCATE-JOURNAL CO.
IS A NON-PROFIT COMPANY.

__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - File online, calculators, forms, and more
http://tax.yahoo.com




From robin at jessikat.fsnet.co.uk  Sun Apr 13 18:06:57 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Sun, 13 Apr 2003 23:06:57 +0100
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>
 <3e970abc$0$49101$e4fe514c@news.xs4all.nl>
 <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>
 <1e740$3e99a2f6$516049d2$15395@nf1.news-service.com>
 <2+aM4KAFkam+EwV+@jessikat.fsnet.co.uk>
 <m37k9y2l08.fsf@mira.informatik.hu-berlin.de>
Message-ID: <m5b4piAB+dm+Ewn+@jessikat.fsnet.co.uk>

In article <m37k9y2l08.fsf at mira.informatik.hu-berlin.de>, Martin v.
L?wis <martin at v.loewis.de> writes
.....
>In CORBA, you need to keep open all connections from which you expect
>a response. So if your allocated socket descriptor pool is exhausted,
>you first close those connections that have no outstanding requests.
>If that is still insufficient, you also close the socket on which you
>have been waiting for a response longest, and tell the application
>that this connection has timed out.
>
>Regards,
>Martin
so how do master slave implementations handle thousands of slaves? If
the master is really a 'master' and not a SETI type server it would seem
that these resource limits might play a part. How do GRID systems work
this?
-- 
Robin Becker



From rick.ratzel at siliconmetrics.com  Thu Apr 10 18:35:33 2003
From: rick.ratzel at siliconmetrics.com (Rick L. Ratzel)
Date: 10 Apr 2003 17:35:33 -0500
Subject: ANN: elmer 1.0.2 - run Python from C or Tcl
Message-ID: <823ckqvv4a.fsf@gt6.siliconmetrics.com>

I've released elmer 1.0.2 to SourceForge:

http://elmer.sourceforge.net

   * Elmer allows Python code to run from C, as if it was written in C.
   * Elmer allows Python code to run from Tcl, as if it was written in Tcl.

   Elmer is not a Python-to-C or Python-to-Tcl translator. Instead, elmer
generates "glue" code that embeds an interpreter for one type of language into
an extension for another.  Elmer allows function calls and various data types,
both native and custom, to pass transparently between two languages...in the
current version, those languages are Python and either Tcl or C.

   Elmer is still considered ALPHA, and the online help and examples still need
alot of work (the "test" directory in the distribution is probably a better
place for examples at this point).  Any feedback is greatly appreciated.




From tjreedy at udel.edu  Sun Apr  6 12:35:58 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Sun, 6 Apr 2003 12:35:58 -0400
Subject: Python & Music: Performance question
References: <1759581e.0304060218.164d4cd4@posting.google.com>
Message-ID: <qVmdnVkRYYTsyA2jXTWcog@comcast.com>

"David Asorey ?lvarez" <forodejazz at yahoo.es> wrote in message
news:1759581e.0304060218.164d4cd4 at posting.google.com...
> Hello, group.
>
> I'm developing a small music program, and I have some doubts.
>
>    I want define a class, Note, which represents a note. It has two
> attributes, pitch and duration (quarter note, half note, etc). There
> will be a lot of instances of this object, maybe hundreds of notes,
> and the program needs access to all of this objects and read its
> attributes.

An alternative would be to simply model a note as a tuple (pitch,
note).
More compact and faster.

Terry





From lekishvili at python.qartu.com  Sun Apr  6 07:21:18 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 6 Apr 2003 04:21:18 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>	 <mailman.1049530929.9225.python-list@python.org>	 <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org>
Message-ID: <a3b5b608.0304060321.6f17ab@posting.google.com>

Ian Bicking <ianb at colorstudy.com> wrote in message news:<mailman.1049571010.25625.python-list at python.org>...
> On Sat, 2003-04-05 at 12:49, Giorgi wrote:
> > The question is that the only way to establish Python as a language
> > for corporate level web-programming is to have the standard approach
> > if not introduced, then at least supported by the Python consortium...
> 
> I don't think that's entirely true.  It's difficult for *Python* to be a
> language for certain areas of web programming (perhaps corporate), but
> certain Python environments could still achieve that.  The corporation
> would choose an environment, not a language, and if a good environment
> is written in Python then indirectly they'll have chosen Python.
> 

That's what I wanted to say.

> But I'm not saying it wouldn't be nice to have more cohesion... that's
> just not the way it is now.  It's harder to make it happen, because
> unlike Java or .NET, the Python community is not a top-down community. 
> Guido has some authority, but I think he's almost as much a decision
> process as a decider -- he decides on things where there's some level of
> consensus in the community, and doesn't make dictates otherwise.  The
> standard library has some authority, but its fairly neutral on these
> things.  Besides that there's no authority at the moment.
> 
Well, maybe we should establish an advisory board? :)

The issue is that an experoenced programmer will hardly ever change
his favorite framework; however, a newbie will have a standard
starting point. I am not a web-programmer and the only reason I went
relatively deep is my task to make a Georgian-language based site for
Python. I have a standard cgi module. At least it's there, but...

Therefore, I don't think that the consensus of the whole Python
community is really required...

> The next step I can envision would be for Twisted to become the
> canonical infrastructure, where higher-level web frameworks are all
> built ontop of it.  In that way we can start to build consensus one
> piece at a time.
> 

Ok, but who is (are) taking over?



From ebolonev at front.ru  Tue Apr  1 06:43:03 2003
From: ebolonev at front.ru (Egor Bolonev)
Date: Tue, 1 Apr 2003 22:43:03 +1100
Subject: HTMLView widget for Tkinter
Message-ID: <15436660484.20030401224303@front.ru>

Hello all,

I need to show html e-mail in my Tkinter program. Is it possible?


Regards,
Egor





From arne at rfc2549.org  Wed Apr 30 08:07:03 2003
From: arne at rfc2549.org (Arne Schwabe)
Date: Wed, 30 Apr 2003 14:07:03 +0200
Subject: Can urllib be used with nonblocking sockets?
Message-ID: <877k9c2nq0.fsf@r2d2.rfc1149.org>

Hi,

I have written a little python program that uses non-blocking sockets
with select and I now want to urrlib in my program and I cannot find a
method in urllib that would allow me to tell urllib that it should use
non blocking sockets.

Is there a easy solution to this? One solution would to do implement
something like a non blocking urllib but I don't want to do that.

Arne



From sjmachin at lexicon.net  Tue Apr 22 04:27:36 2003
From: sjmachin at lexicon.net (John Machin)
Date: 22 Apr 2003 01:27:36 -0700
Subject: When not to use an RE -- an example
References: <c76ff6fc.0304191459.2adf73a@posting.google.com> <BC1pa.23530$5f4.9990@twister.nyroc.rr.com>
Message-ID: <c76ff6fc.0304220027.9042e49@posting.google.com>

"Chris" <password.april03.20.netman at spamgourmet.com> wrote in message news:<BC1pa.23530$5f4.9990 at twister.nyroc.rr.com>...
> What's wrong with:
> 
>     def findrepeat(s):  return len(s) > 1 and not filter(lambda x: x !=
> s[0], s)
> 
> Chris
> 

What's wrong with it? It rates almost as high on the ugliness scale as
the crazy RE that I posted. A guess on efficiency: It's slow.
Rationale: My minimalist proposition (repeated below) takes two
substrings and then compares them, giving up as soon as a non-repeat
is detected [most likely case in my application]. Yours takes no
substrings, but has the overhead of calling the lambda function for
each & every character in the string -- it doesn't terminate early.

> > def repeats2(s):
> >    return len(s) > 1 and s[1:] == s[:-1]



From jkraska at san.rr.com  Sat Apr 19 11:49:21 2003
From: jkraska at san.rr.com (Courageous)
Date: Sat, 19 Apr 2003 15:49:21 GMT
Subject: Stackless 3.0 alpha 1 at blinding speed
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org>
Message-ID: <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com>

>No, he was not. He believes continuations are an insane feature,

Continuations are a _Cool_ feature, Christian. But try not to worry
about the rejection. If you pause and think for a while, really
think, they actually don't fit well with Python's mandate of being
a "simple, approachable language for everyone."

I'm recalling an email message you and I exchanged where you
expressed surprise and a little bit of incredulity that I
grokked what you did. Probably Python's not the language for a
construct where the author expresses surprise that seasoned
programming veterans "get" his work, you know?

So while I am guessing, I suspect the teams final decision
probably came down to deciding that continuations were too
complex and rarified a programming instrument to justify core
changes to the python interpreter and libraries.

C//

ps, I am quite disappointed that you're not supporting any kind
of continuation in future versions of stackless.





From anthony at interlink.com.au  Mon Apr  7 05:39:27 2003
From: anthony at interlink.com.au (Anthony Baxter)
Date: Mon, 07 Apr 2003 19:39:27 +1000
Subject: How to protect python code ? 
In-Reply-To: <u8bka.651$001.591065@monger.newsread.com> 
References: <u8bka.651$001.591065@monger.newsread.com>  <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> 
Message-ID: <200304070939.h379dRe00454@localhost.localdomain>

>>> msw wrote
> I'm really interested in this as well, and I think that it's an 
> important issue for python.  Unfortunately, as I'm sure you've 
> discovered by now, it's a question that comp.lang.python is 
> constitionally incapable of answering. 

I think that's a bit harsh. I've not seen anyone in c.l.py who needed
it and who was willing to do the work and release it and maintain it. 
This is open source - you scratch your own itches. For what it's worth, 
the proprietary fore-runner to Zope (Principia) had something like this. 
I have no idea if the work they did on this was ever released. I know of 
at least two other people who've done something similar with "encrypted" 
byte codes, but they never released their work (I don't think either of 
them still use this.)

If it's an important issue to you, then consider contributing 
(effort, money, whatever) to make it happen. Simply saying on c.l.py
"this is important to me" doesn't achieve much.

Anthony
-- 
Anthony Baxter     <anthony at interlink.com.au>   
It's never too late to have a happy childhood.





From peter at engcorp.com  Wed Apr  2 09:20:45 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 09:20:45 -0500
Subject: Socket timeouts for HTTP connections (httplib)?
References: <9df3d134.0304020550.7e32bf6b@posting.google.com>
Message-ID: <3E8AF1BD.CB2B8789@engcorp.com>

nick wrote:
> 
> There were rumblings about Guido implementing this at a socket level
> in June 2002, but I've not been able to find anything since.
> Presumably this has not yet risen up into httplib either because the
> python.org documentation doesn't mention it.

Search for the word "socket" in this page:

http://www.python.org/dev/doc/devel/whatsnew/node18.html

-Peter



From uptonothing at yahoo.com  Mon Apr  7 12:35:13 2003
From: uptonothing at yahoo.com (Mike)
Date: Mon, 7 Apr 2003 12:35:13 -0400
Subject: Updating CVS
Message-ID: <5Nhka.139$N5.25540@mencken.net.nih.gov>

Hello,

I wrote the following python web script to update a CVS directory and I
don't receive any errors when executing it; however, the directory doesn't
update.

#!/usr/bin/python

import cgi, os, urllib, sys

print 'Content-type: text/html\n'

sys.stderr = sys.stdout

os.environ["CVSROOT"] = "/path/to/CVS"
os.chdir("/directory/that/needs/updating")
print os.system("/usr/bin/cvs -q update -P -A -d")
...

Could someone suggest a way that I might get this script to work.

Thanks,
Mike





From gh at ghaering.de  Thu Apr 24 05:37:12 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Thu, 24 Apr 2003 11:37:12 +0200
Subject: sys.argv problem in Win 2k
In-Reply-To: <bd993a2f.0304240121.7c795cf0@posting.google.com>
References: <bd993a2f.0304240121.7c795cf0@posting.google.com>
Message-ID: <3EA7B048.9070703@ghaering.de>

Anand B Pillai wrote:
> Hi pythonistas
> 
>        All my python scripts on Win 2k does not take command line arguments.
> i.e, len(sys.argv) is always 1. Because of this I cannot use any scripts which
> takes cmd line arguments. [...]

#v+
D:\tmp>copy con: test.py
import sys
print sys.argv
^Z
         1 file(s) copied.

D:\tmp>c:\Python22\python.exe test.py bla blubs
['test.py', 'bla', 'blubs']

D:\tmp>test.py bla blubs
['D:\\tmp\\test.py', 'bla', 'blubs']

D:\tmp>
#v-

Works fine here. Can you show exactly what does not work with a test case?

-- Gerhard





From thomas at xs4all.net  Thu Apr 24 07:27:39 2003
From: thomas at xs4all.net (Thomas Wouters)
Date: Thu, 24 Apr 2003 13:27:39 +0200
Subject: Problems at python.org??
In-Reply-To: <Pine.LNX.4.44.0304240713090.12990-100000@bagend.odc.vzwcorp.com>
References: <Pine.LNX.4.44.0304240713090.12990-100000@bagend.odc.vzwcorp.com>
Message-ID: <20030424112739.GN9242@xs4all.nl>

On Thu, Apr 24, 2003 at 07:15:48AM -0400, Scherer, Bill wrote:

> Pointing my browser at http://www.python.org, I receive:

> """
> No more connections can be made to this remote computer at this 
> time because there are already as many connections as the 
> computer can accept.
> """

> Has there been a sudden Python Popularity Explosion?

Hm, no... The machine is slightly more busy than usual, being fully loaded
for a change, but it doesn't look like the socket or listen limits were
reached. The load was a temporary thing, it's ~20% now and on its way
down, and I didn't have problems connecting even when the load was high. I'm
also not sure how your browser can tell you the connection limit was
reached, since I don't think that's information TCP can give :) But perhaps
it was a network glitch. Do you still have problems connecting ?

-- 
Thomas Wouters <thomas at xs4all.net>

Hi! I'm a .signature virus! copy me into your .signature file to help me spread!




From gh at ghaering.de  Mon Apr 28 20:42:54 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 29 Apr 2003 02:42:54 +0200
Subject: In which versions is list.sort stable?
In-Reply-To: <b8kg8g$1d10$1@ns.felk.cvut.cz>
References: <mailman.1051479463.462.python-list@python.org>
	<mailman.1051562029.16746.python-list@python.org>
	<b8kg8g$1d10$1@ns.felk.cvut.cz>
Message-ID: <3EADCA8E.6040101@ghaering.de>

Frantisek Fuka wrote:
> Can you please explain to a newbie what this thread means?
> 
> Should I not use SORT in Python 2.2.2? Docs do not mention anything 
> about "unstability". Are there any other "bacis" functions (and SORT 
> seems pretty basic to me) that I should avoid???

I didn't know what a "stable sort" is, either. So I looked it up:

http://www.wikipedia.org/wiki/Stable_sort

"""
A stable sort is a sort algorithm that does not change the relative 
order of elements that have equal key values. This is important for some 
algorithms (for example, radix sort).
"""

So it's not that Python's sort implementation is unstable in the sense 
of broken :-)

HTH,

-- Gerhard




From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr 11 05:56:30 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 11 Apr 2003 09:56:30 GMT
Subject: pickle and security
References: <pan.2003.04.10.18.52.25.740636@physics.purdue.edu> <m3wui2gjuw.fsf@mira.informatik.hu-berlin.de>
Message-ID: <slrnb9d4ad.1m0.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <m3wui2gjuw.fsf at mira.informatik.hu-berlin.de>, Martin v. L?wis wrote:
> "Jason Miller" <jmiller at physics.purdue.edu> writes:
> 
>> As I understand it, the major security issue with unpickling untrusted
>> sources is that it may cause python to instantiate objects, calling
>> constructors that could do just about anything.  If I only want to
>> unpickle objects that are not class instances, and (using cPickle) I set
>> find_global to None, are there any security concerns that remain?
> 
> Depending on your version of cPickle, it also contains a call to eval,
> to unpickle a string. This is believed to be safe (as only safe
> strings are passed to eval), but you may want to review that specific
> fragment of code.

What about the attack suggested in
<mailman.1049909846.24938.python-list at python.org> (which is to use
"".__class__.__class__.__subclasses__("".__class__.__base__) to access
supposedly inaccessible classes)?



From aleax at aleax.it  Tue Apr 29 16:23:59 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 20:23:59 GMT
Subject: Newbie: ( if good in self: print "Foobar" )
References: <uN1ra.15662$kc1.13753@nwrddc02.gnilink.net> <%r4ra.19022$K35.649255@news2.tin.it> <i1dra.3534$TR3.612@nwrddc04.gnilink.net>
Message-ID: <zbBra.33363$K35.1059795@news2.tin.it>

David Broadwell wrote:

> BTW: Thank you, You do some good responces to the group. I especially
> liked the 'forgot ()' thread solution even though i'm sure you got flames
> for it. I'll have to read that 'primer' on metaclasses ... loooks usefull
> and potentially code saving.

Even though I'm on the forefront of those who say metaclasses aren't
all THAT complicated -- I suggest you FIRST get a good grasp of classes 
and instances and other related aspects of Python (e.g., dicts), THEN
move on to metaclasses.  You can grasp by far most of Python's OO without
reaching all the way to metaclasses -- THEN, when you do have that grasp,
yes, metaclasses can indeed be a useful way to refactor & reuse code.


Alex




From nav at adams.patriot.net  Fri Apr  4 10:37:29 2003
From: nav at adams.patriot.net (Nick Vargish)
Date: 04 Apr 2003 10:37:29 -0500
Subject: Simulating Qubits?
Message-ID: <yyyu1deuvd2.fsf@adams.patriot.net>

An interesting article on quantum computing, which I found via
Slashdot is here: 

http://www.economist.com/printedition/displayStory.cfm?Story_ID=1682086

It's very light on technical detail, the original paper (which I'm
printing and hope to read over the weekend) is here:

http://arxiv.org/abs/cs/0103009

I'm thinking that it might be interesting to write a module that
simulates the computational abilities of qubits. The Economist article
specifically mentions object-oriented approaches as key to integrating
qubit logic with traditional computing programs.

Okay, at this point I'm just musing aloud. Has anyone else been
thinking about this, or perhaps done some actual work on the problem?

I will hopefully have more to say after I've read the paper, assuming
I'm not just plain out baffled by it.

Nick

-- 
# sigmask.py  ||  version 0.2  ||  2003-01-07  ||  Feed this to your Python.
print reduce(lambda x,y:x+chr(ord(y)-1),'Ojdl!Wbshjti!=obwAqbusjpu/ofu?','')



From mertz at gnosis.cx  Tue Apr  8 20:08:10 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 08 Apr 2003 20:08:10 -0400
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
 <+6xk+kKkXQeJ092yn@gnosis.cx> <mailman.1049829736.3719.python-list@python.org>
 <45yk+kKkXAlY092yn@gnosis.cx> <mailman.1049835314.7390.python-list@python.org>
 <E_Ika.1677$d8.1053@nwrdny02.gnilink.net>
Message-ID: <qR2k+kKkXQNS092yn@gnosis.cx>         

Carl Banks <imbosol-1049844119 at aerojockey.com> wrote previously:
|> Ahh... what I MEANT to write was, of course:
|>    '"%s"' % repr(aString)[1:-1].replace('"',r'\"')
|
|You misspelt
|     '"%s"' % repr(aString)[1:-1].replace('"',r'\"').replace(r"\'","'")

I'm not sure that's so.  The OP wanted "just like" repr(), but with
surrounding double quotes rather than single quote.  That seems like an
underspecification of whether single quotes are escaped.  If left
escaped, the result is just as Python-friendly (albeit slightly less
human-eyeball-friendly), so the simpler version is not obviously wrong.

Yours, Lulu...

--
 mertz@   _/_/_/_/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY:_/_/_/_/ v i
gnosis  _/_/                    Postmodern Enterprises         _/_/  s r
.cx    _/_/  MAKERS OF CHAOS....                              _/_/   i u
      _/_/_/_/_/ LOOK FOR IT IN A NEIGHBORHOOD NEAR YOU_/_/_/_/_/    g s





From richie at entrian.com  Thu Apr  3 07:42:28 2003
From: richie at entrian.com (Richie Hindle)
Date: Thu, 03 Apr 2003 13:42:28 +0100
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>
Message-ID: <qp5o8voto8li6p854klt7vo6dhon69hk2o@4ax.com>

[Drew]
> I submitted a pre-PEP on overloadable assignment. 

[Anders]
> Take a step back and tell us what problem you are trying to solve.
> Whatever it is, I'm sure we can think of a better solution than having
> assignment depend on whatever object, if any, happened to be
> previously bound to the same name.

I have a use case for this.  My PyMeld library lets you manipulate HTML as
though it were a collection of Python objects, like this:

>>> from PyMeld import Meld
>>> page = Meld('''<html><body>
... <textarea id="message">Type a message.</textarea>
... </body></html>
... ''')
>>> page.message = "New content."     # Change the content of an element.
>>> print page
<html><body>
<textarea id="message">New content.</textarea>
</body></html>

So the line `page.message = "New message"` changes the content of the tag
with id 'message'.  In this context, it makes sense that assigning to an
object *changes* that object, rather than *replacing* it with another.

What you can't do with PyMeld is this:

>>> message = page.message
>>> message = "Another new message"
>>> print page
<html><body>
<textarea id="message">Another new message.</textarea>
</body></html>

There are workarounds, but over loadable assignment would make this much
neater.

Disclaimer: I don't think this is sufficient reason for adding over
loadable assignment to Python (especially if it would slow down every
assignment of anything to anything, which seems likely), but you asked for
a use case and I believe this is one.  8-)

-- 
Richie Hindle
richie at entrian.com




From missive at frontiernet.net  Tue Apr 22 19:10:07 2003
From: missive at frontiernet.net (Lee Harr)
Date: Tue, 22 Apr 2003 23:10:07 GMT
Subject: methods and vars in zope objects
References: <c42e062e.0304211805.4ced1175@posting.google.com>
Message-ID: <slrnbab50i.86.missive@localhost.my.domain>

In article <c42e062e.0304211805.4ced1175 at posting.google.com>, Erik Lechak wrote:
> Hello all,
> 
> I have been beating my head against zope for a few days now.  I am
> writing python scripts, and I can't seem to find what methods and
> member variables go along with the various zope objects.  I would like
> to be able to use dir(obj), obj.__class__ or type(obj) like I do when
> I am using python everywhere else.  But It won't let me.  The
> documentation is lengthy; I read throught it, and I must have missed
> it.    I am looking at the code now ... but it's just raising my blood
> pressure.
> 
> Is there a simple script, document, manual, or button that will list
> the methods and member vars available to a certain object?  Something
> like dir().
> 
> Thanks,
> Erik Lechak


Dieter Maurer has this really cool thing called doc finder.
That may do what you need.

Also, you should find a way to read and ask questions at the 
zope at zope.org list. I follow it on gmane.




From mwh at python.net  Thu Apr 17 04:14:15 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 17 Apr 2003 08:14:15 GMT
Subject: Catching Save - newbie
References: <3e9bd039_2@news.estpak.ee> <mailman.1050405748.15567.python-list@python.org> <3e9d1cdd_2@news.estpak.ee> <3E9D5372.494AB8D9@engcorp.com> <7h3of36r3a2.fsf@pc150.maths.bris.ac.uk> <3E9D671A.C2D1CE4D@engcorp.com> <7h3fzoiqv82.fsf@pc150.maths.bris.ac.uk> <3E9D8B1B.24050D34@engcorp.com> <7h3brz6qmsc.fsf@pc150.maths.bris.ac.uk> <3E9DB4AC.E94829D9@engcorp.com> <3e9e5242_2@news.estpak.ee>
Message-ID: <7h3u1cxpmlr.fsf@pc150.maths.bris.ac.uk>

Vahur Lokk <vlokk at hot.ee> writes:

> Peter Hansen wrote:
> 
> > Michael Hudson wrote:
> 
> Ooops!
> I remember my firstever post to linux mailing list caused a month-long 
> flamewar there ;-)
> Did I do it again?

If this is your idea of a flame war, I suggest you try
comp.lang.perl.misc :-)

As far as I can tell Peter & I are merely vigorously agreeing with
each other...

Cheers,
M.

-- 
  I love the way Microsoft follows standards.  In much the same
  manner that fish follow migrating caribou.           -- Paul Tomblin
               -- http://home.xnet.com/~raven/Sysadmin/ASR.Quotes.html



From bdesth at removethis.free.fr  Sat Apr 26 19:56:03 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Sat, 26 Apr 2003 23:56:03 +0000
Subject: serveur TFTP
References: <e4d8f90f.0304240622.64cf7f8e@posting.google.com> <mailman.1051196258.26264.python-list@python.org>
Message-ID: <3EAB1C93.2000506@removethis.free.fr>

Gerhard Haering wrote:
> alex wrote:
> 
>> j'ai recup?r? une librairie TFTP ecrit en python mais j'avoue avoir un
>> peu de mal ? l'utiliser.comment faire un serveur TFTP avec? j'essai
>> d'identifier les methodes que je pourrais appliquer a un objet
>> mon_serveur par exemple :
>> [...]
> 
> 
> Ouf! C'est beaucoup de code et je ne sait rien de TFTP. Si tu lis ce 
> group un peu (ce qui est recommand? avant d'y ?crire ;-) tu va trouver 
> que la langue pr?fer?e est l'anglais. Est-ce que tu peut reposer ta 
> question en anglais?
> 
> Pour la discussion de mati?res pythonesques en francais, je recommende 
> le groupe fr.comp.lang.python.
> 
> En esp?rant que mon francais soit compr?hensible,

Tr?s compr?hensible... Dommage que tu ne participe pas plus souvent aux 
discussions sur f.c.l.python !-)

Laotseu




From max at alcyone.com  Tue Apr 22 05:17:10 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 02:17:10 -0700
Subject: pickling a class definition
References: <uy923x83o.fsf@mail.afm.dtu.dk>
Message-ID: <3EA50896.CBA88543@alcyone.com>

Brian Elmegaard wrote:

> I am pickling instances of a class, but in the unpickling script it
> seems that I have to know how the class was defined. I cannot
> pickle the class definition in one script and unpickle in another.
> 
> Is that correct?

When you pickle a class object or instance, you're not really pickling
the whole class definition, you're pickling a reference to the class
name (and instance attributes in the case of an instance, obviously). 
So Python cleverly creates an object and then sets everything up the way
the original one was.

The unpickling process literally tries to access the class the way you
normally would if you were referencing it -- this is what gets pickled. 
So if you pickle it by referencing it one way, and then unpickle it
another way, this is probably where your error is coming from.

So, for instance, say I have a file main.py, and in that file I define a
Widget class, and in main.py I have a pickleAWidget function which
creates a Widget and then pickles it to a file.  Now I have another
script, other.py, which I want to unpickle that pickled Widget.  So I
import main (because that's where Widget is), then try to unpickle the
object but get an error.  That's because in other.py, the Widget class
is main.Widget, not Widget, and so the unpickler can't find it.

So the long and short of it is, when you're pickling and unpickling
objects, make sure that you reference them in the same way on both sides
of the fence.  In this main.py/other.py example, you might make a third
module, util.py, which would contain the actual Widget definition, which
then both main.py and other.py would import (so that to both the class
would be util.Widget) and then they live happily ever after.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Honesty has nothing to hide
\__/ Joi
    Lsystem / http://www.alcyone.com/pyos/lsystem/
 A Lindenmayer systems explorer in Python.



From jmeyer at pdi.com  Tue Apr  8 14:16:26 2003
From: jmeyer at pdi.com (Jim Meyer)
Date: 08 Apr 2003 11:16:26 -0700
Subject: exceptions == errors?
In-Reply-To: <16018.64044.25718.688696@montanaro.dyndns.org>
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
	 <16018.5597.933725.584314@montanaro.dyndns.org>
	 <1049817240.3158.883.camel@utonium.pdi.com>
	 <16018.64044.25718.688696@montanaro.dyndns.org>
Message-ID: <1049825786.3174.993.camel@utonium.pdi.com>

Hello!

On Tue, 2003-04-08 at 09:34, Skip Montanaro wrote:
> Here's some timeit output against Python CVS (just cvs up'd today since
> Jeremy's and Tim's recent foray into the garbage collector).  The first set
> of trials tries both ways to test for the nonexistent attribute
> str.__nonzero__:
> 
>     >>> t = timeit.Timer(stmt="try:\n  str.__nonzero__\nexcept AttributeError:\n  pass")
>     >>> t.timeit()
>     17.419214010238647
>     >>> u = timeit.Timer(stmt="if hasattr(str, '__nonzero__'):\n  pass")
>     >>> u.timeit()
>     4.5997350215911865
> 
> The next shows access to int.__nonzero__, which does exist:
> 
>     >>> t = timeit.Timer(stmt="try:\n  int.__nonzero__\nexcept AttributeError:\n  pass")
>     >>> t.timeit()
>     1.469681978225708
>     >>> u = timeit.Timer(stmt="if hasattr(int, '__nonzero__'):\n  pass")
>     >>> u.timeit()
>     2.2288650274276733

Very cool. It's faster to try than to test for an attribute when the
attribute exists; if it doesn't, it's not. Thanks for the data as well
as the pointer to timeit, which I hadn't seen before. So, revisiting
your earlier example:

On Mon, 2003-04-07 at 17:20, Skip Montanaro wrote:
>     class Finished(exception): pass
> 
>     try:
>         for foo in some_list:
>             for bar in some_other_list:
>                 ...
>                 if conditions_met:
>                     raise Finished
>                 ...
>     except Finished:
>         pass

...and granting that the alternatives are:

  for foo in some_list:
    for bar in some_other_list:
      ...
      if conditions_met:
        break
    if conditions_met:
      break

...which is ugly or:

  fooCtr = 0
  barCtr = 0
  while not conditions_met :
    if not barCtr < len(some_other_list) :
      fooCtr += 1
      barCtr = 0
    if not fooCtr < len(some_list) :
      break
    ...

...which seems less ugly but more clunky, I think I'll tend to prefer
the exception-based code from now on and reserve the alternatives as
performance optimizations.

Thanks for the education!

--j
-- 
Jim Meyer, Geek at Large                                  jmeyer at pdi.com





From aahz at pythoncraft.com  Wed Apr 16 01:42:00 2003
From: aahz at pythoncraft.com (Aahz)
Date: 16 Apr 2003 01:42:00 -0400
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com> <ugham-t2f.ln1@grendel.myth> <fi%la.7870$T34.222571@news2.tin.it> <lhvmm-7c7.ln1@grendel.myth>
Message-ID: <b7iqf8$6qv$1@panix1.panix.com>

In article <lhvmm-7c7.ln1 at grendel.myth>,
Jim Richardson  <warlock at eskimo.com> wrote:
>On Sat, 12 Apr 2003 21:11:39 GMT,
> Alex Martelli <aleax at aleax.it> wrote:
>> Jim Richardson wrote:
>>> On Sat, 05 Apr 2003 20:13:45 +0100,
>>>  Simon Faulkner <news at titanic.co.uk> wrote:
>>>> 
>>>> I have a list of about 5000 numbers in a text file - up to 14 digits
>>>> each.
>>>> 
>>>> I need to check for duplicates.
>>>> 
>>>> What would people suggest as a good method?
>>> 
>>> In python, just stuff them all in a dictionary, any repeats, will be
>>> eliminated. But this is rather crude and probably slow. But it would
>>> work.
>> 
>> Anything but slow!  Python dictionaries are quite fast.  But removing
>> duplicates is not the same as 'checking for duplicates' -- Simon
>> might rather want (e.g.) a list of all numbers that WERE in fact
>> duplicate.  A script that plays with a Python dict is still no doubt
>> the right solution, but it's hard to write one without more precise
>> specifications regarding what is desired.
>
>yeah, I didn't look at the check for part, I just parsed it as get rid
>of... <sigh> must need a brain upgrade. 
>
>I don't know how fast/slow the dict would be to tell the truth, it just
>doesn't seem that "elegant" and elegance, is often (wrongly I know)
>associated with speed.

Python dicts are canonically elegant *and* fast.  They're even
moderately efficient with memory usage (thanks to the way Python does
objects and bindings).  Don't forget that Python uses dicts internally
for much of its own management; Python dicts are one of the most highly
optimized data structures around.

Note that in Python 2.3 an answer using sets might be better, but sets
are just a thin layer over dicts.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From max at alcyone.com  Tue Apr 22 17:17:39 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 14:17:39 -0700
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org>
Message-ID: <3EA5B173.9420324C@alcyone.com>

Ian Bicking wrote:

> I've been using True and False for a while now, and I put this at the
> top of my files:
> 
> True, False = 1==1, 0==1
> 
> But people seem concerned about this with Python 2.3.  I don't think
> they should be concerned, while I'm reassigning True and False, I'm
> not
> redefining them.

There was talk about instituting warnings for rebinding True and False
to any value, like is currently done for None:

Python 2.3a2 (#1, Feb 19 2003, 20:06:49) 
[GCC egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> None = 'asdf'
<stdin>:1: SyntaxWarning: assignment to None

If done, then a blanket statement like yours above would start
generating warnings eventually; they'd be harmless warnings, but they'd
be unnerving and annoying.

> import sys
> if sys.version_info < (2, 3):
>     True, False = 1==1, 0==1

This would be okay (although the untyped version of the True and False
builtins were introduced in 2.2, not 2.3).  The solution I've adopted
is:

	try:
	    True
	except NameError:
	    True, False = 1, 0

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ There are defeats more triumphant than victories.
\__/ Montaigne
    Discord / http://www.alcyone.com/pyos/discord/
 Convert dates from Gregorian to Discordian.



From evan at 4-am.com  Mon Apr 28 11:45:00 2003
From: evan at 4-am.com (Evan Simpson)
Date: Mon, 28 Apr 2003 10:45:00 -0500
Subject: please comment on technologies
In-Reply-To: <3EABE03B.41C50F07@hotmail.com>
References: <slrnbalplq.254.luc@trillian.dont-panic.info> <HqCqa.8931$3M4.241359@news1.tin.it> <slrnbalv50.254.luc@trillian.dont-panic.info> <3EABE03B.41C50F07@hotmail.com>
Message-ID: <3EAD4C7C.3000101@4-am.com>

Alan Kennedy wrote:
> If you want to use the best and most pythonic markup templating language
> in existence (IMHO ;-) look no further than Template Attribute
> Lanaguage, aka TAL. This was created by the BDFL (can't get more
> pythonic than that), and solves all of the problems mentioned above. TAL
> used to be a ZOPE-only thing, but now there are multiple standalone
> implementations available. A list of them can be found in this Usenet
> posting. 

As Official Cheerleader for ZPT (and by extension, TAL) I must point out 
that it was actually created by a team at Zope Corp. that *included* 
Guido, and is based on ideas from both Hiperl?gica's HiperDOM and 
Enhydra's XMLC.  A maintained list of implementations (plus pointers to 
documentation) can be found at:

http://dev.zope.org/Wikis/DevSite/Projects/ZPT/FrontPage

Cheers,

Evan @ 4-am






From skip at pobox.com  Sat Apr 12 09:38:35 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sat, 12 Apr 2003 08:38:35 -0500
Subject: readline clear buffer
In-Reply-To: <FWRla.4654$T34.116245@news2.tin.it>
References: <slrnb9edf0.tds.luc@grizzly.dont-panic.info>
        <FWRla.4654$T34.116245@news2.tin.it>
Message-ID: <16024.5851.431152.541638@montanaro.dyndns.org>

    Alex> Right: it seems that Python's readline.c does not expose
    Alex> interfaces to rl_delete_text, rl_kill_text, or other GNU Readline
    Alex> functions that would allow such behavior.

GNU readline has a huge API (I count over 340 extern declarations in
readline/{readline.h,history.h}, the two programmer-visible #include files).
I believe the approach all along has been to expose just those parts of the
API necessary for the current task at hand.  I realize this means the
interface will be incomplete, but it's not clear that most of it is
necessary for common tasks.  So, yes, submit a patch which includes wrappers
for the functions you're interested in.

Skip




From staschuk at telusplanet.net  Tue Apr 29 16:37:04 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 14:37:04 -0600
Subject: Dr. Dobb's Python-URL! - weekly Python news and links (Apr 29)
In-Reply-To: <vatflcjn7dum95@corp.supernews.com>; from staschuk@telusplanet.net on Tue, Apr 29, 2003 at 06:04:28PM -0000
References: <vatflcjn7dum95@corp.supernews.com>
Message-ID: <20030429143703.C1336@tibia.amotlpaa.bogus>

Quoth I:
  [...]
>     Jack Diederich exhibits two real-life uses for metaclasses:
>     registering classes with a factory, and named prototypes.
>         <http://groups.google.com/groups?th=296a0181247593f3>

It looks like Google has lost this thread since I assembled the
list of links.  (In fact, I can't find on Google any post to
c.l.py from 27 April...)  The other Google groups links in the
previous list are still good, as of the time of this post.

Jack Diederich's post can be found in the python.org archives:
    <http://mail.python.org/pipermail/python-list/2003-April/160318.html>

-- 
Steven Taschuk                                                   w_w
staschuk at telusplanet.net                                      ,-= U
                                                               1 1




From robin at jessikat.fsnet.co.uk  Sun Apr 13 12:55:20 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Sun, 13 Apr 2003 17:55:20 +0100
Subject: tuple weakref
References: <vwPqrLAR9Um+EwW5@jessikat.fsnet.co.uk>
 <m38yuepdva.fsf@mira.informatik.hu-berlin.de>
Message-ID: <LtFzMJA4ZZm+EwDh@jessikat.fsnet.co.uk>

In article <m38yuepdva.fsf at mira.informatik.hu-berlin.de>, Martin v.
L?wis <martin at v.loewis.de> writes
>Robin Becker <robin at jessikat.fsnet.co.uk> writes:
>
>> What's the correct incantation to get a weak reference to a tuple? The
>> rather obvious wrapper approach fails as the wrapper dies immediately
>> 
>> ie weakref.ref(Wrapper((....)))
>
>What is 'Wrapper'?
>
>Tuples are not weakly-referencable, unless you change the source code
>of tupleobject.c.
>
>Regards,
>Martin

class Wrapper:
        def __init__(self,p):
                self.p = p

Well I actually want to do
        
        self.parent = weakref.ref(parent)

inside a class __init__, unfortunately parent is a tuple (or None).
Strictly speaking with gc I can do without the weakref as gc will
resolve and break any loop. As you say tuples (and the None object)
aren't weakrefable (for some reason).

The trivial silly solution is to attempt to hide the tupleness inside
another class, but that won't work as the wrapper dies immediately.

The tuples are coming from pyRXP's tuple tree representation.  
-- 
Robin Becker



From bbfoto2002 at netscape.net  Tue Apr  1 22:23:17 2003
From: bbfoto2002 at netscape.net (BBFoto)
Date: Wed, 02 Apr 2003 03:23:17 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <3E88C1D8.3070305@motorola.com>
Message-ID: <3e8a57d0.55986091@news-server>

I copied your statement, changed the path and tried it.

"Syntax error" was the result.

Thanks for trying to help.

BB

On Mon, 31 Mar 2003 16:31:52 -0600, Stephen Boulet
<stephen.boulet at motorola.com> wrote:

>You could fire up a "cmd" console, and type in:
>
>   C:\python22\python "C:\program files\pysol\pysol.py"
>
>Adjust the paths to fit your setup. Or you could associate ".py" files with 
>the python executable.
>
>-- Stephen
>
>BBFoto wrote:
>> This question is so basic and I am so lacking in this area, that I
>> will completely understand if I am ignored.  
>> 
>> I received a cd from PC World with a game called PySol, among other
>> things.  It says I need Python to play it.  It gives a reference, so I
>> get Python and I load the game to my HD.  (Win98 platform)   Nowhere
>> in the game does it tell me how to get it to run.  It refers to "drop
>> it on Python."  I have no idea what that means.  
>> 
>> I run the Python executable and type in "PySol" at the prompt on the
>> black screen and get "name not defined."  I can't find a .bat or .exe
>> file for PySol and I am totally unfamiliar with Python.
>> 
>> If anyone could give me a clue, it would be greatly appreciated!
>> 
>> BBfoto
>
>




From ecastro at calarts.edu  Tue Apr 22 14:06:26 2003
From: ecastro at calarts.edu (Eduardo Castro)
Date: Tue, 22 Apr 2003 11:06:26 -0700 (PDT)
Subject: Installing MySQL-python-0.9.2 on Solaris 2.9
In-Reply-To: <3EA58024.1050909@ghaering.de>
Message-ID: <Pine.GSO.4.21.0304221101360.22586-100000@muse.calarts.edu>

Hi, has anyone had problems installing MySQL-python-0.9.2 on Solaris 2.9?

I am getting the following error when I run 'setup.py build'

ld: fatal: library -lmysqlclient: not found


This is part of the error messages:

building '_mysql' extension
skipping _mysql.c (build/temp.solaris-2.9-sun4u-2.2/_mysql.o up-to-date)
gcc -shared build/temp.solaris-2.9-sun4u-2.2/_mysql.o -L/usr/lib/mysql
-L/usr/local/lib/mysql -L/usr/loca/mysql/lib -L/usr/local/mysql/lib/mysql
-L/usr/local/mysql/include -L/usr/bin
-Wl,-R/usr/local/lib:/usr/openwin/lib:/usr/dt/lib -lmysqlclient -lz -o
build/lib.solaris-2.9-sun4u-2.2/_mysql.so
ld: fatal: library -lmysqlclient: not found
ld: fatal: File processing errors. No output written to
build/lib.solaris-2.9-sun4u-2.2/_mysql.so
collect2: ld returned 1 exit status
error: command 'gcc' failed with exit status 1


Thanks

Eduardo Castro
California Institute of the Arts.







From googlegroups at tekNico.net  Sat Apr 19 11:56:57 2003
From: googlegroups at tekNico.net (Nicola Larosa)
Date: 19 Apr 2003 08:56:57 -0700
Subject: Web Hosting
References: <2rjna.171$hT2.182473@news2.news.adelphia.net> <20030419005630.O28553-100000@eden.ispol.com>
Message-ID: <6fb43e5d.0304190756.126d3bd5@posting.google.com>

> More and more hosting companies pop up that offer FreeBSD jail(8) based
> "Virtual Servers". You can compile and run anything your heart desires
> because you get root access to the "server" and it's reasonably priced.

There's even one of those in Italy, with a very good price too:

Areaserver
http://www.areaserver.it/

A direct shell access is very useful to correct a number of system
problems too, apart from programming.

DISCLAIMER: I have no interest nor relationship with Energit, I'm just
a satisfied customer.



From max at alcyone.com  Thu Apr 24 16:48:52 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 24 Apr 2003 13:48:52 -0700
Subject: Problems at python.org??
References: <mailman.1051183057.25161.python-list@python.org>
Message-ID: <3EA84DB4.4EC1FF12@alcyone.com>

"Scherer, Bill" wrote:

> Pointing my browser at http://www.python.org, I receive: ...

I don't.

> Has there been a sudden Python Popularity Explosion?

Any such problems are almost certain to be intermittent and short-lived.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Magnificent desolation.
\__/ Buzz Aldrin (on the Moon)
    Church / http://www.alcyone.com/pyos/church/
 A lambda calculus explorer in Python.



From fuka at fuxoft.cz  Fri Apr 25 17:46:50 2003
From: fuka at fuxoft.cz (Frantisek Fuka)
Date: Fri, 25 Apr 2003 23:46:50 +0200
Subject: Forgetting "()" when calling methods
Message-ID: <b8cadp$1fj9$1@ns.felk.cvut.cz>

When I try to call methods, I sometimes forget to include the 
parentheses. Instead of:

if object.isGreen():
	do something...

i sometimes write:

if object.isGreen:
	do something...

If I understand it correctly, the if statement in this case tests if 
pointer to hasParent method is non-zero, which is of course always True 
so no error is reported and "do something..." always gets executed, so 
the application behaves in quite different way than I expected.

You can say to me "Don't forget to always include the parenteses" but 
I'm still curious if this cannot be somehow configured, so that I get 
error when I try to just access the method pointer (".isGreen") instead 
of calling the method (".isGreen()"). I understand that there has to be 
the possibility of accessing the method pointers but I usually don't 
need it and for a beginner like me this makes the applications very hard 
to debug.




From gh at ghaering.de  Wed Apr 16 10:00:58 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Wed, 16 Apr 2003 16:00:58 +0200
Subject: Pythonic way of web-programming
In-Reply-To: <7kCdnQb9MqHQ2gCjXTWcow@speakeasy.net>
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org> <a3b5b608.0304051049.7048d76@posting.google.com> <mailman.1049571010.25625.python-list@python.org> <%5Oja.10776$7w2.9784@nwrddc01.gnilink.net> <mailman.1049653811.11491.python-list@python.org> <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <f5mcnWIBkPg9aQajXTWcpA@speakeasy.net> <mailman.1050470255.12198.python-list@python.org> <7kCdnQb9MqHQ2gCjXTWcow@speakeasy.net>
Message-ID: <3E9D621A.70902@ghaering.de>

A.M. Kuchling wrote:
> On 16 Apr 2003 00:16:49 -0500, 
> 	Ian Bicking <ianb at colorstudy.com> wrote:
> 
>>One option would be to use an interface with adapters, where if say the
>>Quixote interface was used, Webware would have an adapter from that
>>interface to the Webware HTTP{Request,Response} interface... though I'm
>>not sure how interesting request and response object are.  
> 
> 
> Perhaps they're not very interesting, but that seems the only level that
> could be feasibly standardized.
> 
> I'm willing to host a mailing list for web discussion on my DSL-connected
> machine. 

I'd be interested in joining that list.

> That might be easier than jumping through the SIG hoops.  If
> you're interested, just suggest a name ('web'?) for the list and I'll create
> it. 

web-frameworks?

> I can't do CVS, but maybe we could just use a sandbox/ directory in the
> Python CVS tree.

Or, for that matter, any other Sourceforge project. I admin a few ones 
we could (ab)use for this purpose. This has the benefit that the 
commiters need not be Python committers.

> [...] Yes, but usually those areas of duplication don't actually require a lot of
> maintenance effort after the initial implementation.  We're four months into
> 2003, and Quixote's HTTPRequest and Response haven't had a significant
> change yet.  The changes have all been to higher-level things like the
> publishing logic.

We could also discuss the feasability of creating a common interface for 
form handling and session management.

-- Gerhard





From mathias-usenet at valpo.de  Wed Apr 23 09:11:42 2003
From: mathias-usenet at valpo.de (Mathias Waack)
Date: 23 Apr 2003 13:11:42 GMT
Subject: i18n in a library
References: <b85hjt$6l2na$1@ID-15245.news.dfncis.de> <jdspa.47509$LB6.1087058@news1.tin.it> <b85l43$dhf$1@zeppelin.rz.uni-potsdam.de>
Message-ID: <b863ee$6ndio$1@ID-15245.news.dfncis.de>

"Martin v. L?wis" <martin.vonloewis at hpi.uni-potsdam.de> wrote:

>> I think what you want is gettext.translation, which offers you
>> finer-grained control than gettext.install.  I.e., within the
>> modules of library L, you could use
>>
>>     import gettext
>>     _ = gettext.translation('L').ugettext
>>
>> so that in those modules function _ will look things up in
>> domain 'L', rather than in whatever domain is "installed"
>> globally. 

Unfortunatly I need some more code (because of some historical
reasons) to setup the translation. Its very ugly to add these
lines to each source file. I don't like copy'n'paste as
programming style. 

> Alternatively, gettext.dgettext is the Python version of the
> equivalent C API, so you could also do
> 
> def _(msg):
>    return gettext.dgettext('L', msg)
> 
> This has the added advantage of returning the string itself if
> no catalog can be found (instead of giving IOError as
> gettext.translation does)

Thus I have no chance to see if the translation fails. 

But your advices (and a short view into gettext.py) lead me to
another solution: 

class Translation:
    def __init__(self, module):
        localedir = ... # its a bit special in our environment
        language = ...  # see above
        try:
            self.moduleCatalog = \
              gettext.translation(module,localedir,[language]) 
        except:
            self.moduleCatalog = gettext.NullTranslations()
            self.moduleCatalog._catalog = {}
        try:
            self.libCatalog = \
              gettext.translation('library',localedir,[language]) 
        except:
            self.libCatalog = gettext.NullTranslations()
            self.libCatalog._catalog = {}

        import __builtin__
        __builtin__.__dict__['_'] = self.gettext
    def gettext(self,message):
        return self.moduleCatalog._catalog.get(message) \
               or self.libCatalog._catalog.get(message,message)

(The c'tor has some overhead which is needed somewhere else)
This looks first into the module's catalog and if it fails into
the libraries catalog. Don't know if its nice, but it works... 

Mathias



From tim.one at comcast.net  Tue Apr 29 12:29:57 2003
From: tim.one at comcast.net (Tim Peters)
Date: Tue, 29 Apr 2003 12:29:57 -0400
Subject: The behavior of List.insert
In-Reply-To: <b8m8dn$b03$1@pegasus.csx.cam.ac.uk>
Message-ID: <BIEJKCLHCIOIHAGOKOLHCEPJFHAA.tim.one@comcast.net>

[York]
> given
>
> mylist = [1,2,3,4,5],
>
> mylist.insert(-1,99) will make mylist to be [99,1,2,3,4,5]
>
> just wonder why not let it to be [1,2,3,4,99,5]

If you read the docs, you'll see that this is the documented behavior, in
Python 2.2.2 and earlier.

However, that behavior has changed in Python 2.3:

Python 2.3b1 (#40, Apr 25 2003, 10:15:10) [MSC v.1200 32 bit (Intel)] on
win32
Type "help", "copyright", "credits" or "license" for more information.
>>> mylist = [1, 2, 3, 4, 5]
>>> mylist.insert(-1, 99)
>>> mylist
[1, 2, 3, 4, 99, 5]
>>>





From mwh at python.net  Thu Apr 24 06:40:27 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 24 Apr 2003 10:40:27 GMT
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3ea7b12a@news.012.net.il>
Message-ID: <7h3k7dkkwko.fsf@pc150.maths.bris.ac.uk>

"Avner Ben" <avner at skilldesign.com> writes:

> The first language I learnt was COBOL. That took even longer to overcome!

Of coure, Dijkstra has a quote about that, too.  Let me prod the
sigmonster:

-- 
  The use of COBOL cripples the mind; its teaching should, therefore,
  be regarded as a criminal offence.
           -- Edsger W. Dijkstra, SIGPLAN Notices, Volume 17, Number 5



From toms at f2s.com  Thu Apr  3 02:19:59 2003
From: toms at f2s.com (Tom Saunders)
Date: Thu, 3 Apr 2003 08:19:59 +0100
Subject: web server in python
Message-ID: <b6gnau$d7q$1@news.freedom2surf.net>

i am curerntly trying to make simple dynamic site in python which i intend
to expand on. i can make the html using python but i have a few questions.
using the BaseHTTPServer class, is it then possible to write all the code
for a site inside one python program, which then serves based on requests
from a website?
ive been having trouble finding information about this and i would be
greatfull if someone could give me an example of how to start.
Hope you can help

Tom





From pobrien at orbtech.com  Sun Apr  6 19:11:47 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 18:11:47 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <GCEDKONBLEFPPADDJCOEKEINJFAA.whisper@oz.net>
References: <GCEDKONBLEFPPADDJCOEKEINJFAA.whisper@oz.net>
Message-ID: <m2y92n5ih8.fsf@localhost.localdomain>

"David LeBlanc" <whisper at oz.net> writes:

> > The main reason I started PyAlaMode is that I wanted the same
> > autocompletion and calltips that I've got in the PyCrust shell, but
> > while I'm editing a file.  Those features are working now, and are
> > pretty sweet, if I do say so myself, since they are based on dynamic
> > namespaces updated whenever you want, one per buffer, with no
> > conflicts no matter how many buffers you have open, with automatic
> > reloading of imported modules.
> 
> Scite (based on Scintilla) has calltips and auto-completetion and is
> friendly towards Python. It's multi-platform too. I've often thought a
> python-scriptable Scite would be a great thing. I'd bet that wrapping (boost
> or swig) the main scite (C++) class would give you a lot of the same
> primatives as emacs. Graft on user keystroke assignment and you're well on
> the way...

PyCrust and PyAlaMode are both based on Scintilla as well.  In
wxPython the Scintilla wrapper is called wxStyledTextCtrl.  The
difference between PyAlaMode and Scite is that the autocompletion and
calltips in PyAlaMode are dynamically generated based on the namespace
of the module being edited whenever the user decides to update the
namespace.

In short, I'm already doing what you're suggesting.  ;-)

But the primitives in Scintilla aren't quite the same as the
primitives in Emacs, so there is still work to be done in taking this
to the next level.

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------




From aahz at pythoncraft.com  Tue Apr  8 09:20:10 2003
From: aahz at pythoncraft.com (Aahz)
Date: 8 Apr 2003 09:20:10 -0400
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049731219.27404.python-list@python.org> <b6teqm$956ai$1@ID-169208.news.dfncis.de> <mailman.1049776891.5548.python-list@python.org>
Message-ID: <b6uiaa$1os$1@panix2.panix.com>

In article <mailman.1049776891.5548.python-list at python.org>,
Sean 'Shaleh' Perry  <shalehperry at attbi.com> wrote:
>
>and it is.  However the reason most of the emacs modes, vim syntax
>definitions, etc all don't quite work is they are regex based.  To
>really get where the structure editor needs to be would require a
>parser for each item type.  Not a trivial task.

That's why we're using Python.  ;-)
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From dcengija_remove_ at inet.hr  Sat Apr 12 15:44:26 2003
From: dcengija_remove_ at inet.hr (Davor Cengija)
Date: Sat, 12 Apr 2003 21:44:26 +0200
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com> <b766fi$p3b$1@newstree.wise.edt.ericsson.se> <16411664.REEpLcBOk5@lothar.cengija.com> <amfd9vsugg1hd0cd8qrjdm033opugi5i6e@4ax.com>
Message-ID: <1218043.6iNpJ2gbPL@lothar.cengija.com>

Nomad wrote:

> On Fri, 11 Apr 2003 13:26:32 +0200, Davor Cengija
> <dcengija_remove_ at inet.hr> wrote:
> 
>>Now it looks 'respectable' and 'worth of
>>trying but not for serious work since it doesn't have setup.exe with EULA
>>and other things -- so it isn't a real program'. Now I'm looking for a
>>free install-shield like software.
> 
> Have a look at "Inno Setup" from "http://www.innosetup.com".  Very
> nice, and looks like an install-shield/wise install to the passing
> eye.

Inno Setup is excellent, thanks! It even provides the Uninstall feature. I 
finally became The Real Programmer!

> 
>>So, it isn't the quality of a software what only counts. Actually, I'm not
>>sure that the quality is the most important thing these days.
> 
> Of course not -- it's got to be bloated, but pretty...

Python+wxWindows combination looks pretty. It would help (in such a 
situation) to have the dropdown menus with shadows etc (a la Windows XP), 
but then I'd be required to sell my software. I mean, you cannot have a 
good program (read: pretty looking program) which is distributed for free. 
"Something fishy is going on. Don't install it, it's a virus!"

 :-)

-- 
Davor Cengija, dcengija_remove_ at inet.hr



From duncan at NOSPAMrcp.co.uk  Thu Apr 10 05:40:38 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 10 Apr 2003 09:40:38 +0000 (UTC)
Subject: Adding item in front of a list
References: <mailman.1049902101.26281.python-list@python.org>
Message-ID: <Xns93596C6F1EA0Eduncanrcpcouk@127.0.0.1>

Tim Peters <tim.one at comcast.net> wrote in
news:mailman.1049902101.26281.python-list at python.org: 

> Guido says he also wishes list.insert() had been defined with the
> arguments in the opposite order, so that list.insert(object) could
> have a natural default index argument of 0.  I'd like to change that
> too, but it's clearly too late for that one.
> 
I don't actually see the problem with this. list.insert() currently takes 
exactly two arguments, therefore there would be no backwards compatibility 
issues with allowing it to take one or two arguments where a single 
argument acts as though you had given two but the first defaulted to 0. 
Just like 'range' does in fact!

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From arosado at softhome.net  Wed Apr 16 09:22:25 2003
From: arosado at softhome.net (Andres Rosado)
Date: Wed, 16 Apr 2003 09:22:25 -0400
Subject: Python and Schools
Message-ID: <5.2.0.9.0.20030416092224.03b7abc0@mail.softhome.net>

On 05:06 AM 4/11/2003 -0400, python-list-request at python.org feed this fish 
to the sharks:
>- if you learn programming with Python and later have to use Java or
>Visual Basic, it will be the most frustrating experience for the young
>fellows.

Pardon my ignorance, but how could it be a frustating experience?


-- 
Andres Rosado
Email: andresr at despammed.com
ICQ: 66750646
Homepage: http://andres980.tripod.com/
Get my PGP key at http://andres980.tripod.com/pgp-key.asc

"Cat's tough. He went out fighting."
         -- Depth Charge, on Cheetor's apparent demise,
            "Feral Scream" Part 1  





From matthew at newsgroups.com  Tue Apr  1 18:43:00 2003
From: matthew at newsgroups.com (matthew)
Date: Wed, 02 Apr 2003 11:43:00 +1200
Subject: Question on passing function as argument
In-Reply-To: <d111eafa.0303312121.22bb903@posting.google.com>
References: <b6abiv$b6t$1@lust.ihug.co.nz> <d111eafa.0303312121.22bb903@posting.google.com>
Message-ID: <b6d80c$cho$1@lust.ihug.co.nz>

> Hope that helps!
> xtian

It did! Thanks alot. matthew.




From deja-google at wotevah.com  Tue Apr 15 13:12:57 2003
From: deja-google at wotevah.com (L.C.)
Date: 15 Apr 2003 10:12:57 -0700
Subject: more informative stack traces
References: <e0bf36b5.0304132100.4324c3b8@posting.google.com> <3e9ab723$0$66212$edfadb0f@dread11.news.tele.dk>
Message-ID: <e0bf36b5.0304150912.2067649a@posting.google.com>

You are correct, and I underestimated the number of people using Python 
on Windows :) I'll try to add a registration mechanism for output control 
and make syslog and stdout registered clients to it, so that Windows will 
at least have stdout working. I'll repost the link when done.

Just as a historical note, this code was written at our company to solve 
the problem of exception reporting in daemon-type programs, on Linux, with
minimal impact on the main code base. So once it did that, it was kind of
frozen that way :)

--
L.C. (Laurentiu C. Badea)

"Anders J. Munch" <andersjm at dancontrol.dk> wrote in message news:<3e9ab723$0$66212$edfadb0f at dread11.news.tele.dk>...
> "L.C." <deja-google at wotevah.com> wrote:
> > I am proposing a different style of stack trace which I found more 
> > useful in debugging problems with live code. The main change involves
> > showing the actual parameter values in function calls, gdb-style.
> > It sounds simple because it really is, but it saves lots of guesswork
> > when navigating through the stack to see what caused a problem.
> > 
> > An implementation is available here: http://www.wotevah.com/code/log.py
> > To enable, just import it in your main program.
> > 
> > It requires syslog (thus Unix-only) because it seemed to be the natural
> > mechanism for reporting errors and warnings, although it will also use 
> > stderr if you are on a terminal.
> 
> May I suggest you refactor your code into two modules: a
> platform-independent one that inspects tracebacks and generates the
> traceback text in some sort of container or generator.  And a
> syslogging module one with the same interface as you have now, which
> is a client of the first module.  And perhaps a third module, a client
> that uses portable stderr output.
> 
> Not only will this give you a fighting chance of convincing people on
> syslog-less platforms of the merits of your approach, but it will also
> benefit the structure of your code.  Hardwiring the output mechanism
> puts obstacles in the way of applications such as CGI scripts and GUI
> traceback presenters, where consoles and logfiles may be irrelevant
> because user interaction goes through other channels.
> 
> - Anders



From riknessa at hotmail.com  Fri Apr 18 05:43:21 2003
From: riknessa at hotmail.com (Rikko)
Date: Fri, 18 Apr 2003 11:43:21 +0200
Subject: [Q] can PySerial work with a USB to serial adapter ?
References: <xZYma.26921$c42.17977@fe01.atl2.webusenet.com>
Message-ID: <PEPna.40424$c42.5599@fe01.atl2.webusenet.com>

> I'm trying to use PySerial with Python 2.2.2 on an XP system
> with a USB to RS232 adapter.
> Upon opening the port (making an instance of the Serial class)
> the system return the error message "unable to open file",
> thereby referring to the COM1 port used.

FIXED, figured it out myself.
a usb->serial adapter driver installs itself on the next free comport
number.
For my notebook that meant that it was installed AFTER the built-in modem,
so that was COM3.
Rik.







From duncan at NOSPAMrcp.co.uk  Thu Apr 10 12:38:31 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 10 Apr 2003 16:38:31 +0000 (UTC)
Subject: Pyrex: wrapping C struct as Python class
References: <b743fp$k40$1@bob.news.rcn.net>
Message-ID: <Xns9359B35917C18duncanrcpcouk@127.0.0.1>

"Edward C. Jones" <edcjones at erols.com> wrote in 
news:b743fp$k40$1 at bob.news.rcn.net:

> In C code, I would use "malloc", "free" and "squareit" with Tiny. What 
> is the best way to wrap this in Pyrex so Python code can do
> 
>      t = Tiny(77)
>      print t.ival
>      print t.squareit()
>      del t
> 
> 
Wrapping the function is easy enough, but I think exposing the fields is a 
bit harder. Here is what I came up with:
(note minor corrections to your example)

----- tiny.h ----
typedef struct {
    int iVal;
} Tiny;

int squareit(Tiny *t);

----- tiny.c ----
#include "tiny.h"

int squareit(Tiny *t)
{
    return t->ival * t->ival;
}
----- pyTiny.pyx ----
cdef extern from "tiny.h":
        ctypedef struct c_Tiny "Tiny":
            int iVal

        cdef extern int c_squareit "squareit" (c_Tiny *t)

cdef class Tiny:
    cdef c_Tiny t

    def __init__(self, initial):
        self.t.iVal = initial

    def squareit(self):
        return c_squareit(&self.t)

    def __getattr__(self, name):
        if name=='iVal':
            return self.t.iVal
        return object.__getattribute__(self, name)

    def __setattr__(self, name, val):
        if name=='iVal':
            self.t.iVal = val
        else:
            object.__setattr__(self, name, val)
----- setup.py ---
from distutils.core import setup
from distutils.extension import Extension
from Pyrex.Distutils import build_ext

setup(
  name = 'Tiny',
  ext_modules=[ 
    Extension("pyTiny", ["pyTiny.pyx", "tiny.c"]),
    ],
  cmdclass = {'build_ext': build_ext}
)
----- test.py ---
from pyTiny import Tiny

t = Tiny(77)
print t.iVal
print t.squareit()
t.iVal = 66
print t.iVal
print t.squareit()
del t
---- phew! all done ---

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From cnetzer at mail.arc.nasa.gov  Sat Apr 19 18:32:06 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 19 Apr 2003 15:32:06 -0700
Subject: Sorting without transitivity
In-Reply-To: <20030419213405.GA588@obstler>
References: <20030419213405.GA588@obstler>
Message-ID: <1050791526.589.31.camel@sayge.arc.nasa.gov>

On Sat, 2003-04-19 at 14:34, Frank Niessink wrote:

> How can I get the desired behavior? Should I implement my own
> sort algorithm? Any pointers?

The sort algorithm does NOT compare every object against every other,
which I assume means it can't work like you want, because B and A may
never be compared, and they can't be ordered with respect to each other
based on their relation to C.  You may need to do something like a
bubblesort, where EVERY element is compared against every other
(actually, there are faster sorts than bubblesort that'll do what you
want; someone else can probably help out here.  David Eppstein?  Can you
suggest one?).

Also, note that in your testSortPoints, asking for index(a) or index(b)
is equivalent to asking for index(c).  So, even if a and b are sorted
with respect to each other, if c comes before either, your test may fail
because index(a) == index(b) == index(c).

>From section 2.2.6.4 (Mutable Sequence Types) of the Python reference
guide:
   s.index(x)  -   return smallest i such that s[i] == x

So, you will have to make a list of id #'s and find the index of the
based on their id()s.

Note that by having a list of only 'a' and 'b' instances, the sorting
WILL work.  Throwing a 'c' in there will mess everything up. (try it)

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From johnochiltree at blueyonder.co.uk  Sun Apr 27 03:17:48 2003
From: johnochiltree at blueyonder.co.uk (John Ochiltree)
Date: Sun, 27 Apr 2003 08:17:48 +0100
Subject: Problems making Python-2.3b1
Message-ID: <HuLqa.2812$BL6.1867@news-binary.blueyonder.co.uk>

Hello group,
I hope someone can help me here. I'm having trouble building Python-2.3b1 on
SuSE 8.2. The ./configure didn't seem to object to anything but running the
make bombs out with:

running build
running build_ext
Traceback (most recent call last):
  File "./setup.py", line 1110, in ?
    main()
  File "./setup.py", line 1105, in main
    scripts = ['Tools/scripts/pydoc']
  File "/home/johnoch/build/Python-2.3b1/Lib/distutils/core.py", line 149,
in setup
    dist.run_commands()
  File "/home/johnoch/build/Python-2.3b1/Lib/distutils/dist.py", line 907,
in run_commands
    self.run_command(cmd)
  File "/home/johnoch/build/Python-2.3b1/Lib/distutils/dist.py", line 927,
in run_command
    cmd_obj.run()
  File "/home/johnoch/build/Python-2.3b1/Lib/distutils/command/build.py",
line 107, in run
    self.run_command(cmd_name)
  File "/home/johnoch/build/Python-2.3b1/Lib/distutils/cmd.py", line 333, in
run_command
    self.distribution.run_command(command)
  File "/home/johnoch/build/Python-2.3b1/Lib/distutils/dist.py", line 927,
in run_command
    cmd_obj.run()
  File
"/home/johnoch/build/Python-2.3b1/Lib/distutils/command/build_ext.py", line
269, in run
    self.build_extensions()
  File "./setup.py", line 99, in build_extensions
    self.detect_modules()
  File "./setup.py", line 844, in detect_modules
    self.detect_tkinter(inc_dirs, lib_dirs)
  File "./setup.py", line 941, in detect_tkinter
    for dir in tcl_includes + tk_includes:
TypeError: unsupported operand type(s) for +: 'NoneType' and 'list'
make: *** [sharedmods] Error 1
johnoch at isis:~/build/Python-2.3b1>


Any ideas gratefully received.

John Ochiltree



From santoshi_reddy at hotmail.com  Tue Apr  1 20:19:55 2003
From: santoshi_reddy at hotmail.com (Santoshi Reddy)
Date: Tue, 01 Apr 2003 20:19:55 -0500
Subject: (no subject)
Message-ID: <F132PvoybY9epACm9NZ0000ea2e@hotmail.com>

hi I am new to python.can someone tell me what _int_ is?





_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE*  
http://join.msn.com/?page=features/virus





From dave at pythonapocrypha.com  Tue Apr  8 19:38:03 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Tue, 8 Apr 2003 16:38:03 -0700 (PDT)
Subject: Palindrome finder: help find ways to speed up?
In-Reply-To: <BAB8B22D.2022%dogen@mindspring.com>
References: <080420031230508991%camel@oasis.com> <mailman.1049824453.880.python-list@python.org>
 <BAB8B22D.2022%dogen@mindspring.com>
Message-ID: <Pine.LNX.4.51.0304081628220.1208@ns0.tbye.com>

On Tue, 8 Apr 2003, Jon Brown wrote:

> Steven,
>
> Actually the program does exactly what you describe, except from the inside
> out because it starts in the middle, and it cuts out of the loop if the
> characters don't match.
>
> To look for pals it spins off big, big numbers of combinations, true enough.
> That's what behind the need for speed. Since Python is inherently on the
> pokey side, I expected the slowness but wonder if I'm missing a cool way to
> pick up the pace. But I suspect the solution will involve C++, and I've
> never linked C/C++ with Python.

I wouldn't jump to C/C++ yet - a good algorithm in a slow language often
beats a lesser algorithm in a fast language. Reducing the number of
combinations is very important, as was already pointed out, but you should
also optimize your palindrome checker. For example, rather than testing
character by character you can test if a string is a palindrome by doing
something like (untested):

def IsPal(s):
  r = list(s)
  r.reverse()
  r = ''.join(r)
  c = len(s)/2
  return s[:c] == r[:c]

In Python 2.3 you could even do:

def IsPal(s):
  c = len(s)/2
  return s[:c] == s[::-1][:c]

Also, you should probably call lower() on your input and use strings'
translate method to elminate punctuation and whitespace because

"And E.T. saw waste DNA"

is typically considered a palindrome.

-Dave




From python at dylanreinhardt.com  Fri Apr  4 12:24:39 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 04 Apr 2003 09:24:39 -0800
Subject: Pythonic way of web-programming
In-Reply-To: <a3b5b608.0304040904.146dfff5@posting.google.com>
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
Message-ID: <1049477078.10854.35.camel@ida.dylanreinhardt.com>

On Fri, 2003-04-04 at 09:04, Giorgi wrote:

> - Is Python a real web-programming language at all? I mean, has
> somebody ever used it for real, commercial tasks? 

Yes and yes.

> - Is there a need to design a genuine python framework for
> web-programming? 

Zope is a highly Pythonic framework for building web-facing
applications.  It's been quite good for at least a couple years now...
good enough for enterprise-level systems, good enough for e-commerce.

It's not trivial to learn, but it's well worth it if you need something
more robust than CGI-style web programming tools.  


> Or, maybe, ZOPE is the genuine pythonic way? 

Perhaps... it is highly object-oriented and readily extensible with
Python.  Zope is hardly the only way, but it is worth a good look.

Some docs: http://www.zope.org/Documentation/Books/ZopeBook/2_6Edition

HTH,

Dylan






From mcfletch at rogers.com  Sat Apr 19 22:43:08 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sat, 19 Apr 2003 22:43:08 -0400
Subject: PyOpenGL glClearColor() trouble
In-Reply-To: <20030419215150.18878.00000431@mb-m18.aol.com>
References: <20030419215150.18878.00000431@mb-m18.aol.com>
Message-ID: <3EA2093C.6020606@rogers.com>

This is a "feature" of the Tk widget defined in OpenGL.TK.__init__.py .  
Basically, the tkinter widget is giving you a "simplified" environment 
for your OpenGL code, so it does silly things like setting a background 
color based on attributes of the Tk widget.  These attributes are:

        self.r_back = 1.
        self.g_back = 0.
        self.b_back = 1.

You can set those directly, call self.ogl.set_background( r,g,b ), or 
override the ogl widget's tkRedraw method at your preference (or use a 
real GUI library ;) ).  However, save for the last option, you won't 
actually avoid the glClearColor call.  (I'm rather surprised if you 
care, it's very seldom that such a call winds up having a significant 
overhead in an application).

Enjoy yourself,
Mike

TheDustbustr wrote:

>from OpenGL.GL import *
>import OpenGL.Tk as Tk
>  
>
...

>The way the script is set up right now (second glClearColor commented), glClear
>clears the screen to the default ClearColor (1,0,1,0 aka pink).  I can only
>make the ClearColor black (0,0,0,0) when I call glClearColor in the ogl.redraw
>function... This is inconvienent and I shouldn't need to be called every frame.
>  
>
... 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From oren-py-l at hishome.net  Sat Apr  5 04:32:18 2003
From: oren-py-l at hishome.net (Oren Tirosh)
Date: Sat, 5 Apr 2003 04:32:18 -0500
Subject: Why deleting large variables doesn't free up any memory?
In-Reply-To: <b6kp24$1l46$1@nntp6.u.washington.edu>
References: <169a7283.0304040936.69e2b9a1@posting.google.com> <b6khbt$ftf$1@panix2.panix.com> <slrnb8rmbn.rs.do-not-spam-ben.hutchings@tin.bwsint.com> <b6kp24$1l46$1@nntp6.u.washington.edu>
Message-ID: <20030405093218.GA16960@hishome.net>

On Fri, Apr 04, 2003 at 08:13:56PM +0000, Donn Cave wrote:
> Quoth Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com>:
> | In article <b6khbt$ftf$1 at panix2.panix.com>, Aahz wrote:
> |> In article <169a7283.0304040936.69e2b9a1 at posting.google.com>,
> |> sdieselil <sdieselil at yahoo.com> wrote:
> |>>
> |>> I'm using Python 2.2.2 under FreeBSD. When I create a large variable
> |>> (using for example "a = range(1000000)") 'top' command indicates that
> |>> Python ate 12MB of swap space. But when I execute "del a" nothing
> |>> happens, it doesn't release these 12MB! Using "gc.collect()" also
> |>> doesn't help. What's the problem? What happened to Python's garbage
> |>> collector?
> |> 
> |> That's got nothing to do with Python; generally speaking, once the OS
> |> allocates memory to a process, the process keeps it until it exits.
> |> However, if you allocate and de-allocate big chunks of memory, you'll
> |> notice that Python's size stays static -- the memory gets re-used.
> |
> | Actually it's an issue for the heap manager, not the OS.  Under Windows
> | and Unix at least, it is possible for a process to return memory to
> | the OS.  In practice allocation is done in chunks of at least a page
> | and a page can't be freed until every object in it has gone.  Further-
> | more the memory will probably be needed again, and if it isn't it will
> | be paged out to disk.  So most heap managers don't bother to do this,
> | except for very large allocations.
> 
> Now that's interesting.  Could you provide a reference, say a man page
> for a system call, for how a user process would return memory, on a
> UNIX platform?  (Memory that doesn't just happen to be at the top of
> the data segment, of course.)

If the memory was allocated using an anonymous mmap instead of sbrk you 
can return it to the system using munmap.

Take a look at dlmalloc. http://gee.cs.oswego.edu/dl/html/malloc.html
This malloc implementation was written by Doug Lea and used in glibc and 
several other places.  It handles large allocations above a certain 
threshold (256k by default) using an anonymous mmap.

I recommend reading the dlmalloc code to everyone. It'll make you a better 
programmer.

    Oren




From snag at jool.po  Thu Apr 17 11:52:27 2003
From: snag at jool.po (kpop)
Date: Thu, 17 Apr 2003 16:52:27 +0100
Subject: More
Message-ID: <teBna.9109$pK2.11130@news.indigo.ie>

Hi
I usualy print strings the normal way
print x, but if the string is too big it will
go by the screen too fast and i wont be
able to see anything. I notice the python
internel help system hasa thing like the
unix prgram more, where you press enter
to go down. How can i use that in my own
programs? Is there a function or a module.
I am using windows so i dont have more.
Thanks

reposted because i got the date wrong last time









From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Mon Apr 14 14:14:45 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Mon, 14 Apr 2003 20:14:45 +0200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <3E9AF8BD.BC918E5E@engcorp.com>
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <b7di5e$dd6gj$2@ID-169208.news.dfncis.de> <3E9A4E32.F179FDDC@alcyone.com> <3e9af32e$0$49104$e4fe514c@news.xs4all.nl> <3E9AF8BD.BC918E5E@engcorp.com>
Message-ID: <3e9afaa2$0$49106$e4fe514c@news.xs4all.nl>

Peter Hansen wrote:
> Irmen de Jong wrote:
> 
>>[offtopic]
>>
>>Erik Max Francis wrote:
>>
>>
>>>>>>Course(lowEarthOrbit, lowMarsOrbit).transfers()
>>>
>>>[<ExtractionTransfer @ 0x82893ac (out of `low orbit around `Earth'')>,
>>><HohmannTransfer @ 0x812cef4 (Hohmann from `Earth' to `Mars')>,
>>><InsertionTransfer @ 0x82adb64 (into `low orbit around `Mars'')>]
>>>
>>
>>Is this actual code? What does this do? I'm interested.
> 
> 
> Given that Python is an "agile language", it seems very likely that
> it must be an implementation of Robert Zubrin's XP-style "Mars Direct" 
> program for a manned mission to Mars.  
> 
> <0.9 wink but I wish it were lower-ly y'rs> 
> 
> -Peter

I *was* serious...
If it is real code I suspect it somehow determines a path for a
free moving object in space to transfer from one orbit to another
orbit around different bodies in the solar system.
At least that is what a quick google for "Hohmann orbit transfer"
tells me. And I'm still interested ;-)

--Irmen




From claird at lairds.com  Thu Apr  3 09:25:58 2003
From: claird at lairds.com (Cameron Laird)
Date: Thu, 03 Apr 2003 14:25:58 -0000
Subject: Zope style (was: tuning our site (was: help! advocacy
	resources needed fast))
References: <mailman.1046819836.25704.python-list@python.org> <irdbl-1ds.ln1@news.lairds.org> <v8mjffpuph1oc2@corp.supernews.com> <mailman.1049323265.22242.python-list@python.org>
Message-ID: <v8oh3m1f28gl74@corp.supernews.com>

In article <mailman.1049323265.22242.python-list at python.org>,
Dylan Reinhardt  <python at dylanreinhardt.com> wrote:
>On Wed, 2003-04-02 at 12:54, Cameron Laird wrote: 
>> I find
>> it really interesting when one technology is unequivocally su-
>> perior to another.  I think we have that with DTML vs. Python
>> scripts.
>
>[ jumping in mid-stream ] 
>
>DTML and Python Scripts are different tools, each best-suited to a
>particular thing.  Determining which is better obscures a more important
>point: they're each good tools for a particular set of problems.
>
>DTML's problem space is quite a bit smaller, but that doesn't make it a
>bad tool... just one that should be applied more selectively than it
>typically is. :-)
>
>> There are a couple of pertinent facts.  First, much of the
>> passion "Web designers" are supposed to have for "templates"
>> comes from the supposition that they want stuff to look like 
>> HTML, so they can use their "HTML editors".  Is *anyone* doing
>> that with Zope, though?  
>
>It's not really about the visual editors.  DTML is largely incompatible
>with HTML editors... or so I hear. :-)
>
>The purpose is more basic than that: one is a presentation tool, one is
>a logic tool.
>
>Templating is appropriate when 80-90% of what you want to do is simply
>write out text.  With templating, there is no need to *say* that text
>should be written out... everything you type is returned unless you
>specify otherwise.
>
>When you're performing logic, the reverse is true... most of what you
>type *produces* your content and nothing is returned aside from that
>which is *explicitly* returned.
>
>Templating provides the least cluttered, most concise way of expressing
>*integrating* static and dynamic content.  Python represents the most
>concise way of expressing logical operations.  Each problem will consist
>of some part logic and some part presentation.  Most of my Zope projects
>use a blend of templating and scripting.
>
>> Finally, while your dtml-let is an effective style, 
>
>I almost never use dtml-let.  In my book, it's one of the first signs
>that you're trying to do something DTML isn't well-suited for.  Another
>is use of nested dtml-ifs more than a couple layers deep.
>
>> projects always, always, 
>> always seem to need more parametrization; managers say, "I want
>> it just like that, except it needs to {apply to a different
>> continent,show the current contents of the order,show the 
>> expanded display when the user has selected that, ...}."
>
>If that requirement is thought of as parameterization, DTML is a less
>worthy tool than Python scripts, no doubt.  But if that same set of
>requirements is thought of as "skinning" DTML is a great tool.
>
>DTML's usefulness depends, to some degree, on your use of Acquisition. 
>Your overall site design strategy might have a lot to do with why one
>seems more powerful than the other for any given task.
>
>Dylan
>
>

You make several propositions.  They partition themselves
into two categories:  those I think are important, and with
which I strongly agree; and several I've said myself in the
past.

Talk about Zope might seem better hosted on the mailing list;
I think there's value in keeping it here in clp, though, as
the real issues are ones faced by all Web and Python developers.

You are absolutely right that acquisition is at the heart of
this.  acquisition is very much like object orientation, in its
calculus of inheritance (although I think I have a theorem that
acquisition is harder to get right--more on that, later), and,
of course, notationally.  acquisition is also at least as dif-
ficult to design correctly.  Part of what I'm saying is that,
while I understand how acquisition boosts DTML's expressive
power quite satisfyingly, the domain of applicability of DTML
is so small that it's always--in my experience--more costly to
set up a rational acquisition structure than to work with Python
scripts (themselves occasionally boosted by acquisition).

Object-orientation has the same problems; OO just has more benefits,
and so it can be a net gain.

The symptom of acquisition's liabilities is complexity.  It appears
to me simply to take too much discipline to keep DTML-oriented
acquisition hierarchies from becoming incomprehensible.

Understand, I'm skeptical about the fine divisions of labor that
appear to prevail on some projects:  a Web application designer,
a Web architect, a Web security administrator, several Web visual
designers, and so on.  Yes, if there is that much complexity in the
workflow, acquisition and DTML match the work rather well.  I'm un-
impressed by what I know of such teams, though.

You write aptly of templating as presentation vs. algorithm.  Maybe
a more moderate expression of my views is this:  whereas I began
Zoping with the anticipation that some large portion--80%?  50%
20%--of site content would appear in DTMLs, I find the best balance
is shockingly low, perhaps no more than 5%.  This seems to be true
even for "brochureware" sites.  

Maybe the real issue is just the asymmetry I feel; I much prefer to
add """-coded presentation to a script, than <dtml-ish algorithms
to DTML.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From missive at frontiernet.net  Tue Apr  1 18:03:05 2003
From: missive at frontiernet.net (Lee Harr)
Date: Tue, 01 Apr 2003 23:03:05 GMT
Subject: Creating instance
References: <b6d191$j7s$1@news-reader12.wanadoo.fr>
Message-ID: <JUoia.305$8h6.197@news02.roc.ny.frontiernet.net>

In article <b6d191$j7s$1 at news-reader12.wanadoo.fr>, Salvatore wrote:
> Hello,
> 
> I would like to create an instance of class
> from the string name of the class
> 
> class Foo:
> 	pass
> 
> classname = 'Foo'
> instance = create_instance('Foo')
> 
> 


I think it depends on which namespace the class lives in,
and where you want to get at it from.

In the interactive interpreter, you might do:

class Foo:
    pass
instance = locals()['Foo']()



If the class lived in a module, say FooMod.py you could do:

# FooMod.py
import FooMod

class Foo:
    pass
instance = getattr(FooMod, 'Foo')()




From BPettersen at NAREX.com  Mon Apr 21 12:53:12 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Mon, 21 Apr 2003 10:53:12 -0600
Subject: Is this a super bug?
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE02E@admin56.narex.com>

> From: Michele Simionato [mailto:mis6 at pitt.edu] 
> 
> "Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message 
> news:<mailman.1050811779.16487.python-list at python.org>...
> 
> > <snip> this is the sequence that "should" happen:
> > 
> > 1  super(ff, f)[0]  # f == self
> > 2  superinst(type:<class ff>, obj:f)[0]
> > 3  si.  getitem  (0)
> > 4    si.  getattr  ('  getitem  ')
> > 5    .. return str.  dict  ['  getitem  '].  get  (f)
> > 6    boundmethod(str.  getitem  , f)
> > 7  bmeth(0)
> > 8   'a'
> 
> I don't understand what you are trying to say here ... 

Special method lookup goes through the class, and since the class is
Super, not obj.__class__, they can't be trapped through __getattr__.
(And for some reason they can't be trapped in a metaclass either...?)

[...]
> You may do that, but I don't like it. I like the fact that
> 
> super(cls,self)[index]
> 
> is *not* equivalent to
> 
> super(cls,self).__getitem__(index)

Most languages only define super.m(...) in object foo (class Foo,
superclass Foobase), and then to be equivalent to calling
Foobase::m(foo, ...).

Since Python has special attribute lookup rules for some attributes, the
normal extension would be to define super as an object with data
attributes from foo, and method attributes from Foobase. But that is not
even correct, it has to have the method attributes from Foo, except when
calling foo.m will call Foobase::m, but if Foobase::m calls m2, it will
start lookup in Foo.

One property this has is that given an empty subclass:

  class Foo : Foobase: pass

and a program context C with a hole, C[*] (i.e. C gives the value of the
program once you put a value in the hole), then:

   C[self] == C[super]

proving properties in this half-baked scheme is much harder.

> The present behaviour is the natural one, when you realize that super
> objects are attribute descriptors. In other words:
> 
> super(cls,self) is *not* self

You're objectively correct, however I'm missing the part where you're
saying this is good because...

> In you case 'self' is a string and has a __getitem__ method; 
> nevertheless, since super(cls,self) is a descriptor and not a string,
I 
> would not expect it to have a __getitem__ method, nor the other
methods of a string. 

... yet you want it to behave like if it had e.g. a find method... why
the inconsistency?

> What happens in reality (disclaimer: I don't know nothing about the C 
> implementation, I am guessing here) is that super(cls,self) has a 
> __getattribute__ method that call the methods of the string 
> object: the 
> fact that you can write
> 
> super(cls,self).__getitem__(index)
> 
> does *not* means that super(cls,self) has a __getitem__ method.

No but it means that when I ask for one it will get one..., first,
recall that 

  obj.meth.__get__(None, obj) is an unbound method

while

  m = obj.meth.__get__(ObjClass, obj) is a bound method

when saying obj.meth, you're doing the latter which means:

  1. if the instance contains the bound method, return it.
  2. if the class contains an unbound method with this name
     bind it to the object (create a bound method object
     and return it.

And then the special case semantics:

  obj[x]

calls ObjClass.__getitem__(self, x) directly, not going through the
steps above (thus my confustion that I can't trap it in the metaclass).
Thus, if I ask for __getitem__ by name, we follow the general rules, if
the __getitem__ semantics is invoked indirectly through subscripting, we
follow the special rules.

> I agree this can be confusing at first. The solution would be to
> document ``super`` and the attribute descriptors. 

No, the first thing is to define what we would like to use this concept
for, then we can decide which properties it must have, and then we can
come up with an implementation that is useful, not merely confusing.

> Maybe one of these days I will collect my notes and fill this gap,
> but this would require at least 20-30 pages, so it would take some
> time ...

That would be interesting to read... If I get time I'll try to implement
my semantics... The metaclass issue should make for a lot of typing...

-- bjorn




From ckea25d02 at sneakemail.com  Mon Apr 14 22:21:27 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 15 Apr 2003 14:21:27 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2n0itfrft.fsf@mycroft.actrix.gen.nz>
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz>
Message-ID: <b7fpmq$ht9r$1@ID-169208.news.dfncis.de>

Paul Foley wrote:
> Don't forget the ~20000 other variables that you'd have to pass around
> as explicit arguments to *every single function*.

Paul, you're suffering from an extremely bad case of
tunnel thinking.

You wouldn't pass around 20,000 variables. You would
pass around a very few objects (in most cases probably
just one, i.e. a buffer) from which everything else you
need can be gotten to if and when you need it.

> Exactly as with the non-lexical variables that Python sports right
> now.  This has nothing whatsoever to do with dynamic scoping.

I'm not sure what you think "lexical" means, because
according to most people's definition, Python's scoping
is very much lexical, for both locals and globals.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From aahz at pythoncraft.com  Wed Apr 23 15:27:31 2003
From: aahz at pythoncraft.com (Aahz)
Date: 23 Apr 2003 15:27:31 -0400
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <tyfhe8xqhmy.fsf@pcepsft001.cern.ch> <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com> <b86lc3$tnd$1@plover.com>
Message-ID: <b86pf3$cqs$1@panix3.panix.com>

In article <b86lc3$tnd$1 at plover.com>,
Mark Jason Dominus <mjd at plover.com> wrote:
>In article <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4 at 4ax.com>,
>Courageous  <jkraska at san.rr.com> wrote:
>>Courageous as usual deleted someone else's attribution:
>>>
>>>But he regrets the inclusion of a crippled lambda, rather than the
>>>inclusion of support for anonymous functions (IIRC).
>>
>>Anyway, I guess what I'm saying is that I suspect that Guido
>>didn't have any good alternatives. 
>
>I don't believe that's correct.  At an early Open Source Conference,
>perhaps around 1998 or so, I was present at a conversation between
>Guido and Tom Christiansen.  Tom asked Guido why Python didn't have
>closures, and Guido's reply (and I think this is an exact quote) was
>"Closures aren't important."
>
>I remember it because I was so astonished by the obtuseness of the answer.

Two points:

* I'd be mildly surprised if that were an exact quote.  Python 2.1 was
released 4/2001, with support for nested scopes (which allows building
closures of sorts).  That's roughly two years for Guido to change his
mind, which is a short timeframe for him.  ;-)

* Many of the things for which programmers in other languages use
closures are better done in Python as classes.  He's generally
antipathetic to borrowing idioms from other languages that make it easier
to write unpythonic code -- There's Only One Way.  Like other leaders in
the Open Source community, Guido is occasionally prone to making
obnoxious comments Just Because.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From mis6 at pitt.edu  Wed Apr 30 09:28:28 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 30 Apr 2003 06:28:28 -0700
Subject: Question about accessing class-attributes.
References: <mailman.1051212998.14005.python-list@python.org> <2259b0e2.0304250413.4be8ee45@posting.google.com> <7h3bryulp3e.fsf@pc150.maths.bris.ac.uk> <of2uen32.fsf@python.net> <7h37k9ilm5v.fsf@pc150.maths.bris.ac.uk> <2259b0e2.0304251030.30fd1f94@posting.google.com> <7h3y91tiszi.fsf@pc150.maths.bris.ac.uk>
Message-ID: <2259b0e2.0304300528.34b6f522@posting.google.com>

Michael Hudson <mwh at python.net> wrote in message news:<7h3y91tiszi.fsf at pc150.maths.bris.ac.uk>...
> Then you get this kind of weirdness:
> 
> />> class C(type):
> |..     def __eq__(self, other):
> |..         return 1
> \__ 
> />> class A:
> |..     __metaclass__ = C
> \__ 
> />> class B:
> |..     __metaclass__ = C
> \__ 
> ->> d = {}
> ->> d[A] = 1
> ->> d[B] 
> Traceback (most recent call last):
>   File "<input>", line 1, in ?
> KeyError: <class '__main__.B'>
> ->> A == B
> 1
> 
> Then I suspect whether a instance of C gets A out of the dictionary
> depends on accidents of memory allocation and hash table
> implementation.  That's not a good thing.
> 
> CHeers,
> M.


Really, I don't see where is the problem. As far as I understand (correct me
if I am wrong) dictionaries are indexed trough the identifier of their keys
i.e. hash(A) == id(A). In this example, since A is not B, it is clear to me 
that d[B] is not defined and cannot work.

This issue has nothing to do with metaclasses:

class O(object): 
    def __eq__(self, other):
        return 1
 
a=O(); b=O()

assert id(a)==hash(a)
assert id(b)==hash(b)

assert a==b
assert a is not b

d={a:1}
d[a] #=>1
d[b] # error since d[b] is not defined, b is not a

It seems that everybody (including myself) sees more problems in 
metaclasses than really are.


                                                 Michele



From aleax at aleax.it  Sat Apr 12 15:52:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 19:52:39 GMT
Subject: sorting strings written to files
References: <2745895.1049870828@dbforums.com>
Message-ID: <b8_la.5508$LB6.155037@news1.tin.it>

<posted & mailed>

onefatman wrote:

> Can someone help me?
> 
> I have a small phone directory .dat file that has the entries
> written to it as strings with three fields in each, separated by \t
> and ended by \n.
> 
> e.g.
> phbook.write("Joe\tNobody\t62907243\n")
> phbook.write("Daniel\tStevens\t62962328\n")
> phbook.write("Ellen\tWay\t62910758\n")
> phbook.write("Jason\tGreig\t62314587\n")
> 
> it has a function that allows you to write new entries, but i need to

How does a .dat file "have a function"?  That's not clear to me...

> know how to sort them by first name, and then by last name when the
> first names are the same.

With Python, you can read your file into memory as a list of lists of
fields, sort the list, then write it out again.  The following function
for example would do it, if called with the filename of your .dat
file (e.g. the string "C:/ofm/wot.dat" if that's where's your file
lives -- remember you can use plain slashes, handier than backlashes \n,
even if you run on Windows machines):

def sortofm(filename):
    filein = open(filename)
    listoflists = [ line[:-1].split('\t') for line in filein ]
    filein.close()

    listoflists.sort()

    fileou = open(filename, 'w')
    fileou.writelines(['\t'.join(line) for line in listoflists])
    fileou.close()


In this case, the natural ("lexicographical order") sort of lists
of lists of strings does exactly what you want, because the first
name is the first field, and the last name the second field.  In
more general cases, you'd want to use the "decorate-sort-undecorate"
idiom abundantly explained in a zillion previous posts to this NG
(use Goggle Advanced Group Search for those), the "Python Cookbook",
"Python in a Nutshell", and no doubt many, many other places too;-).


> Also, does anyone know how to delete details and change details? it
> would be a big help. thanks.

Not sure what the "details" are supposed to be.  In general, once
you've read your .dat file into memory as a list of lists (as in
the first three statements of the above function sortofm) and
before you write it back (as in the last three statements thereof),
you can perform any alteration (that leaves listoflists' items in
the same general shape) using Python indexing, list methods, etc.


Alex




From simon at brunningonline.net  Tue Apr 15 04:23:57 2003
From: simon at brunningonline.net (Simon Brunning)
Date: 15 Apr 2003 01:23:57 -0700
Subject: Cryptography (was Re: How to protect python code ?)
References: <3E8BE455.4000904@novagrid.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com> <OIqdnV1p3aljSw-jXTWcpg@comcast.com> <b7bvft$os4$1@panix1.panix.com> <b7ecno$n7m@newton.cc.rl.ac.uk> <R7qdnYcmzPGtawejXTWcrg@speakeasy.net>
Message-ID: <314b29e9.0304150023.10bce745@posting.google.com>

"A.M. Kuchling" <amk at amk.ca> wrote in message 
> Classified establishments certainly have more money to spend, so I expect
> they have big computers and custom-designed chips for trying keys, but I
> doubt they have a super factoring algorithm or powerful attacks against 
> block ciphers.

Possibly true of *most* classified establishments, but I gather that
the PSU has developed a technique enabling them to



From andrew-pythonlist at puzzling.org  Thu Apr 17 05:28:15 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Thu, 17 Apr 2003 19:28:15 +1000
Subject: A better unittest
In-Reply-To: <20030416145011.E4266@tibia.amotlpaa.bogus>
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net> <1EadnQb6ZY_RDwCjXTWcog@comcast.com> <20030416145011.E4266@tibia.amotlpaa.bogus>
Message-ID: <20030417092815.GA22467@frobozz.local>

On Wed, Apr 16, 2003 at 02:50:11PM -0600, Steven Taschuk wrote:
> Quoth Terry Reedy:
>   [...]
> > I like having the actual values too.  It should make debugging an
> > unexpected failure easier.  But what if each value is more that a few
> > bytes?  Example (from my future plans): two lists of all permutations
> > of range(8), one produced by an obvious recursion and the other by a
> > much less obvious iteration.
> 
> Now and then I replace
>     self.assertEquals(longlist1, longlist2)
> with
>     map(self.assertEquals, longlist1, longlist2)
> 
> It's not quite ideal; in some cases it would be nice to see the
> index of the mismatch too, for example.

map(self.assertEquals, enumerate(longlist1), enumerate(longlist2))

[Requires Python 2.3, of course]

-Andrew.





From tismer at tismer.com  Mon Apr 21 21:10:50 2003
From: tismer at tismer.com (Christian Tismer)
Date: Tue, 22 Apr 2003 03:10:50 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <b81tg9$hhs$1@main.gmane.org>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <mailman.1050953314.21065.python-list@python.org> <b81od0$gcc$1@news.hccnet.nl> <b81tg9$hhs$1@main.gmane.org>
Message-ID: <3EA4969A.80203@tismer.com>

Mike Rovner wrote:
> The Python problem with continuation is IMHO they downgrade the language
> level instead of upgrading it. They are so 'basic' that, like macros, tend
> to make Python too fluid and thus unmaintainable and bad for understanding.

Very good point!

- chris

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From claird at lairds.com  Thu Apr 24 15:02:02 2003
From: claird at lairds.com (Cameron Laird)
Date: Thu, 24 Apr 2003 19:02:02 -0000
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <F3Epa.51023$LB6.1152908@news1.tin.it> <e9570f37.0304240556.50465c66@posting.google.com> <k_2dnckpjsVftTWjXTWcpQ@speakeasy.net>
Message-ID: <vagd5alkn2903c@corp.supernews.com>

In article <k_2dnckpjsVftTWjXTWcpQ at speakeasy.net>,
A.M. Kuchling <amk at amk.ca> wrote:
>On 24 Apr 2003 06:56:51 -0700, 
>	Graham Fawcett <graham__fawcett at hotmail.com> wrote:
>> Of course we have no infrastructure for this (at least for the ones
>> that cost money). But I'm sure we could create one if the need arose.
>
>Two sites were created for this purpose, SourceExchange and Cosource.  Both
Also Collab.Net and a few others, but, yes,
there were FAR less successful than expected.

One of the characteristics of software is that
we deal with enormous temporal ranges (from
microseconds to minutes, in a typical appli-
cation; as engineering goes, that's a big 
factor).  It's impressive that the effici-
ency of our work structures can be almost 
equally far removed from the smooth output of
some of its work products.
>are now dead and gone, which should tell you something.  Such a tip-jar
>system produces trivial amounts of money, $20 or $100. If you had $50K to
>spend, you could hire a developer to work on something for 40 hours a week,
>but that amount of money is not found in a tip jar.
			.
		[stuff I don't follow,
		but perhaps is true]
			.
			.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From kkto at csis.hku.hk  Tue Apr  1 08:02:39 2003
From: kkto at csis.hku.hk (Isaac To)
Date: 01 Apr 2003 21:02:39 +0800
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <slrnb8hrv9.mk0.chasm@galileo.rift> <vHcia.39470$Jg1.934856@news1.tin.it>
Message-ID: <7iu1di74ls.fsf@enark.csis.hku.hk>

>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:

    Alex> Julian Tibble wrote: ...
    >> his C++ program, since I'm not sure I can write a better hash table
    >> than the g++ STL developers.

    Alex> I'm SURE you can write a hash-table that's a better hash-table
    Alex> than C++'s std::map, because the latter CANNOT be a hash table --
    Alex> it must rely on ordering of keys, and is normally a red-black
    Alex> tree;-).

But g++ does provide a hash table, although not many people know it.

#include <iostream>
#include <ext/hash_map>
using namespace std;
using namespace __gnu_cxx;

int main() {
  hash_map<long, int> hm;
  for (int i = 0; i < 10000; ++i)
    hm[i*i] = i;
  for (int i = 0; i < 10000; ++i)
    if (hm[i])
      cout << i << ": " << hm[i] << endl;
}

Regards,
Isaac.



From hjwidmaier at web.de  Sun Apr 13 03:40:29 2003
From: hjwidmaier at web.de (Hans-Joachim Widmaier)
Date: Sun, 13 Apr 2003 09:40:29 +0200
Subject: Terminal Emulation
Message-ID: <pan.2003.04.13.07.40.04.733053@web.de>

All of my embedded software offers a means of debugging via a serial
interface. You just hook a terminal onto it and start introspection. These
days, of course, the "terminal" is usually a computer running a terminal
emulator, and the embedded software allows you to do more than just view
memory contents, e. g., download a new software. Of course, I handle this
download with a python program. Now you have 2 programs: the terminal
emulator and the download utility. Not too much hassle under unix, as the
downloader ignores that the serial port is already used by the terminal
emulator, it just does its job and leaves the port in the same state it
had found it, the emulator doesn't even notice.

This doesn't work under Windows, though. So I thought about creating a
terminal emulator with some added functionality. Some thinking revealed
that I didn't know a widget that was handy for the job. Some more
thinking gave me an idea for a unix solution: xterm in slave mode.
Obviously, this only works where I don't really need it. A google search
showed that some others also had a use for such a terminal emulation
widget, but none seems to exist.

Not knowing wxWindows, pyQt and what-have-you, the Tkinter Text widget is
about the only thing left. I'd have to grab all keyboard events away from
it (yuck) and do the escape sequence interpretation by myself (not too
bad, as I only need a subset of the whole vt100/vt320 stuff). The Text
widget just seems way too "heavy" for this.
Handling serial input/output as well as GUI events gives me problems
enough already. ;-)

Has anyone found a nicer solution in the meantime? No problems if it meant
using wxPython or pyGTK. (No pyQt, though, as it's not free under Windows.)

I'd be grateful for any suggestions.

Hans-Joachim



From martin at v.loewis.de  Sun Apr  6 13:46:14 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 06 Apr 2003 19:46:14 +0200
Subject: executable won't run on another machine
References: <v90mh6ppf9bc09@corp.supernews.com>
Message-ID: <m3u1dbmsd5.fsf@mira.informatik.hu-berlin.de>

"Lance" <lbrannma at cablespeed.com> writes:

> I created an executable using McMillan's Installer. It runs great on the
> development machine. I copied it to another machine and it doesn't run. I
> type the executable name from a command prompt, hit return, and get another
> command prompt.
[...]
> Any suggestions will be appreciated.

Most likely you are lacking some modules which installer failed to
find (e.g. because they are triggered by computed imports, as PyXML
does them, or the codecs). It works on your development machine
because the modules are still there.

Normally, you should get an exception. Perhaps you are catching all
exceptions somewhere? I recommend you remove this code, to see the
traceback.

I also recommend to use strace to find out what other modules it tries
to load.

Regards,
Martin




From peter at engcorp.com  Wed Apr 16 15:14:44 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 15:14:44 -0400
Subject: How many of you are Extreme Programmers?
References: <mailman.1050513152.32206.python-list@python.org>
Message-ID: <3E9DABA4.B2B53153@engcorp.com>

Mike Brenner wrote:
> 
> Most of the previous answers stated that they
> do the continual and thorough testing portion of XP.
> 
> Continual and thorough testing was not new to
> XP, and XP did not invent that. We, too, do a lot of that.

It's been pointed out, and repeatedly, and usually by the XP
practitioners themselves (and certainly by those who "formulated"
the approach) that there's nothing new in XP.  XP merely draws
on excellent practices that have been used, sometimes in 
isolation, on other projects and in other methodologies.  If
anything, it's merely the mix which is new.

Note, however, that TDD itself appears to be largely a novel
idea.  I've seen at least one person claim to have done it himself
in the past many years ago, but I've also seen psychics claim 
to have predicted major world events only _weeks_ after they 
actually occurred...

> To be Xtreme, you also need to overcome the social barriers,
> quite a difficult feat. Getting people to see the advantage,
> for example, of two people working together at all times
> is challenging (to the territorial ego of almost all
> developers).

Although the social aspects are not easy, I reject the idea
that almost all developers have territorial egos.  Territoriality
was by far the *least* significant social issue we encountered
in becoming an XP team.  In fact, I'd go so far as to say that 
in most cases (in my experience) people welcome not having to
feel sole responsibility for a given piece of code.  Not always,
but very often.

-Peter



From aleax at aleax.it  Wed Apr 30 12:30:31 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 16:30:31 GMT
Subject: fcntl in windows
References: <k2rvav06nhjgtr578fa5ostftfl69rethp@4ax.com> <mailman.1051717847.12855.python-list@python.org>
Message-ID: <HSSra.42028$K35.1214886@news2.tin.it>

Jp Calderone wrote:

> On Wed, Apr 30, 2003 at 03:26:29PM +0000, Afanasiy wrote:
>> Is there an fcntl in Windows? A stub, a wrapper, etc?
> 
>   http://python.org/doc/lib/module-fcntl.html

The page you indicate says "Availability: Unix", and indeed
trying:

import fcntl

on Windows gives an ImportError: No module named fcntl


As far as I know, nobody's written a fcntl-emulator for
Windows -- the functionality is clearly available (a bit
in msvcrt, all of it in win32all, or maybe also through
ctypes) but the interfaces are, alas, quite different.


Alex




From westernsam at hotmail.com  Fri Apr  4 04:45:17 2003
From: westernsam at hotmail.com (sam)
Date: 4 Apr 2003 01:45:17 -0800
Subject: My Object pool isnt working
Message-ID: <292c8da4.0304040145.480bcacc@posting.google.com>

Hi,

I need to pool some database connections so i thought I'd write a
object pool. I thought I'd use a design I've used in Java before -
having a locked and unlocked hashtable and keeping references of the
connections in two states (i.e. allocated and unallocated). The
problem is that the total number of connections always seems to rise.
Also the total number of connections is less than the sum of the
lengths of the hashtables (dictionaries) which shouldn't happen. I
believe the problems are connected. As I said I've used this design
before in Java and its been fine. The only difference I can think of
(barring shoddy implementation, which I can't rule out) is that
Hashtables in Java are thread safe, but I'm not sure if the built in
dict type in python is.

I've decided to use jonpy's dbpooling module
(http://sourceforge.net/projects/jonpy) instead of writing my own now,
but I'd still like to know whats going on. I have tested on python
1.52 and python 2.2 (both on win2k) and get the same result. All
thoughts appriciated

Regards

Sam Owen

<code>
from Threading import *
import time
import random
import sys

class Object:
    def sayHello(self, caller):
        print "hello", self, caller 
        time.sleep(int(random.random() * 5))

class ObjectPool:
    LOCKED= {}
    UNLOCKED= {}
    sema_connect = Semaphore(1)
    sema_return = Semaphore(1)
    expiretime = 60 #30 mins
    totalConns=0
    numAllocated = 0
    numReturned = 0

    def getConnection(self):
        try:
            self.sema_connect.acquire()
            if len(self.UNLOCKED):
                for ts in self.UNLOCKED.keys() :
                    if ts + self.expiretime < time.time():
                        print "deleting ..."
                        self.totalConns = self.totalConns-1
                        del self.UNLOCKED[ts]
                    else:
                        print "allocating ..."
                        self.numAllocated = self.numAllocated+1
                        o = self.UNLOCKED[ts]
                        self.LOCKED[o]=1
                        del self.UNLOCKED[ts]
                        return o
            print "creating ..."
            self.numAllocated = self.numAllocated+1
            self.totalConns = self.totalConns+1
            o = Object()
            self.LOCKED[o] = 1
            return o
        finally:
            self.sema_connect.release()    

    def returnConnection(self, con):
        try:
            self.sema_return.acquire()
            if self.LOCKED.has_key(con):
                print "returning"
                self.numReturned = self.numReturned +1
                self.UNLOCKED[time.time()]=con
                del self.LOCKED[con]
                return
            else:
                print con, "was not found in LOCKED ?????"
        finally:
            self.sema_return.release()

class ObjectUser(Thread):
    def __init__(self, pool):
        self.pool = pool
        Thread.__init__(self)
    def run(self):
        while 1:
            o = pool.getConnection()
            o.sayHello(self)
            pool.returnConnection(o)
            time.sleep(int(random.random() * 5))
        
if __name__ == '__main__':
    pool = ObjectPool()
    for i in range(2):
        thread = ObjectUser(pool)
        thread.start()
    while 1:
        print "LOCKED --> ", len(pool.LOCKED), "\nUNLOCKED -->",
len(pool.UNLOCKED), "\ntotal conns", pool.totalConns
        print "allocates", pool.numAllocated, "\nreturns",
pool.numReturned
        time.sleep(5)
</code>

<sample output>
creating ...
LOCKED -->  1
UNLOCKED --> 0
total conns 1
allocates 1
returns 0
creating ...
hello <__main__.Object instance at 8797a8> <ObjectUser(Thread-2,
started)>
hello <__main__.Object instance at 879790> <ObjectUser(Thread-1,
started)>
returning
allocating ...
hello <__main__.Object instance at 879790> <ObjectUser(Thread-1,
started)>
returning
LOCKED -->  1
UNLOCKED --> 1
total conns 2
allocates 3
returns 2
allocating ...
hello <__main__.Object instance at 8797a8> <ObjectUser(Thread-2,
started)>
returning
returning
allocating ...
hello <__main__.Object instance at 8797a8> <ObjectUser(Thread-2,
started)>
returning
allocating ...
hello <__main__.Object instance at 8797a8> <ObjectUser(Thread-2,
started)>
creating ...
hello <__main__.Object instance at 85efe8> <ObjectUser(Thread-1,
started)>
LOCKED -->  2
UNLOCKED --> 0
total conns 3
allocates 7
returns 5
returning
returning
allocating ...
hello <__main__.Object instance at 8797a8> <ObjectUser(Thread-1,
started)>
returning
</sample output>



From max at alcyone.com  Fri Apr 25 06:51:32 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 25 Apr 2003 03:51:32 -0700
Subject: OT: losing memories (was Re: How many of you are Extreme 
 Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <oafgavg8e0pf512712g4aenldqktvmb0dr@4ax.com> <I78qa.30798$y3.2515733@news010.worldonline.dk>
Message-ID: <3EA91334.A7A1D8F2@alcyone.com>

Max M wrote:

> If only 10% of the brain is used, why don't we hear:
> 
> "Well he was shot in the head, but the bullet only went through the
> 90%
> of the brain that he doesn't use. So no sweat mate."

In all seriousness, there are cases of severe brain injury (volume wise)
which result in surprisingly little actual impairment of abilities. 
There's obviously the Phineas Gage case (who, after having been involved
in a railroad construction accident had a very large piece of metal fire
through his brain, destroying most of the left frontal part of his
brain) retained most of his mental acuities but had some emotional
behaviorial changes).  More dramatically, there was a recent case (I saw
it on either Nova or Scientific American Frontiers; the latter sounds
more likely) where a young man (whose name I don't know, I could dig it
up) had lost literally more half the volume of his brain (as in, the
shape of his head was quite clearly abnormal due to the lack of skull
and brain tissue) due to some sort of severe trauma and showed very
little sign of significant impairment.

The human brain is really an amazing organ.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The actor is not quite a human being -- but then, who is?
\__/ George Sanders
    Bosskey.net: Unreal Tournament 2003 / http://www.bosskey.net/ut2k3/
 A personal guide to Unreal Tournament 2003.



From exarkun at intarweb.us  Wed Apr 30 12:25:28 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 30 Apr 2003 12:25:28 -0400
Subject: question on the compile built in function.  Filename ignored?
In-Reply-To: <7rSra.706428$S_4.744795@rwcrnsc53>
References: <7rSra.706428$S_4.744795@rwcrnsc53>
Message-ID: <20030430162528.GD26204@meson.dyndns.org>

On Wed, Apr 30, 2003 at 04:01:07PM +0000, Alex wrote:
> I am fairly to new to Python, so please bear with me.
> 
> I am trying to use the built in function compile.  The filename argument
> seems to be ignored no matter what I try.  According to the docs, it should
> be used in the output error message.  Here is a simple example:
> 
> Python 2.2.2 (#1, Apr 16 2003, 22:00:16)
> [GCC 3.2.1 20021207 (Gentoo Linux 3.2.1-20021207)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> >>> compile(':\n', 'abc', 'exec')
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "<string>", line 1
>     :
>     ^
> SyntaxError: invalid syntax
> 
> 
> From the docs I would expect the line 
> 
>   File "<string>", line 1
> 
> to read
> 
>   File "abc", line 1
> 
> 
> What am I missing?

  SyntaxErrors are caught earlier than most other errors.  If you compile
something that is syntactically correct but raises some other exception,
you'll see the filename you specified in the traceback you get when you exec
the result of the compile() call.

  Jp

-- 
Where a calculator on the ENIAC is equipped with 18,000 vacuum tubes and
weighs 30 tons, computers in the future may have only 1,000 vacuum tubes and
weigh only 1.5 tons.    -- Popular Mechanics, March 1949
-- 
 up 41 days, 13:04, 5 users, load average: 0.13, 0.16, 0.09




From aleax at aleax.it  Tue Apr 29 12:34:16 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 16:34:16 GMT
Subject: Function to convert fetchone() result from a list to a dictio nary
References: <mailman.1051626442.19787.python-list@python.org>
Message-ID: <cQxra.20637$3M4.586766@news1.tin.it>

sismex01 at hebmex.com wrote:

>> From: Alex Martelli [mailto:aleax at aleax.it]
>> Sent: Tuesday, April 29, 2003 4:09 AM
>> 
>> So here's a step-by-step version (using SPACES, not TABS:-)...:
>> 
>> def cfd(cursor):
>>     values = cursor,fetchone()
>>     if values is None:
>>         return None
>>     keys = [d[0] for d in cursor.description]
>>     result_dict = dict(zip(keys, values))
>>     return result_dict
>>
> 
> I don't think there's need to explicitly return None,

No, there isn't.  However, stylistically speaking, I
*HATE* functions that along some execution paths have
explicit "return <expr>" statements and along others
have bare "return" or "just fall off the end" -- in my
code-maintenance experience, such functions are just
too likely to have bugs due to the original developer
having _ignored_ the possibility of alternate execution
paths -- so every time I find myself examining one,
I need to look at it long and hard to make SURE the
coder's intentions are indeed exactly expressed.

MUCH better to have an *explicit* "return None" along
the execution paths that really want to do that, MOST
particularly when that aspects of a function's specs
is so anti-intuitive as here (why None when it seems
SO much more natural to return an empty dict!? -- must
be a very explicit specification mandating it, I guess).

> nor to assign result_dict only to immediately return

Absolutely -- that, as I explained, was a didactical
device, to help a newbie understand better what was
going on.  I do prefer "return dict(..." etc myself.

> it, but then, that's just me ;-)
> 
> 
> def cfd(cursor):
>     values = cursor.fetchone()
>     if values:

Are you sure that the specs of cursor.fetchone() make
it IMPOSSIBLE that it returns e.g. an empty list or
tuple?  If that happens, this function is buggy in the
sense that it doesn't implement the same semantics as
the original one -- the original one would return an
empty dict when cursor.fetchone() returned an empty
sequence, this one would return None instead then.

I consider it horrid style to check "if values:" when
what one REALLY means is "if values is not None:".

>         keys = [d[0] for d in cursor.description]
>         return dict(zip(keys,values))

Sure, this part is OK (and suggested it myself as my
own preferred alternative in a followup).


>> Of course, you can bunch this up more, or less, according to
>> taste -- whatever you find clearest and most readable.  E.g.,
>> the last three lines could be instead:
>> 
>>     return dict(zip([d[0] for d in cursor.description],values))
>> 
>> if you're into one-liners!-)
> 
> Ach! But Perly stuff is so unmaintainable... ;-)

Yes, excessive terseness can damage maintenability.  But so
can the lack of explicitness, particularly about None!-)


Alex




From harri.pasanen at trema.com  Thu Apr 10 12:55:25 2003
From: harri.pasanen at trema.com (Harri Pasanen)
Date: Thu, 10 Apr 2003 18:55:25 +0200
Subject: python 2.3a2 Win32 expected failures in regression tests?
In-Reply-To: <16021.39020.38164.886501@montanaro.dyndns.org>
References: <mailman.1049910795.4707.python-list@python.org> <e2d925a8.0304100803.46e63240@posting.google.com> <16021.39020.38164.886501@montanaro.dyndns.org>
Message-ID: <200304101855.25404.harri.pasanen@trema.com>

On Thursday 10 April 2003 18:14, Skip Montanaro wrote:
<snip>
>
> Misc/NEWS is Python's ChangeLog, available online here:
> 
http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/*checkout*/python/python/dist/src/Misc/NEWS?rev=HEAD&content-type=text/plain
>

Thanks.

Are there CVS snapshot tarballs available?  Trying to log in, I'm 
getting

cvs [login aborted]: reading from server: Connection reset by peer

-Harri




From michaels at rd.bbc.co.uk  Wed Apr  2 07:28:31 2003
From: michaels at rd.bbc.co.uk (Michael Sparks)
Date: Wed, 02 Apr 2003 13:28:31 +0100
Subject: intro to OOP using Python, tutorial?:newbie
References: <3E8A5AE4.1080902@netscape.net>
Message-ID: <b6el8v$7d7$2@nntp0.reith.bbc.co.uk>

vector wrote:

> is there an introduction to OOp using python tutorial doc,file,pdf etc
> anywhere on the net. 

How to Think Like a Computer Scientist: Learning with Python 

... is a gentle introduction to Python, OO and various other things.
It's available as HTML, PDF, or LaTeX from:
   http://greenteapress.com/thinkpython.html

Or even dead tree form from bookshops.


Michael.
-- 
Michael.Sparks at rd.bbc.co.uk    
British Broadcasting Corporation, Research and Development
Kingswood Warren, Surrey KT20 6NP

This message (and any attachments) may contain personal views
which are not the views of the BBC unless specifically stated.





From martin.vonloewis at hpi.uni-potsdam.de  Tue Apr 22 15:16:20 2003
From: martin.vonloewis at hpi.uni-potsdam.de (Martin v. Löwis)
Date: Tue, 22 Apr 2003 21:16:20 +0200
Subject: missing X11.h
References: <mailman.1051029574.24925.python-list@python.org>
Message-ID: <b844e3$92i$1@zeppelin.rz.uni-potsdam.de>

> while compiling the source for python 2.2.2 the compilation was unable
> to locate X11/Xlib.h when compiling the file _tkinter.c.  is there a
> package that i need to install in order to get the headers so that i
> can compile X window programs?

Yes.

To find out what the name of the package is, you should tell us what
operating system you use first.

Regards,
Martin





From news1NOSPAM at awakened-lands.net  Mon Apr 21 23:18:13 2003
From: news1NOSPAM at awakened-lands.net (Michael Steuer)
Date: Tue, 22 Apr 2003 13:18:13 +1000
Subject: Creating a stand alone Python Service (win32)
References: <PA$jhlyBDHA.3088@lime.oak-park.us> <3EA2AB83.B4A95A14@engcorp.com>
Message-ID: <AEYXd3HCDHA.2220@lime.oak-park.us>

Hi.

:: Are you certain that it didn't work anyway?  That message comes when
:: the python.exe (or your service?) is run in a context where it is
:: unable to find the file site.py.  It's just a warning though, not
:: a serious failure, unless something later depends on site.py
:: and its normal functioning.
::
:: (To see this in action, set PYTHONHOME somewhere incorrect, then
:: run python using the full path.)
::
:: If you want to make it go away, you might be able to include a
:: dummy site.py in your package, or perhaps somehow invoke the -S
:: option on startup, although to be honest I'm not sure how all
:: this interacts with py2exe so at this point I'm rambling....

You are right, it is working anyway.

I was trying to start the service with a batch file, but it seems that the
service created with py2exe can't be started with "<servicename> -start"
(py2exe allows "-install" and "-remove"), but even though it doesn't really
know the "start" option, it still says "... Starting service - this may take
several seconds - please wait ... Could not start the service - error 997".

That was what made me think that the service is not working and given the
message I got when registering PythonService.exe, I thought it was caused by
this.

After seeing your message, I tried again and used the Service Management
Console to start the service and everything worked just fine. I should have
use "net start <servicename>" in my batch file form the start. That would
have saved me a lot of trouble.

Thanks for pointing out that the message I got was simply a warning and I
appologize for not checking this properly before I asked this question in
the first place !

--?
Cheers,

Mike.

PS: remove NOSPAM from my reply address to get my actual email address.





From mcfletch at rogers.com  Sun Apr  6 11:41:46 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 06 Apr 2003 11:41:46 -0400
Subject: properties and get/set methods
In-Reply-To: <v908s7echa191b@news.supernews.com>
References: <mailman.1049581810.25706.python-list@python.org> <ad052e5c.0304052016.4c67355c@posting.google.com> <mailman.1049604615.22077.python-list@python.org> <v908s7echa191b@news.supernews.com>
Message-ID: <3E904ABA.9050501@rogers.com>

John Roth wrote:

>"Mike C. Fletcher" <mcfletch at rogers.com> wrote in message
>news:mailman.1049604615.22077.python-list at python.org...
>  
>
...

>>In these cases, you're using properties to simplify the modelling of
>>domain objects, or to provide some *generic* functionality over and
>>above a regular value get/set.  You don't want to have to specify that
>>functionality for every property, you just want to say "this is one of
>>_those_ properties".  The code bloat (not to mention mindless
>>duplication of code) when you get up to 15 or 20 properties for an
>>object would be ridiculous.
>>    
>>
>
>Well, I suppose since this is Sunday, it's appropriate to bring
>religion into it.
>
Well, unless you're Jewish, Muslim, Atheist, Hindu, Buddhist, or any of 
the other faiths where Sunday is largely irrelevant ;) .

Seriously though, yes, this is a "religious" argument.  The morality 
going along these lines:

    * Needless duplication of code creates bugs and increases
      maintenance costs
    * Explicit is better than implicit
    * Practicality beats purity, but value purity when you find it
    * Beliefs change

Properties as they were originally designed have a "canon" of use which 
says they're for wrapping getter/setter methods up into objects that 
mediate access to an attribute via x.y notation.  That's fine.  It's 
useful sometimes, though not spectacularly often in my experience.

When people start using properties heavily, the definition migrates 
closer to "describing a property/attribute slot of an object" with the 
access mediation of getter/setters being a side-effect of that 
description.  So after a while, the people using them start seeing them 
in a different light than those for whom they are still just for 
wrapping getter/setter methods.

The generic characteristics and systemic functions of these new 
properties get captured in the new systems (e.g. as sub-classes of 
property objects) in such a way as to dramatically simplify some types 
of programming.  For some, the new properties *become* the interface to 
the canon as they are the only property type used in day-to-day 
programming.  In response, the canon of "properties" either expands to 
embrace the new usage, or fractures/is redefined.

Just an example...

class ElevationGrid( nodetypes.Geometry ):
    PROTO = 'ElevationGrid'
    zDimension = field.newField( 'zDimension', 'SFInt32', 0, 0)
    xSpacing = field.newField( 'xSpacing', 'SFFloat', 0, 0.0)
    creaseAngle = field.newField( 'creaseAngle', 'SFFloat', 0, 0.0)
    normal = field.newField( 'normal', 'SFNode', 1, node.NULL)
    solid = field.newField( 'solid', 'SFBool', 0, 1)
    ccw = field.newField( 'ccw', 'SFBool', 0, 1)
    texCoord = field.newField( 'texCoord', 'SFNode', 1, node.NULL)
    height = field.newField( 'height', 'MFFloat', 0, list)
    color = field.newField( 'color', 'SFNode', 1, node.NULL)
    colorPerVertex = field.newField( 'colorPerVertex', 'SFBool', 0, 1)
    xDimension = field.newField( 'xDimension', 'SFInt32', 0, 0)
    zSpacing = field.newField( 'zSpacing', 'SFFloat', 0, 0.0)
    normalPerVertex = field.newField( 'normalPerVertex', 'SFBool', 0, 1)

that's a node which has default-values for each field, as well as 
per-field watchability, introspection, type-checking and coercion.  
(Incidentally, each property shows up as a single entry in documentation 
systems (not as 39 accessor methods + 13 properties)).  Sure, you can do 
that with getter/setter methods, but... well... why?  Using property 
sub-classes is the solution that seems right to me.

As with all religions, feel free to join or not to join :) .
Unity in diversity and all that.
And, heck, we need a good lynching once in a while,
keeps the blood fresh better'n leaches ;) :D .

Feel the Fizz ;) ,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From skip at pobox.com  Sun Apr  6 13:53:01 2003
From: skip at pobox.com (Skip Montanaro)
Date: Sun, 6 Apr 2003 12:53:01 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m21y0fa5kq.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain>
        <7xn0j3fsdu.fsf@ruckus.brouhaha.com>
        <m21y0fa5kq.fsf@localhost.localdomain>
Message-ID: <16016.27005.335657.809556@montanaro.dyndns.org>

    Patrick> The Scintilla text control (written in C and wrapped in
    Patrick> wxPython by a control called wxStyledTextCtrl) handles the
    Patrick> on-screen representation and user-manipulation of the text.  So
    Patrick> that part is actually easy.  Just read in a file, feed it to
    Patrick> Scintilla, and go.  The next step is wrapping all this in an
    Patrick> Emacs-like framework.

If Scintilla doesn't distinguish between Emacs's notion of buffers (a chunk
of text being edited) and windows (a view onto a buffer) you may find the
task harder than it would first appear.

Skip




From drs at ecp.cc  Mon Apr 21 00:53:56 2003
From: drs at ecp.cc (drs)
Date: Mon, 21 Apr 2003 04:53:56 GMT
Subject: Zope and FreeBSD
References: <e786d63d.0304201806.4c5293a6@posting.google.com>
Message-ID: <EPKoa.2113$%_3.1657859@newssrv26.news.prodigy.com>

"Wenshan Du" <python at dohao.org> wrote in message
news:e786d63d.0304201806.4c5293a6 at posting.google.com...

> I've built a website with Zope 2.5.1 and FreeBSD 4.3, it's better than
> other solutions.
>
> I don't know why, is zope not stable on FreeBSD?

You should upgrade to the latest FreeBSD (4.8) before you do anything else.

-d





From mcfletch at rogers.com  Sun Apr 20 12:01:56 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Sun, 20 Apr 2003 12:01:56 -0400
Subject: PyOpenGL glClearColor() trouble
In-Reply-To: <20030420092541.18878.00000270@mb-m19.aol.com>
References: <mailman.1050806732.26162.python-list@python.org> <20030420092541.18878.00000270@mb-m19.aol.com>
Message-ID: <3EA2C474.1010404@rogers.com>

TheDustbustr wrote:

>>use a real GUI library ;)
>>    
>>
>Which library do you suggest?
>  
>
I normally use wxPython, though PyGame is very popular among the gaming 
folks, and GLUT is popular for those who want to be able to follow 
common samples and documentation written for C.  (Honestly, though, I 
almost always use OpenGLContext when I'm creating OpenGL code, so it all 
works the same and I normally forget which of those three I'm using.)

BTW, I was mostly joking here.  I just don't like the Tk widget's 
approach (or Tkinter's look), (and I hate maintaining the widget,) but 
lots of people use it quite happily :) .  Tk does have a slightly more 
"bare-bones" widget in __init__.py, see the online pydocs for 
documentation on it.

>>I'm rather surprised if you care, it's very >seldom that such a call winds up
>>    
>>
>having a >significant overhead in an application
>I *believe* that this is also the source of another problem (my textures don't
>map to their poly's, my glColor3f() calls don't do anything), is this possible?
> I'm also using an event driven interface, could that be a problem?
>  
>
glClearColor() is an entirely seperate variable in the state machine 
from glColor3f().  It's really unlikely that they are doing anything to 
each other (if they really are it's likely a bug in your OpenGL 
implementation).

Textures not mapping to polygons is often a matter of:

    * forgetting to specify mapping coordinates
    * (comparatively rare) forgetting to specify texture mapping style
      (i.e. GL_REPEAT)
    * forgetting glBindTexture(GL_TEXTURE_2D, self.texture)
    * forgetting glEnable(GL_TEXTURE_2D)
    * not actually loading the image into the appropriate textureID

glColor3f similarly requires some setup to do "what you expect" in some 
cases, though it's normally less inclined to be messed up by aberrant state.

You can see a sample of how to setup a texture for rendering in 
OpenGLContext.texture and OpenGLContext.scenegraph.imagetexture.

Event driven interfaces are not a problem.  wxPython and GLUT are both 
event driven.  PyGame is too, it just has you write the event loop :) .

Good luck,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From achrist at easystreet.com  Mon Apr  7 14:30:49 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Mon, 07 Apr 2003 11:30:49 -0700
Subject: Python & Music: Performance question
References: <1759581e.0304060218.164d4cd4@posting.google.com> <1759581e.0304070549.3b3a4caa@posting.google.com>
Message-ID: <3E91C3D9.E274C521@easystreet.com>

Is there anything in Python like the Midi support in the java
sdk?  (under MS Windows) It's pretty easy in java to send midi data 
straight from the program to the midi sequencer or synthesizer
(no need to write a file or deal with some external player).   
Anything like that for Python?

Al
     

David Asorey ?lvarez wrote:
> 
> forodejazz at yahoo.es (David Asorey ?lvarez) wrote in message news:
> <1759581e.0304060218.164d4cd4 at posting.google.com>...
> > Hello, group.
> >
> > I'm developing a small music program, and I have some doubts.
> >
> >  [...]
> 
>   Thanks to all.
>   I've tried some test with the first version and there is not any
> problem, even with thousands of notes:
> 
> class Note:
>    def __init__(self, pitch, duration):
>       self.pitch = pitch
>       self.duration = duration
>    def change_pitch(factor): pass
>    def change_duration(factor): pass
> 
> if __name__ ==  '__main__':
>     import time
>     notes = []
>     t1 = time.clock()
>     for x in range(10000):
>         notes.append(Note(x, x))
>     t2 = time.clock()
>     print t2 - t1
> 
> Pentium 800, 128 Ram, W2K: 0.16 s
> Pentium 133, 40 Ram, Linux (woody): 1.67 s
> 
> For my purposes, this are acceptable values.
> 
> Again, thanks for your advices.
> 
> David



From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr  3 11:51:25 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 3 Apr 2003 16:51:25 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com>
Message-ID: <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq at 4ax.com>, carroll at tjc.com wrote:
> On Tue, 1 Apr 2003 23:52:52 -0500, David Sfiligoi
><webmaster at quanta1.world--vr.com> wrote:
> 
>>I built a small script that use the xover function in the nntplib module.   
>>The problem that I came across is that what xover can return a huge tuple 
>>when there is 1000s of article in a newsgroup(which is frequent)
>>
>>testxover_resp,testxover_subs = s.xover(start,end)
>>
>>On my system it's not an issue I have 768 Mb of RAM.... but I have to 
>>believe that there is a way to optimise this while keeping all this simple.
>>
>>How can I limit the amount of memory xover would take. the other I did an 
>>xover of a huge group and the python process was taking about 650Mb Res. 
>>memory.  
> 
> This is a tough one; the problem is, if you were actually issuing the
> XOVER command to the NNTP server, you could handle it line-by-line as
> it come in.  But nntplib does that on your behalf and hands you the
> whole tuple, filled with all the headers.
>
> So as long as you're using nntplib instead of rolling your own (and I
> think that using nntplib the way to go), you're going to have the
> constraint that you have to have the resources to handle that big
> tuple.
<snip>

Or you can edit nntplib to add the option of storing XOVER output to a
file.  It doesn't look like this would be too hard.  The 'file' could
be an object that processes each line as it comes, rather than a real
file.  Then send your change to nntplib back to the author and hope it's
accepted into the official version.



From ulope at gmx.de  Wed Apr  2 10:19:00 2003
From: ulope at gmx.de (Ulrich Petri)
Date: Wed, 2 Apr 2003 17:19:00 +0200
Subject: I need help
References: <31ef0f43.0304011338.7ae3d45b@posting.google.com>
Message-ID: <b6ev17$4k8eb$1@ID-67890.news.dfncis.de>

"Jewlz408" <Jewlz408 at aol.com> schrieb im Newsbeitrag
news:31ef0f43.0304011338.7ae3d45b at posting.google.com...
> I need a lot of help
Indeed

> What do you know about Arrays
Everything and nothing.

<sniped basic code>

>
> Thankyou so much for your time
> Please help me!

1. Wrong NG
2. Double post

> E-mail me with your answer please at
> Jewlz408 at aol.com

No. Post here, read here





From jcollins at boulder.net  Tue Apr 29 17:29:32 2003
From: jcollins at boulder.net (jcollins)
Date: Tue, 29 Apr 2003 15:29:32 -0600
Subject: Tiny (with respect to memory) Python
In-Reply-To: <b8mc2o0c8m@drn.newsguy.com>
References: <b8mc2o0c8m@drn.newsguy.com>
Message-ID: <3EAEEEBC.9010906@boulder.net>

pacquets at newsguy.com wrote:

>I can't find any recent postings about this, so I hope someone can help me out
>here:
>
>I ported 2.0 to our weird little real-time OS awhile ago, and I found the memory
>consumption pretty unacceptable for an embedded device (for our low-end models,
>anyway).  Some stats:
>
> * enter the interpreter, add two numbers, exit.  High-water mark:  153K
>
> * enter again, do the following & exit:
>
>  
>
>>>>import os
>>>>os.listdir('.') 
>>>>        
>>>>
>
>high-water mark:  850K
>
>This is not to throw stones;  I think Python's a wonderful language and I'd like
>all our users to have it.  But I really need to bring that number down
>drastically for the low-end products.  I accept that that might require *really*
>draconian measures.  What's the best way to do that?
>
>Some thoughts:
>
>1) there is Pippy, which is now on SourceForge and seems pretty well abandoned. 
>I haven't figured out yet what they did to bring the footprint down.
>
I put a source tar file in the downloads section ("Reduced Python") a 
few weeks ago.  It contains all of the footprint reducing features of 
Pippy, which it turns out can also reduce base heap usage.  Other 
aggressive measures, such as moving dictionaries off  the heap to 
storage memory, running .pyc code directly from storage, etc., aren't 
applicable to the desktop and so aren't included.   Just do a 
./configure --help to display a list of features that may be removed at 
compile time.  FYI, this version of Python is 1.5.2+, just pre-unicode 
so it should be very similar to the 2.0 version.


-- 
Jeffery Collins (http://www.boulder.net/~jcollins)






From harri.pasanen at trema.com  Thu Apr 10 12:03:43 2003
From: harri.pasanen at trema.com (Harri Pasanen)
Date: 10 Apr 2003 09:03:43 -0700
Subject: python 2.3a2 Win32 expected failures in regression tests?
References: <mailman.1049910795.4707.python-list@python.org>
Message-ID: <e2d925a8.0304100803.46e63240@posting.google.com>

Tim Peters <tim.one at comcast.net> wrote in message news:<mailman.1049910795.4707.python-list at python.org>...

> ... 2.3a2 is so old now that it would be better to
> test current CVS Python.

I poked around Python CVS, but I did not find any Changes.log, or
equivalent, to say what has changed since 2.3a2, or from 2.2.2 to
that. What would be the best way to find out?

I chasing an obscure bug where OmniORB's omniidl fails when I build it
using Python 2.3a2 on W2K, while on Linux it works.  On W2K it works
with Python 2.2.2.   Some of its internal import's seem to fail
silently, providing partially imported modules, but only for 2.3a2,
W2K combo.   The culprit is either me, omniorb, or python, and I'm
trying to pin it down.  I seem to recall there were changes to the
import mechanism since 2.2.2?

Regards,

Harri



From sholden at holdenweb.com  Sat Apr 12 10:29:50 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sat, 12 Apr 2003 14:29:50 GMT
Subject: Nice quote
References: <just-60BB53.18302728032003@news1.news.xs4all.nl> <pan.2003.03.28.20.18.26.483935@jerf.org> <b7251f$1ma$2@atlantis.news.tpi.pl>
Message-ID: <ypVla.61025$It5.53369@news2.central.cox.net>

"Jarek Zgoda" <jzgoda at gazeta.usun.pl> wrote in message
news:b7251f$1ma$2 at atlantis.news.tpi.pl...
> Jeremy Bowers <jerf at jerf.org> pisze:
>
> >>   """The Python people also piped to say "everything's just
> >>      fine here" but then they always do, I really must learn that
> >>      language."""
> >
> > Would somebody amplify on this please? What XML library do we have
that's
> > easier to use then DOM or SAX for real world tasks that doesn't require
> > reading the whole file into memory to work?
>
> Is it a problem to read whole file to parse the XML? Really? I'm happy
> using ReportLab's pyRXP and effbot's ElementTree (loosely based on
> J.Clark's expat parser...) - they both use similar approach to build
> document structure. And both are fast, with pyRXP being faster but with
> less intuitive interface.
>

It is a problem when the file is several  gigabytes long, I suppose, and
people are for some reason using XML to transfer data structures that long
according to reports on this group.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From andrew at acooke.org  Mon Apr 28 12:46:51 2003
From: andrew at acooke.org (andrew cooke)
Date: Mon, 28 Apr 2003 12:46:51 -0400 (CLT)
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
In-Reply-To: <UXara.22428$K35.745468@news2.tin.it>
References: <mailman.1051532982.25352.python-list@python.org> 
     <UXara.22428$K35.745468@news2.tin.it>
Message-ID: <1246.10.2.0.2.1051548411.squirrel@127.0.0.1>

Alex Martelli said:
> It's an important design principle, which I first saw proposed in print
> by Andrew Koenig in his landmark book "C Traps and Pitfalls": *all bounded
> loops should be first-bound-included, last-bound-excluded* -- and which I
> will therefore call "Koenig's Principle" although I have never seen the
[...]

hmmm.  also appeared in his "ruminations on c++".  is there a subset of
his books that covers all his wisdom? ;o)

andrew

-- 
http://www.acooke.org/andrew




From graham__fawcett at hotmail.com  Tue Apr 22 11:12:33 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 22 Apr 2003 08:12:33 -0700
Subject: Should I prefer an external database
References: <u3ckbymri.fsf@mail.afm.dtu.dk> <mailman.1051001675.31208.python-list@python.org>
Message-ID: <e9570f37.0304220712.37a3109d@posting.google.com>

Ian Bicking <ianb at colorstudy.com> wrote in message news:<mailman.1051001675.31208.python-list at python.org>...

> Now, that all said, you might want an external database with an
> object-relational mapper.  Using SQLObject (plug: http://sqlobject.org),
> you'd do it like:
> 
> from SQLObject import *
> 
>  [snipped code]


Oh, that's very interesting, Ian. I wasn't aware of the project. 

Do you have, or are you working toward, support for transactions?

I wonder, too, if you support Postgresql as a backend and if so, is
there any plan to use its inheritance features for mapping Python
subclasses. Just curious...

-- Graham



From tim.one at comcast.net  Thu Apr 17 11:36:19 2003
From: tim.one at comcast.net (Tim Peters)
Date: Thu, 17 Apr 2003 11:36:19 -0400
Subject: fixedpoint.py vs Python 2.1
In-Reply-To: <$uJOMKAPtqn+Ew9c@jessikat.demon.co.uk>
Message-ID: <BIEJKCLHCIOIHAGOKOLHEEDOFGAA.tim.one@comcast.net>

[Robin Becker]
> I'm having a bit of difficulty with the fixedpoint module in Python 2.1.
>
> I get the following error
>
> Python 2.1.3 (#35, Apr  8 2002, 17:47:50) [MSC 32 bit (Intel)] on win32
> Type "copyright", "credits" or "license" for more information.
> >>> from fixedpoint import FixedPoint as fp
> >>> fp(fp(.175,3)*fp(22,2),2)
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "fixedpoint.py", line 368, in __mul__
>     return _mkFP(self._roundquotient(n1 * n2, _tento(p)), p, type(self))
>   File "fixedpoint.py", line 497, in _mkFP
>     f = FixedPoint()
> TypeError: object of type 'type' is not callable
> >>>
>
> is there some reason for passing type(self) around rather than
> self.__class__ etc.

This must have come from the SourceForge version of FixedPoint, so ask
there.

Find a version of the code that predates the SF one, and it won't have this
problem (it didn't originally use type(self) or self.__class__).

I don't know whether the SF folks intended to support Pythons earlier than
2.2.





From mertz at gnosis.cx  Mon Apr  7 13:31:17 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Mon, 07 Apr 2003 13:31:17 -0400
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com>
 <mailman.1049389032.4904.python-list@python.org>
  <u8bka.651$001.591065@monger.newsread.com>
Message-ID: <lXbk+kKkXYqd092yn@gnosis.cx>         

msw <markscottwright at hotmail.com> wrote previously:
|The only answers you will get in this group are:
|1. intellectual property law is sufficient
|2. no protection scheme is perfect

Y'know.  If a whole lot of people who are a lot smarter than you give
the same answer to a question (repeatedly), it might be worth
contemplating that there really IS something to that answer.

(the only time -I- have disagreed with the Timbot, I had the comfort of
knowing the Martellibot was on my side :-)).

Yours, Lulu...

--
Keeping medicines from the bloodstreams of the sick; food from the bellies
of the hungry; books from the hands of the uneducated; technology from the
underdeveloped; and putting advocates of freedom in prisons.  Intellectual
property is to the 21st century what the slave trade was to the 16th.




From max at alcyone.com  Fri Apr  4 17:43:57 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 04 Apr 2003 14:43:57 -0800
Subject: What are the differences between _weak_ references and "normal" 
 references?
References: <169a7283.0304040901.4bc3531b@posting.google.com> <mailman.1049477291.1423.python-list@python.org>
Message-ID: <3E8E0AAD.913D16F5@alcyone.com>

Jp Calderone wrote:

>   Weak references to objects don't increment the refcount for that
> object.

Right.  And an importang thing to point out from someone with a C++
background is that weak references understand when their referent is
gone, and won't run off into segmentation fault land.  (In C++, you can
delete the referent, but the reference itself will have no idea and will
gladly dereference it.)

>>> import weakref
>>> class C: pass
... 
>>> c = C() # create an object
>>> 
>>> c1 = weakref.ref(c) # and a weak reference to it
>>> c1
<weakref at 0x81593a4; to 'instance' at 0x815a214>
>>> c1() # call the reference to get the referent
<__main__.C instance at 0x815a214>
>>> del c # now the weak reference is dead
>>> c1
<weakref at 81593a4; dead>
>>> c1()
>>> c1() is None # now the referent is None
1

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If the sky should fall, hold up your hands.
\__/ (a Spanish proverb)
    The laws list / http://www.alcyone.com/max/physics/laws/
 Laws, rules, principles, effects, paradoxes, etc. in physics.



From gmuller at worldonline.nl  Fri Apr 25 13:15:25 2003
From: gmuller at worldonline.nl (GerritM)
Date: Fri, 25 Apr 2003 19:15:25 +0200
Subject: Python Package Index @ python.org down?
References: <3ea96bce$0$49106$e4fe514c@news.xs4all.nl>
Message-ID: <b8bqeu$qtd$1@reader1.tiscali.nl>

"Irmen de Jong" <irmen at -NOSPAM-REMOVETHIS-xs4all.nl> schreef in bericht
news:3ea96bce$0$49106$e4fe514c at news.xs4all.nl...
> I get a Python stacktrace when trying to access the Python Package Index,
> http://www.python.org/pypi
>
> "OperationalError: database is locked"
>
> Tried three times. Anything wrong over there?
>
> --Irmen
>
The full error message I get is:

Traceback (most recent call last):
  File "/usr/local/apache/cgi-bin/pypi", line 46, in ?
    handler.run()
  File "/usr/local/pypi/lib/pypi/webui.py", line 108, in run
    self.store.open()
  File "/usr/local/pypi/lib/pypi/store.py", line 480, in open
    self.cursor = self.conn.cursor()
  File "/usr/local/pypi/lib/sqlite/main.py", line 592, in cursor
    return Cursor(self)
  File "/usr/local/pypi/lib/sqlite/main.py", line 199, in __init__
    self.con._begin()
  File "/usr/local/pypi/lib/sqlite/main.py", line 543, in _begin
    self.db.begin()
OperationalError: database is locked

 regards Gerrit

--
www.extra.research.philips.com/natlab/sysarch/






From aahz at pythoncraft.com  Tue Apr  8 09:24:35 2003
From: aahz at pythoncraft.com (Aahz)
Date: 8 Apr 2003 09:24:35 -0400
Subject: super question
References: <fd1c664e.0304071056.31a0d377@posting.google.com> <b6soav$7p2$1@panix1.panix.com> <Qdpka.2889$Zu2.2627@news01.roc.ny.frontiernet.net>
Message-ID: <b6uiij$2u7$1@panix2.panix.com>

In article <Qdpka.2889$Zu2.2627 at news01.roc.ny.frontiernet.net>,
Lee Harr  <missive at hotmail.com> wrote:
>In article <b6soav$7p2$1 at panix1.panix.com>, Aahz wrote:
>> 
>>     class b(object):
>>         pass
>
>From the super docs ;o)  it says you might do this:
>
>    class C(B):
>        def meth(self, arg):
>            super(C, self).meth(arg)
>
>but why not just do:
>
>    class C(B):
>        def meth(self, arg):
>            B.meth(self, arg)
>
>Is there any difference?
>Any advantage either way?

The advantage comes when you rename B or when you have multiple base
classes for C.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From BPettersen at NAREX.com  Sun Apr 13 19:24:21 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Sun, 13 Apr 2003 17:24:21 -0600
Subject: win32all odbc/MS SQLServer/bigint
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEB1@admin56.narex.com>

> From: M.-A. Lemburg [mailto:mal at lemburg.com] 
> 
> Bjorn Pettersen wrote:
> >>From: M.-A. Lemburg [mailto:mal at lemburg.com] 
> >>
> >>Bjorn Pettersen wrote:
> > 
> > [...]
> > 
> 
> The odbc module returns a long in case floor(x) == x, but the
> interfacing at C level is done using a C double.

Ah, I see. Our account IDs are a combination of bitfields (thus using
all 64 bits) so they're always large enough for this to be true...

[...]

> Sure, but __int64 is only available in VC C++ AFAIK. Many
> compilers have a "long long" type which could be used, but
> then again, how do you know whether the ODBC driver was
> compiled with the same C type and layout as the application
> using it ?

I haven't found a compiler without a long long type, but then I haven't
checked e.g. the Palm or Cray's, so I'm assuming they're out there :-)
This isn't a concern for the win32all odbc module of course, but all
cross platform projects I've seen so far has a #define LONGLONG to
either long long or __int64 somewhere (with an appropriate comment
<wink>).

If you call

  SQLColAttribute(hstmt, n, SQL_DESC_TYPE, NULL, 0, NULL,
(SQLPOINTER)&sqlType);

and sqlType == SQL_BIGINT, that would be a good indication for result
sets, and if

  SQLBindParameter(hstmt, num, SQL_PARAM_INPUT,
                   SQL_C_SBIGINT, SQL_BIGINT, 
                   sizeof(__int64), 0, 
                   (void*)val, sizeof(__int64), &m_nullInd);

returns either SQL_SUCCESS or SQL_SUCCESS_WITH_INFO that ought to be
good enough for parameters?

> It's one of those MS things again... the half cooked, "works
> for me" kind of attitude.

And here I thought it was the "everyone else did it that way, so we'll
do it different just because we can" attitude <wink>.

[...]

> Seriously, performance on the Python processing side is usually
> not an issue (mxODBC is fast); it's network latency, database and
> query optimizations that do matter.

All true, but when all that is taken care of you have to sqeeze where
there's juice left :-) (the reason we're using Python is that the
algorithm calls for ~200K dict inserts and lookups times 10,000 -- i.e.
Python should be faster than C++ :-)

-- bjorn




From bkelley at wi.mit.edu  Wed Apr  9 11:42:20 2003
From: bkelley at wi.mit.edu (Brian Kelley)
Date: Wed, 09 Apr 2003 11:42:20 -0400
Subject: killing child processes
In-Reply-To: <39ee6fa9.0304090148.745fdc46@posting.google.com>
References: <39ee6fa9.0304090148.745fdc46@posting.google.com>
Message-ID: <3e943db1$0$3937$b45e6eb0@senator-bedfellow.mit.edu>

A partial answer to your question.  If you have their PID (process 
handles) then you can delete a windows process using 
win32process.TerminateProcess(pid, 0)

# start the process
hProcess, hThread, dwPid, dwTid = win32process.CreateProcess(...)

# sometime later
win32process.TerminateProcess(hProcess, 0)

For more info look here...
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dllproc/base/terminateprocess.asp

reinhard wrote:
> I've created a process (some batchfile) which in turn creates some
> long running child processes using the win32api CreateProcess call
> (declaring them as a group).
> 
> Is there a way to terminate (kill) all the child processes in the
> group?
> 
> - reinhard nadrchal





From pobrien at orbtech.com  Sun Apr  6 11:13:31 2003
From: pobrien at orbtech.com (Patrick K. O'Brien)
Date: 06 Apr 2003 10:13:31 -0500
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <m3r88g9iqt.fsf@mira.informatik.hu-berlin.de>
Message-ID: <m2brzjacbo.fsf@localhost.localdomain>

martin at v.loewis.de (Martin v. L?wis) writes:

> pobrien at orbtech.com (Patrick K. O'Brien) writes:
> > Here is the background.  I'm working on a file editor (called
> > PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> > PyCrust code base.  I've got a basic version working, and it supports
> > editing multiple files and all the usual, basic features.  But I want
> > it to be as customizable and extensible as Emacs.  In fact, I want to
> > model it after Emacs, so that elisp code could be rewritten in Python
> > and work with PyAlaMode.  To do that, I've got to support the same
> > primitives as Emacs; expose files, buffers, regions, and windows the
> > same way Emacs does; handle the Emacs keybindings and ability to remap
> > keys, the same way Emacs does, etc.
> 
> I think you don't have much design choices then, if you want to expose
> the very *same* functionality. In fact, you have to chose whether you
> want to make it compatible with Emacs or with XEmacs, and you have to
> specify a specific release you target, as each release has its own API
> modifications.

Since I'm only looking for 95% compatibility, I should think a recent
release will be fine.  I'm using GNU Emacs 21.2.1.

> I recommend that you don't; there are certain aspects in Emacs that
> are there for historical reasons only and that should be dropped. My
> favourite here is the Emacs Mule representation, which represents
> characters in multi-byte form, using ISO 2022 escape codes to switch
> between character sets. I recommend that you use Unicode
> instead. This, of course, means that an entire suite of elisp cannot
> be readily ported to your editor, but must be rewritten.

I agree that warts like this should be removed.  And I don't expect to
be able to convert elisp in a completely automated fashion.  But the
API and basic behavior (is that redundant?) need to be as similar as
possible.  Which means how files, buffers, the minibuffer, windows,
etc. are created and exposed to elisp code needs to be understood and
modeled in Python.  I have an understand from a user point of view,
but I need a better understanding from a desing and implementation
point of view in order to code it in Python.

> I would suggest a bottom-up strategy rather than a top-down strategy:
> Take your favourite elisp package (say, gnus, or python-mode), and try
> to make it run. 

I agree.  In fact, python-mode is my first target.

> To do so, write an automated translator that translates all required
> elisp files to Python code. Don't make that translator
> perfect. Instead, make it just translate the files you are looking at,
> and extend it as you find it fails to translate existing elisp code.
> Special-case builtin functions in the translator as appropriate
> (e.g. 'require).

Are there translators out there?  I looked and couldn't find any.  I
wonder if Pymacs could be useful for this.

> Make the translator translate all elisp code. You will then find that
> the package doesn't work, because there are primitive functions which
> are defined in C, not in elisp. Try to write these functions in
> Python, using the Scintilla and Python API.
> 
> That approach will require constant refactorization, but it will tell
> you early on what you can expect and what is impossible to achieve.

That's pretty much the approach I planned to take, after I got some of
the more basic pieces in place.  For example, file handling.  If you
try to save a file and another process has modified it, Emacs warns
you.  Where is that kind of code kept in Emacs?  Is it in elisp, or
part of the core Emacs written in C?  Is there a document anywhere
that gives an overview of the design of Emacs that would answer these
kinds of questions?

-- 
Patrick K. O'Brien
Orbtech      http://www.orbtech.com/web/pobrien
-----------------------------------------------
"Your source for Python programming expertise."
-----------------------------------------------



From sross at connectmail.carleton.ca  Sat Apr 26 14:13:57 2003
From: sross at connectmail.carleton.ca (Sean Ross)
Date: Sat, 26 Apr 2003 14:13:57 -0400
Subject: Some Python Code
References: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com> <mailman.1051378901.13527.python-list@python.org>
Message-ID: <SZzqa.6651$Qg2.1045753@news20.bellglobal.com>

"andrew cooke" <andrew at acooke.org> wrote in message
news:mailman.1051378901.13527.python-list at python.org...
>
> why not put the code on a web page somewhere?  that way it will eventually
> get googled and people will find it when they need something like that and
> search for it, and they can download it.
>
> cheers,
> andrew
>


True enough. I may do so eventually, however, at the moment I do not have a
web page and I wasn't looking to start one. I'll look in to it.
Thanks for your suggestion,
Sean





From ckea25d02 at sneakemail.com  Tue Apr  8 01:20:58 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 17:20:58 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1049776891.5548.python-list@python.org>
References: <m2u1dcw22s.fsf@localhost.localdomain> <mailman.1049731219.27404.python-list@python.org> <b6teqm$956ai$1@ID-169208.news.dfncis.de> <mailman.1049776891.5548.python-list@python.org>
Message-ID: <b6tllv$8uchs$2@ID-169208.news.dfncis.de>

Sean 'Shaleh' Perry wrote:
> However the reason most of the emacs modes, vim syntax 
> definitions, etc all don't quite work is they are regex based.  To really get 
> where the  structure editor needs to be would require a parser for each item 
> type.  Not a trivial task.

No, but it would be a much better foundation to build
this stuff on, I think.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr  3 11:42:09 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 3 Apr 2003 16:42:09 GMT
Subject: How to obtain python file path
References: <20030401163212631-0500@news.covad.net> <slrnb8jusc.5u1.ramen@lackingtalent.com> <20030401165546102-0500@news.covad.net> <slrnb8lr3q.15s.do-not-spam-ben.hutchings@tin.bwsint.com> <20030402115351679-0500@news.covad.net>
Message-ID: <slrnb8op30.174.do-not-spam-ben.hutchings@tin.bwsint.com>

Please don't top-post.

In article <20030402115351679-0500 at news.covad.net>, Dan Grassi wrote:
> This works only if the python script is imported, not if it is executed 
> directly.

Well then:

    def mydir():
        import os.path, sys
        if __name__ == '__main__':
            filename = sys.argv[0]
        else:
            filename = __file__
        return os.path.abspath(os.path.dirname(filename))

> Gee, I'm surprised that this is such a difficult thing to do yet 
> conceptually simple.

I don't think it is all that simple, but maybe I've just got used to the
idea that files don't know (and don't need to know) their names.



From trentm at ActiveState.com  Mon Apr 14 15:00:31 2003
From: trentm at ActiveState.com (Trent Mick)
Date: Mon, 14 Apr 2003 12:00:31 -0700
Subject: debug build of python2.2 for win
In-Reply-To: <m2adetospe.fsf@mother.paradise.lost>; from jdhunter@ace.bsd.uchicago.edu on Mon, Apr 14, 2003 at 01:10:53PM -0500
References: <m2adetospe.fsf@mother.paradise.lost>
Message-ID: <20030414120031.A27228@ActiveState.com>

[John Hunter wrote]
> 
> Any one know where I can get a debug build of python 2.2.2 for win32?

http://ftp.activestate.com/ActivePython/etc/
    OR
ftp://ftp.activestate.com/ActivePython/etc/

Cheers,
Trent

-- 
Trent Mick
TrentM at ActiveState.com




From davecook at nowhere.net  Tue Apr 22 05:22:51 2003
From: davecook at nowhere.net (David M. Cook)
Date: Tue, 22 Apr 2003 09:22:51 GMT
Subject: Should I prefer an external database
References: <u3ckbymri.fsf@mail.afm.dtu.dk>
Message-ID: <LR7pa.144615$It5.84466@news2.central.cox.net>

In article <u3ckbymri.fsf at mail.afm.dtu.dk>, Brian Elmegaard wrote:

> I am working on a script for organizing a program for a conference. I
> started out using a connection to mySQL, but then I thought that it
> would be easier to keep it within python. So I implemented a class:
> 
> class Paper:
>         def __init__(title,author,session):
>                 self.title=title
>                 self.author=author
>                 self.session=session

You might want to take a look at SQLObject

http://sqlobject.org

or IndexedCatalog (which sits atop ZODB):

http://www.async.com.br/projects/IndexedCatalog/

Dave Cook



From mlh at furu.idi.ntnu.no  Sat Apr 26 13:34:20 2003
From: mlh at furu.idi.ntnu.no (Magnus Lie Hetland)
Date: Sat, 26 Apr 2003 17:34:20 +0000 (UTC)
Subject: Another itertool function?
Message-ID: <slrnbalgos.am0.mlh@furu.idi.ntnu.no>

The need for the following function just cropped up in a parser I'm
writing; it seems sufficiently general to perhaps be a candidate for
itertools -- and I can't see any way of implementing it with the
current itertools functions:

def weave(*iterables):
    iterables = map(iter, iterables)
    while iterables:
        for it in iterables[:]:
            try:
                yield it.next()
            except StopIteration:
                iterables.remove(it)

Another possible name (which is a bit more unwieldy) would be
'roundrobin'. Another one still is iweave (to keep the naming
convention, even though here there is no clash).

This function is somewhat similar to izip, but goes through the
iterables in a round-robin fashion, dropping them as they become
empty.

I used this with items that had indices (or timestamps) associated
with them, and for each iteration I added the new ones to a heap. Then
I could get the one with the lowest index (or timestamp) etc.

I expect there would have to be more use-cases to justify it, but I
just thought I'd air the idea...

-- 
Magnus Lie Hetland               "Nothing shocks me. I'm a scientist." 
http://hetland.org                                   -- Indiana Jones



From imbosol-1050078176 at aerojockey.com  Fri Apr 11 14:30:29 2003
From: imbosol-1050078176 at aerojockey.com (Carl Banks)
Date: Fri, 11 Apr 2003 18:30:29 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com>   <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk>
Message-ID: <9RDla.592$Gb1.345@nwrdny01.gnilink.net>

Alexander Schmolck wrote:
> Carl Banks <imbosol-1049990205 at aerojockey.com> writes:
> 
>> What you're missing is that Globals are Evil(tm), and furthermore, that if
> 
> Oh, so *that's* what I was missing -- silly me.
> 
>> your program depends on changing a global in a dynamic scope, it badly needs
>> to be redesigned. (To wit, that functions that refer to the global you want
>> to change need to be changed to accept it as an argument.)
> 
> You did notice, of course, that you don't only need to change that particular
> function foo, but also everything (directly or indirectly) calling foo? As
> well as everything that is called (directly or indirectly) by foo and that
> uses this global state information? Including 3rd party code and builtin
> constructs of the language?

Yes.  Since the idea here is to rewrite the Emacs in Python, you have
to rewrite all of this anyways.


>> I think dynamic scpoing is a wholly bad idea, and simplifying a very
>> misguided programming practice doesn't make it any less so.
> 
> How much time did you actually spend thinking about global variables and
> dynamic scope or what I wrote before you posted?
> 
> Why are global variables evil?
[snip reasons globals are evil]

I didn't say globals are evil.  I said Globals are Evil(tm).  Big
difference.  (You should learn the meaning of (tm); it means the
writer fully realizes that it's only a general truth, even though
others consider it an absolute truth.)


> Ever wondered why python (and about any non-functional language) still has
> global variables (and makes use of them in core modules such as sys)?
> 
> Because certain aspects of a program are orthogonal in a way that is best
> captured by state that is not explicitly passed around (even if by chance it
> would be possible to do it that way, because you can anticipate and correctly
> parameterize all state and have full control over all the involved code (see
> above)).

Well, in another part of this thread, I argued against total
elimination of global variables, much as you are arguing now.  I do
not oppose global variables in general.

I do oppose global variables where there are situations where it feels
necessary to change it and restore it.  That tells you right away that
your functions should not use a global, but should have it passed in.
Designing such a function to refer to the global anyways is Bad
Programming(tm).


[snip uses of globals]
> All this should *not* require rewriting each and every function and method
> call! How do you propose to address these issues (bearing in mind error
> handling and multithreading)?
> 
> Maybe there is a superior way to handle all or some of those issues (AOP,
> Monads, god knows what), but you haven't mentioned it.

These issues should be handled by global variables if they indeed are
global in nature.  The stuff you mentioned usually is.


> Similarly, you have given no indication why you think lexically
> scoped global state better address these problems than dynamically
> scoped state (or why its downsides such as new sources for bugs and
> complicating the language would outweigh the benefits). 

I have not said lexical scoping handles the issues you have referred
to better than dynamic scoping.  I have said that the issues you refer
are only issues in the presense of Bad Programming(tm).  Anyone using
globals discriminately will not have such issues to deal with.

But to answer your question: yes, dyanmic scoping is so awful it
doesn't justify this particular benefit, especially when the benefit
only arises in Bad Programming(tm).  In fact, dynamic scoping is so
awful is doesn't justify any benefits.

It's not just me; meny, many people regard dynamic scoping as a bad
mistake.

I don't wish to go into detail here.  The main problems are: It forces
you to use defensive naming schemes, even for your local variables.
You cannot determine the variables in effect just by looking at a
function in context.  You cannot do lexical closures.  It's less
efficient.


> Rather you
> just pretended the issue didn't exist and I was just being stupid
> for not realizing this to start with.

Don't be presumptuous.


-- 
CARL BANKS



From dave at pythonapocrypha.com  Tue Apr  8 00:14:25 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Mon, 7 Apr 2003 21:14:25 -0700 (PDT)
Subject: How to protect python code ?
In-Reply-To: <3E921A4C.1020908@hotmail.com>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org>
 <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org>
 <3E921A4C.1020908@hotmail.com>
Message-ID: <Pine.LNX.4.51.0304072103470.31663@ns0.tbye.com>

On Tue, 8 Apr 2003, msw wrote:

> Dave Brueck wrote:
> > On Mon, 7 Apr 2003, msw wrote:
> >
> > Not true - check the archives. In the past people, including myself, have
> > posted the general steps to do to "protect" the code. It's not worth the
> > effort in most cases, IMO, but answers *have* been given.
>
> If so, I was unable to find it, and I think this thread proves my point.

And your failed search proves my point:

http://tinyurl.com/91al

(found it in <30 seconds!)

I'm sure there are plenty of people who read this group who would be happy
to do contract work for you to code this up for you, but all the people
who are willing to do it for you for free have spoken.

I take that back - if you act on some of the suggestions you've received
so far (as in, try them out, for example), and then come back and say,
"Ok, I've tried X.  Along the way I learned Y. Now I'm stuck on Z." then
you'll probably get additional help. But I mean, really, you need to meet
people halfway - you've gotten free pointers and hints from busy people
who don't even see the *value* in trying to solve that problem that way -
a sign of real courtesy IMO.

Have a nice day,
-Dave




From sholden at holdenweb.com  Thu Apr 17 09:20:09 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Thu, 17 Apr 2003 13:20:09 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com> <3e9d9bb9$0$151$a1866201@newsreader.visi.com> <a8b7f07a.0304161314.361a80fa@posting.google.com> <3e9dca74$0$170$a1866201@newsreader.visi.com>
Message-ID: <dSxna.114901$It5.77194@news2.central.cox.net>

"Grant Edwards" <grante at visi.com> wrote in message
news:3e9dca74$0$170$a1866201 at newsreader.visi.com...
> In article <a8b7f07a.0304161314.361a80fa at posting.google.com>, A. Lloyd
Flanagan wrote:
>
> >> > python << DONE
> >> > print 2
> >> > for i in (1,4):
> >> >      print i
> >> > DONE
> >>
> >> But how do you do it in a Makefile?
> >
> > Err.  Put it in a script file, and call the script file from make?
>
> That's one solution, but now you've got another file to keep
> track of.
>
> > I tried sticking it in a Makefile, can't get it to work.  So far.
>
> How 'bout this:
>
> all:
>         python2 -c $$'for i in range(5):\n print i\n\n'
>
>
> I think that the $'<string>' notation is a bash-ism.  If that's
> a problem, then
>
>         echo -e "for i in range(5):\n print i\n\n" | python2
>
> should also work.  Adjust the echo options appropriately to get
> interpretation of \n to work.
>
> There may be other ways to get Make to interpret \n in a
> string...
>
Now try putting an import before the "for" statement ...

regards
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/



> --
> Grant Edwards                   grante             Yow!  I want to dress
you
>                                   at               up as TALLULAH BANKHEAD
and
>                                visi.com            cover you with VASELINE
and
>                                                    WHEAT THINS...





From frithiof.jensen at remove_this.jensen.tdcadsl.dk  Tue Apr 15 06:31:17 2003
From: frithiof.jensen at remove_this.jensen.tdcadsl.dk (Frithiof Andreas Jensen)
Date: Tue, 15 Apr 2003 12:31:17 +0200
Subject: Sv: Python 2.2.2 & Tix on win32 - What must I do?
References: <b7666e$oto$1@newstree.wise.edt.ericsson.se> <mailman.1050061697.14672.python-list@python.org> <3e995853$0$24701$edfadb0f@dread14.news.tele.dk> <mailman.1050340346.14102.python-list@python.org>
Message-ID: <3e9bded2$0$24644$edfadb0f@dread14.news.tele.dk>

Raseliarison nirinA <nirina at mail.blueline.mg> skrev i en
nyhedsmeddelelse:mailman.1050340346.14102.python-list at python.org...

> ... win**** binary installation issues?
> How about installing a new and free operating system?
> Tix is a part of tcl package with most Linux distribution.

....That doesn't run CouterStrike, Day Of Defeat and Warcraft III?

I have room & money for *one* standard PC in my office; perhaps later, once
it becomes possible to buy a decent case for the VIA Epia boards, I shall
get a Linux box again!

Generally, Python is so good that I do not really need a Linux box
anymore...whatever works on Wintel will work equally well on Linux too,
unlike Java.

> Why wait?

I shall not, Tix has some really neat features for the users such as Bubble
Help but the first release needs Compatibility with users most: they got
Win32 laptops and little else. A second, third or whatever version can use
Tix...

> I'm sure you will run your Linux box before win binary will be
> released. I always think, i may be wrong, that installing from source is
> better than installing binary package.

I *hate* installing from src - often there is some rotten dependency with
some obscure piece of software or environment that sends me searching the
net for hours.

The *best* thing that happened to open source was package managers and
'canned installs'.





From Jewlz408 at aol.com  Tue Apr  1 16:36:05 2003
From: Jewlz408 at aol.com (Jewlz408)
Date: 1 Apr 2003 13:36:05 -0800
Subject: I need a lot of help!
Message-ID: <31ef0f43.0304011336.1d2777ac@posting.google.com>

I need a lot of help
What do you know about Arrays
How many storage Locations will be allocated for each of the given arrays?
1) Dim A(1 to 10)
2) Dim B(-3 to 5)
3) Dim C$( 12)
4) Dim D( 1 to 4, -2 to 2)
5) Dim E(4,6)

And What does the following programs do with its output?

Dim Table( 1 to 5, 1 to 3)
For I = 1 to 5
????? For J = 1 to 3
????? Let Table(I,J) = 1
????? Next J
Next I
End

What does the program do?
And what is the output?


Thankyou so much for your time 
Please help me!



From amk at amk.ca  Thu Apr 17 14:57:20 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Thu, 17 Apr 2003 13:57:20 -0500
Subject: Tutorial - section 8.5 -- User Defined Execptions?
References: <b7jauh$arj$1@ctb-nnrp2.saix.net> <3E9D35FC.FCAD17C9@alcyone.com> <b7mp2u$3k8$0@216.39.172.122> <3E9EEB54.682751BC@alcyone.com>
Message-ID: <bgCdne6erfqNZAOjXTWcoA@speakeasy.net>

On Thu, 17 Apr 2003 10:58:44 -0700, 
	Erik Max Francis <max at alcyone.com> wrote:
> issue.  I don't know what Guido would say about its use now, but there
> seems pretty near total consensus that repr(x) looks a lot better (and
> clearer) than `x`.

`x` is mentioned in his Python Regrets talk as one thing he would like to
take out.

--amk



From duncan at NOSPAMrcp.co.uk  Thu Apr 24 05:55:54 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 09:55:54 +0000 (UTC)
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <NKxpa.85054$T34.2112670@news2.tin.it> <698f09f8.0304231141.5d458eb0@posting.google.com> <vYDpa.50997$LB6.1152215@news1.tin.it> <b88akc$31dm$1@ns.felk.cvut.cz>
Message-ID: <Xns93676D68C5375duncanrcpcouk@127.0.0.1>

Frantisek Fuka <fuka at fuxoft.cz> wrote in 
news:b88akc$31dm$1 at ns.felk.cvut.cz:

> Many thanks to everyone. Can I get the "new" string functions 
> explanation somewhere? I have Python 2.2.2 but the current documentation 
> only seems to mention the old-style STRING module...
> 
See section 2.2.6.1 of the library reference documentation.
http://www.python.org/doc/current/lib/string-methods.html

P.S. Please can you try to avoid top posting as it makes it hard to follow 
the thread.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From rmunn at pobox.com  Mon Apr  7 16:45:21 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 07 Apr 2003 20:45:21 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain> <x7smsvh7kr.fsf@guru.mired.org>
Message-ID: <slrnb9344o.46e.rmunn@localhost.localdomain>

Mike Meyer <mwm at mired.org> wrote:
> Robin Munn <rmunn at pobox.com> writes:
>>     xover(start, end, open('foo.txt', 'w'))
>> 
>> See? File-like object gets created, passed to your function, and then
>> automatically garbage-collected (and thus closed) when your function
>> returns.
> 
> The garbage collection happening when your function returns is
> implementation specific. It works that way in CPython, but not
> Jython. You should always explicitly close files to make sure they get
> closed - and the output flushed, in this case - when you're done with
> them.

Good point, thanks. Ignore what I said above; what you should do is:

    output_file = open('foo.txt', 'w')
    xover(start, end, output_file)
    output_file.close()

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From rmunn at pobox.com  Wed Apr  2 12:02:27 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 02 Apr 2003 17:02:27 GMT
Subject: python & shared hosting
References: <3e88ef38$0$16524$a729d347@news.telepac.pt>
Message-ID: <slrnb8lgtc.ccp.rmunn@localhost.localdomain>

Jorge Neto <vispersona at hotmail.com> wrote:
> The shared hosting company only has cgi-perl but I would
> like to include some (little) web application on a shared hosting site.
> 
> Anyone knows if it's possible to code the web page with python and make 
> it run via the apache cgi ?
> 
> (I'm still learn python s take it easy on me)

You will probably want to read the cgi module documentation:

    http://www.python.org/doc/current/lib/module-cgi.html

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From aleax at aleax.it  Sun Apr 13 05:27:20 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 09:27:20 GMT
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <3E985DC8.9060907@thuule.pair.com> <mailman.1050178039.28676.python-list@python.org> <bdb5b58f.0304121731.1b1312f4@posting.google.com> <mailman.1050199162.26980.python-list@python.org>
Message-ID: <Y3ama.9044$T34.283440@news2.tin.it>

Alan McIntyre wrote:

> I wasn't interested in the time to execute the entire 200k append
> operations; as Mr. Martelli pointed out earlier, it's O(n^2), but the +=
> append (considered by itself) appears to be O(n).

Right: each append is O(N), so (since building up a long string of
small pieces requires O(N) appends) the overall operation is O(N**2).


Alex




From news at titanic.co.uk  Wed Apr 16 18:50:23 2003
From: news at titanic.co.uk (Simon Faulkner)
Date: Wed, 16 Apr 2003 23:50:23 +0100
Subject: MySQL CGI
Message-ID: <tnhr9vo5a0g86su1cadi77c4v1ilt5knq1@4ax.com>

I am looking for a Python cgi script which, given a MySQL table, will
display a list of the data in the table and allow add, edit, delete of
that data as that's how 90% of my projects start.

Does anything spring to mind or should I roll my own?

TIA

Simon



From drs at ecp.cc  Mon Apr 28 02:26:02 2003
From: drs at ecp.cc (drs)
Date: Mon, 28 Apr 2003 06:26:02 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com>
Message-ID: <_P3ra.4234$%_3.3369939@newssrv26.news.prodigy.com>

<achrist at easystreet.com> wrote in message
news:3EA97D61.3CC9FE74 at easystreet.com...
> Any suggestions of the shortest path from a Python program to RTF
> output?

if you haven't yet seen it, MS publishes the rtf specs:

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnrtfspec/html/rtfspec.asp





From eppstein at ics.uci.edu  Thu Apr 17 12:41:57 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Thu, 17 Apr 2003 09:41:57 -0700
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050529915.4666.python-list@python.org> <_Fzna.27993$LB6.643892@news1.tin.it>
Message-ID: <eppstein-F15082.09415717042003@news.service.uci.edu>

In article <_Fzna.27993$LB6.643892 at news1.tin.it>,
 Alex Martelli <aleax at aleax.it> wrote:

> Steven Taschuk wrote:
>    ...
> > trade-off as being between, say, an algorithm which runs in O(n)
> > time but needs O(n^2) memory to do it, and an alternative which
> 
> I don't think there can be any such algorithm -- in O(n) time,
> the algoritm cannot USE more than O(n) memory, so how could it
> NEED it...?  The reverse, sure, but not this way, I think.

A simple example would be an algorithm that takes as input a sequence of 
values that are all (for whatever reason) guaranteed to be integers in 
range(n**2), and tests whether any two of them are the same:

def anysame(seq):
    allocate array A, of dimension n**2, WITHOUT INITIALIZING IT
        (is this possible in pure python?)
    for i,x in enumerate(seq):
        if A[x] is an integer and 0 <= A[x] < i and A[A[x]] == x:
            return True
        A[x] = i
    return False

Initializing A would require too much time, but the algorithm is capable 
of working with an uninitialized array.  The usual trick for reducing 
the memory requirements of such an algorithm is to replace the array by 
a hash table, and of course in Python the hash table (dictionary) based 
duplicate detection algorithm is much more preferable, but this makes 
theoreticians unhappy by turning worst-case time bounds into randomized 
expected time bounds.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From staschuk at telusplanet.net  Sat Apr 26 13:15:27 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 26 Apr 2003 11:15:27 -0600
Subject: Forgetting "()" when calling methods
In-Reply-To: <oprn8ulhb1mu3s0q@news.cis.dfn.de>; from km-news1@onlinehome.de on Sat, Apr 26, 2003 at 05:57:07PM +0200
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <8CydnSQn2I0ulDejXTWcpA@comcast.com> <WRsqa.9333$K35.305964@news2.tin.it> <oprn8ulhb1mu3s0q@news.cis.dfn.de>
Message-ID: <20030426111526.A30093@tibia.amotlpaa.bogus>

Quoth Klaus Meyer:
> I also have made this mistake sometimes.
> 
> On reason, i think, is the mix of function-call and values, example:
> 
> import time
> print time.time()
> print time.timezone
> 
> You have to remember the doc of the module carefully to make no mistake.

Hm.  Yes and no.  In many cases such a mistake will make the code
fail dramatically with an informative exception.  For example, if
you forget that time.timezone is not a function and try to call
it, Python tells you exactly what you did wrong:

    >>> import time
    >>> time.timezone()
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    TypeError: 'int' object is not callable

So from here you should think: 'int' object is not callable -> I
tried to call an 'int' object -> the code called time.timezone ->
time.timezone is an 'int' object -> time.timezone is not a
function.  At that point you might go digging in the documentation
to find out exactly what it is, but if you just forgot for a
moment, this ought to remind you adequately.

Similarly, if you forget that time.time is a function, then:

    >>> time.time + 5*60*60
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    TypeError: unsupported operand types for +: 'builtin_function_or_method' and 'int'

And again: 'unsupported operand type for +' -> I added two things
which can't be added -> one is a builtin_function_or_method and
the other is an int -> 5*60*60 is an int -> time.time must be a
function -> I have to write "time.time()".

There are cases, of course, in which the problem passes silently
and just makes your program behave strangely.  I don't think
there's any good technological solution here (see below); my
feeling is that one just needs good troubleshooting skills.
Systematic investigation of what your code is and is not doing,
maybe sprinkling print and/or assert statements throughout the
code to gather data, forming and testing hypotheses about what
might be wrong...

(Aside: A lot of questions of the form "why doesn't this work?"
get answered here by programmers who are experienced enough that
the problem is obvious, and who consequently might have trouble
explaining how the querent could have figured it out themselves.
It would be nice to be able to point people at a good on-line
debugging tutorial for beginners; anybody know of such a thing?)

> In this case time() must be a function, because it has to read the actual 
> timer value, but must timezone be a value?

This makes it possible to write
    time.timezone = 5*60*60 # set local timezone to EST
for example.  (Actually I've never tried this, so I'm not sure it
works.)  Otherwise we'd have to have two functions gettimezone()
and settimezone(), which is slightly less convenient.

> My question about style:
> Should a Module/Class etc. return values, if possible, or better use 
> function to return values?

Depends on the circumstances.  If it feels like just a value then
it should be a value; if it feels like the result of an action or
computation then it should be a function's return value.

(The point of properties imho is that they let you take things
which (for implementation reasons perhaps) are actually actions
and make them feel like values for the user.  Sometimes this is
quite a win.)

> If PyChecker could check such cases this would be very helpful.

In general it's very hard (impossible, for practical purposes) to
determine ahead of time what a given name will refer to when a
given line of code is executed -- or even what type of object it
might refer to.  I don't see how pychecker could determine that
time.timezone() is wrong; it could easily be right.

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From mh at habanerotech.net  Tue Apr  1 04:33:47 2003
From: mh at habanerotech.net (Mike Hoegeman)
Date: Tue, 01 Apr 2003 09:33:47 GMT
Subject: dual porting PyQt widget contents and/or python variables into a xml dom 
 tree?
Message-ID: <3E895E58.732E29E@habanerotech.net>

i've got an application where i have lot of various PyQt widgets that
i would like to "dual port" into a xml dom tree node that i define at widget creation time.

In other words, 
	- whenever a widget changes value i'd like a node in the dom tree to 
        get updated with the new value of the widget (whatever that may be).

	-  similarly, i'd like the widget to be updated
        anytime some part of the program alters the widgets 
        associated dom node.

i could sublass every PyQt widget in the universe and put the dual porting code
where needed. I would guess this involves setting up one or more reciprocal signal/slot connections between the widget and some kind of xml dom object.  
This approach seems a bit heavy handed though..

Has Anybody else out there had the same need but found a simpler, way to accomplish this feat?

Doing something this is pretty easy in tcl/tk with variable
traces etc.. Is there something comparable to variable traces in python?

-- 
Mike Hoegeman	Habanero Technologies, LLC
http://habanerotech.net/personal-pages/mh/contact.html



From sholden at holdenweb.com  Mon Apr 14 09:58:27 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Mon, 14 Apr 2003 13:58:27 GMT
Subject: How to free a temporary clob in DCOracle2?
References: <b71t5c$d3f$1@news.Stanford.EDU>
Message-ID: <78zma.83473$It5.53885@news2.central.cox.net>

"Daniel" <dl_un-spam_rubin at yahoo.com> wrote in message
news:b71t5c$d3f$1 at news.Stanford.EDU...
> How do I free the temporary clob when I'm done with it in DCOracle2? I get
> an error when I try to close the temporary clob:
>
> >>> clob = connection.LobLocator('SQLT_CLOB')
> >>> clob.write('<f>foo bar</f>')
> 30
> >>> d='insert into foo values (XMLTYPE.createXML(:m))'
> >>> connection.execute(d, clob)
> 4
> >>> clob.close()
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> dco2.DatabaseError: (22289, 'ORA-22289: cannot perform %s operation on an
> unopened file or LOB')
>

I know nothing about this, so it's just a stab in the dark: maybe you need
to commit the transaction before you close your clob, or close your clob
before you execute your SQL?

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From Patrick.Bussi at space.alcatel.fr  Thu Apr 10 06:44:53 2003
From: Patrick.Bussi at space.alcatel.fr (Patrick.Bussi at space.alcatel.fr)
Date: Thu, 10 Apr 2003 12:44:53 +0200
Subject: =?iso-8859-1?Q?R=E9f._:_Re:_limitation_in_urllib.urlretrieve(url,?=
	=?iso-8859-1?Q?_fn)_=3F=3F?=
Message-ID: <C1256D04.003B4943.00@vzmta01.netfr.alcatel.fr>




---------------------- Envoy? par Patrick Bussi/ALCATEL-SPACE le 10/04/2003
12:44 ---------------------------
(Embedded image moved to file: pic03042.pcx)
Patrick Bussi
(Embedded image moved to file: pic16905.pcx)
10/04/2003 12:44
(Embedded image moved to file: pic11087.pcx)

Pour :    VanL <vlindberg at verio.net>
cc :
Objet :   R?f. : Re: limitation in urllib.urlretrieve(url, fn) ??

I have had similar hang up in a slightly different case using

f=urllib2.urlopen(url).read()

The read() was stuck because the server did not accept to answer to a client
which had not cookie support. I don't know if it could explain the case with
urlretrieve().

I suggest you see if urllib2 + ClientCookie modules can cover your needs (
http://wwwsearch.sourceforge.net/ClientCookie/)

Hope this helps.
Patrick
---
Patrick Bussi

Any opinions expressed are my own and not necessarily those of my Company.




VanL <vlindberg at verio.net> on 10/04/2003 01:26:17

Pour :    python-list at python.org
cc :   (ccc : Patrick Bussi/ALCATEL-SPACE)
Objet :   Re: limitation  in urllib.urlretrieve(url, fn) ??


-------------- next part --------------

There isn't a general python limitation;  I have used urllib.urlretrieve
to automate the downloading of files as large as 50 MB with
python2.2.2/freebsd 4.3.  Can't comment about Solaris, tho.

VanL

HP wrote:
> For example,
> url = 'http://www.londonstockexchange.com/techlib/word/ser/Live8102.doc';
> fn  = '/local/disk/file_name';
>
> urllib:urlretrieve(url, fn) hangs in the air !!
>
> This happens for many of the files in the above
> web site.  Those files have one thing in common:
> their file size is larger than 200K.
>
> Is there a limit on the file size in the function:
> urllib:urlretrieve(url, fn) ??
>
> (I am running python.2.2.2 on Sun Solaris 5.8).
>
> HP

--
http://mail.python.org/mailman/listinfo/python-list





-------------- next part --------------
A non-text attachment was scrubbed...
Name: pic03042.pcx
Type: application/octet-stream
Size: 1275 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030410/d4d9d4dc/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pic16905.pcx
Type: application/octet-stream
Size: 1275 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030410/d4d9d4dc/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pic11087.pcx
Type: application/octet-stream
Size: 1275 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030410/d4d9d4dc/attachment-0002.obj>

From twegener at fastmail.fm  Wed Apr 30 03:13:23 2003
From: twegener at fastmail.fm (Tim Wegener)
Date: Wed, 30 Apr 2003 16:43:23 +0930
Subject: Removing last line while appending to file
References: <20030430152608.3c51bc77.twegener@fastmail.fm>
Message-ID: <20030430164323.256ee71a.twegener@fastmail.fm>

Someone emailed me directly with the solution: use open(filename, "r+")
This solved my problem. 
(I was using the wrong mode for open())

Tim

On Wed, 30 Apr 2003 15:26:08 +0930
Tim Wegener <twegener at fastmail.fm> wrote:

> Hi, 
> 
> Is there a simple way to remove the last line (or arbitrary trailing
> portion) of a file, when appending to it? 
> 
> I want to write a log file in xml, which requires root elements around
> all the log data. This means that when I want to append to the log I
> must first remove the trailing </log> element. 
> 
> I have tried using file.seek() and file.tell() to no avail. I would
> prefer not to parse the entire file each time I want to simply append an
> entry. 
> 
> Thanks, 
> Tim
> 
> 


-- 
----------------------------------------------------------------
RADLogic Pty Ltd                Phone: +61-8-8362-3255
Suite 11, 15 Fullarton Road     Fax:   +61-8-8362-3166
Kent Town                       Email:  twegener at radlogic.com.au
South AUSTRALIA, 5067           WWW: http://www.radlogic.com.au/
----------------------------------------------------------------



From ckea25d02 at sneakemail.com  Wed Apr 16 00:18:01 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 16:18:01 +1200
Subject: do python's nifty indentation rules spell the death of   one-liners?
In-Reply-To: <b7ik4p$1d2t1$1@ID-169208.news.dfncis.de>
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org> <b7ik4p$1d2t1$1@ID-169208.news.dfncis.de>
Message-ID: <b7iksv$1ba93$1@ID-169208.news.dfncis.de>

I wrote:
> I just tried this in a Makefile, and it seems to work:
> 
> all:
>     python -c '\
>     for i in range(10):\
>      print i'

Oops, looks like I spoke too soon, because this
*doesn't* work:

all:
	python -c '\
	print "Hi"\
	for i in range(10):\
	  print i'

It seems that make puts the whole command on one line,
no matter how you lay it out in the Makefile. Bugger.

But do you really have to put the actual Python in
the Makefile? You could have a file full of Python
functions, one for each snippet, and call them
using things like:

   python -c 'import stuff; stuff.spam()'

I just tested that in the Makefile, and it
*does* work!

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From nikolai.NOJUNK at micon.no  Fri Apr  4 17:17:52 2003
From: nikolai.NOJUNK at micon.no (Nikolai Kirsebom)
Date: Sat, 05 Apr 2003 00:17:52 +0200
Subject: Invalid format character in string
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com> <mailman.1049493668.26508.python-list@python.org>
Message-ID: <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>

Hi,
The problems arise where the HTML editor used (to produce the file in
the first place) has put various expressions of %; and %>.  These are
HTML codings the creator of the HTML file does not really know about.
As an example, I've just told them (the users creating the HTML file)
that they should write the text "%(name)s" where they want the name
generated and "%(age)s" where they want the age generated.  Then my
script is run for a set of person objects producing a set of HTML
files with the correct names and corresponding ages. 

On Fri, 4 Apr 2003 15:59:48 -0600, Skip Montanaro <skip at pobox.com>
wrote:

>
>    Nikolai> ... since the text is actually a HTML file, it may contain a
>    Nikolai> lot of characters with % (typically cell width definitions of
>    Nikolai> tables).  If these are contained in the file, the string format
>    Nikolai> statement complains about illegal format character.  
>
>Just double any literal % characters, e.g.:
>
>    <table width="%(width)s%%">
>
>Skip




From Norbert.Klamann at klamann-software.de  Wed Apr 16 07:50:35 2003
From: Norbert.Klamann at klamann-software.de (Norbert Klamann)
Date: 16 Apr 2003 04:50:35 -0700
Subject: AOP in Python (was: Post-OOP: not just an "unfortunate acronym"?)
References: <pan.2003.04.15.21.27.33.796814@mn.rr.com> <slrnb9p5bk.vg8.ramen@lackingtalent.com>
Message-ID: <65c27c6b.0304160350.c0cbc67@posting.google.com>

Dave Benjamin <ramen at lackingtalent.com> wrote in message news:<slrnb9p5bk.vg8.ramen at lackingtalent.com>...
[snip]
> Does anyone have an example of this newfangled "Aspect Oriented" paradigm's
> techniques implemented in Python? 

[snip]

Look for transwarp or PEAK as it is called nowadays .

HTH

Norbert



From bignose-hates-spam at and-zip-does-too.com.au  Tue Apr  8 00:52:13 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Tue, 08 Apr 2003 04:52:13 GMT
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <pan.2003.04.07.23.02.00.101458@jerf.org>
Message-ID: <slrnb94llc.p4.bignose-hates-spam@rose.localdomain.fake>

On Tue, 08 Apr 2003 04:02:07 GMT, Jeremy Bowers wrote:
> It is interesting that all the Python programs I've seen that are
> broken up into files at all (and not just 100+KB hunks of text) are
> plopped into one directory.  

What are your thoughts on Mailman?  It, at least, is a complex
application broken into subdirectories by functional area.

Sadly it has executables that use an 'import paths' hack to get around
this irritating Python limitation.  Fine for executables that are placed
in the path, but lousy for within-the-module unit tests.

> I'd PEP it but given that nobody else is uncomfortable polluting this
> part of the namespace it seems the odds of it going anywhere are slim.

Given the responses I've received in this newsgroup and elsewhere, I'd
say you are certainly not alone with being uncomfortable with namespace
pollution as an 'import' hack.  Please feel free to PEP it.

> In particular, I have found this helpful in putting all the unit tests
> in separate directories named "tests" under the things they are
> actually testing

Which makes a mockery of the elegance of 'if __name__ == "__main__"' for
unit testing.

-- 
 \       "[On the Internet,] power and control will shift to those who |
  `\       are actually contributing something useful rather than just |
_o__)                                 having lunch."  -- Douglas Adams |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From steve at ferg.org  Tue Apr  8 18:31:25 2003
From: steve at ferg.org (Stephen Ferg)
Date: 8 Apr 2003 15:31:25 -0700
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
Message-ID: <b16e4ef7.0304081431.50c78595@posting.google.com>

def dq(s):
	"""enclose a string in double quotes
	"""
	return '"' + s + '"'

s = "abc"
sInDoubleQuotes = dq(s)



From koenig at v-i-t.de  Mon Apr 21 08:44:43 2003
From: koenig at v-i-t.de (Mirko Koenig)
Date: Mon, 21 Apr 2003 14:44:43 +0200
Subject: importing modules with the same name but from different directories
References: <b7vhad$1t6$1@news.tu-darmstadt.de> <3EA34ED6.F0361E7D@engcorp.com>
Message-ID: <b80p3r$7sp$1@news.tu-darmstadt.de>

On Mon, 21 Apr 2003 03:52:22 +0200, Peter Hansen wrote:

> Mirko Koenig wrote:
>> 
>> In every modul directory is a diretory for the language files. These
>> files are used to display the modules label in different languages. eg.
>> 
>> .../modules/module1/language/german.py
>> .../modules/module1/language/english.py
>> .../modules/module2/language/german.py
>> .../modules/module2/language/english.py
>> 
>> My problem is:
>> If is start modul1 from within its modul directory ( .../modules/modul1
>> ) And modul1 calls modul2, then modul2 doesn't load its own language
>> file. Instead it loads the langugae file of modul1.
> 
> Although it's not 100% clear to me what you are trying to do, I'd like
> to suggest a couple of alternatives to what you seem to be trying, which
> might be simpler.
> 
> Option 1 is to just use the execfile() builtin function to execute the
> necessary files, rather than trying to use "import".
> 
> Option 2 is to rename the files to something other than .py and load
> them as regular files, then decode them with exec().

The file are named with .py because there is a dict in them, where the
text is stored, eg.
text = { 1: "hello", 2:"how are you" }

I load this dict from the needed file and have the language i want.

Because the modules are python too, i thought i don't have to use exec.
Instead i can import the module and start it from within the other modul.
Is this worse than using exec/execfile?

Mirko Koenig



From mwm at mired.org  Mon Apr  7 10:53:15 2003
From: mwm at mired.org (Mike Meyer)
Date: Mon, 07 Apr 2003 14:53:15 GMT
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain> <x7smsvh7kr.fsf@guru.mired.org> <w5Wk+ks/KT/A089yn@the-wire.com>
Message-ID: <x77ka69x5v.fsf@guru.mired.org>

mwilson at the-wire.com (Mel Wilson) writes:

> In article <x7smsvh7kr.fsf at guru.mired.org>, Mike Meyer <mwm at mired.org> wrote:
> >The garbage collection happening when your function returns is
> >implementation specific. It works that way in CPython, but not
> >Jython. You should always explicitly close files to make sure they get
> >closed - and the output flushed, in this case - when you're done with
> >them.
>    That's why I prefer to think of Jython as a different
> language.  The idea of a service that can clean up and quit
> when the last client stops using it is just too appealing.

I don't think you have any guarantees that CPython won't change
behavior on you. 

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From martin at v.loewis.de  Thu Apr 24 16:03:19 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 24 Apr 2003 22:03:19 +0200
Subject: True/False
References: <mailman.1051036237.30103.python-list@python.org> <3EA5B173.9420324C@alcyone.com> <mailman.1051056037.8077.python-list@python.org> <3EA5DE1E.22704437@alcyone.com> <m3r87tdbo3.fsf@mira.informatik.hu-berlin.de> <mailman.1051081416.21801.python-list@python.org> <m34r4pgh33.fsf@mira.informatik.hu-berlin.de> <Xns93676070D39E9duncanrcpcouk@127.0.0.1>
Message-ID: <m3znmfek8o.fsf@mira.informatik.hu-berlin.de>

Duncan Booth <duncan at NOSPAMrcp.co.uk> writes:

> > For Python 2.x, this does not really matter much. Yes, you do get
> > references to Py_True/Py_False, but they are still just integer
> > objects with the value 1/0.
> 
> Not entirely accurate. In Python 2.3, True and False have type bool which 
> is a subclass of int, so this could make a difference.

Right; I meant x < 3.

> Even in Python 2.1 & 2.2 there can be a subtle difference. If, for example, 
> you use the win32 extensions to access COM objects, then calling a COM 
> method that accepts a variant type with 1 as a parameter will pass an 
> integer, but 1==1 will make it pass a boolean. The win32 com support did 
> some strange tricks with object identity to tell integers and booleans 
> apart.

I see. Sounds very hacky.

Regards,
Martin



From mwh at python.net  Fri Apr  4 06:06:37 2003
From: mwh at python.net (Michael Hudson)
Date: Fri, 4 Apr 2003 11:06:37 GMT
Subject: Severe c-module problem
References: <b6jo6v$715$1@nets3.rz.RWTH-Aachen.DE>
Message-ID: <7h3he9ejzcz.fsf@pc150.maths.bris.ac.uk>

Rolf Wester <wester at ilt.fhg.de> writes:

> Hi,
> 
> I have a c-Module "pyopt" with a function:
> 
> static PyObject *_wrap_c_nlo_init_Hs(PyObject *self, PyObject *args) {
>      PyObject *resultobj;
>      char * arg0;
> 
>      cout << "in init_Hs" << endl;
> 
>      if(!PyArg_ParseTuple(args,(char *)"ss:c_nlo_init_Hs", &arg0))
                                          ^^                 ^^^^^
                                  specifies two arguments      |
                                                               |
                                         you only pass a pointer for one

>              return NULL;
> 
>      cout << "in init_Hs: " << arg0 << endl;
> 
>      c_nlo_init_Hs(arg0);
>      Py_INCREF(Py_None);
>      resultobj = Py_None;
>      return resultobj;
> }
> 

HTH,
M.

-- 
  Sufficiently advanced political correctness is indistinguishable
  from irony.                                           -- Erik Naggum



From mertz at gnosis.cx  Fri Apr 11 18:46:09 2003
From: mertz at gnosis.cx (David Mertz)
Date: Fri, 11 Apr 2003 18:46:09 -0400
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
 <d116fbae.0304100704.68ef4833@posting.google.com>
 <c48$3e960655$516049d2$3498@nf1.news-service.com>
 <3e96efac$0$49107$e4fe514c@news.xs4all.nl>
 <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk>
 <3e96f481$0$49099$e4fe514c@news.xs4all.nl>
 <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>
 <3e970abc$0$49101$e4fe514c@news.xs4all.nl>
 <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>
 <3e972e43$0$49109$e4fe514c@news.xs4all.nl>
Message-ID: <xW0l+kKkXIiA092yn@gnosis.cx>                       

As much as I like Pyro, it doesn't quite fit into the roundup Mike and
Uche wrote.  Or at least not entirely. Raw sockets, CORBA, XML-RPC, and
SOAP are all *language independent*, while Pyro is Python-specific.  For
many projects, accessing the same services in different languages is not
an issue, but in some projects it matters.  Unfortunately, if that
specification matters, Pyro is ruled out (as is, for example, Java RMI).

This isn't to say that Irmen's time and size data isn't interesting and
worth knowing.  But just to point out that the article wasn't simply
snubbing Pyro for no reason.

Yours, David...

--
 mertz@   _/_/_/_/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY:_/_/_/_/ v i
gnosis  _/_/                    Postmodern Enterprises         _/_/  s r
.cx    _/_/  MAKERS OF CHAOS....                              _/_/   i u
      _/_/_/_/_/ LOOK FOR IT IN A NEIGHBORHOOD NEAR YOU_/_/_/_/_/    g s





From grante at visi.com  Thu Apr 17 11:19:44 2003
From: grante at visi.com (Grant Edwards)
Date: 17 Apr 2003 15:19:44 GMT
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <b7kd74$jqm$1@panix2.panix.com> <hr9s9v8pgqr154h15lsjbk1pjs5u5n6n65@4ax.com> <Inyna.27603$LB6.635831@news1.tin.it> <b7merr$mmn$1@panix2.panix.com>
Message-ID: <3e9ec610$0$178$a1866201@newsreader.visi.com>

In article <b7merr$mmn$1 at panix2.panix.com>, Aahz wrote:

> That's not the main reason, though.  The main reason is that
> Guido believes that control structures should stand out
> clearly.

Sure.  Except the OP was talking about a Makefile.  Makefiles
aren't supposed to be readable.  :)

-- 
Grant Edwards                   grante             Yow!  Gee, I feel kind of
                                  at               LIGHT in the head now,
                               visi.com            knowing I can't make my
                                                   satellite dish PAYMENTS!



From aussie2010 at yahoo.com  Tue Apr  1 17:57:13 2003
From: aussie2010 at yahoo.com (Chris)
Date: 1 Apr 2003 14:57:13 -0800
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <3E8974E5.88B861D7@engcorp.com>
Message-ID: <435d40e0.0304011457.56e32bd0@posting.google.com>

Peter Hansen <peter at engcorp.com> wrote in message news:<3E8974E5.88B861D7 at engcorp.com>...
> Chris wrote:
> > 
> > I would like to request that Python support multi-line 
> > comments (e.g. /* */ in C).
> 
> There are at least two reasons why someone would want to do
> that, and of course it's been suggested before, but what's
> your rationale behind this?  There are generally simple 
> answers that address each possible use case, but perhaps you've
> actually got an original idea about this...
> 
> -Peter

What are the reasons for not having multi-line comments? 

I just remembered that IDLE will color-code """ as a string and not a
comment - even when it is intended only as a comment.



From xnedra at aa-ha.com  Wed Apr 23 10:05:33 2003
From: xnedra at aa-ha.com (Xnedra)
Date: 23 Apr 2003 07:05:33 -0700
Subject: Setting the Title of the script window
References: <1814331d.0304230219.3c6a82de@posting.google.com> <akupa.83076$T34.2073301@news2.tin.it>
Message-ID: <1814331d.0304230605.6a5a4f3@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote: 
> Unfortunately (in this context), "unix" does not have a single way
> to implement "text windows" -- there are many such beasts!
> 
> Fortunately, most such beasts emulate to some extent the oldest of
> them all, 'xterm'.  To set the window title under xterm, you must
> print to the terminal a string with some escape sequences --
> 
> def set_xterm_title(title):
>     print '\033]0;%s\007' % title
> 
> This will also work on a KDE "Konsole" and no doubt several other
> terminal emulators -- but, I think, not ALL of them.

Thanks for your answer....i'll give it a try :)

Katrin



From fredrik at pythonware.com  Tue Apr 29 07:04:14 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Tue, 29 Apr 2003 13:04:14 +0200
Subject: Passing Named Paramater To A C Func?
References: <3EAE57E5.5010202@pa.press.net>
Message-ID: <b8lm5d$4tl$1@main.gmane.org>

John Abel wrote:

> I'm writing a C function, which will have a Python wrapper, and I wish
> to be able to pass it an optional, named parameter.  For example:
>
> def funcName( dirName, fileName=None ):
>
>    if fileName is None:
>        Some Other Stuff
>
>    cCall( dirName, fileName )
>
> Is the best to handle this in the Python wrapper, or is there a way of
> handling these via C.  None of the extending documents seem to mention
> anything about this.

to handle named arguments on the C level, use ParseTupleAndKeywords.
here's an example, taken from the regular expression engine:

static PyObject*
pattern_match(PatternObject* self, PyObject* args, PyObject* kw)
{
    ...

    PyObject* string;
    int start = 0;
    int end = INT_MAX;
    static char* kwlist[] = { "pattern", "pos", "endpos", NULL };
    if (!PyArg_ParseTupleAndKeywords(args, kw, "O|ii:match", kwlist,
                                     &string, &start, &end))
        return NULL;

you also have to set the METH_KEYWORDS flag in the PyMethodDef
table:

static PyMethodDef pattern_methods[] = {
    {"match", (PyCFunction) pattern_match, METH_VARARGS|METH_KEYWORDS},
    ...

:::

in your case, I'm not sure it's really worth the effort, since you
need to look at the argument at the Python level anyway.

</F>







From nirina at mail.blueline.mg  Fri Apr 11 06:28:33 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Fri, 11 Apr 2003 14:28:33 +0400
Subject: string() - Is this some kind of sick joke?
References: <36dbf824.0304091142.527af834@posting.google.com> <1dpc9v4beon9upq2sd3clfb2plh7nlqm7p@4ax.com>
Message-ID: <008e01c30015$669a3040$fba2383e@raseliar>

Python22
>>> y="'hello'"
>>> help(y.strip)
Help on built-in function strip:
strip(...)
    S.strip() -> string
    
    Return a copy of the string S with leading and trailing
    whitespace removed.


Python23
>>> y="'hello'"
>>> help(y.strip)
Help on built-in function strip:
strip(...)
    S.strip([sep]) -> string or unicode
    
    Return a copy of the string S with leading and trailing
    whitespace removed.
    If sep is given and not None, remove characters in sep instead.
    If sep is unicode, S will be converted to unicode before stripping


So I think you do have Python2.3 to use full feature of strip

nirinA







From a.schmolck at gmx.net  Thu Apr 10 08:55:22 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 10 Apr 2003 13:55:22 +0100
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <yfs7ka3iom2.fsf@black132.ex.ac.uk> <b1ab591.0304100042.1188e4bf@posting.google.com>
Message-ID: <yfsn0iyh5qd.fsf@black132.ex.ac.uk>

s.thuriez at laposte.net (sebastien) writes:

> Alexander Schmolck <a.schmolck at gmx.net> wrote in message news:<yfs7ka3iom2.fsf at black132.ex.ac.uk>...
> Thanks for the program. I have tested it with a test matrix that I
> have and that I took from IDL help (Interactive data language).
> 
> The matrix is :
> 
> ((2,1,3),
> (4,2,3),
> (4,1,0),
> (2,3,3),
> (5,1,9))
> 
> 
> The principal component given by IDL is 
> 
> ((0.87, -0.7, 0.69),
> (0.01, -0.64, -0.66),
> (0.49, 0.32, -0.30))
> 
> Runnnig the same matrix with the program that you have kindly provided
> gives :
> 
> ((-0.58, -0.67, 0.45),
> (-0.23, -0.40, -0.88),
> (-0.77, -0.62, -0.074))
> 
> Do you know why is there such a difference ??

I'm not sure what you computed with IDL but it sure doesn't look like
principal components of anything to me (the result matrix you quote is not
even orthogonal).

Matlab's svd (with [U,S,V] = svd(your_matrix)) gives pretty much the same V as
what I posted (you didn't by any chance just truncate (rather than round) the
values gave above?):

  -0.5812   -0.6728    0.4577
  -0.2346   -0.4001   -0.8859
  -0.7792    0.6223   -0.0747


'as




From ckea25d02 at sneakemail.com  Tue Apr  1 19:58:05 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 12:58:05 +1200
Subject: Reminiscencationizings (Re: Questions on linked lists)
In-Reply-To: <x73cl2jezh.fsf@guru.mired.org>
References: <kl8ia.27315$OV.127379@rwcrnsc54> <5vonl-vm3.ln1@beastie.ix.netcom.com> <x73cl2jezh.fsf@guru.mired.org>
Message-ID: <b6dc3f$4a3qv$1@ID-169208.news.dfncis.de>

Mike Meyer wrote:
> Geez, this is odd memory day. I once had an instructor say we could
> write an assembler in "any structured language"... I wrote a SNOBOL preprocessor 
> - in SNOBOL, of course

Oh, wow. Mention of SNOBOL brings back fond memories of
submitting SNOBOL code on punched cards to a batch
system running on a Burroughs B6900...

The computer centre meticulously charged for seconds
of CPU time used. Being an interpreter, SNOBOL programs
were rather expensive to run. I think it cost me about
50 cents per run, and we were given a budget of about
$2 each for the assignment, so we didn't get many goes
to get it right!

The lecturer warned that we would lose marks if the
program crashed. On the last run, mine produced the
correct output and then crashed right at the end. So,
I cut the error messages off the bottom of the listing
and handed it in. I don't think he ever suspected,
although I wonder whether he noticed that the last
page was a bit shorter than the rest...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From aleax at aleax.it  Tue Apr 15 03:17:23 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 07:17:23 GMT
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it> <ouadesfkn2.fsf@andlx-anamika.cisco.com>
Message-ID: <7mOma.20244$T34.649927@news2.tin.it>

Ganesan R wrote:

>>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:
> 
>> Note that any Python function call always passes VALUES -- i.e.
>> there is no difference between your call to sys.getrefcount in
>> the above snippet and one that directly does sys.getrefcount(1).
>> In either case, it's the VALUE (1) that you're asking info about;
>> it makes no difference HOW you obtain the value to pass (I guess
>> this is probably already clear to you, but, just in case...).
> 
> Then, I find this confusing
> 
> 
===========================================================================
> Python 2.2.1 (#1, Jul 29 2002, 23:15:49)
> [GCC 2.95.4 20011002 (Debian prerelease)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import gc
>>>> a = "hello world"
>>>> gc.get_referrers(a)
> [{'__builtins__': <module '__builtin__' (built-in)>, '__name__':
> [{'__main__', 'gc': <module 'gc' (built-in)>, '__doc__': None, 'a': 'hello
> [{world'}]
>>>> b = "hello world"
>>>> gc.get_referrers(a)
> [{'a': 'hello world', 'b': 'hello world', 'gc': <module 'gc' (built-in)>,
> [{'__builtins__': <module '__builtin__' (built-in)>, '__name__':
> [{'__main__', '__doc__': None}]
>>>> gc.get_referrers("hello world")
> [('hello world', None)]
>>>> 
> 
===========================================================================
> 
> Why am I getting a different list for the last call?

Because you're asking about a distinct (even though equal) value,
AKA object.  Python MAY but doesn't HAVE TO reuse the same object
when you use several literals indicating equal immutables -- in
practice, currently, it does for small integers, but (in Python
2.2) not for such strings as you used here.  Python can always
change this in future releases, since this is strictly an issue of
implementation -- optimizing one way or another, no more than that.

See:

[alex at lancelot ox]$ python2.2
Python 2.2.2 (#1, Oct 24 2002, 11:43:01)
[GCC 2.96 20000731 (Mandrake Linux 8.2 2.96-0.76mdk)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> a='hello world'
>>> b='hello world'
>>> id(a), id(b), id('hello world')
(135683376, 135688960, 135689328)
>>>

[alex at lancelot ox]$ python2.3
Python 2.3a2+ (#11, Mar 28 2003, 12:17:11)
[GCC 3.2 (Mandrake Linux 9.0 3.2-1mdk)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> a='hello world'
>>> b='hello world'
>>> id(a), id(b), id('hello world')
(1076642784, 1076642864, 1076643144)
>>>

In 2.2, we see Python using different copies of the literal string
value in all three cases.  In 2.3, we see indications of a different
strategy -- one value (object, identity, copy, call it as you wish)
is being used for the first two occurrences, a distinct one for the
third occurrence.  Strictly issues of optimization strategies on
the part of different Python implementations, no more.


Keep in mind that EQUALITY and IDENTITY are different concepts.  For
immutables, equality is what you generally care about -- identity is
shifty and implementation-dependent since such optimizations are
permissible.  Thus, you shouldn't test "a is b" nor "a is 'hello world'",
but rather you should use the == (equality check) operator in lieu
of the is (identity check) operator for such tests.

References are to specific values (specific objects, specific identities,
specific copies, call them as you wish), thus all considerations about
reference tracking need to keep this concept in mind.


One last indication (not playing a role here, but you're sure to run
into this as you play with id and the like...): id(...) is unique
as long as an object exists, but as soon as the object is destroyed,
its identity may be reused.  So, don't get tricked by such artefacts
as the following:

[alex at lancelot ox]$ python2.2
Python 2.2.2 (#1, Oct 24 2002, 11:43:01)
[GCC 2.96 20000731 (Mandrake Linux 8.2 2.96-0.76mdk)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> id('hello world')
135683352
>>> id('hello world')
135682152
>>> id('hello world')
135683192
>>> id('hello world')
135683352
>>> id('hello world')
135682736
>>>

it may SEEM like the first and fourth literals use the same value, distinct
from the 2nd and 3rd and 5th -- but they don't -- the first one has been
destroyed, and happens to get recycled, at the time you create the 4th (in
other builds you'll no doubt get different results for this -- in 2.2, the
allocator used is more system-dependent than in 2.3).


Morals: don't care too deeply about identity of immutables...


Alex




From mikael at isy.liu.se  Wed Apr  2 02:45:15 2003
From: mikael at isy.liu.se (Mikael Olofsson)
Date: Wed, 2 Apr 2003 09:45:15 +0200
Subject: Precidence of unary plus/minus relative to exponentiation
In-Reply-To: <D7A3CFD7825BD6119B880002A58F06C202F2C760@groexmb02.pfizer.com>
References: <D7A3CFD7825BD6119B880002A58F06C202F2C760@groexmb02.pfizer.com>
Message-ID: <20030402094515.3f865867.mikael@isy.liu.se>

On Tue, 1 Apr 2003 12:40:21 -0500
"Warnes, Gregory R" <gregory_r_warnes at groton.pfizer.com> wrote:
> The current operator precedence places unary plus and minus higher
> than exponentiation.  This leads the counterintuitive result:
> 
> 	>>> -2.0**2.0
> 	-4.0
> 	
> Python's operator predicence is causing this to be interpreted as:
> 
> 	>>> -(2.0**2.0)
> 
> rather than the normal mathematical precidence, which would give
> 
> 	>>> (-2.0)**2.0
> 
> Is there a good reason for the Python order of precedence?

I would say that the behaviour is what I expect. I would never expect
-2.0**2.0 to mean (-2.0)**2.0. Nobody I've met in a mathematical context
(engineers and mathematicians) have ever assumed your interpretation.
Where have you seen that interpretation?

/Mikael

-----------------------------------------------------------------------
E-Mail:  mikael at isy.liu.se
WWW:     http://www.dtr.isy.liu.se/dtr/staff/mikael               
Phone:   +46 - (0)13 - 28 1343
Telefax: +46 - (0)13 - 28 1339

         /"\
         \ /     ASCII Ribbon Campaign
          X      Against HTML Mail
         / \

This message was sent by Sylpheed.
-----------------------------------------------------------------------
Link?pings kammark?r: www.kammarkoren.com





From BPettersen at NAREX.com  Mon Apr 21 11:59:18 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Mon, 21 Apr 2003 09:59:18 -0600
Subject: __setitem__ puzzle
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE02A@admin56.narex.com>

> From: Anton Vredegoor [mailto:anton at vredegoor.doge.nl] 
> 
> I have subclassed the builtin list type to accept tuples for indexing.
> I am just experimenting with the code to see what I can do with it, so
> asking for what the code will be used for in order to provide a better
> way to do it will be answered with 'ni' :-)
> 
> There's a problem with __setitem__ however. What's the proper way to
> write this class?
> 
> Anton
> 
> class tlist(list):
>     
>     def __getitem__(self,x):
>         y = self[:]
>         for z in x:
>             y = y[z]
>         return y
> 
>     def __setitem__(self,x,val):
>         y = self[:]
          root = y         # save 1st level mutable items
>         for z in x[:-1]:
>             y = y[z]
>         y[x[-1]] = val
          self[:] = root   # update 1st level mutable items

[...]
>     L[(1,1,0)] = 0

Is equivalent to L[1,1,1], and you can also do fun things like L[1:2,
4:9]. Advice: special case one tuples (L[n]).

-- bjorn




From maxk at NtOoSuPrAoM.edu  Fri Apr  4 00:16:13 2003
From: maxk at NtOoSuPrAoM.edu (Maxim Khesin)
Date: Fri, 04 Apr 2003 05:16:13 GMT
Subject: about NLTK in Python
References: <97e89849.0304032019.28a612ad@posting.google.com>
Message-ID: <xy8ja.4470$n95.1427485@twister.nyc.rr.com>

> >>> from nltk.token import *
> Traceback (most recent call last):
>   File "<interactive input>", line 1, in ?
> ImportError: No module named nltk.token
> >>>
> C:\\Python22\\nltk1.0
add the
C:\\Python22\\nltk1.0
to a .pth file in C:\\Python22\\
folder





From mcfletch at rogers.com  Wed Apr  2 10:38:37 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Wed, 02 Apr 2003 10:38:37 -0500
Subject: Best practice for object attributes?
In-Reply-To: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk>
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk>
Message-ID: <3E8B03FD.70904@rogers.com>

Just... can't... resist... question... on... modelling... domain... 
objects...

class Shrubbery1( propertied.Propertied ):
    """Representation of a Shrub, surprisingly"""
    name = common.StringProperty(
       "name", """The name of the intrepid gardener""",
       defaultValue = "",
    )
    flowering = common.BooleanProperty(
       "flowering", """Whether this is carniverous flowering Blood Root""",
       defaultValue = 0,
    )
    knightOwner = basic.BasicProperty(
       "knightOwner", """The knight which own the shrub's mortgage""",
       baseType = Knight,
       # stupid example, just asks the Knight type for default instance
       # probably looking up who owns the plot of land on which we sit
       defaultFunction = lambda prop, client: 
Knight.getDefaultKnight(client),
       setDefaultOnGet = 0, # until explicitly assigned, will always 
lookup the default
    )

Wendal = Shrubbery1( name = "Wendal", knightOwner = SirHenry )
Grace = Shrubbery1( name = "Grace" )
# Grace's knightOwner is now getDefaultKnight(Grace) at the time you ask 
for it

The advantages of this way being:

    * GUIs can introspect the property to figure out all sorts of
      wonderful things, such as how to coerce values to the correct
      types, or what type of "selection" widgets to show for "knightOwner"
    * The properties show up in your standard pydoc documentation
      (though admitedly, only the __doc__ values are useful in that case)
    * You can define functions for retrieving the defaults, so that
      run-time lookups occur to figure out who is the current default despot

Of course, then you need the basicproperty package, but doesn't 
_everybody_ want that? ;)

See also:

    * Chaco's Traits library (including delegation)
    * Zope 3's Field objects (deep in the bowels of Zope, (with cool
      little "bootstrap fields" describing them))
    * OpenGLContext's vrml fields library (sub-domain specific (i.e. 3D
      graphics))

(I intuit the existence of a few others (A.B.Strakt's, the Chandler 
people's), but they aren't available yet AFAIK)

Enjoy,
Mike  

Michael Sparks wrote:

>Hello,
>
>
>If I want to model objects with 3 attributes based on the behaviour of
>python 2.2/2.3 I'ved noticed I can use either of the following idioms:
>
>class shrubbery1(object):
>   name = None
>   knightowner = "ni"
>   flowering = False
>
>or:
>class shrubbery2(object):
>   def __init__(self):
>      self.name = None
>      self.knightowner = "ni"
>      self.flowering = False
>
>In both cases I gain objects which have the same 3 attributes.
>Differences between
>them I find are:
>   * shrubbery1 has 3 class attributes as well, with the initial values
>     indicated. This isn't the case with shrubbery2.
>
>     >>> [x for x in dir(shrubbery1) if not x[0] == "_"]
>     ['flowering', 'knightowner', 'name']
>
>     >>> [x for x in dir(shrubbery2) if not x[0] == "_"]
>     []
>
>   * shrubbery2 has 3 entries added to self.__dict__ which act as the
>     attribute stores. This is not the case with shrubbery1.
>
>     >>> shrubbery1().__dict__
>     {}
>     >>> shrubbery2().__dict__
>     {'flowering': False, 'name': None, 'knightowner': 'ni'}
>
>
>Personally I find the former (shrubbery1) clearer, but I'm posting to
>find out what people think is generally the better of the two. Do people
>find any inherent advantages to one versus the other? (I note that the
>shrubbery1 also makes the names of object attributes available to
>metaclasses as well)
>
>Thanks for any comments,
>
>
>Michael.
>  
>
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From exarkun at intarweb.us  Wed Apr 30 11:49:49 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 30 Apr 2003 11:49:49 -0400
Subject: fcntl in windows
In-Reply-To: <k2rvav06nhjgtr578fa5ostftfl69rethp@4ax.com>
References: <k2rvav06nhjgtr578fa5ostftfl69rethp@4ax.com>
Message-ID: <20030430154949.GC26204@meson.dyndns.org>

On Wed, Apr 30, 2003 at 03:26:29PM +0000, Afanasiy wrote:
> Is there an fcntl in Windows? A stub, a wrapper, etc?

  http://python.org/doc/lib/module-fcntl.html

  Jp

-- 
Seduced, shaggy Samson snored.
She scissored short.  Sorely shorn,
Soon shackled slave, Samson sighed,
Silently scheming,
Sightlessly seeking
Some savage, spectacular suicide.
                -- Stanislaw Lem, "Cyberiad"
-- 
 up 41 days, 11:04, 5 users, load average: 0.14, 0.11, 0.03




From xxdigitalhellxx at hotmail.com  Wed Apr 16 16:56:36 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Wed, 16 Apr 2003 14:56:36 -0600
Subject: Big-O notation
In-Reply-To: <a8b7f07a.0304161158.29c0ef6b@posting.google.com>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com> <a8b7f07a.0304161158.29c0ef6b@posting.google.com>
Message-ID: <b7kg3r$pvt$1@heckle.cc.uregina.ca>

Peter van Kampen <news at datatailors.xs4all.nl> wrote in message 
news:<slrnb9q90a.mhg.news at datatailors.com>...
>I think I almost 'get it'. Except who or what decides what Big-O a
>certain algorithm has. Is that by 'hunch' or by running benchmarks or
>are there still other ways to recognize 'Big-O's (except for the
>formal computations that I understand are not very common)? 

I'll try to give a simple explanation. Big-O is the upper bound of the time 
it will take to complete the function.

There's usually a variable factor in determining how long an operation will 
take.

Inserting into a hash table:
O(1) - constant time. This means that the amount of time it takes to insert 
into a hash table is not dependent on the size of the table. This is because 
the position to insert the item is calculated (the hash) and then inserted. 
This, of course, changes when there starts being collisions.

Inserting into the first free spot of an array:
O(n) - linear time. n is the size of the array. This means when you cycle 
through the array you have to go through n items (at the most).

And it carries on. I forget exactly how to calculate them, but you can 
understand the simple ones.

Here's some really simple code to show different Big-O's

n = 24

# O(1)
# constant time
print n

# O(n)
# linear time
for x in range(0,n):
     print x

# O(n^2)
# quadratic time
for x in range(0,n):
     for y in range(0,n):
         print x, y

# O(x^n)
# exponential time
for y in range(0, x**n):
     print y

Keep in mind that Big-O is the *upper bound*. These all illustrate the 
maximum time. Any real code is most likely going to reach it's "destination" 
before it hits the end (though not always).

-- 
Hope this didn't make you more confused than ever,
Ryan




From mwh at python.net  Wed Apr 16 16:13:16 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 16 Apr 2003 20:13:16 GMT
Subject: How many of you are Extreme Programmers?
References: <1050508138.3e9d7b6a1d520@mcherm.com> <mailman.1050509219.31421.python-list@python.org>
Message-ID: <7h37k9uqjz8.fsf@pc150.maths.bris.ac.uk>

Christopher Blunck <blunck at gst.com> writes:

> It's quite funny, but if you sit an XP and RUP guy down at the same table
> you'll find that they agree with most principals.  They'll disagree on some
> things (like design), but I've never seen them disagree on iterative 
> development, frequent testing, frequent integration, etc.  They may disagree
> on the frequency of "frequent testing" (XP favoring daily testing)

*Daily*?  When I'm doing TDD (which isn't as often as I should) I
usually run the tests many, many times an hour...

Cheers,
M.

-- 
  Ya, ya, ya, except ... if I were built out of KSR chips, I'd
  be running at 25 or 50 MHz, and would be wrong about ALMOST
  EVERYTHING almost ALL THE TIME just due to being a computer!
                                              -- Tim Peters, 30 Apr 97



From do-not-spam-ben.hutchings at businesswebsoftware.com  Mon Apr 28 14:35:31 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 28 Apr 2003 18:35:31 GMT
Subject: A new and very robust method for doing file locking over NFS?
References: <lcu1cw525a.fsf@gaffa.mit.edu> <slrnbafu8h.ek.do-not-spam-ben.hutchings@tin.bwsint.com> <lck7dix0iv.fsf@gaffa.mit.edu>
Message-ID: <slrnbaqt3j.ao.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <lck7dix0iv.fsf at gaffa.mit.edu>, Douglas Alan wrote:
> Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> writes:
> 
>> The usual solution for creating a lock file in an NFS-safe way is:
> 
> Yes, this is the method described in the Red Hat man page for the
> open() system call, which I mentioned, but it doesn't solve the
> unlocking problem, which remains, I believe, as I described.

You're quite right.  There probably isn't a solution.  Welcome to NFS.



From dreed at capital.edu  Sun Apr 20 00:05:43 2003
From: dreed at capital.edu (Dave Reed)
Date: Sun, 20 Apr 2003 00:05:43 -0400
Subject: red hat 8.0 and python
In-Reply-To: <3EA21326.7070100@sympatico.ca>
References: <3EA21326.7070100@sympatico.ca>
Message-ID: <200304200005.43862.dreed@capital.edu>

On Saturday 19 April 2003 23:25, Peter wrote:
> I've just installed Red hat 8.0, and although python 2.2.1 comes with 
> it, i cannot use anything which uses the Tkinter library (this 
includes 
> IDLE). I a few other packages work, and the interpreter runs fine. I 
> suppose that this might be due to the setup, but I cannot figure it 
out. 
>   Could it be that Tkinter module does not come with the python 
> distribution which came with Red Hat 8.0? I read that it should, but I 
> searched my whole computer for anything with ?tk? in the name with no 
luck.


It's on one of the CDs - install the tkinter-2.2.1-17 rpm file.


> I downloaded the rpm's for the python 2.2.2 (from www.python.org) but 
> simple "rpm -i" or "rpm -U" installing gives me errors; conflicts with 
> the currently installed version. I am not quite sure whether I should 
> try uninstalling what is there now because I've read that the a few 
> programs need it to run.
> 
> Any suggestions? Also if anyone knows of a webpage listing the 
> step-by-step instructions on how to install/uninstall/upgrade the 
python 
> distribution, please reply. I have read the info provided on 
> www.python.org, but it is not detailed enough for me.

If you just want Tkinter, install the above mentioned RPM rather than
going to the trouble of upgrading.

Dave





From mwilson at the-wire.com  Wed Apr 16 07:35:24 2003
From: mwilson at the-wire.com (Mel Wilson)
Date: Wed, 16 Apr 2003 07:35:24 -0400
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net>
 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
 <mailman.1050438927.20366.python-list@python.org>
 <yu99y92bh10b.fsf@europa.research.att.com>
 <mailman.1050457621.2105.python-list@python.org>
 <slrnb9q90a.mhg.news@datatailors.com>
Message-ID: <8/Tn+ks/KXWc089yn@the-wire.com>

In article <slrnb9q90a.mhg.news at datatailors.com>,
Peter van Kampen <news at datatailors.xs4all.nl> wrote:
>In article <mailman.1050457621.2105.python-list at python.org>, Mike C.
>Fletcher wrote:
>> Andrew Koenig wrote:
>>
>> Okay, so now we have "The Practical Coder's Guide to big-O Notation" :)
>> .  We rock ;) ,
>> Mike
>
>You do, you know...c.l.p really does.
>
>I think I almost 'get it'. Except who or what decides what Big-O a
>certain algorithm has. Is that by 'hunch' or by running benchmarks or
>are there still other ways to recognize 'Big-O's (except for the
>formal computations that I understand are not very common)?
>
>for example
>
>bigO_N(i):
>    for n in xrange(i):
>        print n
>
>Would be O(N) right?

Yes.  A slightly more concrete example:

def bigO_N (a_list):
    for a in a_list:
        print a

where run-time depends pretty much on the len (a_list).


>bigO_N2(i,j):
>    for n in xrange(i):
>        for o in xrange(j):
>            print i,j
>
>This would be O(N**2)? Actually O(i*j) I suppose but if i==j N**2
>applies.

Again:

def bigO_N2 (a_list):
    for n in a_list:
        for m in a_list:
            print n, m

I guess two different lists might be used, just as different
i and j values were used above.  The run-time depends on the
product of two sizes, so it's called order N-squared.  The
O-notation is intentionally sloppy .. scarcely better that a
hand-wave, but it's a useful hand-wave.


>But these are not really 'algorithms'.

   Yes they are.

>                                       Most algorithms are a lot more
>complex even when reduced to it's essentials.

   Most algorigthms have a lot of processing details that
can distract you from regarding the bare-bones structure.
The O notation is only about that structure:
 - hit one element of the data, ignoring the rest
 - a pass over the data
 - a pass over the data for each data element
 - a pass over the data for each permutation of elements
 - something else

>                                              I understand why
>quicksort is better than bubblesort but I don't see an obvious way to
>label quicksort O(??).

   Divide-and-Conquer algorithms tend to be O(n log n).
Quicksort works by dividing the data set up into a high set
and a low set, and then quicksorting each subset separately.
Making a pass over the set to divide it into high and low is
O(n) .. just like the O(n) examples above in this post.  The
number of times you will divide the set in two is roughly
log2(n).  So performing log(n) copies of a process that's
O(n) makes for a process that itself is O(n log n).

        Regards.        Mel.



From francisgavila at yahoo.com  Mon Apr 28 23:20:22 2003
From: francisgavila at yahoo.com (Francis Avila)
Date: Mon, 28 Apr 2003 23:20:22 -0400
Subject: Dynamic repeat counts in struct format strings?
Message-ID: <varrqma1u6bocb@corp.supernews.com>

I suppose this message is just wondering aloud, but:

Why hasn't the struct module grown a feature to specify the repeat count of
an element based on the value of the previous element, like perl's
(http://www.perldoc.com/perl5.6/pod/func/pack.html)?

What I mean is, why can't I do something like this (borrowing perl syntax of
'/'):

>>> struct.unpack('!h/s', '\x00\x05Hello')
('Hello',)

Or, more generally:

>>> data = '\x02\x00\x04\x00\x05'
>>> struct.unpack('!b/h', data)
(4, 5)

The '/' specifies that the previous element functions as a repeat count for
the next element, much like how we can specify a numeric repeat count
currently.  This last snippet would be equivalent to what we have to do now:

>>> s = struct.calcsize('!b')
>>> i = struct.unpack('!b', data[:s])[0]  # i == 2
>>> struct.unpack('!'+`i`+'h', data[s:])  # fmt == '!2h'
(4, 5)

Even better would be:

>>> data = '\x02\x00\x03abc\x00\x04Test'
>>> struct.unpack( '!b(h(s))', data)
('abc', 'Test')

In other words, x(fmt), where x is repeat count of format string fmt, and x
is not included in the output.

And for pack():

>>> struct.pack('!b(h(s))', 'abc', 'Test')
'\x02\x00\x03abc\x00\x04Test'

(Implementing a pack() that acted like this would be quite a bit harder,
though, I think, since the semantics have changed. pack() as it is now is
very dumb--every element maps directly to something you pass it.)

And while I'm on it, I might as well mention that it would be nice if there
were a variant of unpack() that returned the unused data as the last element
in the tuple, instead of throwing an exception. (It's easy enough to write
one's self, though.)


Now, I'm not advocating that struct become perl's /pack and /unpack (struct
is _far_ cleaner), but being able to specify a repeat count within the data
itself would be a very useful feature, making it easier to wrap some generic
'data structure' type interface around struct (like the xstruct module),
faster (less python string/list fiddling and repeated calls to pack or
unpack), and generally less fiddly.

Is there any specific reason why this type of functionality never found its
way into struct?  Is it the same reason that pascal strings include the
length byte in the length (or is that just how pascal strings are
implemented)?  Is there some fundamental reason why calcsize() always has to
be deterministic (i.e., always knows exactly, given any format string, what
the length of the data will be?)  Even pascal strings require that you
explicitly name the length of the string when packing (it silently truncates
or null-pads the string you pass to it).  With this feature you wouldn't
even need 'p'!  It would be expressed as 'B/s' or 'B(s)'.


(Of course, everyone's just going to say, hey, great idea, now go implement
it!)

--
Francis Avila




From aleax at aleax.it  Wed Apr 16 11:38:50 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 15:38:50 GMT
Subject: pattern match
References: <mailman.1050504872.31313.python-list@python.org>
Message-ID: <eOena.23056$LB6.547891@news1.tin.it>

Gabor Nagy wrote:

> I have a string, like '(FOO(1) BAR("2"))'
> 
> I'd like to have a { 'FOO':1, 'BAR':'2' }
> 
> Is there an easy way to do this, or do I have to write parsing, maybe use
> regexps, or something like that?

Some parsing is obviously needed -- turning a string that respects
some syntax into "more meaningful objects" (in some appropriate
context for "meaningful";-) is the DEFINITION of parsing.

You may or may not have to "write parsing", depending on what syntax
you need to deal with in your input strings -- and regular expressions
may or may not be the best tools for the conceptually lower parts of
the parsing job ("tokenizing").  If the tokens in your strings happen
to respect exactly the same rules as Python's tokens, in particular,
you can use the excellent tokenize module in the standard library...:

>>> st = 'FOO(1) BAR("2")'
>>> import tokenize
>>> import cStringIO
>>> for x in tokenize.generate_tokens(cStringIO.StringIO(st).readline):
...   print x
...
(1, 'FOO', (1, 0), (1, 3), 'FOO(1) BAR("2")')
(50, '(', (1, 3), (1, 4), 'FOO(1) BAR("2")')
(2, '1', (1, 4), (1, 5), 'FOO(1) BAR("2")')
(50, ')', (1, 5), (1, 6), 'FOO(1) BAR("2")')
(1, 'BAR', (1, 7), (1, 10), 'FOO(1) BAR("2")')
(50, '(', (1, 10), (1, 11), 'FOO(1) BAR("2")')
(3, '"2"', (1, 11), (1, 14), 'FOO(1) BAR("2")')
(50, ')', (1, 14), (1, 15), 'FOO(1) BAR("2")')
(0, '', (2, 0), (2, 0), '')
>>>

as you see, you need to provide generate_tokens a "readline-like"
function -- when what you have to start with is a string, wrapping
a cStringIO.StringIO around the string and passing its bound
method .readline is generally simplest.

Each item x returned by looping on generate_tokens is a 5-items
tuple.  x[0] is an arbitrary code -- 1 means an identifier, 50
means punctuation, 2 a number, etc -- you can learn a little bit
more by checking the dictionary tokenize.tok_name.  x[1] is the
substring of the input string that corresponds to the token.
Then, you get indications of WHERE the token is in the input
string (where it starts and ends in terms of lines and columns)
and the input string itself -- stuff we don't care about.

So here's a perhaps more readable way to show what's going on...:

>>> for x in tokenize.generate_tokens(cStringIO.StringIO(st).readline):
...   print '%12s %r'%(tokenize.tok_name[x[0]], x[1])
...
        NAME 'FOO'
          OP '('
      NUMBER '1'
          OP ')'
        NAME 'BAR'
          OP '('
      STRING '"2"'
          OP ')'
   ENDMARKER ''
>>>

this still doesn't give you the dict you want, of course.
But then, we don't yet know what syntax your strings are in --
you haven't told us!  If the syntax is simple enough, and in
particular if you don't need to diagnose ERRORS in the input
strings, the task might perhaps be trivial...:

>>> def makedict(st):
...   thedict = {}
...   for x in tokenize.generate_tokens(cStringIO.StringIO(st).readline):
...     whattok = tokenize.tok_name[x[0]]
...     if whattok == 'NAME': name=x[1]
...     elif whattok in ('NUMBER','STRING'): thedict[name]=eval(x[1])
...   return thedict
...
>>> makedict(st)
{'FOO': 1, 'BAR': '2'}
>>>

here I'm ignoring punctuation, and just assuming that, apart from
punctuation and whitespace, the string is a sequence alternating
identifiers and values, where all values are numbers or strings.

Of course, it isn't very hard to perform more thorough checks, so
that you can identify mistakes in the input string -- or accept a
richer syntax for said input string -- but you'll really need to
tell us more about the latter's syntax before we can offer any
further help in these respects.


Alex




From jkraska at san.rr.com  Wed Apr 16 11:31:16 2003
From: jkraska at san.rr.com (Courageous)
Date: Wed, 16 Apr 2003 15:31:16 GMT
Subject: Big-O notation
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net> <mailman.1050438927.20366.python-list@python.org> <yu99y92bh10b.fsf@europa.research.att.com> <mailman.1050457621.2105.python-list@python.org> <slrnb9q90a.mhg.news@datatailors.com>
Message-ID: <8qtq9vgg7it0b5udsup2aqtpl7nqhao1hd@4ax.com>

>I think I almost 'get it'. Except who or what decides what Big-O a
>certain algorithm has.

It's mathematics. There's neither opinion nor "benchmarks".

C//




From tungwaiyip at yahoo.com  Fri Apr 25 13:56:10 2003
From: tungwaiyip at yahoo.com (Tung Wai Yip)
Date: Fri, 25 Apr 2003 10:56:10 -0700
Subject: Exception and finally question
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com> <b89rd1$2oij$2@agate.berkeley.edu>
Message-ID: <1tliav4uoelidf0oke36vvtttb3mn43e9e@4ax.com>

On Thu, 24 Apr 2003 23:18:57 +0000 (UTC), Benjamin Hartshorne
<ben at hal.rescomp.berkeley.edu> wrote:

>Tung Wai Yip <tungwaiyip at yahoo.com> wrote:
>> Hello,
>
>> How to I access the exception object in a handler?
>
>> try:
>>     raise NameError, 'HiThere'
>> except NameError:
>>     #how to print the message 'Hi There' here?
>
>try:
>  raise NameError, "HiThere'
>except NameError, whoshere:
>  print whoshere, "how are you?"
>
>-ben

Thank you. I didn't realize this syntax because I have it mixed up
with the tuple version.

 except (RuntimeError, TypeError, NameError):

By the way, mixing tuple with the exception target would look more
confusing (at least to me with Java background).

 # I guess we can code this way
 except (RuntimeError, TypeError, NameError), e:

 # this probably won't mix.
 except ,e:


tung




From alan_salmoni at yahoo.com  Tue Apr 22 05:21:44 2003
From: alan_salmoni at yahoo.com (Alan James Salmoni)
Date: 22 Apr 2003 02:21:44 -0700
Subject: "pythonic"
References: <glns9v0r21a1c88m7jn2fuvh3kdvkq5urh@4ax.com>
Message-ID: <b68f71a7.0304220121.ada9cec@posting.google.com>

I must confess that I thought that Pythonic was a state of mind:
something more like Zen than any strict definition could provide.

Let's just say that something Pythonic may only be defined by example.
"You know it when you see it" sort of stuff!

Alan James Salmoni
SalStat Statistics
http://salstat.sunsite.dk

Afanasiy <abelikov72 at hotmail.com> wrote in message news:<glns9v0r21a1c88m7jn2fuvh3kdvkq5urh at 4ax.com>...
> Too many times now I have seen the term "pythonic" used to vaguely
> describe something which is never explicitly defined, as if it cannot be.
> 
> Now it seems I can just replace "pythonic" with "my idea of utopia".
> 
> I do not know if this ignorance, laziness, or a meme, but I would like
> to know if there a real definition of "pythonic". Anyone?



From aahz at pythoncraft.com  Mon Apr 28 09:57:34 2003
From: aahz at pythoncraft.com (Aahz)
Date: 28 Apr 2003 09:57:34 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <mailman.1051376980.15852.python-list@python.org> <b8ekv0$o3e$1@panix2.panix.com> <mailman.1051490148.16068.python-list@python.org>
Message-ID: <b8jc0e$q9$1@panix1.panix.com>

In article <mailman.1051490148.16068.python-list at python.org>,
Steven Taschuk  <staschuk at telusplanet.net> wrote:
>Quoth Aahz:
>> In article <mailman.1051376980.15852.python-list at python.org>,
>> Steven Taschuk  <staschuk at telusplanet.net> wrote:
>>>
>>>It would be nice to be able to point people at a good on-line
>>>debugging tutorial for beginners; anybody know of such a thing?)
>> 
>> Doing a Google for "debugging tutorial" pulls this as the top link:
>> http://heather.cs.ucdavis.edu/~matloff/debug.html
>
>   [...]
>Of course, if I'd wanted to know what Google thinks, I'd have
>asked Google.  Perhaps you were wondering why I didn't; I'm hoping
>that one of the experienced teachers who posts here might make a
>recommendation backed by their experience.  That would carry much
>more weight (with me, at least) than Google's page rank.

What I didn't mention is that I used to go to UC Davis; while I never
took a class from Professor Matloff, I know him by reputation.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From sholden at holdenweb.com  Tue Apr 15 09:32:31 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Tue, 15 Apr 2003 13:32:31 GMT
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it> <ouadesfkn2.fsf@andlx-anamika.cisco.com> <7mOma.20244$T34.649927@news2.tin.it>
Message-ID: <PRTma.99121$It5.64393@news2.central.cox.net>

"Alex Martelli" <aleax at aleax.it> wrote ...
> Ganesan R wrote:
>
> >>>>>> "Alex" == Alex Martelli <aleax at aleax.it> writes:
> >
> >> Note that any Python function call always passes VALUES -- i.e.
> >> there is no difference between your call to sys.getrefcount in
> >> the above snippet and one that directly does sys.getrefcount(1).
> >> In either case, it's the VALUE (1) that you're asking info about;
> >> it makes no difference HOW you obtain the value to pass (I guess
> >> this is probably already clear to you, but, just in case...).
> >
> > Then, I find this confusing
> >
[source of confusion snipped along with 2.2 example]
>
> [alex at lancelot ox]$ python2.3
> Python 2.3a2+ (#11, Mar 28 2003, 12:17:11)
> [GCC 3.2 (Mandrake Linux 9.0 3.2-1mdk)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> >>> a='hello world'
> >>> b='hello world'
> >>> id(a), id(b), id('hello world')
> (1076642784, 1076642864, 1076643144)
> >>>
>
> In 2.2, we see Python using different copies of the literal string
> value in all three cases.  In 2.3, we see indications of a different
> strategy -- one value (object, identity, copy, call it as you wish)
> is being used for the first two occurrences, a distinct one for the
> third occurrence.  Strictly issues of optimization strategies on
> the part of different Python implementations, no more.
>
Just in case somebody comes across this message out of context and wonders
what on Earth you're talking about, I feelI should point out that the
Martellibot has here made an uncharacteristic mistake and assumed that
1076642784 == 1076642864, which in most universes is incorrect.

>
> Keep in mind that EQUALITY and IDENTITY are different concepts.  For
> immutables, equality is what you generally care about -- identity is
> shifty and implementation-dependent since such optimizations are
> permissible.  Thus, you shouldn't test "a is b" nor "a is 'hello world'",
> but rather you should use the == (equality check) operator in lieu
> of the is (identity check) operator for such tests.
>
> References are to specific values (specific objects, specific identities,
> specific copies, call them as you wish), thus all considerations about
> reference tracking need to keep this concept in mind.
>
>
> One last indication (not playing a role here, but you're sure to run
> into this as you play with id and the like...): id(...) is unique
> as long as an object exists, but as soon as the object is destroyed,
> its identity may be reused.  So, don't get tricked by such artefacts
> as the following:
>
> [alex at lancelot ox]$ python2.2
> Python 2.2.2 (#1, Oct 24 2002, 11:43:01)
> [GCC 2.96 20000731 (Mandrake Linux 8.2 2.96-0.76mdk)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> >>> id('hello world')
> 135683352
> >>> id('hello world')
> 135682152
> >>> id('hello world')
> 135683192
> >>> id('hello world')
> 135683352
> >>> id('hello world')
> 135682736
> >>>
>
> it may SEEM like the first and fourth literals use the same value,
distinct
> from the 2nd and 3rd and 5th -- but they don't -- the first one has been
> destroyed, and happens to get recycled, at the time you create the 4th (in
> other builds you'll no doubt get different results for this -- in 2.2, the
> allocator used is more system-dependent than in 2.3).
>
And the rest is just as true as it always was: I fell into this trap writing
a web server, where a form instance contained its id() value when
represented in HTML to try to trap incorrect submissions. Took me a while to
work out why the id() values were being repeated!

>
> Morals: don't care too deeply about identity of immutables...
>
That's only one moral.

nitpicking-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From dino_levy at mindlessSPAM.com  Mon Apr 21 07:28:17 2003
From: dino_levy at mindlessSPAM.com (Dino Levy)
Date: Mon, 21 Apr 2003 13:28:17 +0200
Subject: Mail and text extraction
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com> <3EA34DAE.EC67C1E3@engcorp.com>
Message-ID: <pan.2003.04.21.13.28.15.303279.2294@mindlessSPAM.com>

On Mon, 21 Apr 2003 03:47:26 +0200, Peter Hansen wrote:

 
> Why not download the whole message, then scan it after downloading? You
> provide no good reason, so far, for why you wouldn't just do this much
> simpler step.  Keep the design simple, modular, clean, and you can do
> most of what you want with simple, modular, existing solutions in
> standard modules.  Why wouldn't you do that?
> 
> -Peter

Well yes, that's antoher way of doing it, but I wanted to make it done in
one breath, like connecting and immediately parsing the text, but it's
obviously not achievable:)
I don't want MIME, I really need POP3 connections...
Ok, thanks anyway



From ckea25d02 at sneakemail.com  Tue Apr  1 20:07:54 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 02 Apr 2003 13:07:54 +1200
Subject: Comments
In-Reply-To: <435d40e0.0304011445.3abf38ff@posting.google.com>
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <3E8974E5.88B861D7@engcorp.com> <435d40e0.0304011445.3abf38ff@posting.google.com>
Message-ID: <b6dclq$49i5m$2@ID-169208.news.dfncis.de>

Chris wrote:
> I wanted to comment out code in a large if statement that was getting an error.

Another way to disable code is

   if 0:
     ...

This has the advantage that it will work even if the
code to be commented out itself contains any triple
quoted literals.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mhammond at skippinet.com.au  Mon Apr  7 00:19:50 2003
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Mon, 07 Apr 2003 04:19:50 GMT
Subject: Python ODBC interface crash on connection failure to MS SQL
In-Reply-To: <48e1ebc2.0304061945.3039c772@posting.google.com>
References: <48e1ebc2.0304061945.3039c772@posting.google.com>
Message-ID: <G%6ka.368$1h6.2207@news-server.bigpond.net.au>

Sophy wrote:

> Hi,
> 
> I would appreciate ur advice on this problem that i am facing. 
> I am running a python application and using the MSSQL database through
> a python ODBC interface. I find that the python ODBC interface crashes
> whenever there is a connection failure to the MSSQL. I get an "unable
> to read memory" message coming up and a print message "disconnected"
> which i later found in the ODBC.cpp source file in the method
> odbcPrintError() which handles the dbi errors.
> I use the same with MySQL and dont seem to have any problems with
> connection failure handling.
> 
> Can somebody advice me as to how i can resolve this issue.
> 

That printf() is strange - but unfortunately may not be near the crash. 
     If you have no access to MSVC etc, you could try using 
win32api.Apply() to catch the win32 exception - not a great option though.
Mark.




From tim at bladerman.com  Tue Apr  8 03:40:51 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Tue, 08 Apr 2003 07:40:51 GMT
Subject: How to wait for a click?
References: <af26e1d6.0304072256.2567f2e6@posting.google.com>
Message-ID: <72vka.4601$sJ3.15509@newsb.telia.net>

"Bill Mill" <llimllib at f2o.org> wrote in

> In the new structure, a referee class calls the genmove() function
> of a HumanGoPlayer class and expects as a return value a move. Herein
> lies the problem; how do I have genmove() (which has a reference to
> the board) wait for a click and return the move? I know I could have
> it check some variable in the board class every x milliseconds, but
> that's very inelegant. Is there a better way? Am I thinking about the
> problem wrong?

I am not sure I got your problem right but I give it a try anyway.
One way would be to let your genMove() function be really short and dont
expect a returnvalue. Instead it is just used as an indikator so that the
player knows that it is his turn (either the AI or Human player). Then the
player in turn cals a function in the structure callled
structure.hereIsMyMove(move)

Means the player always make a callback to the structre.

Another way would be to completley skip the genMove function and let players
call the structre anytime with a function
structure.hereIsMyMove(move,myId)
Then the structure must always check if I had the right to do the actual
move and if so, make the move.

None of these, I think, are nice looking but they should work and they are
slightly better then polling a variable every so milliseconds.

Good luck.

Tim





From fortepianissimo at yahoo.com.tw  Fri Apr 11 14:42:20 2003
From: fortepianissimo at yahoo.com.tw (Fortepianissimo)
Date: 11 Apr 2003 11:42:20 -0700
Subject: intern'ed strings and deepcopy()
Message-ID: <ef08387c.0304111042.73a16e68@posting.google.com>

I can't seem to find the exact description between these two in the
Library Reference - so am I correct in assuming that deepcopy() will
not actually duplicate an intern'ed string?

I have experimented with simple code but just want to get a definitive
answer on this one. Thanks.



From max at alcyone.com  Fri Apr 11 20:05:30 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 11 Apr 2003 17:05:30 -0700
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk>
Message-ID: <3E97584A.5E7FD112@alcyone.com>

Alexander Schmolck wrote:

> No it isn't. The idea of symbols (viz interned strings) is their
> uniqueness
> (amongst other things, establishing identity is faster then
> establishing
> string equality). So I don't think the OP's question is unreasonable:
> you
> might well want to deepcopy some container that contains symbols (and
> you'd
> like to preserve their identity).

My point was that if you find yourself wanting to make a truly unique
copy of a string (an immutable object), then there is almost certainly
something wrong with your design.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ It is human nature to think wisely and act foolishly.
\__/ Anatole France
    Bosskey.net: Quake III Arena / http://www.bosskey.net/q3a/
 A personal guide to Quake III Arena.



From tim.one at comcast.net  Thu Apr 24 11:26:50 2003
From: tim.one at comcast.net (Tim Peters)
Date: Thu, 24 Apr 2003 11:26:50 -0400
Subject: What means exactly "Memory error"?
In-Reply-To: <200304241406.20833.b.maryniuk@forbis.lt>
Message-ID: <BIEJKCLHCIOIHAGOKOLHCEBCFHAA.tim.one@comcast.net>

[Bo M. Maryniuck. gets a traceback ending with
    File "/usr/lib/python2.2/site-packages/_xmlplus/dom/Event.py", 
        line 93, in initEvent
    self.bubbles = canBubbleArg
    MemoryError
]

I can't help with your app, but I can tell you what MemoryError means.  Python raises MemoryError when:

1. A call to the system malloc() or realloc() returns NULL (there's
   not enough memory to proceed).

or

2. *Before* calling the system malloc() or realloc(), Python determines
   that the value of the argument it would need to pass is too big to
   fit in a variable of the platform's type size_t, so that it's not
   even possible to try asking malloc/realloc for that much memory (those
   take an argument of type size_t).

As with any failure at such a low level in the system, the possible causes are too numerous to try to guess.  The most likely cause is that you've simply run out of memory (#1 above) -- but that's more a symptom than an explanation.  Don't expect it to be easy to resolve.





From jepler at unpythonic.net  Thu Apr 24 08:09:39 2003
From: jepler at unpythonic.net (jepler at unpythonic.net)
Date: Thu, 24 Apr 2003 07:09:39 -0500
Subject: eval function
In-Reply-To: <3EA7CF53.1080100@gemplus.com>
References: <3EA7CF53.1080100@gemplus.com>
Message-ID: <20030424070938.A19928@unpythonic.net>

On Thu, Apr 24, 2003 at 01:49:39PM +0200, Franck Bui-Huu wrote:
> Hy,
> 
> I would like to use eval function with my own dicts.

You can't.  There's a partially-working patch somewhere for this (I wrote
it) but the performance impact was deemed too great for it to be included
in core Python (the performance impact was between 3% and 5% as far as I
recall, even when the feature was not used.  The performance impact is much
greater if the feature *is* used, though I don't remember a figure)

Try looking for a related patch on the python patch tracker.  You can
recompile Python with this patch added, if you know how.

Jeff




From aleax at aleax.it  Tue Apr 15 18:27:25 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 22:27:25 GMT
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net> <30uo9vo46k53v4to6d65koh72nqf0p8nvn@4ax.com>
Message-ID: <hH%ma.26675$T34.794125@news2.tin.it>

carroll at tjc.com wrote:

> On Tue, 15 Apr 2003 15:37:39 -0400 (EDT), wilfred_s at webtv.net (W S)
> wrote:
> 
>>Hello,
>>
>>I find the material, as presented, in the book Programming Python
>>frustrating and extremely difficult to understand.
>>
>>Are the nutshell and cookbook O'Reilly Python books easier to read?
> 
> I really dislike the "Programming Python" book.[1]  After playing
   ...
> (The "case study" issue I describe above, which I consider a "bug,"
> the above review considers to be a feature: "Many of the example
> scripts and programs build on one another from chapter to chapter,

Yes, there seems to be an irretrievable split on this issue, for
both authors and readers of programming books: some thrive on
"significant" examples, i.e. programs that DO something useful,
complicated, etc, while others far prefer examples that focus on
the specific issue being examined, almost inevitably "toy"
examples to cut out issues that are "locally" irrelevant.

For example, among the C++ classics, Stroustrup's book tries hard
to give "significant" examples -- I've always far preferred Lippman's
book, with his "toy" examples, which, to me, illustrated C++'s
various aspects so much more clearly and sharply (Eckel's "Thinking
in C++" is much in the same vein, and I also appreciated it highly).

Since I've thus shown I belong in the "toy examples" brigade, you
can bet that's emphatically what I put in the Nutshell (and that
I didn't like "Programming Python" either, back when I was learning
Python, much as I appreciate Mark Lutz for other works, such as
the Pocket Reference and Learning Python) -- not that a Nutshell
has space for "large examples" anyway, of course (unless, like was
done for the Java one, you split off the examples into a separate
book -- quite a different issue;-).

The Cookbook is quite different, since the mix of examples was
essentially decided by the Python community -- what David Ascher
and I did was essentially a job of _selection_ (and adding a
few snippets to illustrate some things the community had basically
neglected -- as well of course as discussion/comments/etc).

So, in the Cookbook, you'll find plenty of examples that are just 
snippets illustrating very specific points, as well as "toy-level", 
simple things, AND a few major, multi-page "recipes" that do a 
significant, complex job.

However, for both the Cookbook AND the Nutshell, I did do my best 
to consider *NON*-sequential readers -- people who just want
to delve deep into ONE specific subject and get the gist of it --
as well as people who enjoy reading the book, or some chapters in
it, sequentially.  How well I managed to, of course, it's not for
me to say -- you can find several reviews of both the Cookbook
and the Nutshell, both on the web and specifically in this NG
(google will help you find them -- "Advanced Group Search" for
stuff in this newsgroup, of course); and you can look at the
sample chapters for each work, from O'Reilly's site, or Amazon,
or no doubt other sites yet; and make up your own mind...!


Alex




From tyler at cg1.org  Fri Apr 18 20:38:06 2003
From: tyler at cg1.org (Tyler Eaves)
Date: Sat, 19 Apr 2003 00:38:06 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <Gu_na.34381$LB6.779929@news1.tin.it>
Message-ID: <OT0oa.690$l75.32@tornadotest1.news.pas.earthlink.net>

<posted & mailed>

Alex Martelli wrote:

> Tim Gahnstr?m /Bladerman wrote:
> 
>> 
>> "Evan Simpson" <evan at 4-am.com> wrote in
>>> Guy Middleton wrote:
>>> > >>> reduce(lambda a, b: a+b, [y[1] for y in x])
>>> >
>>> > But I don't really like lambda, is there a better way to do this?
>>>
>>> What's wrong with:
>>>
>>> t = 0
>>> for y in x:
>>>      t += y[1]
>>>
>>> The extra variable?  The number of lines?
>> 
>> I am really curious about that to, I would most definitley say that this
>> is the most pythonic way. It is simple and easily readabel by anyone and
>> I am sure it is just as fast any of the other way.
> 
> Quite.  Performance is easy to measure, rather than guess at:
> 
> [alex at lancelot python2.3]$ python -O timeit.py -s"lst=[('a',1), ('b',2),
> ('c',3)]*100" -s'import operator' 'x=reduce(lambda x,y: x+y[1], lst, 0)'
> 1000 loops, best of 3: 235 usec per loop
> 
> [alex at lancelot python2.3]$ python -O timeit.py -s"lst=[('a',1), ('b',2),
> ('c',3)]*100" -s'import operator' 'x=reduce(operator.add,[y[1] for y in
> lst])'
> 1000 loops, best of 3: 277 usec per loop
> 
> [alex at lancelot python2.3]$ python -O timeit.py -s"lst=[('a',1), ('b',2),
> ('c',3)]*100" -s'import operator' 'x=0' 'for y in lst: x+=y[1]'
> 10000 loops, best of 3: 103 usec per loop
> 
> Using lists 10 or 100 times as long seems to show O(N) behavior for
> each case, as could be expected.  So, the simplest (and I would agree
> thereby most Pythonic) approach is over twice as fast as the fancy ones
> (measurements on 2.2.2 seem to show similar results, all approaches
> being about 20% slower than in 2.3, uniformly).
> 
> 
> Alex

Wow, I feel vindicated now for never trying to wrap my brain around stuff
like list comprehensions. *g*  




From arnaud at tribu.ch  Wed Apr 30 09:42:52 2003
From: arnaud at tribu.ch (Arnal)
Date: Wed, 30 Apr 2003 15:42:52 +0200
Subject: Three questions about Tkinter
Message-ID: <BAD59F7C.B1B%arnaud@tribu.ch>

Hi,

I'm new at python programming (as well as english writing!), and I'm trying
to make a simple text editor.

I'm using Tkinter, but I've the following problems:

-I can't figure out how to use a printer to print under Windows.

-I use a "text" control (I dislike the "widget" term, sorry), but I can't
resize it on the fly (the width and height properties are measured in
characters and lines, not in pixels). My current workaround is to make the
window not resizable but it would be nice to do it.

-I've a sub menu for recents files. When I read in my pref file to list
every file paths stored in it, I would like to add them in the recents
submenu. I can't find how to do it. I tried, inside a "while" with the
variable i:

MRecents.insert_command(0,label=GetFileName(file),accelerator=acc,command=Op
enRecent(i))

MRecents.insert_command(0,label=GetFileName(file),accelerator=acc,command=Op
enRecent i)

(MRecents is a global variable representing my submenu, GetFileName is a
function, no special meaning in the problem,acc is defined). In the first
example, the OpenRecent function is called directly, that's the normal
behavior, and there is a syntax error in the second example.
How to define an array of menu items with all calling the same function with
a parameter?



Has someone any information on one of these problems? It would be great.
Thanks in advance.




From johnroth at ameritech.net  Wed Apr  9 19:00:49 2003
From: johnroth at ameritech.net (John Roth)
Date: Wed, 9 Apr 2003 19:00:49 -0400
Subject: string() - Is this some kind of sick joke?
References: <36dbf824.0304091142.527af834@posting.google.com>
Message-ID: <v99998l6b9ae37@news.supernews.com>

"Daniel Gowans" <chlump at yahoo.com> wrote in message
news:36dbf824.0304091142.527af834 at posting.google.com...
> There is some sick joke in the python 2.2.2 documentation I have.
> This is located in the ActivePython documentation:
>
> strip(s[, chars])
>   Return a copy of the string with leading and trailing characters
> removed. If chars is omitted or None, whitespace characters are
> removed. If given and not None, chars must be a string; the characters
> in the string will be stripped from the both ends of the string this
> method is called on.
>
> The same goes (in the docs) for rstrip and lstrip.
>
> But does it actually work that way?  No!
>
> Observe from the 2.2.2 interpreter (and 2.2.1 in HPUX):
>
> >>> y="'hello'"
> >>> import string
> >>> string.strip(y,"'")
> Traceback (most recent call last):
>   File "<interactive input>", line 1, in ?
> TypeError: strip() takes exactly 1 argument (2 given)

This is incorrect. The two operand version is a *function* in the string
module, the one operand version is a *method* on the string
type. My copy of the documentation is quite explicit about
the difference. It's consistent with all of the string functions/
methods.

For the string *methods*, see Sequence Types under
Built In Types.

What you gave in the example was the method on the
string type "string.strip([chars])". This only takes one
operand because the string is the string you're calling the
method on.

> >>>
>
> What gives?  Did I get the bogus version?  Is this wishful thinking
> documentation?  I also tried  y.strip("'")  with the same results.

Another poster said that this does appear to be a bug.

John Roth


>
> If this can't be done this way (oh, how I love chomp()) then is there
> a simple, elegant way to do it?

The correct way is to use the method.

John Roth





From wedjlok at attbi.com  Mon Apr 28 15:02:04 2003
From: wedjlok at attbi.com (Eric Anechiarico)
Date: Mon, 28 Apr 2003 19:02:04 GMT
Subject: saving output to file
Message-ID: <MUera.658292$L1.187296@sccrnsc02>

I am still not as proficient as I would like to be in Python, so this may be
a trivial question to most.

I want to have the output I get in my IDE window (using MacPython 2.2.2
primarily in Carbon) to be written to a file instead of having to copy the
output and then paste it to a text file myself.  How would I go about doing
this?

Mainly I would be using the code to do this with programs that calculate
outputs of large numbers (primes, Fibonacci, etc.) and they sometimes
overflow the buffer and I cannot get back very far to view the beginning.
Make sense?  Basically since I am doing this in the IDE window anyway.

Any suggestions?

Thanks!





From michael at stroeder.com  Thu Apr 17 10:09:03 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Thu, 17 Apr 2003 16:09:03 +0200
Subject: Pythonic way of web-programming
In-Reply-To: <slrnb9qlmb.djs.jon+usenet@snowy.squish.net>
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <B8Pma.20554$T34.656139@news2.tin.it> <mailman.1050471334.22000.python-list@python.org> <0gqdnfw_XuYm3gCjXTWcow@speakeasy.net> <slrnb9qlmb.djs.jon+usenet@snowy.squish.net>
Message-ID: <v4p2n-ci3.ln1@nb2.stroeder.com>

Jon Ribbens wrote:
> 
> You're wrong - FastCGI is excellent.

Hmm, sometimes it's acting a little bit strange...some hangs...

Ciao, Michael.




From blunck at gst.com  Tue Apr 29 18:14:41 2003
From: blunck at gst.com (Christopher Blunck)
Date: 29 Apr 2003 15:14:41 -0700
Subject: Where did you learn to unit test?  How did you learn?
References: <1147e466.0304290638.287687f6@posting.google.com>      <200304290836.42804.shalehperry@attbi.com> <mailman.1051634984.9521.python-list@python.org>
Message-ID: <1147e466.0304291414.546f57f4@posting.google.com>

"andrew cooke" <andrew at acooke.org> wrote in message news:<mailman.1051634984.9521.python-list at python.org>...
> imho help with refactoring is by far the biggest bonus of static typing
> (in a language like java; in more functional languages refactoring tends
> to be more about adding higher order functions which, again, are helped by
> static typing, but the experience is somewhat different).  this may depend
> on your development environment, of course - at the moment i'm working
> with a somewhat ugly codebase that needs a lot of refactoring.

That's a very good point.  And it's something that I do miss in
Python.  I guess you could argue that if you have solid unit test
coverage that this becomes a non-issue (the classes that use the old
interface will no longer work correctly).


> a key point to making tests work in java is abstracting out test
> specifications.  most of the work of test classes is adding frameworks
> that allow cases to be specified in a compact manner.  once that is done,
> adding test cases is very simple (for example, i verify trees from a
> parser against xml specifications - the framework is fairly complex, but
> once done it's easy to test many different cases).  this is particularly
> important in java because the language itself (the syntax) is fairly
> inflexible and verbose.

Very funny because I just made this point not more than 1 hour ago to
a coworker.  He was struggling through writing the first unit test in
a subsystem that has never been unit tested.  His frustration
surrounded your exact argument - adding the initial unit test
framework is difficult, but makes incremental test development
trivial.  Point in case is that his TestData class (which produces
objects that he uses to test) is about 500 lines of code.  His actual
unit test is about 20 - 40 lines of code.  But now that the framework
is there, lots more 20 - 40 line unit tests can be written.  This is
(of course) in Java, but it translates to python as well.


> this approach (lead by example) seems to me much better than forcing
> people to use tests.  after all, they're not always appropriate.  and
> there's no surer way of making someone dislike somethig than forcing them
> to do it.  also, generally, leading by example seems to get you a "better"
> team.

Yup.  And of course - there's only so much you can teach a team that
isn't interested in learning.  I'd add a corollary to your statement:
it's equally important to know when you've done all you can for a team
and it's time to move on.



From anton at vredegoor.doge.nl  Mon Apr 28 05:45:46 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Mon, 28 Apr 2003 11:45:46 +0200
Subject: Python-2.3b1.exe install error on W98se
Message-ID: <b8itdt$53h$1@news.hccnet.nl>

A message box popups almost immediately:

<message box>
Install

Could not initialize installation. File size expected=1972043840, size
returned=9096924. The parameter is incorrect.

		OK
</message box>

There are situations where a good simulation is organizationally
invariant to the real thing. For example a good simulation of a chess
program must play chess just as well as the real chess program. Please
keep this in mind if directing me to sourceforge in order to file a
bug report :-)

Anton




From aleax at aleax.it  Fri Apr 25 18:51:12 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 25 Apr 2003 22:51:12 GMT
Subject: Dealing with files...
References: <pan.2003.04.25.22.39.14.432753@soul.net>
Message-ID: <AZiqa.5752$3M4.154723@news1.tin.it>

lost wrote:

> Hi guys!
> I'm starting to deal with files, and need to know how to deal directly
> with the filesystem. what I'm looking for, is similar to this pseudo code:
> 
> for each_file in "/var/path/to/files/":
> print each_file #show the NAME of the file
> write the NAME of the file    into file_list.txt
> print "done"
> 
> This would print the name of each file in /var/path/to/files (for example,
> a list of html files), and then add the name of the file into
> file_list.txt.
> 
> Any help on this would be very much appreciated! Thanks.

import os

file_list = open('file_list.txt', 'a')

for each_file in os.dir('/var/path/to/files/'):
    print each_file
    file_list.write(each_file)
    file_list.write('\n')

file_list.close()

print 'done'


Here, I'm assuming you want to APPEND the names of the files,
one per line, to the textfile whose name you indicated.

Also, I'm assuming that you want to see subdirectories (if any)
as well as files.

It's trivially easy to make each of these assumptions in some
OTHER direction, of course, but, unless you give precise specs,
people who try to help you will have to just take their best
guesses at what you might have meant.


Alex




From jarausch at skynet.be  Sun Apr 27 16:44:52 2003
From: jarausch at skynet.be (Helmut Jarausch)
Date: Sun, 27 Apr 2003 22:44:52 +0200
Subject: 2.3b1 and Tk8.4.2
In-Reply-To: <b8hcd5$f5d$07$1@news.t-online.com>
References: <3eac2d62$0$594$ba620e4c@reader1.news.skynet.be> <b8hcd5$f5d$07$1@news.t-online.com>
Message-ID: <3EAC4144.9080105@skynet.be>

Martin v. L?wis wrote:
> Helmut Jarausch wrote:
> 
>> Is there a fix other than keeping Tk8.3 just for Python?
> 
> 
> I recommend to analyse the problem first.
> 
> Why is there a reference for TkSetClassProcs in _tkinter.so?
> _tkinter.c does not call that function.
> 

Yes, sorry. I've just noticed that tkinter is linked against libBLT
which I hadn't rebuilt against my current Tcl/Tk

Now it builds just fine.

Many thanks,
Helmut.


-- 
Helmut Jarausch

Lehrstuhl fuer Numerische Mathematik
RWTH - Aachen University
D 52056 Aachen, Germany




From sholden at holdenweb.com  Sat Apr 19 09:45:11 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Sat, 19 Apr 2003 13:45:11 GMT
Subject: Why functional Python matters
References: <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk> <3E9EB3E9.1040208@removethis.free.fr> <b7m546$2a83$1@news.peterlink.ru> <lyptnlxq1n.fsf@cartan.de> <m24r4xdyd6.fsf@mycroft.actrix.gen.nz> <lyd6jlxkpi.fsf@cartan.de> <usft9v01l8fcdg8erdl2kishg0cnr1t9dp@4ax.com> <lyadepp1ck.fsf@cartan.de> <1lt0avslergps3bdv1lpvr48kef6t03640@4ax.com> <87d6jjjuzu.fsf@darkstar.cartan> <bdh1avcltqircf31cnkfro4thkafe2pamq@4ax.com>
Message-ID: <Hpcoa.134462$It5.21768@news2.central.cox.net>

"Thaddeus L Olczyk" <olczyk at interaccess.com> wrote in message
news:bdh1avcltqircf31cnkfro4thkafe2pamq at 4ax.com...
> On 19 Apr 2003 00:33:57 +0200, Nils Goesche <ngo at cartan.de> wrote:
>
> >Thaddeus L Olczyk <olczyk at interaccess.com> writes:
> >
> >> While it's true that in math and CS, many definitions evolve
> >> over time ( ones that come to mind right now are
> >> function/procedure [which were distinct, now considered the
> >> same thing], fiber bundles, sheaves, topological space ... ),
> >
> >Actually, the resp. definitions of fiber bundles, sheaves and
> >topological spaces /haven't/ changed much over time.  If they
> >have changed at all, once they were called by that names.
> >
> When Gauss spoke of topological spaces they were more
> like smooth manifolds ( smooth Hausdorff manifolds to be more
> precise ). The definition changed over a period of time till
> it settled on it's present definition mid 40's ( IIRC).
>
> Steenrod discusses the problems of varying definitions
> of fibre bundles in his book.
>
> I'll skip sheaves since I don't remember the details.
> >> that's all just bullshit thrown out to distract people from
> >> *your stupidity*.
> >
> >Darn, I had hoped I'd get away with it one more time.
> >
> Don't you mean twice, as it's clear that this reply is
> another attempt at distraction, totally ignoring the
> issues.
>
> >> Thaddeus L. Olczyk, PhD
> >
> >I was somewhat hoping that now, having finally obtained your
> >degree,
> Sorry to disillusion you, but I got my doctorate many years ago.
> I stopped keeping track when it passed ten.
>
> Perhaps you will learn from this mistake. Making unfounded
> assumptions will often cause you to make dumb errors. Keep
> this in mind when you start doing serious programming, as
> it wil save you from a lot of  errors in your code.
>
> >which is seemingly very important to you,
> Is it? I suppose so. A doctorate is not something just anyone
> can accomplish. It takes a dedicated effort over a period of years,
> something most people don't have the patience for.
>
> Most of the time I really don't think about it. It's just something
> I did in the past.
>
>
> >you'd have
> >become a bit more relaxed.  I was wrong.
> You seem to be wrong quite a bit of the time. Perhaps you would
> not be wrong as much if you did not make unfounded assumptions.
> --------------------------------------------------
> Thaddeus L. Olczyk, PhD
> Think twice, code once.

See what happens when the Lispers invade (or did we invade them? - either
way, it didn't take long to descend to the stage of ugly name-calling). I
didn't think an 11-year-old would be able to *get* a Ph.D. If it's just
something he did in the past, why does it appear at the foot of every email
he sends out?

glad-to-stick-to-a-friendlier-environment-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From jhaveri at aludra.usc.edu  Wed Apr  2 05:03:46 2003
From: jhaveri at aludra.usc.edu (Jinal Jhaveri)
Date: Wed, 2 Apr 2003 02:03:46 -0800
Subject: problem installing tkinter
Message-ID: <Pine.GSO.4.33.0304020202140.6367-100000@aludra.usc.edu>

Hi All ,
I am facing a big problem while installing tkinter

I have itcl-3.2 already installed.
While installing tix I get following error
error: Failed dependencies:
        libitcl3.2.so is needed by tix-8.1.4-89
        libitk3.2.so is needed by tix-8.1.4-89
Now due to me not able to install tix, I cannot even install tkinter.
Can anyone plzz guide me
thank you
Jinal


                        ALWAYS AIM HIGH^(HIGH)

The strongest people are not those who always win but those who stand back up when they fall




From rmunn at pobox.com  Fri Apr  4 19:04:41 2003
From: rmunn at pobox.com (Robin Munn)
Date: Sat, 05 Apr 2003 00:04:41 GMT
Subject: Invalid format character in string
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com> <mailman.1049493668.26508.python-list@python.org> <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com>
Message-ID: <slrnb8rics.k4j.rmunn@localhost.localdomain>

I rearranged your text to put the response AFTER what you're responding
to, which is the natural reading order. Please don't top-post.

Nikolai Kirsebom <nikolai.NOJUNK at micon.no> wrote:
> On Fri, 4 Apr 2003 15:59:48 -0600, Skip Montanaro <skip at pobox.com>
> wrote:
>>
>>    Nikolai> ... since the text is actually a HTML file, it may contain a
>>    Nikolai> lot of characters with % (typically cell width definitions of
>>    Nikolai> tables).  If these are contained in the file, the string format
>>    Nikolai> statement complains about illegal format character.  
>>
>>Just double any literal % characters, e.g.:
>>
>>    <table width="%(width)s%%">
>>
>>Skip
> 
> The problems arise where the HTML editor used (to produce the file in
> the first place) has put various expressions of %; and %>.  These are
> HTML codings the creator of the HTML file does not really know about.
> As an example, I've just told them (the users creating the HTML file)
> that they should write the text "%(name)s" where they want the name
> generated and "%(age)s" where they want the age generated.  Then my
> script is run for a set of person objects producing a set of HTML
> files with the correct names and corresponding ages. 

If the original HTML file is not under your control, then what you
probably want to do is run it through a search-and-replace to double any
% characters that are meant to be taken literally. I.e.,

    import re

    html_str = '... %(width) wide, 50% off this week only! ...'
    quoted_str = re.sub(r'%(?!\()', '%%', html_str)
    print quoted_str

This would produce:

'... %(width) wide, 50%% off this week only! ...'

Notice the use of a negative look-ahead assertion (the '(?!...)' syntax)
to only match % characters that aren't followed by an open parenthesis.

By the way, please don't top-post like that; it makes multiple-person
conversations very hard to read. See http://www.caliburn.nl/topposting.html
for more reasons why bottom-posting is the accepted style.

Hope this helps!

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From exarkun at intarweb.us  Thu Apr  3 11:50:25 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 3 Apr 2003 11:50:25 -0500
Subject: os.chmod() question ?
In-Reply-To: <bd993a2f.0304022308.19f36fec@posting.google.com>
References: <bd993a2f.0304022308.19f36fec@posting.google.com>
Message-ID: <20030403165025.GB30464@meson.dyndns.org>

On Wed, Apr 02, 2003 at 11:08:24PM -0800, Anand B Pillai wrote:
>
> [snip - differeing behavior of os.chmod on Windows and Linux]
> 
> Do I need to pass an octal integer as the second argument ?

  No.  But you do need to pass the right bits, and these are most commonly
selected by typing in an integer in octal base.  ;)  You seem to be aware
that 0777 is not the same as 777, so go ahead and throw the 0 in front. 
This should be required on Linux as well as Windows, so I'm not sure how you
got the results you did on Linux (I get -r----x--t when passing 777 to
chmod).

  Jp

> 
> Thanks
> Anand Pillai
> -- 
> http://mail.python.org/mailman/listinfo/python-list

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 14 days, 12:00, 4 users, load average: 0.00, 0.03, 0.00




From aahz at pythoncraft.com  Sat Apr  5 10:28:00 2003
From: aahz at pythoncraft.com (Aahz)
Date: 5 Apr 2003 10:28:00 -0500
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <pan.2003.04.05.09.49.45.708544.878@thomas-guettler.de>
Message-ID: <b6msm0$1un$1@panix1.panix.com>

In article <pan.2003.04.05.09.49.45.708544.878 at thomas-guettler.de>,
Thomas Guettler  <pan-newsreader at thomas-guettler.de> wrote:
>
>If you want it to be 100% Python you can use medusa: A webserver written
>in python.

These days, you should use Twisted instead of medusa.
http://www.twistedmatrix.com/

At PyCon, I realized an hour before my presenation that I'd neglected to
set up a web server for my threads tutorial (which includes a spider),
and I was able to set up Twisted's web server in about fifteen minutes.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From eichin at metacarta.com  Tue Apr 29 18:51:42 2003
From: eichin at metacarta.com (eichin at metacarta.com)
Date: 29 Apr 2003 18:51:42 -0400
Subject: POD vs. reST for standalone writing?
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <mailman.1051590229.26633.python-list@python.org>
Message-ID: <7gy91ssyrl.fsf@pikespeak.metacarta.com>

> That's significant: "somewhat (and I think only a little) easier to
> read".  I designed reStructuredText to be a *lot* easier to
> read. Opinions vary of course, I think mostly determined by what
> you're used to.  (Some people think in XML, others dream in TeX.  Not
> me.)

Oh, very much so - and what I dream in is often much influenced by
what I've been coding/speaking recently... which is why POD struck a
cord after doing a few months of python programming, actually (the
whole "oh, just indent the literal stuff" part, the rest is very
light.) (15 years ago, I could type LaTeXish equations faster than
many people could write them on paper, and certainly faster than they
appeared on most chalkboards :-)  Not today, though.)

And I certainly see the value of biasing towards the reader - decades
of C and C++ justify significant overreaction in that regard :)  At
the same time, the need to encourage casual writers is still there.

The particular writability issue that comes to mind is the use of
underlining characters, quoting from quickstart.html:

Chapter 1 Title
===============

Section 1.1 Title
-----------------

Subsection 1.1.1 Title
~~~~~~~~~~~~~~~~~~~~~~

and the description "The underline/overline must be at least as long
as the title text."  My first problem with this is that if changing
the content causes the markup to need changing, the markup changes
will be a bit more error prone.  The next is that level for a given
punctuation type is based on, if I'm not misinterpreting this, what
punctuation usage has been seen *in this document*, which makes moving
selections among a small set of documents a little harder, unless you
have an explicit style guide (in which case, why not represent it
outright in the code?)

POD here would just say

=head1 Chapter 1 Title

=head2 Section 1.1 Title

=head3 Section 1.1.1 Title

(ignoring that in both reST and POD, you'd really just say "Title" for
all three and leave the numbering/layering to the stylesheets.)

Perhaps this is just one corner of the notation - but I'm looking for
a case that simply doesn't have all that much notation, or that much
to explain to the end user.

(more to come, apologies for stopping in the middle, there have been a
wealth of good comments here...)



From tjreedy at udel.edu  Wed Apr 16 13:36:03 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Wed, 16 Apr 2003 13:36:03 -0400
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <IrOdnUSmPonOxwCjXTWcpA@comcast.com> <n0iqy1zg.fsf@python.net>
Message-ID: <1EadnQb6ZY_RDwCjXTWcog@comcast.com>

"Thomas Heller" <theller at python.net> wrote in message
news:n0iqy1zg.fsf at python.net...
======================================================================
> FAIL: test_failUnlessEqual (__main__.FailingTests)
> --------------------------------------------------------------------
--
> TestFailed: 0 != 1
>   File "xunit.py", line 12, in test_failUnlessEqual
>     self.failUnlessEqual(self.a, self.b)

I like having the actual values too.  It should make debugging an
unexpected failure easier.  But what if each value is more that a few
bytes?  Example (from my future plans): two lists of all permutations
of range(8), one produced by an obvious recursion and the other by a
much less obvious iteration.

Hmmm.  What I would really like is a structure comparison function
that spit out the first difference found  so I did not have to print
each and compare by eye.  That is not directly in the scope of your
problem, but it would be the most useful succinct display.

Terry J. Reedy






From roy at panix.com  Tue Apr  1 13:08:59 2003
From: roy at panix.com (Roy Smith)
Date: 1 Apr 2003 13:08:59 -0500
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <x7of3qjh4g.fsf@guru.mired.org>
Message-ID: <b6ckjr$903$1@panix2.panix.com>

Mike Meyer  <mwm at mired.org> wrote:
> The trick to getting good performance out of interpreted languages is
> doing things to take advantage of those libraries.

This is a very true statement, and one which I think is
under-appreciated by a lot of people.

Over the past couple of years, I've been part of a group that wrote
about 18 KLOC in TCL.  One thing we found over and over again is that
if you care about speed, you need to throw away a lot of your
pre-conceptions about what's fast and what's not.  This is as true in
Python or Perl or Java as it is in TCL. To a certain extent, it's even
true of raw C++ vs. STL.

For example, let's say you're building up a list and at the end, you
want to know how many elements you added to the list.  The obvious
thing to do would be to keep track of how many elements you add as you
build the list.  It certainly seems like it should be more efficient
than building the list and then calling llength (that's TCL for
len()).  In fact, that's true in C or C++, but not true in most
interpreted languages.  It certainly wasn't true in TCL (we did
measurements).

I just tried it in Python, and (not surprisingly), found that it's not
true in Python either.  The program at the end of this post builds a 2
million element list two different way.  One way, it counts the
elements as it appends them.  The other way, it just appends them and
calls len() on the result.  When I run it on my 700 MHz Pentium Linux
box, I get:

-----------------------------------------
The first list is 2000000 long
It took 7.678 seconds to build, 0.000 to count, 7.678 total
The second list is 2000000 long
It took 9.103 seconds to build and count it
-----------------------------------------

Once you've performed the experiment, the reason is obvious.  The
naive thought that llength, or len(), would have to walk the list to
count the elements in linear time is just plain wrong; in both cases
it's got the length stored somewhere and can return it in constant
time.  By building the list, you've already paid for the cost of
computing its length.

The bottom line is exactly what Mike said: in interpreted languages,
the more you can push down into the libraries and interpreter core,
the better off you are most of the time.  And profiling helps too :-)

--------------------------------------

#!/usr/bin/env python

import time

n = 2000 * 1000

t0 = time.time()
l = []
for i in range (n):
    l.append(i)

t1 = time.time()
length = len(l)
t2 = time.time()

print "The first list is %d long" % length
print "It took %3.3f seconds to build, %3.3f to count, %3.3f total" \
      % (t1-t0, t2-t1, t2-t0)

del l

t0 = time.time()
l = []
length = 0
for i in range (n):
    l.append(i)
    length += 1
t1 = time.time()

print "The second list is %d long" % length
print "It took %3.3f seconds to build and count it" % (t1 - t0)



From peter at engcorp.com  Mon Apr  7 08:51:28 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 07 Apr 2003 08:51:28 -0400
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no> <3E915BA0.A4B5286@engcorp.com> <pan.2003.04.07.11.36.03.78556@idi.ntnu.no>
Message-ID: <3E917450.CBD5CDD9@engcorp.com>

Nils Grimsmo wrote:
> 
> On Mon, 07 Apr 2003 07:06:08 -0400, Peter Hansen wrote:
> > When do you think/hope run() will be called?  I don't think that
> > method is looked for by wxPython...  Did you base this on
> > some sample code?
> 
> run() is called when you push the button. run() sets the label for the
> button every second for four seconds, but the button label is not updated
> before you return from run().
> 
> that is not the behavior i want. any suggestions?

Ah, yes, clearly run() is called from your own code.  It was
so obvious I didn't see it.  :-)

Well, this time I actually tested your program, under Windows with
Python 2.2.2 and wxPython 2.3.3.1 and it works just fine.  I click 
the button and see foo/bar/foo/bar over four seconds.   

What platform are you on, and what versions?

-Peter



From ianb at colorstudy.com  Mon Apr  7 13:50:11 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 07 Apr 2003 12:50:11 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <Pine.GSO.4.51.0304071626360.8306@techunix.technion.ac.il>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	 <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
	 <b6qs04$7eqi1$1@ID-169208.news.dfncis.de>
	 <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
	 <mailman.1049693590.23386.python-list@python.org>
	 <m28yum76rt.fsf@localhost.localdomain>
	 <Pine.GSO.4.51.0304071626360.8306@techunix.technion.ac.il>
Message-ID: <1049737811.9724.235.camel@lothlorien>

On Mon, 2003-04-07 at 08:58, Beni Cherniavsky wrote:
> Cooperative super calls will be needed for things like keymap lookup
> ot accomodate the need to look in more than one keymap.
> Alternatively, you acn create a single default method to do keymap
> lookup and it will dispatch to all appropriate methods / dicts of each
> class involved that can then be written without bothering with
> cooperation.

When hooks fail, events can often work better.

  Ian






From andrew-pythonlist at puzzling.org  Tue Apr 29 00:38:15 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Tue, 29 Apr 2003 14:38:15 +1000
Subject: POD vs. reST for standalone writing?
In-Reply-To: <b8kuet$ftn$1@panix1.panix.com>
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com> <mailman.1051584944.30800.python-list@python.org> <b8kuet$ftn$1@panix1.panix.com>
Message-ID: <20030429043815.GB13813@frobozz.local>

On Tue, Apr 29, 2003 at 12:18:37AM -0400, Aahz wrote:
> In article <mailman.1051584944.30800.python-list at python.org>,
> Andrew Bennetts  <andrew-pythonlist at puzzling.org> wrote:
> >
> >  - ReST isn't extensible enough; how do you add markup for, e.g. API
> >    references?  How do you plug that into your docutils installation?  You
> >    can't.  (The Lore markup for APIs is 
> >        <code class="API">twisted.internet.reactor.run</code>
> >    )
> 
> You're wrong::
> 
>     :API:`twisted.internet.reactor.run`
> 
> Granted that isn't currently trivial to add, but it's definitely
> possible.

Hmm, interesting.  I thought the :Foo: syntax was for definition lists, not
inline in text?  I guess what I'm asking is, what would be the ReST
equivalent of:

    <p>To start the reactor, use the <code class="API"
    base="twisted.internet">reactor.run</code> method.  To stop it, use
    <code class="API" base="twisted.internet">reactor.stop</code>.</p>

[The optional base attribute provides the fully qualified name to allow the
text to be linked to the right part of the API docs, but doesn't affect the
final display.]

And how would I make the ReST tools link that to my API documentation
generated with pydoc, epydoc, or whatever?

(Btw, don't get me wrong; I *like* ReST for some uses.  I just think it's
better suited to wikis than to reference manuals and the like.  I don't
really like ReST's liberal use of punctuation in large doses.)

-Andrew.





From mal at lemburg.com  Sun Apr 13 09:08:06 2003
From: mal at lemburg.com (M.-A. Lemburg)
Date: Sun, 13 Apr 2003 15:08:06 +0200
Subject: win32all odbc/MS SQLServer/bigint
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEB0@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DDEB0@admin56.narex.com>
Message-ID: <3E996136.3070607@lemburg.com>

Bjorn Pettersen wrote:
>>From: M.-A. Lemburg [mailto:mal at lemburg.com] 
>>
>>Bjorn Pettersen wrote:
> 
> [...]
> 
> 
>>The win32all odbc module fetches BIGINTs as floats. That's
>>where you probably get your corruption from...
> 
> 
> That's what I thought from reading the code, but I didn't want to imply
> it publically in case I was wrong. I'm sure there are areas where loss
> of precision is acceptable, but in business applications there is a
> rather strong reaction when this happens, especially without
> documentation, or warning, and the fact that the Python type in the
> result set is a long not a float...

The odbc module returns a long in case floor(x) == x, but the
interfacing at C level is done using a C double.

>>>I looked at the mxODBC documentation, and it says it's doing 
>>>a conversion [automatic] to string for bigint data... (again 
>>>too slow). I'm pretty flummoxed, the ODBC C calls for 
>>>bigint/__int64 are pretty straight forward(?)
>>
>>They are ? __int64 is not exactly portable across platforms...
> 
> AIX, IRIX, SGI, Linux, SUN, HP, z/OS, OS/390, and Win* all support a 64
> bit integer type (I did the research at my last company, and only for
> the platforms we supported... -- I'm sure there are exceptions, but ODBC
> isn't a lowest common denominator specification by design).

Sure, but __int64 is only available in VC C++ AFAIK. Many
compilers have a "long long" type which could be used, but
then again, how do you know whether the ODBC driver was
compiled with the same C type and layout as the application
using it ?

It's one of those MS things again... the half cooked, "works
for me" kind of attitude.

>>BTW, conversion from strings to Python longs is done in C,
>>so it's not as slow as you may think.
> 
> I haven't looked at the Py_LongFromLONGLONG(sp?) algorithm, so you may
> be right. It was too slow when I wrote our odbc wrapper library in
> C++...

It's called PyLong_FromLongLong(), but only avialable if the
C compiler knows about long longs and only in Python 2.0 and
above.

It is faster than the PyLong_FromString() API which mxODBC
currently uses, but only if the ODBC driver does actually
manage BIGINTs as long longs internally (sending these
as strings over the wire is not uncommon, so you'd win on
the Python side, but lose on the ODBC driver side).

>>>As an added bonus, someone had the pleasure of showing me 
>>>that a simple select/fetchall (i.e. all C code for Python), 
>>>was over three times faster in C#..., and the win32all odbc 
>>>module has crashed python somewhere deep down in sql three 
>>>times in the last two hours (we're expecting this app to 
>>>run for ~4 days...). Looks like it's back to C++ (*sigh*).
>>
>>Have you tried mxODBC on this ? We have many customers running
>>mxODBC under Zope on heavily loaded sites for weeks... so far
>>without any complaints about its stability.
> 
> I will definitely try it later today (your reputation precedes you, but
> upper management is cheap, especially when something doesn't come from
> MS <groan>). My other options are wrapping our db lib as a Python
> extension (mostly done, but not DB API compliant), or recoding in C++
> (should be fast, I've got a working prototype :-)
> 
> Btw., are any of your customers running with 3Tb of collections
> ("someone didn't pay their bill") data, with throughput requirements and
> hard deadlines ("if you're not done by noon you must pay us back
> $50K"/hr) <wink>? (sorry if I come across too harsh, not my intention,
> but after going through the joy of formally proving my program could do
> no harm to either the data or the throughput of the server [oh, and
> being called at 3am on Sunday], my sense of humor seems to have gone
> into hiding -- maybe next week I'll have a beer and laugh about it :-)

I've heard similar things before :-)

Seriously, performance on the Python processing side is usually
not an issue (mxODBC is fast); it's network latency, database and
query optimizations that do matter.

> -- bjorn
> 
> ps: In case you were wondering, we're not doing this on live production
> data because we don't know what we're doing...

-- 
Marc-Andre Lemburg
eGenix.com

Professional Python Software directly from the Source  (#1, Apr 13 2003)
 >>> Python/Zope Products & Consulting ...         http://www.egenix.com/
 >>> mxODBC, mxDateTime, mxTextTools ...        http://python.egenix.com/
________________________________________________________________________
EuroPython 2003, Charleroi, Belgium:                        72 days left





From the_3_project at yahoo.com  Tue Apr 29 08:29:21 2003
From: the_3_project at yahoo.com (Joe Green)
Date: 29 Apr 2003 05:29:21 -0700
Subject: _winreg newbie question
References: <a5b86892.0304282156.208b81b1@posting.google.com> <a5b86892.0304290139.6870fb2b@posting.google.com> <Xns936C72D87CE5Dduncanrcpcouk@127.0.0.1>
Message-ID: <a5b86892.0304290429.5dfcd888@posting.google.com>

Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in message news:<Xns936C72D87CE5Dduncanrcpcouk at 127.0.0.1>...
> the_3_project at yahoo.com (Joe Green) wrote in 
> news:a5b86892.0304290139.6870fb2b at posting.google.com:
> 
> >    if (key is "LOCAL_MACHINE"):
> 
> What do you think this line does? It doesn't compare two strings to see if 
> they are equal.
> 
> Perhaps you meant:
> 
>      if key == 'LOCAL_MACHINE':

YES! That's what I'm trying to do :)

> 
> The 'is' operator checks whether two expressions both result in exactly the 
> same object. This is not something you want to do very often.

Very well, thanks for the hint.

> Usually objects can be equal without being the same object. (You can use 
> 'is' to check whether an object 'is None' but that is because there is 
> guaranteed only to be one None object in the system.)
> 

Makes sense.

> Some other style points (feel free to disagree/ignore):

Oh no, this is very helpful stuff. However, I was trying to minimize
the code and put in a bare minimum to demonstrate the problem -- this
is not really how the real implementation looks.

> 
> Never use 'from <module> import *' unless the module documentation insists 
> you should (_tkinter does). It is likely to come back and bite you later.
> 

Ah.. yes, normally I don't. In cases such as this, I do because I know
there will not be conflicts.

> Get rid of the unnecessary parentheses in your 'if' statements. Python is 
> not C.
> 

They are for my sanity. I do come from a C++ background :) I don't
usually use them unless there are logical and's and or's in the if
statement. Again, this was just for the demonstration of the problem.

> There are several ways to avoid that unsightly list of if...elif 
> statements. Perhaps the easiest would be:
> 
>     def IterKeys(self, key, subkey=None):
>         regkey = getattr(self, key, None)
>         ...
> 
> and then use regkey in place of 'self.reg[mykey]'.
>  

Ah. Now that's something I can incorporate. Let's see how that works.

> Calling sys.exit when you cannot create the class seems a bit drastic. You 

Yeah of course, drastic measures were needed to emphasise :)

> might prefer to throw a more specific exception instead, or even just 

Right. That's what the actual implementation does.

> remove your try...except block and handle the exception at an outer level. 
> As a general point an except block that catches every exception should make 
> sure it does something useful (such as logging or displaying the error that 
> actually occurred).

I do both whenever possible. Did I say I come from a Linux background?
We like them logs :)

Thanks a lot!



From skip at pobox.com  Fri Apr  4 19:28:25 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 4 Apr 2003 18:28:25 -0600
Subject: Socket timeouts for HTTP connections (httplib)?
In-Reply-To: <459f9b68.0304041442.58cf392e@posting.google.com>
References: <9df3d134.0304020550.7e32bf6b@posting.google.com>
        <459f9b68.0304041442.58cf392e@posting.google.com>
Message-ID: <16014.9001.221220.26142@montanaro.dyndns.org>

    Matt> I have recently posted a patch to sf regarding ftplib, which has
    Matt> the same problem (no way to use ne wtimeout functionailty).

As an alternative, can you check out Python from the current CVS and see if
it works for you?  I checked in a patch about a week ago.

Skip




From mis6 at pitt.edu  Mon Apr 28 13:31:52 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 10:31:52 -0700
Subject: Question about accessing class-attributes.
References: <0kifav8dfcbu0t8j4qrrkq6ag7gcg1s062@4ax.com> <Xns9367885737E62duncanrcpcouk@127.0.0.1> <EARpa.120$3M4.2728@news1.tin.it> <Xns93679D1BF4A2Cduncanrcpcouk@127.0.0.1> <2259b0e2.0304241414.675b1e14@posting.google.com>
Message-ID: <2259b0e2.0304280931.69efbab7@posting.google.com>

mis6 at pitt.edu (Michele Simionato) wrote in message news:<2259b0e2.0304241414.675b1e14 at posting.google.com>...
> Duncan Booth <duncan at NOSPAMrcp.co.uk> wrote in message news:<Xns93679D1BF4A2Cduncanrcpcouk at 127.0.0.1>...

> > <snip> what happens if 
> > you want to combine these and get e.g. a Bunch which counts its instances?
> > 
> > I haven't tried it to see what happens, maybe you just need to indirect 
> > type.__new__ calls through super, and make a new submetaclass with the old 
> > metaclasses as bases, but I'd like to know if there are issues here or if 
> > I'm just worrying needlessly.
> 
> I did quite a lot of experiments on multiple inheritance of metaclasses,
> and as far as I can tell it works quite well. You should not worry, except
> in the case in which you want to multiple inherits from a metaclass and
> a normal class by redefining __new__. Then I have found a nasty subtility
> that made me loose an afternoon. You can find the solution if you Google
> on the newsgroup under "another super wart" (read only my last post, the
> first was wrong). This only happens in very complicate hierarchies, anyway.

I take back what I said. I have just downloaded Python 2.3b1 and discovered
the the super bug I was talking about has been fixed!
Now the following script will work as expected:

"Super bug fixed in 2.3b"

class Counter(object):
    counter=0 
    def __new__(cls,*args,**kw):
        cls.counter+=1; print "Creating instance #%s of %s" % (
            cls.counter,cls)
        return super(Counter,cls).__new__(cls,*args,**kw)

class MetaCounter(Counter,type):
    pass 

class C(Counter): #=> Creating instance #1 of <class '__main__.MetaCounter'>
    __metaclass__=MetaCounter

C() #=> Creating instance #1 of <class '__main__.C'>

assert super(Counter,C).__new__ is object.__new__ # True in 2.3
# assert super(Counter,C).__new__ is type.__new__  # True in 2.2

This works in Python 2.3b, where the superclass of Counter with respect to
the MRO of C is object; does not work in previous versions, where super
computes the superclass of Counter with respect to the MRO of
type(C)=MetaCounter, which is type and the script would raise a
TypeError: type.__new__(C): C is not a subtype of type

Now super retrieves correctly even properties therefore I would say it
works pretty nice. At this point, there is little reason to wonder about 
using multiple inheritance and metaclasses ;)



                                Michele



From rob at zoism.org  Sat Apr  5 11:33:35 2003
From: rob at zoism.org (Rob Brown-Bayliss)
Date: 06 Apr 2003 04:33:35 +1200
Subject: segfault with imp.load_module()
In-Reply-To: <000701c2fb58$c8065ff0$8201fd0a@hallows>
References: <1049534591.20683.22.camel@musicbox>
	 <000701c2fb58$c8065ff0$8201fd0a@hallows>
Message-ID: <1049560434.22975.0.camel@musicbox>

On Sat, 2003-04-05 at 21:49, John Abel wrote:
> Could this not have to do with your use of file as a variable?

I dont see how that should be a problem?

-- 

*  Rob Brown-Bayliss
*  =================	
*      zoism.org






From mcfletch at rogers.com  Wed Apr  9 19:10:32 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Wed, 09 Apr 2003 19:10:32 -0400
Subject: Python application deployment?
In-Reply-To: <14bd1399.0304091426.1b08e005@posting.google.com>
References: <14bd1399.0304091426.1b08e005@posting.google.com>
Message-ID: <3E94A868.9030509@rogers.com>

There's FAQ's for this, but here's the short summary:

    Use py2exe or McMillan Installer, both generate stand-alone apps
    with just those parts of Python you need

    Package the results up however you like (e.g. I use the free Inno
    Setup package) and distribute.

Results w/ py2exe tend to be fairly good in my experience.  You wind up 
with "normal" seeming download sizes for non-trivial applications (e.g. 
4 or 5 MB for a small wxPython GUI application (wxPython is fairly 
monolithic when built, so it takes a lot of space) with ZODB database 
support and various win32-specific functions).  For non-GUI apps that's 
going to be considerably smaller.

HTH,
Mike

kkennedy wrote:

>Is there a way to deploy the Python "runtime" with your Python
>scripts?  I am only concerned with Windows (since most Linux distros
>would already have it installed).  I would like it to be smaller than
>the 7 MB download available from python.org, and be able to be
>installed transparent to the user.  In other words, it would be nice
>to be able to simply unzip the files needed to run my scripts without
>the user having to download and install the full Python windows setup,
>and then download and install my Python scripts seperately.
>
>Are there any tools to make this doable by the common programmer?
>  
>

-- 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From fuka at fuxoft.cz  Wed Apr 23 09:44:57 2003
From: fuka at fuxoft.cz (Frantisek Fuka)
Date: Wed, 23 Apr 2003 15:44:57 +0200
Subject: Testing if string contains a substring
Message-ID: <b865dq$28gi$1@ns.felk.cvut.cz>

I want a simple test to see if str contains any occurences of substr. I 
am currently using this:

import string
if string.find(str,substr) + 1:
     ...do something...

But it somehow looks "ugly" to me. Is there more "standard" way to test 
this?




From martin at v.loewis.de  Tue Apr 22 18:19:41 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 23 Apr 2003 00:19:41 +0200
Subject: True/False
References: <1051036194.697.177.camel@lothlorien> <16037.36122.922342.405340@montanaro.dyndns.org> <1051038682.702.191.camel@lothlorien> <mailman.1051039416.29975.python-list@python.org>
Message-ID: <m33ckatbsy.fsf@mira.informatik.hu-berlin.de>

Skip Montanaro <skip at pobox.com> writes:

> It might.  There was a thread on python-dev or c.l.py about this particular
> item within the last couple months.  I don't recall if any conclusions were
> reached.

The conclusion is that assignments to True/False cannot be banned *yet*.
Over time, it will happen (although the time might well be 10 years).

Regards,
Martin



From pinard at iro.umontreal.ca  Sat Apr  5 17:21:33 2003
From: pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois_Pinard?=)
Date: 05 Apr 2003 17:21:33 -0500
Subject: Input from a file as a command line argument
In-Reply-To: <oqfzow7g3a.fsf@titan.progiciels-bpi.ca>
References: <3E8EDBDA.7020701@pbezone.net>
	<pan.2003.04.05.16.31.02.87447.878@thomas-guettler.de>
	<oqfzow7g3a.fsf@titan.progiciels-bpi.ca>
Message-ID: <oq7ka87fgy.fsf@titan.progiciels-bpi.ca>

[Fran?ois Pinard]

>    import sys
>    line = sys.stdin.readline():
>    while line:
>        print line
>        line = sys.stdin.readline()

Before you all tease me with this, I know the first `:' is spurious! :-)

-- 
Fran?ois Pinard   http://www.iro.umontreal.ca/~pinard




From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr 24 12:23:46 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 24 Apr 2003 16:23:46 GMT
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <m3y92cvolo.fsf@mira.informatik.hu-berlin.de> <k7dwsp9l.fsf@morpheus.demon.co.uk>
Message-ID: <slrnbag3si.ek.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <k7dwsp9l.fsf at morpheus.demon.co.uk>, Paul Moore wrote:
> martin at v.loewis.de (Martin v. L?wis) writes:
> 
>> Microsoft has added API to support output of arbitrary Unicode
>> characters on the console, regardless of the code page
>> (WriteConsoleW).
> 
> Thanks. I'll investigate this.
> 
>> This only works if the console uses a Unicode font (Lucida Sans
>> Unicode).
> 
> How do I get this? I have Lucida Sans Unicode installed, but in the
> console properties screen, the only font I can choose is Lucida
> Console...

Lucida Console is just a dumb name for Lucida Sans Typewriter, the
non-proportional version of Lucida Sans.



From peter.maas at mplusr.de  Thu Apr 10 08:58:36 2003
From: peter.maas at mplusr.de (Peter Maas)
Date: Thu, 10 Apr 2003 14:58:36 +0200
Subject: ASTVisitor
Message-ID: <b73pks$t0t$1@swifty.westend.com>

I am trying to employ ASTVisitor to convert a python expression to
postfix form. Following the documentation, I created my own visitor
class with methods called visit<class>, where <class> represents a
syntax node like And, Or, ...:

class cVisitor:
     def printNode(self, node, level):
         ....
     visitExpression = printNode
     visitOr = printNode
     visitAnd = printNode
     ...

When using the visitor

cv = cVisitor()
av = compiler.visitor.ASTVisitor()
print av.preorder(tree, cv)

only the Expression node was handled, i.e. the code above didn't
descend down the tree. According to the docs the return value
of visit... should decide wether to visit the node's subtree but
my code didn't bother.

The code of the ASTVisitor class (methods preorder, dispatch)
shows no indication of tree traversal which would explain my
failure.

Have I missed something or have I found a dark corner of python
with some cobwebs? :) I can do the recursion myself but I thought
this to be the responsibility of ASTVisitor. I would appreciate
hints to working example code. Until now I haven't found any on
the web.

Mit freundlichen Gruessen,

Peter Maas

-------------------------------------------------------------------
Peter Maas, M+R Infosysteme, D-52070 Aachen, Hubert-Wienen-Str. 24
Tel +49-241-93878-0 Fax +49-241-93878-20 eMail peter.maas at mplusr.de
-------------------------------------------------------------------




From news1NOSPAM at awakened-lands.net  Sun Apr 20 06:41:16 2003
From: news1NOSPAM at awakened-lands.net (Michael Steuer)
Date: Sun, 20 Apr 2003 20:41:16 +1000
Subject: Creating a stand alone Python Service (win32)
Message-ID: <PA$jhlyBDHA.3088@lime.oak-park.us>

Hi.

I want to use a Python service I wrote as a stand alone service, meaning
that it should be possible to install the service on a client without having
to install Python first. As I did with other Python (non service)
applications before, I used py2exe to convert the service into a stand alone
application. This "stand alone" service runs fine on my development machine,
but it won't work on a computer without Python installed (respectively the
win32 extensions for Python), as it needs the PythonService.exe to be
registered on the client.

I tried to register the PythonService.exe on the client, but (after adding
the DLLs it needs) I got the following message when running
"PythonService.exe /register" (enclosed in *'s to separate it from the rest
of this message):

*********************************
Registering the Python Service Manager ...
'import site' failed; use -v for traceback
*********************************

Is there a way to get "PythonService.exe" to run on a client without
installing Python ?

If that is not possible, how could I change my service so that it doesn't
need PythonService.exe anymore ?

Thanks for your help in advance.

--?
Cheers,

Mike.

PS: remove NOSPAM from my reply address to get my actual email address.





From sholden at holdenweb.com  Mon Apr 14 10:07:33 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Mon, 14 Apr 2003 14:07:33 GMT
Subject: newbie: strange python problem
References: <pan.2003.04.10.12.27.06.600319@somedomain.com> <3E956A20.E3321A5E@engcorp.com> <pan.2003.04.10.13.07.59.235430@somedomain.com>
Message-ID: <Fgzma.83541$It5.37164@news2.central.cox.net>

"maxx" <nospam at somedomain.com> wrote in message
news:pan.2003.04.10.13.07.59.235430 at somedomain.com...
> On Thu, 10 Apr 2003 08:57:04 -0400, Peter Hansen wrote:
>
> > Check the output of "ls -la /" manually... the first line does
> > *not* have more than two items in it...  if you were to change
> > the first line above to this, it would work, though I suspect
> > there is another option to "ls" which would also suppress
> > that extra output.
> >
> >> for line in os.popen('ls -la /').readlines()[1:]:
> >
> > -Peter
>
> Right! A simple mistake, but those are the ones you make taking your
> first steps in a new programming language.
>
Sure. But the subject line should really have been "Python: strange newbie
problem" :-)

Welcome to c.l.py, maxx.

regards
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/
Did you miss PyCon DC 2003?            Would you come to PyCOn DC 2004?






From theller at python.net  Thu Apr 17 09:20:47 2003
From: theller at python.net (Thomas Heller)
Date: 17 Apr 2003 15:20:47 +0200
Subject: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com>
Message-ID: <y929gt00.fsf@python.net>

"Harald Schneider" <h_schneider at marketmix.com> writes:

> Hi,
> 
> I created a NT service under with py2exe under Python 2.1
> The service ran without any problems.
> 
> Then I upgraded to Python 2.2 and py2exe for 2.2 and now I have a big
> problem:
> All functions that access registry keys from within the service fail.
> When I run my application outside the Service class, everything works well.
> 
> Any idea?

Whithout any further information, one can only guess.
Permission problems? Did you not only move from Python 2.1 to 2.2,
but also from NT to 2000?

What error do you get?

Thomas



From aleax at aleax.it  Tue Apr 15 08:40:15 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 12:40:15 GMT
Subject: instance creatiion as an "event"
References: <mailman.1050095779.7490.python-list@python.org> <pan.2003.04.11.23.20.04.65410@webone.com.au> <mailman.1050117319.5102.python-list@python.org> <b7gt8r$em0$1@panix1.panix.com>
Message-ID: <P4Tma.22981$T34.691702@news2.tin.it>

Aahz wrote:

> In article <mailman.1050117319.5102.python-list at python.org>,
> Steven Taschuk  <staschuk at telusplanet.net> wrote:
>>Quoth Simon Burton:
>>> On Fri, 11 Apr 2003 17:15:32 -0400, Arthur wrote:
>>>>
>>>> The only thing I seem to be missing is some kind of "hook" to trigger a
>>>> general update cycle of all existing class instances upon the creation
>>>> of a new class instance at the interactive prompt
>>> 
>>> class Member:
>>> members = []
>>> def __init__(self):
>>> for m in Member.members:
>>> m.notify(self)
>>> Member.members.append(self)
>>> def notify(self,other):
>>> print self,"a new member:",other
>>
>>Though you will probably want to use weak references in
>>Member.members, unless you intend that all instances of the class
>>live as long as the class does.
> 
> Blech.  It's essentially impossible to delete a class as long as
> instances of it exist.

Of course (the instances are keeping the class alive through their
__class__ attributes), but what does that have to do with Steven's
suggestion, making it worth a "blech"?  If Member.members is coded
as above, then you have circular reference loops involving the
class and each instance so just about everything becomes sort of
"immortal" -- with weak references in Member.members, no circular
reference loops, no problems (or, am I missing something...?).


Alex




From pc451 at yahoo.com  Wed Apr 16 13:30:09 2003
From: pc451 at yahoo.com (Peter Clark)
Date: 16 Apr 2003 10:30:09 -0700
Subject: Degree symbol (UTF-8 > ASCII)
Message-ID: <f7199550.0304160930.6cfeca65@posting.google.com>

I'm working with a xml document which doesn't include an encoding, so
it defaults to UTF-8. Of course, all of the text is ASCII, and likely
to remain so. I would like to insert the degree symbol (chr(176)), but
because this is outside the bounds (chr(128) is the limit), Python
raises an XML error. What's the simplest way of getting including an
unadorned degree symbol? Again, it's not necessary to preserve the
UTF-8 encoding, but I'm not quite certain as to how to tell Python
that the XML document is plain ASCII (or I guess in the case of the
degree symbol, latin-1).
    Thanks,
    :Peter



From staschuk at telusplanet.net  Sat Apr 26 15:10:49 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 26 Apr 2003 13:10:49 -0600
Subject: Another itertool function?
In-Reply-To: <slrnbalgos.am0.mlh@furu.idi.ntnu.no>; from mlh@furu.idi.ntnu.no on Sat, Apr 26, 2003 at 05:34:20PM +0000
References: <slrnbalgos.am0.mlh@furu.idi.ntnu.no>
Message-ID: <20030426131049.E30093@tibia.amotlpaa.bogus>

Quoth Magnus Lie Hetland:
  [...]
> def weave(*iterables):
>     iterables = map(iter, iterables)
>     while iterables:
>         for it in iterables[:]:
>             try:
>                 yield it.next()
>             except StopIteration:
>                 iterables.remove(it)
> 
> Another possible name (which is a bit more unwieldy) would be
> 'roundrobin'. Another one still is iweave (to keep the naming
> convention, even though here there is no clash).

I wrote an equivalent function a while back, too; I called it
"interleave".

(I can't remember what I needed it for, oddly.)

-- 
Steven Taschuk                                                   w_w
staschuk at telusplanet.net                                      ,-= U
                                                               1 1




From jjl at pobox.com  Tue Apr 22 14:58:47 2003
From: jjl at pobox.com (John J. Lee)
Date: 22 Apr 2003 19:58:47 +0100
Subject: another python style question -- copying a data file
References: <zTepa.127177$Vzu.30273@news02.bloor.is.net.cable.rogers.com>
Message-ID: <87znmi4avs.fsf@pobox.com>

guy at obstruction-no-spam.com (Guy Middleton) writes:
[...]
> I want to copy a data file, but if the file doesn't already exist, I don't
> care.  I do care about any other errors, such as wrong permissions.
> 
> I have the following code, is this a typical Python way to do this?
> 
>     try:
>         os.stat(DATAFILE)
>         os.system("cp %s %s.new" % (DATAFILE, DATAFILE))
>     except OSError, arg:
>         # ENOENT is ok, we didn't do the copy
>         # anything else is a problem
>         if arg.errno != errno.ENOENT:
>             raise

I'd probably 'look before I leap' (simply because it's simple and the
first thing that comes to mind), assuming the race condition is not
important:

if os.path.exists(DATAFILE):
    shutil.copy(DATAFILE, DATAFILE+".new")  # portable, & no subshell!

You might also want to use os.path.splitext(DATAFILE)[0] + ".new"
instead of DATAFILE+".new", to strip off any old extension first.


John



From duncan at NOSPAMrcp.co.uk  Tue Apr 29 08:30:55 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 29 Apr 2003 12:30:55 +0000 (UTC)
Subject: Web programming and a different "type" problem
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com> <slrnbarhfb.67t.jon+usenet@snowy.squish.net> <o0i1o-i07.ln1@nb2.stroeder.com>
Message-ID: <Xns936C889D8B676duncanrcpcouk@127.0.0.1>

Michael Str?der <michael at stroeder.com> wrote in
news:o0i1o-i07.ln1 at nb2.stroeder.com: 

> Jon Ribbens wrote:
>> e.g. <input type="checkbox" name="foo*">
>                                        ^^^
> Is the * guaranteed to work?
> 
> http://www.w3.org/TR/html401/interact/forms.html#h-17.4
> 
> name        CDATA
> 
> http://www.w3.org/TR/html401/types.html#type-cdata
> 
> ------------------------------ snip ------------------------------
> ID and NAME tokens must begin with a letter ([A-Za-z]) and may be
> followed by any number of letters, digits ([0-9]), hyphens ("-"),
> underscores ("_"), colons (":"), and periods (".").
> ------------------------------ snip ------------------------------

Read it again. In an input tag, the type of the 'name' attribute is 
'CDATA'. The paragraph you quote is about tokens of type 'ID' or 'NAME' 
which is not relevant here.

The documents you refer to are actually slightly inconsistent. According to 
the text, the 'id' attribute has the type 'NAME', but the DTD actually uses 
the type 'ID' for the 'id' attribute. The 'NAME' type is used only for the 
attributes 'name' and 'http-equiv' in a 'META' tag.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From gregory.p.green at boeing.com  Mon Apr  7 11:53:49 2003
From: gregory.p.green at boeing.com (Green, Gregory P)
Date: Mon, 7 Apr 2003 08:53:49 -0700
Subject: Problem setting COM property using win32com
Message-ID: <E1B4F80EB080194BBAD3092392F511B3DADEEA@XCH-NW-13.nw.nos.boeing.com>

I am creating the object like so:

   attrs = win32com.client.Dispatch("SmRecList.SmRecordList")

I hadn't noticed the dynamic. I guess I can try to force early binding.

-----Original Message-----
From: Paul Prescod [mailto:paul at prescod.net]
Sent: Monday, April 07, 2003 12:17 AM
To: Green, Gregory P
Cc: python-list at python.org
Subject: Re: Problem setting COM property using win32com


I don't claim to be a win32com expert but I notice two things in your 
message that made me suspicious.

Greg Green wrote:
> I have a commercial com api that I'm trying to use with python. I am
> trying to set a property of the com object. The property is called
> Value, and I know enough to use the SetValue function. But that
> doesn't work.
> 
> Here is the code:
>     attrs.SetValue("CN_ID", 0, "0000005")
> 
> Here is the error:
> Traceback (most recent call last):
>   File "example.py", line 49, in ?
>     prmidattrs.SetValue("CN_ID", 0, "0000005")
>   File "C:\PYTHON22\lib\site-packages\win32com\client\dynamic.py", line 438, in __getattr__
>     raise AttributeError, "%s.%s" % (self._username_, attr)
> AttributeError: SmRecList.SmRecordList.SetValue
> 
> If I look at the makepy generated module, I see the method.

Why are you getting an exception in "dynamic.py" if you are using "makepy"?

How are you creating the prmidattrs object?

  Paul Prescod





From koenig at v-i-t.de  Tue Apr 22 13:04:56 2003
From: koenig at v-i-t.de (Mirko Koenig)
Date: Tue, 22 Apr 2003 19:04:56 +0200
Subject: pygettext but no msgfmt
Message-ID: <b83snm$1ms$1@news.tu-darmstadt.de>

Hi

I tried to use gettext for internationalizing but i can't find the msgfmt
tool.

I have debian woody with python2.1 and python 2.2 installed.
In the debian archive the file msgfmt.py isn't listed either.

How can i convert the po files then?

Mirko Koenig



From aleax at aleax.it  Sun Apr 13 12:52:19 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 16:52:19 GMT
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org> <435d40e0.0304101653.1eb11a6c@posting.google.com> <mailman.1050087439.11600.python-list@python.org> <435d40e0.0304130212.5659bd4d@posting.google.com>
Message-ID: <7Bgma.8749$LB6.237526@news1.tin.it>

Chris wrote:
   ...
> Thanks Inyeol. Be nice if they gave a small example of using finditer.

How small?  E.g., here's how you could rewrite (a simpler version of)
findall using finditer (warning, untested):

def findall(are, instring):
    return [mo.group(0) for mo in are.finditer(instring)]

Too small?  Then perhaps you could mention a larger one...?


Alex





From LogiplexSoftware at earthlink.net  Thu Apr 10 13:15:50 2003
From: LogiplexSoftware at earthlink.net (Cliff Wells)
Date: 10 Apr 2003 10:15:50 -0700
Subject: Python compared toJava
In-Reply-To: <3E95A2A7.238BFA6C@engcorp.com>
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
	 <3E95A2A7.238BFA6C@engcorp.com>
Message-ID: <1049994950.13005.54.camel@software1.logiplex.internal>

On Thu, 2003-04-10 at 09:58, Peter Hansen wrote:
> Blaktyger wrote:
> > 
> > I recently read that the Python interpreter is like the Java Virtual Machine.
> > Is that true?
> 
> Yes!  In fact, it's exactly the same thing!  We just use a different name
> for it to avoid the wrath of Sun Microsoftsystems, and because we kind of like
> snakes.
> 
> Note: see http://www.catb.org/~esr/faqs/smart-questions.html before posting
> again please. 

Peter, I hate to chastise you, but you are making unfounded assumptions
here.  The OP said nothing about wanting to ask smart questions.  You
just *assumed* he did.  How insensitive.

-- 
Cliff Wells, Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308  (800) 735-0555 x308





From nhodgson at eb2.net.au  Mon Apr 14 20:00:08 2003
From: nhodgson at eb2.net.au (Neil Hodgson)
Date: Tue, 15 Apr 2003 00:00:08 GMT
Subject: Using property in classic class may not work
Message-ID: <cYHma.5800$hF.46142@news-server.bigpond.net.au>

    The "property" feature in Python 2.2 appears very helpful to me and 
seems to work on both classic and new-style classes. However, pychecker 
gives this error message:

DemonTerm.py:563: Using property (frame) in classic class DemonFrame may 
not work

    The class in question is derived from a wxWindows class, so I can't 
make it derive just from object. The "What's New in Python 2.2" document 
only mentions property wrt new-style clases. Can I depend on property 
working for classic classes? Is there a workaround such as deriving from 
object as well?

    Neil




From tjreedy at udel.edu  Fri Apr 11 09:33:15 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Fri, 11 Apr 2003 09:33:15 -0400
Subject: collecting data from file
References: <3E9671A3.2050004@sympatico.ca> <3E96B6BC.E9DBE74F@engcorp.com>
Message-ID: <MM2dneyWRoYWXwujXTWcow@comcast.com>

"Peter Hansen" <peter at engcorp.com> wrote in message
news:3E96B6BC.E9DBE74F at engcorp.com...
> Mustafa Celik wrote:
> >
> > I want to scroll thru a file;
> >    * find lines that match a string (e.g. HELLO) on the 2nd column
> >    * add up the 4th column (an integer) on each matching line, say
the
> > variable is TOTAL
> >    * subtract the 4th column from TOTAL if another is string (e.g.
BYE)
> > is hit on 2nd column of a line
> >
> > Any tips?
>
> Yes, post an example showing what you intend.  The instructions are
> decent, but implementable only by making a few assumptions about
what
> you mean.

In particular, how are the 'columns' specified.  Your examples imply
that you mean variable-length data fields rather than character
columns.  If so, are they separated by  <space>, <tab>, a mixture of
the two, <comma>, or something else.

>  Also, are there any error conditions to be handled, or do
> you guarantee the input is always perfect (e.g., always at least
four
> columns, and so on)?

Missing data is the great bugaboo of data analysis.  Ignoring that
possibility

sum = 0
for line in file('mydata'):
  fields = line.split() # assume whitespace
  if field[1] == 'HELLO':
    sum += field[3]
  elif field[1] == 'BYE'
    sum += field[3]

Terry J. Reedy





From andrew-pythonlist at puzzling.org  Mon Apr 21 21:39:47 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Tue, 22 Apr 2003 11:39:47 +1000
Subject: Twisted 1.0.4
In-Reply-To: <006901c3081c$8240d7e0$0200a8c0@b>
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b>
Message-ID: <20030422013947.GB7666@frobozz.local>

On Mon, Apr 21, 2003 at 06:41:42PM +0300, Oktay Safak wrote:
> I see that Buildbot is GPL'ed. I am not very knowledgable in these licensing
> stuff but I wonder whether this would be a problem for Twisted users.

Buildbot is a seperate project to Twisted, so its licence has no relevance
to Twisted and Twisted's users.  Similarly, Twisted has some C extensions
which are typically built with gcc, but gcc's licence (GPL) also has no
relevance to Twisted.

That said, Buildbot and CVSToys (http://purl.net/net/CVSToys) have been
extremely useful to the Twisted team, and it's worth taking a look at them.
The fact that they will require you to install Twisted, thus furthering its
world domination, is just incidental, really ;)

-Andrew.





From cnetzer at mail.arc.nasa.gov  Tue Apr  8 20:06:38 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 08 Apr 2003 17:06:38 -0700
Subject: usage of assert and -O
In-Reply-To: <005e01c2fe26$608868c0$0ba587d9@jaws>
References: <005e01c2fe26$608868c0$0ba587d9@jaws>
Message-ID: <1049846798.601.38.camel@sayge.arc.nasa.gov>

On Tue, 2003-04-08 at 16:27, Matthew Russell wrote:
> Hi all,
>  
> I understand that when you invoke the interpreter with the -O flag
> assert statments are not executed - is this 100% correct? (as i belive
> it to be)

Yes, it is correct.

>  
> and if so, shouldn't the use of assert (if not assert itself) be
> depricated?

No.  Why?  I must assume you think the assertions should always be
evaluated, which is definitely NOT the case.  It is *trivial* to define
your own assert functions that does not get optimized away, if that is
what you want.

ie.

def always_assert( predicate ):
    if not predicate:
        raise AssertionError
 
There IS a use for an assert that can be removed during optimization.  I
often use asserts for internal consistency checking that cannot be
easily replaced with unit tests.  Once I have everything working, the
asserts (almost by definition) become redundant, and I don't necessarily
still want or need them to be evaluated in production code.

Unit tests are great at telling you WHAT functions have broken in your
code.  asserts placed judiciously in the code can help you zero in on
WHY things have broken.

I tend to define my own asserts that I don't want to go away (ie. they
become part of the API).  One such function is an argument_assert() that
I use to make sure that arguments given to a function are valid (ie.
greater than or equal to zero when appropriate, or within a predefined
range, etc.)

-- 
Bay Area Python Interest Group - http://www.baypiggies.net/

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From martin at v.loewis.de  Tue Apr 29 14:52:13 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 29 Apr 2003 20:52:13 +0200
Subject: 2.3b1 sys.path anomaly on W2K
References: <b8mfgc$ll8$1$830fa795@news.demon.co.uk>
Message-ID: <m3wuhd6srm.fsf@mira.informatik.hu-berlin.de>

"Stephen D Evans" <sde at recombinant.demon.co.uk> writes:

> Where does the 'C:\\WINNT\\System32\\python23.zip' come from in sys.path?
> Does anyone else get this? Is it meant to be there?

>From the zipfile import mechanism, yes, and yes. See

http://www.python.org/peps/pep-0273.html

Regards,
Martin



From matthias.kramm at stud.tu-muenchen.de  Mon Apr  7 04:29:50 2003
From: matthias.kramm at stud.tu-muenchen.de (Matthias Kramm)
Date: Mon, 7 Apr 2003 10:29:50 +0200
Subject: Help: "undefined reference to `PyArg_ParseTuple'" when execising 'Extending Python with C or C++ '
In-Reply-To: <mailman.1049636769.32638.python-list@python.org>
References: <mailman.1049636769.32638.python-list@python.org>
Message-ID: <b6rctu$7ea$04$1@news.t-online.com>

On Sun, Apr 06, 2003 at 09:45:28PM +0800, Lu Jian Xiong wrote:
> [root at localhost user0]# gcc -I /usr/include/python2.2/ -L /usr/lib/python2.2/c
> onfig/ -l python2.2 spammodule.c
> /tmp/ccIl3dpq.o: In function `spam_system':
> /tmp/ccIl3dpq.o(.text+0x16): undefined reference to `PyArg_ParseTuple'
> /tmp/ccIl3dpq.o(.text+0x47): undefined reference to `Py_BuildValue'
> collect2: ld returned 1 exit status

Try putting the "-l python2.2" part after "spammodule.c", so that
the above command reads
    gcc -I/usr/include/python2.2/ -L/usr/lib/python2.2/config/ spammodule.c -lpython2.2

Btw. what do you need the "-L/usr/lib/python2.2/config/" for?

Greetings

Matthias





From mis6 at pitt.edu  Tue Apr 29 17:21:56 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 29 Apr 2003 14:21:56 -0700
Subject: Question about accessing class-attributes.
References: <mailman.1051637056.20383.python-list@python.org>
Message-ID: <2259b0e2.0304291321.18e37af6@posting.google.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message news:<mailman.1051637056.20383.python-list at python.org>...
> > From: Alex Martelli [mailto:aleax at aleax.it] 
>  [...]
> Here's the case I'm confused about:
> 
> >>> class meta(type):
> ...   def   getattr  (cls, name):
> ...     if name == '  len  ':
> ...        print "meta.  getattr  ('  len  ')"
> ...        return lambda: 42
> ...     else:
> ...        print 'meta.  getattr  ', name
> ...        return name
> ...
> >>> class S(object):
> ...     metaclass   = meta
> ...
> >>> S.  len  ()
> meta.  getattr  ('  len  ')
> 42
> >>> len(S)
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> TypeError: len() of unsized object
> >>>
> 
> I was told that special method "foo(x, arg)" was implemented as
> "type(x).  foo  (x, arg)", which doesn't seem to be the case always...
> Compare:
> 
> >>> class meta(type):
> ...   def   len  (cls):
> ...     return 42
> ...
> >>> class S(object):
> ...     metaclass   = meta
> ...
> >>> S.  len  ()
>  42
> >>> len(S)
>  42
> >>>
> 
> So, it looks like it's looking up   len   in the metaclass, but not
> falling back on   getattr   when it isn't there? 

I am sure Alex is going to answer, but since you asked me a related
question in another thread ...

Finally, now I see where is the source of the confusion:

len(S) fails because meta.__len__ is NOT defined in
you first example. The point is that __getattr__ works on instances of
meta, NOT on meta itself, therefore S.__len__ is recognized thanks
to __getattr__, but NON meta.__len__
In order to recognize meta.__len__ you must define __getattr__ on the
metaclass of meta, i.e. the meta-metaclass of S:

class metameta(type):
    def __getattr__(cls, name):
        if name == '__len__':
            print "metameta.__getattr__('  len  ')"
            return lambda cls: 42
        else:
            print 'metameta.__getattr__', name
            return name

class meta(type):
    __metaclass__=metameta
    
class S(object):
    __metaclass__ = meta

print type(S).__len__(S) 

with output

metameta.__getattr__('  len  ')
42


I must say, however, that len(S) is not *literally* type(S).__len__(S):

print len(S)

Traceback (most recent call last):
  File "<stdin>", line 18, in ?
TypeError: len() of unsized object

I think Python looks if __len__ is defined in the meta-class and
if not, it does raise an exception. It does not know that there is
a __getattr__ in the meta-metaclass, unless you invoke the metaclass literally
as in type(S).__len__(S).

I did not understood your point in the other thread, I didn't imagine
you where thinking about meta-metaclasses. As you see, there are 
situations where they may be useful, i.e. in the customization of 
metaclasses, but most people don't think at this level yet ;)

Cheers,

                       Michele



From spam at mattbergin.co.uk  Sun Apr 13 12:51:21 2003
From: spam at mattbergin.co.uk (Matt Bergin)
Date: Sun, 13 Apr 2003 17:51:21 +0100
Subject: Reference Tracking
In-Reply-To: <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk>
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk>
Message-ID: <b7c4i8$f6o$1@wisteria.csv.warwick.ac.uk>

Matt Bergin wrote:
> I don't think there's any risk of reference cycles. I should have been 
> more clear about what the objects are. I'm trying to write a class to 
> wrap MySQLdb and hand out database cursors to worker threads.

I should stress this is for self-education. I know there are already 
modules which do this, but I don't like to use code that I don't 
understand properly.

-- 
Matt Bergin




From peter at engcorp.com  Fri Apr 25 19:05:26 2003
From: peter at engcorp.com (Peter Hansen)
Date: Fri, 25 Apr 2003 19:05:26 -0400
Subject: Divided by a common language (was: OO + Python Developers Wanted)
References: <mailman.1051183297.29110.python-list@python.org> <yyyznmezekm.fsf@adams.patriot.net> <vais01o44vtc0d@corp.supernews.com> <yyyvfx2zdmo.fsf@adams.patriot.net> <mailman.1051298731.5020.python-list@python.org>
Message-ID: <3EA9BF36.E2685CC9@engcorp.com>

Steven Taschuk wrote:
> 
> Quoth Nick Vargish:
> > claird at lairds.com (Cameron Laird) writes:
> >
> > > >Which "City" would that be?
> > > <URL: http://www.cityoflondon.gov.uk/ >
> >
> > Oh, then he should have specified, "The City", and I would have known
> > he meant Mother London. Here in Amerika, we have a number of urban
> > locales that consider themselves "the City"...
> 
> Typical American arrogance!  Sheesh!
> 
> "The City" is obviously Toronto.  (Unless you're in the west.)

Ahem(*)...  You know better.  Toronto is _never_ to be mentioned
except in close association with the adjective "world-class".  

Mel will be by to have a talk with you later.

-Peter

(*) That was a simple clearing of the throat.  No SARS here.  Nosirree.
Move right along... nothing to <cough> see here folks.



From alanmk at hotmail.com  Wed Apr 30 06:35:32 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Wed, 30 Apr 2003 11:35:32 +0100
Subject: HTTP application streams a response to the client gradually - how   
 deal with it?
References: <Xns936CB5697A6D4cpl19ghumspamgourmet@62.153.159.134> <3EAEC97C.1EADEBF7@hotmail.com> <Xns936D814A89Fcpl19ghumspamgourmet@62.153.159.134> <3EAF956F.CF1D9A26@hotmail.com> <Xns936D7A9DFE6E5cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <3EAFA6F4.3BD5D4FC@hotmail.com>

Harald Massa wrote:

> The exact problem is:
> 
> the data has to be processed one chunk at a time. Without the response
> being complete.
> 
> Example: The chat server sends:
> 
> Bart says: nice to see you all here
> Miriam says: yes, I also enjoy the talk
> Lisa says: cool evening, yeah
> ...
> 
> the response channel stays open. All the time. When I wait for it to
> finish, so that I can start processing the Response-Objekt - no good.

Ouch, that's kind of messy, isn't it?

One hack you could use to make it work is to take
httplib.HTTPconnection, find the code that reads chunks from the socket,
and add in a callback to a function of yours. Not pretty, but it would
work.

> And while the responses get in from the server, I also have to do
> additional POST requests
> 
> POST user=Icke says="i also think it's a nice evening"

As you mentioned, that can be solved with threads.

> But: it seems that the server does not
> finish his response. So until httplib.read() has a timeout ...

In 2.3, sockets can have timeouts. But I'm not sure how that would
benefit you (unless you were operating single threaded)?

So, how about this for a proposed solution (bodge-job)

1. Have one thread that opens a (hacked) httplib.HTTPconnection to the
server, and always keeps it open, continually reading chunks and passing
them to a function call of yours.

2. Another thread listens for keyboard input, and transmits it to the
server over a separate httplib.HTTPconnection that is opened and closed
as required.

>> functionality you need, and (mostly) bug free, having been used by
>> thousands of people over the years. Rolling your own HTTP protocol
>> would be an enormous amount of work, due to the wide range of weird and
>> wonderful ways in which HTTP is used and abused.

> Yeah ... and I think I have just stumbled into a very abusive one.

You can say that again.

> Thanks for all your feedback... maybe there is an easier way?

The bodge-job I mentioned above is the simplest I can think of.

regards,

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From nikolai.NOJUNK at micon.no  Fri Apr  4 16:50:47 2003
From: nikolai.NOJUNK at micon.no (Nikolai Kirsebom)
Date: Fri, 04 Apr 2003 23:50:47 +0200
Subject: Invalid format character in string
Message-ID: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com>

I have a file (HTML) containing a set of %(name)s entries.  When I run
a string format statement of the content of the file towards a
dictionary object, the values for the corresponding keys are replaced
correctly.  However, since the text is actually a HTML file, it may
contain a lot of characters with % (typically cell width definitions
of tables).  If these are contained in the file, the string format
statement complains about illegal format character.  I've
(temporarily) solved it by replacing the sequences '%;' and '%>' with
something else and then restore the sequences after the string format
statement.  It works, but I'm sure there is a more elegant (and not
hard-coded) solution.

Any help is appreciated.
Nikolai Kirsebom




From tjreedy at udel.edu  Sun Apr 27 01:45:45 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Sun, 27 Apr 2003 01:45:45 -0400
Subject: Looping around a simple list of tuples
References: <3EAB7213.4030209@magma.ca>
Message-ID: <ILqdnQj4BsM_8zajXTWcpQ@comcast.com>

"Martin d'Anjou" <point14 at magma.ca> wrote in message
news:3EAB7213.4030209 at magma.ca...
> I have a list of tuples, each tuple is a (scalar , list)
>
> a=[("r1",[1,2,3]),("r2",[8,9,10]),("r5",[12,14,16])]
>
> for el in a:
>     print el
>     for r,l in el:
>        print "R "+`r`
>        print "L "+`l`
>
> ('r1', [1, 2, 3])
> R 'r'
> L '1'
> Traceback (innermost last):
>    File "<stdin>", line 3, in ?
> ValueError: unpack list of wrong size
>
> In my mind, the 'el' implicit variable is a tuple, and the for r,l
loop should
> just work but it does not. How come?

len(el) == 2, so inner loop (tries to) execute twice
first time, r,l = 'rl' (len == 2), so r= 'r', l='l', as printed
second time, r,l = [1,2,3] (len=3) does not work for reason stated.

perhaps you really meant

for r, l in a:
     print "R "+`r`
     print "L "+`l`
# produces
R 'r1'
L [1, 2, 3]
R 'r2'
L [8, 9, 10]
R 'r5'
L [12, 14, 16]

Terry J. Reedy





From bjorn at algonet.se  Sun Apr 13 05:53:38 2003
From: bjorn at algonet.se (=?iso-8859-1?q?Bj=F6rn=20Fahller?=)
Date: Sun, 13 Apr 2003 11:53:38 +0200
Subject: extending/embedding with triggers
Message-ID: <200304131153.38833.bjorn@algonet.se>

Hello all,

Sorry for probably wanting to do something rather backwards...

I want to use Python as the scripting language in my program, but I'm facing 
two problems that I don't understand how I can solve with the public 
extention/embedding API. Perhaps it's doable inside Python itself.

I would like to express, in my python code, something like:

def func1():
  ...

def func2():
  ...

trig1 = mymodule.trigger(condition1)
trig2 = mymodule.trigger(condition2)
trigger.condition([(trig1,func1),(trig2,func2)])
... do something here, with many calls to mymodule

trigger.disarm([trig1,trig2])

The idea is that mymodule will some time call a function that will make the 
conditions held by the triggers become true. When they do, I want the 
functions to be executed. I cannot use a normal "if" statement, since I do 
not know when the trigger will become true, it can be anytime between 
"trigger.condition" and "trigger.disarm". Typically the functions are small 
and will, for example, acknowledge the condition to the module (but I want to 
control how it will be acknowledged in the script.)

The second problem is similar in that it too deals with triggers, but it also 
adds parallellism. In this case, most of the execution is driven by triggers.

def path1():
 trigger.awit_one_shot(mymodule.trigger(condition1))
 mymodule.somefunc()
 trigger.awit_one_shot(mymodule.trigger(condition2))
 mymodule.otherfunc()

def path2():
  trigger.await_one_shot(mymodule.trigger(condition3))
  mymodule.alternativefunc()


mymodule.do_something()
path.parallell([path1,path2])
mymodule.finish_something()

The thing here is that the my module will do all the work in both path1 and 
path2, between "do_something" and "finish_something", but the order between 
them is undefined and can be path1 before path2, the reversed, or 
interleaved. With a good generic problem to the first, this could be solved 
with threading, but I'd rather not need to make "mymodule" thread-aware.
   _
/Bjorn.





From gh at ghaering.de  Fri Apr 25 11:28:04 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Fri, 25 Apr 2003 17:28:04 +0200
Subject: pythong referencing/aliasing
In-Reply-To: <ylcqa.9116$h%2.963225@read1.cgocable.net>
References: <ylcqa.9116$h%2.963225@read1.cgocable.net>
Message-ID: <3EA95404.5050105@ghaering.de>

Howe wrote:
> How does Python handles variable references/aliases?
> 
> Example:
> 
> a = 1

A new int object with the value 1 is created. The variable a references 
this int object.

> b = (REF)a

In Python: b = a

b now refers to the same object a does.

> b = 2

A new int object with the value 2 is created. b now references this int 
object instead of the one with the value 1.

Objects not referenced any more (like this one) will eventually be 
garbage-collected. [1]

> print a
> 
> In this case, since "b" is just a reference to "a", changes to "b"
> automatically change "a".
> 
> Does python have this ability?

Sure. This, however, only works with mutable objects. ints (and strings, 
and tuples) are immutable. An example for a mutable object would be a 
list (or a class instance). Try this, for example:

C:\src\python\dist\src>python
'import site' failed; use -v for traceback
Python 2.3a2+ (#27, Apr 23 2003, 21:13:49)
[GCC 3.2.2 (mingw special 20030208-1)] on mingw32_nt-5.11
Type "help", "copyright", "credits" or "license" for more information.
 >>> a = [3]	# create a list with one element, 3
 >>> b = [4]	# create a list with one element, 4
 >>> a, b	# display both
([3], [4])
 >>> b = a	# b references the object a refers to; nothing else 
references the [4] list, so it will be garbage collected, eventually
 >>> a,b		# display both
([3], [3])
 >>> a.append(4)	# append the item 4 to the list a (and b) refers to
 >>> a,b		# print both
([3, 4], [3, 4])
 >>> ^Z

-- Gerhard

[1] ints are a bad excample, because they're special-cased in the 
CPython implementation.




From syver-en+usenet at online.no  Mon Apr  7 10:21:04 2003
From: syver-en+usenet at online.no (Syver Enstad)
Date: 07 Apr 2003 16:21:04 +0200
Subject: Debugging embedded python
References: <u4r5bgyfx.fsf@online.no> <b6q6s1$1c9$00$1@news.t-online.com>
Message-ID: <ur88e4cdr.fsf@online.no>

"Martin v. L?wis" <martin at v.loewis.de> writes:

> Syver Enstad wrote:
> > I have the python interpreter embedded in a daemon/service (this is
> > win2k so it's a service). I'd really like to be able to debug the
> > python code that is running. I was thinking of popping up a console
> > window to be able to run pdb, how does one go about it to accomplish
> 
> > this? Are there any code doing similar things or articles.
> 
> I recommend that you find a non-daemon mode of operation, and support
> debugging there. This has tradition on Unix (atleast): many daemons
> support a do-not-fork-into-background-do-not-fork-many-processes mode,
> to simplify debugging.

Yes, I have found out how to make the process to behave more like an
ordinary python process, but I need to pop up a console window or
something similar that let's me inspect the python code. I first
thought about using PyCrust for this, but the service I am running in
fails to import wxPython.wx. I'll have a look at Python Programming on
Win32. I think it had something on creating a python console in a
running program.

I've also had a look at the HapDebugger. It seems pretty nice, I am
looking into how to embedd the client side of it into an multithreaded
application that calls python functions instead of running a single
script. It doesn't seem that undoable, I'll see where it leads me.

-- 

Vennlig hilsen 

Syver Enstad



From adalke at mindspring.com  Wed Apr 23 22:43:41 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 23 Apr 2003 20:43:41 -0600
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org> <GXupa.83382$T34.2079161@news2.tin.it> <5jadeg285s.fsf@xoxy.net>
Message-ID: <b87iqe$e9v$1@slb4.atl.mindspring.net>

Charlie Reiman:
> Are there any dangers in something like this:
>
> a = [1, infimum, infimum]
> a.sort()
>
> Should the __cmp__ methods return 0 for equality (identity, in this
> case)?

As Alex pointed out, it should.

However, I recall a post by Tim Peters some years ago saying
that he fixed the sort code so it would finish, even in the face of
absurdities like this.

But I'm too lazy to find the post, or even test the code.  ;)

                    Andrew
                    dalke at dalkescientific.com





From maschio_77 at hotmail.com  Wed Apr  2 04:53:35 2003
From: maschio_77 at hotmail.com (Federico)
Date: Wed, 2 Apr 2003 11:53:35 +0200
Subject: underline a letter in a button to use "ALT+key" shortcut with Tkinter
Message-ID: <b6ebu8$bp2$1@lacerta.tiscalinet.it>

With Tkinter how can I underline a letter in the text in a button and set an
"ALT+Key" shortcut for that button? is there some library I can use?
Thanks





From mike at bindkey.com  Mon Apr  7 14:48:18 2003
From: mike at bindkey.com (Mike Rovner)
Date: Mon, 7 Apr 2003 11:48:18 -0700
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org> <XSXja.346441$3D1.186570@sccrnsc01>
Message-ID: <b6sh4l$i0h$1@main.gmane.org>

"Bryan" <belred1 at yahoo.com> wrote in message
news:XSXja.346441$3D1.186570 at sccrnsc01...
> mike,
>
> you mentioned small c libs you use pyrex.  what about small but complex
data
> types?  we have a structures that contain pointers to structures that

Sure Pyrex has its limits. My picture was general. In each particular case I
weight pro and contra.

But it's especially nice to wrap C with Pyrex because it VERY easy to adapt
interface to pythonic way.
Say I had a lib with (void*) handle parameter to pass almost every data
structure and a bunch of different real structures.
With Pyrex an ability to redefine was a plus - hard things like swig and c++
will not do in that case (nicely I mean).
Complex (struct inside struct) data was no problem.
I also use a script to convert an .h file to Pyrex - it was simple 15-liner
__to run once__:

#!/usr/bin/env perl
/^\s*{\s*$/ && next;
/^\s*} T.*$/ && next;

s/;//;
s/^\t//;

s|/\*|\#|;
s|^ *\*|\#|;
s| *\*/$||;

s/^\s*typedef +struct +SSX([A-Z_]+)/cdef struct TSX$1:/; $n=$1 if $1;
s/^\s*struct/\ncdef struct $n:/;
s/^\s*}/\t$n/;

print;

WBR,
Mike







From martin at v.loewis.de  Sun Apr 27 04:18:24 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 27 Apr 2003 10:18:24 +0200
Subject: Python 2.3b1 + Tkinter on Solaris 2.6
References: <mailman.1051406621.31763.python-list@python.org>
Message-ID: <m3d6j8tktr.fsf@mira.informatik.hu-berlin.de>

Jeff Bauer <jbauer at rubic.com> writes:

> Prior to filing a bug report, I wanted to check
> if there's anything obvious I'm overlooking with
> getting tk8.4.2 working with Python2.3b1 on
> Solaris 2.6.

You are overlooking something, but it is not obvious, and it is likely
not a bug in Python.

> *** WARNING: renaming "_tkinter" since importing it failed:
>   ld.so.1: ./python: fatal: relocation error: file
>   /usr/local/lib/libtk8.4.so: symbol __eprintf: referenced symbol not
> found

__eprintf is traditionally a symbol from libgcc.a, and gets into
object files whenever the source code has assert().

It appears that your Tk installation was made with a gcc version that
had assert implemented with __eprintf. Could it be that you have
changed gcc versions since?  If so, you need to recompile Tk (or,
alternatively, link python and/or _tkinter.so with the gcc version
that was used to link Tk)

Regards,
Martin



From Yu.Wang at synopsys.com  Thu Apr 17 02:07:36 2003
From: Yu.Wang at synopsys.com (Yu Wang)
Date: Thu, 17 Apr 2003 14:07:36 +0800
Subject: get pid from popen3 or get redirection in execl 
In-Reply-To: <3E9E0FBA.6050503@indy.rr.com>
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com> <bDSma.22753$T34.687158@news2.tin.it> <slrnb9o8s8.iea.news@datatailors.com> <mailman.1050427771.31620.python-list@python.org> <3E9E0FBA.6050503@indy.rr.com>
Message-ID: <3E9E44A8.3080203@synopsys.com>

Hi,all
   Does "execl", or whatever this family, support redirection of output?
   Say, I have a script: ( on my Linux7.2,C shell environment,with 
python2.2.1)

command="somecommand"
arg="-i somefile > output file"
execl(command,arg)

what's wrong with this?

And,
    how could I get child pid from "popen3(somecommand)".

Thanks to you.
Regards.





From ckea25d02 at sneakemail.com  Wed Apr 16 00:05:07 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 16:05:07 +1200
Subject: do python's nifty indentation rules spell the death of   one-liners?
In-Reply-To: <mailman.1050452786.8192.python-list@python.org>
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org>
Message-ID: <b7ik4p$1d2t1$1@ID-169208.news.dfncis.de>

Gerhard H?ring wrote:
> Try:
> 
> python -c """
> print 2
> for i in (1, 4):
>     print i
> """
> 
> It seems to work under a bash script at least.

I just tried this in a Makefile, and it seems to work:

all:
	python -c '\
	for i in range(10):\
	 print i'

(There's a tab at the beginning of each line, and
*no* extra whitespace after that apart from what
Python needs for indentation.)

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From bokr at oz.net  Tue Apr  8 21:26:51 2003
From: bokr at oz.net (Bengt Richter)
Date: 9 Apr 2003 01:26:51 GMT
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org> <kauu6b.f2d1.ln@boundary.tundraware.com> <v960rni1esvo94@corp.supernews.com>
Message-ID: <b6vssr$jek$0@216.39.172.122>

On Tue, 08 Apr 2003 17:14:31 -0000, claird at lairds.com (Cameron Laird) wrote:

>In article <kauu6b.f2d1.ln at boundary.tundraware.com>,
>Tim Daneliuk  <tundra at tundraware.com> wrote:
>			.
>			.
>			.
>>As a side note, this question was ultimately motivated by an entirely
>>different sort of "pure math" question: Does (integer) 1 = 1.0?
>>Engineers, physicists, and computer scientists all would
>>say "no" because of the precision problem with the latter number.
IMO there is no "precision problem" with the unique abstract number
normally associated with "1.0" or "1" or "ord('\x01')" or "I" or "one"
or "ein" or "uno" or "+1e0" or ... ;-)

>>I am unclear, though, on whether a mathematician understands
>>1.0 to mean 1.0000000... (a single number) or the _neighborhood_
>>around 1 with one digit precision.
>			.
>			.
>			.
>I'm a (former) mathematician.  I've always taken "1.0"
>to be a local, linguistic act, requiring me to infer
>the speaker's context.  "1.0" means different things in
>the mouths of different people.
>
Well put. Of course, any sign whose inferred purpose was to
get your attention can be substituted for "1.0" in the above,
generalizing from speech to include other communicative acts ;-)

>The open interval (0.95, 1.05) is one of the plausible
>interpretations.  It's not among the first half-dozen
>likely to come to my mind, though; maybe in the top 
>ten ...  

Half open [.095, 1.05) or [1.0, 2.0) seem more useful to me
if the intent is to imply intervals, but a single exact number
is the interpretation I prefer (and is in any case necessary
as a reference for the definition of the intervals).

Which brings me to my point: IMO the distinction between
abstract entities vs. their various concrete representations
is crucial to communicating clearly about ... well, just about
anything.

E.g., concretely, you can use your hands to represent floating
point .01 to 99. very simply: Use your fingers to represent
the digits in BCD, and expose one thumb to indicate the decimal
point position. Or if you like binary, you can take advantage
of your opposable thumb to put one between selected fingers
for .00000001 through 10000000. base 2. Or you can use various
squiggly marks made visible to the communicatee. Or you can tap
out a pattern on someone's knee, say in morse code, or do the same
with audible beeps, or hand someone a tray with several bowls
of pebbles, etc., etc. Not to mention intermediate hidden (to
human senses) physical representations that play a role in
storage and transmission of numeric information, such as in computers.

The point is, a single abstract entity can have any number of
representations. "Floating point number" can thus be interpreted
to indicate a selected state of one of the many possible concrete
floating point representation mechanisms, together with the
exact abstract numerical value represented.

If you want to refer to an exact interval related to the exact value,
that requires additional information, which you can supply implicitly
or explicitly by context (cf Cameron's comment above). IMO a single
exact abstract value typically retains it primacy as an interpretation
of the representation, whatever additional definitions extend relative
to that through some special convention of interpretation and implicit
information.

Another point that may become clearer if one thinks about abstractions
vs representations is that computers necessarily implement concrete
transformations of _representations_ -- i.e., they do not directly implement
abstract mathematical operations. Pouring the contents of two bowls of pebbles
into a single empty bowl can represent integer addition, but when a bowl
overflows, that is a special effect of the _implementation_ of addition, not of
addition as an abstract integer operation.

The special effects encountered in transforming floating point representations
to implement abstract mathematical operations are also mostly properties of
the representation mechanisms, not of the abstract math, but they happen at
many points _throughout_ the interval of real numbers that contains the domain
of operands and range of results with exact floating point representations. The
special effects don't just happen at the extremes, as with integers.
This makes it trickier to think about.

The special effects derive from the necessity of choosing a representation
of an exact number as a result where a representation of the exact mathematical
result may not be available. Often it _is_ available, and the result _is_ an exact
representation of the mathematically correct result, but often, too, that's not possible
for a particular representation mechanism.

I.e., the representations of two numbers in floating point can correspond exactly
to intended abstract operands, but it is possible that the implementation of addition
will transform the FPU state to yield a resulting FPU representation (which necessarily
has a specific exact abstract numerical interpretation), but whose abstract value is not
identical to the expected mathematically correct value in the abstract realm.

[OT] BTW, when thinking about such realities as wars and foreign populations, or immediate
surroundings, including one's body, etc., it is also interesting to note that there are
representation issues involved. I.e., (IMO) one's subjective experience of reality is an experience
of one's state of mind -- however that arose -- and thus is an experience of a _representation_
of reality, not of reality itself[1]. And perception of reality has a context, like perception of "1.0".

People are so full of context ;-) Otherwise how could they disagree so violently about reality?
(Of course the above reflects my internal representations of reality and results from mechanisms
for transforming those representations ;-)

[1] Except insofar as the state of mind itself is a part of reality, of course.

Regards,
Bengt Richter



From blunck at gst.com  Tue Apr 29 10:38:34 2003
From: blunck at gst.com (Christopher Blunck)
Date: 29 Apr 2003 07:38:34 -0700
Subject: Where did you learn to unit test?  How did you learn?
Message-ID: <1147e466.0304290638.287687f6@posting.google.com>

Hi all-

This is my second generic question posted to c.l.p (the first was 'How
many of you are Extreme Programmers?') but it seems like a somewhat
appropriate discussion for the group.

I've noticed in my development that Python programmers (in general)
write more unit tests than their java counterparts.  If I download a
project off of jakarta.apache.org, it very rarely has unit tests. 
However, when I download a python module, it frequently has unit
tests.

I'm in a position where I'm trying to spread the testing bug to my
team.  It's remarkable to watch folks who were once viciously opposed
to unit testing adopt it  as being absolutely critical during
application development.  While this is encouraging, it is far from
the common case, and I'd like to improve that.

What advice can you offer on spreading the testing bug around?  

What success stories do you have?  

What practices did you follow?  

Did you beat people over the head with it until they wised up?  

Were you softer in your approach?  

What worked?  

What were the long term implications?

How quickly did unit testing catch on?

Or did unit testing never catch on?



Any insight you can offer would be appreciated.


-c



From aleax at aleax.it  Sun Apr 27 04:27:09 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 27 Apr 2003 08:27:09 GMT
Subject: please comment on technologies
References: <slrnbalplq.254.luc@trillian.dont-panic.info> <HqCqa.8931$3M4.241359@news1.tin.it> <slrnbalv50.254.luc@trillian.dont-panic.info>
Message-ID: <xvMqa.9914$3M4.267925@news1.tin.it>

luc wastiaux wrote:

> In article <HqCqa.8931$3M4.241359 at news1.tin.it>, Alex Martelli wrote:
>>> The project would be developped by a group of 4 to 6 CS students
>>> (including me), over a three months timespan (but not full time, we have
>>> to go to classes and other stuff)
>> 
>> The most valuable thing you'll learn will be how to structure your
>> work.  I heartily recommend incremental development (most crucial
>> features first), test-driven design, pair programming, and a healthy
>> diffidence towards "big design up front" - you may get the temptation
>> of designing in some big chunks of infrastructure preparing for future
>> enhancements... that may never come.
> 
> In short, you recommend following extreme programming guidelines, did I
> understand your advice correctly ?

Not necessarily ALL of them.  "Customer on site and part of the team" may 
be particularly hard here for example -- if your "customer" is a professor
he may not WANT to be part of your team for undestandable reasons.  And
while "40 hours weeks" is a wonderful rule, it's hard to stick to when one
is a student in some engineering discipline -- what with the courseload
and all, I remember my time as a student as one of EIGHTY-hours weeks,
month in month out (looking back a quarter century, I can't believe I ALSO
managed to have something of a social life -- guess I must not have slept
for a few years on end;-).

But I'm definitely recommending some variation on *Agile* development
methodologies, and particularly recommending *AGAINST* the "waterfall"
model of development that you may have been taught and, in my opinion
and experience, is a disaster.  In "waterfall", you're supposed to do
all analysis first; then, all design; then, all the coding; and then
all the testing.  It just plain don't work.  It's better to plan for
incremental, iterative development from the start.  Do SOME analysis,
identify an absolutely minimal set of features you MUST have at all
costs, i.e., in your case, a set such that you'll fail your course if
you don't have AT LEAST those features.  Put all the "oh yeah and it
would be nice to have THIS too" on one side for now -- index cards do
work quite nicely for that (XP uses them, but so do many other approaches,
particularly agile ones).

Having identified an absolutely-minimal, will-die-without-these set of
features, then do SOME design -- as little as you can get away with:
i.e., design the minimal infrastructure you MUST have to implement the
features in the minimal set.  This is where some (at least preliminary)
technology choices MUST be made, too; for the purpose you'll probably
need "spikes", where two of you go off on the side for a bit (could be
hours or a couple of days) and try out some technological possibility
to ascertain it's well understood and stable and rich enough for you.

Avoid "big design up front" -- don't make the infrastructure rich and
powerful to support "future features" that may never come to be -- that's
an EASY trap to fall into.  What you DON'T want is to have the deadline
come and by that time have an analysis that supports 4 times as many
features as you could sensibly implement in the given time, a halfway-
done infrastructure that might support all of those 4-times-as-many
features plus as many again if they were ever dreamed up, and about 1/10
the working application code and about 1/100 the unit and acceptance
tests you should have.  I'm being light-hearted, but this IS what happens
in some software projects, particularly but not exclusively ones where the
participants lack actual in-the-trenches experience.

Divide the work up into SMALL increments, and make sure every one of
you is reasonably familiar with all the subsystems as well as the
overall architecture.  Pair programming, with ever-shifting pairs,
helps here.  One thing I've found helps things move forwards, even
though XP doesn't support it, is to have each of you take "primary
architectural responsibility" for some of the subsystems and modules
you have identified -- and don't forget to nominate one of you as
"overall architect".  This means that, while consensus is seeked, when
consensus cannot be reached the person who wears the architect hat
for the specific subsystem has to make the decision and get the whole
project moving again -- decisions can later be reversed, at the cost
of some rework, but I've found that trying to do EVERYthing by consensus
can slow things down excessively (a variant of "analysis paralysis").

*DON'T SKIMP ON TESTS*.  Once you THINK you have a design for some
subsystem, feature, module, class, function -- *WRITE TESTS FOR IT*
before you start coding it.  The tests will fail at first -- of course
they must, as the subsystem/etc ain't there yet.  But they DRIVE your
coding, acting as "executable docs and specifications" to some extent.

They're better than discursive docs/specs because they're executable --
thus necessarily unambiguous (not necessarily COMPLETE -- so, adding
SOME discursive material in terms of comments and docstrings is good
too -- but don't be surprised if NON-executable stuff fails to get
maintained and updated as your understanding grows... sadly, it almost
inevitably happens all of the time).

So -- I'm not specifically recommending XP for your needs (in some
areas, such as the advice to have "chief architects", my advice runs
counter to XP's) -- but I _AM_ recommending an Agile and deliberately
DESIGNED process for your work, AND avoidance of "waterfall".


>>> Spyce today and it looks like I could use it for the web interface.
>> 
>> Sure, that's one possibility.  Personally I'd choose Webware, which
>> offers an amount of flexibility I appreciate, but Spyce's OK too.
> 
> I will have a look a webware. I was looking forward to using spyce because
> it features <?php style tags to embed code into HTML templates, does
> webware have that ?

Webware does let you embed Python in HTML -- even though more often 
than not this is an inferior approach, Webware's designers are well
aware of the unhealthy fascination the approach holds for people
coming from PHP or ASP, and therefore support it.  But that's just
part of Webware's flexibility.  Servlets and templating are the ways
you should REALLY be using, IMHO.  Anyway, Webware supports them
all (AND lets you connect your code to the webserver with commendable
flexibility too -- so you can use one model when you're developing
and unit-testing and another for more extensive tests, for example,
without changing your code, just the "admin" procedures that you
use for starting it up).


>>> Then I have this idea about using XSL for templating, the idea is not
>>> only to separate the data from the representation, but also allow
>>> programmers to "leech off" the search engine (it's easier to do this if
>>> the output is XML).
>> 
>> The ability for other programmers to send "queries" to your search
>> engine and get XML output is good.  However, personally, I would avoid
>> XSL (unless you guys are already well familiar with it -- or NEED to
>> get very familiar with it anyway to pursue your studies).  It's simpler
>> to have separate URLs, or a form field requesting XML or HTML output,
>> and just use different templates with the same contents -- doing your
>> templating in Cheetah (or preppy/Spyce/whatever).
> 
> Ok, I thought XSL was the definitive answer to all templating needs, but
> you don't seem very excited about it ?

I guess I shouldn't put my name to this -- it's not a "respectable"
thing to say, and might inveigle me in flamewars I'd rather avoid -- 
but I think XSL is an approach that makes things harder for you, and
is therefore less pragmatically useful, than ad-hoc templating based
on mostly-procedural approaches.  It's worth learning it (particularly
if you have no other experience with non-procedural languages!) because
it expands your mind -- but I personally wouldn't want to have "mastery
of XSL" as a do-or-die prereq for any crucial project I'm responsible
for.  Perhaps I'm just not looking at it the right way.

> thanks a lot for all your comments and advice!

You're welcome!


Alex




From rmunn at pobox.com  Fri Apr  4 12:29:57 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 04 Apr 2003 17:29:57 GMT
Subject: split encloser
References: <mailman.1049417184.1536.python-list@python.org>
Message-ID: <slrnb8qr8o.iu8.rmunn@localhost.localdomain>

Jason Tiller <jtiller at sjm.com> wrote:
> Hi, "Aussie", :)
> 
> On 3 Apr 2003 aussie2010 at yahoo.com wrote:
> 
>> string.split() takes a delimiter and works fine as long as the
>> delimiter isn't part of the data fields. But frequently they are.
>> e.g. 'John Doe,135 South Main St.,#122, Springfield, Iowa' or
>>       ' so long goodbye see ya'
> 
>> Because the fields can contain the delimiter in some cases, an
>> encloser is usually used (typically "") to handle those fields.
> 
>> The above strings would be written:
>>
>> 'John Doe,"135 South Main St., #122", Springfield, Iowa'
>>    and
>> '"so long" goodbye "see ya"'
> 
>> I don't understand regular expressions but I was wondering if anyone
>> that did knew of a way to get re.split() to handle "enclosers" as
>> used above.
> 
> Hmm.  I am not yet a knowledgable user of Python's regex features, but
> I *do* know Perl's pretty well.  With Perl, you might split up your
> fields like this:

[snip regex-based solution]

My rule of thumb regarding regular expressions is that if you can't grok
the *entire* regex in fifteen seconds or less, you're almost certainly
better off with some other solution. Split the regex, use string
methods, anything -- just don't use any regex so complicated that it
takes you more than fifteen seconds to understand it. I'm sure Jason
Tiller's solution is a clever one, but I took one look at that regex and
said, "Too hard -- there's got to be a better way."

In this case, I believe that better way is to use the csv module. It
will be in the standard library in Python 2.3; meanwhile, you can
download and use the csv module written by Object Craft:

    http://www.object-craft.com.au/projects/csv/

As a parting shot, here's a quote from Jamie Zawinski:

    Some people, when confronted with a problem, think "I know,
    I'll use regular expressions." Now they have two problems.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From howard at eegsoftware.com  Tue Apr  1 01:20:42 2003
From: howard at eegsoftware.com (Howard Lightstone)
Date: Tue, 01 Apr 2003 06:20:42 GMT
Subject: using a private embedded python
References: <HpRha.17229$OV.116887@rwcrnsc54> <b693js$jos$1@zeppelin.rz.uni-potsdam.de> <REXha.316269$S_4.271878@rwcrnsc53> <b6aktn$3br9c$1@ID-169208.news.dfncis.de>
Message-ID: <Xns934FE35301466howardeegsoftwarecom@66.75.162.198>

"Greg Ewing (using news.cis.dfn.de)" <ckea25d02 at sneakemail.com> wrote in
news:b6aktn$3br9c$1 at ID-169208.news.dfncis.de: 

> Bryan wrote:
>> my
>> concern is what you mentioned...,  that other company products will
>> want to use python or the python scripts i'm executing will want a
>> module i haven't frozen.   so this option unfortunately seems too
>> restrictive. 
> 
> There are tools for Windows that will package up an application
> together with its own private interpreter into a self-unpacking
> archive. Usually they only include the modules used by that
> application, however. If your application is intended to be
> scripted in Python by the end user, you'd want to include all of
> the standard library. I don't know whether these tools have an
> option for that, possibly they do, you'd have to investigate.
> 
[snip]
 
> > can you have two separate
> > versions of 2.2.2 on the machine at the same time?  on windows? on
> > unix? 
> 
> On Unix, yes, certainly, as long as you take care to install
> them in different directories.
> 
> On Windows, you can certainly have *different* version numbers
> installed system-wide simultaneously, but I don't know about
> different copies of the *same* version. It must be possible
> to have multiple "private" ones, since the abovementioned tools
> manage it somehow.
> 
Well, I have an embedded implementation (of 2.2.2) on the same Windows 
machine as my vanilla 2.2.2 install.  I 'installed' a copy of 2.2.2 in a 
subdirectory of my application (just like the release!).  To start my app, 
I have a small program which sets up the "Path"s and exec's my main 
application which initializes a Python interpreter instance linked to the 
normal .DLLs/libraries present down there somewhere in my application tree.  
The small program is necessary on Windows because the Windoze loader tries 
to find all the linked DLLs at load time.... it's too late to muck with 
paths at that point.  All the libraries, demos, and docs are there for use, 
just not in the "normal" places.   Pmw.Blt even works!

The only fly in the ointment is PythonWin and it's siblings.  I NEED them 
but they make significant use of registry entries.  For my release package, 
I monitored the registry changes when installing into my application 
directory and made a cut/paste registry updater that I run as part of the 
install process.  Oh well......



From cnetzer at mail.arc.nasa.gov  Tue Apr 15 22:27:41 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 15 Apr 2003 19:27:41 -0700
Subject: Big-O notation (was: Re: Python and Schools)
In-Reply-To: <b7id3q$3s4$1@news.hccnet.nl>
References: <qGqla.61$Tj7.91655@news.uswest.net>
	 <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
	 <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
	 <v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
	 <PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
	 <3E9C6C62.8010808@rogers.com>
	 <mailman.1050441947.6447.python-list@python.org>
	 <b7id3q$3s4$1@news.hccnet.nl>
Message-ID: <1050460061.11984.67.camel@sayge.arc.nasa.gov>

On Tue, 2003-04-15 at 18:51, Anton Vredegoor wrote:

> Although it's important to not set too severe limitations I have the
> impression this salesman is underestimating the problem.

Oops. I never was cut out for sales.  :)  Thanks for correcting me.

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From usenet at microtonal.co.uk  Fri Apr 11 06:50:20 2003
From: usenet at microtonal.co.uk (Graham Breed)
Date: 11 Apr 2003 03:50:20 -0700
Subject: BEGINNER: converting integer to short month name?
References: <OF47A94545.30289BFA-ON85256D03.006E4C80-85256D03.006FA823@seminole.fl.us> <mailman.1049921355.15625.python-list@python.org> <b73066$evv$1@slb9.atl.mindspring.net> <3E950823.DE77C48D@alcyone.com>
Message-ID: <1d22f12e.0304110250.278ebb3c@posting.google.com>

I get

Python 2.2 (#1, Dec 31 2001, 15:21:18)
[GCC 2.95.3-5 (cygwin special)] on cygwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import calendar
>>> calendar.month_abbr[0]
Segmentation fault (core dumped)



From ckea25d02 at sneakemail.com  Sun Apr  6 23:50:08 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 15:50:08 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
Message-ID: <b6qs04$7eqi1$1@ID-169208.news.dfncis.de>

Carl Banks wrote:
> As an example, take the section on buffers.  First function is
> "current-buffer".  
> Next three functions are "save-current-buffer", "with-current-buffer",
> and "with-temp-buffer".  These are actually not functions, but special
> forms and macros.

In a truly Pythonic design, I think I would do away with
the notion of a current-buffer as a dynamically scoped
global variable. Instead, the buffer to be operated on
would be passed as a parameter to anything which does
things to a buffer (or they would be methods of Buffer).

You may need to have a current-buffer and associated
functions for elisp compatibility, but they should be
implemented on top of a more Pythonic layer.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From aleax at aleax.it  Sat Apr 12 14:38:23 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 18:38:23 GMT
Subject: Macros in Python? (PEP 310)
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org> <b770t8$2b8$00$1@news.t-online.com> <b771jb$ac6$03$1@news.t-online.com> <a8b7f07a.0304111546.3e5a5636@posting.google.com> <wJRla.3135$LB6.88227@news1.tin.it> <b79jii$njr$06$1@news.t-online.com>
Message-ID: <z2Zla.5338$LB6.149344@news1.tin.it>

Dominic wrote:

>> def wrapfunc(func, pre=mkprint("hi"), post=mkprint("bye")):
>>     def wrapped(*args, **kwds):
>>         pre()
>>         try: return func(*args, **kwds)
>>         finally: post()
>>     return wrapped
> 
> Ups, I am surprised that wrapped still has access
> to func I thought that this would only work
> with Python's lambda. My fault.

Python's lambda gives *NO* extra power compared
with Python's named functions (except the ability
to keep functions unnamed, of course;-).  And that
isn't your fault, but lambda's:-).

> The pre/post calls are only executed once
> when the file is loaded so this should be
> handled with care...

To be precise: the *default values* for pre
and post are computed exactly once in this
approach.  The CALLS to pre and post are in
fact executed each time wrapped is executed.

> However I do not need to pass pre/post as
> parameters, so it does not matter very much.

OK, you can then hardcode them inside wrapped
rather than have them as parameters to wrapfunc
with default values.

 
> The only drawback is that the code needs
> to be separated into a fuction unless the
> generator "hack" is used. So wrapfunc seems
> to be the optimum so far.

Not sure what "generator hack" you're considering.
A generator is a function, so how would using a
generator save the need to "separate into a
function" the code to be wrapped?  Not sure what
you mean by "separating" anyway -- you do need to
*NAME* the code, then pass the name you've chosen
to give -- and must take a modicum of care with
respect to bindings of local variables, which
become free variables (that a function can't
rebind in Python...) -- but basically the NAMING is 
really the only "overhead" that would be avoidable
if Python had unnamed blocks (the scoping probably
would remain even WITH hypothetical unnamed blocks).


Alex




From gregory.p.green at boeing.com  Tue Apr  8 20:39:04 2003
From: gregory.p.green at boeing.com (Green, Gregory P)
Date: Tue, 8 Apr 2003 17:39:04 -0700
Subject: Problem setting COM property using win32com
Message-ID: <E1B4F80EB080194BBAD3092392F511B3DADEFF@XCH-NW-13.nw.nos.boeing.com>

The suggestion of replacing default_interface with the class name worked. At least I didn't get an error. I immediately ran into the same sort of problem (AttributeNotFound) with another object of another class. If I can solve that the same way, then I can verify that it worked for sure.

-----Original Message-----
From: Paul Prescod [mailto:paul at prescod.net]
Sent: Monday, April 07, 2003 3:42 PM
To: Green, Gregory P; python-list at python.org
Subject: Re: Problem setting COM property using win32com


Bare in mind that I'm not a COM expert. We should probably be talking 
about this on the python-win32 list. But I see something strange in your 
makepy code...

Green, Gregory P wrote:
> 
> # This CoClass is known by the name 'SmRecList.SmRecordList'
> class SmRecordList(CoClassBaseClass): # A CoClass
> 	# An object used for storing data. A collection of headers and records, each composed of a collection of nodes.  In turn, each node can contain information of various types.
> 	CLSID = pythoncom.MakeIID("{C85E7014-8B4F-11D1-8E20-00A02498EA3C}")
> 	coclass_sources = [
> 	]
> 	coclass_interfaces = [
> 		'{C85E7013-8B4F-11D1-8E20-00A02498EA3C}',
> 		'{C85E7015-8B4F-11D1-8E20-00A02498EA3C}',
> 	]
> 	default_interface = '{C85E7013-8B4F-11D1-8E20-00A02498EA3C}'

In my makepy'd classes, default_interface is never a string.

class ResourceManager(CoClassBaseClass): # A CoClass
	# ResourceManager Class
...	default_interface = _ResourceManager

# This CoClass is known by the name 'XMetaL.Global'
class Global(CoClassBaseClass): # A CoClass
....
	default_interface = _Global

That's probably why you're getting the "cannot call string" problem. 
Look at the __init__ that's causing a problem. It probably looks like this:

class CoClassBaseClass:
	def __init__(self, oobj=None):
		if oobj is None: oobj = pythoncom.new(self.CLSID)
		self.__dict__["_dispobj_"]=self.default_interface(oobj)

I don't know how your default_interface got to be a string (perhaps a 
genpy bug, or hand-tweaking, or something else I don't understand), but 
I'd suggest you replace it with something like:

default_interface = _SmRecordList

or maybe

default_interface = ISmRecordList

  Paul Prescod





From aleax at aleax.it  Tue Apr 22 17:47:10 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 23:47:10 +0200
Subject: converting list back to string
In-Reply-To: <7E5271B6.6049CB9F.00A631DC@netscape.net>
References: <7E5271B6.6049CB9F.00A631DC@netscape.net>
Message-ID: <200304222347.10918.aleax@aleax.it>

On Tuesday 22 April 2003 07:30 pm, regnivon at netscape.net wrote:
> Alex,
>
> Thank you for responding to my post.  I've included a copy of my original
> post because you said that it wasn't included on your feed.  Here it is:

Thanks!  Putting python-list back in, though...


> my code:
>
> import xreadlines
>
> input = open('c:/python22/programs/linetest.txt', 'r') # 3 lines of
> text
>
> output = open('c:/python22/programs/output.txt', 'w')
>
> for line in input.xreadlines():
>     remove = line.split()
>     remove[0:1] = []
>     output.write(remove)
>
> output.close()
> input.close()

If what you want is to remove the first field and let the rest of
each line unchanged, best might be (in Python 2.2 or better):

input = open('c:/python22/programs/linetest.txt', 'r')
output = open('c:/python22/programs/output.txt', 'w')

# loop directly on the file open for reading -- xreadlines is obsolescent
for line in input:
    # split selectively so most whitespace in the line's not disturbed
    first_field, rest_of_line = line.split(None, 1)
    output.write(rest_of_line)

output.close()
input.close()

i.e., this approach involves no lists at all.

> my quest:  i know that output.write(remove) is not correct.  you
> cannot use the 'write' method on a list.  is there a better way to

You could use output.writelines(remove), but that wouldn't give you
back your lost whitespace;-).

> this part of the program.  By the way, I have purchased the Python Cookbook
> and have been learning a lot about programming.  I appreciate you taking
> the time to help write the book and to answer my question.  I'm certainly a

Hope you're liking the Cookbook!  It may be aimed at programmers a bit
more experienced than you currenlty are, but I hope this just means you'll
be enjoying it for longer;-).

> newbie programmer, but Python seems to be the way to go. Scott

Absolutely -- there's no looking back from it!-).


Alex





From claird at lairds.com  Wed Apr  2 15:54:07 2003
From: claird at lairds.com (Cameron Laird)
Date: Wed, 02 Apr 2003 20:54:07 -0000
Subject: Zope style (was: tuning our site (was: help! advocacy resources needed fast))
References: <mailman.1046819836.25704.python-list@python.org> <mailman.1047414442.16521.python-list@python.org> <v81d26sb3vf170@corp.supernews.com> <irdbl-1ds.ln1@news.lairds.org>
Message-ID: <v8mjffpuph1oc2@corp.supernews.com>

In article <irdbl-1ds.ln1 at news.lairds.org>,
Kyler Laird  <Kyler at news.Lairds.org> wrote:
>claird at lairds.com (Cameron Laird) writes:
>
>>Also, I don't believe *at
>>all* arguments about how Web designers need to work in
>>DTML; it would interest me a great deal to try to thrash
>>that out with someone who sees it differently.
>
>Are you baiting me?  Yes, I still use DTML for forms that
>I know are going to be maintained by other people.  With
>enough of the right "dtml-let" statements, it's usable.
>
>I could do just as well in Python using """ blocks.  That
>would also allow me to avoid having to make a break to
>Python when the functionality required gets too much.
>
>So...I'm no longer willing to argue for DTML.  I use it
>for historical reasons only.
>
>--kyler

Nah, I'm not baiting.  I'm just such a fair-minded guy I find
it really interesting when one technology is unequivocally su-
perior to another.  I think we have that with DTML vs. Python
scripts.

There are a couple of pertinent facts.  First, much of the
passion "Web designers" are supposed to have for "templates"
comes from the supposition that they want stuff to look like 
HTML, so they can use their "HTML editors".  Is *anyone* doing
that with Zope, though?  It would surprise me.  OK, I'm regu-
larly surprised; this one would surprise me a *lot*.  The
ergonomics of that particular combination are just all wrong.

Second, I strongly agree:  write
  source_block_that_Web_monkeys_can_edit = """
    <!-- There's a whole bunch of HTML here. -->
    <h2>a heading</h2>
	 ...
  """
and the non-programmers on the team can cope.

Finally, while your dtml-let is an effective style, I end up
frustrated when I use it, because projects always, always, 
always seem to need more parametrization; managers say, "I want
it just like that, except it needs to {apply to a different
continent,show the current contents of the order,show the 
expanded display when the user has selected that, ...}."  I'm
going to be scripting.  I might as well face it from the start.

That's my sentiment, at least.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From duncan at NOSPAMrcp.co.uk  Thu Apr 24 04:24:28 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 08:24:28 +0000 (UTC)
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <1azna.115402$It5.22896@news2.central.cox.net> <tyfhe8xqhmy.fsf@pcepsft001.cern.ch> <btqu9vkh7u9nj6v25uqevibqhlaq0m8br4@4ax.com> <b86lc3$tnd$1@plover.com> <ccc7084.0304231430.6b81effb@posting.google.com> <mailman.1051139802.16916.python-list@python.org>
Message-ID: <Xns93675F5FC17Cduncanrcpcouk@127.0.0.1>

"andrew cooke" <andrew at acooke.org> wrote in 
news:mailman.1051139802.16916.python-list at python.org:

> a related question - why does queue have to be passed to helper() in the
> code below?  i tried to define it inside walktree (but outside helper)
> since it's global to the function, but python objects saying that the
> local variable "queue" in helper is accessed before it's defined.  this
> seems to be an example of where closures would work perfectly...

Well, I don't know what you did but the problem was your's not Python's. I 
tried your code and it ran, I then added queue = [basepath] immediately 
before the def helper, and removed the parameter queue and the [basepath] 
from the call. I made no other changes. It still runs.

Perhaps you had some other version of the code where you tried assigning to 
queue inside the helper function? That would give the problem you describe, 
but simply mutating it as the code you posted does is fine.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From mwh at python.net  Mon Apr 14 02:54:29 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 14 Apr 2003 06:54:29 GMT
Subject: Parsing strings?
References: <698f09f8.0304132224.6a28c76f@posting.google.com>
Message-ID: <7h3znmtpo0r.fsf@pc150.maths.bris.ac.uk>

tweedgeezer at hotmail.com (Jeremy Fincher) writes:

> Is there any way to turn a string (in Python format, escapes and all)
> into an actual Python string without using eval()?
> 
> I.e., I'd like to turn the four-character string "\\x00" into the
> one-character Python string "\x00" without using eval().  Is there any
> way to do that?

Don't think so -- and if there is, Lib/compiler/transformer.py would
like to know...

Cheers,
M.

-- 
  Wise frogs would take sanctuary in the cool moistness of the cat's
  milk-dish, from where they would watch you.  It's worrying being
  watched by a milk-dish.               -- Tanuki the Raccoon-dog, asr



From imbosol-1050127854 at aerojockey.com  Sat Apr 12 02:40:42 2003
From: imbosol-1050127854 at aerojockey.com (Carl Banks)
Date: Sat, 12 Apr 2003 06:40:42 GMT
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050120565.3489.python-list@python.org>
Message-ID: <KxOla.1317$ah6.258@nwrdny03.gnilink.net>

Christian Tanzer wrote:
> Dynmaic scoping is used for lots of things in Emacs. With a different
> design many of those uses could certainly be replaced by a simpler
> design. I doubt that one could easily replace all of them, though.
> 
> As someone else pointed out, dynamic scoping is used to change values
> of functions called indirectly -- thus converting them into arguments
> passed around doesn't really work.

No, but lexical closures can do this just fine:

    def a():
        value_to_be_passed_indirectly = some_calculation()
        def b():
            do_something_with(value_to_be_passed_indirecty)
        call_function_that_calls_callback(b)


Not to mention you can do similar stuff with a class:

    def a(self):
        closure = B(some_calculation())
        call_function_that_calls_callback(closure.b)
 
    class B:
        def __init__(self,value):
            self.value_to_be_passed_indirectly = value
        def b(self):
            do_something_with(self.value_to_be_passed_indirecty)


The fact is, arguments have been passed indirectly in Python for a
long time, happily, and without dynamic scoping.  It will not be hard
to replace this use of dynamic scoping.


> One could use a
> temporary-global-binding approach (or better temporary-object-binding
> [as in temporarily overriding a buffer attribute]), but without special
> forms like `with-` style macros it would look rather clumsily.

Even if you did want to do this, it would hardly be clumsy to use a
try ... finally.


-- 
CARL BANKS



From hat at se-46.wpa.wtb.tue.nl  Mon Apr  7 07:52:20 2003
From: hat at se-46.wpa.wtb.tue.nl (Albert Hofkamp)
Date: Mon, 7 Apr 2003 11:52:20 +0000 (UTC)
Subject: Python and web programming .. my way ?
References: <odfr8v8uuaf9vcgdom0vfmue7p5h4dv6or@4ax.com>
Message-ID: <slrnb92pjk.fnk.hat@se-46.wpa.wtb.tue.nl>

On Fri, 04 Apr 2003 19:31:23 +0200, Svenne Krap <usenet2002 at krap.dk> wrote:
> Hi.
> 
> My question is, how does you handle page generation nicely in python ?

Have a look at the Web Framework Shootout for several implementations
from

http://colorstudy.com/docs/shootout.html

(nice doc btw!)

Albert
-- 
Unlike popular belief, the .doc format is not an open publically available format.



From blunck at gst.com  Wed Apr 16 10:30:46 2003
From: blunck at gst.com (Christopher Blunck)
Date: 16 Apr 2003 07:30:46 -0700
Subject: How many of you are Extreme Programmers?
Message-ID: <1147e466.0304160630.5b510724@posting.google.com>

Was just reading a thread about how python implements protected and
private methods.  I found JP's response quite interesting (this is
something I've heard numerous friends of mine say of the language):

[In Python] everything is permissible, but not
everything is opportune.  And that is very much Python's philosophy:
rather than focusing on trying to make some things impossible (and
generally failing -- a simple cast in most implementations of C++ lets
you blast away any "protected" that a silly library designer may have
tried to impose on you;-), Python empowers and trusts the programmer.


The natural response a non-Python programmer has to this statement is,
"what?!  you __trust__ the programmer?!  that doesn't work in most
environments."  Despite the ignorance this response demonstrates, I
none-the-less thought about the statement and the answer I came up
with was "we write lots of really thorough tests to demonstrate
functionality."

That got me to thinking:  continuous testing is a cornerstone of the
XP methodology.  <generalization>And Python programmers typically
write lots of tests</generalization> (at least more than their Java
counterparts from what I've seen).  That being said, can that
generalization be extended to "Python programmers subscribe to the XP
methodology"?

So how many of you guys use XP processes?



From duncan at NOSPAMrcp.co.uk  Wed Apr  9 04:48:47 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 9 Apr 2003 08:48:47 +0000 (UTC)
Subject: STRING TO LIST
References: <mailman.1049843053.19630.python-list@python.org> <sNicneNGn5EZ6g6jXTWcpQ@comcast.com>
Message-ID: <Xns9358627B73228duncanrcpcouk@127.0.0.1>

"Terry Reedy" <tjreedy at udel.edu> wrote in 
news:sNicneNGn5EZ6g6jXTWcpQ at comcast.com:

>> I need to change a variable from string to list.
>> > >>> x=str("[[12,13],[14,15]]")
>> > must be a list like this:
>> > [[12, 13], [14, 15]]
> 
>>>> eval('[[12,13],[14,15]]' )
> [[12, 13], [14, 15]]
> 

While this works, the original poster should be aware that unless he/she 
has full control over the origin of the string, it could have nasty side 
effects, since the string could do things like opening files or even 
executing arbitrary Python code.

Of course this may not be relevant, but in case it is, a better solution 
is:

>>> x=str("[[12,13],[14,15]]")
>>> print eval(x, {'__builtins__': {}})
[[12, 13], [14, 15]]
>>> 

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From bdesth at removethis.free.fr  Mon Apr 21 20:20:44 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 22 Apr 2003 00:20:44 +0000
Subject: Why functional Python matters - A kind of conclusion
References: <mailman.1050527910.25653.python-list@python.org>	 <3EA08A7D.3010206@removethis.free.fr> <mailman.1050945519.4137.python-list@python.org>
Message-ID: <3EA48ADC.4000207@removethis.free.fr>

Cliff Wells wrote:
> On Fri, 2003-04-18 at 16:30, laotseu wrote:
> 
> for x in range(100):
> 
>>   print "I shall not say that Lisp is a functional language anymore"
> 
> 
> Better yet:
> 
> for x in range(1000):
>     print "I shall not invite those rude %$&!s at c.l.lisp to c.l.py anymore"
> 
> ;)
> 

Yes, I do apologize for this too. As said in another post in this 
thread, I had forgotten why I stopped reading c.l.lisp.

Now I remember :(

Laotseu




From romany at actimize.com  Sun Apr 13 00:46:00 2003
From: romany at actimize.com (Roman Yakovenko)
Date: Sun, 13 Apr 2003 07:46:00 +0300
Subject: win32 COM test question
Message-ID: <91BFE89EFFA2904E9A4C3ACB4E5F2DF5020575@exchange.adrembi.com>

http://www.hps1.demon.co.uk/users/andy/pyvb/spamsrv.htm

And next time try to search for the answer with google ( python com
visual basic )

> -----Original Message-----
> From: Gabor Nagy [mailto:linux42 at freemail.c3.hu] 
> Sent: Friday, April 11, 2003 1:57 PM
> To: Python
> Subject: win32 COM test question
> 
> 
> Hi!
> 
> I have downloaded the win32all extension from
> http://starship.python.net/crew/mhammond/win32/Downloads.html
> 
> I have followed the description in
> http://www.python.org/windows/win32com/QuickStartServerCom.html
> 
> and created a python script, which when ran, registered my 
> test server.
> 
> So far, so good.
> 
> Now I'd like to test my class from anything but python. On the above
> mentioned page the "Testing our Class" paragraph says this:
> 
> For the purposes of this demonstration, we will test the class using
> Visual Basic. This code should run under any version of Visual Basic,
> including VBA found in Microsoft Office. Any COM compliant 
> package could
> be used alternatively.
> This is not a tutorial in VB. The code is just presented! Run 
> it, and it
> will work!
> 
> The problem is, there is no code on the page, and I don't 
> know how to call
> the exported functions of my new class
> 
> So could someone write me the missing VB example? (I don't 
> know VB or COM,
> so I couldn't figure it out myself, but I simply want to put 
> this call in
> an existing VB program)
> 
> Regards,
> Gabor
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list
> 




From abpillai at lycos.com  Tue Apr 29 10:06:37 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 29 Apr 2003 07:06:37 -0700
Subject: OO + Python Developers Wanted
References: <mailman.1051183297.29110.python-list@python.org> <yyyznmezekm.fsf@adams.patriot.net>
Message-ID: <bd993a2f.0304290606.51a83b53@posting.google.com>

Which is the "City" ?  Which country ?
Is it a contract job ? What after that ?

Anand Pillai

Nick Vargish <nav at adams.patriot.net> wrote in message news:<yyyznmezekm.fsf at adams.patriot.net>...
> "Neeta Ramchandani" <nramchandani at harveynash.com> writes:
> 
> > This is a role for an Investment Bank in the City, for duration of
> > 3-6 months. Rates are OPEN!!
> 
> Which "City" would that be?
> 
> Nick



From glenfant at NOSPAM.bigfoot.com  Mon Apr  7 05:37:36 2003
From: glenfant at NOSPAM.bigfoot.com (Gilles Lenfant)
Date: Mon, 7 Apr 2003 11:37:36 +0200
Subject: OpenFlow, Help Please!!!
References: <K87ka.16$DD6.2390@vicpull1.telstra.net>
Message-ID: <b6rg5g$d9l$1@biggoron.nerim.net>

"Andrew H - Lisasoft" <noemail at forme.com> a ?crit dans le message de news:
K87ka.16$DD6.2390 at vicpull1.telstra.net...
> Hi thanks for reading this.
>
> I'm trying to use OpenFlow (running on Zope). I've no idea how to even
> open the application. Please help me. Is there a tutorial or some kind
> of documentation that will help get me started? I've looked everywhere
> but can't find anything.
>
>
> Any help would be GREATLY appreciated.
>

http://www.openflow.it/EN/index_html

Didn't you find at above URL docs.
There's a downloadable demo here too.

--Gilles




From SherriceGreen at 2by2.net  Fri Apr  4 08:43:46 2003
From: SherriceGreen at 2by2.net (SherriceGreen at 2by2.net)
Date: Fri, 04 Apr 2003 08:43:46 (PST)
Subject: PLEASE REMOVE ME FROM YOUR MAILING LIST
Message-ID: <200304040841578.SM01308@tracy>









From marco.rossini at gmx.ch  Wed Apr 23 17:40:34 2003
From: marco.rossini at gmx.ch (marco)
Date: Wed, 23 Apr 2003 23:40:34 +0200
Subject: ascii txt to LaTeX
Message-ID: <3ea7085a$0$1032$5402220f@news.sunrise.ch>

i wrote a python program for converting an word wrapped(!) ascii text into 
LaTeX format. it actually doesn't do much, just finds(!) and places 
paragraphs (\n\n), replaces apostrophes and umlauts. it does routine stuff: 
why do boring things myself if my computer can do them for me?

feel free to use it (GPL). also i would appreciate critique on my code.

probably someone has made such (or: a better) program before, but i don't 
care, it was easy for me to do and it's all _I_ need.

        marco

#!/usr/bin/python

# Converts a regular word-wrapped ascii text into formated LaTeX
# detects paragraphs, replaces apostrophes, replaces umlauts

# Program written by Marco Rossini <marco.rossini at gmx.ch>
# Copyright: GPL

from sys import argv
from sys import exit
from math import ceil
from string import find
from string import join
from string import strip
from string import replace
from string import whitespace
from string import punctuation

if len(argv) != 2: exit("txt2latex: Argument error!")

try: f = file(argv[1],"r")
except: exit("txt2latex: File not found!")


# GET THE MAXIMAL NUMBER OF CHARACTERS PER LINE
array = f.readlines()
linelength = 0
for i in range(len(array)):
    array[i] = strip(array[i])
    if len(array[i]) > linelength: linelength = len(array[i])

# GUESS IF IT'S A PARAGRAPH BREAK
for i in range(len(array)-1):
    nleft = linelength - len(array[i])
    nright = find(array[i+1]," ")
    if nright == -1: nright = len(array[i+1])
    # if it is, append \n\n to the line, else a space
    if nright+1 <= nleft:
        if len(array[i]) > 0: array[i] += '\n\n'
    else:
        array[i] += ' '

# the lines are joined, the text is NOT word wrapped anymore
text = join(array,"")

# Replace apostrophes intelligent(ly?)
i = 0;
while i < len(text):
    # ... before a word
    if text[i] == '\"' and find(whitespace,text[i-1]) >= 0:
        text = text[:i] + "``" + text[i+1:]
        i += 1
    # ... before a word (single)
    if text[i] == '\'' and find(whitespace,text[i-1]) >= 0:
        text = text[:i] + "`" + text[i+1:]
    # ... after a word, no punctuation
    if text[i] == '\"' and find(whitespace,text[i+1]) >= 0:
        text = text[:i] + "''" + text[i+1:]
        i += 1
    if text[i] == '\"' and find(punctuation,text[i+1])>= 0:
        text = text[:i] + "''" + text[i+1:]
        i += 1
    i += 1

# Here replacements for umlauts. modify if you like to.
text = replace(text,"?","\\\"{a}") -A
text = replace(text,"?","\\\"{A}") -A
text = replace(text,"?","\\\"{o}") -A
text = replace(text,"?","\\\"{O}") -A
text = replace(text,"?","\\\"{u}") -A
text = replace(text,"?","\\\"{U}") -A

# handle output
print text




From member27857 at dbforums.com  Wed Apr  9 21:47:06 2003
From: member27857 at dbforums.com (onefatman)
Date: Thu, 10 Apr 2003 01:47:06 +0000
Subject: sorting strings written to files
References: <2745895.1049870828@dbforums.com>
Message-ID: <2749980.1049939226@dbforums.com>

They're strings separated by \n so list.sort doesn't work on them. and
when i add a new entry, all i'm doing is opening the file in "a" mode
and adding one to the end so i cant manually sort them.

--
Posted via http://dbforums.com



From rmunn at pobox.com  Tue Apr 15 11:42:29 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 15 Apr 2003 15:42:29 GMT
Subject: Python FAQ (Was: Re: Confusion: assignments create object references.)
References: <mailman.1050080058.23175.python-list@python.org> <3E970241.483698C7@engcorp.com> <slrnb9kveb.1q1.rmunn@localhost.localdomain> <b7gsfk$98n$1@panix1.panix.com>
Message-ID: <slrnb9nlc0.493.rmunn@localhost.localdomain>

Aahz <aahz at pythoncraft.com> wrote:
> In article <slrnb9kveb.1q1.rmunn at localhost.localdomain>,
> Robin Munn  <rmunn at pobox.com> wrote:
>>
>>In fact, the Python FAQ needs some trimming, IMHO. With the "Programming
>>in Python" section containing over 100 questions, few are going to have
>>the patience to sit down and read the entire list to see if their
>>question is on the list. Thus we get the same questions over and over in
>>the newsgroup. The FAQ is not serving its purpose.
>>
>>I am not suggesting we get rid of the information currently contained
>>in the FAQ, but we should probably refactor the FAQ into two levels: one
>>for the truly frequently-asked questions, and one for "Here are more
>>questions that crop up from time to time; look through this for useful
>>nuggets of information." I think a shorter FAQ would be read a lot more
>>often.
> 
> "What 'we', white man?"
> 
> The python.org webmasters would love for someone to do this work.  Any
> volunteers?....

<mutter>Should have known better than to speak up...</mutter>

You had to ask that, didn't you? Just when I'm finally about to have a
bit of Free Time(tm).

I'll work on this. If anyone has suggestions, please send them to me.
Include the words "Python FAQ" in the subject, please.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From bokr at oz.net  Thu Apr 17 19:59:46 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 23:59:46 GMT
Subject: Downloading TV listings with urllib
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4> <mailman.1050609705.22139.python-list@python.org>
Message-ID: <b7nf5i$i8d$0@216.39.172.122>

On Thu, 17 Apr 2003 21:59:54 +0200, =?ISO-8859-1?Q?Gerhard_H=E4ring?= <gh at ghaering.de> wrote:

>Josh,
>
>there seems to be a problem with your email account that you might want 
>to investigate:
>
>Reporting-MTA: dns; mail.ghaering.de
>Arrival-Date: Thu, 17 Apr 2003 21:47:39 +0200 (CEST)
>
>Final-Recipient: rfc822; jslm at earthlink.net
>Action: failed
>Status: 5.0.0
>Diagnostic-Code: X-Postfix; host mx03.earthlink.net[207.217.120.78] 
>said: 550
>     jslm at earthlink.net...User unknown
      ^^^^
>
Where'd that come from? The OP header I received starts with

Xref: news comp.lang.python:254370
Path: news.alt.net!news-out.visi.com!petbe.visi.com![...etc.]
From: Josh <mlsj at earthlink.net>
            ^^^^
Newsgroups: comp.lang.python

Just noticed. Not involved ;-)

Regards,
Bengt Richter



From mathieu.rohon at bull.net  Wed Apr  9 08:08:35 2003
From: mathieu.rohon at bull.net (mat)
Date: 9 Apr 2003 05:08:35 -0700
Subject: Python & Remote Access Database
Message-ID: <a6b6f195.0304090408.21003c3a@posting.google.com>

hi everyone!

I have some trouble to find a way to connect and find informations in
a Remote Access Database, from my Python program which running under
linux Red Hat 7.3.

the fact are:
-the access Database is on a WinNT4 server
-my python script is under Linux Red Hat 7.3, with Python 1.5
-both pc are ine the same domain

I saw different module : ODBCmx, win32com... but i don't know which
one to use, and how to use it.

does anyone have ever met those problem?
does anyone have found a solution?

please help

thanks

mat



From daniel at dittmar.net  Tue Apr  1 17:00:07 2003
From: daniel at dittmar.net (Daniel Dittmar)
Date: Wed, 02 Apr 2003 00:00:07 +0200
Subject: How to obtain python file path
In-Reply-To: <20030401165546102-0500@news.covad.net>
References: <20030401163212631-0500@news.covad.net> 	<slrnb8jusc.5u1.ramen@lackingtalent.com> <20030401165546102-0500@news.covad.net>
Message-ID: <b6d260$lg$1@news.eusc.inter.net>

Dan Grassi wrote:
> Not quite, that only provides the path to the initial .py file.
> 
> With a directory structure:
> fa
>     a.py
>     fb
>         b.py
> 
> cd .../fa
> execute a.py
> where a adds fb to sys.path
> a imports b.py
> 
> I need b.py to report it's path which would be .../fa/fb

 >>> import string
 >>> string.__file__
'D:\\Python22\\lib\\string.py'

Daniel




From aahz at pythoncraft.com  Thu Apr 17 10:46:48 2003
From: aahz at pythoncraft.com (Aahz)
Date: 17 Apr 2003 10:46:48 -0400
Subject: Python Thread Question
References: <bd993a2f.0304170534.245aa2d5@posting.google.com> <mailman.1050587609.6119.python-list@python.org>
Message-ID: <b7meoo$mf6$1@panix2.panix.com>

[Mostly good answers from Jp, but one correction is needed]

In article <mailman.1050587609.6119.python-list at python.org>,
Jp Calderone  <exarkun at intarweb.us> wrote:
>On Thu, Apr 17, 2003 at 06:34:11AM -0700, Anand B Pillai wrote:
>> 
>>  4. Can I safely say that multithreading will improve my application
>>     performance if it has similar work to do on many resources at the
>>     same time ? (egs: a web parser/ spider/ a disk-to-disk file copier/
>>     directory synchronizer) Or does it depend upon the nature of the
>>     task at hand ?
>
>  Nope.  Threading costs you context switches.  A multi-threaded app
>using blocking IO will appear "faster" than a single-thread app doing
>the same, but both take longer to run than a single written app using
>non-blocking IO.
>
>  Context switches are usually relatively cheap compared to what your
>app is actually doing, though, so the difference between doing blocking
>IO in a multi-thread app and non-blocking IO in a single-thread app
>isn't always obvious.  (Other things associated with multi-threaded
>apps, such as deadlocks and race conditions are, though ;)

This isn't quite true on a multi-CPU box.  Because I/O releases the GIL,
the OS can schedule threads to run on all available CPUs.  So it's
certainly possible for a threaded app to outrun non-blocking I/O.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From ianb at colorstudy.com  Sun Apr  6 16:56:48 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 06 Apr 2003 15:56:48 -0500
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2k7e78j5s.fsf@localhost.localdomain>
References: <m2u1dcw22s.fsf@localhost.localdomain>
	 <6ee58e07.0304061144.35c5ab92@posting.google.com>
	 <m2k7e78j5s.fsf@localhost.localdomain>
Message-ID: <1049662608.650.91.camel@lothlorien>

On Sun, 2003-04-06 at 15:28, Patrick K. O'Brien wrote:
> Actually, we don't have to clone everything about the interface.  I'm
> using wxPython, so the interface is pretty and more easily configured.
> But some people actually like the minibuffer, so we'll need to support
> something like it and something less ugly.  Heck, I'm actually running
> three different versions of the interface now, and I only wrote the
> code this week (minus the almost two year's worth of code in PyCrust).

No, don't change the interface!  It's one of the best parts about
Emacs... it doesn't use all the stupid crap that most GUIs put in.  The
minibuffer is vastly superior to popup dialogs.  The keyboard focus is
simple and actually works, unlike most GUIs.  M-x actually works, where
menus for that many commands wouldn't work well at all.

I came to appreciate Emacs more when I looked at THE
http://humane.sf.net/the/ -- an editor made by Jef Raskin, former Mac
interface designer.  The interface he describes, without realizing it,
is Emacs, and THE is just a poor implementation of Emacs.  I point out
the similarities in:
https://sourceforge.net/forum/forum.php?thread_id=796533&forum_id=193382

Now, I'm not saying Emacs couldn't be improved.  But it's a far, far
better starting point that "modern" text editors.  I'm thoroughly
convinced that modality is the superior interface for most situations,
and I think GUIs are leaning more in that direction as time goes on.  I
think this is why vi remains popular -- even though I personally don't
like it -- because it's modal.  At least anything that's keyboard
friendly should be modal, and a text editor that isn't keyboard friendly
is clearly stupid.

But I'd still like a Python Emacs to make experimentation with UI
possible.  It would also be neat if it could be an environment for
application development, like Emacs is, like THE wants to be... Oberon
was like this too, and while I didn't like many parts of Oberon, the
overall environment was really neat.  Squeak keeps trying to move this
way too.  I'd love to see an environment where there was less boundary
between module, one-off program, and end-user application.  But ignore
such high ideas, because they will distract and overwhelm :)  The editor
is a great starting point for any number of things, but if you get to
thinking about all those things it'll keep you from starting the
foundation ;)

  Ian






From carroll at tjc.com  Tue Apr  8 13:10:05 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Tue, 08 Apr 2003 10:10:05 -0700
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain> <6j8u8v0hjh3n9dm1jjflu3a970ada6qho4@4ax.com> <slrnb933p7.46e.rmunn@localhost.localdomain>
Message-ID: <0au59v0n82d016vs3mai634d4088uf9f0k@4ax.com>

On Mon, 07 Apr 2003 20:39:13 GMT, Robin Munn <rmunn at pobox.com> wrote:

>The two approaches you describe in the paragraph above are really the
>same approach, in a *slightly* different disguise. When you say "if the
>caller invoked it with the wrong object type, he'll get an error," I'm
>guessing that your idea of the "wrong object type" is an object that's
>not a subclass of "file". But in Python, the "wrong object type" would
>be... (drumroll)... Any object type that didn't have a method named
>"write"!

I don't look at it that way.  The doc for the method specifies that
the parameter is (if not a string) a file object.  That means that the
wrong object type is any object type that isn't a file object.

Now, it happens that the only other method used for this object is the
write method, but that's not guaranteed.  If someone called the method
using an object that is not a file object, but that has a write method
(e.g., a StringIO object), sure, it would work.  For now.  In this
release.  But if the nntp method were changed in the future to do
something else with the file object that actually depended on it being
a file object, it would fail.  

I submit that the failure is because the user is not passing it an
object of the type that it is documented to require, because it says
pretty plainly that it requires a file object, not merely any object
that includes a method named write.  Using a non-file object is just
something you can get away with.

>The key concept to understand is this: in Python, types usually don't
>matter. What matters is interfaces. If you have a function that expects
>to be passed a file object, for instance, someone could pass it a
>StringIO object (http://www.python.org/doc/2.2/lib/module-StringIO.html)
>instead. 

FWIW, nntplib will work with StringIO (or anything else that has a
write method) -- not because it does interface checking on the object
passed, but because, after it's type-checked whether it's a string
(and if it's a string, it creates a file object with open() on the
file named in the string), it assumes that any non-string object it's
been passed is a file object (if you rely on the docs) or a object
that supports write() (if you read the code), and goes ahead and uses
write().  If the object doesn't have a write method, you'll get an
exception.

In any event, I'm not planning on re-writing NNTPlib's file support --
I'm just extending its existing file support to some additional NNTP
commands.




From daniel.dittmar at sap.com  Thu Apr 24 11:48:08 2003
From: daniel.dittmar at sap.com (Daniel Dittmar)
Date: Thu, 24 Apr 2003 17:48:08 +0200
Subject: php equivalents?
References: <OHSpa.8996$h%2.915957@read1.cgocable.net>
Message-ID: <b890vo$rpj$1@news1.wdf.sap-ag.de>

Kevin Howe wrote:
>> register_shutdown($func)
>
> This is basically a "cleanup" function. Just before a script ends, it
> will invoke all functions registered using register_shutdown(). This
> can be used to close database connections, clear sessions, save data,
> etc.

As the notion of 'script' could be very different between PHP and Python,
have a look at the try-finally statement. The code in the finally part is
executed, whether the rest of the code executes normally or is aborted
through an exception.

>> var_export($var)
>
> This function converts a variable to a text string of valid code
> which can be reevaluated at any time. Python has pickle/cPickle, and
> also has modules that can dump/load to XML, however, none of these
> create actual code, but instead create a formatted data string.

repr () is the equivalent function, at least for the builtin types.

Daniel






From graham__fawcett at hotmail.com  Thu Apr  3 12:25:23 2003
From: graham__fawcett at hotmail.com (Graham Fawcett)
Date: 3 Apr 2003 09:25:23 -0800
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com>
Message-ID: <e9570f37.0304030925.542d8721@posting.google.com>

christophe grimault <christophe.grimault at novagrid.com> wrote in message news:<3E8BE455.4000904 at novagrid.com>...
>
> An other way is to move some parts to a C API ext module and deliver 
> only the .so,
> but this is quite time consuming ...

You might also consider Pyrex, which could let you write the sensitive
stuff in nearly-Python. Pyrex will translate into C for you, then
compile your module into an .so.

Extracting and decompiling Python code from a binary created with
McMillan installer/py2exe/etc. is a pretty trivial task, in the sense
that there are tools that will do the work for you, and give you in
return a near-letter-perfect copy of your original source.

I recommend reading Alex Martelli's respose to a similar inquirer a
while back:

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&selm=3mrw9.71925%24TD1.3188780%40news2.tin.it

-- Graham



From steinhardt at inpro.de  Tue Apr  8 05:35:28 2003
From: steinhardt at inpro.de (x-herbert)
Date: 8 Apr 2003 02:35:28 -0700
Subject: ftp through proxy/firewall
Message-ID: <39c0c7.0304080135.41226ffe@posting.google.com>

Hi,

I can?t download or upload a file per FTP through a proxy :-((

How can i insert in the script the proxy information e.g. host =
192.168.0.99 and port 2121 ?

Thanks & CU

x-herbert



From lekishvili at python.qartu.com  Sat Apr  5 13:49:46 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 5 Apr 2003 10:49:46 -0800
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <mailman.1049530929.9225.python-list@python.org>
Message-ID: <a3b5b608.0304051049.7048d76@posting.google.com>

> 
> Right now, no, there isn't any Right Way to do Python on the web.  Zope
> might be the closest thing, but there is no consensus.  I can't really
> figure out what a consensus would be, but I imagine there's at least
> some things we can come to consensus on... maybe underlying
> infrastructure, or ways we can share code (like making templating
> separate from the frameworks).  There's still a lot of work to be done
> on that path, though.
> 

The question is that the only way to establish Python as a language
for corporate level web-programming is to have the standard approach
if not introduced, then at least supported by the Python consortium...



From adalke at mindspring.com  Wed Apr 30 14:27:29 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 30 Apr 2003 12:27:29 -0600
Subject: Raising objects
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it> <mailman.1051678851.21661.python-list@python.org> <OFLra.36816$K35.1143840@news2.tin.it> <mailman.1051719837.14426.python-list@python.org>
Message-ID: <b8p4au$dh7$1@slb9.atl.mindspring.net>

Steven Taschuk:
> But is there a problem with being able to raise anything?  Perhaps
> some class of easy-to-produce bugs that I'm not seeing?

Here's my guess.  Currently to catch any exception you need to do

try:
   f()
except:
  ...look in sys.exc_info()...

Calling sys to do that is a bit of a wart, but it's the only
solution since current Python doesn't have a unified base
type.

However, if all exceptions were derived from Exception
then you could do

except Exception, e:
   ... use e ...

This cleans up the wart at the expense of not allowing
any old object to be thrown.

Will Python ever require that all classes be derived
from 'object'?  If so, then the generic "catch everything"
would be

except object, e:
   ... use e ...

But that won't happen soon.

                    Andrew
                    dalke at dalkescientific.com





From martin at v.loewis.de  Fri Apr 18 17:18:22 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 18 Apr 2003 23:18:22 +0200
Subject: Which documentation for 2.2?
References: <mailman.1050685171.4364.python-list@python.org>
Message-ID: <m3fzoffqsh.fsf@mira.informatik.hu-berlin.de>

Paul Berkowitz <berkowit at silcom.com> writes:

> Is this an updated (corrected?) documentation for the same Python
> 2.2 release as the other (in which case why is the earlier set still
> published), or is it documentation for a 2.2p1 update of Python
> itself (in which case why doesn't it say so on the main page?)

It is an update to the documentation.

Regards,
Martin



From tanzer at swing.co.at  Mon Apr  7 02:18:20 2003
From: tanzer at swing.co.at (Christian Tanzer)
Date: Mon, 07 Apr 2003 08:18:20 +0200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: Your message of "06 Apr 2003 19:40:51 +0200."
             <m33ckvo76k.fsf@mira.informatik.hu-berlin.de>
Message-ID: <E192PxU-0007S4-00@tswings.swing.cluster>

martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=) wrote:

> pobrien at orbtech.com (Patrick K. O'Brien) writes:
>
> > Are there translators out there?  I looked and couldn't find any.
>
> Parsing S-Expressions is fairly trivial - they are designed to be
> readily parsable. For a full translator, you need to define the
> language mapping first, i.e. how do you want to represent atoms,
> function names, etc.

I expect the main difficulty to be the difference in scoping
models -- the dynamic scoping of Elisp is probably pretty hard to
simulate in Python without incurring major performance problems. And
that's used by lots of Elisp code, so skipping it is not an option if
you want to have comparable behavior.

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92





From sjmachin at lexicon.net  Sun Apr 27 03:23:05 2003
From: sjmachin at lexicon.net (John Machin)
Date: 27 Apr 2003 00:23:05 -0700
Subject: Reading a non-standard floating point format
References: <mailman.1051400681.24630.python-list@python.org> <b8fd9r$6fo$0@216.39.172.122>
Message-ID: <c76ff6fc.0304262323.7350200c@posting.google.com>

bokr at oz.net (Bengt Richter) wrote in message news:<b8fd9r$6fo$0 at 216.39.172.122>...
> On Sat, 26 Apr 2003 19:41:43 -0400, "Tim Peters" <tim_one at email.msn.com> wrote:
> [...]
> >There's no way to guess without examples, though.  If you can, show us the
> >48-bit byte strings for these numbers:
> >
> >    0.0
> >    1.0
> >    1.25
> >    2.9e-39
> >    1.7e38
> >
> >That should be enough to fill in the missing pieces.
> >
> Maybe include
>     -1.25
> also?
> And can minus zero be generated?
> Are numbers sign/magnitude or?
> 

I would ask for 1.125 as well as 1.25 (to be sure of bit order within
a byte[*]) and also 1.03125 (1+1/32, 8 bits onwards from 1.25) to be
sure of byte order within the 48-bit stucture.

The OP might also consider posting on alt.sys.alpha-micro -- surely
somebody there will be able to provide better docs than the quoted
excerpt from the AlphaBasic manual. You could ask about access to the
AMs that are still running -- it could be very handy to mount a
"chosen-plaintext" attack on the problem i.e. write an AlphaBasic
program which wrote various chosen f.p. numbers to a "random access"
(i.e. binary file) then ogle it with the aid of a hex dumper.

* ... bit order within a byte: don't laugh, we're in 1980s
microcomputer territory here; no f. p. h/ware, all done in ASM by
heavy bit-bashing. I recall different variants of "standard" 32-bit
CRCs -- the variants arising from whether the bits were shifted in
from the left/right/high-end/low-end/east/west/...

Hope this helps,
John



From cliechti at gmx.net  Thu Apr 10 16:16:25 2003
From: cliechti at gmx.net (Chris Liechti)
Date: 10 Apr 2003 22:16:25 +0200
Subject: Writing a TCP/IP chat server
References: <Pine.LNX.4.44.0304101906590.8243-100000@servalan.ecs.soton.ac.uk>
Message-ID: <Xns9359E2D2C5500cliechtigmxnet@62.2.16.82>

Ian Partridge <iop199 at ecs.soton.ac.uk> wrote in
news:Pine.LNX.4.44.0304101906590.8243-100000 at servalan.ecs.soton.ac.uk: 

> Hi,
> 
> I am trying to write a chat server, which users will telnet to in
> order to interact with each other.
> 
> What is the best way to do this?
> 
> I have searched the web and found three different approaches:
> 
> 1. Use the socket libraries and select to multiplex between the 
> connections.
> 
> 2. Use the SocketServer module (although I can't find a decent
> tutorial on how to use this module)
> 
> 3. Use the asyncore module (I only found one example doing this, and I
> couldn't understand it).

 4. use twisted  http://twistedmatrix.com/
where you can:
4a. wirite your own protocol with a few lines of code
or
4b. use whats already in the libraries

5. use an existing server, e.g. there are different IRC servers.
there is no reason to reinvent the wheel unless ouy do it for education ;-)

> What's the the simplest way to approach the problem?

5. is simplest, for fun 4. (and 1. if you want to port it to C for some 
reason, but trust me, there isn't one to do so ;-)

chris
 
> Any tips gratefully received.
> 
> TIA,
> Ian
> 
> 



-- 
Chris <cliechti at gmx.net>




From aleax at aleax.it  Tue Apr 29 09:46:51 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 13:46:51 GMT
Subject: os.spawn
References: <Xns936C9D9E91B1Bthewolffman@194.109.133.29>
Message-ID: <fnvra.19907$3M4.571598@news1.tin.it>

Jeroen Wolff wrote:

> Hi,
> 
> I want to run some shell commands from within python.
> Wat i normal do at the shell prompt is:
> 'flow-cat <dirname> | flow-stat <param> > outputfile.txt'
> 
> I like to do this from within python and wait for it to finish and the
> process the outputfile.txt
> Is this possible? The problem is with '|' and '>" piping....

Why is this a problem?  Doesn't everything run as you wish
when you do:

import os
os.system('flow-cat dirname | flow-stat param > outputfile.txt')

???


Alex




From staschuk at telusplanet.net  Tue Apr 22 22:24:21 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 22 Apr 2003 20:24:21 -0600
Subject: Regarding Dictionaries in python
In-Reply-To: <200304221903.24570.andy@wild-flower.co.uk>; from andy@wild-flower.co.uk on Tue, Apr 22, 2003 at 07:03:24PM +0100
References: <3EA455DB.9000300@antares.enst-bretagne.fr> <200304221903.24570.andy@wild-flower.co.uk>
Message-ID: <20030422202421.B1462@tibia.amotlpaa.bogus>

Quoth Andy Jewell:
  [...]
> Dictionaries are not ordered entities - check the Tutorial.  [...]
> Numeric items are always added in ascending order (as far as my little 
> experiments showed anyway):  
> 
> >>> d={}
> >>> d[1]="one"
> >>> d
> {1: 'one'}
> >>> d[2]="two"
> >>> d
> {1: 'one', 2: 'two'}
  [etc.]

This happy sorting of int keys works in simple cases, but it is
very fragile and should not be relied on.

When it works, it's because (a) ints are their own hash values,
(b) entries are by preference placed at index hash(key) in the
underlying table, and (c) iteration over the dict occurs in the
order of the underlying table.

But even if the keys' hash function preserves order (as ints'
does, if we ignore negative values), the ordering can be broken if
hash(key) is larger than the table size (which is always a power
of 2, I think):

    >>> d = {6: 'foo'}
    >>> d[65536] = 'bar' # stored at index 65536 % tablesize == 0
    >>> d
    {65536: 'bar', 6: 'foo'}

Further, when a new key's hash is larger than the table size, it
might collide with an existing entry; in this case the new entry
has to be placed elsewhere.  The selection of the elsewhere is
complicated (see comments in Objects/dictobject.c), and can easily
break the sorting.  This kind of breakage depends not only on the
hash values, but on the order of insertion:

    >>> d = {}
    >>> d[23] ='foo'
    >>> d[7] = 'bar'  # collides with 23
    >>> d             # appears sorted anyway
    {7: 'bar', 23: 'foo'}

    >>> d = {}
    >>> d[7] = 'bar'
    >>> d[23] = 'foo' # collides with 7
    >>> d             # dis-sorted
    {23: 'foo', 7: 'bar'}

Moreover, this kind of breakage may go unnoticed until more
elements are added to the dict:

    >>> d = {}
    >>> d[2] = 'foo'
    >>> d[18] = 'bar' # collides with 2
    >>> d             # appears sorted anyway, until...
    {2: 'foo', 18: 'bar'}
    >>> d[5] = 'snee'
    >>> d
    {2: 'foo', 18: 'bar', 5: 'snee'}

In other words, do not be deceived by simple tests of this matter!

-- 
Steven Taschuk                                        staschuk at telusplanet.net
"Study this book; read a word then ponder on it.  If you interpret the meaning
 loosely you will mistake the Way."         -- Musashi, _A Book of Five Rings_




From vze4rx4y at verizon.net  Mon Apr 14 19:31:43 2003
From: vze4rx4y at verizon.net (Raymond Hettinger)
Date: Mon, 14 Apr 2003 23:31:43 GMT
Subject: Iterator vs Lists    Was: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi> <Oqwma.11524$LB6.319489@news1.tin.it>
Message-ID: <zxHma.16806$Gb1.16713@nwrdny01.gnilink.net>

[Martelli bot]
> d=dict.fromkeys(range(200))''' 'for x in d.keys(): pass'
> d=dict.fromkeys(range(200))''' 'for x in d.iterkeys(): pass'
> 10000 loops, best of 3: 28.2 usec per loop

> That's most definitely NOT a significant performance difference,
> of course, but calling iteration on iterkeys "a lot faster" than
> iteration on keys would appear to be substantially misleading, at
> least without a lot of qualifications and hedging!-)

Here are a few thoughts about the performance differences between
looping over an iterator versus building a list and looping over it.

* Constructing a list is especially fast when the underlying object
   can report its length.  That enables the list constructors to
pre-allocate
   the entire list at once.  If not, the construction requires a series of
   appends and an occasional resize/copy operation.

* Iterators have a tiny constant size while lists take space proportional
   to the length of the list.  The part that is not obvious is that looping
   over the iterator re-uses the same memory location again an again.
   So the relevant data is almost always in the hardware memory cache.
   The same is also true for small lists.  However, large lists cannot all
   fit in the cache and accessing them will be slower than their iterator
   based counterparts.  As processor speeds race ahead of memory
   speeds, the performance difference will become more pronounced
   over time.

* It takes time to allocate Python objects, but the smart allocator saves
   some of that time when it can reuse recently discarded objects.  This
   technique works very well with iterators which tend to continuously
   create and abandon many same sized objects.  However, list based
   approaches keep a reference to each element preventing their reuse.

   Sometimes the difference is tiny; sometimes it is huge.  There were some
   examples where itertools.izip() outperformed zip() by about ten to one.
   Timbot's subsequent optimizations to zip() made this much less likely.
Still,
   izip() generally performs at least a little better than its list based
counterpart.

The executive summary:
For small datasets, iterator and list based approaches have similar
performance.
For larger datasets, iterators save both time and space.


Raymond Hettinger








From aleax at aleax.it  Tue Apr  1 04:40:52 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 01 Apr 2003 09:40:52 GMT
Subject: Is Python overhyped (just like Java)?
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <uznnb43a8.fsf@boost-consulting.com>
Message-ID: <E8dia.53612$i26.1638078@news2.tin.it>

David Abrahams wrote:
   ...
>> One example I give is a task for which C++ is quite suited, with its
>> standard library -- reading a text file, breaking it into
>> whitespace-separated 'words', building an index from each word to
>> the line numbers on which it appears, and showing the index with
>> words in alphabetical order, one per line, each followed by the line
>> numbers on which it appears.
> 
> Boy, that sure sounds like a job for Python to me.  Something like
> "high-performance linear algebra with sparse matrices" sounds more
> like a job for which C++ is "quite suited".

The C++ standard library has no primitives to deal with sparse
matrices, while it does have some for handling text, etc, etc.
If you have other libraries available, besides the standard ones,
both C++ and Python could be fine.  To WRITE such libraries, I'd
choose C++ (or maybe Python + Numeric, but that's iffier) -- but
applications are another thing.


>> excellence of Python's intrinsics (dictionaries, lists, strings)
>> versus the lesser quality of C++'s library implementation (maps,
>> vectors, strings) -- with different implementations, you may see
>> different ratios.
> 
> Really?  Are you sure that time wasn't sunk into I/O?  "Read all the
> lines from a file" is a primitive in Python, but not in C++.

Add files to the list of Python intrinsics above, I guess!

> Also, it's only fair to point out that C++ doesn't have one library
> implementation -- each compiler comes with its own.  Which
> implementation(s) were you testing?  Oh, I see it was probably
> libstdc++ that comes with gcc 3.1.  Well, IIRC GCC-3.1 was well known
> to optimize poorly (it took them some time to figure out how to
> integrate all the new optimizations that came in with GCC-3, and they
> actually made things slower for a while), and I wouldn't exactly say
> that libstdc++ is the fastest implementation.

Yes, I used the gcc and libraries that come for free with Linux.  I'll
be quite happy to benchmark with any other free compiler and implementation
of the C++ standard library, of course!

> In other words, there's nothing intrinsic about C++ which says that it
> has to be twice as slow as Python at this job.  I don't think it's
> fair to talk about "the lesser quality of C++'s library
> implementation."
> 
> Aren't you glad I'm asking these questions now, instead of at your
> talk in Oxford? ;-)

No, I'll be quite happy to debate this in Oxford, too.  Hmm, better
rush -- my plane's leaving soon...!


> <snip>
> 
>> The simplest and most flexible C++ you can write is still way
>> bigger, more complicated, and less flexible than the most refined
>> and optimized Python code it may make sense to write -- it's as
>> simple as this.
> 
> I think when you're comparing the use of the core language I have to
> agree.  However, when it comes to functionality provided by libraries,
> it really depends on the quality and availability of the library
> interface.  That said, I think Python probably still has many more
> useful libraries than C++ does.

I'm talking of the STANDARD libraries only.  If you open the field
to third-party libraries, you've basically comparing infinities;-)


Alex




From adamv at nwlink.com  Thu Apr 17 04:47:18 2003
From: adamv at nwlink.com (Adam Vandenberg)
Date: Thu, 17 Apr 2003 01:47:18 -0700
Subject: HTML form pre-populating functionality or hints on building my own
References: <mailman.1050550654.3607.python-list@python.org>
Message-ID: <v9sqhq3gj3b48d@corp.supernews.com>

> I'm wanting to find (or make) a module that will take a HTML file, locate
the
> input-related form tags (<INPUT, <TEXTAREA, <SELECT, <OPTION) and modify
them...

Here's a sample that parses some HTML and handles some form elements,
refilling them on POST-back:
http://flangy.com/dev/tinkertoys/ultraform.py
(Source: http://flangy.com/dev/tinkertoys/ultraform.py?viewsource=1 )

But unfortunately... it uses "special tags" for the forms (the t.whatever
tags.)
It would probably be easy enough to modify it to look for the regular form
tag names instead.
I specifically went with special tags because I wanted to abstract a control
from the HTML used to implement it.

I'm using the standard module sgmllib as the parser. Feel free to take the
code and run with it.
(Note that production could will want to be a lot more thorough about
cgi.escape'ing data that's taken from the form and written back out to the
page.)

Take care,
Adam Vandenberg
http://flangy.com






From michael at stroeder.com  Thu Apr 17 10:06:57 2003
From: michael at stroeder.com (=?ISO-8859-1?Q?Michael_Str=F6der?=)
Date: Thu, 17 Apr 2003 16:06:57 +0200
Subject: Pythonic way of web-programming
In-Reply-To: <0gqdnfw_XuYm3gCjXTWcow@speakeasy.net>
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <B8Pma.20554$T34.656139@news2.tin.it> <mailman.1050471334.22000.python-list@python.org> <0gqdnfw_XuYm3gCjXTWcow@speakeasy.net>
Message-ID: <11p2n-ci3.ln1@nb2.stroeder.com>

A.M. Kuchling wrote:
> 
> The FastCGI protocol is really complicated because it supports multiplexing 
> several different transactions over one socket.  Data is broken up into
> multiple packets, each with a unique transaction ID, and the process on the
> other end can then reassemble them and hand them out to different threads.
> However, I'm not sure the Apache mod_fastcgi supports this and few client
> modules seem to, so all of that complexity buys nothing in most cases.

I'm using a multi-thread FastCGI web application with mod_fastcgi. Or maybe 
I did not get your point.

Ciao, Michael.




From graham at effectif.com  Mon Apr 14 14:33:53 2003
From: graham at effectif.com (Graham Ashton)
Date: Mon, 14 Apr 2003 19:33:53 +0100
Subject: __slots_ and inheritance
References: <3E9631B5.5070505@irl.cri.nz> <b7340h0r7m@enews1.newsguy.com> <yfsistmh4ec.fsf@black132.ex.ac.uk> <a8b7f07a.0304120352.75b53dd8@posting.google.com> <6EWla.4594$LB6.132415@news1.tin.it>
Message-ID: <pan.2003.04.14.18.33.53.118906@effectif.com>

On Sat, 12 Apr 2003 15:53:38 +0000, Alex Martelli wrote:

>> Alexander Schmolck <a.schmolck at gmx.net> wrote in message
>> news:<yfsistmh4ec.fsf at black132.ex.ac.uk>...
>>  
>>> __slots__ is the one addition in python2.2 that really annoys me. It
>>> seems an ill-conceived jumble of conceptually unrelated things (bug
>>> avoidance, optimization, safety, encapsulation) that screws up
>>> reflection big time for little apparent gain. As Knuth said ...
> 
> I think it's an important memory optimization, PERIOD.  The other
> conceptually unrelated things (bug avoidance -- I can't even SEE the
> others!) aren't in __slots__' own design (even though a commentator
> who's usually quite perceptive seems to have seen the "bug avoidance"
> feature that wasn't really there).

>From http://www.python.org/doc/2.2.1/whatsnew/sect-rellinks.html:

    Finally, it's possible to constrain the list of attributes that can be
    referenced on an object using the new __slots__ class attribute.
    Python objects are usually very dynamic; at any time it's possible to
    define a new attribute on an instance by just doing obj.new_attr=1.
    This is flexible and convenient, but this flexibility can also lead to
    bugs, as when you meant to write obj.template = 'a' but made a typo
    and wrote obj.templtae by accident.

    A new-style class can define a class attribute named __slots__ to
    constrain the list of legal attribute names.

This thread has pointed out rather well that __slots__ don't really behave
as suggested above in anything but the simplest of situations. It wasn't
until I read the __slots__ section in the nutshell book that I realised
that this wasn't the case. Consequently, I think the above paragraphs are
misleading. Shouldn't they be fixed?

--
Graham Ashton



From robin at jessikat.fsnet.co.uk  Fri Apr 11 14:36:38 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Fri, 11 Apr 2003 19:36:38 +0100
Subject: module for doing unix "find"-like things?
References: <7LDla.1483$Pq2.135830754@newssvr14.news.prodigy.com>
Message-ID: <hklpO1B2swl+Ewv9@jessikat.demon.co.uk>

In article <7LDla.1483$Pq2.135830754 at newssvr14.news.prodigy.com>, Mark
Harrison <mh at pixar.com> writes
>Is there such a thing?  I didn't see such a thing, but
>"find" is not the easist thing to search for.
>
>Specifically, I would like to replace an external call to
>    for i in `find $top -type l -maxdepth 1`; ...
>
>Many TIA,
>Mark
>
there used to be a find module which is now in lib-old (I think), but
that didn't do anything other than search for files which match a glob
pattern.

-- 
Robin Becker



From aahz at pythoncraft.com  Wed Apr 16 16:08:04 2003
From: aahz at pythoncraft.com (Aahz)
Date: 16 Apr 2003 16:08:04 -0400
Subject: One-liner blocks (was Re: accessing a functions var from a subfunction.)
References: <a3cae241.0304140616.1833e5c4@posting.google.com> <iCzma.16405$T34.518484@news2.tin.it>
Message-ID: <b7kd74$jqm$1@panix2.panix.com>

In article <iCzma.16405$T34.518484 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>
>def test():
>    class Bunch: pass
>    loc = Bunch()
>    loc.count = 0
>    def inc_count():
>        loc.count += 1
>    inc_count()
>    inc_count()
>    print loc.count

Note that Guido deeply regrets permitting one-line blocks; this is
something he may try fixing for Python 3.0.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

Why is this newsgroup different from all other newsgroups?



From goodger at python.org  Sat Apr 26 18:25:18 2003
From: goodger at python.org (David Goodger)
Date: Sat, 26 Apr 2003 18:25:18 -0400
Subject: Reminder to PEP authors
Message-ID: <3EAB074E.9040102@python.org>

There are several PEPs with "Draft" status which are ripe for review.
PEP 1 states:

     Once the authors have completed a PEP, they must inform the PEP
     editor that it is ready for review.  PEPs are reviewed by the BDFL
     and his chosen consultants, who may accept or reject a PEP or send
     it back to the author(s) for revision.

     Once a PEP has been accepted, the reference implementation must be
     completed.  When the reference implementation is complete and
     accepted by the BDFL, the status will be changed to "Final".

PEP authors, please keep your PEPs up to date.  When you think the PEP
is ready for review, please send a note to <peps at python.org> stating
"PEP X is ready for review".  Otherwise the PEP may remain in "Draft"
limbo indefinitely.  It is the PEP author's responsibility to move the
process forward:

     Each PEP must have a champion -- someone who writes the PEP using
     the style and format described below, shepherds the discussions in
     the appropriate forums, and attempts to build community consensus
     around the idea.

Authors with CVS check-in privileges are welcome to check in their own
content changes.  Others should send updates to <peps at python.org>
(please make updates to the latest text from CVS).

-- 
David Goodger                    <http://starship.python.net/~goodger>
Python Enhancement Proposal (PEP) Editor <http://www.python.org/peps/>

(Please cc: all PEP correspondence to <peps at python.org>.)





From bdesth at removethis.free.fr  Tue Apr 15 19:50:31 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 15 Apr 2003 23:50:31 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
Message-ID: <3E9C9AC7.9090706@removethis.free.fr>

Dave Benjamin wrote:
> It has been my increasing concern that Python is headed in a decidedly
> object-oriented direction, ignoring the benefits it currently offers to
> functional programmers. 

(snip a long and interesting post)

Agree++

Even for OO programmers, functionnal features in Python are IMHO a great 
plus, and BTW functionnal and OO paradigm does not have to conflict 
(that would be functionnal vs imperative and object vs procedural). CLOS 
is one of the great system objects out here, and it's been implemented 
on top of a functionnal language, so...

In fact, I'd rather regret that Python is not more functional *and* 
object oriented !-)

Guido, if you hear us... !-)

My 2 eurocents
Laotseu




From member27735 at dbforums.com  Mon Apr  7 06:48:24 2003
From: member27735 at dbforums.com (nitintz)
Date: Mon, 07 Apr 2003 10:48:24 +0000
Subject: Importing problem
Message-ID: <2737271.1049712504@dbforums.com>

hi,

I am new to the python.
Actually i am using PyQt .. and my problem is i wish to import a Qt
project in my PyQt program.
Can anyone help me in it?

Thanks in Advance
Nitin

--
Posted via http://dbforums.com



From postmaster at earthlink.net  Thu Apr 24 17:29:09 2003
From: postmaster at earthlink.net (Daniel T.)
Date: Thu, 24 Apr 2003 21:29:09 GMT
Subject: MacPython-OSX (wxPython build) question.
Message-ID: <postmaster-343A8F.17285824042003@nnrp05.earthlink.net>

I pulled the binary off of wxPython "MacPython-OSX-2.3a2-2" and 
installed it, but when I go to run the PythonIDE, it quits and I get the 
following in my console log:

Traceback (most recent call last):
  File 
"/Applications/MacPython-2.3/PythonIDE.app/Contents/Resources/PythonIDE.p
y", line 51, in ?
    import PythonIDEMain
  File 
"/Library/Frameworks/Python.framework/Versions/2.3/Mac/Tools/IDE/PythonID
EMain.py", line 419, in ?
    PythonIDE()
  File 
"/Library/Frameworks/Python.framework/Versions/2.3/Mac/Tools/IDE/PythonID
EMain.py", line 36, in __init__
    Wapplication.Application.__init__(self, 'Pide')
  File 
"/Library/Frameworks/Python.framework/Versions/2.3/Mac/Tools/IDE/Wapplica
tion.py", line 28, in __init__
    FrameWork.Application.__init__(self)
  File 
"/Library/Frameworks/Python.framework/Versions/2.3/lib/python2.3/plat-mac
/FrameWork.py", line 115, in __init__
    self.makemenubar()
  File 
"/Library/Frameworks/Python.framework/Versions/2.3/Mac/Tools/IDE/Wapplica
tion.py", line 269, in makemenubar
    self.makeusermenus()
  File 
"/Library/Frameworks/Python.framework/Versions/2.3/Mac/Tools/IDE/PythonID
EMain.py", line 142, in makeusermenus
    os.mkdir(path)
OSError: [Errno 13] Permission denied: '/Scripts'

So my question is, why is it trying to access/create a "/Scripts" folder 
while creating the menus? This shouldn't be happening should it?



From spam at magnetic-ink.dk  Sat Apr 12 13:03:23 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Sat, 12 Apr 2003 17:03:23 +0000 (UTC)
Subject: DNS lookup
References: <eTFaa.3697$Fc.114939@tornado.fastwebnet.it> <b4fb0l$u0q$03$1@news.t-online.com> <qQWla.732$5n6.120@news.cpqcorp.net>
Message-ID: <038ed2ee-0a8b-4e9d-82f1-8726a235e302@news.szn.dk>

LaMont Jones wrote:

> And if I need asynchronous NS/SOA lookups?

You could look into the ADNS module:

	<http://dustman.net/andy/python/adns-python>
	<http://packages.debian.org/python-adns>


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From revanna at mn.rr.com  Tue Apr  1 10:43:00 2003
From: revanna at mn.rr.com (Anna)
Date: Tue, 01 Apr 2003 15:43:00 GMT
Subject: list index question
References: <e62b26d3.0304010613.39f98991@posting.google.com> <pan.2003.04.01.15.38.54.181811@mn.rr.com>
Message-ID: <pan.2003.04.01.15.42.18.291186@mn.rr.com>

On Tue, 01 Apr 2003 09:38:54 +0000, Anna wrote:

> 
>>>> list1=[(100,0), (10,5), (100,10), (100,0)] print list1

Have I mentioned that I *hate* it when my mailreader does this to a simple
copy/paste from IDLE... 

That print statement should, of course, have been on a separate line...

Ugh.
Anna



From sjmachin at lexicon.net  Sat Apr  5 18:39:40 2003
From: sjmachin at lexicon.net (John Machin)
Date: 5 Apr 2003 15:39:40 -0800
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
Message-ID: <c76ff6fc.0304051539.7819161b@posting.google.com>

Simon Faulkner <news at titanic.co.uk> wrote in message news:<1lau8vcctfn4g2pi9k5df3km7qgg5r72sv at 4ax.com>...
> I have a list of about 5000 numbers in a text file - up to 14 digits
> each.
> 
> I need to check for duplicates.
> 
> What would people suggest as a good method?
> 
> Simon

OTTOMH, untested, caveat lector, YMMV, etc etc:

Methods based on sorting:

1. On Unix or equivalent:
sort <thelist.txt | uniq -c | grep "^1[^0-9]"
[there may be an option for uniq to output only the dups -- RTFM]

2. suck the data into a spreadsheet program, sort the column (A), in
cell B1 put a formula =if(A1=A2,"***","") [Excel notation], propagate
that down column B

3. seeing you did ask this on c.l.py:

alist = file("thelist.txt").readlines()
alist.sort()
for k in xrange(len(alist)-1):
   if alist[k+1] == alist[k]:
      print "Dup", alist[k+1],

Method based on hashing:

seen = {}
for line in file("thelist.txt"):
   if line in seen:
      print "Dup", line,
   seen[line] = 1

WARNING: you don't say whether your numbers are real or integer. If
they are real, then only method 2 has any chance of pointing out that
"123", "123.", "1.23E2" and "123.00" are duplicates [if indeed you
regard them as duplicates]. The other methods are based on exact
textual comparison [including, in the Python methods, the "\n"!!]

So you might want to do a variant of method 3:

tolerance = 0.00001 # suit yourself
alist = []
for line in file("thelist.txt"):
   anum = float(line.strip())
   alist.append(anum)
alist.sort()
for k in xrange(len(alist)-1):
   if abs(alist[k+1] - alist[k]) < tolerance:
      print "Dups", alist[k], alist[k+1]

HTH,
John



From jdhunter at ace.bsd.uchicago.edu  Tue Apr  1 10:15:17 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Tue, 01 Apr 2003 09:15:17 -0600
Subject: What a Deal << 2GB and 1GB >> USB Flash Memory Drive
In-Reply-To: <23c792e99172f4592374b34ae05431a7@news.teranews.com> ("lorabanks"'s
 message of "Mon, 31 Mar 2003 19:26:31 GMT")
References: <23c792e99172f4592374b34ae05431a7@news.teranews.com>
Message-ID: <m2vfxyxn96.fsf@mother.paradise.lost>

>>>>> "lorabanks" == lorabanks  <rzrtkxlp at qxqsjvlo.com> writes:

    lorabanks> Great Deals on USB Flash Memory Drive
>>>>> Price: 2GB - $779
>>>>> Price: 1GB - $389

    lorabanks> with and with Free Shipping *%$
    lorabanks> http://www.ebestshoppers.com #%&


I wonder if this made it past the spam filter because of the >>>> ...

John Hunter





From lbrannma at cablespeed.com  Sun Apr  6 12:47:35 2003
From: lbrannma at cablespeed.com (Lance)
Date: Sun, 6 Apr 2003 09:47:35 -0700
Subject: executable won't run on another machine
Message-ID: <v90mh6ppf9bc09@corp.supernews.com>

Hi All,

I created an executable using McMillan's Installer. It runs great on the
development machine. I copied it to another machine and it doesn't run. I
type the executable name from a command prompt, hit return, and get another
command prompt.

I'm running RedHat 8.0 on both machines. Both have the workstation
installation. The development machine is dual boot, with XP on the other
partition.

Any suggestions will be appreciated.

Thanks,
Lance





From hat at se-46.wpa.wtb.tue.nl  Tue Apr  8 11:16:03 2003
From: hat at se-46.wpa.wtb.tue.nl (Albert Hofkamp)
Date: Tue, 8 Apr 2003 15:16:03 +0000 (UTC)
Subject: Logging into CVS using Python
References: <ShBka.143$N5.26584@mencken.net.nih.gov>
Message-ID: <slrnb95ptl.io7.hat@se-46.wpa.wtb.tue.nl>

On Tue, 8 Apr 2003 10:47:13 -0400, Mike <uptonothing at yahoo.com> wrote:
> How do I use Python to login into CVS? Here's what I have tried, but it has
> failed. I'm not even sure that I'm using the correct methods.

Doing 'cvs -d :pserver:mike at server.com:1000/path/to/CVS login' on the
command-line should work. If it does, the method is correct :-)

> import popen2
> 
> pipe_in, pipe_out = popen2.popen2("cvs -d
>:pserver:mike at server.com:1000/path/to/CVS login")

If mike and PASSWORD are a correct combination at server.com, it may be
easier to convince rsh or ssh to log you in without entering a password.
That also has the advantage of not having your password in a Python file.

> 
> s = pipe_in.readline()
> 
> print s
> 
> pipe_out.write("PASSWORD")
> 
> pipe_out.flush()
> 
> pipe_in.close()
> pipe_out.close()
> 
> The code just results in "cvs login: authorization failed." Any help would
> be appreciated.

Two things may be wrong (or maybe even both are wrong):
- cvs asks for the password using /dev/tty, which is directly to your
  terminal, instead of through your pipe.
- After the flush(), you close the pipe. This means that cvs at the other
  end has no stdin or stdout any more.
  On unix, this usually means that the process is killed. This may even be
  happening before PASSWORD is received at the other end.

You may want to have a look at the pexpect package, which is built to do
what you are trying to do (i.e. running an interactive session through a
pipe). I have very good experiences with it.



Albert
-- 
Unlike popular belief, the .doc format is not an open publically available format.



From gh at ghaering.de  Sat Apr 19 05:28:57 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Sat, 19 Apr 2003 11:28:57 +0200
Subject: Python and SSL sockets
In-Reply-To: <LIELKILHEABECBGHPOOJCEEHCIAA.johnson@vdc.equadriga.com>
References: <LIELKILHEABECBGHPOOJCEEHCIAA.johnson@vdc.equadriga.com>
Message-ID: <3EA116D9.3020603@ghaering.de>

Elester Johnson wrote:
> Hi Friends,

People posting 9 KB of HTML (and in your case, even imgages, too) for a 
two-line question to a mailing list are definitely NOT my friends.

Please stop sending that junk NOW.

> How do I transfer an HTML file from a windows machine to a linux 
> machine (server), using SSL sockets.

I'd suggest to try uploading the file via HTTPS. Because Pre-2.3 Windows 
versions of Python aren't built with SSL support, you'll need my 
_socket.pyd replacement:

http://pypgsql.sourceforge.net/misc/python-ssl.html

-- Gerhard




From peter at engcorp.com  Wed Apr  2 06:20:03 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 06:20:03 -0500
Subject: i cant remember that python regex viewer
References: <pan.2003.04.02.11.03.07.614824@asd.com>
Message-ID: <3E8AC763.C89B8B03@engcorp.com>

John Does wrote:
> 
> i cant remember that python regex viewer:

I suspect you mean this:

-------- Original Message --------
From: stuff at nonpoint.mailzilla.net (Phil)
Subject: Kodos 0.8 - a Regular Expression GUI tool
Newsgroups: comp.lang.python.announce

Kodos 0.8 has been released.

Kodos is a Python GUI for creating, testing and debugging regular
expressions for the Python programming language.

More information, documentation and screenshots are available here:
http://kodos.sourceforge.net/

Kodos can be downloaded from it's Sourceforge project page:
https://sourceforge.net/projects/kodos/



From jane.doe at acme.com  Sat Apr  5 09:35:44 2003
From: jane.doe at acme.com (Jane Doe)
Date: Sat, 05 Apr 2003 16:35:44 +0200
Subject: [Newbie] Opinions on GUI and distribution
References: <17nr8vgmif2df7dr8dd5saut781b1fr044@4ax.com> <mailman.1049489168.24646.python-list@python.org>
Message-ID: <8aqt8vg87rj0se2jsde8vsmdp8r33189oe@4ax.com>

On Fri, 4 Apr 2003 15:45:14 -0500, Jp Calderone <exarkun at intarweb.us>
wrote:
>  Easy.  Distutils, py2exe, McMillian Installer, and several others give you
>a wide variety of installation approaches.

I'm on my way... I checked the FAQ for this ng before posting, but
only found some old ones on www.faqs.org.

>  Some of them are, certainly.  I think that this is an area where the
>differences tend to come down mainly to personal taste.

As long as the main widgets are available, either by calling Windows'
native set or installing some custom ones, our customers should be
happy. Provided custom sets don't come in big install programs.

>Bottom line, I doubt you'll find an application for which Python's runtime performance disqualifies it as a
>possible implementation language.

Great :-)

Thx a bunch
JD.



From oblivious at web.de  Fri Apr 11 02:12:13 2003
From: oblivious at web.de (Dominic)
Date: Fri, 11 Apr 2003 08:12:13 +0200
Subject: Macros in Python?
In-Reply-To: <m2znmymbaj.fsf@yahoo.com>
References: <b71qde$q71$07$1@news.t-online.com> <slrnb9b01m.108.do-not-spam-ben.hutchings@tin.bwsint.com> <b74bem$oi3$02$1@news.t-online.com> <m2znmymbaj.fsf@yahoo.com>
Message-ID: <b75m9b$kr$07$1@news.t-online.com>

Johann Hibschman wrote:
> Dominic <oblivious at web.de> writes:
> 
> 
>>>Alternatively, you can curry the above function:
>>>    def with_root_privs(func):
>>>        def call(func, *args, **kwargs):
>>>            save_euid = os.geteuid()
>>>            os.seteuid(0)
>>>            try:
>>>                func(*args, **kwargs)
>>>            finally:
>>>                os.seteuid(save_euid)
>>>        return lambda *args, **kwargs: call(func, *args, **kwargs)
> 
> 
>>Thanks. This one is also very good. :-)
>>Interesting idea to use a closure to build
>>a function wrapper.
> 
> 
> Hm.  Why not just do:
> 
> def with_root_privs(func):
>     def call(*args, **kwargs):
>         save_euid = os.geteuid()
>         os.seteuid(0)
>         try:
>             func(*args, **kwargs)
>         finally:
>             os.seteuid(save_euid)
>     return call
> 
> I don't see what that extra layer of indirection is gaining you.
Well, you could use the lambda construct to create
parametrized methods. In this example it does not matter
though.
Like a Lisp closure the lambda expression binds the
variables within the function call. Unlike Lisp
they cannot be changed lateron as far as I know.


def make_adder(value):
    def add(a,value):
       return a+value
    return lambda a: value+x

add3=make_adder(3)
print add(3)

I haven't tried it, but should work.

Ciao,
  Dominic


> 
> -Johann




From frobozz_electric at hotmail.com  Sun Apr 27 11:19:22 2003
From: frobozz_electric at hotmail.com (Sean Ross)
Date: Sun, 27 Apr 2003 11:19:22 -0400
Subject: Some Python Code
References: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com> <mailman.1051378901.13527.python-list@python.org> <2UBqa.5766$2g5.860168@news20.bellglobal.com>
Message-ID: <2ySqa.135$Jy3.40861@news20.bellglobal.com>

Hi.

For anyone who downloaded the Genetic Algorithm zip file from my site,
please note: The GAGraph.py file that was bundled with that package was a
very early, non-functioning version of the program that was included by
mistake. This has now been replaced by the working code. It's not
impressive, but it did what I needed it to do.
Sorry for any inconvenience

Sean





From ckea25d02 at sneakemail.com  Wed Apr  2 19:43:15 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 03 Apr 2003 12:43:15 +1200
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <mailman.1049296805.12278.python-list@python.org>
References: <mailman.1049296805.12278.python-list@python.org>
Message-ID: <b6fvj9$518n3$1@ID-169208.news.dfncis.de>

sismex01 at hebmex.com wrote:
> No, the correct writing of 39 in roman numerals is "XXXIX",
> the proposed rule for only 3 of X, I or C in a single number
> is premature, or incomplete.
> 
> It should read "3 consecutive X, I or C digits".

You shouldn't need to explicitly state any such rule
if the rest of the rules are formulated properly.

The best way would be to specify a grammar in some
form, e.g. a regular expression.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ramen at lackingtalent.com  Thu Apr 24 04:43:55 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Thu, 24 Apr 2003 08:43:55 -0000
Subject: In praise of PythonCard
References: <3EA1CB22.4010002@earthlink.net> <mailman.1050793771.18312.python-list@python.org> <QxAoa.540832$3D1.296696@sccrnsc01> <a5415304.0304210544.1342b436@posting.google.com> <3EA490F5.30606@removethis.free.fr>
Message-ID: <slrnbaf8vd.gb9.ramen@lackingtalent.com>

In article <3EA490F5.30606 at removethis.free.fr>, laotseu wrote:
> [..snip parallel between pythoncard and smalltalk..]
> 
> (oh shit, I'm sure I've just said another stoopidity that will end up in 
> a long cross-post with c.l.smalltlak now)

LOL - I'm surprised nobody's invoked the Mac vs. PC holy war yet. You guys
know that PythonCard is a rehash of Macintosh technology, right? <running>

Sounds great! What's a GUI?

Cheers,
Dave



From danny at iamlearning.com  Wed Apr  2 13:40:19 2003
From: danny at iamlearning.com (danny at iamlearning.com)
Date: Wed, 2 Apr 2003 12:40:19 -0600 (CST)
Subject: How to compile a python script to *.pyo
Message-ID: <2977.143.166.82.240.1049308819.squirrel@iamlearning.com>

I'm new on this, please bear with we.

Let's say I have this script:

--------------------------
#!/usr/bin/env python
x = 'Hello Python'

print x,

--------------------------

How can I compile this script to *.pyo?

Thanks in advance,

Danny






From peter at engcorp.com  Thu Apr 10 08:57:04 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 10 Apr 2003 08:57:04 -0400
Subject: newbie: strange python problem
References: <pan.2003.04.10.12.27.06.600319@somedomain.com>
Message-ID: <3E956A20.E3321A5E@engcorp.com>

maxx wrote:
> 
> Simple code (I think):
> for line in os.popen('ls -la /').readlines():
>     direntry = line.split()
>     print direntry
>     print direntry[0]
>     print direntry[1]
>     print direntry[2]
> ---
> 
> No list-items with an index above 1 are printed. They produce a
> 'IndexError: list index out of range' error... which is strange, because
> the list 'direntry' is 9 items large, so listitems up to index 8 should be
> printed... right?

Check the output of "ls -la /" manually... the first line does
*not* have more than two items in it...  if you were to change
the first line above to this, it would work, though I suspect
there is another option to "ls" which would also suppress
that extra output.

> for line in os.popen('ls -la /').readlines()[1:]:

-Peter



From claird at lairds.com  Fri Apr  4 09:30:50 2003
From: claird at lairds.com (Cameron Laird)
Date: Fri, 04 Apr 2003 14:30:50 -0000
Subject: computer science with python
References: <m3istvtodl.fsf@moroni.pp.asu.edu> <slrnb8qjjf.i81.rmunn@localhost.localdomain>
Message-ID: <v8r5oqot9tk7a2@corp.supernews.com>

In article <slrnb8qjjf.i81.rmunn at localhost.localdomain>,
Robin Munn  <rmunn at pobox.com> wrote:
			.
			.
			.
>I've never read _Computer Science with Perl_, so I don't know how
>similar or dissimilar this book might be, but _How to Think Like a
>Computer Scientist: Learning with Python_ has to be pretty close to what
>you're looking for, at least. You can find it at:
>
>    http://www.ibiblio.org/obp/thinkCSpy/
			.
			.
			.
In format and approach, they have almost nothing to do with each
other; <URL: http://www.unixreview.com/documents/s=7822/ur0302l/ >
has a few details.  The high quality of editing both enjoy con-
verges their effect to a surprising degree.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From Peter.Havens at Level3.com  Fri Apr 11 13:21:42 2003
From: Peter.Havens at Level3.com (Havens, Peter)
Date: Fri, 11 Apr 2003 11:21:42 -0600
Subject: Confusion: assignments create object references.
Message-ID: <0C642BE4EE085149B52839125D4FC9D41AD515@idc1exc0002.corp.global.level3.com>

// -----Original Message-----
// From: Mike C. Fletcher [mailto:mcfletch at rogers.com]
//
// Because Python isn't C ;) .  Seriously, what you want to 
// think about is namespaces and names.  A name is a simple 
// mapping from a string to a (smart (i.e. reference-counted))
// PyObject * (under the covers).  That is, a name holds a 
// reference (pointer) to an object.  That's all it ever 
// holds a reference to.  It never holds a reference to another name  
// "slot"(though it can hold the same string value as a name, 
// but that's still just a string object).

Eureka! This makes perfect sense to me now!! Thank you kindly for your help. Genius makes the incredible seem obvious. You're a genius :>

Pete




From adalke at mindspring.com  Fri Apr 11 13:43:24 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Fri, 11 Apr 2003 11:43:24 -0600
Subject: Python and Schools
References: <qGqla.61$Tj7.91655@news.uswest.net> <pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de> <Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134> <14e4hpwd3g3bf$.1qbn5rtun7k7j$.dlg@40tude.net> <3E96D270.B8E99B27@engcorp.com>
Message-ID: <b76ult$plu$1@slb5.atl.mindspring.net>

Mads Orbesen Troest wrote:
> > but really; all the self me here and self me there is - to me at last -
> > exceedingly clumsy and a constant source of frustration.

Peter Hansen
> >>> import this

 ... from the interpreter and you'll see a list of 'Pythonic' guidelines.
More specifically, "explicit is better than implicit".  See the 'self'
and you know where the variable comes from.  Compare to C++
where you don't know (without some looking) if the variable is
an attribute or a variable in the compilation unit.

                    Andrew
                    dalke at dalkescientific.com





From peter at engcorp.com  Tue Apr  8 19:11:46 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 08 Apr 2003 19:11:46 -0400
Subject: help modules Python
References: <mailman.1049747592.21028.python-list@python.org> <b6uu8h$dom$1@slb3.atl.mindspring.net>
Message-ID: <3E935732.4E411863@engcorp.com>

Matt Gerrans wrote:
> 
> Looks like you just need the latest Python (I think object was introduced
> with version 2.0, but others on this list would know more precisely [snip]

Even better, *anyone* can find this out by checking out Andrew Kuchling's 
pages here:

  http://www.amk.ca/python/

-Peter



From gh at ghaering.de  Thu Apr 17 13:58:43 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Thu, 17 Apr 2003 19:58:43 +0200
Subject: Downloading TV listings with urllib
In-Reply-To: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4>
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4>
Message-ID: <3E9EEB53.1010207@ghaering.de>

Josh wrote:
> Can someone give me general pointers on how it could be done? I am trying 
> to go to tvguide.com and download the TV listings for my area. The problem 
> is how to specify my location and cable provider. I used Proxomitron to 
> look at the traffic generated while downloading the listings with my 
> browser. When I used httplib with:
> 
> host = 'www.tvguide.com'
> pathn = '/Listings/index.asp'
> 
> def tvg():
>     try:
>         h = httplib.HTTPConnection(host)
>         h.putrequest('GET', pathn)
>         h.putheader('Accept', 'text/html')
>         h.putheader('Accept', 'text/plain')
>         h.putheader('User-Agent:', 'Mozilla/5.0 (Windows; U; Windows NT 
> 5.1; en-US; rv:1.3) Gecko/20030312')
>         h.putheader('Accept:', 'application/x-shockwave-
> flash,text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/p
> lain;q=0.8,video/x-mng,image/png,image/jpeg,image/gif;q=0.2,*/*;q=0.1')
>         h.putheader('Accept-Language:','en-us,en;q=0.5')
>         h.putheader('Accept-Encoding:', 'gzip,deflate,compress;q=0.9')
>         h.putheader('Accept-Charset:', 'ISO-8859-1,utf-8;q=0.7,*;q=0.7')
>         h.putheader('Keep-Alive:', '300')
>         h.putheader('Cookie', 
> 'SITESERVER=ID=fd7d0ca25acdd2439ab3f4b48b0827b4; GBA=18; 
> TVGID=9CE64992B85F4F9A82B57EBC092D11B2; nat=0; ServiceID=75347; zip=63130; 
> ptfc=yfki')
>         h.endheaders()
>         err = h.getresponse()
>     	  return err
>     except socket.error, er:
>         print 'socket error ', er
> 
> when I use err.read() to read the object returned I get everything in Hex.

My guess is that you get what you're asking for - gzip-compressed data.

-- Gerhard




From mwh at python.net  Thu Apr 10 06:23:09 2003
From: mwh at python.net (Michael Hudson)
Date: Thu, 10 Apr 2003 10:23:09 GMT
Subject: a size-intelligent progress bar?
References: <1PYka.1140$RX1.361@newsfep3-gui.server.ntli.net> <mailman.1049945595.1665.python-list@python.org>
Message-ID: <7h3y92izm5v.fsf@pc150.maths.bris.ac.uk>

Steven Taschuk <staschuk at telusplanet.net> writes:

> Quoth Tom Chance:
> > I'm making a shell-based program that includes a progress bar (like wget's
> > one). At the moment, I've got it printing an equals sign every 2% so I get
> > a progress bar 50 "="s wide, which fits my Konsole perfectly. This isn't
> > very handy for other shell sizes though :)
> > 
> > Does anyone have, or can anyone point me in the direction of a way of making
> > a decent progress bar that will automagically detect the width of the shell
> > open, even resizing it mid-"progress"?
> 
> I think the way to determine the width of the current terminal is
> terminal-dependent.  You could use curses.

Or check environment variables, or use ioctl.  I've posted code to do
this many, many times to this newsgroup, though not for a while.

> If memory serves, when a window gets resized the process(es?)
> associated with it gets sent a WINCH signal.  I don't know whether
> curses traps this for you; if so, just keep rechecking the
> terminal width via curses at appropriate moments.

Just checking the terminal size each time you redraw the progress bar
should be sufficient.

Cheers,
M.

-- 
  It is time-consuming to produce high-quality software. However,
  that should not alone be a reason to give up the high standards
  of Python development.              -- Martin von Loewis, python-dev



From db3l at fitlinxx.com  Tue Apr 29 14:56:45 2003
From: db3l at fitlinxx.com (David Bolen)
Date: 29 Apr 2003 14:56:45 -0400
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org> <1bf4033f.0304290627.40f7d831@posting.google.com> <uptn56xxi.fsf@fitlinxx.com> <mailman.1051638646.15981.python-list@python.org>
Message-ID: <uhe8h6sk2.fsf@fitlinxx.com>

"andrew cooke" <andrew at acooke.org> writes:

> komodo has a gui builder these days (don't know how good it is).
> http://www.activestate.com/Products/Komodo/more_information.plex

Good point, I should have mentioned Komodo.  For me, the GUI drawback
is that it's Tk based (so same look and feel issue as I had with
Pythonworks).

But Komodo has an even bigger strike in my testing - it's still just a
huge user of system resources, even in its latest release, which has
definitely improved from earlier releases.  Not sure how much of this
is due to using the Mozilla base or not.

I won't claim load times are anywhere near the best measure, but it's
one data point - at least on my machine it takes over 45s (and 30+MB
of memory) just to start up (20+s even on immediately subsequent
attempts).  That compares to 10s (10MB) for VB 6.0 and 9s (11MB) for
Visual Studio .NET (on first loads), not to mention the other Python
IDEs I've tried.

-- David



From tjreedy at udel.edu  Sat Apr 19 21:15:23 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Sat, 19 Apr 2003 21:15:23 -0400
Subject: Is this a super bug?
References: <mailman.1050791852.9118.python-list@python.org>
Message-ID: <cO2dnb3olutAbDyjXTWcoQ@comcast.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> wrote in message
news:mailman.1050791852.9118.python-list at python.org...
M:\python>python
Python 2.3a2 (#39, Feb 19 2003, 17:58:58) [MSC v.1200 32 bit (Intel)]
on
win32
Type "help", "copyright", "credits" or "license" for more information.
>>> class ff(str):
...    def foo(self):
...      print 'str.__getitem__', str.__getitem__(self, 0)
...      print 'super(ff, self).__getitem__', super(ff,
self).__getitem__(0)
...      print 'super(ff, self)[0]', super(ff, self)[0]
...
>>> f = ff('asdf')
>>> f.foo()
str.__getitem__ a
super(ff, self).__getitem__ a
super(ff, self)[0]
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "<stdin>", line 5, in foo
TypeError: unsubscriptable object
----------
I suspect that printing type(super(ff,self)) might answer your
question.

TJR





From spam at magnetic-ink.dk  Wed Apr  9 03:08:09 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Wed, 9 Apr 2003 07:08:09 +0000 (UTC)
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <6268469c-8aa0-4573-a457-15b0b5bff8cb@news.szn.dk>

Jim Shady wrote:

> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string.
> For example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?

E.g.:

#v+

def longest_str(s):
    cache = {}
    words = s.split('/')
    for word in words:
        wlen = len(word)
        if cache.has_key(wlen):
            cache[wlen].append(word)
        else:
            cache[wlen] = [word]
        # end if
    # end for
    return cache[max(cache.keys())]
# end def longest_str

#v-

Apply sanity checks as needed.


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From maxm at mxm.dk  Thu Apr 24 03:16:25 2003
From: maxm at mxm.dk (Max M)
Date: Thu, 24 Apr 2003 09:16:25 +0200
Subject: Data-driven testing
In-Reply-To: <3EA74EC0.4BAB336@engcorp.com>
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com>
Message-ID: <X8Mpa.30197$y3.2398527@news010.worldonline.dk>

Peter Hansen wrote:

> For throwaways, I don't write quite so many tests, and in many cases if 
> it's only a "small" throwaway script (e.g. one or two subroutines only, 
> plus a little mainline) I will write a test or two for the subroutines 
> (because I'm doing TDD, not because I want to have unit tests to make 
> sure my script works properly... if you see the distinction) but never
> for the mainline.  


While I do understand the principles of tdd, testing first and all that, 
I have a hard seing why it should be an advantage to do it in the early 
stages of a project.

Normaly if a am entering a new domain, where I don't have a clear 
picture of the requirements, I try out different approaches. Each 
approach using a different combination of ideology, objects and functions.

This is all to get a feel of the best approach in this particular 
domain. Some times I try out a *lot* of approaches before I end up with 
a particular elegant design.

At which point I am ready to "solve the problem". It is at this point I 
can see the point of unit testing.

In tdd do you really use testing for the first "creative" fase also?


-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From python at dylanreinhardt.com  Wed Apr 23 21:22:47 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 23 Apr 2003 18:22:47 -0700
Subject: *Newbie Question* --> Is it viable to distribute Python apps
	over the Internet?
In-Reply-To: <25a672ee.0304231444.1ce0dd1b@posting.google.com>
References: <25a672ee.0304210914.1cc2049c@posting.google.com>
	 <3ea434fd$0$180$a1866201@newsreader.visi.com>
	 <25a672ee.0304211504.4f3cd067@posting.google.com>
	 <mailman.1050968618.16005.python-list@python.org>
	 <va956b925cj1df@corp.supernews.com>
	 <25a672ee.0304231444.1ce0dd1b@posting.google.com>
Message-ID: <1051147367.22265.37.camel@ida.dylanreinhardt.com>

On Wed, 2003-04-23 at 15:44, Denis Agar wrote:
> Ok, try to ignore the thickness of my skull, but you're saying that it
> is possible to create progams using mfc42.dll, a GUI that most people
> have already? I have the win32 extensions and am in the process of
> trying to understand them. (PS: Anyone got some (s)ample code they
> could throw my way?)

You might find this walkthrough helpful/useful/amusing:

http://www.onlamp.com/pub/a/python/excerpts/chpt20/?page=5

Dylan





From mwh at python.net  Mon Apr 14 02:52:55 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 14 Apr 2003 06:52:55 GMT
Subject: Implementing Metatypes in C
References: <mailman.1050275480.8590.python-list@python.org>
Message-ID: <7h34r51r2nt.fsf@pc150.maths.bris.ac.uk>

Phil Thompson <phil at river-bank.demon.co.uk> writes:

> 1. Is my understanding correct, or is there a more correct way to do what I 
> want that avoids the problem?

Not sure -- I *think* you're correct.

> 2. If my solution is correct, can the structure defined in typeobject.c be 
> published in a header file as the comment says? Please?

Already done for 2.3.

Cheers.
M.

-- 
  Not only does the English Language borrow words from other
  languages, it sometimes chases them down dark alleys, hits
  them over the head, and goes through their pockets.   -- Eddy Peters



From mertz at gnosis.cx  Fri Apr 11 13:45:48 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Fri, 11 Apr 2003 13:45:48 -0400
Subject: Confusion: assignments create object references.
References: <mailman.1050078618.3242.python-list@python.org>
 </6vl+kKkX4RI092yn@gnosis.cx>
Message-ID: <M9vl+kKkXkqW092yn@gnosis.cx>         

"Havens, Peter" <Peter.Havens at Level3.com> wrote previously:
|I am confused about when assignments create references to objects as
|opposed to copying the value of those objects.

This is an easy one to answer:  Assignment ALWAYS creates references to
objects (in Python).

Yours, Lulu...

--
    _/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: Postmodern Enterprises _/_/_/
   _/_/    ~~~~~~~~~~~~~~~~~~~~[mertz at gnosis.cx]~~~~~~~~~~~~~~~~~~~~~  _/_/
  _/_/  The opinions expressed here must be those of my employer...   _/_/
 _/_/_/_/_/_/_/_/_/_/ Surely you don't think that *I* believe them!  _/_/





From km-news1 at onlinehome.de  Thu Apr 24 18:19:18 2003
From: km-news1 at onlinehome.de (Klaus Meyer)
Date: Fri, 25 Apr 2003 00:19:18 +0200
Subject: Tkinter - tkFileDialog - need help
Message-ID: <oprn5mygrymu3s0q@news.cis.dfn.de>

Hi,

my script uses a directory-select box.
But it's not possible to find an select a directory that has attribut 
"hidden" inside the dialog-box (askdirectory).

It's possible to make "hidden" directory viewable? How? Thanks!

Win2000:

import tkFileDialog
tkFileDialog.askdirectory(initialdir="C:/", title="Test")


-- 
Gru? - regards Klaus :-)



From bvdpoel at kootenay.com  Mon Apr  7 17:26:09 2003
From: bvdpoel at kootenay.com (Bob van der Poel)
Date: Mon, 07 Apr 2003 14:26:09 -0700
Subject: Python & Music: Performance question
In-Reply-To: <ehkka.6614$y3.832923@news010.worldonline.dk>
References: <1759581e.0304060218.164d4cd4@posting.google.com> <1759581e.0304070549.3b3a4caa@posting.google.com> <3E91C3D9.E274C521@easystreet.com> <ehkka.6614$y3.832923@news010.worldonline.dk>
Message-ID: <3e91ed7c_1@dns.sd54.bc.ca>

Max M wrote:
> achrist at easystreet.com wrote:
> 
>>Is there anything in Python like the Midi support in the java
>>sdk?  (under MS Windows) It's pretty easy in java to send midi data 
>>straight from the program to the midi sequencer or synthesizer
>>(no need to write a file or deal with some external player).   
>>Anything like that for Python?
> 
> 
> Not really no. Especially not something that is cross-platform.
> 
> It really is a pitty, as Python would be exceptionally good for 
> algorithmic composition.
> 

Just to get a plug in for my program, MMA, 
http://www.kootenay.com/~bvdpoel/music.html, I had the same 
considerations. I ended up just having the program create midi files and 
use an external player to play them. But, it would be nice to have a 
python "play this as a midi stream" function.

-- 
Bob van der Poel ** Wynndel, British Columbia, CANADA **
EMAIL: bvdpoel at kootenay.com
WWW:   http://www.kootenay.com/~bvdpoel





From DavidMLewis.NO at SPAM.mac.com  Tue Apr 15 16:41:42 2003
From: DavidMLewis.NO at SPAM.mac.com (David Lewis)
Date: Tue, 15 Apr 2003 20:41:42 GMT
Subject: Printing Columns
References: <mailman.1050426466.11198.python-list@python.org>
Message-ID: <150420031341416871%DavidMLewis.NO@SPAM.mac.com>

In article <mailman.1050426466.11198.python-list at python.org>, Lindstrom
Greg - glinds <Greg.Lindstrom at acxiom.com> wrote:

> Hello-
> I have been working on a little project and have come across an interesting
> (to me) problem.  I have a list of stuff that I would like to print to
> stdout in columns.  Now, it's no problem to print it out as follows:
> 
> myList = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H']
> 
>  A  B  C
>  D  E  F
>  G  H
> 
> but I would like
> 
>  A  D  G
>  B  E  H
>  C  F
> 
This should give you a start:

>>> x = ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h']
>>> x[0: :3]
['a', 'd', 'g']
>>> x[1: :3]
['b', 'e', 'h']
>>> x[2: :3]
['c', 'f']
>>> 

Best,
   David



From staschuk at telusplanet.net  Thu Apr 17 18:41:29 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 16:41:29 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <b7n4hk$2g1u$1@nntp6.u.washington.edu>; from donn@u.washington.edu on Thu, Apr 17, 2003 at 08:58:28PM +0000
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu>
Message-ID: <20030417164129.D10011@tibia.amotlpaa.bogus>

Quoth Donn Cave:
> Quoth Steven Taschuk <staschuk at telusplanet.net>:
> | Quoth Tim Peters:
> |> [...] What to do with each
> |> method is in large part art, and the teachable seem to do OK learning by
> |> example.  [...]
> |
> | I wonder, then, whether the docs would benefit from being
> | augmented with a few examples.
> |
> | Donn -- would that improve their clarity in your view?
> 
> No, but if _replaced_ with a few examples, maybe.  I mean, if you
> accept my contention that the documentation misrepresents the
> actual usage of these functions, it's hard to remedy that with
> an example.  Unless maybe hostile examples, like a repr that
> spews out pages of state, but I assume that would not be the idea.

*chuckle*  Indeed.  (Though in my coding "pages of state" is very
uncommon; it really only occurs for large containers, which have
very simple internal structure.)

Looking back over your posts in this thread, I see a few kinds of
objection to the current documentation.

First, there's the description of __repr__ and __str__ in terms of
fuzzy properties their output should have -- "informal" vs
"official", "convenient" vs "complete", and the like; you feel
that these descriptions are unclear.  For this problem, a few
examples would certainly help to clarify the intended meanings of
these terms.  It might also help to give examples of where these
representations would normally be used -- for debugging output,
for normal logging, for display to the user of the application.

Second, I have some sense that you feel these terms are not merely
unclear but misleading or wrong.  Is this true?  If so, do you
have alternative fuzzy terms in mind (the "data" vs the "object"),
or would you prefer that the documentation avoid fuzzy terms
entirely, and only give examples?

Third, you (and others) specifically object to the notion that if
possible one should have eval(repr(x)) == x.  Would you want this
notion to be omitted entirely?  Mentioned as possible and
sometimes good?  Described as highly desirable, though sometimes
infeasible?  (Imho the docs currently do this last, though perhaps
they could do so more clearly.)

Have I missed anything?

(I solicit comments from everybody, of course, not just Donn.)

  [...]

-- 
Steven Taschuk                                                 o- @
staschuk at telusplanet.net                                      7O   )
                                                               "  (




From jack at performancedrivers.com  Tue Apr  1 15:31:11 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Tue, 1 Apr 2003 15:31:11 -0500
Subject: PEP 313 - Roman numeral literals
In-Reply-To: <Pine.LNX.4.51.0304011257400.15952@ns0.tbye.com>; from dave@pythonapocrypha.com on Tue, Apr 01, 2003 at 01:01:14PM -0800
References: <x7y92uhzpw.fsf@guru.mired.org> <Pine.LNX.4.51.0304011257400.15952@ns0.tbye.com>
Message-ID: <20030401153111.B936@localhost.localdomain>

On Tue, Apr 01, 2003 at 01:01:14PM -0800, Dave Brueck wrote:
> On Tue, 1 Apr 2003, Mike Meyer wrote:
> 
> > PEP: 313
> > Title: Adding Roman Numeral Literals to Python
> 
> Before we submit this to the BDFL, I propose that we wait a short while to
> give the community an opportunity to debate its merits, followed by a
> vote, just so Guido can get idea of where the community stands. Further, I
> propose that Aahz be the official vote counter. Does anybody have any
> preferences as to the voting method we should use?

Yes, I propose we reuse an idea from the ternary vote.
In order to demostrate understanding of the proposal your preferences should
be listed in order, with the enumeration done in valid roman numerals.

-jackdied




From max at alcyone.com  Wed Apr 16 21:56:34 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 18:56:34 -0700
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com> <3e9daad2$0$49115$e4fe514c@news.xs4all.nl> <f7199550.0304161732.48c75c42@posting.google.com>
Message-ID: <3E9E09D2.F3BF1F5F@alcyone.com>

Peter Clark wrote:

>     Since the output is meant to be read to be displayed by a font
> which is in essentially latin-1 encoding, I need to restrict the
> manner in which the degree symbol is displayed to one byte. Yet I
> cannot get it to behave, even though 'print chr(176) works perfectly
> fine at the prompt. My suspicion is that the default encoding of the
> system is messing python up somewhere along the way--is there any way
> to tell it to just print the stupid character and not be concerned
> with the output?

I've come into this conversation late, but could it be that what's
confusing you is that UTF-8 and Latin-1 are not the same thing?  It
sounds like you want Latin-1 but are asking for UTF-8.  UTF-8 is an
octet representation of Unicode which uses escape sequences and the like
to represent eight-bit information; Latin-1 is an eight-bit encoding. 
Both have the property that pure-ASCII data will be represented without
modification, but they aren't the same beast.  If you're converting to
UTF-8 and are puzzled why 8-bit data is expanding to multiple
characters, then chances are UTF-8 isn't what you wanted.

[where u is a Unicode string representing the degree symbol]
>>> u.encode('latin-1')
'\xb0'
>>> u.encode('utf-8')
'\xc2\xb0'
>>> print u.encode('latin-1')
[the degree symbol]

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ It was involuntary.  They sank my boat.
\__/ John F. Kennedy (on how he became a war hero)
    Bosskey.net: Return to Wolfenstein / http://www.bosskey.net/rtcw/
 A personal guide to Return to Castle Wolfenstein.



From drs at ecp.cc  Tue Apr 22 18:54:52 2003
From: drs at ecp.cc (drs)
Date: Tue, 22 Apr 2003 22:54:52 GMT
Subject: Can someone send me the latest Gadfly?
References: <200304220934.38013.scott_list@mischko.com> <mailman.1051030174.3152.python-list@python.org>
Message-ID: <0Ljpa.2547$%_3.2056417@newssrv26.news.prodigy.com>

"Gerhard H?ring" <gh at ghaering.de> wrote in message
news:mailman.1051030174.3152.python-list at python.org...
> Scott Chapman wrote:
> > www.chordate.com appears to be seriously down.  Their DNS is resolving
to
> > ns1.rackshack.net.  I called Rackshack and they don't have a zone file
for
> > chordate.com.
> >
> > Chordate hosts the Gadfly database and I would like a copy of it on my
RH8
> > system but can't download it.  Google isn't helpful in finding another
> > location to get it from.
> >
> > Can someone help out here?
>
> http://sourceforge.net/projects/gadfly
> http://sourceforge.net/project/showfiles.php?group_id=662
>

I am very confused ... You obviously did a reasonable amount of research to
find this, and you are also clearly a sophisticated internet user ... a
google search on "gadfly python" turned up the db as the second result, and
an informational page linking to the db as the first.  what google searches
did you try?

-d





From guy at obstruction-no-spam.com  Sat Apr 19 16:12:16 2003
From: guy at obstruction-no-spam.com (Guy Middleton)
Date: Sat, 19 Apr 2003 20:12:16 GMT
Subject: arg-passing style question
Message-ID: <A4ioa.87812$Vzu.36931@news02.bloor.is.net.cable.rogers.com>

I want to pass two values to a function that expects a single argument.

So, I could wrap them in either a list or a tuple.  Is either way preferred
over the other?


queue = Queue.Queue()
a = "foo"
b = {"key": "value"}
queue.put((a,b))	# this works
queue.put([a,b])	# this works too



From robin at jessikat.fsnet.co.uk  Thu Apr 24 08:09:14 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Thu, 24 Apr 2003 13:09:14 +0100
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org>
 <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk>
 <mailman.1050587068.2733.python-list@python.org>
 <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com>
 <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3ea7b12a@news.012.net.il>
 <7h3k7dkkwko.fsf@pc150.maths.bris.ac.uk>
Message-ID: <Rpu197AqP9p+Ew5w@jessikat.demon.co.uk>

In article <7h3k7dkkwko.fsf at pc150.maths.bris.ac.uk>, Michael Hudson
<mwh at python.net> writes
>"Avner Ben" <avner at skilldesign.com> writes:
>
>> The first language I learnt was COBOL. That took even longer to overcome!
>
>Of coure, Dijkstra has a quote about that, too.  Let me prod the
>sigmonster:
>
I think that key board layout is also a pain, but I disagree that things
can't be unlearned. I got used to having the esc key on the right hand
side of the keyboard and then switched, it was painful, but eventually
my viing got better again. Those reflex things are hard, but certainly
seem to be conquerable. I'm having the same sort of experience moving
back and forth to a Mac wrt ctrl-C vs splat-C etc 
-- 
Robin Becker



From aahz at pythoncraft.com  Fri Apr  4 12:32:47 2003
From: aahz at pythoncraft.com (Aahz)
Date: 4 Apr 2003 12:32:47 -0500
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
Message-ID: <b6kfjv$4cc$1@panix2.panix.com>

In article <a3b5b608.0304040904.146dfff5 at posting.google.com>,
Giorgi <lekishvili at python.qartu.com> wrote:
>
>- Is Python a real web-programming language at all? I mean, has
>somebody ever used it for real, commercial tasks? Please, do not
>consider my question too trivial. There is no language equally
>suitable for all tasks, however, Python seems to me a language
>naturally tailored for web-programming; this or that way, I have more
>than two years of experience of professional programming in Python.

No, Python is not a real web-programming language (in comparison to, say,
PHP).  However, Python is a real programming language, period; with
Python, it is possible to construct useful frameworks for programming Web
applications, and Zope has served as a proof of this for many years.

The reason why there are so many solutions is because there are tradeoffs
for ease vs. power for each type of web application -- and there are
many types of web applications.  The "only one way" Pythonic principle
just doesn't apply.

Probably the "standard" answers these days for the basic building block
of a web application is either Zope (if you want a full-blown framework
for content management) or Quixote (if you want a relatively simple
system that feels like programming in Python).
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From andersjm at dancontrol.dk  Mon Apr  7 03:14:47 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Mon, 7 Apr 2003 09:14:47 +0200
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org> <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk> <x7el4jibzl.fsf@guru.mired.org> <Kp2ja.2589$y3.426276@news010.worldonline.dk> <x7znn6gkfn.fsf@guru.mired.org>
Message-ID: <3e912546$0$10390$edfadb0f@dread11.news.tele.dk>

"Mike Meyer" <mwm at mired.org> wrote:
> > No limit on programmer time?
> 
> You just have to hire more programmers.

That only works if your project isn't late <0.7 wink>

- Anders





From tjreedy at udel.edu  Sun Apr 27 01:52:30 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Sun, 27 Apr 2003 01:52:30 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com> <3EAB6A14.707AD826@alcyone.com>
Message-ID: <ahSdnaSulLiu7TajXTWcow@comcast.com>

"Erik Max Francis" <max at alcyone.com> wrote in message
news:3EAB6A14.707AD826 at alcyone.com...
> Jeremy Fincher wrote:
>
> > Perhaps the answer is to remove the __nonzero__ method on
> > functions/methods.  Logically, what purpose does it serve?
>
> # f holds a callable object or None
> if f:
>     f(...)

or
print f and f() or 'cannot do that' # assumes boo(f()) always True

I consider it a feature that all objects have truth value (except for
perversely written user classes), so that conditional expressions and
statements always work without exception.

TJR





From dino_levy at mindlessSPAM.com  Sat Apr  5 10:52:44 2003
From: dino_levy at mindlessSPAM.com (Dino Levy)
Date: Sat, 05 Apr 2003 17:52:44 +0200
Subject: POP3 and selective download/deletion
References: <pan.2003.04.05.12.12.25.511680.2271@mindlessSPAM.com> <pan.2003.04.05.16.22.31.832230.878@thomas-guettler.de> <luCja.11925$It5.5735@news2.central.cox.net> <86d6k1ar0u.fsf@fibble.nospam.dev.null>
Message-ID: <pan.2003.04.05.17.52.43.14616.1268@mindlessSPAM.com>

On Sat, 05 Apr 2003 17:43:45 +0200, Robert Pluim wrote:

>> Unfortunately the POP3 protocol doesn't allow you download the headers
>> of a message without downloading the whole thing. IMAP4 is much more
>> capable, but it's also a very much more complex protocol.
> 
> Que?  You've never heard of the TOP command?

Hehe, I heard of it, but it doesn't help much in this case. Steve thank
you for your link, I'll certainly find something useful there...
Does anyone have another idea?



From staschuk at telusplanet.net  Wed Apr 30 22:36:40 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 30 Apr 2003 20:36:40 -0600
Subject: Raising objects
In-Reply-To: <%OYra.27334$3M4.759346@news1.tin.it>; from aleax@aleax.it on Wed, Apr 30, 2003 at 11:16:11PM +0000
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it> <mailman.1051678851.21661.python-list@python.org> <OFLra.36816$K35.1143840@news2.tin.it> <mailman.1051719837.14426.python-list@python.org> <%OYra.27334$3M4.759346@news1.tin.it>
Message-ID: <20030430203640.B12032@tibia.amotlpaa.bogus>

Quoth Alex Martelli:
> Steven Taschuk wrote:
  [...]
> > This function does report that strings and lists are new-style.
> > Aren't they?
> 
> If by "new-style" you mean "anything EXCEPT classic classes" then
> yes, of course -- strings, lists, numbers, etc, aren't classic
> classes.  But then it's simpler to check for types.ClassType, no?

Ah yes, the type/class distinction.  I frequently forget that.

So what then is the distinction between new-style classes and
types?  (Other than that types are not quite reliably subclassable
yet, or so I hear.)

  [...]
> Moreover, if what you raise is always derived by Exception, the
> Exception base class can ensure all sort of nice error-reporting

A good point.

But usually in Python we do not care whether an object inherits
from a specific subclass; we only care whether it has the
attributes we need it to have.  Why should the error-reporting
facilities of Exception be any different?

In other words, why should we stop a programmer who *doesn't* want
those error-reporting facilities from raising some other object?
Isn't trusting the programmer pythonic?

> > So, suppose that strings were special-cased for backwards
> > compatibility:
  [...]
> > Is there a problem with this scheme?
> 
> The horrid complexity and specialcasing is problematic enough
> for me.  Allowing such generality in what is raised buys you
> nothing pragmatically useful AND you pay a price in complexity.
> No thanks!!!

I agree the pragmatic gain is negligible.  (Except for a
simplification of the rules for exceptions; I'll explain this
below.)

Forget the special-casing of strings; I overlooked that we need
the identity test (not an equality test, as I erroneously wrote)
so that
    raise ValueError # the class, not an instance of it!
will be caught by an up-stack
    except ValueError:
clause.  (My excuse for forgetting this is that I always raise
instances.)

I'd like to get rid of this; it seems conceptually wrong to me
that a class and its instances are interchangeable in this
respect.  The identity test which makes it work made sense back
when exceptions were strings (as you explained so cogently), but
imho it's now a (minor) wart for backwards compatibility.

So, what I'd like to see eventually is this:

    Raise anything you like.

    An except clause catches an exception e if and only if it
    names a type of which e is an instance.

This seems simpler than what you would want to see eventually:

    Raise any subclass of (New)Exception, or an instance of such
    a subclass.

    An except clause catches an exception if and only if it
    names either a type of which e is an instance, or an object
    which is identical to e.

(Hopefully I've correctly understood what you're saying.)  Either
of these is, of course, simpler than the present rules.

The only disadvantage of my ideal scenario that I can see is that
it introduces a new kind of "forgot the parentheses" bug;
    raise ValueError
would raise an exception which would go right past
    except ValueError:
and only be caught by the very unlikely
    except type:
or a bare except.

To get from the present rules to my ideal scenario, the identity
test has to be removed, and while we're at it, the syntax
    raise foo, bar
should go too.  (Again, it made sense back when exceptions were
strings.)  And perhaps sys.exc_* would have to be revised as well.
The path seems straightforward:

    Step 0.  Revise the Tutorial to advocate "raise ValueError(...)"
    instead of "raise ValueError, ...".  Mention the latter only as
    an old way of doing things.

    Step 1.  Issue warnings when an exception is caught by the
    identity test, and when the "raise foo, bar" syntax is used.
    (These warnings could be phased in via __future__ if desired.)

    Step 2.  Remove the identity test; catch only by the isinstance
    test.  Remove the "raise foo, bar" syntax.  Revise sys.exc_*
    if appropriate.

    Step 3.  Make Exception a new-style class.

    Step 4.  Allow anything to be raised.

(I haven't thought about the e in
    except ValueError, e:
yet.)

Hm... maybe I should write a PEP.

-- 
Steven Taschuk                                7\ 7'Z {&~         .
staschuk at telusplanet.net                        Y r          --/hG-
                                            (__/ )_             1^1`




From s.thuriez at laposte.net  Thu Apr  3 04:43:33 2003
From: s.thuriez at laposte.net (sebastien)
Date: 3 Apr 2003 01:43:33 -0800
Subject: FTP : get  dir name with space caracters embedded  + is ther something quicker than LIST
Message-ID: <b1ab591.0304030143.52422a3e@posting.google.com>

Hi ,

I have 2 problems regarding FTP methods.


My GOAL :
---------

I am looking for a specific file on a huge ftp server. The structure
of the ftp server is not known. I have to determine which are
directories and which are files so that it can recursively go into the
subdirectories to scan the files (it is multi threaded).

It is working well but I have 2 problems.


 My trouble :
-----------------

It am using the ftp.retrlines('LIST',liste.append)
and analyse which one are directories since they have the letter d at
the beginning of the line.

1) The LIST METHOD IS EXTREMLY SLOW !!!

In certain directories there is thousand of file and the LIST method
is extremely slow.
It works but is there other method to get the directories without
having to do a LIST ?

2) DIRECTORY NAME WITH SPACES  

LIST is slow but it works. One other problem is the directory name in
which there is spaces : like
"data file today".
 I can do a :
elements= line.split(" "), 

but the data before my directory name has not the same format : for
example :

.....
may 22   2001     name of directory
mar 15    02:00   other directory name
.....

I do not know how to get the directory name simply.

One method would be to get a len(elements) and if len(elements) > 9
for example on my server. All the last elements would be part of
directory name. It is not so clean since it may depend on the FTP
server configuration.

Do you have any idea ?

Thanks for having taken the time to read my mail.

Sebastien.



From aleax at aleax.it  Sat Apr 19 15:25:45 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 19 Apr 2003 19:25:45 GMT
Subject: Testing methods
References: <9d8ac130.0304190938.755d4b4c@posting.google.com> <b7s2r9$11ka$1@news.peterlink.ru>
Message-ID: <Zohoa.37253$LB6.847098@news1.tin.it>

Anton Muhin wrote:
   ...
> try:
>      foo.absent_method()
> except (AttributeError, TypeError), e:
>      print "Failed. Message @%s@" % e, "Type: %s" % e.__class__.__name__
> else:
>      print "OK"

Careful!!!  There's a LITTLE trap here.  If foo DOES have absent_method
BUT that method's body propagates one of these exceptions, this will
catch the propagated exception AS IF it had been generated in the attempt
to access / call it.  You can split the access-and-call in two to help:

try:
    bound = foo_absent_method
    if not callable(bound):
        raise TypeError, "type %s is not callable" % type(bound)
except (AttributeError, TypeError), e:
    print "Failed. Message @%s@" % e, "Type: %s" % e.__class__.__name__
else:
    bound()
    print "OK"

this isn't QUITE satisfactory here, because, when compared to your
code, it doesn't correctly catch the case where foo.absent_method IS
both present and callable BUT requires at least one mandatory
argument.  Unfortunately, catching that particular case (which is
going to raise TypeError when bound is called) requires analyzing the
traceback to see if the TypeError is raised right at the call operation
or INSIDE the execution of bound -- too much trouble to be worth it,
IMHO.  Opinions may differ on this.  But I want to emphasize the
need to keep AS LITTLE AS POSSIBLE inside the try clause, to avoid
accidentally "masking" unexpected errors that should propagate; in
particular, it's OFTEN worth splitting an access-and-call such as

    someobject.amethod(args)

into a just-access-it part to keep inside the try clause:

    bound = someobject.amethod

and a separate just-call-it part to keep inside the else clause:

    bound(args)

so that errors caused INSIDE the body of bound don't accidentally
get trapped and reported as if they were something else (i.e. as
if they were errors caused by ACCESSING someobject.amethod).


Alex




From eltronic at juno.com  Tue Apr 29 09:50:33 2003
From: eltronic at juno.com (eltronic at juno.com)
Date: Tue, 29 Apr 2003 09:50:33 -0400
Subject: Tkinter exit problem in Win98
Message-ID: <20030429.095109.-177423.0.eltronic@juno.com>

> > Has anyone encountered a problem with Tkinter apps 
> shutting down properly in Wi98?

use pythonw.exe  yourTk.py
or python.exe -i yourTk.py
or ren someTk.py someTk.pyw which opens with pythonw

ive changed my double click open of py files to an editor
too avoid the problem. you can also add 
python.exe -i %1 %*  somewhere in the registry
to always open py file in the interactive mode.

if you python -i some.py from a dos window
when the program finishes your left with a prompt
>>>
if you ^Z  control+Z to exit python you also exit the
dos window. that seems too harsh.
the pdb debugger is much better, it accepts a q to quit.


> That is the oldes Python-Bug, see www.python.org/sf/216289
> Nobody has been able so far to study in detail what is causing
> the problem and how it could be fixed.

is it still a problem in python2.3 with Tk8.4, anyone?






e

________________________________________________________________
The best thing to hit the internet in years - Juno SpeedBand!
Surf the web up to FIVE TIMES FASTER!
Only $14.95/ month - visit www.juno.com to sign up today!




From peter at engcorp.com  Mon Apr  7 23:13:54 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 07 Apr 2003 23:13:54 -0400
Subject: How to protect python code ?
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <v938h26mdl9s8f@corp.supernews.com>
Message-ID: <3E923E72.AA57E192@engcorp.com>

Cameron Laird wrote:
> 
> Gratuitous factual tangent:  'might surprise folks
> how often acquirers of software companies essentially
> discard source code.  They want customer lists and
> relations, and occasionally the engineering staff,
> but existing programs?  Those are nearly valueless.

Source code with decent unit and acceptance tests is an asset.
Source code without decent tests is a liability.

-Peter Hansen



From lekishvili at python.qartu.com  Mon Apr 21 13:13:16 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 21 Apr 2003 10:13:16 -0700
Subject: Python and .NET
Message-ID: <a3b5b608.0304210913.4ab7f176@posting.google.com>

Hi all!

I wonder if somebody gives me a hint where to find a tutorial (well,
sort of) of how to integrate Python-based modules into a .NET
application.
More preciously, I would like to have _model_ written in Python (as I
have done the lib), while _controller_ and _view_  should be in C#.

Any hints would be gratefully acknowledged.

Best wishes,
Giorgi

P.S. I have found Kobra. I suppose this is the only option for today.
Unfortunately, I couldn't download it from the following URL:
http://home.attbi.com/~chetangadgil//DotNetWrapperForPython.htm



From carroll at tjc.com  Thu Apr  3 03:33:39 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Thu, 03 Apr 2003 00:33:39 -0800
Subject: nntplib: how to access NNTP response codes on exceptions?
Message-ID: <cirn8vsv4bhtp82n1tfil3e2kkounpgvav@4ax.com>

The documentation for nntplib lists a number of exceptions, usually
raised when the NNTP server responds with an error message of some
kind.  When this happens, the NNTP server supplies a numeric response
code indicating the nature of the error.

For instance, if you issue a HEAD 123 for a nonexistent article, the
NNTP server will respond:

   423 No Such Article In Group

where "423" is the response code that means "no such article."

Using nntplib, if I use

 (resp, article_number, msg_id, headers) = s.head('123')

I get the exception nntplib.NNTPTemporaryError as documented in
http://www.python.org/doc/current/lib/module-nntplib.html .  But
NNTPTemporaryError is raised for any response code in the range
400-499.  Is there a way of determining the particular response code?
Ordinarily, it would be the first three characters in the first
element ("resp", in my example) returned from the nntplib method
(whether head, stat, etc.).  But because of the exception, the
assignment to the tuple never occurs, so I can't check it.

I want to distinguish between 423 errors, which are expected and which
I can ignore, and anything else that comes up.



From frithiof.jensen at remove_this.jensen.tdcadsl.dk  Sun Apr 13 08:32:54 2003
From: frithiof.jensen at remove_this.jensen.tdcadsl.dk (Frithiof Andreas Jensen)
Date: Sun, 13 Apr 2003 14:32:54 +0200
Subject: Sv: Python 2.2.2 & Tix on win32 - What must I do?
References: <b7666e$oto$1@newstree.wise.edt.ericsson.se> <mailman.1050061697.14672.python-list@python.org>
Message-ID: <3e995853$0$24701$edfadb0f@dread14.news.tele.dk>

Raseliarison nirinA <nirina at mail.blueline.mg> skrev i en
nyhedsmeddelelse:mailman.1050061697.14672.python-list at python.org...

> You must have Tix widgets installed.
> http://tix.sourceforge.net/
> and maybe also the Tix library

Easier said than done ;-)

The Tix home page makes some vague statements on installing Tixapps and
Idestudio for binary installs yet the home-pages for said applications refer
back to tix.sourceforge.net ...

Perhaps I shall wait until Tix is properly integrated into Python; I do not
wish do damage my development environment.





From ramen at lackingtalent.com  Tue Apr 15 21:16:50 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Wed, 16 Apr 2003 01:16:50 -0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <slrnb9p4oo.vg8.ramen@lackingtalent.com> <mailman.1050453385.17091.python-list@python.org>
Message-ID: <slrnb9pbol.vg8.ramen@lackingtalent.com>

In article <mailman.1050453385.17091.python-list at python.org>,
Jp Calderone wrote:
>   I think you meant "partial lexical scoping" ;)  Dynamic scoping is what
> Python used to do.  Since it isn't quite lexical scoping, it's best to just
> stick with the term "nested scopes".

Ahh... this sounds familiar. Thanks, and sorry for the confusion. =)

Now, just to be thorough, the reason Python doesn't support "full lexical
scoping" is...? I'm guessing that it doesn't apply to classes (thus still
necessitating the use of self.*). Anything else?

Thanks,
Dave



From ckea25d02 at sneakemail.com  Tue Apr  8 21:27:41 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 09 Apr 2003 13:27:41 +1200
Subject: [Slightly OT]: More on ints and floats
In-Reply-To: <e0bf36b5.0304072228.4b47a93a@posting.google.com>
References: <fjss6b.1021.ln@boundary.tundraware.com> <e0bf36b5.0304072228.4b47a93a@posting.google.com>
Message-ID: <b6vsc9$9k1kk$1@ID-169208.news.dfncis.de>

L.C. wrote:
> I would ask a different question. What exactly would be the BENEFIT 
> of doing everything in FP ?

Simpler hardware and instruction set?

On the Burroughs 6000 series, integers were just denormalised
floats, and there was only one set of ADD, SUB etc. instructions
which would take any combination of ints, floats and doubles
and Do The Right Thing. (There were tag bits which distinguished
floats from doubles, but not ints from floats -- they were
both just "numbers".)

Not sure whether it actually made the hardware any simpler
(Burroughs hardware tended, in fact, to be insanely complicated)
but Burroughs obviously thought it was a good idea.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From nessus at mit.edu  Fri Apr 25 07:46:32 2003
From: nessus at mit.edu (Douglas Alan)
Date: Fri, 25 Apr 2003 07:46:32 -0400
Subject: A new and very robust method for doing file locking over NFS?
References: <lcu1cw525a.fsf@gaffa.mit.edu> <slrnbafu8h.ek.do-not-spam-ben.hutchings@tin.bwsint.com>
Message-ID: <lck7dix0iv.fsf@gaffa.mit.edu>

Ben Hutchings <do-not-spam-ben.hutchings at businesswebsoftware.com> writes:

> The usual solution for creating a lock file in an NFS-safe way is:

Yes, this is the method described in the Red Hat man page for the
open() system call, which I mentioned, but it doesn't solve the
unlocking problem, which remains, I believe, as I described.

|>oug



From martin at v.loewis.de  Wed Apr 30 15:28:27 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 30 Apr 2003 21:28:27 +0200
Subject: Python 2.3b1 and _tkinter.Tcl_Obj
References: <b8p30s$be4$1$8302bc10@news.demon.co.uk>
Message-ID: <m3wuhbg4ys.fsf@mira.informatik.hu-berlin.de>

"Stephen D Evans" <sde at recombinant.demon.co.uk> writes:

> Is this Python or Tk/Tcl? Is it a bug or a feature?
> 
> There are some inconsistencies in the types returned by some Tkinter widgets
> when using cget() or configure() e.g.
>     widget.cget('background')
> Menu widgets return a Tcl_Obj (typename='border'), Buttons return a string.

The inconsistency is in Tk. Tk widgets can chose whether they use the
Tcl object API or the Tcl string API; apparently, some widgets chose
one API, and other widgets the other.

If this is a problem for you, you have the following options:

- set Tkinter.wantobjects to False before creating the first
  tkapp object. This will give you the Python 2.2 behaviour.
- convert Tcl_Obj objects to strings using str()
- convert Tcl_Obj objects to Unicode using unicode()
- convert Tcl_Obj objects using 
  ASCII-if-possible-otherwise-Unicode-if-possible-otherwise-bytes
  algorithm of Python 2.2 using Tcl_Obj.string

If you feel you need to take one of these approaches, please let us
know, so we can see whether backwards compatibility needs to be
improved.

Regards,
Martin



From mertz at gnosis.cx  Tue Apr 15 03:37:31 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Tue, 15 Apr 2003 03:37:31 -0400
Subject: Speeding up: s += "string"
Message-ID: <7a7m+kKkX0KR092yn@gnosis.cx>         

The topic of the O() problems of the "s += 'string'" operation has been
discussed in several recent threads.  The aspect that seems not to have
really come up is "why not FIX it?"

This particular slowness is not merely academic.  It's bitten me.  In
one of his posts, I think the Martellibot even said it bit him (in the
mists of past time).  Probably most Python programmers fell into this
trap at least once.  I got past it, and now usually use the
"lst.append(string); "".join(lst)" solution.  But the idea of making a
string longer by... well, adding more to it... seems really obvious.
Almost the "one obvious way to do it."

It seems to me that the behavior CAN be improved.  If strings were to
use a preallocation strategy--similar to what lists do--MANY string
appends could avoid a memory copy on the original string.  The intial
value could stay where it was, and the extra characters could live in
the already-available contiguous space.  Obviously, the utilized length
would need to be stored somewhere.  Once in a while, of course, the
preallocation would fill up; in that case you'd need to do a costly
strcopy()... but this could be avoided MOST of the time, ISTM.

I quite confess to not *really* understanding optimization issues.
Doing the preallocation right could be complicated.  It's not like you
want to preallocate a megabyte for every one-char string.  I suppose
you'd want to determine how often a particular string already underwent
appends, and what size strings were typically appended.  And since
strings are immutable, it's not the self-same *objects* that underwent
the change, but the underlying contiguous memory areas.  Still... it
doesn't seem undoable.

Yours, Lulu...

--
---[ to our friends at TLAs (spread the word) ]--------------------------
Echelon North Korea Nazi cracking spy smuggle Columbia fissionable Stego
White Water strategic Clinton Delta Force militia TEMPEST Libya Mossad
---[ Postmodern Enterprises <mertz at gnosis.cx> ]--------------------------





From mwm at mired.org  Tue Apr  1 12:31:45 2003
From: mwm at mired.org (Mike Meyer)
Date: Tue, 01 Apr 2003 17:31:45 GMT
Subject: PyQt QTableItem inheritance...
Message-ID: <x77kaejf9b.fsf@guru.mired.org>

I've already tried posting this to the pyKDE mail list, and gotten no
response.

I'm trying to create an item in a QTable that uses the
QDoubleValidator on the input edit field. The only way I can see to do
this is subclass QTableItem, and override createEditor with a method
that sets the validator, like so:

class Float_Field(QTableItem):

    def __init__(self, table, data):
        print "Creating float field."
        QTableItem.__init__(self, table, QTableItem.OnTyping, str(data))

    def createEditor(self):
        print "createEditor"
        editor = QLineEdit(self.text(), self.table().viewport())
        self.connect(editor, SIGNAL("textChanged(QString &)"),
                     self.table(), SLOT("doValueChanged"))
        editor.setValidator(QDoubleValidator(self))
        return editor



While I get the output from __init__, I never see the output from
createEditor. Any help would be appreciated.

	Thanks,
	<mike

-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From david at nospam-freenet.org.nz  Wed Apr 23 02:50:59 2003
From: david at nospam-freenet.org.nz (David McNab)
Date: Wed, 23 Apr 2003 18:50:59 +1200
Subject: Help - building Linux standalones
Message-ID: <pan.2003.04.23.06.50.59.658420@nospam-freenet.org.nz>

Hi,

I've been battling to build an app as a linux standalone.

Using McMillan Installer, I've successfully created a standalone for
Windows which works to expectations.

But in Linux-land, we get into dramas with system library versions.
(I'm on debian, using glibc v2.3. Not all distros use this version, such
as mandrake, which uses 2.2).

For one, McMillan Installer doesn't include system libraries in the
distribution directory. The 'standalones' which it generates require the
same version of system libs to be present on the target system - not an
acceptable outcome.

Switching to the 'freeze' utility, we get a bunch of sources that
compile/link/run ok, as long as they're linked against shared system libs,
and the same version of these libs is present on the target. Same problem.
But when linked against static libc.a, m.a, pthread.a etc, the resulting
program just segfaults.

Also, I've tried to build Python 2.2.2 to itself link the system libs
statically, but I've walked into a minefield here - the generated python
interpreter crashes. 

If anyone has managed to build python progs into Linux standalones that
don't require specific sys lib versions, can you please tell me what
you've done and how you've done it?

Thanks most kindly in advance for any help
Regards
David




From theller at python.net  Thu Apr  3 01:43:22 2003
From: theller at python.net (Thomas Heller)
Date: 03 Apr 2003 08:43:22 +0200
Subject: Python commands
References: <3e8bb9c5$1_1@news.iprimus.com.au>
Message-ID: <7kacm67p.fsf@python.net>

"Paul Dexter" <dextech at terrigal.net.au> writes:

>     Hi all,
>                 I have just discovered Python and I was wondering if anyone
> could help me find an equivalent command in Python for the "OUT" and "INP"
> commands in Qbasic. I wish to convert a Qbasic program to Python.
>     The following snippet is an example of the task I wish to perform;
> 
> base0 = &H378:
> 
> interface:
> OUT base0, &HA0
> OUT base0, &H80
> OUT base0 + 2, &H8
> 
> V = (INP(base0 + 1) AND &H10) * &H8
> 
>     The base address is 378 (lpt) and the "interface" subroutine outputs
> data to the port (378) and "V" is the input data statement I have used. I
> also use the "GETKEY", "LOCATE" and "DRAW" commands in Qbasic
> 
>     Does anyone know how to perform the same tasks/commands in Python or are
> they even available. I have come across other Basic computer language
> variations that do not support the commands I require.
> 
>     Thankyou for any help,
>                                         Paul

Usually port I/O instructions cannot be used from user mode programs,
even on Windows (NT, 2000, XP). This may be different on win95 or win98.


However, the ctypes module allows calling functions in dlls, and this
is what I get on WindowsXP (windll.msvcrt loads the MS C runtime
library, and _outp is the function):

>>> from ctypes import windll
>>> windll.msvcrt._outp(0x378, 0xA0)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
WindowsError: exception: priviledged instruction
>>>

http://starship.python.net/crew/theller/ctypes.html

Thomas



From altis at semi-retired.com  Thu Apr 10 16:20:10 2003
From: altis at semi-retired.com (Kevin Altis)
Date: Thu, 10 Apr 2003 13:20:10 -0700
Subject: Writing a TCP/IP chat server
References: <Pine.LNX.4.44.0304101906590.8243-100000@servalan.ecs.soton.ac.uk>
Message-ID: <9jkla.522$Aa.86418@news.uswest.net>

A simple chat implementation with a GUI. This client is also a server and
can dispatch messages to multiple recipients by simply providing their IP
addresses by separating each address with a comma in the provided text
field.

  http://pythoncard.sourceforge.net/samples/chat.html

You may also want to look at Twisted.

ka

"Ian Partridge" <iop199 at ecs.soton.ac.uk> wrote in message
news:Pine.LNX.4.44.0304101906590.8243-100000 at servalan.ecs.soton.ac.uk...
> Hi,
>
> I am trying to write a chat server, which users will telnet to in order to
> interact with each other.
>
> What is the best way to do this?
>
> I have searched the web and found three different approaches:
>
> 1. Use the socket libraries and select to multiplex between the
> connections.
>
> 2. Use the SocketServer module (although I can't find a decent tutorial on
> how to use this module)
>
> 3. Use the asyncore module (I only found one example doing this, and I
> couldn't understand it).
>
> What's the the simplest way to approach the problem?
>
> Any tips gratefully received.
>
> TIA,
> Ian
>





From juneaftn at REMOVETHIShanmail.net  Thu Apr 24 03:45:38 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Thu, 24 Apr 2003 16:45:38 +0900
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <X8Mpa.30197$y3.2398527@news010.worldonline.dk>
Message-ID: <b884m3$182$1@news.hananet.net>

"Max M" <maxm at mxm.dk> wrote in message
news:X8Mpa.30197$y3.2398527 at news010.worldonline.dk...
> Peter Hansen wrote:
>
> > For throwaways, I don't write quite so many tests, and in many cases if
> > it's only a "small" throwaway script (e.g. one or two subroutines only,
> > plus a little mainline) I will write a test or two for the subroutines
> > (because I'm doing TDD, not because I want to have unit tests to make
> > sure my script works properly... if you see the distinction) but never
> > for the mainline.
>
>
> While I do understand the principles of tdd, testing first and all that,
> I have a hard seing why it should be an advantage to do it in the early
> stages of a project.
>
> Normaly if a am entering a new domain, where I don't have a clear
> picture of the requirements, I try out different approaches. Each
> approach using a different combination of ideology, objects and
functions.
>
> This is all to get a feel of the best approach in this particular
> domain. Some times I try out a *lot* of approaches before I end up with
> a particular elegant design.
>
> At which point I am ready to "solve the problem". It is at this point I
> can see the point of unit testing.
>
> In tdd do you really use testing for the first "creative" fase also?
>

In XP, they call that kind of exploration SpikeSolution. You can do SS with
or without TDD.

When I do TDD, I end up with a good design which I wouldn't have discovered
without TDD. Every time I do TDD I'm surprised at the final design. I
didn't expect the design. Moreover, it comes out with a different design
when TDD the same program slightly differently again. Often, it's more
creative than I am.

That said, I still do SS without TDD sometimes, but I love the explorative,
generative, and discovering power of TDD.




From ianb at colorstudy.com  Tue Apr 29 23:31:32 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 29 Apr 2003 22:31:32 -0500
Subject: Web programming and a different "type" problem
In-Reply-To: <m365owlu6v.fsf@ferret.object-craft.com.au>
References: <mailman.1051532982.25352.python-list@python.org>
	 <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com>
	 <slrnbarhfb.67t.jon+usenet@snowy.squish.net>
	 <b8lj2q$l14$1@newshost.accu.uu.nl>
	 <m3fzo1v561.fsf@ferret.object-craft.com.au>
	 <mailman.1051654786.4587.python-list@python.org>
	 <m365owlu6v.fsf@ferret.object-craft.com.au>
Message-ID: <1051673492.698.214.camel@lothlorien>

On Tue, 2003-04-29 at 19:12, Dave Cole wrote:
> So what Albatross does is allow (require) you to specify when a field
> should return a list.  If you build a form which contains multiple
> inputs with the same name ("radio" is an obvious exception) then it
> will raise an exception if you do not use the "list" attribute on all
> of the instances of the field in the form.

How do you specify that you expect a list?  Does it happen at runtime,
when you fetch the value from the request with an I'm-expecting-a-list
method?  Or is there some more declarative schema involved?

> In our experience you can get subtle and annoying bugs in your
> application as a result of incorrect handling of fields which
> sometimes return multiple values.

Indeed, but I've usually found that the bugs were from form generation,
not from form processing.  It might be better if the error is signaled
earlier and in a more pleasant way, but IMHO throwing away duplicate
values when they aren't expected just covers up bugs.  But I don't think
that's what you're doing in this case (though it sounds like jonpy does
this, for instance)...?

> When the </al-form> tag is processed, the record of fields is encoded
> into a hidden field called __albform__.  The field value is MD5 signed
> by combining the encoded value with a server side secret and the sign
> is combined with the __albform__ value.

Again out of curiosity -- what exactly goes into the __albform__ value? 
I.e., what information are you saving there?

  Ian






From skywalkerpackage at hotmail.com  Tue Apr 15 06:34:03 2003
From: skywalkerpackage at hotmail.com (Johannes Eble)
Date: Tue, 15 Apr 2003 10:34:03 GMT
Subject: Visio like library for Python?
References: <3e93e50f.676072541@news.online.de> <r7s79vghrcufpsostmv8avr7ncu2985ja7@4ax.com>
Message-ID: <3e9bded1.100826623@news.online.de>

Hi John,

thanks very much for your answer. I have had a look on Open Office. It
is very interesting, but also still very complex to use Open Office
modules and the UNO architecture.

After thinking about the pro and cons, I think that it is easier for
me to use WxPython, Object Graphics Library, and perhaps PyUt as other
posters have suggested (see my other answers).


Regards


Johannes






From claird at lairds.com  Mon Apr 21 21:03:07 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 22 Apr 2003 01:03:07 -0000
Subject: *Newbie Question* --> Is it viable to distribute Python apps
 over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <3ea434fd$0$180$a1866201@newsreader.visi.com> <25a672ee.0304211504.4f3cd067@posting.google.com> <mailman.1050968618.16005.python-list@python.org>
Message-ID: <va956b925cj1df@corp.supernews.com>

In article <mailman.1050968618.16005.python-list at python.org>,
Gerhard H?ring  <gh at ghaering.de> wrote:
>Denis Agar wrote:
>> I guess I understand the situation then. Is there any way to use a GUI
>> that most people already have such as the ones employed by Visual
>> Basic or Visual C++?
>
>VB's GUI code is likely in the VB runtime. With "VC GUI" you likely mean 
>the MFC classes, which are usually in the MFC42.DLL. It can be used from 
>Python via the win32 extensions, which are a relatively slim download.
>
>But looking at the included examples, I wouldn't want to develop a GUI 
>using the win32 extensions, as it's basically hardcore win32 API 
>programming.
			.
			.
			.
Let's note, though, that there are several contributors
to c.l.p who speak with comfort about having delivered
applications to customers coded in Python against MFC.
Gerhard and I haven't, but apparently it's not infeasible.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From a.schmolck at gmx.net  Sat Apr 19 19:44:22 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 20 Apr 2003 00:44:22 +0100
Subject: When not to use an RE -- an example
References: <c76ff6fc.0304191459.2adf73a@posting.google.com>
Message-ID: <yfs7k9qjbmx.fsf@black132.ex.ac.uk>

sjmachin at lexicon.net (John Machin) writes:

> I needed a check for strings consisting of repeated characters -- like
> when users type "ZZZZZZZ" instead of "UNKNOWN" into a database field.
> After implementing the obvious overlapping-substring comparison, I got
> to thinking how this could be done with REs. The following resulted:
> 
> import re
> repeats1 = re.compile(r"^(?:(.)(?=\1))+\1\Z", re.DOTALL).match
> def repeats2(s):
>    return len(s) > 1 and s[1:] == s[:-1]
> for testvalue, expected in zip(
>    ['','x','xx','xxx','xxxxxx','xy','xxy','xyy','\n\n\n','aaa\n'],
>    [0,  0,  1,   1,    1,       0,   0,    0,    1,       0     ]):
>    print repr(testvalue), not not repeats1(testvalue),
> repeats2(testvalue), expected
> 
> Pergly/phugly, eh? Note the effort required to ensure the newline
> cases worked.


What's wrong with:

>>> matchRepetition = re.compile(r'(.)\1+\Z', re.DOTALL).match
>>> map(bool, map(matchRepetition,
...  ['','x','xx','xxx','xxxxxx','xy','xxy','xyy','\n\n\n','aaa\n']))
[0, 0, 1, 1, 1, 0, 0, 0, 1, 0]

?

At least it would seem to pass your testcases.

'as



From andrew.markebo at telia.com  Thu Apr  3 05:31:07 2003
From: andrew.markebo at telia.com (Andrew Markebo)
Date: Thu, 03 Apr 2003 10:31:07 GMT
Subject: FTP : get  dir name with space caracters embedded  + is ther
 something quicker than LIST
References: <b1ab591.0304030143.52422a3e@posting.google.com>
Message-ID: <m37kab27pw.fsf@telia.com>

Regarding question 1, speed.. are you sure it is python and not the
ftp-server that takes time to send you the filenames?

What speed do you get if you use the nlst command instead (can be
found on the ftp object)

| 2) DIRECTORY NAME WITH SPACES  
|
| LIST is slow but it works. One other problem is the directory name in
| which there is spaces : like
| "data file today".
|  I can do a :
| elements= line.split(" "), 
|
| but the data before my directory name has not the same format : for
| example :
|
| .....
| may 22   2001     name of directory
| mar 15    02:00   other directory name
| .....

Implementation error ;-) Maybe should try line.split(" ", 3), telling
it only to split the 3 first spaces, the rest should probably stick
together.. Or do a regexp that can handle the driectory output.. 

           /Andy

-- 
 The eye of the beholder rests on the beauty!



From rmunn at pobox.com  Tue Apr  1 10:06:38 2003
From: rmunn at pobox.com (Robin Munn)
Date: Tue, 01 Apr 2003 15:06:38 GMT
Subject: Breaking out of a while loop with a key press?
References: <3e885ad5@mail.hmgcc.gov.uk> <slrnb8gc0t.783.rmunn@localhost.localdomain> <wM0ia.14517$_14.288@nwrdny02.gnilink.net>
Message-ID: <slrnb8ilo9.9hc.rmunn@localhost.localdomain>

Carl Banks <imbosol-1049135446 at aerojockey.com> wrote:
> Robin Munn wrote:
>> Richard <richardd at hmgcc.gov.uk> wrote:
>>> Hi,
>>> 
>>> Can anyone suggest a simple way of detecting a key press and existing the
>>> program? I have a program which at the moment uses a while True: statement.
>>> However this is not ideal as I have to close down the console window to
>>> exist the program. Can anyone tell me the proper way of doing this? So that
>>> a CTRL-C, for example, can quit the program correctly?
>> 
>> You could use a "try: ... except KeyboardInterrupt: ..." block to detect
>> Ctrl-C. Although KeyboardInterrupt *should* by default not be caught, so
>> a Ctrl-C *should* be able to shut down your program.
> 
> In this case, it shouldn't.  He's running his program in a console
> window created within Python.  The idea of a console is
> (theoretically) to give a program complete control of this input;
> therefore, console windows often disable KeyboardInterrupt and treat
> Control-C like any regular keystroke.

Ah. I was assuming something like the Linux console, or the Windows
"console" that's almost-but-not-entirely like a Unix console in
behavior...

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From staschuk at telusplanet.net  Fri Apr 18 12:44:38 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 18 Apr 2003 10:44:38 -0600
Subject: Teaching : Python, Scheme, Java...
In-Reply-To: <roy-08BA75.10145718042003@malibu.unice.fr>; from roy@unice.fr on Fri, Apr 18, 2003 at 10:14:57AM +0200
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr> <mailman.1050609281.18019.python-list@python.org> <roy-08BA75.10145718042003@malibu.unice.fr>
Message-ID: <20030418104438.A709@tibia.amotlpaa.bogus>

Quoth Jean-Paul Roy:
  [...]
> Let me cite the Normalization Report on Scheme:
> 
> "Programming languages should be designed not by piling feature on top 
> of feature, but by removing the weaknesses and restrictions that make 
> additional features appear necessary."
> 
> And so the wile or for... loops could be defined in Python as in Scheme 
> with the tail recursion optimzation.

Let me paraphrase another source:  "A programming language should
be as simple as possible, but no simpler."  :)

  [...]
> My teaching experience is that students knowing loops through the while 
> construct actually *understand* the concept of iteration when they see 
> that it is modelized by a tail recursive function which transforms a 
> "vector of loop variables" (a state vector). Then, it is only a GOTO 
> which passes arguments. I'm really happy to teach that way.

A nice point.

I wrote earlier that tail recursion is an obfuscated way of
writing iteration; I see now that I misunderstood what you're
teaching.

For teaching simply how to program, explicit loop syntax is better
than tail recursion imho, since it makes the programmer's intent
more explicit.  For teaching theory, tail recursion is better,
since it makes the state vector more explicit (as an argument list
to the tail-recursive function, rather than implicit in the
environment).

-- 
Steven Taschuk                            staschuk at telusplanet.net
Every public frenzy produces legislation purporting to address it.
                                                  (Kinsley's Law)




From dave at boost-consulting.com  Thu Apr 24 22:50:43 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Thu, 24 Apr 2003 22:50:43 -0400
Subject: boost.python - Virtual Functions with Default Implementations
References: <3c539fdc.0304241407.54dac3dc@posting.google.com>
Message-ID: <ubryvguik.fsf@boost-consulting.com>

sam at insidegabriel.com (Sam) writes:

> I'm trying to get virtual functions with default implementations to be
> exposed to Python.  When I follow the examples given on the boost.org
> site...
> http://www.boost.org/libs/python/doc/tutorial/doc/virtual_functions_with_default_implementations.html
>
> ...I get the following error:
> c:\projects\publiclibs\boost\boost_1_29_0\boost\python\detail\def_helper.hpp(131)
> : error C2440: 'return' : cannot convert from 'int (__thiscall
> giPython::BaseWrap::*const )(void)' to 'const char *'
>         There is no context in which this conversion is possible
>         c:\projects\publiclibs\boost\boost_1_29_0\boost\python\detail\def_helper.hpp(130)
> : while compiling class-template member function 'const char
> *__thiscall boost::python::detail::def_helper<int (__thiscall
> giPython::BaseWrap::*)(void),struct
> boost::python::detail::not_specified,struct
> boost::python::detail::not_specified>::doc(void) const'
>
> Anyone have any ideas?  Here is my code:

Yeah, I'm pretty sure the tutorial page you're looking at is only
valid for Boost 1.30.0, and you're using 1.29.0.  Check out the
tutorial pages in your Boost installation
($BOOST_ROOT/libs/python/doc) and see if they give the same
instructions...

BTW, I suggest you post Boost.Python questions to the C++-sig:
http://www.python.org/sigs/c++-sig/

HTH,
-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From alloydflanagan at attbi.com  Tue Apr 15 09:27:16 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 15 Apr 2003 06:27:16 -0700
Subject: Using property in classic class may not work
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it>
Message-ID: <a8b7f07a.0304150527.5b0b1187@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<ifPma.15344$LB6.409776 at news1.tin.it>...
> Neil Hodgson wrote:
> 
> >     The class in question is derived from a wxWindows class, so I can't
> > make it derive just from object. The "What's New in Python 2.2" document
> > only mentions property wrt new-style clases. Can I depend on property
> > working for classic classes? Is there a workaround such as deriving from
> > object as well?
> 
> No, and yes, respectively -- deriving from object means your class
> isn't classic any more, so properties do work as intended there.  Vide:
> 
... (big chunck cut) ...
> as you see, this does work as intended.
> 
> 
> Alex

However, the last time I tried to subclass from both wxFrame and
object, wxPython freaked out (that's a technical term for "I don't
remember what happened, but boy, it sure didn't work").
Not sure about the latest version of wxPython, though -- there's
binaries for python 2.3 out by now.  I'll have to try it again..



From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr 24 12:56:53 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 24 Apr 2003 16:56:53 GMT
Subject: Help - building Linux standalones
References: <pan.2003.04.23.06.50.59.658420@nospam-freenet.org.nz>
Message-ID: <slrnbag5qk.ek.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <pan.2003.04.23.06.50.59.658420 at nospam-freenet.org.nz>,
David McNab wrote:
> Hi,
> 
> I've been battling to build an app as a linux standalone.
> 
> Using McMillan Installer, I've successfully created a standalone for
> Windows which works to expectations.
> 
> But in Linux-land, we get into dramas with system library versions.
> (I'm on debian, using glibc v2.3. Not all distros use this version, such
> as mandrake, which uses 2.2).
<snip>

glibc maintains very good backward binary compatibility, so there should
be no need to redistribute it.  Doing so will not only bloat your program
but will also obligate you to distribute glibc source if you distribute
the program commercially.  You *should* be able to install an older
version of the glibc development files and build with them.  Then your
program would work with that or any newer version.  Perhaps you can
install libc6-dev from potato - if that's still available - or woody.
(I wouldn't advise doing development using anything but stable.)



From mustafa.celik at sympatico.ca  Fri Apr 11 03:41:23 2003
From: mustafa.celik at sympatico.ca (Mustafa Celik)
Date: Fri, 11 Apr 2003 03:41:23 -0400
Subject: collecting data from file
Message-ID: <3E9671A3.2050004@sympatico.ca>

Hi,
I want to scroll thru a file;
   * find lines that match a string (e.g. HELLO) on the 2nd column
   * add up the 4th column (an integer) on each matching line, say the 
variable is TOTAL
   * subtract the 4th column from TOTAL if another is string (e.g. BYE) 
is hit on 2nd column of a line

Any tips?





From eppstein at ics.uci.edu  Sun Apr 20 20:40:11 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Sun, 20 Apr 2003 17:40:11 -0700
Subject: Good Python generators example?
References: <paDoa.1611$hT2.1228371@news2.news.adelphia.net> <yu99y9243nn1.fsf@europa.research.att.com>
Message-ID: <eppstein-01B6A0.17401120042003@news.service.uci.edu>

Robert> Can someone point me to a nice concise Python 2.2 compatible
Robert> example of the use of generators?

There are a bunch of them at
http://aspn.activestate.com/ASPN/Cookbook/Python?kwd=Algorithms

A couple of mine:

Sieve of Eratosthenes prime number generator
http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/117119

Use of generators in the middle of an algorithm for finding farthest 
pairs of points
http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/117225

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From andy at reportlab.com  Sat Apr 19 17:36:14 2003
From: andy at reportlab.com (Andy Robinson)
Date: Sat, 19 Apr 2003 22:36:14 +0100
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com> <a3b5b608.0304080302.c26ad68@posting.google.com> <af6b5bdd.0304181001.558da6f@posting.google.com>
Message-ID: <k8g3av4mkn20e3q59fa7ngmfddg5mavs7b@4ax.com>

On 18 Apr 2003 11:01:48 -0700, bitbucket at safe-mail.net (poiboy) wrote:

>
>Neglecting the inherent joy of an education, I believe the cost of using XSLT
>exceeds what benefits it may have (portability, and..) and is likely more
>expensive than successive rewrites with other template systems.
> 

Many have said that to me in private or on the phone.  Few have
been brave enough to say it on Usenet. Bravo :-)

- Andy Robinson



From avner at skilldesign.com  Thu Apr 24 05:40:56 2003
From: avner at skilldesign.com (Avner Ben)
Date: Thu, 24 Apr 2003 11:40:56 +0200
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com>
Message-ID: <3ea7b12a@news.012.net.il>

"Dave Benjamin" <ramen at lackingtalent.com> wrote in message
news:slrnbaf7bi.gb9.ramen at lackingtalent.com...
> In article <3E9FFCB5.79AEA22E at engcorp.com>, Peter Hansen wrote:
> > Aahz wrote:
> >>
> >> In article <mailman.1050587068.2733.python-list at python.org>,
> >> Jp Calderone  <exarkun at intarweb.us> wrote:
> >> >
> >> >--
> >> >It is practically impossible to teach good programming style to
> >> >students that have had prior exposure to BASIC: as potential
> >> >programmers they are mentally mutilated beyond hope of
> >> >regeneration.        -- Dijkstra
> >>

The first language I learnt was COBOL. That took even longer to overcome!

    Avner.





From no.spam at address.com  Tue Apr  8 02:42:29 2003
From: no.spam at address.com (Edmund)
Date: Tue, 08 Apr 2003 06:42:29 GMT
Subject: Tools for refactoring code
Message-ID: <4gr49v0pepuj0p4vj4t31s3mmgbo017tej@4ax.com>

I have an application that needs to be refactored, and it would be
useful to know where and how often functions and objects/methods are
being used. Is there some kind of Python source code cross-referencer
out there? I looked, but can't find one.

In general, are there any tools to help the refactoring process?

...Edmund.



From paul at boddie.net  Tue Apr 22 06:36:25 2003
From: paul at boddie.net (Paul Boddie)
Date: 22 Apr 2003 03:36:25 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com> <6ee58e07.0304061758.35070e7c@posting.google.com> <a3b5b608.0304080302.c26ad68@posting.google.com> <af6b5bdd.0304181001.558da6f@posting.google.com> <k8g3av4mkn20e3q59fa7ngmfddg5mavs7b@4ax.com>
Message-ID: <23891c90.0304220236.4d0196fb@posting.google.com>

Andy Robinson <andy at reportlab.com> wrote in message news:<k8g3av4mkn20e3q59fa7ngmfddg5mavs7b at 4ax.com>...
> On 18 Apr 2003 11:01:48 -0700, bitbucket at safe-mail.net (poiboy) wrote:
> 
> >Neglecting the inherent joy of an education, I believe the cost of using XSLT
> >exceeds what benefits it may have (portability, and..) and is likely more
> >expensive than successive rewrites with other template systems.

But XSL is primarily a transformation language and not a templating
system, especially since templating systems often have domain-specific
features that reduce the verbosity. Moreover, it makes most sense to
use XSL(T) when you're working in an environment with the necessary
XML and XSL tools already in place, and where it isn't necessarily
going to be easier to write program code to do the transformations.
For example, I've used XSLT to process XML documents that have been
generated from various Java libraries, and that processing would have
been tedious to do by either modifying those Java libraries, or by
writing some Java classes and generating the output from the code.

> Many have said that to me in private or on the phone.  Few have
> been brave enough to say it on Usenet. Bravo :-)

I'm not sure. Search for "XSLT overkill group:comp.text.xml" on
Google, for example. :-)

Paul



From aleax at aleax.it  Tue Apr 22 08:10:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 22 Apr 2003 12:10:39 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net> <nCUoa.41676$LB6.941312@news1.tin.it> <pan.2003.04.22.03.08.58.526973@mn.rr.com>
Message-ID: <3japa.44301$LB6.997057@news1.tin.it>

Anna wrote:

> On Mon, 21 Apr 2003 16:02:27 +0000, Alex Martelli wrote:
> 
> 
>> As a bonus -- sum([]) now returns 0, while in my original design it
>> raised
>> a ValueError, just like e.g. max([]) does.  I thought raising an error
>> was best in the name of "errors shouldn't pass silently" and "in the face
>> of
>> ambiguity, refuse the temptation to guess".  But -- once we decree that
>> "sum is about numbers", there's no real ambiguity, so we can have the
>> little extra handiness of having sum([]) not be an error;-).
> 
> Hooray!
> 
> Sum() is a goodness! And sum([]) returns 0 is a great goodness.
> Thanks Alex.

You're welcome, but for "sum([]) returns 0" you should thank Guido
and Tim (it would raise a ValueError if I had implemented it my way;-).

[Which may help explain why I use and love a language architected by
Guido (with lots of input from Tim & many others, of course), rather
than one architected by my own self...;-)]

To be properly thankful that we're not using an Alex-designed
language, we might want to look at a simple example from the
domain-specific language BBL (Bridge-Bidding Language) which I
designed and implemented in the '80s [well before I met Python --
but I was quite familiar with Icon, REXX and other highly
readable very-high-level languages, so that's no excuse;-)].

Here's the BBL expression to determine whether the current hand
is suitable for a weak-notrump opener...:

& (                      # conditions for a weak-notrump opener
    s ( 1 g 4  3  3  3   # shape 4333 (any), or...
        2 g 4  4  3  2   #       4432 (any), or...
        3   3- 3- 3- 5   #       5332 w 5 clubs, or...
        4   3- 3- 5  3-  #       5332 w 5 diamonds;
      )
    < ( 11                    # moreover, a 12-14 range for
        \+ SHDC c( 4 3 2 1 0) # normal Milton-Work point-count
        15
      )
  )

There -- aren't you glad I don't do language design any more?-)


Alex




From peter at engcorp.com  Mon Apr 28 08:13:19 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 28 Apr 2003 08:13:19 -0400
Subject: Python-2.3b1.exe install error on W98se
References: <b8itdt$53h$1@news.hccnet.nl>
Message-ID: <3EAD1ADF.27E9E4A5@engcorp.com>

Anton Vredegoor wrote:
> 
> A message box popups almost immediately:
> 
> <message box>
> Install
> 
> Could not initialize installation. File size expected=1972043840, size
> returned=9096924. The parameter is incorrect.
> </message box>

Hmm.... clearly the download was unsuccessful.  In fact, Python 2.3b1
*has* grown to over 1.8G, and you didn't wait long enough to download
the whole thing.  (Are you on dialup or something?)  The reason for
the new, much larger installation package, in addition to all the new
features which were crammed in at the last minute, is that the PSU



From rogue9 at ntlworld.com  Fri Apr 25 03:28:00 2003
From: rogue9 at ntlworld.com (Rogue 9)
Date: Fri, 25 Apr 2003 08:28:00 +0100
Subject: Creating dynamically named lists/dictionaries
References: <ydFpa.15$IO5.10324@newsfep2-gui.server.ntli.net>
Message-ID: <_V5qa.37$rA3.25@newsfep3-gui.server.ntli.net>

On Thu, 24 Apr 2003 00:22:18 +0100, Rogue 9 wrote:

> Hi all,
snip snip
> 	I hope I have explained my 'problem' sufficiently for someone to help or
> to at least point me in the right direction to work towards a solution.
> 	Yours hopefully,
> 		Rogue9
	Thanks to all who answered - I've printed off the replies and worked
through them (didn't fully understand it but got the gist anyway).I've
managed to get the idea of what to do next and made progress in the
program in that I've got functions to display a user chosen individual
week and a function which will allow me to calculate the median for each
ball by position for all results or a user chosen range.
	Again please accept my thanks as you have removed a boulder in my road to
understanding.
	Cheers,
	Lol McBride



From fusion at thuule.pair.com  Sat Apr 12 14:41:12 2003
From: fusion at thuule.pair.com (Alan McIntyre)
Date: Sat, 12 Apr 2003 14:41:12 -0400
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
Message-ID: <3E985DC8.9060907@thuule.pair.com>

Just off the cuff here, it seems to me that the time required to do 
"string += string" is going to be proportional to the size of the 
original string whether you do it in Python or with the STL's 
std::string or with C strings (so long as you don't know beforehand how 
long the final string will be).

Now I'm inclined to do a comparison just for my own curiosity's 
sake...maybe I'll be back with graphs in just a bit. :)

Mads Orbesen Troest wrote:
> Hi;
> 
> Given that strings are immutable, is it exceedingly slow (and memory 
> spiking) to do a lot of "string += string" operations in one's code? IOW, 
> does each and every += evaluation lead to a new concatenated copy of the 
> previous string (now freed for garbage collection) and the new string, so 
> that the longer the string to which stuff is appended is, the longer times 
> each += operation takes?
> 
> TIA,
>    /\/\\ads Orbesen Troest






From BPettersen at NAREX.com  Wed Apr 23 15:32:38 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 23 Apr 2003 13:32:38 -0600
Subject: Exception.__init__ needed in derived class?
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DE0C7@admin56.narex.com>

> From: george young [mailto:gry at ll.mit.edu] 
> 
> [python-2.3a1]
> Does class Exception have a constructor __init__ that needs
> to be called?  I couldn't find this in the docs.  I see many
> uses of inheriting from Exception without invoking
> Exception.__init__(self), but nobody ever *said* that this 
> was safe to leave out...
[...]

I can't find it documented anywhere either... In any case, Exception
looks like this:

  class Exception:
     def __init__(self, args):
         self.args = args
     def __str__(self):
         if len(self.args) == 0: return ''
         if len(self.args) == 1: return str(self.args[0])
         return str(self.args)
     def __getitem__(self, k):
         return self.args[k]

(quickly translated, there might be others [I couldn't follow the trail
of populate_methods..] :-)  I don't think I've seen __getitem__
documented either...

-- bjorn




From bluetigger at web.de  Tue Apr 22 04:57:01 2003
From: bluetigger at web.de (Andreas KLostermann)
Date: Tue, 22 Apr 2003 10:57:01 +0200
Subject: Python 2.2.2 and 2.3a2 compile problems on Mingw32
Message-ID: <200304220857.h3M8v0q29980@mailgate5.cinetic.de>

I wanted to compile Python on Mingw32. That's a GNU C++ Compiler for Windows and it can configure...  The "make" process begins promissing, but fails at "posixmodule.c"...
This Problem is the same for 2.2.2 and 2.3a2.
Can someone help me?

Output:

$ make
gcc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -I. -I./Include -DHAVE_CONFIG_H                   -c ./Modules/posixmodule.c -o Modules/posixmodule.o
Modules/posixmodule.c:278: warning: function declaration isn't a prototype
Modules/posixmodule.c: In function `posix_ttyname':
Modules/posixmodule.c:777: warning: implicit declaration of function `ttyname'
Modules/posixmodule.c:777: warning: assignment makes pointer from integer withou                t a cast
Modules/posixmodule.c: In function `posix_mkdir':
Modules/posixmodule.c:1230: too many arguments to function `mkdir'
Modules/posixmodule.c: In function `posix_utime':
Modules/posixmodule.c:1429: warning: implicit declaration of function `utime'
Modules/posixmodule.c: In function `posix_fork':
Modules/posixmodule.c:1902: warning: implicit declaration of function `fork'
Modules/posixmodule.c: In function `posix_getegid':
Modules/posixmodule.c:1988: warning: implicit declaration of function `getegid'
Modules/posixmodule.c: In function `posix_geteuid':
Modules/posixmodule.c:2003: warning: implicit declaration of function `geteuid'
Modules/posixmodule.c: In function `posix_getgid':
Modules/posixmodule.c:2018: warning: implicit declaration of function `getgid'
Modules/posixmodule.c: In function `posix_getppid':
Modules/posixmodule.c:2131: warning: implicit declaration of function `getppid'
Modules/posixmodule.c: In function `posix_getuid':
Modules/posixmodule.c:2177: warning: implicit declaration of function `getuid'
Modules/posixmodule.c: In function `posix_kill':
Modules/posixmodule.c:2207: warning: implicit declaration of function `kill'
Modules/posixmodule.c: In function `posix_wait':
Modules/posixmodule.c:3364: warning: implicit declaration of function `wait'
Modules/posixmodule.c: In function `posix_pipe':
Modules/posixmodule.c:3867: warning: implicit declaration of function `pipe'
make: *** [Modules/posixmodule.o] Error 1

______________________________________________________________________________
UNICEF bittet um Spenden f?r die Kinder im Irak! Hier online an
UNICEF spenden: https://spenden.web.de/unicef/special/?mc=021101





From hammer at maxwell.com  Sun Apr 13 23:01:21 2003
From: hammer at maxwell.com (Maxwell Hammer)
Date: Sun, 13 Apr 2003 23:01:21 -0400
Subject: Question regarding indirect function calls
References: <pan.2003.04.13.20.13.21.504295@maxwell.com>
Message-ID: <pan.2003.04.14.03.01.20.731919@maxwell.com>

In reply to: Jack & Alex:

Thanks for your advice, I did get it working using "getattr" as you
suggested.

Great! now I have another questions, since I have your attention :-)

What is the most efficient way of accessing global variables across
modules?

I have main.py which parses a config file and sets up config values
appropriately. Different values are required by different functions in
other modules.

I have tried defining the config values as global in main, AND in the
other modules - but does not work (get undefine errors).
main.py:
  global config_values
  config_values = []
...
 load config_values
...

Module x
 global config_values

def function1():
  try to access config_values[0] fails - not found

So what I am currntly doing is passing what ever config values are needed
through the function call
  x.function1(config_values, etc)

How would you do this??

Once again thanks for help with indirect calls.




From marcus at infa.abo.fi  Tue Apr  8 07:12:07 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 8 Apr 2003 13:12:07 +0200
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <mailman.1049778853.27420.python-list@python.org>
Message-ID: <slrnb9580p.7og.marcus@infa.abo.fi>

>> I continue to be amazed that there's simply no way to say
>> 
>>   import "../foo/ook.py"
>> 
>> as should be possible to implement any well-architected class hierarchy.
>
>  It's simple.  Code should not depend directly on code higher than it in
>the package hierarchy.  If it does (and it isn't one of a few extremely rare
>cases), it is misdesigned.
>
>  Code shallow in the hierarchy should manage required references to objects
>required by code deep in the hierarchy by passing them around explicitly. 
>Any other solution is just bunk.

Could you explain why using import via PYTHONPATH is different than
import via current directory?

E.g define a class in file foobar/t2.py:

class A:
	pass


And then, use it from foobar/test.py:

from foobar import t2 as t1
import t2

print id(t1.A)
print id(t2.A)

Why are the id:s different? That is, why do I get two copies of the A
class?  This leads (IMHO) to the fact that one _never_ should use
"import X", (witness isinstance(t1.A(), t2.A) returning 0...)
everything should go via PYTHONPATH. Which in turn means that
you have to set PYTHONPATH for all programs. ?

Some insight into why this is done as it is would be helpful.

-- 
Marcus Alanen
maalanen at abo.fi



From joshway_without_spam at myway.com  Fri Apr 25 12:58:13 2003
From: joshway_without_spam at myway.com (Joshua Marshall)
Date: Fri, 25 Apr 2003 16:58:13 +0000 (UTC)
Subject: pythong referencing/aliasing
References: <ylcqa.9116$h%2.963225@read1.cgocable.net>
Message-ID: <b8bpf5$hmj$1@ginger.mathworks.com>

Kevin Howe <khowe at perfnet.ca> wrote:
> How does Python handles variable references/aliases?

> Example:

> a = 1
> b = (REF)a
> b = 2
> print a

> In this case, since "b" is just a reference to "a", changes to "b"
> automatically change "a".

> Does python have this ability?

You're basically asking two questions, and both answers are "no".
There is no way to associate two variables the same address, and there
is no REF/address-of operator in Python.

As others have mentioned, you might be able to use lists to give you
the behavior you're looking for.



From gh at ghaering.de  Wed Apr 23 16:27:07 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 23 Apr 2003 22:27:07 +0200
Subject: Curses on Win32
In-Reply-To: <b86r8p$27n$1@gateway.northgrum.com>
References: <b86r8p$27n$1@gateway.northgrum.com>
Message-ID: <3EA6F71B.7040206@ghaering.de>

Dang Griffith wrote:
>    I'm running 2.3a2 and decided to spearmint with curses, only to find
> it's not available for import on Win32.  Is/was curses available in 2.2?
>  Is there an alternative text user interface library available?

Please excuse me just referring you to recent discussions of this topic:

http://groups.google.com/groups?q=group:comp.lang.python+curses+win32&scoring=d

-- Gerhard




From ckea25d02 at sneakemail.com  Mon Apr 14 23:20:36 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 15 Apr 2003 15:20:36 +1200
Subject: Complexity of standard Python data structures (Re: Speed of string
 += string)
In-Reply-To: <slrnb9lsbe.4pa.marcus@infa.abo.fi>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi> <Oqwma.11524$LB6.319489@news1.tin.it> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it> <slrnb9lsbe.4pa.marcus@infa.abo.fi>
Message-ID: <b7ft5m$hcs0$1@ID-169208.news.dfncis.de>

Marcus Alanen wrote:
> As an example of "unfinished" interfaces: there is no list.clear()
> function, even though dict.clear() exists.

I don't think you need one, because it would be
equivalent to

   del lst[:]

and probably slower as well.

> either Python is doing something very very odd to get
> both methods fast, or then a list is really just a vector and
> list.insert() is O(N) --- something which I would've expected a
> "warning" about.

Giving "warnings" about that sort of thing might scare
people more than is warranted. While technically it's true
that list.insert is O(N) in CPython, you need a very large
list before this becomes apparent, because the data shuffling
is done by such a tight C loop that most of the time it's
practically instantaneous compared to all the machinations
the intepreter is going through to get to each of your
list.insert calls. So, *most* of the time, it's *effectively*
O(1).

I'm not saying that the O() behaviours of these things should
be kept a secret or anything. But you need to interpret them
in the context of how Python works generally. With a C++
mindset, someone might see that list.insert is O(N) and think
"Omigosh! It's O(N)! I can't possibly use that, it'll be far
too slow!" when, in fact, it would be a lot *faster* to use
it instead of coming up with some fancier solution that has
what looks like a better O() behaviour...

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From max at alcyone.com  Fri Apr 25 22:03:56 2003
From: max at alcyone.com (Erik Max Francis)
Date: Fri, 25 Apr 2003 19:03:56 -0700
Subject: Exception and finally question
References: <imqgavovcf95pdqgcj44ia80cntvn5k4lc@4ax.com> <3EA86EDA.EFD023C0@alcyone.com> <b8coru$u8m$0@216.39.172.122>
Message-ID: <3EA9E90C.BA1E9C2F@alcyone.com>

Bengt Richter wrote:

> Hm, how would you catch all exceptions raised as strings (e.g., raise
> "ho", "hum")
> and print both string and 2nd arg? Do I have to dig with inspect?
> It might be convenient if the system would convert such exceptions to
> a standard one,
> so we could write e.g.,
>     except  OldStringException, e:
>         print e, e.args
> and see the particular string and also a message arg.

The way I handled them in EmPy was to do an except Exception, e: ... and
if that got passed up, do an except: e = sys.exc_type.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ I am a gentlemen:  I live by robbing the poor.
\__/ George Bernard Shaw
    Bosskey.net: Return to Wolfenstein / http://www.bosskey.net/rtcw/
 A personal guide to Return to Castle Wolfenstein.



From cnetzer at mail.arc.nasa.gov  Thu Apr  3 19:46:52 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 03 Apr 2003 16:46:52 -0800
Subject: <KeyPress>, focus, and Pmw widgtes
In-Reply-To: <4378fa6f.0304031602.33c3bcee@posting.google.com>
References: <4378fa6f.0304031602.33c3bcee@posting.google.com>
Message-ID: <1049417211.598.33.camel@sayge.arc.nasa.gov>

On Thu, 2003-04-03 at 16:02, Marc wrote:
> Hi,
> 
> I was trying to use a keypress to trigger an event in a Gui. Reading
> through past postings I discovered that the widget has to have focus
> to actually register the event. However, even setting the focus on the
> widget component I am unable to trigger the event. Obviously doing
> something wrong, I look to the masses for help. Here is my code:

First of all, in the future, please post a working example, complete
with import statements, class names, mainloop() etc., so that I or
others do not have to spend our time typing it all in.  You will have
better luck with responses, I'd wager.

Next, the event type you are looking for is '<KeyRelease>', not
'<KeyPress-Up>'.

Third, the Pmw widget probably intercepts (or at least, doesn't forward)
these event types.  You can try the following:

instead of:

    self.fps.bind( '<KeyPress-Up>', self.getTBPS)

do this:

    self.fps.component('entry').bind( '<KeyRelease>', self.getTBPS)

This sets the binding on the actual entry Tkinter component of the Pmw
mega widget.  This will get you started.  However, it is perhaps not the
best way to do what you want.

For one thing, it may possibly interfere with the operations of the
EntryField widget as a whole (I can't remember if new bindings override
old ones).  But mainly, it won't be triggerred for key-repeat events,
which you may actually want.

Instead, you should read about the validator options of an EntryField
widget.  This can be made to call a function whenever the contents of
the entry widget are modified (where by keypress, or cut-n-paste, etc.)

So, instead of "validate=None", when you create the widget, try
"validate=self.getTBPS".  You will need to return a True or False value
to indicate whether the validation actually worked.  You will also need
to change other things, but that will get you started.  Use the
facilities that Pmw provides; the author spent a lot of time making them
all work reasonably well.


-- 
Bay Area Python Interest Group - http://www.baypiggies.net/

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)






From tlavoie at acm.org  Thu Apr 10 12:00:51 2003
From: tlavoie at acm.org (Tim Lavoie)
Date: 10 Apr 2003 11:00:51 -0500
Subject: ANN: Capon 0.1 - a software build framework
References: <mailman.1049901964.20840.clpa-moderators@python.org> <b72fo5$bq6$1@panix1.panix.com>
Message-ID: <87ptnuxryk.fsf@theasylum.dyndns.org>

>>>>> "aahz" == aahz  <aahz at pythoncraft.com> writes:

    aahz> So how does Capon differ from SCons?

Well, the name has funnier connotations, for one.   :)



From aleax at aleax.it  Tue Apr 15 04:18:22 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 08:18:22 GMT
Subject: Using property in classic class may not work
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au>
Message-ID: <ifPma.15344$LB6.409776@news1.tin.it>

Neil Hodgson wrote:

>     The "property" feature in Python 2.2 appears very helpful to me and
> seems to work on both classic and new-style classes. However, pychecker
> gives this error message:
> 
> DemonTerm.py:563: Using property (frame) in classic class DemonFrame may
> not work
> 
>     The class in question is derived from a wxWindows class, so I can't
> make it derive just from object. The "What's New in Python 2.2" document
> only mentions property wrt new-style clases. Can I depend on property
> working for classic classes? Is there a workaround such as deriving from
> object as well?

No, and yes, respectively -- deriving from object means your class
isn't classic any more, so properties do work as intended there.  Vide:

>>> class X:
...   def __init__(self, x=23): self.xx=x
...   def getx(self): return 1000+self.xx
...   def setx(self,x): self.xx=x
...   x=property(getx,setx)
...
>>> x=X()
>>> x.x
1023
>>> x.x=99
>>> x.x
99

The assignment hasn't been caught by setx, so x.xx is still
23 and x.x is now the integer 99, not a property any more.  But:

>>> class Xob(X,object): pass
...
>>> x=Xob()
>>> x.x
1023
>>> x.x=99
>>> x.x
1099
>>>

as you see, this does work as intended.


Alex




From martin at v.loewis.de  Sun Apr 27 04:14:41 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 27 Apr 2003 10:14:41 +0200
Subject: Problems making Python-2.3b1
References: <HuLqa.2812$BL6.1867@news-binary.blueyonder.co.uk>
Message-ID: <m3he8ktkzy.fsf@mira.informatik.hu-berlin.de>

John Ochiltree <johnochiltree at blueyonder.co.uk> writes:

> Any ideas gratefully received.

It's a typo in setup.py. Replace

        if (tcllib is None or tklib is None and
            tcl_includes is None or tk_includes is None):
            # Something's missing, so give up
            return

with

        if (tcllib is None or tklib is None or
            tcl_includes is None or tk_includes is None):
            # Something's missing, so give up
            return

Alternatively, install the Tcl/Tk header files if you also want to
build Tkinter.

Please do report bugs at sf.net/projects/python in the future; a bug
report in comp.lang.python *will* get lost.

Regards,
Martin



From BPettersen at NAREX.com  Mon Apr  7 17:33:11 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Mon, 7 Apr 2003 15:33:11 -0600
Subject: socket question(3) 
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDDAA@admin56.narex.com>

> From: Rick Y [mailto:msg_2222 at yahoo.com] 
> 
> installed openssl package from sunfreeware. error:
> 
> ./viewcvs-install
> Traceback (most recent call last):
[...]
> ImportError: ld.so.1: python: fatal: libgcc_s.so.1:
> open failed: 
> No such file or directory
[...]

i.e. it can't find libgcc_s.so.1 on LD_LIBRARY_PATH (not a Solaris guy
anymore, so I couldn't tell you where to get it if you don't have it
installed...)

-- bjorn




From selwyn at home.net.nz  Tue Apr 22 02:25:08 2003
From: selwyn at home.net.nz (selwyn)
Date: Tue, 22 Apr 2003 18:25:08 +1200
Subject: problem starting python interpreter in emacs (win98)
Message-ID: <V65pa.4289$AB5.573029@news02.tsnz.net>

system: win98
python version 2.2.2
emacs  21.2.1
..........
hi all,

I am having diffculty getting the python interpreter to run under emacs,
either from the menu or with C-c ! Syntax highlighting and .py file 
recognition works fine.

When I try strating the interpreter with the above Emacs complains:
"Searching for program: No such  file or directory, python"

I have set the environment variable  HOME to C:\ which is where .emacs
resides, and I have byte compiled python-mode.el which sits in C:\Python22

My .emacs file is as follows:

(setq load-path (cons "C:\Python22" load-path))
(setq auto-mode-alist
        (cons '("\\.py$" . python-mode) auto-mode-alist))
(setq interpreter-mode-alist
        (cons '("python" . python-mode)
              interpreter-mode-alist))
(autoload 'python-mode "python-mode" "Python editing mode." t)
(global-font-lock-mode t)
(setq font-lock-maximum-decoration t)

(require 'cua)
(CUA-mode t)

..........
I guessed that (setq interpreter-mode-alist...etc is telling Emacs where
the interpreter is located, and so I tried adjusting the value of
"python" to "C:\Python22\python" and "C:\Python22\python.exe".

However I am slightly confused about when Emacs reloads .emacs. Do I
have to restart windows everytime I make a change, or is there a method
for reloading while Emacs is running?

Any help greatly appreciated.

Cheers,
Selwyn






From roy at panix.com  Fri Apr 18 12:41:20 2003
From: roy at panix.com (Roy Smith)
Date: Fri, 18 Apr 2003 12:41:20 -0400
Subject: Classes in Python
References: <3EA01A9B.4030209@vnet.net>
Message-ID: <roy-B8E873.12412018042003@reader1.panix.com>

Bill Martin <wcmartin at vnet.net> wrote:
> I'm wondering about the value of allowing a class definition like this:
> 
>              class C:
>                pass

The most common use I can think for an empty class like that is for 
defining exceptions.

> Now I can define a = c() and b = c(), then say a.x1 = 1.2, b.x2 = 3.5 or 
> some such. If I try to print a.x2 or b.x1, I get an exception message 
> basically saying those member variables don't exist.

Of course :-)

When you said a = c(), you generated an instance of c.  Likewise, b = 
c() generated a second instance.  These both inherit everything that the 
base class has (which in this case, is nothing).

Next, you said, a.x1 = 1.2, creating an attribute x1 in instance a.  
This doesn't create an attribute x1 in the entire class, just in the 
particular instance a.  So, when you tried to access b.x1, you got an 
error.

You can see what attributes any object has with the dir() function.  It 
might be instructive to try playing around with code like this and watch 
what happens:

class c:
   pass

dir(c)

a = c()
dir(a)
a.x = 1
dir(a)
dir(c)

c.y = 2
dir(c)
b = c()
dir(a)
dir(b)
dir(c)

Extra credit for deranged minds: figure out a way to make:

a = c()
b = c()

try:
   print b.x
except AttributeError:
   print "wugga wugga"

a.x = "foo"

try:
   print b.x
except AttributeError:
   print "wugga wugga"

print "wugga wugga" and then print "foo".  Here's an example to prove it 
works (on Python 2.2):

bash-2.05a$ ./deranged.py 
wugga wugga
foo



From edcjones at erols.com  Sun Apr 20 21:57:30 2003
From: edcjones at erols.com (Edward C. Jones)
Date: Sun, 20 Apr 2003 21:57:30 -0400
Subject: Good Python generators example?
In-Reply-To: <paDoa.1611$hT2.1228371@news2.news.adelphia.net>
References: <paDoa.1611$hT2.1228371@news2.news.adelphia.net>
Message-ID: <b7vj7h$su1$1@bob.news.rcn.net>

Robert Oschler wrote:
> Can someone point me to a nice concise Python 2.2 compatible example of the
> use of generators?

def walk(top, hidden=False):
     """ Walk through a tree of directories"""
     if not hidden and os.path.basename(top)[0] == '.':
         return
     if os.path.exists(top):
         yield top
     if not os.path.isdir(top):
         return
     names = listdir(top)
     if names == None:
         return
     for name in names:
         fullname = os.path.join(top, name)
         for name2 in walk(fullname, hidden):
             yield name2




From paul at fxtech.com  Tue Apr 15 14:42:15 2003
From: paul at fxtech.com (Paul Miller)
Date: Tue, 15 Apr 2003 13:42:15 -0500
Subject: create a new instance in C of a Python subclass
Message-ID: <36ko9vknj9mm6m1ad18r3dm51q3t02pe03@4ax.com>

I've got a C extension type that is subclassed in Python. In C, I want to
create a new instace of the subclass, which I have a pointer to. What
magical incantation will allow this? I'm assuming some combination of
PyObject_New and PyObject_Init are required, using the ob->ob_type pointer?





From mcfletch at rogers.com  Thu Apr 10 01:37:01 2003
From: mcfletch at rogers.com (Mike C. Fletcher)
Date: Thu, 10 Apr 2003 01:37:01 -0400
Subject: Proper categorization and metadata for Python rpms?
Message-ID: <3E9502FD.2060300@rogers.com>

I'm looking at how one goes about creating RPM distributions for Python 
packages (using distutils). The documentation doesn't give any 
recommendations regarding the proper specification of dependencies or 
provided features, so I am posting here with my current plans, in the 
hope that some Linux people will be kind enough to correct that 
understanding.

Here is my current plan:

    PyOpenGL is dependent on quite a few libraries (and one tool),
    Python 2.2.2, Numeric Python, Tkinter, OpenGL, GLU, GLUT (and SWIG
    1.3.13).

        I specify the libraries as "requires" options to the bdist_rpm
        command:

            'requires':'python,OpenGL,glut,tkinter,tk,Numeric'

        Not sure how to specify the requirement for SWIG (and
        particularly that version). Note that the only RPM'd version of
        Numeric seems to be v20 (we are sensitivie to that annoying
        shuffling of a structure introduced between v21 and v22, so we
        would like to be able to specify Numeric-22 or something like
        that (and then have an RPM available for it, of course)).

        Similarly, not sure what to specify regarding dependencies on
        source tarballs (e.g. for header files) or the like for
        building.  I'm assuming you can specify something in
        build-requires, but I'm not sure what.

    PyOpenGL also provides quite a number of facilities, which I would
    like the user to be able to specify separately, particularly as we
    may eventually break out the various components into multiple RPMs. 
    I'm particularly interested in what I should call these
    "capabilities" provided by the package.  Here's what I'm currently
    planning:

        'provides':'python-OpenGL,python-GLU,python-GLUT,python-GLE,tk-togl'

    That is, all of the items which are python-specific (i.e. python
    libraries) are given the python- prefix, while the tk component
    (togl) is given the tk- prefix (togl doesn't currently have a
    non-python RPM AFAICS). 

However, when I look at the few Python libraries I find with RPMs 
(Numeric, Tkinter, wxPython, pychecker), the capabilities that are 
declared are just "Numeric", "tkinter", "wxPython" and 'pychecker'.  Is 
this a general RPM policy?  Is there a mechanism for de-cluttering the 
namespace?

General guidance on how to properly build RPMs for Python packages would 
be helpful too (e.g. is it best to do so as root or as an unprivelidged 
user? Is there anything needed in the shell environment to ensure that 
the RPMs are generally usable?)

bdist_wininst is so much easier, really :) ,
Mike

_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/







From abpillai at lycos.com  Thu Apr  3 02:08:24 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 2 Apr 2003 23:08:24 -0800
Subject: os.chmod() question ?
Message-ID: <bd993a2f.0304022308.19f36fec@posting.google.com>

Hi pythonfolks,
 
    What numerical argument does os.chmod() accept on windoze .
For a unix box, I used to give the following for files:

chmod - 777 => All rwx permissions enabled (-rwxrwxrwx)
chmod - 644 => RW for owner, r for group/world (-rw-r--r--)

I attempted os.chmod() command on Windoze on similar lines
on a read-only file.

-r-xr-xr-x   1 DS\apillai      DS\Domain+Users       0 Apr  3 12:00 test

Python 2.2.1 (#34, Apr  9 2002, 19:34:33) [MSC 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import os
>>> os.chmod('test', 777)
>>> ^Z

This was the result.
-r-xr-xr-x   1 DS\apillai      DS\Domain+Users       0 Apr  3 12:00 test

Now,

Python 2.2.1 (#34, Apr  9 2002, 19:34:33) [MSC 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import os
>>> os.chmod('test', 644)
>>> ^Z

This is the result.
-rwxrwxrwx   1 DS\apillai      DS\Domain+Users       0 Apr  3 12:00 test

    Apparently, os.chmod() works differently from original chmod() command.
On a unix box, the results were totally different.

Do I need to pass an octal integer as the second argument ?

Thanks
Anand Pillai



From peter at engcorp.com  Mon Apr 28 14:27:13 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 28 Apr 2003 14:27:13 -0400
Subject: problem with string.lower function
References: <b8jrgf$ashp7$1@ID-145503.news.dfncis.de>
Message-ID: <3EAD7281.D1BB6D90@engcorp.com>

Don Low wrote:
> 
> def filefind(word, filename):
>         count = 0
>         try:
>                 fh = open(filename, 'r')
>         except:
>                 print filename, ":", sys.exc_info()
> 
>         allLines = fh.readlines()
>         fh.close()
> 
>         for eachLine in allLines:
>         test = string.lower(eachLine) // change is here
>         if string.find(test, word) > -1:
>                 count = count + 1
>                 print eachLine,
> 
> When I run the script, however, I get the following:
> 
>  File "2bin/./fgrepwc.py", line 37
>                  if string.find(eachLine, word) > -1:
>              ^
> SyntaxError: invalid syntax

Your indentation is almost certainly messed up.  Please make
sure you are using *spaces only*, or that you are using tabs
consistently (but really, it's better to avoid TABs!) or you
might get the above types of problem.

If you are not mixing in TABs, then the problem is simply that
the lines after the "for" statement are not indented at all,
but in Python indentation is used to denote blocks so you *must*
indent all lines in the subblock by the same amount, and more than
the for statement's indentation, or you'll have broken code.

-Peter



From drew at astro.pas.rochester.edu  Fri Apr  4 07:45:03 2003
From: drew at astro.pas.rochester.edu (Drew Moore)
Date: 4 Apr 2003 04:45:03 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <7h3znn7k308.fsf@pc150.maths.bris.ac.uk> <3E8CBC89.733D01D1@alcyone.com> <b6iigl$mk3$0@216.39.172.122> <3E8CF4B5.1E221DD7@alcyone.com>
Message-ID: <9383481e.0304040445.a3752dd@posting.google.com>

Erik Max Francis <max at alcyone.com> wrote in message news:<3E8CF4B5.1E221DD7 at alcyone.com>...
> Bengt Richter wrote:
> 
> > So how do you feel about properties? Seems like the OP's proposal
> > would
> > open up another way to implement at least the assignment part. Perhaps
> > there should be an analogous __getvalue__ method for the right hand
> > side.
> > (Already ducking ;-)

I had the same thought..

> 
> I really should have said "rebinding" instead of "rebinding/attribute
> assignment."  I misspoke, sorry.
> 
> I know
> 
> 	something = somethingElse
> 
> is always going to mean rebinding, and I like that..

Can we follow up on this a little?
I'm trying to understand this concern in more detail..

I'd like this PEP to go one of two ways..
either implementation,
or a good solid technical reason why its a bad idea.

so far, the response seems to be:
"it is technically possible, but my scripts might
run slower.. and it just makes me nervous."

I suspect there is some basis for this nervousness, but
I still haven't seen an example of how this proposed feature
would break things.

since no current code uses __assign__,
nothing would break immediately.

however, once it started to be used.. what would go wrong??

nothing?   ## in your dreams, kiddo
all hell would break loose? ## far more likely.

What are use cases when
something = somethingElse

with overloadable assignment

will give real problems in a real-world code example?

would help() immediately break?
would inspect immediately break?
would pydoc immediately break?

I'm an electrical engineer..
I have never won any popularity
contests with my programming skills (or lack thereof)...

still,
I'd like to see this PEP die for a good technical reason.
help me out here!!

Drew



From aleax at aleax.it  Wed Apr 16 05:13:40 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 09:13:40 GMT
Subject: find non-empty value in a dictionary
References: <b7ib8b$3rk$1@lust.ihug.co.nz> <mailman.1050457707.7350.python-list@python.org>
Message-ID: <899na.20916$LB6.519252@news1.tin.it>

Dylan Reinhardt wrote:

> What about:
> 
> if x:
>    <stuff>
> elif y:
>    <stuff>
> 
> Etc, etc.
> 
> Seems to work fine with a float value of 0.  There may be some edge case
> where this doesn't work, but if there is, I haven't found it.
> 
> If you're just looking for the result (not which variable had the
> non-zero value), you could do something like (untested):
> 
> result = max([d[var] / d['P'] for var in ['X','Y','Z']])

This will fail if the non-zero entry is < 0 ...


Alex




From claird at lairds.com  Tue Apr 15 08:05:15 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 15 Apr 2003 12:05:15 -0000
Subject: Speeding up: s += "string"
References: <mailman.1050393203.4704.python-list@python.org> <UoPma.20694$T34.658757@news2.tin.it>
Message-ID: <v9ntbr3ioli8d3@corp.supernews.com>

In article <UoPma.20694$T34.658757 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
			.
			.
			.
>If I wanted to pursue this kind of things, I think I'd try putting
>a Boost Python wrapper on top of STL's "rope" implementation of
>strings (note: STL and the Standard C++ Library are DIFFERENT things).
>
>This would give mutable strings that allow, not just s1+=s2, but all 
>sort of modifications, with very different performance characteristics
>from what one normally thinks of.  Measuring performance of text
>processing with those vs normal Python strings might be interesting...
>
>
>Alex
>

I'm not optimistic.  I'll be more precise:  the problem
is real, and good solutions are sensitive to data char-
acteristics.  It's easy to improve performance of the
s1 += s2 operations that characterize e-mail filtering
while simultaneously degrading unacceptably the same
operation embedded in an XML filter.  I am close to
asserting the conjecture that every solution harms too
many existing legitimate users.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr  3 12:04:39 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 3 Apr 2003 17:04:39 GMT
Subject: os.chmod() question ?
References: <bd993a2f.0304022308.19f36fec@posting.google.com>
Message-ID: <slrnb8oqd5.174.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <bd993a2f.0304022308.19f36fec at posting.google.com>,
Anand B Pillai wrote:
> Hi pythonfolks,
>  
>     What numerical argument does os.chmod() accept on windoze .
> For a unix box, I used to give the following for files:
> 
> chmod - 777 => All rwx permissions enabled (-rwxrwxrwx)
> chmod - 644 => RW for owner, r for group/world (-rw-r--r--)
> 
> I attempted os.chmod() command on Windoze on similar lines
> on a read-only file.
<snip> 
>     Apparently, os.chmod() works differently from original chmod() command.
> On a unix box, the results were totally different.
> 
> Do I need to pass an octal integer as the second argument ?

The Python interpreter can't interpret literals in different bases
according to which function you call, so you must put a '0' at the
beginning of octal numbers.

Anyway, Windows file permissions work in a totally different way
from Unix permissions.  The only simple flag is the R (read-only)
attribute.  In NT/2000/XP all other permissions are controlled by
an ACL.  In toy Windows everything else is always permitted.

Under Windows, I think os.chmod() ignores everything except the
owner-writable flag (0200) and sets the R attribute to the inverse
of that.  Cygwin will show permissions as either 0777 or 0555
depending on the state of the R attribute.

So 777 decimal = 01411 octal which means the R attribute is set
and the permissions read as 0555, and 644 decimal = 01204 octal
which means the R bit is cleared and the permissions read as 
0777.  All clear?



From martin at v.loewis.de  Sat Apr 19 03:03:01 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 19 Apr 2003 09:03:01 +0200
Subject: make test fails on NetBSD 1.6
References: <ca3cce40.0304181809.5f704e9b@posting.google.com>
Message-ID: <m3y9279dga.fsf@mira.informatik.hu-berlin.de>

carlggb at terc.edu (Carl Alexander) writes:

> I'm new to Python.  Building 2.2.2 on NetBSD 1.6 seems to go fine, but I get
> this from make test.  It's not quite clear to me how I should go about asking
> "someone to teach regrtest.py about which tests are expected to get skipped 
> on netbsd1".

If you consider yourself a NetBSD expert, you should determine which
of these tests should have worked: e.g. should it be expected that
NetBSD supports the en_US locale (in which case test_locale should
have worked); should it be expected that the system supports fork with
threads (in which test_fork1 should be executed), and so on.

If you have a list of tests that you would expect to be executed, find
out why the tests that got skipped did so; for the rest, submit a
patch to sf.net/projects/python to make an entry in
Lib/test/regrtest.py to list the tests that are expected to be skipped.

You should be concerned about the failure of test_threading; it means
that your thread library doesn't work as Python expects it to work. In
that case, it is best to either fix the problem, or compile Python
with --without-threads.

Regards,
Martin



From aleax at aleax.it  Sat Apr 26 14:51:10 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 26 Apr 2003 18:51:10 GMT
Subject: In which versions is list.sort stable?
References: <slrnbalhft.bc1.mlh@furu.idi.ntnu.no>
Message-ID: <yyAqa.8330$3M4.231032@news1.tin.it>

Magnus Lie Hetland wrote:

> Has list.sort always been stable? Or, if not, in which Python versions
> is it not stable?

list.sort has not been stable in 1.5.2, 2.0.*, 2.1.*, 2.2.*.  I do not
know about previous versions.  Note that list.sort may once again cease
being stable (in 2.4.* or further versions) if the timbot dreams up one
that's even better on _other_ criteria -- stability's not going to
constrain his algorithm choices.


Alex




From spam at magnetic-ink.dk  Wed Apr  2 05:13:51 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Wed, 2 Apr 2003 10:13:51 +0000 (UTC)
Subject: Python equivalent of _tempnam?
References: <3e8ab45b$1@mail.hmgcc.gov.uk>
Message-ID: <80173e13-af17-4641-b1a5-590ec21b1dbe@news.szn.dk>

Richard wrote:

> I have been told that a way of doing this in C is by using
> _tempnam( ). However I want to write this in Python, so is
> there an equivalent function?

See the tempfile module.


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From logistix at zworg.com  Sat Apr 12 21:31:29 2003
From: logistix at zworg.com (logistix)
Date: 12 Apr 2003 18:31:29 -0700
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <3E985DC8.9060907@thuule.pair.com> <mailman.1050178039.28676.python-list@python.org>
Message-ID: <bdb5b58f.0304121731.1b1312f4@posting.google.com>

Alan McIntyre <fusion at thuule.pair.com> wrote in message news:<mailman.1050178039.28676.python-list at python.org>...
> Alan McIntyre wrote:
> > Now I'm inclined to do a comparison just for my own curiosity's 
> > sake...maybe I'll be back with graphs in just a bit. :)
> 
> Here's the times required for adding 200k two-character strings to a 
> single string (in steps of 1000 += operations):
> 
> http://norfolkgraphics.com/testing/test.png
> 
> The green line is for Python, red for C++.  It looks kind of O(n)'ish 
> for both of them.
> 
> Each point in the graph corresponds to the time required to perform 1000 
> += operations (the Python version of the code is below).  I dumped the 
> output of the Python and C++ programs to text files and used SciPy to 
> plot them.
> 
> The total run time for the Python version was 194.069 seconds, the C++ 
> std::string version was 157.386 seconds.   I would have done a plain C 
> version but my curiosity ran out and it's too nice outside. ;)
> 
> Full disclosure: The C++ program was compiled with Borland C++ Builder 
> 5, all optimizations turned on; the Python program is just as you see it 
> below, run on Python 2.2.2 (binary install release from python.org). 
> YMMV.  IANAL.
> 
> ----------------------------------------------------------
> import time
> 
> x = []
> t = []
> s = ''
> 
> starttime = time.time()
> 
> for i in range(200):
>      ts = time.time()
>      for j in range(1000):
>          s += 'aa'
> 
>      x.append(i)
>      t.append(time.time()-ts)
                           ^^^

This is wrong.  You should be doing t.append(time.time()-starttime). 
That will show how much more time 1k, 2k, 3k etc operations will take
compared to 1k.

 Right now you're only timing over and over for groups of 1000.  If
this was linear [O(n)], you'd get a flat line on the graph with your
current equation.



From sross at connectmail.carleton.ca  Tue Apr  1 18:39:05 2003
From: sross at connectmail.carleton.ca (Sean Ross)
Date: Tue, 1 Apr 2003 18:39:05 -0500
Subject: Questions on linked lists
References: <kl8ia.27315$OV.127379@rwcrnsc54> <eD8ia.4560$9I.350859@news20.bellglobal.com> <2tdpl-3i3.ln1@beastie.ix.netcom.com>
Message-ID: <dppia.1860$DD6.490810@news20.bellglobal.com>

Hi.


"Dennis Lee Bieber" <wlfraed at ix.netcom.com> wrote in message
news:2tdpl-3i3.ln1 at beastie.ix.netcom.com...
> Sean Ross fed this fish to the penguins on Monday 31 March 2003 08:33
> pm:
>

....um...fish?...
whatever...


> > http://www.ibiblio.org/obp/thinkCSpy/chap17.htm
> >
> > This is a chapter on linked lists, implemented in Python.
> >
> > It covers wrappers, helpers, preconditions, and invariants.
> >
> > Hope this helps,
>
>         <blink> Well, I'm surprised (after my previous response)...
>

Actually, as far as I know, my posting was made before your previous
response, so your response had no bearing on the content of my message. As
such,  your surprise is surprising to me ;)


>
>         Though I begin to wonder if this may not be the same source the
> original poster was confused by... After all, the terminology is
> identical (whereas I could see some other text having minor variations)
>
>

It's not possible to tell from the original poster message (see below)
whether they were using the tutorial I suggested, they did however ask for
"a good explanation of this subject esp in the context of Python". My
suggestion matched this criteria, IMHO. As far as having identical
terminology, well, other than 'cargo', all of the other terms are quite
common, AFAIK, when discussing linked-lists.


Kevin Reeder wrote:
>The tutorial I'm working through has a section on Linked Lists that
>has left me puzzled. I believe I got the basic concept of nodes, embedded
>references, and cargo, but when the discussion turned to wrappers,
>helpers, preconditions and invariants, I started taking on water. Can
anyone
>here point me to documentation with a good explanation of this
>subject esp in the context of Python?


OK, then. Have a nice day,
Sean





From dberlin at dberlin.org  Sat Apr 12 20:52:43 2003
From: dberlin at dberlin.org (Daniel Berlin)
Date: Sat, 12 Apr 2003 20:52:43 -0400 (EDT)
Subject: Speed of string += string
In-Reply-To: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net>
Message-ID: <Pine.LNX.4.53.0304122050120.11576@dberlin.org>


On Sat, 12 Apr 2003, Mads Orbesen Troest wrote:

> Hi;
>
> Given that strings are immutable, is it exceedingly slow (and memory
> spiking) to do a lot of "string += string" operations in one's code? IOW,
> does each and every += evaluation lead to a new concatenated copy of the
> previous string (now freed for garbage collection) and the new string, so
> that the longer the string to which stuff is appended is, the longer times
> each += operation takes?

Yes.
Try the following:

x = ""
for i in xrange(300000):
	x += 'a' * 80

versus

import array
y = array.array('c')
for i in xrange(300000):
	y.fromstring('a' * 80)
x = y.tostring()


You'll find the first one takes minutes, the second, 4 seconds.




From gustav at morpheus.demon.co.uk  Mon Apr 14 18:11:34 2003
From: gustav at morpheus.demon.co.uk (Paul Moore)
Date: Mon, 14 Apr 2003 23:11:34 +0100
Subject: Displaying Unicode on the console (Windows)
References: <182bcf76.0304140744.1b3d43ee@posting.google.com> <m3y92cvolo.fsf@mira.informatik.hu-berlin.de>
Message-ID: <k7dwsp9l.fsf@morpheus.demon.co.uk>

martin at v.loewis.de (Martin v. L?wis) writes:

> Microsoft has added API to support output of arbitrary Unicode
> characters on the console, regardless of the code page
> (WriteConsoleW).

Thanks. I'll investigate this.

> This only works if the console uses a Unicode font (Lucida Sans
> Unicode).

How do I get this? I have Lucida Sans Unicode installed, but in the
console properties screen, the only font I can choose is Lucida
Console...

Paul.
-- 
This signature intentionally left blank



From skip at pobox.com  Mon Apr  7 20:43:04 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 7 Apr 2003 19:43:04 -0500
Subject: exceptions with spaces don't work??
In-Reply-To: <b6t48i$mvd$1@news.doit.wisc.edu>
References: <b6t48i$mvd$1@news.doit.wisc.edu>
Message-ID: <16018.6936.291043.692184@montanaro.dyndns.org>

    Paul> If I raise an exception with a string, as in:
    Paul>     raise 'somestring'
    Paul> then my statement
    Paul>     except 'somestring':
    Paul> will catch it unless there are spaces....
    ...
    Paul> Is this a bug??

Dunno, however, note that string exceptions have been deprecated for a long
time (since 2.0 or 2.1 I think) and will completely go away at some point.
If there's a bug there I doubt fixing it will be real high priority.

Skip





From anthony.evershed at npl.co.uk  Mon Apr  7 05:15:31 2003
From: anthony.evershed at npl.co.uk (CTony)
Date: 7 Apr 2003 02:15:31 -0700
Subject: COM and MS Word
Message-ID: <2eb85cfb.0304070115.6aecb644@posting.google.com>

It appears that some data files I need are in WordPerfect (.dat)
format, and I need to get them into Python for processing and later
entry into a database. I haven't been able to find a way of
interfacing to WordPerfect directly, but MS Word 2000 can read these
files...

Here's the snag, though. All the examples of Python/COM/Word usage
I've seen have involved pushing data into Word, rather than extracting
data *from* it. I know that the data field I need is x lines below a
known heading, so getting to the field is no problem:

finder = wd.Selection.Find
finder.Text = 'foo'
finder.Execute()
for n in range(x):
    wd.Selection.MoveDown
wd.Selection.HomeKey

get you to the beginning of the field, no problems. How, however, can
I select (block mark) and extract that line of text?

If it helps, I'm using Python 2.2, with win32all build 148, on Windows
NT 4 (though the resulting program will be used on Windows 2000 as
well)...

Many thanks,

Tony Evershed



From duncan at NOSPAMrcp.co.uk  Thu Apr 24 11:00:32 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 24 Apr 2003 15:00:32 +0000 (UTC)
Subject: php equivalents?
References: <OHSpa.8996$h%2.915957@read1.cgocable.net>
Message-ID: <Xns9367A1EA7F9D5duncanrcpcouk@127.0.0.1>

"Kevin Howe" <khowe at perfnet.ca> wrote in
news:OHSpa.8996$h%2.915957 at read1.cgocable.net: 

> I'm switching from PHP to Python, and am wondering if there are Python
> equivalents to the following two PHP functions:
> 
>> register_shutdown($func)
> 
> This is basically a "cleanup" function. Just before a script ends, it
> will invoke all functions registered using register_shutdown(). This
> can be used to close database connections, clear sessions, save data,
> etc.

Use the atexit module to register functions that are called on normal 
interpreter shutdown.

> 
>> var_export($var)
> 
> This function converts a variable to a text string of valid code which
> can be reevaluated at any time. Python has pickle/cPickle, and also
> has modules that can dump/load to XML, however, none of these create
> actual code, but instead create a formatted data string.
> 
> Any help is much appreciated.

When would it be advantageous to output executable code instead of pickling 
or marshalling an object? You wouldn't want executable code that an end 
user could modify, and if the end user can't see it why would they care 
about the format.

There is no standard way to produce executable code that will reconstruct 
arbitrary Python objects although to a certain extent the repr function 
will do that for some types (e.g. numbers, and simple tuples, lists and 
dictionaries).

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From rlratzel at siliconmetrics.com  Tue Apr  8 12:46:17 2003
From: rlratzel at siliconmetrics.com (Rick Ratzel)
Date: 8 Apr 2003 09:46:17 -0700
Subject: Calling python from a C function that is embedded in python?
References: <mailman.1049722812.17109.python-list@python.org>
Message-ID: <905fcd91.0304080846.73d1acfe@posting.google.com>

I just released elmer 1.0.2, which might be able to help.  Elmer lets
you write code in Python and run it in either C or Tcl.  It will pass
function calls and data types (both native and custom) between the two
languages...at the moment, you can write a module in Python and call
it from C or Tcl, as if it was written in C or Tcl.

   The online docs need alot of work, but the test cases under "test"
in the distribution is the best place to look for examples (test2
shows running a Python module from a C program, which sounds like what
you want to do).  The code is still considered alpha, and any feedback
is greatly appreciated.  Thanks, and let me know if that helps...

http://elmer.sourceforge.net


VanL <vlindberg at verio.net> wrote in message news:<mailman.1049722812.17109.python-list at python.org>...

> In short:  I want python to be able to call wrapped functions like they 
> are python functions, and a wrapped C library to call wrapped python 
> functions like they were in C.
> 
> Is there any way to do this?



From oblivious at web.de  Tue Apr  8 04:48:35 2003
From: oblivious at web.de (Dominic)
Date: Tue, 08 Apr 2003 10:48:35 +0200
Subject: Comms Protocol design dilemma
In-Reply-To: <mailman.1049738831.30630.python-list@python.org>
References: <mailman.1049738831.30630.python-list@python.org>
Message-ID: <3E928CE3.4090008@web.de>

> number generation involved, I can pretty much rule out asynchronous
> sockets. Is there any better strategy to tackle this problem?

Well asynchronous sockets are fine. I've written a load-test
client in Python which handles 1000 tcp connections simultaneously.
(1000 sockets is the per process limit for non root users)
I've used the python asyncore module. If you use asynchronous
sockets you need to keep your session state explicitly since
you do not have threads. You could use a finite-state-machine
or Python's generators. (See my previous mail on comp.lang.python,
Programming Erlang style in Python. You should certainly
check out the Erlang programming language homepage (for the concepts)
from Erricson since that language has been designed for servers which 
you intend to write.)

Ciao,
  Dominic




From carlggb at terc.edu  Fri Apr 18 22:09:59 2003
From: carlggb at terc.edu (Carl Alexander)
Date: 18 Apr 2003 19:09:59 -0700
Subject: make test fails on NetBSD 1.6
Message-ID: <ca3cce40.0304181809.5f704e9b@posting.google.com>

I'm new to Python.  Building 2.2.2 on NetBSD 1.6 seems to go fine, but I get
this from make test.  It's not quite clear to me how I should go about asking
"someone to teach regrtest.py about which tests are expected to get skipped 
on netbsd1".

[much output skipped]....
test_zipfile
test_zlib
166 tests OK.
1 test failed:
    test_threading
26 tests skipped:
    test_al test_asynchat test_cd test_cl test_curses test_dl
    test_email_codecs test_fork1 test_gdbm test_gl test_imgfile
    test_linuxaudiodev test_locale test_minidom test_nis test_pyexpat
    test_queue test_sax test_socket_ssl test_socketserver
    test_sunaudiodev test_thread test_threaded_import
    test_threadedtempfile test_winreg test_winsound
Ask someone to teach regrtest.py about which tests are
expected to get skipped on netbsd1.
*** Error code 1

Stop.
make: stopped in /usr/local/src/Python-2.2.2



From guy at obstruction-no-spam.com  Sun Apr 20 14:10:28 2003
From: guy at obstruction-no-spam.com (Guy Middleton)
Date: Sun, 20 Apr 2003 18:10:28 GMT
Subject: arg-passing style question
References: <A4ioa.87812$Vzu.36931@news02.bloor.is.net.cable.rogers.com>
Message-ID: <ooBoa.98763$Vzu.41106@news02.bloor.is.net.cable.rogers.com>

In article <A4ioa.87812$Vzu.36931 at news02.bloor.is.net.cable.rogers.com>,
Guy Middleton <guy at obstruction-no-spam.com> wrote:
> I want to pass two values to a function that expects a single argument.
> 
> So, I could wrap them in either a list or a tuple.  Is either way preferred
> over the other?

Tuples appear to be faster, enough to make a difference if I'm doing lots of
these:

>>> setup = """
... a = 1
... b = 2
... def fn(x):
...         pass
... """
>>> print timeit.Timer("fn((a,b))", setup, clock).repeat()
[2.90625, 2.921875, 2.921875]
>>> print timeit.Timer("fn([a,b])", setup, clock).repeat()
[4.5859375, 4.6015625, 4.5859375]
>>> 



From mikael at isy.liu.se  Fri Apr  4 01:50:24 2003
From: mikael at isy.liu.se (Mikael Olofsson)
Date: Fri, 4 Apr 2003 08:50:24 +0200
Subject: Precidence of unary plus/minus relative to exponentiation
In-Reply-To: <D7A3CFD7825BD6119B880002A58F06C202F2C774@groexmb02.pfizer.com>
References: <D7A3CFD7825BD6119B880002A58F06C202F2C774@groexmb02.pfizer.com>
Message-ID: <20030404085024.49248d77.mikael@isy.liu.se>

On Thu, 3 Apr 2003 10:56:35 -0500
"Warnes, Gregory R" <gregory_r_warnes at groton.pfizer.com> wrote:
> Oops.  Again it appears that I was typing with my brain off....  

Don't worry. We all do that once in a while.

/Mikael

-----------------------------------------------------------------------
E-Mail:  mikael at isy.liu.se
WWW:     http://www.dtr.isy.liu.se/dtr/staff/mikael               
Phone:   +46 - (0)13 - 28 1343
Telefax: +46 - (0)13 - 28 1339

         /"\
         \ /     ASCII Ribbon Campaign
          X      Against HTML Mail
         / \

This message was sent by Sylpheed.
-----------------------------------------------------------------------
Link?pings kammark?r: www.kammarkoren.com





From python at dylanreinhardt.com  Sat Apr 19 15:25:15 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 19 Apr 2003 12:25:15 -0700
Subject: Parent references: hot or not?
In-Reply-To: <20030419165923.GA26681@meson.dyndns.org>
References: <1050763980.14721.41.camel@ida.dylanreinhardt.com>
	 <20030419145756.GA25916@meson.dyndns.org>
	 <1050771656.14717.48.camel@ida.dylanreinhardt.com>
	 <1050772061.14712.55.camel@ida.dylanreinhardt.com>
	 <20030419165923.GA26681@meson.dyndns.org>
Message-ID: <1050780314.14722.58.camel@ida.dylanreinhardt.com>

On Sat, 2003-04-19 at 09:59, Jp Calderone wrote:
>   Hope this helps,

Yes, your code example was exactly what I needed to see.

Thanks a bunch!

Dylan






From aahz at pythoncraft.com  Tue Apr 15 22:47:09 2003
From: aahz at pythoncraft.com (Aahz)
Date: 15 Apr 2003 22:47:09 -0400
Subject: Speed of string += string
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <roy-06B012.20145112042003@reader1.panix.com> <bJ2ma.5447$883.4511@nwrdny02.gnilink.net> <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net>
Message-ID: <b7ig7d$ale$1@panix1.panix.com>

In article <iz86drbfq4va$.1r0k6kjvji1h5.dlg at 40tude.net>,
Mads Orbesen Troest  <mads at troest.NEVERMORE.dk> wrote:
>On Sun, 13 Apr 2003 01:05:11 GMT, Carl Banks wrote:
>>
>> Constant.
>> Constant.
>> Constant.
>
>But is that /guaranteed/ (as it is with STL), or does it just happen to
>be that way with the current implementation (however unlikely it is to
>change)?

Both, sort of.  Python uses dicts internally; you can rely on any
changes to achieve roughly equivalent or better performance guarantees.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From aleax at aleax.it  Mon Apr 28 07:28:53 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 28 Apr 2003 11:28:53 GMT
Subject: dynamic function names
References: <20030428102400.21c5074a.bram@nospam.sara.nl> <iP5ra.14084$3M4.382689@news1.tin.it> <Xns936B6D56043E4duncanrcpcouk@127.0.0.1>
Message-ID: <Vf8ra.20839$K35.702350@news2.tin.it>

Duncan Booth wrote:

> Alex Martelli <aleax at aleax.it> wrote in
> news:iP5ra.14084$3M4.382689 at news1.tin.it:
> 
>> See?  "def funcname():" *RE-BINDS* identifier funcname to
>> refer to the function object.
>> 
>> So, what you need is to call function 'function' in module
>> 'new'.  You can first build a temporary function with a def
>> to give you the right code, globals, defaults and closure,
>> and then pass each of these to new.function together with
>> the funcname you want to use:
>> 
>>>>> import new
>>>>> def makefunc(funcname):
>> ...   def temp(): print "this function is named", funcname
>> ...   f = new.function(temp.func_code, globals(), funcname,
>> ...           temp.func_defaults, temp.func_closure)
>> ...   return f
>> ...
>>>>> f = makefunc('foo')
>>>>> f()
>> this function is named foo
>>>>> f.func_name
>> 'foo'
> 
> Note that this code only works on Python 2.3 (because the closure argument
> to new.function didn't exist earlier), and even there if the function
> 'foo' throws an exception the traceback will still call it 'temp'.

Very good points, thanks!  Of course, they can be fixed -- but you
need to prepare a new suitable code-object rather than just reuse
the one from 'temp'.  It's just a bit longer, because a code object
has a LOT of attributes you must pass to make a new one;-).

Given this, I would factor this a bit differently:

import new

def make_named_func(name, func=None):
    if func is None:
        def func(*args): return args
    c = func.func_code
    nc = new.code(c.co_argcount, c.co_nlocals, c.co_stacksize, c.co_flags,
        c.co_code, c.co_consts, c.co_names, c.co_varnames, c.co_filename,
        name,
        c.co_firstlineno, c.co_lnotab, c.co_freevars, c.co_cellvars)
    nf = new.function(nc, func.func_globals, name)
    return nf

f1 = make_named_func('onename')
print f1('fee', 'fie', 'foo', 'fum')

def raiser(): return 1/0

f2 = make_named_func('nametwo', raiser)
print f2()


This works equally well in both Python 2.2 and 2.3.  Of course, it would
be MUCH easier if co_name wasn't a *read-only* attribute of code objects
(we could just assign c.co_name in that case...), but, oh well...;-).


Alex




From ckea25d02 at sneakemail.com  Wed Apr 30 23:32:11 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 01 May 2003 15:32:11 +1200
Subject: Raising objects
In-Reply-To: <b8p4au$dh7$1@slb9.atl.mindspring.net>
References: <453d10b2.0304290750.222fb453@posting.google.com> <ltxra.31737$K35.1001968@news2.tin.it> <mailman.1051678851.21661.python-list@python.org> <OFLra.36816$K35.1143840@news2.tin.it> <mailman.1051719837.14426.python-list@python.org> <b8p4au$dh7$1@slb9.atl.mindspring.net>
Message-ID: <b8q3l9$cjj44$1@ID-169208.news.dfncis.de>

Andrew Dalke wrote:
> However, if all exceptions were derived from Exception
> then you could do
> 
> except Exception, e:
>    ... use e ...
> 
> This cleans up the wart at the expense of not allowing
> any old object to be thrown.

Another related wart could be cleaned up at the same
time: if the traceback were available as an attribute
of the exception, it would eliminate the need for
sys.exc_info altogether.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mail at andreas-penzel.de  Wed Apr 16 07:23:05 2003
From: mail at andreas-penzel.de (Andreas Penzel)
Date: Wed, 16 Apr 2003 13:23:05 +0200
Subject: Problem Cron-Job using Python-Script
Message-ID: <b7jeem$1l5i0$1@ID-69142.news.dfncis.de>

Hello!

I have a Python-Script to create Log-Files und work with them.
Very well.
But if I start the Script with Cron it works not correct.
I often use os.system() to send commands to the shell including paths from
the os.
Could this be the problem?
And if, how to solve?

Thanks, Andreas





From mads at troest.NEVERMORE.dk  Tue Apr  8 12:31:23 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Tue, 8 Apr 2003 18:31:23 +0200
Subject: Loading a class the name of which is only known runtime
Message-ID: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net>

Hi folks;

"And now for something completely different:"

I have a situation in which I, from a CGI script written in Python, want to 
instantiate a class, the name of which is known to me only at runtime 
(because it depends on a CGI parameter).

Is it possible to load (import and instantiate) a class dynamically in 
Python? I have a feeling the answer is Yes, but as I am unfortunately not 
quite the python expert (yet :) forgive me if I ask something obvious.

Thanks in advance,
   /\/\\ads Orbesen Troest



From staschuk at telusplanet.net  Sat Apr 12 10:08:28 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 12 Apr 2003 08:08:28 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>; from mads@troest.NEVERMORE.dk on Sat, Apr 12, 2003 at 02:41:46PM +0200
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>
Message-ID: <20030412080828.A361@tibia.amotlpaa.bogus>

Quoth Mads Orbesen Troest:
  [...]
> Which is the most correct to use; or should I even overload both? The 
> documentation says __str__ is for the, quote, informal representation; 
> whereas __repr__ is for the, quote, official representation. Furthermore, 
> __reptr__ should, quote, look like an expresseion. I'm not quite sure what 
> is meant by these distinctions. Is it that __repr__ is used by pickle and 
> the like, and - thus - should represent the entire object state? [...]

Pickle doesn't use __repr__; it has other, more reliable, ways of
getting object state.

__repr__ is used by: (1) '%r' % foo and repr(foo); (2) '%s' % foo,
str(foo), and print foo if __str__ isn't provided; and (3) the
interactive interpreter, when printing the value of an expression
entered at the prompt.  (And no doubt, other places that I've
forgotten.)

The distinction is not a functional one; it is between different
kinds of information about the object which a string
representation might provide.  The documentation in this case
means just what it says.

An example:  The other day I described here an algorithm for
finding palindromes.  The code I wrote to illustrate it has a
class which represents palindromes-in-progress; its __str__
representation is, for example,
    'a man ... (pa)nama'
(Here 'a man' has matched 'nama', and the next step in building
the palindrome is to find a word or words starting with 'ap'.)
This is an informal, convenient representation.  It is, as it
happens, suitable for display to an end user (say, in a status
message indicating what the palindrome search is looking at
presently).

The __repr__ representation, on the other hand, is
    'NearPalindrome([['a', 'man'], ['panama']], 'pa', 0)'
This representation is nearly useless to an end user; but it gives
a complete account of the object and could be of use to a
programmer.

-- 
Steven Taschuk                          staschuk at telusplanet.net
"Its force is immeasurable.  Even Computer cannot determine it."
                           -- _Space: 1999_ episode "Black Sun"




From simonb at webone.com.au  Tue Apr  8 13:35:19 2003
From: simonb at webone.com.au (Simon Burton)
Date: Wed, 09 Apr 2003 03:35:19 +1000
Subject: Loading a class the name of which is only known runtime
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net>
Message-ID: <pan.2003.04.08.17.35.15.762651@webone.com.au>

On Tue, 08 Apr 2003 18:31:23 +0200, Mads Orbesen Troest wrote:

> Hi folks;
> 
> "And now for something completely different:"
> 
> I have a situation in which I, from a CGI script written in Python, want to 
> instantiate a class, the name of which is known to me only at runtime 
> (because it depends on a CGI parameter).
> 
> Is it possible to load (import and instantiate) a class dynamically in 
> Python? I have a feeling the answer is Yes, but as I am unfortunately not 
> quite the python expert (yet :) forgive me if I ask something obvious.
> 
> Thanks in advance,
>    /\/\\ads Orbesen Troest

[simon at arrow simon]$ python
Python 2.2.2 (#2, Nov 24 2002, 11:41:06) 
[GCC 3.2 20020903 (Red Hat Linux 8.0 3.2-7)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
Make it so.
>>> class foo:
...   pass
... 
>>> 
>>> name = "foo"
>>> 
>>> locals()
{'__builtins__': <module '__builtin__' (built-in)>,
 '__doc__': None,
 '__name__': '__main__',
 'foo': <class __main__.foo at 0x81c302c>,
 'inspect': <module 'inspect' from '/usr/lib/python2.2/inspect.pyc'>,
 'my_print': <function my_print at 0x81c1f44>,
 'name': 'foo',
 'pprint': <module 'pprint' from '/usr/lib/python2.2/pprint.pyc'>,
 'readline': <module 'readline' from '/usr/lib/python2.2/lib-dynload/readline.so'>,
 'rlcompleter': <module 'rlcompleter' from '/usr/lib/python2.2/rlcompleter.pyc'>,
 'string': <module 'string' from '/usr/lib/python2.2/string.pyc'>,
 'sys': <module 'sys' (built-in)>,
 'types': <module 'types' from '/usr/lib/python2.2/types.pyc'>}
>>> locals()[name]
<class __main__.foo at 0x81c302c>
  __doc__=None
  __module__='__main__'
>>> instance = locals()[name]()
>>> instance
<__main__.foo instance at 0x81c4c64>
  __doc__=None
  __module__='__main__'
>>> 




From rmunn at pobox.com  Mon Apr 14 16:41:41 2003
From: rmunn at pobox.com (Robin Munn)
Date: Mon, 14 Apr 2003 20:41:41 GMT
Subject: newbie: strange python problem
References: <pan.2003.04.10.12.27.06.600319@somedomain.com> <3E956A20.E3321A5E@engcorp.com> <pan.2003.04.10.13.07.59.235430@somedomain.com>
Message-ID: <slrnb9lihd.1q1.rmunn@localhost.localdomain>

maxx <nospam at somedomain.com> wrote:
> On Thu, 10 Apr 2003 08:57:04 -0400, Peter Hansen wrote:
> 
>> Check the output of "ls -la /" manually... the first line does
>> *not* have more than two items in it...  if you were to change
>> the first line above to this, it would work, though I suspect
>> there is another option to "ls" which would also suppress
>> that extra output.
>> 
>>> for line in os.popen('ls -la /').readlines()[1:]:
>> 
>> -Peter
> 
> Right! A simple mistake, but those are the ones you make taking your
> first steps in a new programming language.

In addition to this, you may have problems if any filenames had spaces
in them. See Skip Montanaro's reply to your post for a better solution
that would properly handle spaces in filenames.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From martin at v.loewis.de  Fri Apr 18 04:56:45 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 18 Apr 2003 10:56:45 +0200
Subject: Problem in Accessing C++ Class Instance Global Variable (near success, help please!!)
References: <ea0e2298.0304180031.49129224@posting.google.com>
Message-ID: <m38yu89oaa.fsf@mira.informatik.hu-berlin.de>

alanpong at hkstar.com (Alan Pong) writes:

> I has one global c++ class instance variable.  I want to call its
> member
> function through wrapper function which is accessed by python.
> However, as seen in wrapper.cpp, it give me another new variable copy
> so that phython actually accessing another instance.
> 
> Does anybody has idea to fix it?

No, but I have an idea how to further investigate.

>From the code you have shown, it appears that getInstance is not used
at all. So I recommend to delete the functions

> extern "C" CTest * getInstance()
> {
> 	return &gtest;  //***Question is here: a new variable!!! why?***
> }

This one: it isn't used, so your question should be irrelevant.

> static PyObject *
> import_getInstance( PyObject *self, PyObject *args )
> {
>   char *obj = getInstance();
>   return Py_BuildValue("i", obj);
> }

and this one. It uses above function, but, in turn, isn't used,
either. So your program should continue to work without visible
changes if you remove those functions.

If that isn't the case, you have concealed some detail from us that
might be important in understanding the mystery.

Regards,
Martin



From lekishvili at python.qartu.com  Fri Apr  4 07:54:02 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 4 Apr 2003 04:54:02 -0800
Subject: server cluster support in Python?
Message-ID: <a3b5b608.0304040454.736b680b@posting.google.com>

Does Python have the one?

GRTZ,
Giorgi



From mwh at python.net  Mon Apr 28 07:20:14 2003
From: mwh at python.net (Michael Hudson)
Date: Mon, 28 Apr 2003 11:20:14 GMT
Subject: RuntimeError at end of test.autotest in Python 2.3b on windows, tmpfile also fails
References: <mailman.1051367863.3218.python-list@python.org> <3eaaa408$0$49109$e4fe514c@news.xs4all.nl> <3eac5942$0$49098$e4fe514c@news.xs4all.nl>
Message-ID: <7h3llxukgx5.fsf@pc150.maths.bris.ac.uk>

Irmen de Jong <irmen at -NOSPAM-REMOVETHIS-xs4all.nl> writes:

> PS if developers don't run "import test.autotest" what *do* they run?

"make test" (or at least, that's what I do).

Cheers,
M.

-- 
81. In computing, turning the obvious into the useful is a living
    definition of the word "frustration".
  -- Alan Perlis, http://www.cs.yale.edu/homes/perlis-alan/quotes.html



From drifty at bigfoot.com  Fri Apr  4 14:44:35 2003
From: drifty at bigfoot.com (Brett C.)
Date: 4 Apr 2003 11:44:35 -0800
Subject: python-dev Summary for 2003-03-16 through 2003-03-31
Message-ID: <mailman.1049583243.7529.clpa-moderators@python.org>

+++++++++++++++++++++++++++++++++++++++++++++++++++++
python-dev Summary for 2003-03-16 through 2003-03-31
+++++++++++++++++++++++++++++++++++++++++++++++++++++

This is a summary of traffic on the `python-dev mailing list`_ from
March 16, 2003 through March 31, 2003.  It is intended to inform the
wider Python community of on-going developments on the list and to
have an archived summary of each thread started on the list.  To
comment on anything mentioned here, just post to
python-list at python.org or `comp.lang.python`_ with a subject line
mentioning what you are discussing. All python-dev members are
interested in seeing ideas discussed by the community, so don't
hesitate to take a stance on something.  And if all of this really
interests you then get involved and join `python-dev`_!

This is the fourteenth summary written by Brett Cannon (Managed to
keep my sanity as long as A.M. Kuchling did before he stopped doing
the Summaries =).

All summaries are archived at http://www.python.org/dev/summary/ .

Please note that this summary is written using reStructuredText_ which
can be found at http://docutils.sf.net/rst.html .  Any unfamiliar
punctuation is probably markup for reST_ (otherwise it is probably
regular expression syntax); you can safely ignore it (although I
suggest learning reST; its simple and is accepted for `PEP markup`__).
 Also, because of the wonders of programs that like to reformat text,
I cannot guarantee you will be able to run the text version of this
summary through Docutils_ as-is unless it is from the original text
file.

__ http://www.python.org/peps/pep-0012.html

.. _python-dev: http://www.python.org/dev/
.. _python-dev mailing list:
http://mail.python.org/mailman/listinfo/python-dev
.. _comp.lang.python:
http://groups.google.com/groups?q=comp.lang.python
.. _Docutils: http://docutils.sf.net/
.. _reST:
.. _reStructuredText: http://docutils.sf.net/rst.html

.. contents::


.. _last summary: http://www.python.org/dev/summary/2003-03-01_2003-03-15.html

======================
Summary Announcements
======================
PyCon is now over!  It was a wonderful experience.  Getting to meet
people from python-dev in person was great.  The sprint was fun and
productive (work on the AST branch, caching where something is found
in an inheritance tree, and a new CALL_ATTR opcode were all worked
on).  It was definitely worth attending; I am already looking forward
to next year's conference.

I am trying a new way of formatting the Quickies_ section.  I am
trying non-inline implicit links instead of inlined ones.  I am hoping
this will read better in the text version of the summary.  If you have
an opinion on whether the new or old version is better let me know. 
And remember, the last time I asked for an opinion Michael Chermside
was the only person to respond and thus ended up making an executive
decision.

.. _PyCon: http://www.python.org/pycon/


========================
`Re: lists v. tuples`__
========================
__ http://mail.python.org/pipermail/python-dev/2003-March/034029.html

Splinter threads:
    - `Re: Re: lists v. tuples
<http://mail.python.org/pipermail/python-dev/2003-March/034070.html>`__

This developed from a thread from covered in the `last summary`_ that
discussed the different uses of lists and tuples.  By the start date
for this summary, though, it had turned into a discussion on
comparisons.  This occured when sorting heterogeneous objects came up.
 Guido commented that having anything beyond equality and non-equality
tests for non-related objects does not make sense.  This also led
Guido to comment that "TOOWTDI makes me want to get rid of __cmp__"
(TOOWTDI is "There is Only One Way to Do It").

Now before people start screaming bloody murder over the possible
future loss of __cmp__() (which probably won't happen until Python 3),
realize that all comparisons can be done using the six other rich
comparisons (__lt__(), __eq__(), etc.).  There is some possible code
elegance lost if you have to use two rich comparisons instead a single
__cmp__() comparison, but it is nothing that will prevent you from
doing something that you couldn't do before.  There can also be a
performance penalty in sorting in some instances.

This all led Guido to suggest introducing the function before().  This
would be used for arbitrary ordering of objects.  Alex Martelli said
it would "be very nice if before(x,y) were the same as x<y whenever
the latter doesn't raise an exception, if feasible".  He also said
that it should probably "define a total ordering, i.e. the implied
equivalence being equality".


================================
`Fast access to __builtins__`__
================================
__ http://mail.python.org/pipermail/python-dev/2003-March/034243.html

There has been rumblings on the list as of late of disallowing
shadowing of built-ins.  Specifically, the idea of someone injecting
something into a module's namespace that overrides a global (by doing
something like ``socket.len = lambda x: 42`` from the socket module)
is slightly nasty, rarely done, and prevents the core from optimizing
for built-ins.

Raymond Hettinger, in an effort to see how to speed up built-in
access, came up with the idea of replacing opcode calls of LOAD_GLOBAL
and replace them with LOAD_CONST after putting the built-in being
called into the constants table.  This would leave shadowing of
built-ins locally unaffected but prevent shadowing at the module. 
Raymond suggested turning on this behavior for when running Python -O.

The idea of turning this on when running with the -O option was shot
down.  The main argument is that semantics are changed and thus is not
acceptable for the -O flag.  It was mentioned that -OO can change
semantics, but even that is questionable.

So this led to some suggestions of how to turn this kind of feature
on.  Someone suggested something like a pragma (think Perl) or some
other mechanism at the module level.  Guido didn't like this idea
since he does not want modules to be riddled with code to turn on
module-level optimizations.

But all of this was partially shot down when Guido stepped in and
reiterated he just wanted to prevent outside code from shadowing
built-ins for a module.  The idea is that if it can be proven that a
module does not shadow a built-in it can output an opcode specific for
that built-in, e.g. len() could output opcode for calling
PyOject_Size() if the compiler can prove that len() is not shadowed in
the module at any point.

Neil Schemanauer suggested adding a warning for when this kind of
shadowing is done.  Guido said fine as long as extension modules are
exempt.  Now no matter how well the warning is coded, it would be
*extremely* difficult to catch something like ``import X; d =
X.__dict__; d["len"] = lambda x: 42``.  How do you deal with this?  By
Guido saying he has not issue saying something like this "is always
prohibited".  He said you could still do ``setattr(X, "len", lambda x:
42)``, though, and that might give you a warning.  Neil's patch can be
found at http://www.python.org/sf/711448 .


================================
`capability-mediated modules`__
================================
__ http://mail.python.org/pipermail/python-dev/2003-March/034149.html

Splinter threads:
    - `Capabilities <http://mail.python.org/pipermail/python-dev/2003-March/034152.html>`__

The thread that will not die (nor does it look like it will in the
near future; Guido asked to postpone discussing it until he gets back
from `Python UK`_ which will continue the discussion into the next
summary.  I am ending up an expert at capabilities against my will. 
=)

In case you have not been following all of this, capabilities as being
discussed here is the idea that security is based on passing around
references to objects.  If you have a reference you can use it with no
restrictions.  Security comes in by controlling who you give
references to.  So I might ask for a reference to file(), but I won't
necessarily get it.  I could, instead, be handed a reference to a
restrictive version of file() that only opens files in an OSs
temporary file directory.  So, in capabilities-land, executing
``open_file = file`` only works if you have the reference to 'file',
otherwise the assignment fails and you just don't get access.  If that
is not clear, read the `last summary`_ on this thread.  And now, on to
the new stuff...

There were also suggestions to add arguments to import statements to
give a more fine-grained control over them.  But it was pointed out
that classes fit this bill.

The idea of limiting what modules are accessible by some code by not
using a universally global scope (i.e., not using sys.modules) but by
having a specific scope for each function was suggested.  As Greg
Ewing put it, "it would be dynamic scoping of the import namespace".

While trying to clarify things (which were at PyCon thanks to the Open
Space discussion held there on this subject), a good distinction
between a rexec_ world (as in the module) and a capabilities was made
by Guido.  In capabilities, security is based on passing around
references that have the amount of power you are willing for it to
have.  In a rexec world, it is based on what powers the built-ins give
you; there is no worry about passing around code.  Also, in the rexec
world, you can have the idea of a "workspace" where __builtin__ has
very specific definitions of built-ins that are used when executing
untrusted code.

It has been pointed out that rexec can be viewed as a specific
implementation of capabilities.  Since you are restricting what
references code by making only certain references available to the
codes you are using capabilities, just more at a namespace level than
on a per-reference level.

Ka-Ping Yee wrote up an example of some code of what it would be like
to code with capabilities (can be found at
http://mail.python.org/pipermail/python-dev/2003-March/034287.html ).

.. _Python UK: http://www.python-uk.org/
.. _rexec: http://www.python.org/dev/doc/devel/lib/module-rexec.html

=========
Quickies
=========

`tzset`__
    time.tzset() is going to be kept in Python, but only on UNIX.  The
testing suite was also loosened so as to not throw as many
false-negatives.

__ http://mail.python.org/pipermail/python-dev/2003-March/034062.html

`Windows IO`__
    stdin and stdout on Windows are TTYs.  You can get 3rd-party
modules to get more control over the TTY.

__ http://mail.python.org/pipermail/python-dev/2003-March/034102.html

`Who approved PyObject_GenericGetIter()???`__
    Splinter threads: `Re: [Python-checkins] python/dist/src/Modules
_hotshot.c,...`__; `PyObject_GenericGetIter()`__
    Raymond Hettinger wrote a function called
PyObject_GenericGetIter() that returned self for objects that were an
iterator themselves.  Thomas Wouters didn't like the name and neither
did Guido since it was generic at all; it worked specifically with
objects that were iterators themselves.  Thus the function was renamed
to PyObject_SelfIter().

__ http://mail.python.org/pipermail/python-dev/2003-March/034107.html
__ http://mail.python.org/pipermail/python-dev/2003-March/034103.html
__ http://mail.python.org/pipermail/python-dev/2003-March/034110.html

`test_posix failures?`__
    A test for posix.getlogin() was failing for Barry Warsaw under
XEmacs (that is what he gets for not using Vim_ =).  Thomas Wouters
pointed out it only works when there is a utmp file somewhere. 
Basically it was agreed the test that was failing should be removed.

__ http://mail.python.org/pipermail/python-dev/2003-March/034120.html
.. _Vim: http://www.vim.org/

`Shortcut bugfix`__
    Raymond Hettinger reported that a change in `_tkinter.c`_ for a
function led to it returning strings or ints which broke PMW_
(although having a function return two different things was disputed
in the thread; I think it used to return a string and now returns an
int).  The suggestion of making string.atoi() more lenient on its
accepted arguments was made but shot down since it changes semantics. 
If you want to keep old way of having everything in Tkinter return
strings instead of more proper object types (such as ints where
appropriate), you can put teh line ``Tkinter.wantobjects = 0`` before
the first creation of a tkapp object.

__ http://mail.python.org/pipermail/python-dev/2003-March/034138.html
.. __tkinter.c: http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/python/python/dist/src/Modules/_tkinter.c
.. _PMW: http://pmw.sourceforge.net/

`csv package ready for prime-time?`__
    Related: `csv package stitched into CVS hierarchy`__
    Skip Montanaro: Okay to move csv_ package from the sandbox into
the stdlib?
    Guido van Rossum: Yes.

__ http://mail.python.org/pipermail/python-dev/2003-March/034162.html
__ http://mail.python.org/pipermail/python-dev/2003-March/034179.html
.. _csv: http://www.python.org/dev/doc/devel/lib/module-csv.html

`string.strip doc vs code mismatch`__
    Neal Norwitz asked for someone to look at
http://python.org/sf/697220 which updates string.strip() from the
string_ module to take an optional second argument.  The patch is
still open.

__ http://mail.python.org/pipermail/python-dev/2003-March/034167.html
.. _string: http://www.python.org/dev/doc/devel/lib/module-string.html

`Re: More int/long integration issues`__
    The point was made that it would be nice if the statement ``if num
in range(...): ...`` could be optimized by the compiler if range() was
only the built-in by substituting it with something like xrange() and
thus skip creating a huge list.  This would allow the removal of
xrange() without issue.  Guido suggested a restartable iterator
(generator would work wonderfully if you could just get everything
else to make what range() returns look like the list it should be).

__ http://mail.python.org/pipermail/python-dev/2003-March/034019.html

`socket timeouts fail w/ makefile()`__
    Skip Montanaro discovered that using the makefile() method on a
socket cause the file-like object to not observe the new timeout
facility introduced in Python 2.3.  He has since patched it so that it
works properly and that sockets always have a makefile() (wasn't
always the case before).

__ http://mail.python.org/pipermail/python-dev/2003-March/034177.html

`New Module? Tiger Hashsum`__
    Tino Lange implemented a wrapper for the `Tiger hash sum`_ for
Python and asked how he could get it added to the stdlib.  He was told
that he would need community backing before his module could be added
in order to make sure that there is enough demand to warrant the
edition.

__ http://mail.python.org/pipermail/python-dev/2003-March/034191.html
.. _Tiger hash sum: http://www.cs.technion.ac.il/~biham/Reports/Tiger/

`Icon for Python RSS Feed?`__
    Tino Lange asked if an XML RSS feed icon could be added at
http://www.python.org/ for http://www.python.org/channews.rdf .  It
has been added.

__ http://mail.python.org/pipermail/python-dev/2003-March/034196.html

`How to suppress instance __dict__?`__
    David Abrahams asked if there was an easy way to suppress an
instance __dict__'s creation from a metaclass.  The answer turned out
to be no.

__ http://mail.python.org/pipermail/python-dev/2003-March/034197.html

`Weekly Python Bug/Patch Summary`__
    Another summary can be found at
http://mail.python.org/pipermail/python-dev/2003-March/034286.html
    Skip Montanaro's weekly reminder how Python ain't perfect.

__ http://mail.python.org/pipermail/python-dev/2003-March/034200.html

`[ot] offline`__
    Samuele Pedroni is off relaxing is is going to be offline for two
weeks starting March 23.

__ http://mail.python.org/pipermail/python-dev/2003-March/034204.html

`funny leak`__
    Christian Tismer discovered a memory leak in a funky def statement
he came up with.  The leak has since been squashed (done at PyCon_
during the sprint, actually).

__ http://mail.python.org/pipermail/python-dev/2003-March/034212.html

`Checkins to Attic?`__
    CVS_ uses something called the Attic to put files that are only in
a branch but not the HEAD of a tree.

__ http://mail.python.org/pipermail/python-dev/2003-March/034230.html
.. _CVS: http://www.cvshome.org/

`ossaudiodev tweak needs testing`__
   Greg Ward asked people who are running Linux or FreeBSD to execute
``Lib/test/regrtest.py -uaudio test_ossaudiodev`` so as to test his
latest change to ossaudiodev_.

__ http://mail.python.org/pipermail/python-dev/2003-March/034233.html
.. _ossaudiodev: http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/python/python/dist/src/Modules/ossaudiodev.c

`cvs.python.sourceforge.net fouled up`__
    Apparently when you get that nice message from SourceForge_
telling you that recv() has aborted because of server overloading you
can rest assured that people with checkin rights get to continue to
connect since they get priority.

__ http://mail.python.org/pipermail/python-dev/2003-March/034234.html
.. _SF: 
.. _SourceForge: http://www.sf.net/

`Doc strings for typeslots?`__
    You can't add custom docstrings to things stored in typeobject
slots at the C level.

__ http://mail.python.org/pipermail/python-dev/2003-March/034239.html

`Compiler treats None both as a constant and variable`__
    As of now the compiler outputs opcode that treats None as both a
global and a constant.  That will change as some point when assigning
to None becomes an error instead of a warning as it is in Python 2.3;
possibly 2.4 the change will be made.

__ http://mail.python.org/pipermail/python-dev/2003-March/034281.html

`iconv codec`__
    M.A. Lemburg stated that he questioned whether the iconv codec was
ready for prime-time.  There have been multiple issues with it and
most seem to stem from a platform's codec and not ones that come with
Python.  This affects all u"".encode() calls when the codec does not
come with Python.  Hye-Shik Chang said he would get his iconv codec NG
patch up on SF in the next few days and that would be applied.

__ http://mail.python.org/pipermail/python-dev/2003-March/034300.html




From joze_storme at yahoo.com  Sun Apr 13 03:15:27 2003
From: joze_storme at yahoo.com (Joze)
Date: 13 Apr 2003 00:15:27 -0700
Subject: Creating a PyObject* from a C double*
Message-ID: <20cb355e.0304122315.c068128@posting.google.com>

Hi,

As part of a Python extension, i have an array of doubles created in a
C function, which should return a corresponding PyObject* to Python.
The array is meant to be used by Numeric/numarray, but apparently, the
C function cannot directly return a PyArrayObject*.

I see that this question has been asked plenty of times in this
newsgroup before, but i cannot seem to make it work. Basically, my C
function is like this:

static PyObject* my_func( PyObject *self, PyObject *args) {

	/* Here i build a variable X of type double* ...  */
	/* and want to pass it back to Python             */

	return ????;

}

I have tried PyArray_FromDims(), PyArray_FromDimsAndData() but i can't
seem to get it right. Can anyone tell me how this works? I'd be most
grateful.

Thanks a lot !
Joe.



From andrew-pythonlist at puzzling.org  Fri Apr 25 06:02:25 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Fri, 25 Apr 2003 20:02:25 +1000
Subject: Char to ASCII and back ?
In-Reply-To: <roy-CF32FA.11400525042003@malibu.unice.fr>
References: <roy-CF32FA.11400525042003@malibu.unice.fr>
Message-ID: <20030425100225.GI15695@frobozz.local>

On Fri, Apr 25, 2003 at 11:40:06AM +0200, Jean-Paul Roy wrote:
> How do you transform a char (or a one-char string) into the integer 
> ascii code and back ? Say to code a message.

You can use ord and chr for single characters:

    >>> ord('a')
    97
    >>> chr(97)
    'a'

But see also the struct module; it may be what you want for more complicated
packing and unpacking of bytes.

-Andrew.





From evan at 4-am.com  Tue Apr 22 10:56:18 2003
From: evan at 4-am.com (Evan Simpson)
Date: Tue, 22 Apr 2003 09:56:18 -0500
Subject: zope TAL
In-Reply-To: <b83j25$3namt$1@hades.rz.uni-saarland.de>
References: <b83j25$3namt$1@hades.rz.uni-saarland.de>
Message-ID: <3EA55812.4070405@4-am.com>

Uwe Schmitt wrote:
> are there any good resources concerning TAL ?

The best tutorial-style introduction to TAL is probably chapters 9 and 
13 of the work-in-progress version of the Zope Book at:

http://www.zope.org/Documentation/Books/ZopeBook/2_6Edition

For historical perspective and technical specs, see:

http://dev.zope.org/Wikis/DevSite/Projects/ZPT/FrontPage

...and click through to OriginalFrontPage.

Cheers,

Evan @ 4-am






From Bill.Scherer at verizonwireless.com  Wed Apr  9 07:01:40 2003
From: Bill.Scherer at verizonwireless.com (Scherer, Bill)
Date: Wed, 9 Apr 2003 07:01:40 -0400 (EDT)
Subject: creating new clobs in DCOracle2
In-Reply-To: <b6vjpb$45h$1@news.Stanford.EDU>
Message-ID: <Pine.LNX.4.44.0304090655360.23210-100000@bagend.odc.vzwcorp.com>

[P&M]

On Tue, 8 Apr 2003, Daniel wrote:

> I'm using DCOracle2 to access oracle and want to insert data into a clob
> column.  The DCOracle2 docs say that you need to first insert an empty clob
> using EMPTY_CLOB() and then query the table to grab it, and finally, write
> to it using the write method.
>  In JDBC you can do this:
> 
> tempClob = CLOB.createTemporary(conn, true, CLOB.DURATION_SESSION);
> 
> and then you can write data into the tempClob, and then insert that into the
> table. Can't you do something like this in DCOracle2?

Yes. It goes something like this (untested):


cursor.execute("""
insert into 
  mytable (foo, bar, clob) 
  values (1, 2, EMPTY_CLOB())
""")

cursor.execute("""
select clob 
from mytable
where foo = 1
and bar = 1
for update
""")

clob = cursor.fetchone()[0]

clob.write(clobData)

db.commit()




> 
> Thanks,

Your welcome!

> Daniel
> 
> 
> 

-- 
Bill.Scherer at Verizon Wireless
RHCE 807101044903581





From mrsinger at usa.net  Tue Apr  1 10:06:28 2003
From: mrsinger at usa.net (Matt Singer)
Date: 1 Apr 2003 07:06:28 -0800
Subject: list index question
Message-ID: <e62b26d3.0304010613.39f98991@posting.google.com>

I'm sure this is probably a simple question but... I'm a python rookie.

I'm reading an index table from a database of two columns using a fetchall() call. 

If I print this, I get something like

(100, 1)     #results[0]
(100, 10)    #results[1]
(101, 1)
(101, 5)

I want to find the index for (100, 10) without looping.  

Any suggestions gracefully accepted.

Thanks



From juneaftn at REMOVETHIShanmail.net  Wed Apr 30 11:26:16 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Thu, 1 May 2003 00:26:16 +0900
Subject: HTTP application streams a response to the client gradually - how deal with it?
References: <Xns936CB5697A6D4cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <b8opts$801$1@news.hananet.net>

"Harald Massa" <cpl.19.ghum at spamgourmet.com> wrote in message
news:Xns936CB5697A6D4cpl19ghumspamgourmet at 62.153.159.134...
>
> Hello,
>
> I want to connect to an HTTP application, that streams its response to
the
> client gradually.
>
> (it is a "web chat")
>
> From reading documentes and google, I learned that urllib will not fit.
>
> Also httplib --- I only see methods that provide me with whole HTTP-
> responses.
>
> Is there any way to deal with "streamed http" without going to socket-
> programming?
>
>
> Tnx,
>
> Harald
>

I don't know if this is want you are looking for or not. But anyway I hope
this can be helpful to anyone who wants to build a "web chat".

Usually, that kind of client-server intercommunication with browsers is
done via java applet. You can hide it as a 1 pixel somewhere around the
corner of the screen. The applet does the communication with the server,
and the javascript does the presentation using Live Connection.

This is one of the best ways of making a  "web chat", AFAIK.

June




From bdesth at removethis.free.fr  Tue Apr 15 19:49:37 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Tue, 15 Apr 2003 23:49:37 +0000
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com>
Message-ID: <3E9C9A91.8090209@removethis.free.fr>

Dave Benjamin wrote:
> It has been my increasing concern that Python is headed in a decidedly
> object-oriented direction, ignoring the benefits it currently offers to
> functional programmers. 

(snip a long and interesting post)

Agree++

Even for OO programmers, functionnal features in Python are IMHO a great 
plus, and BTW functionnal and OO paradigm does not have to conflict 
(that would be functionnal vs imperative and object vs procedural). CLOS 
is one of the great system objects out here, and it's been implemented 
on top of a functionnal language, so...

In fact, I'd rather regret that Python is not more functional *and* 
object oriented !-)

My 2 eurocents
Laotseu




From peter at engcorp.com  Tue Apr 15 20:45:54 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 15 Apr 2003 20:45:54 -0400
Subject: =?iso-8859-1?Q?R=E9ception?= et =?iso-8859-1?Q?d=E9codage?= d'une 
	trame UDP
References: <3e9c9b1e$0$1509$7a628cd7@news.club-internet.fr>
Message-ID: <3E9CA7C2.2C9730EA@engcorp.com>

Benoit BESSE wrote:
> 
> Bonsoir tout le monde,
> 
> qui a d?j? ?crit un module qui ?coute, re?oit, d?code et insert dans une BD
> une trame UDP.
> Je voudrais d?coder la trame UDP ainsi et l'ins?rer dans une table :
>     - Port Source
>     - Port Destination
>     - Longueur
>     - Somme de contr?le
>     - Donn?es
> 
> Merci par avance.

Hmm... almost looks like another "Twisted has that" but I'm 
not sure what "BD" is so I can't be sure...  ;-)

-Peter



From dave at pythonapocrypha.com  Mon Apr 28 23:48:49 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Mon, 28 Apr 2003 20:48:49 -0700 (PDT)
Subject: GUI testing with python  (SendKeys)
In-Reply-To: <b8kbpf$63a$0@198.102.102.163>
References: <b8c6c9$p2m$0@198.102.102.163> <b8ecgs$47i$1@sirius.dur.ac.uk>
 <b8kbpf$63a$0@198.102.102.163>
Message-ID: <Pine.LNX.4.51.0304282045280.24631@ns0.tbye.com>

On Mon, 28 Apr 2003, pj wrote:

> I took a look at a girder.
> It seemed far excessive for my requirements.
> Frankly I havnt understood what its all about yet.
>
> Just the sendkeys offered by Windows Scripting object isnt enough either. I
> need some mouse manipulation too. So I thought perls Win32::GuiTest was a
> good starting point.
>
> Is there another easier to read description of girder somewhere on the net ?

Hi PJ,

I've been happy with AutoIt (http://www.hiddensoft.com/AutoIt/) as it
simulates keypresses, mouse clicks, and a few other things. There is no
Python binding but it does have an ActiveX interface you can use, so that
works just as well, e.g.:

from win32com.client import Dispatch
c = Dispatch('AutoItX.Control')
c.WinActivate('SomeApp', '')
c.LeftClick(400, 200)
...

-Dave




From PegasusAssocs at aol.com  Mon Apr 28 01:20:34 2003
From: PegasusAssocs at aol.com (PegasusAssocs at aol.com)
Date: Sun, 27 Apr 2003 22:20:34 -0700
Subject: Unique lighting products for your special needs...
Message-ID: <GYNNETADtFlmmlQ080G00005e91@gynnetad.gynnnet.com>

	 Pegasus Associates...	

Greetings!

Pegasus Associates is a lighting company
specializing in attractive, unique lighting
products. Here are seven product lines that
may be of interest to you.

CABLE LIGHTING
We have several choices of low voltage cable
lighting kits - with names like Satellite, Mars,
Venus, UFO, and Mercury - these kits come
complete with cable, cable fixtures, MR16
halogen lamps, transformers, and connecting hardware.
http://www.PegasusAssociates.com/products/Cablelight/cablelt.html 

DISPLAY LIGHTING
Pegasus Associates offers several different
display lights - PAR halogen exhibit lights,
low voltage halogen trade show lights, and
quartz halogen arm lights, most with
adjustable or flexible arms.
http://www.PegasusAssociates.com/products/DisplayLts/displt.html 

FIBER OPTIC LIGHTING
We offer several fiber optic lighting 
products - fiber optic downlight, fiber
optic spot light, fiber optic step light,
fiber optic cabinet light, and fiber
optic star ceiling lights.
http://www.PegasusAssociates.com/products/Fiberoptics/fiberop.html 

RECESSED LIGHTING
You will find a variety of low voltage
recessed lights and line voltage recessed 
lighting fixtures with new construction housings,
remodel housings, and housings for insulated ceilings.
http://www.PegasusAssociates.com/products/MiniRecDownlts/minirecdnlts.ht
ml
<http://www.pegasusassociates.com/products/MiniRecDownlts/minrecdnlts.ht
ml>  

SCONCES
You have a choice of many different
wall sconces - incandescent, fluorescent,
glass, metal, paintable ceramic, some with
metal accents, and some ADA compliant.
http://www.PegasusAssociates.com/products/indexframe3.html
<http://www.pegasusassociates.com/indexframe3.html>  

UNDER CABINET LIGHTING
There are many choices of under
cabinet lights - xenon light fixtures,
a xenon light strip, xenon mini-lights,
and several low-profile fluorescent fixtures.
http://www.PegasusAssociates.com/products/undercab/undercab.html 

UV FILTERS
We have several sizes of Optivex?
UV vilters, UV filtering tube guards,
UV flitering diffusers, UV filtering
plastic film, and SoLux? Daylight
MR16 lamps - perfect for retarding 
photochemical degradation in a
variety of sensitive materials.
http://www.PegasusAssociates.com/products/UVFilters/UVFilter.html 

Cordially,
Tom Farin
Pegasus Associates
412.269.0773 phone
412.264.1099 fax
www.PegasusAssociates.com <http://www.pegasusassociates.com> 
"Unique lighting products for your special needs." 

  _____  

To remove yourself from our list, simply type REMOVE in the subject line
and forward this email to info at BizBulk.com with the email address you
would like removed.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030427/6e355927/attachment.html>

From mayhew at cis.ohio-state.edu  Wed Apr 23 15:19:58 2003
From: mayhew at cis.ohio-state.edu (Michael Mayhew)
Date: Wed, 23 Apr 2003 14:19:58 -0500
Subject: Using the MySQLdb module
Message-ID: <3EA6E75E.7070305@cis.ohio-state.edu>

Greetings,

	I am trying to use the MySQLdb module on a simple database local to my 
personal Linux machine. However, I keep running into trouble when I 
execute SQL on my cursor objects, specifically when I execute the insert 
command. I have been through numerous docs on the Python DB API v2.0 and 
on the module itself but have not turned up anything useful. Any 
assistance would be greatly appreciated. Here is the error message:



Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "/usr/lib/python2.2/site-packages/MySQLdb/cursors.py", line 95, 
in execute
     return self._execute(query, args)
   File "/usr/lib/python2.2/site-packages/MySQLdb/cursors.py", line 114, 
in _execute
     self.errorhandler(self, exc, value)
   File "/usr/lib/python2.2/site-packages/MySQLdb/connections.py", line 
33, in defaulterrorhandler
     raise errorclass, errorvalue
_mysql_exceptions.OperationalError: (1054, "Unknown column 'atr' in 
'field list'")



It seems that the values I am passing in my cursor.execute(SQL) 
statement are actually not being parsed as values but rather as column 
headings. I have tried many different forms of the SQL and keep 
producing the same error message. Here also is a sample structure of the 
SQL command, where c is the cursor object:



sql = "insert into refFlatHum values ( %s, %s, %d, %d )" % ('atr', 
'chr1', 90810, 92190)

c.execute(sql)



TIA, Michael Mayhew




From aleax at aleax.it  Wed Apr 23 17:54:35 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 23 Apr 2003 21:54:35 GMT
Subject: Testing if string contains a substring
References: <b865dq$28gi$1@ns.felk.cvut.cz> <NKxpa.85054$T34.2112670@news2.tin.it> <698f09f8.0304231141.5d458eb0@posting.google.com>
Message-ID: <vYDpa.50997$LB6.1152215@news1.tin.it>

Jeremy Fincher wrote:

> Alex Martelli <aleax at aleax.it> wrote in message
> news:<NKxpa.85054$T34.2112670 at news2.tin.it>...
>> However, such new functionality doesn't get back-ported to previous
>> releases of Python, such as 2.2.*.  In all Python releases from 1,6
>> included to 2.3 excluded, "needle.find(haystack) > 0" is the idiom.
>> Module string is needed if you want to use antediluvian versions of
>> Python, such as 1.5.2, where strings didn't have methods.
> 
> Of course, you  meant "haystack.find(needle) >= 0", right?

Yes, sorry.  See my response to Bengt as to why I was confused
between needle-on-left and haystack-on-left;-).


> I always use (and have seen used) "haystack.find(needle) != -1", myself.

I never saw any added value in that one extra character -- if I did
want to use -1 I would also use > rather than != ...;-)


Alex




From rmunn at pobox.com  Fri Apr  4 10:15:19 2003
From: rmunn at pobox.com (Robin Munn)
Date: Fri, 04 Apr 2003 15:15:19 GMT
Subject: Best practice for object attributes?
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk> <3E8AEC44.2090909@removethis.free.fr> <b6eusn$avn$1@nntp0.reith.bbc.co.uk> <2uGia.77$nG6.43@nwrdny01.gnilink.net> <b6jjkc$b30$1@nntp0.reith.bbc.co.uk>
Message-ID: <slrnb8qjcb.i81.rmunn@localhost.localdomain>

Michael Sparks <michaels at rd.bbc.co.uk> wrote:
> Carl, Jp Calderone, laotseu:
> 
> Thanks for the comments between the 3 of you I now understand the
> behaviour of the following ....
> 
>> python2.3
> Python 2.3a2 (#1, Feb 20 2003, 09:54:48)
> [GCC 3.2] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>1> class A: time='today'
> ...
>>2> B=A()
>>3> B_CB=A()
>>4> B.time='yesterday'
>>5> A.time='tomorrow'
>>6> C=A()
>>7> C.time
> 'tomorrow'          #1#
>>8> B.time
> 'yesterday'         #2#
>>9> B_CB.time
> 'tomorrow'          #3#
> 
> ... to be this ....
> 
> 1 Creates a class with class attribute "time", value 'today'.
> 2 Creates an object "B" with an empty __dict__, no local attributes
> 3 Creates an object "B_CB" with an empty __dict__, no local attributes
> 4 Inserts into "B" a local attribute "time" with value 'tomorrow'
> 5 Changes the class attribute "time" to 'tomorrow'
> 6 Creates an object "C" with an empty __dict__, no local attributes
> 
> 7 The system looks in C for a local attribute time, and does not find
>   one. It then looks in the class to see if it has attribute time, finds
>   one and treats *that* as the value of C.time. 
>   Hence the system gets the value of A.time - specifically 'tomorrow'
>   hence the output on line #1#
> 
> 8 The system looks in B for a local attribute time - it finds one with
>   the value 'yesterday' in B's dict, hence the output on line #2#
> 
> 9 The system looks in B_CB for a local attribute time, and does not find
>   one. It then looks in the class to see if it has attribute time, finds
>   one and treats *that* as the value of B_CB.time. 
>   Hence the system gets the value of A.time - specifically 'tomorrow'
>   hence the output on line #3#

Exactly right.

By the way, having just picked up a copy of _Python in a Nutshell_ by
Alex Martelli, I highly recommend it. Run, don't walk, to your nearest
bookstore and buy yourself a copy. (Or do what I do and order from
Bookpool.com and get about 40% off O'Reilly titles!) Pages 75-76 of
_Python in a Nutshell_ have an excellent discussion of this very
subject.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From bignose-hates-spam at and-zip-does-too.com.au  Sat Apr 12 00:42:51 2003
From: bignose-hates-spam at and-zip-does-too.com.au (Ben Finney)
Date: Sat, 12 Apr 2003 04:42:51 GMT
Subject: Package hierarchies and cross-directory 'import'
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <slrnb9f6js.p2.bignose-hates-spam@rose.localdomain.fake>

On Tue, 08 Apr 2003 00:07:56 GMT, Ben Finney wrote:
>  On Mon, 7 Apr 2003 10:45:58 GMT, Michael Hudson wrote:
> > It is a Fact Of Life that Python packages (and modules therein)
> > need to know where they are in the global module namespace.
>  
> I feel dim for not being able to see the answer, but it seems absurd
> to me that one can't rename a directory and expect the scripts within
> it to continue working without outside help.

Responses have ranged from "Yes, it is a really annoying flaw, I wish it
would be fixed" to "Obviously if you can't do it in Python it shouldn't
be done".

The former, while sympathetic, doesn't help.  The latter is rather
head-in-the-sand.  In the meantime, I need a solution.  So I've hacked
one up.

In each package directory, I'll place a version of this _path.py file.
The module_depth constant is set to the number of directories below the
"root" of the package tree.  Then, each module in the directory needs
only to 'import _path' to ensure the root module directory is in the
module search path.

It's a nasty hack.  I find it to be the best option at present, because:

It could be written more concisely, but I prefer readability.

It's somewhat more maintainable than hard coding 'sys.path.append(
"../.." )' into every two-level deep module (and changing it whenever
that module needs to move to a different level).

It can be copied as-is to any package directory within the tree, and
modifying the module_depth constant.

It's far preferable to having the actual filesystem path hard-coded
somewhere (and having the whole package tree fail when the directory is
renamed or moved to someone else's system).

But, gosh, why on earth is this even necessary?  Importing with relative
paths would be so much simpler, no?


_path.py:
=====
#!/usr/bin/python
# _path.py

""" Path initialisation for modules in this package
"""

import sys
import os.path

# Specify how deep this package is in the tree
module_depth = 2

# Get the directory of this module
this_dir = os.path.dirname( __file__ )

# Get the directory of the root module
root_module_dir = this_dir
for i in range( 0, module_depth ):
    root_module_dir = os.path.join( root_module_dir, '..' )

root_module_dir = os.path.normpath( root_module_dir )

# Append the root module directory to the module search path
if( not root_module_dir in sys.path ):
    sys.path.append( root_module_dir )

#
# End of _path.py
=====

-- 
 \     "A cynic is a man who, when he smells flowers, looks around for |
  `\                                   a coffin."  -- Henry L. Mencken |
_o__)                                                                  |
http://bignose.squidly.org/ 9CFE12B0 791A4267 887F520C B7AC2E51 BD41714B



From claird at lairds.com  Tue Apr  1 11:41:08 2003
From: claird at lairds.com (Cameron Laird)
Date: Tue, 01 Apr 2003 16:41:08 -0000
Subject: else: w/o except: - why not?
References: <mailman.1049211247.3694.python-list@python.org>
Message-ID: <v8jg94pkcdipd1@corp.supernews.com>

In article <mailman.1049211247.3694.python-list at python.org>,
Michael Chermside  <mcherm at mcherm.com> wrote:
			.
			.
			.
>I can't speak for Tim, but my OWN opinion is precisely in alignment
>with what you have written.
>
>I think that a "bare except" is useful in only two ways. The first
>is like a "finally" clause that's skipped if everything works. In
>other words, the following bits of code is (almost[1]) equivalent to 
>your second example:
>
>    everythingWorked = False
>    try:
>        f1()
>        everythingWorked = True
>    finally:
>        if not everythingWorked:
>            f2()
>
>This doesn't come up much, but it's occasionally useful, and your
>version (using bare except with a raise) is a little prettier.
>
>The other situation is running "untrusted plug-in" code of some 
>sort. Something like this:
>
>    try:
>        randomCodeSomebodyUploaded()
>    except:
>        logProblem()
>
>But really, that hardly EVER comes up. I wish that guido had made
>it harder to spell "bare except" because these situations seem so
>unusual. The next best solution is to just avoid it unless you're
>SURE you need it.
			.
			.
			.
Thanks, Michael.

I'm rather obsessive on exceptions; I think programmers 
generally misunderstand them and neglect their importance.
Very, VERY few programming texts do them justice; one ele-
ment of my enthusiasm for Alex's books <URL: http://
www.unixreview.com/documents/s=7822/ur0303j/ > is that he
speaks like an adult about them.  Unless you've made a
disciplined survey of the extant literature, you won't
realize how rare this is.

I'm working on a couple of articles of my own to shift the
balance a bit.

I disagree with you on the frequency of "running 'untrusted
plug-in' code", or perhaps I misunderstand what you mean.
Briefly, I mistrust a LOT of code.  Even when I'm using
well-known, mature extensions, my aim is to have at least
top-level error-handlers that catch *everything*.  Are we
on the same subject?
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From sde at recombinant.demon.co.uk  Wed Apr 30 14:01:00 2003
From: sde at recombinant.demon.co.uk (Stephen D Evans)
Date: Wed, 30 Apr 2003 19:01:00 +0100
Subject: Python 2.3b1 and _tkinter.Tcl_Obj
Message-ID: <b8p30s$be4$1$8302bc10@news.demon.co.uk>

Is this Python or Tk/Tcl? Is it a bug or a feature?

There are some inconsistencies in the types returned by some Tkinter widgets
when using cget() or configure() e.g.
    widget.cget('background')
Menu widgets return a Tcl_Obj (typename='border'), Buttons return a string.

The following code demonstrates using configure() rather than cget()

# Python 2.3b1 (#40, Apr 25 2003, 19:06:24) [MSC v.1200 32 bit (Intel)] on
win32
#
import Tkinter
import _tkinter
import types

# workaround
#Tkinter.wantobjects = 0

for name in Tkinter.__dict__:
    c = getattr(Tkinter, name)
    if isinstance(c, types.ClassType) and Tkinter.Widget in c.__bases__:
        widget = c()
        tup = widget.configure('background')[3:] # last two
        print c, type(tup[0]), type(tup[1])

        assert 2 == len(tup)
        for i in tup:
            assert isinstance(i, (types.StringType, _tkinter.Tcl_Obj))
            assert not isinstance(i, _tkinter.Tcl_Obj) or i.typename ==
'border'
#-----------------------



I found this while fiddling with the Pmw megawidget (http://pmw.sf.net)
tests and demos
under Python 2.3b1

Stephen D Evans





From tim_one at email.msn.com  Sun Apr 13 14:54:27 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Sun, 13 Apr 2003 14:54:27 -0400
Subject: intern'ed strings and deepcopy()
In-Reply-To: <b7c85g$slh$03$1@news.t-online.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCCEPEEGAB.tim_one@email.msn.com>

[Martin v. L?wis]
> ...
> To make my example more correct, I should write
>
> type(a) == type(b) == str and b == a 'implies'
>    id(intern(b+" s")) == id(intern(a+" s"))
>
> so that the strings being interned are temporaries themselves. For some
> reason, this still doesn't trigger the mortality of the interned string
> in 2.3a2, but I haven't investigated why.

obmalloc reuses the freed string space immediately (by design; it knows that
the most recently freed memory is most likely to still be in cache).

To prevent that, you need to trick it into giving the freed memory to
something else that persists.  This should print False in 2.3a2:

def save(s):
    global gunk
    gunk = s
    return 0

a = b = "ab"
print id(intern(a + " s")) == (
    save(a + "xy") or id(intern(b + " s")))

Less obscurely,

a = b = "ab"
print id(intern(a + " s"))
whatever = "cd" + " ef"
print id(intern(b + " s"))

prints two different numbers.

> In any case, there is no *guarantee* (anymore) that the id of an
> interned string stays the same over time; one should store the
> interned string itself, and not its id.

Yup.





From aleax at aleax.it  Tue Apr 15 04:11:13 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 15 Apr 2003 08:11:13 GMT
Subject: Pythonic way of web-programming
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
Message-ID: <B8Pma.20554$T34.656139@news2.tin.it>

Andy Robinson wrote:
   ...
> First, let's note the half-a-dozen different templating schemes
> available.  DTML, Woven, Quixote, YAPTU, preppy.

YAPTU's a toy -- and, as its author, I hope I can assert that with
some credibility!-)  Cheetah, OTOH, is a serious contender for the
"best of breed" Python templating utility (at least when templates
need to be maintained by non-programmers).  I don't know if it's
entirely correct to describe Quixote as a templating tool.  Rather,
I _would_ count such packages as Spyce...

None of this invalidates your main point, of course (indeed, on
the contrary, it might reinforce it).


> be wonderful if this could be systematized in something like the 'db
> api', with some interfaces defined in the standard Python library.

The DB API isn't in the standard library (and until some substantial
amount of reusable CODE is factored out, rather than just somewhat
vague interface specs, I doubt it will be... or _should_ be).  If a
consensus could be reached among a sufficient number of the maintainers
of existing packages, then, sure, some interfaces could be standardized.

However -- look at the XML situation for contrast.  There IS just such
a standard (and parts of it ARE in the standard libraries, while others,
PyXML stuff, integrate as seamlessly as they can with those via some
import tricks).  AND YET -- you guys have developed and maintain PyRXP,
which uses substantially different interfaces -- *AND WELL IT DOES*,
because thanks to the different interfaces (trees of bare tuples, or
lists as the case may be, rather than intricate objects as in DOM) it
can save a LOT of memory, and generally zip along much, much faster.

Aren't there analogies here...?  preppy may be the simplest (and thus,
the best) templating tool for a Python programmer -- starting with a
good knowledge of Python, said programmer needs to learn very little to
start using preppy.  Yet Cheetah's very different templating language
may be best if one needs templates to be handled with 3rd party tools
designed to handle HTML pages, and more generally maintained by people
who do not consider themselves programmers and might strongly resist
"having to learn to program".  Targeting different audiences, is it
surprising that preppy and Cheetah end up with so different interfaces?


> Finally, there's a transport mechanism.  Here one could use
> CGI, Twisted, Zope, BaseHTTPServer or whatever else.

Not sure I'd qualify Zope as a "transport" mechanism!-)  Rather
the choice here is:

  -- a custom server program coded in Python (BaseHTTPServer and
     the like, Medusa, Twisted)

  -- CGI, the only standard ensuring operability with different
     webservers (maybe fast-CGI might live here too...?)

  -- webserver-specific ways to cooperate with existing webservers:
     LRWP for Xitami, mod_python or mod_webkit or PyApache for
     Apache, ASP or lower-level interfaces to MS servers, ...

Can we find enough commonality among such disparate needs to
extract out some common interfaces?  In the DB-API case (and in
the XML one) there were some existing external standards (not
necessarily well-observed, vide the SQL standard, but still maybe
better than nothing) as a reference point.  Besides CGI, and
HTTP itself, what would we build upon here?  Perhaps the Request-
Response model you mention would suffice for reasonably high
level interactions, but much of what one can do e.g. with mod_python
ISN'T "reasonably high level" (or else one would just use
PyApache for CGI-substition-level tasks...!).


> Wouldn't it be nice if one day we could have some standardized
> interfaces, so that one could switch between these without undue pain
> and reduce the confusion people must face?

Yes, it would be VERY nice, but I have no idea about how to make
it happen, or even about how it would look like if it ever did.


Alex




From tug at wilson.co.uk  Tue Apr 29 15:08:39 2003
From: tug at wilson.co.uk (John Wilson)
Date: Tue, 29 Apr 2003 20:08:39 +0100
Subject: Problem subclassing tuple
References: <mailman.1051631814.18940.python-list@python.org> <2MmdnZmgXowvMTOjXTWcpA@comcast.com>
Message-ID: <09c901c30e82$bfe7c240$640a0b0a@handel>

----- Original Message ----- 
From: "Terry Reedy" <tjreedy at udel.edu>
Newsgroups: comp.lang.python
To: <python-list at python.org>
Sent: Tuesday, April 29, 2003 5:48 PM
Subject: Re: Problem subclassing tuple


[snip]

> Because tuples are immutable, one cannot mutate them in an init
> method, even to set the initial values.  So, as you apparently
> suspected, they are initialized in the __new__ method from a sequence.
> This is the source of the exception (see below). tuple.__init__ should
> be a no op.
> 
> >>> class Holder(tuple):
> ...      def __init__(self, *values):
> ...          print "Holder init entered"
> ...
> >>> Holder(1,2)
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> TypeError: tuple() takes at most 1 argument (2 given)
> # init not entered
> >>> Holder((1,2))
> Holder init entered
> (1, 2)
> 
> While Holder.__init__() cannot modify the value of self, it *can* add
> attributes (which surprised me a bit) so it can have a use.
> 
> >>> class Holder(tuple):
> ...     def __init__(self, seq):
> ...         self.y = 1
> ...
> >>> a=Holder((1,2))
> >>> a.y
> 1
> 
> I don't know what more you could do in a __new__ method.

Terry,
    I have found that I can subclass list OK. 

This will do for what I'm after.

    Thanks to all the contributors to this thread for all the info :-))

John Wilson
The Wilson Partnership
http://www.wilson.co.uk





From jdhunter at ace.bsd.uchicago.edu  Thu Apr 10 14:34:52 2003
From: jdhunter at ace.bsd.uchicago.edu (John Hunter)
Date: Thu, 10 Apr 2003 13:34:52 -0500
Subject: pygtk install
In-Reply-To: <3E9273D6.7030500@netscape.net> (vector's message of "Tue, 08
 Apr 2003 17:01:42 +1000")
References: <3E9273D6.7030500@netscape.net>
Message-ID: <m2smsq8alv.fsf@mother.paradise.lost>

Im having problems installing pygtk on my windows XP machine.
I installed
pygtk-1.99.16.win32-py2.2.exe
from
http://www.pcpm.ucl.ac.be/~gustin/win32_ports/
and my test prog
import pygtk
pygtk.require('2.0')
import gtk
gives me this error
Traceback (most recent call last):
   File
"G:\Development\Python\Lib\site-packages\Pythonwin\pywin\framework\scriptutils.py",
line 301, in RunScript
     exec codeObject in __main__.__dict__
   File "E:\Md_prj\md_422\Code\Python\testgtk1.py", line 3, in ?
     import gtk
   File
"G:\Development\Python\Lib\site-packages\gtk-2.0\gtk\__init__.py",
line 39, in ?
     import gobject as _gobject
ImportError: DLL load failed: The specified module could not be found.

I then tried
GTK+-Runtime-Environment-2.2.1.1.exe
and reloaded python etc all the same

-- 
http://mail.python.org/mailman/listinfo/python-list





From lale at fotonation.com  Wed Apr 30 06:49:14 2003
From: lale at fotonation.com (Ilariu Raducan)
Date: Wed, 30 Apr 2003 11:49:14 +0100
Subject: win32 display jpg file
In-Reply-To: <mailman.1051681966.18285.python-list@python.org>
References: <irxra.10972$pK2.14342@news.indigo.ie> <RuDra.7018$lD4.74951@news-server.bigpond.net.au> <mailman.1051681966.18285.python-list@python.org>
Message-ID: <WSNra.11118$pK2.14587@news.indigo.ie>

Thank you all for your quick replays.

I check the last outline and I can share my findings with you

in document  OnOpenDocument add:
	im = Image.open(filename)
         self.dib = ImageWin.Dib(im.mode,im.size)
         self.dib.paste(im)
	#make in redraw
	#self.frame.InvalidateRect()
	#or
	#self.UpdateAllViews()
in View
     def OnDraw(self,dc):
         self.dib.expose(dc.GetHandleAttrib())

Regards,
Lale


Fredrik Lundh wrote:
> Mark Hammond wrote:
> 
> 
>>>Does anybody knows how to display a JPG file in pythonwin and draw lines
>>>over it?
>>
>>With a fair bit of work :)  You would need to use PIL to decode the jpg
>>and convert it to a bitmap.  IIRC, Pythonwin still has a PIL demo in it.
> 
> 
> if you cannot find it, here's an outline:
> 
> when you load the image:
> 
>     from PIL import Image, ImageWin
> 
>     try:
>         im = Image.open(filename)
>     except IOError:
>         # deal with the error
>     else:
>         # check image size; resize if necessary
>         # check the image mode
>         if im.mode not in ("1", "L", "RGB"):
>             im = im.convert(Image.getmodebase(im.mode))
>         dib = ImageWin.Dib(im)
>         # store the "dib" variable somewhere
> 
> in the PAINT handler, just do:
> 
>     dib.expose(dc)
> 
> where "dc" is a device context for the window, cast to an integer.
> according to the PIL docs, "you can use the GetHandleAttrib method
> of the CDC class to get a suitable handle."
> 
> in 1.1.4, you can pass in ImageWin.HWND(wnd) instead of the dc,
> where wnd is a window handle cast to an integer.
> 
> in all versions, you can also use
> 
>     dib.draw(dc, (x0, y0, x1, y1))
> 
> where (x0, y0, x1, y1) is the coordinates for where you want the
> image to appear.
> 
> (expose is the same thing as draw(dc, (0, 0, width, height))
> 
> </F>
> 
> 
> 
> 




From aleax at aleax.it  Mon Apr 14 17:24:39 2003
From: aleax at aleax.it (Alex Martelli)
Date: Mon, 14 Apr 2003 23:24:39 +0200
Subject: Seperating HTML content from logic ( was: Re: Inserting Python commands into HTML documents)
In-Reply-To: <200304141348.18694.scott_list@mischko.com>
References: <zpqdne0Kztcb0AejXTWcow@comcast.com> <xrtma.10655$LB6.301836@news1.tin.it> <200304141348.18694.scott_list@mischko.com>
Message-ID: <200304142324.39351.aleax@aleax.it>

On Monday 14 April 2003 10:48 pm, Scott Chapman wrote:
> On Monday 14 April 2003 00:29, Alex Martelli wrote:
> > Sure, there are many ways in which you can choose to use this
> > debatable "let's mix logic right inside presentation so we can't
> > easily separate them any more" approach (perhaps excusable when
> > one's in a deuced hurry, if one's used to it and thus can deploy
> > it fastest).
>
> I've been dealing with web pages for a while, using the "embed the
> scripting language into the HTML" model.  I'd like to find some good docs
> on how to implement the "lets seperate the logic and presentation" model.

I find Steve Holden's book "Python Web Programming" (New Riders)
pretty good at that, at an instructional level.  But I think there's a
growing body of literature on how to do that with XML for the "logic"
and stylesheets (including XSLT) for the presentation (I admit I'm not
really up to date on that, though).


> I'd also like to know which Python toolkits/frameworks that are aimed at
> implementing this model.
>
> Any pointers?

I would suggest "Webware for Python", and Cheetah for templating (but,
there are good templating alternatives -- Cheetah seems to stand out
as particularly suitable for non-programmers maintaining the presentation
layer[s] while programmers maintain the logic of the scripts behind it).


Alex





From mike at mmrd.com  Wed Apr 30 10:56:40 2003
From: mike at mmrd.com (MAK)
Date: 30 Apr 2003 07:56:40 -0700
Subject: running a PyUnit TestSuite
References: <b8nrlv$dvn$07$1@news.t-online.com>
Message-ID: <ec5a166f.0304300656.1dd30437@posting.google.com>

"Achim Domma" <achim.domma at syynx.de> wrote in message news:<b8nrlv$dvn$07$1 at news.t-online.com>...

> I have a TestSuite which depends on the execution order of the tests. If I
> use the 'main' object provided by unittest.py I can not specify the
> execution order

Yep, ran into the same thing right away.  Since 'main' runs tests in
the alphabetical-sort order of the function name, I solved the problem
by naming all of my test functions like this...

test_01_sometest
test_02_anothertest
...

...and now they run in the order I want.



From aleax at aleax.it  Sat Apr 12 18:17:56 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 22:17:56 GMT
Subject: exceptions == errors?
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com> <roy-AB9C56.20201307042003@reader2.panix.com>
Message-ID: <og0ma.8129$T34.232220@news2.tin.it>

Roy Smith wrote:
   ...
> And, of course, "catch" is spelled "except" in Python :-)

One of the very, VERY few places where I'd prefer the C++ choice --
my fingers and brain keep misspelling 'except' as 'expect' (well
it IS an exception I DO expect, that's why I'm coding a handler...!-)


Alex




From lusifer__ at hotmail.com  Tue Apr 29 09:26:12 2003
From: lusifer__ at hotmail.com (Dagur)
Date: 29 Apr 2003 06:26:12 -0700
Subject: VPython: Using scene.exit to catch Windows' close button
Message-ID: <2e39ca4b.0304290526.c24f243@posting.google.com>

Hello group,

I'm developing an interactive application using the latest VPython
under Python 2.2.2/Windows XP.

I'm having problems using VPythons scene.exit to detect whether the
VPython window's Exit button (the little 'X') has been pressed. The
VPython documentation tells me to set scene.exit = 0, which is
supposed to disable automatically closing the window. I have done
this, however I can't seem to change Vpython's behavior. It still
closes the window immediately, instead of letting my script do its
clean-up routine.

I'd be happy to know if VPython's documentation doesn't match its
behavior, or whether I'm missing something obvious.

Thanks in advance,

Dagur Dagsson



From bh at intevation.de  Wed Apr 16 12:52:44 2003
From: bh at intevation.de (Bernhard Herzog)
Date: 16 Apr 2003 18:52:44 +0200
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <yfsr883b96o.fsf@black132.ex.ac.uk>
Message-ID: <6qd6jmgzab.fsf@salmakis.intevation.de>

Alexander Schmolck <a.schmolck at gmx.net> writes:

[description how dynamic scoping could be put into the python syntax]
> Maybe this is stupid and maybe it doesn't work and no I haven't
> thought deeply about packages and classes so far (which as Beni
> already mentioned seem to render a literal mapping of CL's approach
> unattractive)

Python is often quite explicit about which namespaces a variable come
from (e.g. instance variables require an explicit self; the global
statement). The same approach could be taken with dynamic scopes which
could require an explicit dynamic scope and that can even be implemented
in python itself:

import inspect

class Scope:

    def __init__(self, base, defaults):
        if base is not None:
            self.__dict__.update(base.__dict__)
        self.__defaults = defaults

    def __getattr__(self, attr):
        if self.__defaults is not None:
            try:
                return self.__defaults[attr]
            except KeyError:
                pass

        raise AttributeError, attr

# Name for local variables holding dynamic scopes
scope_var_name = "scope"

def dynscope():
    """Return a new dynamic scope for use in the caller's frame"""
    caller = inspect.currentframe().f_back
    frame = caller.f_back
    while frame:
        base = frame.f_locals.get(scope_var_name)
        if isinstance(base, Scope):
            break
        frame = frame.f_back
    else:
        base = None

    scope = Scope(base, caller.f_globals)
    return scope


# Example usage

# default value for dynamically scoped variable
variable = "default"

def print_variable():
    """Print the value of the dynamically scoped variable 'variable'"""
    scope = dynscope()
    print "dynamically scoped variable:", scope.variable


def intermediate():
    print_variable()

def shadow_variable():
    # define a dynamic scope with a value for 'variable'
    scope = dynscope()
    scope.variable = "shadowed"

    #Call use_variable indirectly
    intermediate()


if __name__ == "__main__":
    # Calling print_variable directly should print "default" as value
    print_variable()
    # Calling print_variable through shadow_variable should print
    # "shadowed" as value
    shadow_variable()



   Bernhard

-- 
Intevation GmbH                                 http://intevation.de/
Sketch                                 http://sketch.sourceforge.net/
MapIt!                                           http://www.mapit.de/



From upclick at upclick.com.br  Mon Apr  7 17:34:43 2003
From: upclick at upclick.com.br (UP Click)
Date: Mon, 7 Apr 2003 18:34:43 -0300
Subject: Contato - Seu site no Aonde e Achei
Message-ID: <41153-22003417213443370@vaio>

	
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030407/beaed746/attachment.html>

From list-python at ccraig.org  Wed Apr 16 11:14:23 2003
From: list-python at ccraig.org (Christopher A. Craig)
Date: 16 Apr 2003 11:14:23 -0400
Subject: Big-O notation (was: Re: Python and Schools)
In-Reply-To: <3E9CBAFE.8016BFB2@alcyone.com>
References: <qGqla.61$Tj7.91655@news.uswest.net>
	<pan.2003.04.11.09.17.46.771369.890@thomas-guettler.de>
	<Xns935A6D886AC5cpl19ghumspamgourmet@62.153.159.134>
	<v9e21mhemm8c2@corp.supernews.com> <b7grsc$63p$1@panix1.panix.com>
	<PLYma.5565$2x2.2919311@dca1-nnrp1.news.algx.net>
	<3E9C6C62.8010808@rogers.com>
	<mailman.1050441947.6447.python-list@python.org>
	<3E9CBAFE.8016BFB2@alcyone.com>
Message-ID: <t1wn0iqtqy8.fsf@kermit.wreck.org>

Erik Max Francis <max at alcyone.com> writes:

> It also means that one doesn't bother distinction between logarithms
> base 10, e, 2, etc.; O(ln n) ~ O(lg n) ~ O(lb n).  Similarly, O(2^n) ~
> O(10^n) ~ O(e^n) too.

The second part of this isn't true.  If a process f(n) has a running
time T(N) for in input of size N, said process is O(N) if there exist
constants "k" and "j" such that O(N)*k <= T(N) for all N>=j

While there does exist a constant (log(A)/log(B)) for which
log_A(N)*k <= log_B(N) for all sufficiently large N, there is
no constant for which (A**N)*k <= B**N for all sufficiently large N
when B>A.

-- 
Christopher A. Craig <list-python at ccraig.org>
"Never let schooling get in the way of your education"  Mark Twain




From bokr at oz.net  Thu Apr 17 12:54:06 2003
From: bokr at oz.net (Bengt Richter)
Date: 17 Apr 2003 16:54:06 GMT
Subject: check to see if a symbol in a string has a binding
References: <kfed6jlv06e.fsf@alumni.cse.ucsc.edu>
Message-ID: <b7mm7e$uch$0@216.39.172.122>

On 17 Apr 2003 04:22:01 -0700, Jinsong Zhao <zhaojs at yahoo.com> wrote:

>Sorry if the question is too simple. I want to see if the name inside
>the string has a binding and callable. For example,
>
>1. func_string="func"
>2. func_string="func_string"
>3. func_string="dir"
>
>So the in case 1, "func" does not have a binding; case 2 "func_string"
>has a binding but not callable; in case 3 "dir" has a binding and is
>callable. How do I proceed based on the content in func_string? This
>is rather baffling.
>
>Any suggestion? Thanks in advance.
>
Insufficient spec. "Has a binding" in what name space?
Here's a possibility, passing the usual suspect namespaces
(not tested beyond what you see below):

 >>> def isCallable(s, *namespaces):
 ...     if not namespaces: raise ValueError, 'need at least one name space arg'
 ...     for ns in namespaces:
 ...         v = ns.get(s)
 ...         if v: return callable(v)
 ...     return 0
 ...
 >>> for fs in 'func func_string dir'.split():
 ...     print '%s: %s' % (fs, ('no', 'yes')[isCallable(fs)])
 ...
 Traceback (most recent call last):
   File "<stdin>", line 2, in ?
   File "<stdin>", line 2, in isCallable
 ValueError: need at least one name space arg

Oops, well that part got tested ;-)

 >>> for fs in 'func func_string dir'.split():
 ...     print '%s: %s' % (fs, ('no', 'yes')[isCallable(fs,vars(),
 ...                                       globals(), __builtins__.__dict__)])
 ...
 func: no
 func_string: no
 dir: yes

Regards,
Bengt Richter



From jbauer at rubic.com  Sun Apr 27 10:28:25 2003
From: jbauer at rubic.com (Jeff Bauer)
Date: Sun, 27 Apr 2003 09:28:25 -0500
Subject: Python 2.3b1 + Tkinter on Solaris 2.6
Message-ID: <3EABE909.75D71230@rubic.com>

> You are overlooking something, but it is not obvious, and it is likely
> not a bug in Python.
> 
>> *** WARNING: renaming "_tkinter" since importing it failed:
>>   ld.so.1: ./python: fatal: relocation error: file
>>   /usr/local/lib/libtk8.4.so: symbol __eprintf: referenced symbol not
>> found
> 
> __eprintf is traditionally a symbol from libgcc.a, and gets into
> object files whenever the source code has assert().
> 
> It appears that your Tk installation was made with a gcc version that
> had assert implemented with __eprintf. Could it be that you have
> changed gcc versions since?  If so, you need to recompile Tk (or,
> alternatively, link python and/or _tkinter.so with the gcc version
> that was used to link Tk)

Martin,

Thanks for your response.  I removed every occurrence of
the tcl/tk libraries from my system, then rebuilt tcl/tk
from source.  I then rebuilt Python2.3b1 from scratch, but
the same error occurs.  :-(

Jeff Bauer
Rubicon Research




From luc at nospam.com  Fri Apr 11 17:38:40 2003
From: luc at nospam.com (luc wastiaux)
Date: Fri, 11 Apr 2003 21:38:40 +0000 (UTC)
Subject: readline clear buffer
Message-ID: <slrnb9edf0.tds.luc@grizzly.dont-panic.info>

Hello, 
when using readline, I can retrieve the contents of the line buffer using 
get_line_buffer()

but is there a way to clear this buffer ? none of the provided functions 
seem to allow that: http://www.python.org/doc/lib/module-readline.html

what I would like to do is TAB-complete a command containing more than one 
word, ie:

/msg user1 <TAB> 
inserts /msg user2

on the command line, and this cyclically given a certain list of users. 
Does anyone have an idea of how to do that?

thanks in advance for any suggestions.

-- 
luc wastiaux
 $ finger luc at info.4002.org



From do-not-spam-ben.hutchings at businesswebsoftware.com  Wed Apr  9 07:14:42 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 9 Apr 2003 11:14:42 GMT
Subject: socket question
References: <mailman.1049735831.15656.python-list@python.org>
Message-ID: <slrnb98052.t8.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <mailman.1049735831.15656.python-list at python.org>, Rick Y wrote:
> how can i enable _sockt module in my solaris python?. 
> i did not build it. Downloaded it from sunfreeware. 
> 
> ./viewcvs-install
> Traceback (most recent call last):
>   File "./viewcvs-install", line 35, in ?
>     import compat
>   File "./lib/compat.py", line 20, in ?
>     import urllib
>   File "/usr/local/lib/python2.1/urllib.py", line 26,
> in ?
>     import socket
>   File "/usr/local/lib/python2.1/socket.py", line 41,
> in ?
>     from _socket import *
> ImportError: ld.so.1: python: fatal: libssl.so.0.9.6:
> open failed: No such file or directory

It looks like you need to get openssl-0.9.6 from there as well.



From nikolai.NOJUNK at micon.no  Sat Apr  5 17:00:11 2003
From: nikolai.NOJUNK at micon.no (Nikolai Kirsebom)
Date: Sun, 06 Apr 2003 00:00:11 +0200
Subject: Invalid format character in string
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com> <mailman.1049493668.26508.python-list@python.org> <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com> <mailman.1049496249.4466.python-list@python.org>
Message-ID: <raku8v8a2sj4rkpvnu1mjmmnjdsv1gq9f8@4ax.com>

On Fri, 4 Apr 2003 16:43:18 -0600, Skip Montanaro <skip at pobox.com>
wrote:

>Okay, how about a preprocessing step like this:
>
>    data = open("htmlfile").read()
>    # double up any % not followed by a left paren
>    data = re.sub(r"%([^(])", r"%%\1", data)
>    sys.stdout.write(data % valdict)
>
>?  It's not perfect, but should be a lot better than the pain you're
>experiencing now.
>
>Skip

Thanks for all the answers.

I think I'll go for the Skip Montanro's suggestion
(double up any % not followed by a left parenthesis). 

Nikolai




From dave at pythonapocrypha.com  Fri Apr 11 18:47:10 2003
From: dave at pythonapocrypha.com (Dave Brueck)
Date: Fri, 11 Apr 2003 15:47:10 -0700 (PDT)
Subject: ASN.1
In-Reply-To: <A128D751272CD411BC9200508BC2194D018853E1@escpl.tcp.com.ar>
References: <A128D751272CD411BC9200508BC2194D018853E1@escpl.tcp.com.ar>
Message-ID: <Pine.LNX.4.51.0304111546280.6997@ns0.tbye.com>

On Fri, 11 Apr 2003, Batista, Facundo wrote:

> I'm searching a ASN.1 parser, to load the structure of a file to memory and
> then be able to parse the file.
>
> Do you know something like this for python?

pySNMP has one. Also check on the Vaults of Parnassus.

-Dave




From nhodgson at bigpond.net.au  Mon Apr  7 08:48:56 2003
From: nhodgson at bigpond.net.au (Neil Hodgson)
Date: Mon, 07 Apr 2003 12:48:56 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <7xn0j3fsdu.fsf@ruckus.brouhaha.com>
Message-ID: <Yseka.1370$1h6.8783@news-server.bigpond.net.au>

Paul Rubin:

> I think you're going to need some C extensions to represent buffers
> similar to how Emacs does it.  It's hard to move blocks of characters
> around in Python with the array module without constantly re-allocating
> the arrays.

   The only addition to array that is really needed to provide all the
building blocks used in Scintilla is a method for moving blocks around in an
array. Others that would be nice to have but are not as important are fast
ways to add space to the end of an array and to copy between arrays.
However, the real time sink in a GUI editor is most commonly in measuring
and drawing the text, not in the speed of the underlying text storage data
structures.

   Here is a copy_internal method that can be added to the array module that
takes (dest, source,len) parameters. When I have time I'll package this up
for inclusion in Python.

static PyObject *
array_copy_internal(arrayobject *self, PyObject *args)
{
 int dest = -1;
 int source = -1;
 int len = -1;
 if (!PyArg_ParseTuple(args, "iii", &dest, &source, &len))
  return NULL;

 if ((dest != source) && (len > 0)) {
  if ((dest < 0) || (source < 0)) {
   PyErr_SetString(PyExc_RuntimeError,
       "can't use negative index in copy_internal");
   return NULL;
  } else if (((dest + len) >= self->ob_size) ||
   ((source + len) >= self->ob_size)) {
   PyErr_SetString(PyExc_RuntimeError,
       "can't index past the end of an array");
   return NULL;
  } else {
   int itemsize = self->ob_descr->itemsize;
   memmove(self->ob_item + dest*itemsize,
    self->ob_item + source*itemsize,
    len*itemsize);
  }
 }

 Py_INCREF(Py_None);
 return Py_None;
}

PyDoc_STRVAR(copy_internal_doc,
"copy_internal(dest,source,len)\n\
\n\
Copy a range within an array to another place within that array.");

// And in the methods table array_methods goes:

 {"copy_internal", (PyCFunction)array_copy_internal, METH_VARARGS,
  copy_internal_doc},

   Neil





From cben at techunix.technion.ac.il  Wed Apr  9 12:23:26 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Wed, 9 Apr 2003 19:23:26 +0300 (IDT)
Subject: How to split string in equally long parts
In-Reply-To: <mailman.1049836334.25959.python-list@python.org>
References: <b6vd0e$i1b$01$1@news.t-online.com> <mailman.1049836334.25959.python-list@python.org>
Message-ID: <Pine.GSO.4.51.0304091921530.1222@techunix.technion.ac.il>

Mike Rovner wrote on 2003-04-08:

> "Heiko Henkelmann" <heiko at hhenkelmann.de> wrote in message
> news:b6vd0e$i1b$01$1 at news.t-online.com...
> > I need to split a string into an array of equally long substrings, like
> > e.g.:
> >
> > '1234567890'   -> ['12', '34','56','78','90']
> >
> > What is the most efficient way to do this?
> >
> > Heiko
>
> Task description is too vague.
>
Indeed, what exactly the OP wants when it doesn't divide precisely?
E.g. when the string is 257 characters long?

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>




From peter at engcorp.com  Wed Apr 16 20:21:11 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 20:21:11 -0400
Subject: Using __repr__ or __str__ for own printable class?
References: <mailman.1050525028.7854.python-list@python.org> <b7km3s$238m$1@nntp6.u.washington.edu> <mailman.1050537782.19056.python-list@python.org>
Message-ID: <3E9DF377.E94BFCB8@engcorp.com>

Steven Taschuk wrote:
> 
> Quoth Donn Cave:
>   [...]
> > It's clear that the docs promote this idea, but they also hint at
> > its half-baked quality - repr() evals to the original value, unless
> > it doesn't.  I don't think the idea is clear at all.
> > [...] The only kind of clarity this documentation possesses is
> > when it lets you read into it what you already thought.  Not the fault
> > of the documentation, though, but of the unsound ideas behind it.
> 
> I'm of the opinion that the ideas are sound; if there is a
> problem, I think it must lie in the documentation.

There may also be some disagreement on appropriate uses, which could
be the cause of the documentation problem, or the result of it (or
neither :-).

> But information-rich and unambiguous representations, however
> useful for (e.g.) debugging, are not always easy to read or
> understand; for the humans then, we'd like an "informal",
> easier-to-read representation, whence __str__.

I've always considered __str__ to be for use in the application,
such as when you have a class which is sometimes asked for a 
string representation for display purposes to a real user, or in
an application log, or something.

I've always considered __repr__ to be for use in *debugging* the
application, where a *debugging* log needs to display information,
or in a traceback.  Since it's next to impossible for most interesting
objects to be reproduced easily from the output of __repr__ if it's
also desirable that the output be meaningful and useful to a human 
reader, I don't buy into the 'eval(repr(x)) == x' theory very much,
other than to note that it's the most obvious thing to do for 
primitive data types.

I've never had any confusion in my own mind about when it's 
appropriate to use one or the other.  Perhaps this perspective 
should be used in the documentation, rather than the focus on 
reproducing objects from their __repr__.  

-and-not-just-because-it's-my-own-view-ly y'rs, ;-)
 Peter



From nirina at mail.blueline.mg  Tue Apr  8 17:49:15 2003
From: nirina at mail.blueline.mg (Raseliarison nirinA)
Date: Wed, 9 Apr 2003 01:49:15 +0400
Subject: Importing problem
References: <2737271.1049712504@dbforums.com>
Message-ID: <019601c2fe2c$ca5e77c0$f7a2383e@raseliar>

"nitintz" wrote:


> 
> hi,
> 
> I am new to the python.
> Actually i am using PyQt .. and my problem is i wish to import a Qt
> project in my PyQt program.
> Can anyone help me in it?

myself i don't use PyQt but i can help you if you give more details.
Have you tried:

from Qt import *

> 
> Thanks in Advance
> Nitin
> 

nirinA







From belred1 at yahoo.com  Sun Apr  6 11:39:35 2003
From: belred1 at yahoo.com (Bryan)
Date: Sun, 06 Apr 2003 15:39:35 GMT
Subject: What are the differences between SWIG and Boost.Python?
References: <169a7283.0304040821.9ad0976@posting.google.com> <mailman.1049495109.20211.python-list@python.org>
Message-ID: <XSXja.346441$3D1.186570@sccrnsc01>

mike,

you mentioned small c libs you use pyrex.  what about small but complex data
types?  we have a structures that contain pointers to structures that
contain pointers to structures. etc.  some structures contain methods, some
data, some with many constants.  imagine a simplified c++ written in c.
everytime we want to write an extension, we always seem to just write it in
c instead of pyrex so we can simply bring in the headers as is.  in pyrex,
don't you have to redefine all the structures?  if pyrex had something like

extern "pyrex"
   #include "my_complicated_c_structures.h"

it would be great.  i haven't dived into pyrex all that much and i know it's
an incredible project, but redefining this header stuff just keeps me away
from it.

bryan

"Mike Rovner" <mike at bindkey.com> wrote in message
news:mailman.1049495109.20211.python-list at python.org...
> FWIW, I decided for myself:
>
> - wrap small C libs with Pyrex,
> - wrap big C libs with SWIG,
> - wrap small C++ libs with boost.python,
> - wrap big C++ with (I still have to do that ;)
>
> Mike
>
> PS.
> For me small is tens, big is hundreds and more
>
> "sdieselil" <sdieselil at yahoo.com> wrote in message
> news:169a7283.0304040821.9ad0976 at posting.google.com...
> > Hi
> >
> > I want to write C extension for Python. I know that there are 2 most
> > popular libraris for this purpose: SWIG and Boost.Python. What are the
> > differences between them? Which one is better for little extensions
> > and which is better for large-scale extensions?
> > --
> > http://mail.python.org/mailman/listinfo/python-list
> >
>
>
>
>





From lbates at swamisoft.com  Fri Apr 25 11:37:28 2003
From: lbates at swamisoft.com (Larry Bates)
Date: Fri, 25 Apr 2003 10:37:28 -0500
Subject: wxTree control in Boa
References: <YRp6a.29664$Ua.2281949@news2.west.cox.net>
Message-ID: <Vh2cnbFnefSnyzSjXTWcqw@comcast.com>

You might want to take a look at dynawin instead.
There is a tree class and quite a few examples.

-Larry

"Hoang" <tr at jotsite.com> wrote in message
news:YRp6a.29664$Ua.2281949 at news2.west.cox.net...
> does anyone have a most simplistic example of a treeview control in the
Boa
> framework that includes displaying of icons, the + or - sign during
> expansion and collapse?  I've gotten the control to work but the visual
> aspects aren't there.  The Boa Explorer has that feature and I was looking
> into the code to see that.  However, I am still a novice at Python and
there
> is just too much code in Explorer.py and ExplorerNodes.py.  I'm sure
someone
> must have fumbled through this while trying out Boa.
>
> thanks
>
>





From andrew.markebo at telia.com  Thu Apr  3 05:25:36 2003
From: andrew.markebo at telia.com (Andrew Markebo)
Date: Thu, 03 Apr 2003 10:25:36 GMT
Subject: nntplib: how to access NNTP response codes on exceptions?
References: <cirn8vsv4bhtp82n1tfil3e2kkounpgvav@4ax.com>
Message-ID: <m3brzn27z3.fsf@telia.com>

| I get the exception nntplib.NNTPTemporaryError as documented in
| http://www.python.org/doc/current/lib/module-nntplib.html .  But
| NNTPTemporaryError is raised for any response code in the range
| 400-499.  Is there a way of determining the particular response code?
| Ordinarily, it would be the first three characters in the first
| element ("resp", in my example) returned from the nntplib method
| (whether head, stat, etc.).  But because of the exception, the
| assignment to the tuple never occurs, so I can't check it.

If you catch it through something like:

 except NNTPTemporaryError, e:

The response is put in e.response, so you can find the code in the
first three characters.. 

You might have missed that parameters can be thrown with the
exceptions, and thats what e is.. an instansiation of a
NNTPTemporaryError class.

                   /Andy

-- 
 The eye of the linker rests on the libs!



From peter at engcorp.com  Tue Apr  8 05:22:28 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 08 Apr 2003 05:22:28 -0400
Subject: passing no arguments to a callback function question
References: <b6tqig$93k$1@lust.ihug.co.nz>
Message-ID: <3E9294D4.A9F95112@engcorp.com>

matthew wrote:
> 
> I'm pretty new to python and most of my stumblings get me somewhere but
> I'm stumped on this. Given the following code/exception how do I avoid
> the exception generated when notify_listeners is called?
> 
>      def print_params():
>         print 'PARAMS: ', params
> 
> gives:-
> 
>    File "D:\Python22\Lib\site-packages\TeePee\notifier.py", line 39, in
> __call__
>      return self.fn(*(self.args + args), **d)
> TypeError: print_params() takes no arguments (1 given)

You *always* need to include the self argument in methods.

Change that to "def print_params(self):" and it will work...  actually,
you may need to use "self.params" in the code too.  Python is *explicit*
about instance attributes, unlike some other languages.

-Peter



From warlock at eskimo.com  Tue Apr  8 05:31:42 2003
From: warlock at eskimo.com (Jim Richardson)
Date: Tue, 8 Apr 2003 02:31:42 -0700
Subject: List of Numbers
References: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>
Message-ID: <ugham-t2f.ln1@grendel.myth>

On Sat, 05 Apr 2003 20:13:45 +0100,
 Simon Faulkner <news at titanic.co.uk> wrote:
> I have a list of about 5000 numbers in a text file - up to 14 digits
> each.
> 
> I need to check for duplicates.
> 
> What would people suggest as a good method?
> 
> Simon

I'd use sort|uniq, but I don't know if that's available for MS type osen. 

In python, just stuff them all in a dictionary, any repeats, will be
eliminated. But this is rather crude and probably slow. But it would
work.

-- 
Jim Richardson         http://www.eskimo.com/~warlock

Linux, because eventually, you grow up enough to be trusted with a fork()



From mwm at mired.org  Tue Apr  1 12:37:38 2003
From: mwm at mired.org (Mike Meyer)
Date: Tue, 01 Apr 2003 17:37:38 GMT
Subject: Questions on linked lists
References: <kl8ia.27315$OV.127379@rwcrnsc54> <5vonl-vm3.ln1@beastie.ix.netcom.com>
Message-ID: <x73cl2jezh.fsf@guru.mired.org>

Dennis Lee Bieber <wlfraed at ix.netcom.com> writes:
>  (The instructor for my data structures 
> class once made the mistake of allowing us to use "any language he 
> could understand" -- which, of those on the college computer, left out 
> SNOBOL and APL -- he received assignments in BASIC, FORTRAN, Assembly, 
> and COBOL, all using statically allocated arrays.)

Geez, this is odd memory day. I once had an instructor say we could
write an assembler in "any structured language". Most people chose C,
Pascal or Algol. I wrote a SNOBOL preprocessor - in SNOBOL, of course
- to let me write the assembler in structured SNOBOL. I was one of the
first ones to finish the assignment.

Today, I could do it faster in Python - but that was many miles ago.

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From tweedgeezer at hotmail.com  Sat Apr 12 05:27:15 2003
From: tweedgeezer at hotmail.com (Jeremy Fincher)
Date: 12 Apr 2003 02:27:15 -0700
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com> <mailman.1049836294.24621.python-list@python.org>
Message-ID: <698f09f8.0304120127.52417f53@posting.google.com>

Francois Pinard <pinard at iro.umontreal.ca> wrote in message news:<mailman.1049836294.24621.python-list at python.org>...
> # Python delimits a string it by single quotes preferably, unless
> # single quotes appear within the string while double quotes do
> # not, in which case it uses double quotes for string delimiters.
> # Checking the string contents, the C code stops at the first NUL.
> # We prefix the string with a single quote and a NUL, this forces
> # double quotes as delimiters for the whole prefixed string.  Then,
> # we get rid of the representation of the single quote and the NUL.
> 
> write('"' + repr("'\0" + value)[6:])

Ah, perfect!  Thanks :)

Jeremy



From achrist at easystreet.com  Fri Apr 25 14:24:33 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Fri, 25 Apr 2003 11:24:33 -0700
Subject: How to write RTF from Python?
Message-ID: <3EA97D61.3CC9FE74@easystreet.com>

Any suggestions of the shortest path from a Python program to RTF
output?  RTF seems to be the best (reasonable combination of 
features and compatability) common denominator for exporting
data so that (windows) users can bring it into their own word
processing program, whatever that may be.  (Any problem with
that?)

Requirements are about like this:

1. Don't need an edit control, just want to write a file, but if
I have to open an edit window, I can.

2. Don't want to do too much fancy.  Text, headings, bulleted
lists,  maybe some variations in font size.

3. Development platform is WinNT, target platform is any reasonably
recent Windows machine (eg either anything newer than Win95 release 1
or Win98 or better). 

TIA,


Al



From do-not-spam-ben.hutchings at businesswebsoftware.com  Fri Apr 11 11:43:10 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 11 Apr 2003 15:43:10 GMT
Subject: popen(), sendmail: Success?
References: <3e95d63f.2635484@news.muenster.de> <slrnb9bpoj.113.news@cipw08.physik.uni-bonn.de> <v29im-ia2.ln1@nb2.stroeder.com> <slrnb9dma7.1vh.news@cipw08.physik.uni-bonn.de>
Message-ID: <slrnb9doke.180.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <slrnb9dma7.1vh.news at cipw08.physik.uni-bonn.de>,
Carsten Gaebler wrote:
> In article <v29im-ia2.ln1 at nb2.stroeder.com>, Michael Str?der wrote:
<snip>
>> stderr *is* captured by the web server and most times written to an error 
>> log. That's where those error messages belong.
> 
> Right. But if you are a really big ISP (and the original poster's ISP *is*
> big) you do not want to let your web servers write error logs.
<snip>

Big enough not to care about errors?!  I would leave that ISP as quickly
as possible.



From andrew-pythonlist at puzzling.org  Sun Apr 27 00:47:40 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Sun, 27 Apr 2003 14:47:40 +1000
Subject: a problem about running php on twisted 1.0.4 under window?
In-Reply-To: <4dbec4b4.0304262018.3ed4a8f3@posting.google.com>
References: <4dbec4b4.0304262018.3ed4a8f3@posting.google.com>
Message-ID: <20030427044739.GH26530@frobozz.local>

On Sat, Apr 26, 2003 at 09:18:51PM -0700, xlp223 wrote:
> my system is 
> window me
> python 2.2.2
> twisted 1.0.4
> php 4.3.1
> mysql
> phpbb 2.0.4
> i want to run phpbb on twisted 1.0.4 under window,

It's usually a better idea to post Twisted questions to
twisted-python at twistedmatrix.com, not all of the Twisted developers read
comp.lang.python...

That said, I'll try answering your question anyway :)

> when i installed above software on window xp home edition,it was successful.
> but when i installed above software on window me,it failed.

Are you installing exactly the same versions of things on both systems?
We've had some problems with people trying to use the CLI executable of PHP 4.3
instead of the CGI executable.  (Or something like that; I'm not a PHP
expert)

> Step of my doing is 
> 1)create dir c:\pydoc and copy c:\python22\doc to c:\pydoc 
>   create c:\pydoc\phpbb2 and copy phpbb2 to it
> 
> 2)edit hello.rpy in c:\python22\scripts and save it in this dir
> from twisted.web import static, twcgi
> 
> class PHPScript(twcgi.FilteredScript):
>     filter = 'C:\\PHP\\php'
> 
> resource = static.File("c:\\pydoc")
> resource.processors[".php"] = PHPScript
> resource.indexNames = ['index.html','index.php']
> 
> 3)run the command:
>   python mktap.py web -p 80 --resource-script=hello.rpy -i index.html,index.php
> 
> 4)run the command:
>   python twistd.py -f web.tap -r win32
> 
> After the fourth step ran,i entered into the ie browser and input
> http://localhost/   it is ok to show the html.
> but while i input http://localhost/phpbb2 ,
> it show 500 error. 

What does the logfile (twistd.log, judging from your commands) say?  It
should record the tracebacks from HTTP 500 errors in there.

-Andrew.





From gh at ghaering.de  Tue Apr 15 20:37:00 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: Wed, 16 Apr 2003 02:37:00 +0200
Subject: Any good donothing?
In-Reply-To: <slrnb9p2vf.t3.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix>
References: <slrnb9p2vf.t3.Dialtone.NOSPAM.ME@vercingetorix.vercingetorix>
Message-ID: <20030416003700.GB22395@mephisto.ghaering.test>

* Somebody who thinks pseudonyms like "Dialtone" are k3wl wrote:
> Hi all. I'm working on a kind of all-in-one cd2mp3 encoder based on lame
> and cdda2wav and pygame.
> 
> A little code snippet is the following (it is already indented once
> because it is part of a bigger class, not so bigger, named cd2mp3).
> My problem is to sync the 2 processes which are launched with the
> os.system() call but without having the main program locked.

>From what I remember other such solutions just use pipes to pipe the
output of cdda2wav into the mp3 encoder. No synchronization problems,
then.

> Is there any other way to make a do_nothing_for_some_time(n) function?

The sleep function from the time module does just this. I doubt it
solves your problems, though. I'd look into pipes and try to use them
using os.popen*. If you want to display the progress in a GUI, you might
also want to select.select or select.poll on the file descriptor in
order to keep the GUI smooth.

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/




From regnivon at netscape.net  Tue Apr 22 09:37:56 2003
From: regnivon at netscape.net (scn)
Date: 22 Apr 2003 06:37:56 -0700
Subject: coverting list back to string
References: <3cf72d52.0304211132.799d8afb@posting.google.com> <3EA44A5F.6ADA09B8@alcyone.com>
Message-ID: <3cf72d52.0304220537.c63e6f1@posting.google.com>

Erik.  This worked and thank you.  However, the lines of text are all
one continuous string instead of line-by-line.  I put an extra space
between ' '. join(remove) and was able to separate the list elements
with a space between each element.  Not sure on how to have each
string written as a separate line. My code probably needs a '\n'
somewhere and I'll play with it to find out.  Again, I've not
programmed much and appreciate your help. Scott

Erik Max Francis <max at alcyone.com> wrote in message news:<3EA44A5F.6ADA09B8 at alcyone.com>...
> scn wrote:
> 
> > my quest:  i know that output.write(remove) is not correct.  you
> > cannot use the 'write' method on a list.
> 
> Presumably you want to join the list back into a string, with something
> like:
> 
> 	' '.join(remove)
> 
> or in general
> 
> 	S.join(L)
> 
> where S is a delimiter string and L is a sequence.  Note that this may
> have an surprising effect since L.split() with no argument will split on
> _any_ block of whitespace, whereas ' '.join(L) will only join the parts
> back with a single space:
> 
> >>> L = 'this has  a   lot    of    spaces'.split()
> >>> ' '.join(L)
> 'this has a lot of spaces'
> 
> This may not bother you, but I thought I'd point it out as a clear
> difference.



From gotyoubacktoo at yahoo.com  Wed Apr 30 02:00:08 2003
From: gotyoubacktoo at yahoo.com (Psybar Phreak)
Date: Wed, 30 Apr 2003 16:00:08 +1000
Subject: looping through a file
Message-ID: <3eaf6668$0$1030$afc38c87@news.optusnet.com.au>

hi all - ive just started with python (after having done java and c) and am
having a little bit of trouble with the script.  Im using as a login script.

There's a users.dat file in the format
user1
user1password
user2
user2password
... etc

at the moment - the code i work, only checks the first and second lines (ie.
details of user1).  but it appears that python doesn't have a do... while...
until loop.


can anyone help?

thanks!!

Ive attached the appropriate part of the code i have working

-------------------------------

#!/usr/local/bin/python

import cgi, string

form = cgi.FieldStorage()

if form.has_key('usernameField') and form.has_key('passwordField'):
         users_username = string.strip(form['usernameField'].value)
         users_password = string.strip(form['passwordField'].value)

InFile = open('users.dat', 'r')

file_username = string.strip(InFile.readline())
file_password = string.strip(InFile.readline())




authorised = 0

while file_username and file_password:
       if file_username == users_username:
              if file_password == users_password: # correct
             authorised = 1
      break
              else:                               # login matched but not
password
                    file_username = InFile.readline()
             file_password = InFile.readline()
       else: # neither match
              file_username = InFile.readline()
              file_password = InFile.readline()





From peter at engcorp.com  Mon Apr 14 04:15:13 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 14 Apr 2003 04:15:13 -0400
Subject: Decoding / Encoding ASN1 BER?
References: <3e9a5fc6@mail.hmgcc.gov.uk>
Message-ID: <3E9A6E11.8B5D9191@engcorp.com>

Richard wrote:
> 
> Can anyone tell me the best way of  encoding and decoding BER ASN1 using
> Python? I'm looking for the most popular solution and if possible any code
> examples.

Which benefits of the solution being "most popular" are of value to you?

What definition of "most popular" do you intend, or are you willing to 
take any definition, such as "most downloaded" or "most widely used in
actual released software", even if it's not defined by the respondent.

What don't you like about the results provided by Google to searches
for things like "Python ASN1", including Google Groups which, one would
think, should provide you with at least one objective way to measure
"most popular".

(I don't really mean these questions facetiously.  I seem to be taking
such questions fairly literally lately, and really can't figure out
what you want from your question.)

I noticed SNMPY appears often in search results, as well as several 
previously-unknown-to-me libraries when I do these searches.  It's not 
clear whether any of these are currently supported (which might be 
something you require as part of a package being "most popular"), 
except _perhaps_ the one which appears to be in Twisted, which I 
posted about only several days ago...  (Twisted is of course very
actively developed, but whether that particular part is supported or
not, or even finished, is another story.)

(And please remember to post to the list/newsgroup once you have
an answer, so others can benefit from your research in the future.)

-Peter



From mlh at furu.idi.ntnu.no  Sat Apr 26 13:46:37 2003
From: mlh at furu.idi.ntnu.no (Magnus Lie Hetland)
Date: Sat, 26 Apr 2003 17:46:37 +0000 (UTC)
Subject: In which versions is list.sort stable?
Message-ID: <slrnbalhft.bc1.mlh@furu.idi.ntnu.no>

Has list.sort always been stable? Or, if not, in which Python versions
is it not stable?

-- 
Magnus Lie Hetland               "Nothing shocks me. I'm a scientist." 
http://hetland.org                                   -- Indiana Jones



From DavidMLewis.NO at SPAM.mac.com  Sat Apr 19 06:50:40 2003
From: DavidMLewis.NO at SPAM.mac.com (David Lewis)
Date: Sat, 19 Apr 2003 10:50:40 GMT
Subject: error in string module documentation?
References: <mailman.1050740611.30741.python-list@python.org>
Message-ID: <190420030350395930%DavidMLewis.NO@SPAM.mac.com>

In article <mailman.1050740611.30741.python-list at python.org>, Michael
Anckaert <xantor at linux.be> wrote:

> Hello all,
> In the string module documentation I read the following:
> 
> 
> lstrip(s[, chars])
>         Return a copy of the string with leading characters removed. If
>         chars is omitted or None, whitespace characters are removed. If
>         given and not None, chars must be a string; the characters in
>         the string will be stripped from the beginning of the string
>         this method is called on.
>         
> But I get the following error from my program:
> 
>   File "./game.py", line 35, in run_game
>     item = string.lstrip(cmd, "look")
> TypeError: lstrip() takes exactly 1 argument (2 given)
> 
> Is there an error in the documentation or is it my error?

It looks like something is getting shadowed. Use of the string module
is no longer recommended, since 2.2. Now, you'd not import the string
module, but say:

   item = cmd.lstrip('look')

Hope this helps.

   David



From aleax at aleax.it  Tue Apr 29 05:09:23 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 09:09:23 GMT
Subject: Function to convert fetchone() result from a list to a dictionary
References: <c31eab9b.0304282332.32799f80@posting.google.com>
Message-ID: <7jrra.18677$3M4.547894@news1.tin.it>

<posted & mailed>

Carsten Gehling wrote:

> I'm fairly new to Python, and have yet to discover alle the little
> details, that may make my programs look leaner and perhaps become
> faster...
> 
> Using the DB API to access databases, I've stumbled upon what I would
> call a shortcoming in the API (IMHO). All the fetch commands return
> the results as lists instead of dictionaries.

Yes.  I think what you REALLY want (even though you have no way to
know you want it:-) is Greg Stein's wonderful dtuple.py, see:

http://www.lyra.org/greg/python/dtuple.py

The best explanation and set of examples I know for this is in Steve
Holden's "Python Web Programming" (New Riders) -- I highly recommend
this to Python newbies (even ones not interested in the web per se:
it's got lot of interesting coverage of database issues, too!), and
if you can't purchase it I think you can at least get its examples
free off the web (start looking at Steve's site, www.holdenweb.com).

When you wrap the result of fetchone into a dtuple.DatabaseTuple,
you can then access it in many ways, including dict-like, and the
performance is pretty good too.

Anyway, back to your specific question:


> So I've made my own conversion function, using the cursor.description
> attribute. It's pretty straight-forward, so I was wondering, if
> there's a smarter way to do this?
> 
> The function is like this:
> -------------------------------
> def cursor_fetchone_dict(cursor):
> result = cursor.fetchone()
> if result == None: return None
> result_dict = {}
> for i in range(len(result)):
> result_dict[cursor.description[i][0]] = result[i]
> return result_dict

Pretty good (except that you're using tabs, which don't show up
when I reply with KDE 3's KNode:-)  You can do a bit better (for
some definition of that term:-) by calling the dict type directly
with a list of key-value tuple -- you can get that from zipping
up a list of keys and a list of values.  The list of values you
already have (it's "result"); the list of keys you can get from
the cursor.description with a list comprehension.

So here's a step-by-step version (using SPACES, not TABS:-)...:


def cfd(cursor):
    values = cursor,fetchone()
    if values is None:
        return None
    keys = [d[0] for d in cursor.description]
    result_dict = dict(zip(keys, values))
    return result_dict

Of course, you can bunch this up more, or less, according to
taste -- whatever you find clearest and most readable.  E.g.,
the last three lines could be instead:

    return dict(zip([d[0] for d in cursor.description],values))

if you're into one-liners!-)


Alex




From rmunn at pobox.com  Wed Apr 16 14:39:37 2003
From: rmunn at pobox.com (Robin Munn)
Date: Wed, 16 Apr 2003 18:39:37 GMT
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net> <yfswui1f3la.fsf@black132.ex.ac.uk> <9RDla.592$Gb1.345@nwrdny01.gnilink.net> <yfsk7dydmxj.fsf@black132.ex.ac.uk> <z0sma.14160$Gb1.10725@nwrdny01.gnilink.net> <m2n0itfrft.fsf@mycroft.actrix.gen.nz> <slrnb9l507.1q1.rmunn@localhost.localdomain> <QBBma.17307$T34.545916@news2.tin.it> <slrnb9o9ob.493.rmunn@localhost.localdomain> <eX%ma.18570$Gb1.6139@nwrdny01.gnilink.net>
Message-ID: <slrnb9qk41.7eb.rmunn@localhost.localdomain>

Carl Banks <imbosol-1050441179 at aerojockey.com> wrote:
> Robin Munn wrote:
>> I suppose I just illustrated how a try ... finally block might be
>> considered "broken and fragile". I tried to use it, and missed some
>> subtleties. I'm still not convinced that a "with"-like statement is
>> necessary in Python, but I think I just went from -1 to -0 on it.
> 
> I wouldn't be so hasty.
> 
> First, you're basing it on a bad example.  Rebinding sys.stdout is bad
> programming, period.  Wanting to save and restore it is even worse.

Since Alex Martelli gave such a lengthy response (thanks Alex!), I'll
limit myself to a counterexample: pipes. I would call using Unix pipes
"rebinding stdout", handled by the shell. What I was trying to do in my
example was do something like pipes, purely in Python. Specifically,
I've got this function that prints something to sys.stdout. And this
function is not under my control. It could be something in the standard
library, or it could be a hook or a callback which will be written later
by someone else -- wherever it comes from, I can't go diving into the
function and change all occurrences of "sys.stdout.write('foo')" to
"myfileobject.write('foo')". But I need to capture the output of that
program for processing (maybe to display it in a GUI window). If it was
a separate process, I'd use a pipe or os.popen(). But it's a Python
function. Therefore I want to rebind sys.stdout temporarily.

How would you suggest capturing output *without* rebinding sys.stdout?
If there's an alternative, I don't know about it and would welcome a
chance to learn a new approach to the problem.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From staschuk at telusplanet.net  Sat Apr  5 13:49:45 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sat, 5 Apr 2003 11:49:45 -0700
Subject: negative values in profile output
In-Reply-To: <20030404130802.F8345@gurthang.twcny.rr.com>; from mboedick@mboedick.org on Fri, Apr 04, 2003 at 01:08:02PM -0500
References: <20030404130802.F8345@gurthang.twcny.rr.com>
Message-ID: <20030405114944.A371@tibia.amotlpaa.bogus>

Quoth Matthew M. Boedicker:
> When using the profile module to profile a long-running program,
> I am getting some negative values:
> 
> 6753628 -4036.027   -0.001 -4036.027   -0.001 gobject_sql.py:36(<lambda>)
> 
> Am I correct to assume these don't have special meaning and are the
> result of something wrapping?

Looks like a bug to me.

I suppose it could be wrapping, if the Python involved doesn't
promote int to long on overflow.  What version are you running?

I wonder also whether it might be a problem in the bias
calculations inside the profiler.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From cere at u.washington.edu  Mon Apr 28 00:10:30 2003
From: cere at u.washington.edu (Cere Davis)
Date: Sun, 27 Apr 2003 21:10:30 -0700
Subject: overloading print behavior
Message-ID: <vapb1f5mcqf0cd@corp.supernews.com>


So I have a function like:

class C:
    def __init__(self):
       self.a="asdf"
       self.b="123"
c=C()
st=["abc",c]

When I:

print st

I get:

['abc', <__main__.C instance at 0x40288fac>]

Which is what I expect,  but I would like it if print or perhaps the 
str() coercion that is happening behind the scenes would walk through my 
C instance
and resolve the name/value paris within the C class and print the 
values.  Something like:

print st

 > 'abc',<C instance>['a=asdf','b=123']

I'm not sure what the best way to present this would be but I'm lead to 
believe that there is a clever way to do this by registering the 
name/value info of
the variables in the C class with __setitem__ /__getitme__ but I am not 
interly sure how this works...or if it works.  Could someone enlighten 
me on this simple concept?

Thanks,
Cere





From smejmoon at yahoo.com  Mon Apr  7 06:58:25 2003
From: smejmoon at yahoo.com (G'irts Kalnins)
Date: 7 Apr 2003 03:58:25 -0700
Subject: python+php HTTP sessions
Message-ID: <83f82a93.0304070258.6b56aedd@posting.google.com>

do you have any experience with mixed systems?

i'm interested of approaches how to keep same session while using php
generated pages and python generated pages. (session ID, session
variables, e.t.c.)

--Smejmoon



From ramen at lackingtalent.com  Tue Apr  1 10:30:55 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 15:30:55 -0000
Subject: Terminal-based text editor with color syntax hilighting
References: <slrnb8hvka.pki.ramen@lackingtalent.com> <mailman.1049169002.28920.python-list@python.org> <b6bti2$c7f$1@news1.radix.net>
Message-ID: <slrnb8j99n.421.ramen@lackingtalent.com>

In article <b6bti2$c7f$1 at news1.radix.net>, Thomas Dickey wrote:
>>> I'm talking about an SSH session, not an XTerm. Text only.)
> 
> ssh is an application that might run in an xterm
> 

Right, I should have been more clear about this. I'm talking about
non-X-capable terminal sessions like SecureCRT, serial connections, Linux
without XFree86, etc. So it needs to be ncurses, s-lang, or something
similar.

> 
> actually XEmacs has done color for a few years.  For vi-clones, elvis and
> vile both do color as well.
> 

Yeah, XEmacs has fairly sophisticated color hilighting I think...

Thanks,
Dave



From peter at engcorp.com  Tue Apr  8 19:24:11 2003
From: peter at engcorp.com (Peter Hansen)
Date: Tue, 08 Apr 2003 19:24:11 -0400
Subject: Tiobe index algorithm now published
References: <b6v8id$qli$1@reader1.tiscali.nl>
Message-ID: <3E935A1B.CB8D97E2@engcorp.com>

GerritM wrote:
> 
> We have had discussions about the "TIOBE Programming Community Index" at
> http://www.tiobe.com/tpci.htm
> One of our consistent critics was that the algorithm was not published nor
> verifiable. Tiobe has now added the (simple) algorithm on this page. They
> also explain which "programming languages" are monitored.
> 
> For all clarity I have no relationship with Tiobe, although both Paul Jansen
> and myself have worked at different periods at Philips Research.

"""Programming languages that have status "A" are considered to be mainstream 
   languages. Status "A+" and "A++" indicate that a programming language is 
   inbetween status "A" and "B"."""

And Python gets a nice-looking A++.

Somebody is clueless though (other than the inherent problems of this sort
of ranking approach).  Since when does A+ or A++ indicate something less
than A?  In what planet's school system did these guys grow up?

-Peter



From sismex01 at hebmex.com  Tue Apr 29 10:19:18 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Tue, 29 Apr 2003 09:19:18 -0500
Subject: Function to convert fetchone() result from a list to a dictio
	nary
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F062330489CF7A@mail.hebmex.com>

> From: Alex Martelli [mailto:aleax at aleax.it]
> Sent: Tuesday, April 29, 2003 4:09 AM
> 
> So here's a step-by-step version (using SPACES, not TABS:-)...:
> 
> def cfd(cursor):
>     values = cursor,fetchone()
>     if values is None:
>         return None
>     keys = [d[0] for d in cursor.description]
>     result_dict = dict(zip(keys, values))
>     return result_dict
>

I don't think there's need to explicitly return None,
nor to assign result_dict only to immediately return
it, but then, that's just me ;-)


def cfd(cursor):
    values = cursor.fetchone()
    if values:
        keys = [d[0] for d in cursor.description]
        return dict(zip(keys,values))


> Of course, you can bunch this up more, or less, according to
> taste -- whatever you find clearest and most readable.  E.g.,
> the last three lines could be instead:
> 
>     return dict(zip([d[0] for d in cursor.description],values))
> 
> if you're into one-liners!-)
> 

Ach! But Perly stuff is so unmaintainable... ;-)

> Alex
>

-gus




From andrew.gregory at npl.co.uk  Thu Apr  3 09:45:25 2003
From: andrew.gregory at npl.co.uk (Andrew Gregory)
Date: 3 Apr 2003 06:45:25 -0800
Subject: Finding field widths of floats for neat printing
Message-ID: <2830c89c.0304030645.51ad149c@posting.google.com>

I want to write lists of floating point numbers to text files in neat
columns with decimal points aligned, right-hand edges straight (rather
than ragged), and without unnecessary trailing zeros, e.g.
1.1    -34.67     0.006  
2.0     -1.05     0.010  
etc.

In order to do this I need to loop through the lists of numbers for
each column to find the field sizes m and n need to achieve this. I
can then produce a format specifier string (fss) for each column,
e.g. fss = "%" + "%i.%if" % (m,n)

Given a number, -34.67 for example is there a SMART Python way of
determining the field sizes? Obviously I need to work at less than the
precision of the machine, and ignore the fact that -34.67 might be
represented as
-34.6699999999999

I know that there are messy arithmetic methods (can use logs) - and
have implemented one.

Andrew.



From camelo at esss.com.br  Wed Apr 16 08:15:20 2003
From: camelo at esss.com.br (Marcelo A. Camelo)
Date: Wed, 16 Apr 2003 09:15:20 -0300
Subject: Variable inside module
In-Reply-To: <IUFma.19201$T34.609834@news2.tin.it>
Message-ID: <000001c30411$d9819850$0d00000a@esss.com.br>

> any future rebinding of bar.var will have 
> no effect on what foo.var is bound to now.
> Yes, there ARE two different names (and 
> variables are just that -- names): foo.var 
> and bar.var.

Alex and Ken, thank you for the clarification.
Cheers, Marcelo A. Camelo





From edcjones at erols.com  Thu Apr 10 11:43:23 2003
From: edcjones at erols.com (Edward C. Jones)
Date: Thu, 10 Apr 2003 11:43:23 -0400
Subject: Pyrex: wrapping C struct as Python class
Message-ID: <b743fp$k40$1@bob.news.rcn.net>

I have a C struct and functions that use it:

tiny.h:

     typedef struct {
         int ival;
     } Tiny;

     int square(Tiny* t);

tiny.c:

     int squareit(Tiny* t)
     {
         return t.ival * t.ival;
     }

In C code, I would use "malloc", "free" and "squareit" with Tiny. What 
is the best way to wrap this in Pyrex so Python code can do

     t = Tiny(77)
     print t.ival
     print t.squareit()
     del t




From Fabrice.Popineau at supelec.fr  Mon Apr 28 09:05:43 2003
From: Fabrice.Popineau at supelec.fr (Fabrice Popineau)
Date: Mon, 28 Apr 2003 15:05:43 +0200
Subject: LSM2003/VHLL Annoucement
Message-ID: <1xzm3h7c.fsf@esemetz.ese-metz.fr>

The 2003 Libre Software Meeting (LSM) will take place  this year in Metz
(North-East  of France) from July,  9th to July, 12th.  I have taken the
responsibility for the VHLL track (Very High Level Languages) originally
initiated by Robert  Strandh  and headed  by   him for the  first  three
editions. I will be pleased to welcome in Metz all people interested in
VHLL and who will contribute to the success of the 2003 LSM.

This track is meant to show how those high level languages such as Lisp,
Scheme and their   derivatives, ML, Python,  etc. increase productivity,
quality, fiability,  ease of maintenance  in the  context of developping
applications.

Developpers who want  to  share  their  experience in this   domain  are
invited to contact me. Subjects may include :
- development of tools or platforms
- comparison between languages
- developper's experiences using VHLL
- ...

You will find more information about the LSM2003 by browsing
http://rencontresmondiales.org. The web site is frequently updated.

Yours sincerely,

-- 
Fabrice Popineau
------------------------
e-mail:       Fabrice.Popineau at supelec.fr  |  The difference between theory 
voice-mail:   +33 (0) 387764715            |  and practice, is that
surface-mail: Supelec, 2 rue E. Belin,     |  theoretically,
	      F-57070 Metz 	           |  there is no difference !



From edcjones at erols.com  Thu Apr  3 17:38:54 2003
From: edcjones at erols.com (Edward C. Jones)
Date: Thu, 03 Apr 2003 17:38:54 -0500
Subject: How to convert certain typedefs to Pyrex?
Message-ID: <b6lq3p$hls$1@bob.news.rcn.net>

How do I treat the following typedefs in Pyrex?

typedef struct _IplTileInfo IplTileInfo;
typedef struct _CvContourScanner* CvContourScanner;
typedef struct CvPOSITObject CvPOSITObject;




From alanmk at hotmail.com  Fri Apr 18 10:05:20 2003
From: alanmk at hotmail.com (Alan Kennedy)
Date: Fri, 18 Apr 2003 15:05:20 +0100
Subject: Random web page?
References: <b7m4ql$ril$1@news.hccnet.nl> <3E9EAA36.BFB1018D@engcorp.com> <IQmdnQg6xagDdQOjXTWcpA@august.net> <b7neqo$knk$1@news.hccnet.nl> <N_Nna.50877$t_2.4682@afrodite.telenet-ops.be> <b7og5c$h8b$1@news.hccnet.nl> <3E9FFE30.DE250F28@engcorp.com>
Message-ID: <3EA00620.38FC726D@hotmail.com>

Peter Hansen wrote:

> I believe it's since version 1.1 that HTTP has required clients to
> supply the domain name in the HTTP request, to allow this kind of
> virtual hosting.  Think of it as nothing more (to use your analogy)
> than putting a name on an envelope *along with* the address, to
> get it to the right person there.  Prior to 1.1 (?) HTTP allowed
> specifying only the street address...

The "host" header became mandatory in HTTP 1.1

http://rfc.net/rfc2616.html#s14.23

It was optional in HTTP 1.0

http://rfc.net/rfc1945.html

Many HTTP 1.0 implementations chose to provide the "host" header anyway,
even though it wasn't mandatory, to facilitate virtual hosting. The most
notable of these is Netscape 4.x. 

I believe that there are very few browsers left out there that don't
supply the "host" header.

You can check browser headers at this address

http://xhaus.com/headers

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan



From aleax at aleax.it  Sun Apr 13 12:28:15 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 16:28:15 GMT
Subject: How to do per-line profiling?
References: <roy-7A3D8E.11045713042003@reader1.panix.com>
Message-ID: <zegma.8647$LB6.234768@news1.tin.it>

Roy Smith wrote:

> Are there any tools to do profiling on a line-by-line basis?

Yes, package hotshot (part of the standard Python library) is just 
such a tool.  Unfortunately, it's somewhat under-documented!  And,
that's one tidbit where I even gave up when I was writing "Python
in a Nutshell"... I only documented the standard profiler, not hotshot.

I do have about half of each of a user guide and reference for
hotshot written -- at draft level; about 1300 words for the UG
and 900 for the Reference -- I had started writing them for Strakt,
but then ended up giving direct help to my colleagues trying to
use hotshot for a very specific and arduous task (profiling an
asynchronous multi-threaded GUI-rich client...) rather than writing
the rest of the general docs for "sane" use cases.  Now, I hear
that with Python 2.3 it may be the case that hotshot doesn't work
reliably any longer for line-by-line profiling, so, I dunno whether
it's worth pursuing the task of documenting it -- maybe people on
the Python core team are better placed to commend on that!


Alex




From skip at pobox.com  Mon Apr  7 21:31:00 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 7 Apr 2003 20:31:00 -0500
Subject: How to protect python code ?
In-Reply-To: <3E921A4C.1020908@hotmail.com>
References: <3E8BE455.4000904@novagrid.com>
        <mailman.1049389032.4904.python-list@python.org>
        <u8bka.651$001.591065@monger.newsread.com>
        <mailman.1049722991.18682.python-list@python.org>
        <3E921A4C.1020908@hotmail.com>
Message-ID: <16018.9812.163477.819191@montanaro.dyndns.org>

    mark> Our customers are unlikely to go after "movl $0x0007, %ax" with a
    mark> hex editor, but are almost certain to mess with code that looks
    mark> like "licenseCount = 5".

This was the first message in which I saw this kind of detail about why you
want to hide the source, though I must admit that I've ignored most of this
thread (it comes up frequently, so long-time readers quickly tire of it).
There's a big difference between, "I don't want my users to shoot them in
the foot", and, "I need to protect our company's intellectual property from
bad guys".  Most people it seems are interested in the IP issue, not the
hole-in-foot issue.  If it's intellectual property protection you're after,
then you *shouldn't* rely on any sort of encryption scheme to protect your
code because all the keys to decrypting the encryption must be buried in the
executable somewhere.  Bad guys will be able to get to your code.

    mark> My frustration with this question is that it seems to offend the
    mark> c.l.p.  readership to such a degress that they collectively decide
    mark> to drown out any meaningful discussion on the matter.

Nope.  It just comes up so often that, like I said, it gets old in a hurry.

Skip




From 201005 at infosys.com  Wed Apr 16 09:20:19 2003
From: 201005 at infosys.com (201005- M.S.Shivakumaraswamy)
Date: Wed, 16 Apr 2003 18:50:19 +0530
Subject: Hi
Message-ID: <FA13E89D0804CB4DB5FB420B70E2C0CC0ADD952E@kecmsg07.ad.infosys.com>

Hi 

 

I am working on windows

But there is no curses module for the windows

Can any one suggest where can I get curses enabled python??

 

Regards

Shiv

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030416/855ab14e/attachment.html>

From danb_83 at yahoo.com  Sat Apr  5 23:16:08 2003
From: danb_83 at yahoo.com (Dan Bishop)
Date: 5 Apr 2003 20:16:08 -0800
Subject: properties and get/set methods
References: <mailman.1049581810.25706.python-list@python.org>
Message-ID: <ad052e5c.0304052016.4c67355c@posting.google.com>

Timothy Grant <tjg at craigelachie.org> wrote in message news:<mailman.1049581810.25706.python-list at python.org>...
> Hi all,
> 
> Given the following class...
> 
> class monty(object):
>     def __init__(self):
>         self.__foo = 0
>         self.__bar = 0
> 
>     def setvar(self):
>         do_some_attribute_generic_thing()
>         set_appropriate_attribute()        
> 
>     def getvar(self):
>         do_some_attribute_generic_thing()
>         return appropriate_attribute()
> 
>     foo = property(getvar, setvar)
>     bar = property(getvar, setvar)
> 
> =====================================
> 
> Is it possible to determine which attribute was used to as the trigger for the 
> call to setvar() or getvar()? so that set_appropriate_attribute() or 
> appropriate_attribute() can set the correct attribute or get the correct 
> attribute?

Why not just use two different getvar and setvar functions?

However, if you absolutely must insist on using only one getter/setter
pair, you could use

class monty(object):
    def __init__(self):
        self.__foo = 0
        self.__bar = 0

    def setvar(self, value, whatCalledIt=None):
        do_some_attribute_generic_thing()
        set_appropriate_attribute()        
 
    def getvar(self, whatCalledIt=None):
        do_some_attribute_generic_thing()
        return appropriate_attribute()
 
    foo = property(lambda self: self.getvar('foo'),
                   lambda self, value: self.setvar(value, 'foo'))
    bar = property(lambda self: self.getvar('bar'),
                   lambda self, value: self.setvar(value, 'bar'))



From grante at visi.com  Thu Apr 17 11:43:55 2003
From: grante at visi.com (Grant Edwards)
Date: 17 Apr 2003 15:43:55 GMT
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com> <3e9d9bb9$0$151$a1866201@newsreader.visi.com> <a8b7f07a.0304161314.361a80fa@posting.google.com> <3e9dca74$0$170$a1866201@newsreader.visi.com> <dSxna.114901$It5.77194@news2.central.cox.net> <3e9ec5c7$0$178$a1866201@newsreader.visi.com> <KEzna.115586$It5.56408@news2.central.cox.net>
Message-ID: <3e9ecbbb$0$199$a1866201@newsreader.visi.com>

In article <KEzna.115586$It5.56408 at news2.central.cox.net>, Steve Holden wrote:

>> > Now try putting an import before the "for" statement ...
>>
>>   echo -e 'import sys\nfor i in range(5):\n sys.stdout.write(str(i)+"\\n")\n\n' | python2
>>
>> Works for me.  Was it supposed to fail?
> 
> We=ll, you might say I was just too lazy to test it for myself :-)

I couldn't figure out how to get a % in there, but I only tried
for about 30 seconds.  There are apparently bits of the shell
quoting syntax/semantics that I still don't understand after 20
years of using Unix...

-- 
Grant Edwards                   grante             Yow!  I need "RONDO".
                                  at               
                               visi.com            



From donn at u.washington.edu  Tue Apr  1 12:24:35 2003
From: donn at u.washington.edu (Donn Cave)
Date: 1 Apr 2003 17:24:35 GMT
Subject: spawning a separate process
References: <b6c2c6$lr7$1@aspen.sucs.soton.ac.uk>
Message-ID: <b6ci0j$2bbc$1@nntp6.u.washington.edu>

Quoth "Mark Light" <light at soton.ac.uk>:
|        I have a python program from which I wish to spawn an os.system call
| (a bash script) that may take several minutes to run - whilst the rest of
| the script carries on. I have tried using bits of code from books - but it
| never seem to work quite right

os.spawnv can do that and has some other attractions as well.
Perhaps that's one of the things you've already tried.  I guess
I may be wasting my time with you, with your obscure dissatisfactions,
but in case anyone else is reading - check out os.spawnv if you
haven't already, mainly because it's much more reliable for commands
computed from input that might have unexpected shell syntax in them.

	Donn Cave, donn at u.washington.edu



From tim.one at comcast.net  Sun Apr 13 01:30:37 2003
From: tim.one at comcast.net (Tim Peters)
Date: Sun, 13 Apr 2003 01:30:37 -0400
Subject: __slots_ and inheritance
In-Reply-To: <1050073407.3e96d93fded70@mcherm.com>
Message-ID: <LNBBLJKPBEHFEDALKOLCKEBNEDAB.tim.one@comcast.net>

[Michael Chermside]
> As Alexander Schmolck has so eloquently explained, use of
> __slots__ (which is intended simply as a speed-and-memory optimization)
> breaks several things...  among them:
>  1. It breaks weakrefs (can't make a weakref to a __slots__ object).
>  2. It breaks pickling (can't pickle a __slots__ object without
>     custom code).
>  3. It breaks introspection (eg: can't call vars() on a __slots__ object).
>  4. It breaks "decoration" with new attributes (can't add new attributes
>       if object has no __dict__).
>
> Of these, #4 was intentional, but I don't see any reason for the others!
> For instance, introspection (most of it anyhow) could be enabled for
> __slots__ objects by modifying vars() (and several other things) to
> check the __slots__ for a list of additional values. Pickling too could
> set and get values listed in __slots__. I haven't any idea what the
> issue with weakrefs would be.
>
> So can anyone out there tell me if these are (a) planned, but not done
> yet, or (b) welcome, patches would be appreciated, or (c) difficult or
> impossible for some reason I'm not aware of.

A mix of all, but mostly #b and #c now.  #2 is repaired in Python 2.3, but
required a new pickle protocol, and lots of work:

    http://www.python.org/peps/pep-0307.html






From gh at ghaering.de  Tue Apr 22 12:49:06 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 22 Apr 2003 18:49:06 +0200
Subject: Can someone send me the latest Gadfly?
In-Reply-To: <200304220934.38013.scott_list@mischko.com>
References: <200304220934.38013.scott_list@mischko.com>
Message-ID: <3EA57282.7010501@ghaering.de>

Scott Chapman wrote:
> www.chordate.com appears to be seriously down.  Their DNS is resolving to 
> ns1.rackshack.net.  I called Rackshack and they don't have a zone file for 
> chordate.com.
> 
> Chordate hosts the Gadfly database and I would like a copy of it on my RH8 
> system but can't download it.  Google isn't helpful in finding another 
> location to get it from.
> 
> Can someone help out here?

http://sourceforge.net/projects/gadfly
http://sourceforge.net/project/showfiles.php?group_id=662

-- Gerhard




From bdesth at removethis.free.fr  Wed Apr 30 11:20:52 2003
From: bdesth at removethis.free.fr (laotseu)
Date: Wed, 30 Apr 2003 15:20:52 +0000
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org> <1bf4033f.0304290627.40f7d831@posting.google.com> <uptn56xxi.fsf@fitlinxx.com>
Message-ID: <3EAFE9D4.10805@removethis.free.fr>

David Bolen wrote:


(snip)

> 
> In terms of Python IDEs, most are code/debug only, no GUI.  In the
> wxPython world, BOA Constructor is getting there but still too much of
> a moving target to my mind, and you have to like the Delphi model.
> PythonCard may be promising but is also early on.  wxDesigner is nice,
> but not integrated into a more general IDE.

There is also wxGlade, a glade-like RAD tool for wxWindows/wxPython.

Laotseu




From peter at engcorp.com  Wed Apr  2 16:18:50 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 02 Apr 2003 16:18:50 -0500
Subject: How to compile a python script to *.pyo
References: <mailman.1049309105.22947.python-list@python.org> <b6fj57$626$1@news-reader11.wanadoo.fr>
Message-ID: <3E8B53BA.DBB0C84C@engcorp.com>

Philippe Lafoucri?re wrote:
> 
> danny at iamlearning.com wrote:
> 
> > I'm new on this, please bear with we.
> >
> > Let's say I have this script:
> >
> > --------------------------
> > #!/usr/bin/env python
> > x = 'Hello Python'
> >
> > print x,
> >
> > --------------------------
> >
> > How can I compile this script to *.pyo?
> >
> > Thanks in advance,
> >
> > Danny
> 
> python -O yourscript.py

Generally this doesn't work.  You have to import
a module (or use other techniques) for it to be compiled.
Python does not store a .pyc or .pyo file for the main 
module, perhaps surprisingly.

-Peter



From abpillai at lycos.com  Sat Apr 12 07:35:00 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 12 Apr 2003 04:35:00 -0700
Subject: mirroring text window to file
References: <3E9783E3.6010108@videotron.ca> <3E978E80.AC7BA047@alcyone.com>
Message-ID: <bd993a2f.0304120335.ae931d2@posting.google.com>

Try this,

 In your program,
 class YourClass:

    ...
    ...

    def write(self, output):
       text_window.insert('end', output)
       log_file.write(output)

 sys.stdout = YourClass
 
This will direct all 'print' output to the
write method of your class which is the 
wrapper for the two methods you outlined.

Anand Pillai

Erik Max Francis <max at alcyone.com> wrote in message news:<3E978E80.AC7BA047 at alcyone.com>...
> Marcel Achim wrote:
> 
> > I was wondering if there's an existing solution other than writing
> > to both a file and an output text window (Tk text widget), as in :
> > 
> >    text_window.insert('end', output)
> >    log_file.write(output)
> 
> Write a wrapper class that does both.



From bokr at oz.net  Tue Apr 15 14:19:08 2003
From: bokr at oz.net (Bengt Richter)
Date: 15 Apr 2003 18:19:08 GMT
Subject: Using property in classic class may not work
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it> <a8b7f07a.0304150527.5b0b1187@posting.google.com> <hRUma.23848$T34.709378@news2.tin.it>
Message-ID: <b7hies$ol8$0@216.39.172.122>

On Tue, 15 Apr 2003 14:40:13 GMT, Alex Martelli <aleax at aleax.it> wrote:

>A. Lloyd Flanagan wrote:
>
>> Alex Martelli <aleax at aleax.it> wrote in message
>> news:<ifPma.15344$LB6.409776 at news1.tin.it>...
>>> Neil Hodgson wrote:
>>> 
>>> >     The class in question is derived from a wxWindows class, so I can't
>>> > make it derive just from object. The "What's New in Python 2.2"
>>> > document only mentions property wrt new-style clases. Can I depend on
>>> > property working for classic classes? Is there a workaround such as
>>> > deriving from object as well?
>>> 
>>> No, and yes, respectively -- deriving from object means your class
>>> isn't classic any more, so properties do work as intended there.  Vide:
>>> 
>> ... (big chunck cut) ...
>>> as you see, this does work as intended.
>>> 
>>> 
>>> Alex
>> 
>> However, the last time I tried to subclass from both wxFrame and
>> object, wxPython freaked out (that's a technical term for "I don't
>> remember what happened, but boy, it sure didn't work").
>> Not sure about the latest version of wxPython, though -- there's
>> binaries for python 2.3 out by now.  I'll have to try it again..
>
>If wxPython keeps freaking, you may have to give up on properties
>if you HAVE to inherit its classes (can't just wrap with automatic
>delegation) and go back to "good old" (yeah right) __setattr__.  Or,
>gently pressure wxPython's maintainers to have them move to
>new-style classes... whatever aspects of old-style classes they
>may be taking for granted at present, migrating to new-style can
>hardly be a major job, in my experience.
>
>
I may be fooling myself here, but a quick experiment suggests being
able to define a new class that inherits from both a classic class
that you can't touch and a new-style one that you can give a property, e.g.:

 Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
 Type "help", "copyright", "credits" or "license" for more information.
 >>> class Old3rdParty:
 ...     def oldMethod(self): print 'oldMethod was called'
 ...
 >>> class NewPropertyCarrier(object):
 ...     def __init__(self, v): self.prop = v
 ...     def _get_p(self): return self._v
 ...     def _set_p(self, v): self._v = v
 ...     prop = property(_get_p, _set_p)
 ...
 >>> class Derived(Old3rdParty, NewPropertyCarrier):
 ...     pass
 ...
 >>> d = Derived(123)
 >>> d
 <__main__.Derived object at 0x007D68E0>
 >>> d.prop
 123
 >>> d.oldMethod()
 oldMethod was called
 >>> d.prop = 456
 >>> d.prop
 456

It seems like prop is surviving as a property, not getting clobbered into an attribute:

 >>> dir(d)
 ['__class__', '__delattr__', '__dict__', '__doc__', '__getattribute__', '__hash__', '__init__',
 '__module__', '__new__', '__reduce__', '__repr__', '__setattr__', '__str__', '__weakref__', '_ge
 t_p', '_set_p', '_v', 'oldMethod', 'prop']
 >>> vars(d)
 {'_v': 456}
 >>> d._v
 456
 >>> d.prop
 456
 >>> d._v = 'sneaky'
 >>> d.prop
 'sneaky'
 >>> d.prop = 'normal'
 >>> d.prop
 'normal'
 >>> d._v
 'normal'

So could wxPython play the role of Old3rdParty above and allow building a Derived with
the properties desired? Or is mixed bases a no-no?

Regards,
Bengt Richter



From robin at jessikat.fsnet.co.uk  Thu Apr 17 09:37:20 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Thu, 17 Apr 2003 14:37:20 +0100
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org>
 <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk> <3E9EA707.4DB53582@engcorp.com>
Message-ID: <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk>

In article <3E9EA707.4DB53582 at engcorp.com>, Peter Hansen
<peter at engcorp.com> writes
>> This sounds remarkably similar to what we commonly call genetic
>> optimisation or evolution strategy selection. There a fitness criterion
>> is applied to randomly selected elements of a pool of solutions, with
>
>The key is that XP is _not_ applying "randomly selected elements".
>The experience and abilities of the programmer are heavily involved
>in the process.
interestingly genetic programming et al seem to perform worse when
attempts are made to 'direct' the evolution. 
-- 
Robin Becker



From aleax at aleax.it  Sat Apr 12 18:14:12 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sat, 12 Apr 2003 22:14:12 GMT
Subject: Handling slices
References: <mailman.1049676910.30971.python-list@python.org> <7h3ptnyio34.fsf@pc150.maths.bris.ac.uk>
Message-ID: <Uc0ma.8119$T34.231657@news2.tin.it>

Michael Hudson wrote:

> Francois Pinard <pinard at iro.umontreal.ca> writes:
> 
>> Hi, Python friends.
>> 
>> I'm currently using Python 2.2.1.  How does one support slices when given
>> to __getitem__?  Suppose I get a slice object.  How can I use it to get
>> a slice of a local array?  In other words, how one does apply a slice?
> 
> In 2.2.1, by hand, I think.

Essentially, yes, though it can be nicer than this...:

> def __getitem__(self, item):
>     if isinstance(item, types.SliceType):
>        if item.step != 1:
>           raise ValueError
>        if item.start is None:
>           start = 0
>        else:
>           start = item.start
>        if item.stop is None:
>           stop = len(self.local_array)
>        else:
>           stop = item.stop
>        return self.local_array[start:stop]

I'd probably take another tack, such as (warning, untested!!!)...:

def applyslice(asequence, aslice):
    seqlen = len(asequence)
    def mkidx(x, default):
        if x is None: return default
        elif x>=0: result = x
        else: result = x + seqlen
        return max(0, min(x, seqlen-1))
    if aslice.step is None: step = 1
    else: step = aslice.step
    if step > 0:
        default_start = 0
        default_stop = seqlen
    else:
        default_start = seqlen-1
        default_stop = -1
    start = mkidx(aslice.start, default_start)
    stop = mkidx(aslice.stop, default_stop)
    return type(asequence)([asequence[z]
            for x in xrange(start, stop, step)
           ])

of course, this needs to be refactored (and how!) to also
be usable for __setitem__ AND __delitem__... and it sure
makes a LOT of sense to specialcase the common case where
step is 1 to delegate right to asequence's slicing...


> I presume you don't need to be told that this is all much nicer in
> 2.3...

Given that built-in sequences now accept slice objects in their
__getitem__, all of the above becomes just a
    return asequence[aslice]
which IS indeed nicer;-).  But you might still need to keep the
applyslice function around in case you need to deal with sequence
types from older extensions, or the like...


Alex




From news at manuelmgarcia.com  Fri Apr 18 19:43:42 2003
From: news at manuelmgarcia.com (Manuel Garcia)
Date: Fri, 18 Apr 2003 23:43:42 GMT
Subject: Are reference cycles bad form?
Message-ID: <3o11avop6vkh3ake8dgbtst93t1or771ct@4ax.com>

I would like to ask the group, are reference cycles bad form?

I am debating whether to give a 'child' object a reference to its
'parent' object during creation of the 'child'.

Before Python had garbage collection, I wouldn't dare make reference
cycles.  Rather than create a reference cycle that I would have to
manually break later, I would just make my algorithms a little smarter
so they could figure out the 'parent' without help from the 'child'.

Python now has garbage collection, and I don't have to worry about
creating cycles, but is it bad form to create these cycles when I can
just make my algorithms a little smarter?

Using the 'weakref' module seems silly.  I have complete confidence
with Python's garbage collection, so why slow things down with extra
indirection?

I will probably refactor the algorithms with and without the 'child'
-> 'parent' reference, and see which I like better.  (Sometimes a
spiderweb of references is more of a crutch than a help.)

But as a general pricipal, is a design without reference cycles better
than a design with them?

(Uh oh, I am having a Visual Basic flashback... shaking and cold
sweats...  ;-)

Manuel



From tim at bladerman.com  Thu Apr 10 15:23:24 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Thu, 10 Apr 2003 19:23:24 GMT
Subject: Python compared toJava
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com> <3E95A2A7.238BFA6C@engcorp.com> <3e95b7d6$1@rmcnt03a.rmc.ca>
Message-ID: <Mwjla.5104$sJ3.17511@newsb.telia.net>

From: "bourassa" <bourassa-m at rmc.ca>
> Peter Hansen wrote:
> > Blaktyger wrote:
> >
> >>I recently read that the Python interpreter is like the Java Virtual
Machine.
> >>Is that true?
> >
> > Note: see http://www.catb.org/~esr/faqs/smart-questions.html before
posting
> > again please.
> >
> Would it have bee so hard to either ignore the post or to give a useful
> url?  Learn some manners before _your_ next post.


Yes!!! I see a flamewar coming towards me over the horizon :-)
Just when "Desert storm II" is fading out of the tv news blacktyger steps in
and gives us something to be upset about.

Lets start calling eachother names right from the start this time...

...you lizzard, me worm...





From ialbert at mailblocks.com  Sat Apr 12 22:14:06 2003
From: ialbert at mailblocks.com (Istvan Albert)
Date: Sun, 13 Apr 2003 02:14:06 GMT
Subject: an ugly file-reading pattern
Message-ID: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>

I have been working with perl and java for quite a
few years now and I am currently actively looking
to replace perl wit something better, a language that
promotes software engineering principles. I was
directed toward python and ruby, I have been
working with both of them for a couple of weeks now.

And now for the python question:

Suppose that I'm looking for a way to
read lines from a text file:

now there is the readlines() method
that slurps everything up (no good for me
since I don't like to build into my systems
implicit assumptions about filesizes)

if I choose to process the file line by line
the "Learning Python" book advises me to use the
following atrocity:

while 1:
	line = file.readline()
	if not line: break

Maybe I'm picky here but having to start an infinite loop
then needing a conditional to  break out of it is
anything but an elegant pattern.

This has left me wondering about python,

cheers,

Istvan.




From tjreedy at udel.edu  Wed Apr 16 09:37:08 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Wed, 16 Apr 2003 09:37:08 -0400
Subject: A better unittest
References: <3ckier9a.fsf@python.net>
Message-ID: <IrOdnUSmPonOxwCjXTWcpA@comcast.com>

"Thomas Heller" <theller at python.net> wrote in message
news:3ckier9a.fsf at python.net...
======================================================================
> FAIL: test_fail (__main__.FailingTests)
> --------------------------------------------------------------------
--
> Traceback (most recent call last):
>   File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 35, in
test_fail
>     self.fail("fail on purpose")
>   File "C:\Python22\lib\unittest.py", line 254, in fail
>     raise self.failureException, msg
> AssertionError: fail on purpose

<versus>

======================================================================
> FAIL: test_fail (__main__.FailingTests)
> --------------------------------------------------------------------
--
> Traceback (most recent call last):
>   File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 35, in
test_fail
>     self.fail("fail on purpose")
> AssertionError: fail on purpose

> What do people think? Is this a useful change, and should I submit a
> patch?

I plan to start using inittest Real Soon Now.  For normal use,  the
irrelevant display of inittest internals constitutes noise and I
believe I would prefer its suppression.  And yet, it is a bit
disconcerting: the last displayed line does *not* raise the displayed
exception.  So there is a visual 'mismatch'.  Two suggestions for
further revision.

1. The phrase 'most recent call last' is not longer true.  Substitute
'inittest internals suppressed' or somesuch.

2. Add some visual indicator of the omission, such as an ellipsis line
'...'.

Traceback (inittest internals suppressed):
  File "C:\tss5\components\_Pythonlib\xunit\xunit.py", line 35, in
test_fail
    self.fail("fail on purpose")
...
AssertionError: fail on purpose

Or maybe a prefix to the last line

>>AssertionError: fail on purpose

This patch effectively assumes that inittest works perfectly now and
after all future patches.  While I would make the suppression the
default, I suspect that normal tracebacks might sometimes be a needed
option.

Terry J. Reedy







From martin at v.loewis.de  Sun Apr  6 13:40:51 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 06 Apr 2003 19:40:51 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <m3r88g9iqt.fsf@mira.informatik.hu-berlin.de> <m2brzjacbo.fsf@localhost.localdomain>
Message-ID: <m33ckvo76k.fsf@mira.informatik.hu-berlin.de>

pobrien at orbtech.com (Patrick K. O'Brien) writes:

> Are there translators out there?  I looked and couldn't find any.  

Parsing S-Expressions is fairly trivial - they are designed to be
readily parsable. For a full translator, you need to define the
language mapping first, i.e. how do you want to represent atoms,
function names, etc.

> That's pretty much the approach I planned to take, after I got some of
> the more basic pieces in place.  For example, file handling.  If you
> try to save a file and another process has modified it, Emacs warns
> you.  Where is that kind of code kept in Emacs?  Is it in elisp, or
> part of the core Emacs written in C?  

It very much depends on the precise problem you are looking at: Some
is elisp, some is C, and the boundaries are weak - just as they are in
Python. 

For this specific problem, if you do C-x C-s, it invokes
save-buffer. C-h f tells you that this is in "files". Looking at the
code in files.el reveals that it eventually calls basic-save-buffer;
you could trace this this further down.

Alternatively, you look at the error message "has changed since
visited or saved". You grep through the emacs source, and find
that it is in files.el, in basic-save-buffer:

	  (or (verify-visited-file-modtime (current-buffer))
	      (not (file-exists-p buffer-file-name))
	      (yes-or-no-p
	       (format "%s has changed since visited or saved.  Save anyway? "
		       (file-name-nondirectory buffer-file-name)))
	      (error "Save not confirmed"))

So the message certainly comes from elisp; the test is
verify-visited-file-modtime. C-h f shows it is a builtin function;
a grep reveals it is in fileio.c, which (essentially) reads

  if (stat (SDATA (filename), &st) < 0)
    {
      /* If the file doesn't exist now and didn't exist before,
	 we say that it isn't modified, provided the error is a tame one.  */
      if (errno == ENOENT || errno == EACCES || errno == ENOTDIR)
	st.st_mtime = -1;
      else
	st.st_mtime = 0;
    }
  if (st.st_mtime == b->modtime
      /* If both are positive, accept them if they are off by one second.  */
      || (st.st_mtime > 0 && b->modtime > 0
	  && (st.st_mtime == b->modtime + 1
	      || st.st_mtime == b->modtime - 1)))
    return Qt;
  return Qnil;

This is, of course, all off-topic for comp.lang.python.

> Is there a document anywhere that gives an overview of the design of
> Emacs that would answer these kinds of questions?

No, because the decision to implement something in C or elisp is not a
design decision - you can elevate functions from C to lisp
(introducing deeper primitives) or move functions form lisp to C at
will, without changing the design.

Regards,
Martin



From exnihilo at NOmyrealCAPSbox.com  Sun Apr  6 20:28:13 2003
From: exnihilo at NOmyrealCAPSbox.com (nihilo)
Date: Mon, 07 Apr 2003 00:28:13 GMT
Subject: doing hundreds of re.subs efficiently on large strings
In-Reply-To: <b65cq3$mi5$0@216.39.172.122>
References: <3E80CE1D.1000200@NOmyrealCAPSbox.com> <b5qu69$t98$0@216.39.172.122> <3E8370F2.2030806@NOmyrealCAPSbox.com> <b65cq3$mi5$0@216.39.172.122>
Message-ID: <3E90C634.1030900@NOmyrealCAPSbox.com>

Bengt Richter wrote:
> 
> Thanks for testing, that is interesting. It would be interesting to see
> the test code you ran, so we could see the reasons for .25 vs .43 (and
> I could revise my reality-model as necessary ;-)
> 
> Regards,
> Bengt Richter

Hi Bengt,

Here is some code that I created to test the two versions.

Original:


import time, re
start = time.clock()
data = '' # get http://news.google.com for sample data to run this on. I 
was just putting this in the file for testing to avoid time 
complications over the network. I've omitted it here because it is more 
than 50KB.

data = data.replace('US and Britain', 'Oceania')
data = data.replace('US and British', 'Oceanian')
data = data.replace('America and Britain', 'Oceania')
data = data.replace('the U.S.A.', 'Oceania')
data = data.replace('the United States of America', 'Oceania')
data = data.replace('the U.S.', 'Oceania')
data = data.replace('the USA', 'Oceania')
data = data.replace('the US', 'Oceania')
data = data.replace('The US', 'Oceania')
data = data.replace('the United States', 'Oceania')
data = data.replace('U.S.A.', 'Oceania')
data = data.replace('United States of America', 'Oceania')
data = data.replace('U.S.', 'Oceania')
data = data.replace(' US', ' Oceania')
data = data.replace('US ', 'Oceania ')
data = data.replace('US,', 'Oceania,')
data = data.replace('US-', 'Oceania-')
data = data.replace('United States', 'Oceania')
data = data.replace('American', 'Oceanian')
data = data.replace('British', 'Oceanian')
data = data.replace('Great Britain', 'Oceania')
data = data.replace('Britain', 'Oceania')
data = data.replace('United Kingdom', 'Oceania')
data = data.replace('European Union countries', 'Oceania provinces')
data = data.replace('European nations', 'Oceania provinces')
data = data.replace('European', 'Oceanian')
data = data.replace('Europe', 'Oceania')
data = data.replace('the EU', 'Oceania')
data = data.replace('Australian', 'Oceanian')
data = data.replace('Australia', 'Oceania')
data = data.replace('Washington Post', 'Ministry of Truth')
data = data.replace('Washington DC', 'Airstrip One')
data = data.replace('Washington D.C.', 'Airstrip One')
data = data.replace('Washington, D.C.', 'Airstrip One')
data = data.replace('Washington, DC', 'Airstrip One')
data = data.replace('Washington', 'Airstrip One')
data = data.replace('WASHINGTON', 'Airstrip One')
data = data.replace('Boy Scouts', 'Junior Anti-Sex League')
data = data.replace('Girl Scouts', 'Junior Anti-Sex League')
data = data.replace('President Hussein', 'Emmanuel Goldstein, Enemy of 
the People')
data = data.replace('Saddam Hussein', 'Emmanuel Goldstein')
data = data.replace('Saddam', 'Goldstein')
data = data.replace('Hussein', 'Goldstein')
data = data.replace('President Bush', 'Big Brother')
data = data.replace('president Bush', 'Big Brother')
data = data.replace('George Bush', 'Big Brother')
data = data.replace('George W. Bush', 'Big Brother')
data = data.replace('President George W Bush', 'Big Brother')
data = data.replace('George W Bush', 'Big Brother')
data = data.replace('Mr. Bush', 'Big Brother')
data = data.replace('Bush', 'Big Brother')
data = data.replace('Donald Rumsfeld', 'O\' Brien')
data = data.replace('Rumsfeld', "O' Brien")
data = data.replace('United Nations', 'bloc nations')
data = data.replace('UN', 'bloc nations')
data = data.replace('U.N.', 'bloc nations')
data = data.replace('coalition allies', 'bloc allies')
data = data.replace(' allies', ' bloc allies')
data = data.replace('Japanese', 'Eastasian')
data = data.replace('Chinese', 'Eastasian')
data = data.replace('Japan', 'Eastasia')
data = data.replace('China', 'Eastasia')
data = data.replace('North Korea', 'Eastasia')
data = data.replace('South Korea', 'Eastasia')
data = data.replace('India', 'Eastasia')
data = data.replace('Hong Kong', 'Eastasia')
data = data.replace('Vietnam', 'Eastasia')
data = data.replace('Singapore', 'Eastasia')
data = data.replace('the Phillipines', 'Eastasia')
data = data.replace('Thailand', 'Eastasia')
data = data.replace('Sri Lanka', 'Eastasia')
data = data.replace('Mexico', 'Oceania')
data = data.replace('Canadian', 'Oceanian')
data = data.replace('Canada', 'Oceania')
data = data.replace('Federal Bureau of Investigation', 'Ministry of Love')
data = data.replace('Department of State', 'Ministry of Truth')
data = data.replace('Pentagon', 'Ministry of Peace')
data = data.replace(' Post', ' Ministry of Truth')
data = data.replace('Russian', 'Eastasian')
data = data.replace('Russia', 'Eastasia')
data = data.replace('Australian', 'Oceanian')
data = data.replace('Australia', 'Oceania')
data = data.replace('Christian Science Monitor', 'INGSOC Science Monitor')
data = data.replace('Christianity', 'INGSOC')
data = data.replace('Washington Post', 'Ministry of Truth')
data = data.replace('Washington&nbsp;Post', 'Ministry&nbsp;of&nbsp;Truth')
data = data.replace('Daily&nbsp;Telegraph', 'Ministry&nbsp;of&nbsp;Truth')
data = data.replace('New&nbsp;York&nbsp;Times', 
'Ministry&nbsp;of&nbsp;Truth')
data = data.replace('International Herald Tribune', 'Ministry of Truth')
data = data.replace('International&nbsp;Herald&nbsp;Tribune', 
'Ministry&nbsp;of&nbsp;Truth')
data = data.replace('BBC', 'Ministry of Truth')
data = data.replace('wartime', 'peacetime')
data = data.replace('congressional', 'Inner Party')
data = data.replace('Senate', 'Inner Party Senate')
data = data.replace('House of Representatives', 'Inner Party House')
data = data.replace('House', 'Inner Party House')
data = data.replace('Congress', 'Inner Party')
data = data.replace('Marines', 'Peace Keepers')
data = data.replace('soldiers', 'peace keepers')
data = data.replace('Republican Guards', 'Resistance Elite Troops')
data = data.replace('Republican Guard', 'Resistance Elite Troops')
data = data.replace('Department of Defense', 'Ministry of Peace')
data = data.replace('Department of Homeland Security', 'Ministry of Love')
data = data.replace('Associated Press', 'Ministry of Truth')
data = data.replace('FBI', 'Ministry of Love')
data = data.replace('CNN', 'Ministry of Truth')
data = data.replace('New York Times', 'Ministry of Truth')
data = data.replace('UPI', 'Ministry of Truth')
data = data.replace('United Press International', 'Ministry of Truth')
data = data.replace('Federal Reserve', 'Ministry of Plenty')
data = data.replace('Homeland Security', 'Motherland Security')
data = data.replace('homeland security', 'motherland security')
data = data.replace('protesters', 'restistance thought criminals')
data = data.replace('prisoner', 'Ministry of Love prison guest')
# data = data.replace(' prison', 'Ministry of Love')
data = data.replace('solitary confinement', 'Room 101')
data = data.replace('anti-war', 'Resistance')
data = data.replace('Anti-war', 'Resistance')
data = data.replace('paramilitary', 'Ministry of Peace')
data = data.replace('para-military', 'Ministry of Peace')
data = data.replace('military', 'Ministry of Peace')
data = data.replace('Military', 'Ministry of Peace')
data = data.replace('Gulf War', 'Gulf Peace Conflict')
data = data.replace('news network', 'Ministry of Truth')
data = data.replace('Al-Qaeda', 'Resistance')
data = data.replace('Al Qaeda', 'Resistance')
data = data.replace('Al-Quaida', 'Resistance')
data = data.replace('al Qaeda', 'Resistance')
data = data.replace('Qaeda', 'Resistance')
data = data.replace('Taliban', 'Resistance')
data = data.replace('Taleban', 'Resistance')
data = data.replace("Royal Air Force", 'Ministry of Peace Air Force')
data = data.replace('air force', 'Ministry of Peace Air Force')
data = data.replace('Invasion', 'Liberation')
data = data.replace('invasion', 'liberation')
data = data.replace('\'shock and awe\'', 'awe-shock')
data = data.replace('\"shock and awe\"', 'awe-shock')
data = data.replace('weapons of mass destruction', 'gifts of mass coercion')
data = data.replace('demonstrators', 'thoughtcriminals')
data = data.replace('protestors', 'thoughtcriminals')

print time.clock() - start

And the modified code is:

import time, re
start = time.clock()
data = '' # same data
dict = { 'US and Britain' : 'Oceania', 'US and British' : 'Oceanian', 
'America and Britain' : 'Oceania','the U.S.A.' : 'Oceania', 'the United 
States of America' : 'Oceania', 'the U.S.' : 'Oceania', 'the USA' : 
'Oceania', 'the US' : 'Oceania', 'The US' : 'Oceania', 'the United 
States' : 'Oceania', 'U.S.A.' : 'Oceania', 'United States of America' : 
'Oceania', 'U.S.' : 'Oceania', ' US' : ' Oceania', 'US ' : 'Oceania ', 
'US,' : 'Oceania,', 'US-' : 'Oceania-', 'United States' : 'Oceania', 
'American' : 'Oceanian', 'British': 'Oceanian', 'Great Britain' : 
'Oceania', 'Britain' : 'Oceania', 'United Kingdom' : 'Oceania', 
'European Union countries' : 'Oceania provinces', 'European nations' : 
'Oceania provinces', 'European' : 'Oceanian', 'Europe' : 'Oceania', 'the 
EU' : 'Oceania', 'Australian' : 'Oceanian', 'Australia' : 'Oceania', 
'Washington Post' : 'Ministry of Truth', 'Washington DC' : 'Airstrip 
One', 'Washington D.C.' : 'Airstrip One','Washingto' : '.C.', 
'Washingto' : 'C', 'Washington' : 'Airstrip One', 'WASHINGTON' : 
'Airstrip One', 'Boy Scouts' : 'Junior Anti-Sex League', 'Girl Scouts' : 
'Junior Anti-Sex League', 'President Hussein' : 'Emmanuel Goldstei', 
'Saddam Hussein' : 'Emmanuel Goldstein', 'Saddam' : 'Goldstein', 
'Hussein' : 'Goldstein', 'President Bush' : 'Big Brother', 'president 
Bush' : 'Big Brother', 'George Bush' : 'Big Brother', 'George W. Bush' : 
'Big Brother', 'President George W Bush' : 'Big Brother', 'George W 
Bush' : 'Big Brother', 'Mr. Bush' : 'Big Brother', 'Bush' : 'Big 
Brother', 'Donald Rumsfeld' : 'O\' Brien', 'Rumsfeld' : 'O\'Brien', 
'United Nations' : 'bloc nations', 'UN' : 'bloc nations', 'U.N.' : 'bloc 
nations', 'coalition allies' : 'bloc allies', ' allies' : ' bloc 
allies', 'Japanese' : 'Eastasian', 'Chinese' : 'Eastasian', 'Japan' : 
'Eastasia', 'China' : 'Eastasia', 'North Korea' : 'Eastasia', 'South 
Korea' : 'Eastasia', 'India' : 'Eastasia', 'Hong Kong' : 'Eastasia', 
'Vietnam' : 'Eastasia', 'Singapore' : 'Eastasia', 'the Phillipines' : 
'Eastasia', 'Thailand' : 'Eastasia', 'Sri Lanka' : 'Eastasia', 'Mexico' 
: 'Oceania', 'Canadian' : 'Oceanian', 'Canada' : 'Oceania', 'Federal 
Bureau of Investigation' : 'Ministry of Love', 'Department of State' : 
'Ministry of Truth', 'Pentagon' : 'Ministry of Peace', ' Post' : ' 
Ministry of Truth', 'Russian' :'Eastasian', 'Russia' : 'Eastasia', 
'Australian' : 'Oceanian', 'Australia' : 'Oceania', 'Christian Science 
Monitor' : 'INGSOC Science Monitor', 'Christianity' : 'INGSOC', 
'Washington Post' : 'Ministry of Truth', 'Washington&nbsp;Post' : 
'Ministry&nbsp;of&nbsp;Truth', 'Daily&nbsp;Telegraph' : 
'Ministry&nbsp;of&nbsp;Truth', 'New&nbsp;York&nbsp;Times' : 
'Ministry&nbsp;of&nbsp;Truth', 'International Herald Tribune' : 
'Ministry of Truth', 'International&nbsp;Herald&nbsp;Tribune' : 
'Ministry&nbsp;of&nbsp;Truth', 'BBC' : 'Ministry of Truth', 'wartime' : 
'peacetime', 'congressional' : 'Inner Party', 'Senate' : 'Inner Party 
Senate', 'House of Representatives' : 'Inner Party House', 'House' : 
'Inner Party House', 'Congress' : 'Inner Party', 'Marines' : 'Peace 
Keepers', 'soldiers' : 'peace keepers', 'Republican Guards' : 
'Resistance EliteTroops', 'Republican Guard' : 'Resistance Elite 
Troops', 'Department of Defense' : 'Ministry of Peace', 'Department of 
Homeland Security' : 'Ministry of Love', 'Associated Press' : 'Ministry 
of Truth', 'FBI' :'Ministry of Love', 'CNN' : 'Ministry of Truth', 'New 
York Times' : 'Ministry of Truth', 'UPI' : 'Ministry of Truth', 'United 
Press International' : 'Ministry of Truth', 'Federal Reserve' : 
'Ministry of Plenty', 'Homeland Security' : 'Motherland Security', 
'homeland security' : 'motherland security', 'protesters': 'restistance 
thought criminals', 'prisoner' : 'Ministry of Love prison guest', 
'solitary confinement' : 'Room 101', 'anti-war' : 'Resistance', 
'Anti-war' : 'Resistance', 'paramilitary' : 'Ministry of 
Peace','para-military' : 'Ministry of Peace', 'military' : 'Ministry of 
Peace', 'Military' : 'Ministry of Peace', 'Gulf War' : 'Gulf Peace 
Conflict', 'news network' : 'Ministry of Truth', 'Al-Qaeda' : 
'Resistance', 'Al Qaeda' : 'Resistance', 'Al-Quaida' : 'Resistance', 'al 
Qaeda' : 'Resistance', 'Qaeda' : 'Resistance', 'Taliban' : 'Resistance', 
'Taleban' : 'Resistance', 'Royal Air Force' : 'Ministry of Peace Air 
Force', 'air force' : 'Ministry of Peace Air Force', 'Invasion' : 
'Liberation', 'invasion' : 'liberation', '\'shockand awe\'' : 
'awe-shock', '\"shock and awe\"' : 'awe-shock', 'weapons of mass 
destruction' : 'gifts of mass coercion', 'demonstrators' : 
'thoughtcriminals', 'protestors' : 'thoughtcriminals' }
pat = re.compile('(Canada|the U\.S\.A\.|air force|International Herald 
Tribune|Associated Press|prisoner|Great Britain|George W Bush|US and 
British|soldiers|Federal Bureau of 
Investigation|Europe|demonstrators|Australia|Congress|Saddam|the 
U\.S\.|Boy Scouts|solitary 
confinement|Taliban|Russian|Thailand|military|Hong 
Kong|congressional|protestors|Department of Defense|House of 
Representatives|Christian Science Monitor|Canadian|President Bush|the 
US|European 
nations|BBC|protesters|Daily&nbsp;Telegraph|Washington|American|invasion|Invasion|Military|Marines| 
US|Rumsfeld|U\.N\.|Gulf War|paramilitary|North Korea|coalition 
allies|House|para-military|US |India|\\\'shock and awe\\\'|Senate|George 
W. Bush|\\\"shock and awe\\\"|Al Qaeda|South Korea|Pentagon|President 
George W Bush|US,| allies|Washington|New York 
Times|wartime|Christianity|Taleban|Bush|New&nbsp;York&nbsp;Times|Al-Quaida|homeland 
security|Washington Post|Washington DC|Australian| Post|Mr\. Bush|United 
Nations|President Hussein|the United States|Homeland Security|Federal 
Reserve|America and Britain|Singapore|president Bush|the 
Phillipines|United States of America|Britain|Donald Rumsfeld|Republican 
Guards|Department of Homeland Security|Qaeda|China|al 
Qaeda|European|Department of State|Washington&nbsp;Post|the EU|George 
Bush|International&nbsp;Herald&nbsp;Tribune|UPI|WASHINGTON|U\.S\.|news 
network|European Union countries|FBI|British|Royal Air Force|United 
States|United Press International|Vietnam|Republican 
Guard|anti-war|CNN|Anti-war|Hussein|Russia|Al-Qaeda|Washington 
D\.C\.|Saddam Hussein|the United States of America|Chinese|Mexico|Girl 
Scouts|U\.S\.A\.|US-|the USA|Japanese|United Kingdom|US and 
Britain|UN|The US|Sri Lanka|Japan|weapons of mass destruction)')
strings = pat.split(data)
for i in xrange(1,len(strings),2):
     strings[i] = dict[strings[i]]
data = ''.join(strings)

I just checked these again, and the results were .052 on average for the 
first version, and .080 on average for the second. The second would 
obviously be a lot faster if I pre-compiled the pattern and cPickled it, 
but I wanted to include compilation time in the test.  The fastest 
version that I have tested is using Alex's replacement suggestion in 
combination with pre-compilation and cPickle, but the naive data = 
data.replace is surprisingly fast.

cheers,

nihilo




From mertz at gnosis.cx  Sun Apr 27 22:30:00 2003
From: mertz at gnosis.cx (Lulu of the Lotus-Eaters)
Date: Sun, 27 Apr 2003 22:30:00 -0400
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz>
 <698f09f8.0304262115.386c2571@posting.google.com>
 <3EAB6A14.707AD826@alcyone.com> <ahSdnaSulLiu7TajXTWcow@comcast.com>
 <7ULqa.9804$3M4.264757@news1.tin.it> <g3qdncyq1Zx2FzGjXTWcrg@comcast.com>
Message-ID: <oIJr+kKkXEwV092yn@gnosis.cx>         

"Terry Reedy" <tjreedy at udel.edu> wrote previously:
|At one time, it was easy for anyone to get 100% on a quiz like 'how
|many of the following lists can Python sort?".  Now, even some
|experienced users, might miss the correct answer for the following
|pair (0, 1, or 2?):
    [1, 1j].sort()
    ['1', 1j].sort()

And to be even trickier, answer the question for the following lists:

    [u'x', 'x', chr(255)].sort()
    ['x', chr(255), u'x'].sort()
    [1j, '1', 1].sort()
    [1, 1j, '1'].sort()


Then try to explain the answer to ANYONE you know.

--
    _/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: Postmodern Enterprises _/_/_/
   _/_/    ~~~~~~~~~~~~~~~~~~~~[mertz at gnosis.cx]~~~~~~~~~~~~~~~~~~~~~  _/_/
  _/_/  The opinions expressed here must be those of my employer...   _/_/
 _/_/_/_/_/_/_/_/_/_/ Surely you don't think that *I* believe them!  _/_/





From theller at python.net  Fri Apr 18 16:40:23 2003
From: theller at python.net (Thomas Heller)
Date: 18 Apr 2003 22:40:23 +0200
Subject: SOLVED: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net> <b7mbnb$pp2$06$1@news.t-online.com> <8yu9gpnb.fsf@python.net> <b7mf9u$2jb$03$1@news.t-online.com> <b7mjlc$6tm$03$1@news.t-online.com> <b7p1mq$i2t$02$1@news.t-online.com>
Message-ID: <7k9rh748.fsf@python.net>

"Harald Schneider" <h_schneider at marketmix.com> writes:

> Strange ... I changed the order of the import statements from
> 
> import win32serviceutil,win32service,win32event,pywintypes
> import os,string,rpcserver,xmlrpclib,confighandler
> from agent import *
> from setuptool import RegistryTool
> 
> to
> 
> from agent import *
> from setuptool import RegistryTool
> import win32serviceutil,win32service,win32event,pywintypes
> import os,string,rpcserver,xmlrpclib,confighandler
> 
> and not it works!
> 
> The registry access is done in a class contained in the agent module. Under
> Python 2.1 the order played no role, under 2.2 it does!
> 

I think I've seen this before, but I'm not sure.  IIRC, using 'from
agent import this, that, whatever' instead of the 'from agent import *'
may also fix the problem.

Usually it is bad style anyway ;-).

Thomas



From cben at techunix.technion.ac.il  Fri Apr 11 08:47:01 2003
From: cben at techunix.technion.ac.il (Beni Cherniavsky)
Date: Fri, 11 Apr 2003 15:47:01 +0300 (IDT)
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1049723958.31192.python-list@python.org>
References: <m2u1dcw22s.fsf@localhost.localdomain> <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>
 <b6qs04$7eqi1$1@ID-169208.news.dfncis.de> <0X7ka.41242$wM6.2435@nwrdny01.gnilink.net>
 <mailman.1049693590.23386.python-list@python.org> <m28yum76rt.fsf@localhost.localdomain>
 <mailman.1049723958.31192.python-list@python.org>
Message-ID: <Pine.GSO.4.51.0304111544290.15707@techunix.technion.ac.il>

Beni Cherniavsky wrote on 2003-04-07:

> Patrick K. O'Brien wrote on 2003-04-07:
>
> > Ian Bicking <ianb at colorstudy.com> writes:
> >
> > > On Mon, 2003-04-07 at 00:23, Carl Banks wrote:
> > > > Third, there are a *ton* of functions in Emacs that operate on a
> > > > certain point in a certain buffer (delete-char, kill-region, etc.).
> > > > These kinds of functions typically get bound to keystrokes and other
> > > > sorts of commands, which a user expects to work in the current buffer.
> > > > If all these functions had to be passed the current buffer as an
> > > > argument, it would just complicate things.
> > >
> > > Maybe instead they should be methods of the buffer, and a keystroke
> > > automatically calls a method on the current buffer.
> >
> > That should be easy enough.
> >
> > > Managing so many methods from diverse sources might be difficult.  I
> > > want generic methods, I keep thinking of places where they'd be mighty
> > > useful... like Mixins but better.
> >
> > This I don't understand.  Could you elaborate?
> >
> I'm not Carl but I had similar thoughts, so I'll try :-).

s/Carl/Ian Bicking/
Sorry for the mis-attribution, I skipped one...

> ...  This bringing of many classes together is probably what Carl
> Banks talked about.

ditto.

-- 
Beni Cherniavsky <cben at tx.technion.ac.il>

Never spend several weeks on a single homework - or you will miss a
nice Linux conference ;-(.




From mk at net.mail.penguinista  Tue Apr  8 05:07:15 2003
From: mk at net.mail.penguinista (=?UTF-8?B?0LTQsNC80ZjQsNC9INCzLg==?=)
Date: Tue, 08 Apr 2003 11:07:15 +0200
Subject: cx_Oracle and Number(11)
References: <3e918724@news.mt.net.mk> <Xns9356EDF45F8E9cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <3e929143@news.mt.net.mk>


>> I have a column in the Oracle database that is defined as
>>   ORDER_ID NUMBER ( 11),
>> but when I select that table with cx_Oracle 2.5 the values in the column
>> show up as floats in the python program.
> 
> Isn't that by definition? Number turns to float, INT to INT?

hmm, what definition?


-- 
??????                                  (jabberid:damjan at bagra.net.mk)

Beware of he who would deny you access to information,
for in his heart, he dreams himself your master



From imbosol-1049649076 at aerojockey.com  Sun Apr  6 14:40:16 2003
From: imbosol-1049649076 at aerojockey.com (Carl Banks)
Date: Sun, 06 Apr 2003 18:40:16 GMT
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <kw_ja.9871$b77.3271@nwrdny02.gnilink.net>

Patrick K. O'Brien wrote:
> If you were crazy enough to think that you could create a program
> along the lines of Emacs, but written in Python, how would you go
> about it?  How would you design the domain model for files, buffers,
> windows, and frames?  How would you allow the same level of
> customizability?  How would you map functions (or methods or whatever)
> to keys?  Any thoughts?
> 
> Here is the background.  I'm working on a file editor (called
> PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> PyCrust code base.  I've got a basic version working, and it supports
> editing multiple files and all the usual, basic features.  But I want
> it to be as customizable and extensible as Emacs.  In fact, I want to
> model it after Emacs, so that elisp code could be rewritten in Python
> and work with PyAlaMode.  To do that, I've got to support the same
> primitives as Emacs; expose files, buffers, regions, and windows the
> same way Emacs does; handle the Emacs keybindings and ability to remap
> keys, the same way Emacs does, etc.
> 
> Sound like fun?


Sounds like a lot of refactoring.

Here's how I would do it.  More experienced programmers, feel free to
correct anything experience says won't work so well.

What I recommend is to read through the Emacs Lisp manual and figure
out the functions you'll want to support.  Obviously, you'll want to
make each type of object (buffer, windows, frames, positions, markers,
keymaps, etc.) a Python class.  When you read through the manual, take
the lisp functions and assign them as method of the appropriate class.
(For now, don't *write* the methods yet; just prototype them.)  This
builds a set of interfaces that the classes should follow.  The point
of this is to give you an idea of exactly what the different objects
in Emacs do, and how they interact.

Don't be too slavish in trying to match the exact behavior of the lisp
functions, and be careful to assign functions to the class they really
belong to.  Don't assign a function to the keymap class just because
it's in the keymap section.

As an example, take the section on buffers.  First function is
"current-buffer".  What class should it belong to?  Certainly not
Buffer.  The current buffer is a global property, so "current_buffer"
should be either a global function, or a method of an application
context class.

Next three functions are "save-current-buffer", "with-current-buffer",
and "with-temp-buffer".  These are actually not functions, but special
forms and macros.  Python has no analogue to these (barring some
obscure metaclass tricks).  Depending on the form, it might be useful
to have a corresponding method in Python, or it might be better to
think about a more Pythonic way to do it.  I haven't thought about it
too much, but off hand, I'd say I'd choose the latter in this case.
In fact, these seem to be mostly a convenience, so I wouldn't bother
to include them.

Next function is "buffer-name".  This obviously belongs in the Buffer
class.  I'd just call it "name".  Perhaps you want to retain the exact
name, though.  That's up to you.

Next function is "rename-buffer".  This should belong in the Buffer
class--however, in Emacs, it only operates on the current buffer.  I
think that's a design flaw; I would fix it.  I would make "rename" a
method of Buffer.

You might want to also include global "buffer_name" and
"rename_buffer" functions that operate on the current buffer.

Next function is "get-buffer".  This should be global, because the
list of buffers is global.

And so on.

Understand what I'm saying?  Eventually, you'll have a nice list of
classes waiting to be fleshed out.  It should be straightforward to do
it.

The next thing you should do it refactor your original PyAlaMode code
to use these classes.  Don't add functionality.  Just implement what
you need to get back to the original behavior of PyAlaMode, but using
the new classes.

You said the keymaps aren't editable?  Add one keymap.  Don't worry
about adding functions to edit keymaps yet.  Say you currently use a
string to hold the buffer.  Change it to use a Buffer object.

Once you've done all this, then it should be straightforward to fill
out the interfaces and add features.


There is one potentially disasterous pitfall you should be aware of.
Emacs does all kinds of strange things with namespaces.  It uses
dynamic binding, unlike Python, which uses lexical binding.  Emacs has
one monolithic global namespace, as opposed to Python, which has an
organized infrastructure of "global" namespaces, that is, modules.
Also, Emacs has weird namespaces like buffer local variables.  One
thing you'll have to do is think about how to sort all these namespace
issues out.

It seems to me that a lot of these issues won't matter much if you
have a good and consistent strategy for organizing stuff and passing
parameters around.


-- 
CARL BANKS



From ckea25d02 at sneakemail.com  Thu Apr  3 19:58:32 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 04 Apr 2003 12:58:32 +1200
Subject: When Python outruns C++ ...
In-Reply-To: <Kp2ja.2589$y3.426276@news010.worldonline.dk>
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org> <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk> <x7el4jibzl.fsf@guru.mired.org> <Kp2ja.2589$y3.426276@news010.worldonline.dk>
Message-ID: <b6ikri$5t6dn$1@ID-169208.news.dfncis.de>

Anders J. Munch wrote:
> Here's a brilliant algorithm for register allocation that I just stole
> from the PSU labs.
> 
> 1. Generate all possible combinations.
> 2. Score each of them according to estimated performance.
> 3. Pick the best scoring combination.
 >
 > Now try with 8 registers and 10 variables.  1,073,741,824 variations.

Of course, before leaking this algorithm to you
(you don't *really* believe they didn't know you
were after it, do you?) the PSU removed the final
and most important step:

   4. Use the time machine to deliver the results
      a few microseconds after the problem was submitted.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From fuka at fuxoft.cz  Mon Apr 28 20:15:12 2003
From: fuka at fuxoft.cz (Frantisek Fuka)
Date: Tue, 29 Apr 2003 02:15:12 +0200
Subject: In which versions is list.sort stable?
In-Reply-To: <mailman.1051562029.16746.python-list@python.org>
References: <mailman.1051479463.462.python-list@python.org> <mailman.1051562029.16746.python-list@python.org>
Message-ID: <b8kg8g$1d10$1@ns.felk.cvut.cz>

Can you please explain to a newbie what this thread means?

Should I not use SORT in Python 2.2.2? Docs do not mention anything 
about "unstability". Are there any other "bacis" functions (and SORT 
seems pretty basic to me) that I should avoid???

Thanks




From exarkun at intarweb.us  Sun Apr 13 19:04:33 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 13 Apr 2003 19:04:33 -0400
Subject: Protected Methods and Python
In-Reply-To: <12aeh974i1kdh.dfbwdn5e18px.dlg@40tude.net>
References: <221d8dbe.0304131058.1ecd1c3c@posting.google.com> <18lb3o2cs1se2$.isoixa5igmj6$.dlg@40tude.net> <12aeh974i1kdh.dfbwdn5e18px.dlg@40tude.net>
Message-ID: <20030413230433.GA17538@meson.dyndns.org>

On Sun, Apr 13, 2003 at 09:26:37PM +0200, Mads Orbesen Troest wrote:
> On Sun, 13 Apr 2003 21:24:24 +0200, Mads Orbesen Troest wrote:
> 
> >> Or may be it is possible to emulated Protected Methods in Python.
> > 
> > It can be "emulated" by prepending the method/attribute name with 2 
> > underscores. 
> 
> Whoops, I was too quick and didn't see that you wrote "protected" and not 
> "private".
> 
> At any rate: the mangling scheme can be used for "protected" methods, since 
> you know the mangled name is _classname__attribute/method. You can then be 
> one of the "evil" persons I talked about in the previous post, and access 
> the "private" member that way.
> 

  Double underscore prefix is for avoiding name collisions in inheritance,
not preventing people from touching things you consider private to the
implementation.

  Language enforced private attributes are a Wrong idea because they
needlessly increase the difficulty of people trying to use code which
employs this feature.

  There are a few possible scenarios:

  1) I am a skillful coder and I am using a library written by a skillful
coder.  The library is bug-free (for the purpose which I use it) and I have
no need to touch the implementation's internal variables.  In this scenario,
runtime protected private attributes are not useful.

  2) I am not a skillful coder and I am using a library written by a
skullful coder.  The library is bug-free (for the purpose which I use it)
but I feel the need to touch the implementation's internal variables anyway. 
Whether or not they are protected by the runtime environment, I will
probably find a way to screw up my application, because I am not skilled. 
In this scenario, runtime protected private attributes are not useful.

  3) I am a skillful coder and I am using a library written by someone who
is not skillful.  The library contains bugs which disrupt my application. 
In this scenario, runtime protected private attributes make it impossible
for me to work around bugs in the library, and so are worse than useless.

  Since Python provides no real runtime protection of private attributes,
the best you can hope to do is make it excessively painful for people using
your code to work around your bugs.  I can't imagine why anyone would want
to do this.

  If an attribute is private to your implementation, document it this way
and prefix it with a single underscore.  Nothing else is necessary.

  high-and-mightily y'rs,

  Jp

-- 
"There is no reason for any individual to have a computer in their
home."
                -- Ken Olson, President of DEC, World Future Society
                   Convention, 1977
-- 
 up 24 days, 19:01, 7 users, load average: 1.07, 1.20, 1.45




From irmen at -NOSPAM-REMOVE-THIS-xs4all.nl  Mon Apr 14 14:09:30 2003
From: irmen at -NOSPAM-REMOVE-THIS-xs4all.nl (Irmen de Jong)
Date: Mon, 14 Apr 2003 20:09:30 +0200
Subject: Displaying Unicode on the console (Windows)
In-Reply-To: <182bcf76.0304140744.1b3d43ee@posting.google.com>
References: <182bcf76.0304140744.1b3d43ee@posting.google.com>
Message-ID: <3e9af967$0$49112$e4fe514c@news.xs4all.nl>

Paul Moore wrote:
> OK. I know this is a common question, and I know the answer is
> basically "it's not as easy as you think" :-) But I'm confused as to
> what I need to do to get a Unicode string to print on the console in
> Windows.
> 
> To use a concrete example, I'd like to print the Euro symbol. A
> Unicode string for this is u'\20a0'
> 
> 
>>>>unicodedata.name(u'\u20a0')
> 
> 'EURO-CURRENCY SIGN'
> 

Hmmm


 >>> import unicodedata
 >>> unicodedata.name(u'\u20ac')
'EURO SIGN'
 >>> unicodedata.name(u'\u20a0')
'EURO-CURRENCY SIGN'
 >>>


I've always used 0x20ac as Euro Sign, ?.
This works nicely in all circumstances, also printing it out on the windows 
console using CP1252 encoding.

I didn't know what 0x20a0 is, so I looked it up [1],[2]:
"The new single currency for member countries of the European Monetary Union 
(EMU) is the euro. The euro character is encoded in the Unicode Standard as 
U+20AC EURO SIGN.

To avoid confusion, the historical character U+20A0 EURO-CURRENCY SIGN has 
been updated with an informative note and a cross reference to U+20AC EURO 
SIGN."

It appears to me that 0x20a0 was intented for the Euro sign at the time
the Euro was not yet a "real" currency, but only a "ECU".

--
[1]  http://www.unicode.org/reports/tr8/#Euro%20Sign
      (although this document has been superseded by a new Unicode standard,
       it pretty much explains what was going on).
[2]  http://www.unicode.org/charts/PDF/U20A0.pdf
      (A nice chart that clearly shows that 0x20a0 is *not* the Euro sign).




From richardd at hmgcc.gov.uk  Tue Apr  1 01:59:44 2003
From: richardd at hmgcc.gov.uk (Richard)
Date: Tue, 1 Apr 2003 07:59:44 +0100
Subject: Breaking out of a while loop with a key press?
References: <3e885ad5@mail.hmgcc.gov.uk>
Message-ID: <3e8938e1$1@mail.hmgcc.gov.uk>

Hi guys,

Thank you all for your suggestions. I'm sorry my question was a bit
ambiguous with regard to the 'console window' reference. I was in fact
referring to a Win2K Command Prompt (cmd.exe).

I will try catching the KeyboardInterrupt exception and check that I have no
bare 'except's anywhere that could be catching this for me.

Thanks for your help...

Richard

"Richard" <richardd at hmgcc.gov.uk> wrote in message
news:3e885ad5 at mail.hmgcc.gov.uk...
> Hi,
>
> Can anyone suggest a simple way of detecting a key press and existing the
> program? I have a program which at the moment uses a while True:
statement.
> However this is not ideal as I have to close down the console window to
> exist the program. Can anyone tell me the proper way of doing this? So
that
> a CTRL-C, for example, can quit the program correctly?
>
> Cheers
>
> Richard
>
>





From carroll at tjc.com  Fri Apr  4 22:31:06 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Fri, 04 Apr 2003 19:31:06 -0800
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com>
Message-ID: <09is8votcf3116a94euite6vhdfffgefko@4ax.com>

On Fri, 04 Apr 2003 15:00:28 -0800, carroll at tjc.com wrote:

>On 3 Apr 2003 16:51:25 GMT, Ben Hutchings
><do-not-spam-ben.hutchings at businesswebsoftware.com> wrote:
>
>>Or you can edit nntplib to add the option of storing XOVER output to a
>>file.  It doesn't look like this would be too hard.  The 'file' could
>>be an object that processes each line as it comes, rather than a real
>>file.  Then send your change to nntplib back to the author and hope it's
>>accepted into the official version.
>
>That's not a bad idea.  David, you should consider taking this
>approach.

David --

I've made an minor modification to nntplib to do this.  So far, I've
modified only the xover method, but I intend to change all the
commands that return lists, and submit it as a proposed change.  Let
me know if you'd like a copy.

Original method works as follows:

   xover(start, end)
       Return a pair (resp, list). list is a list of tuples, one for
       each article in the range delimited by the start and end
       article numbers. Each tuple is of the form (article number,
       subject, poster, date, id, references, size, lines). This is an
       optional NNTP extension, and may not be supported by all
       servers.

Modified method works as follows:

   xover(start, end, [file])
       Return a pair (resp, list). list is a list of tuples, one for
       each article in the range delimited by the start and end
       article numbers. Each tuple is of the form (article number,
       subject, poster, date, id, references, size, lines).  

       If the file parameter is specified, then the output is stored
       to a file rather than to list.  If file is a string, then the
       method will open a file object with that name, write to it, 
       then close it.  If file is a file object, then it will start
       calling write() on it to store the lines returned by XOVER.  If
       file is supplied, then the returned list is an empty list.
  
       This is an optional NNTP extension, and may not be supported
       by all servers.

   



From magnus.therning at wanadoo.nl  Mon Apr 21 14:16:46 2003
From: magnus.therning at wanadoo.nl (Magnus Therning)
Date: Mon, 21 Apr 2003 18:16:46 GMT
Subject: printing email parts of type message/external-body??
References: <ZKzoa.602320$sj7.24103068@Flipper> <3EA2CFF2.1440294E@engcorp.com>
Message-ID: <iAWoa.649943$sj7.26068978@Flipper>

Peter Hansen <peter at engcorp.com> wrote:
> Magnus Therning wrote:
>> 
>> I am having problems with the following piece of code, would appreciate
>> any insight that can be offered!
>> 
>> ---- test.py ------------
>> from email.Message import Message
>> 
>> meb_message = Message()
>> meb_message.add_header('Content-Type', 'message/external-body',
>>     access_type='local-file')
>> meb_message.set_payload('name="test"\n')
>> print meb_message
>> -------------------------
>> magnus at die:[test].$ python test.py
>> 
>> Traceback (most recent call last):
>>    g.flatten(msg.get_payload(0), unixfrom=False)
>>  File "/usr/lib/python2.2/email/Message.py", line 179, in get_payload
>>    raise TypeError, i
>> TypeError: 0
> 
> Almost looks like a bug.  According to email.Message's documentation,
> a TypeError will be raised by get_payload() if an index parameter is
> provided but the payload is a string, as in this case.

But no index is given, so that would indicate a bug, right?

Well, unfortunately it doesn't bring me any closer to a solution :( Are
there no pyton based MUAs that would have run into this problem already?

/M

-- 
Magnus Therning                    (OpenPGP: 0x6A83A7DF)

Discovery consists of seeing what everybody has seen and thinking what
nobody has thought.
     -- Albert Szent-Gyorgyi



From spam at magnetic-ink.dk  Tue Apr  1 12:49:02 2003
From: spam at magnetic-ink.dk (Klaus Alexander Seistrup)
Date: Tue, 1 Apr 2003 17:49:02 +0000 (UTC)
Subject: Comments
References: <435d40e0.0304010014.503f9f8f@posting.google.com> <d2c6a353-612a-4f45-8ddd-bddc04d59e6a@news.szn.dk> <slrnb8jfsq.9n8.philh@cabalamat.uklinux.net>
Message-ID: <41d02ffe-a3d3-4bde-9466-d4b53309242a@news.szn.dk>

Phil Hunt wrote:

>>#v+
> 
> What does this mean?

That's the beginning of a set of "verbatim marks" for the JED editor.

> Or just:
> 
> """
> This is a comment
> """

Right, no assignment to a dummy variable needed.


  // Klaus

-- 
 ><> 	unselfish actions pay back better



From lucio at movilogic.com  Tue Apr  1 14:17:29 2003
From: lucio at movilogic.com (lucio)
Date: Tue, 01 Apr 2003 16:17:29 -0300
Subject: Terminal-based text editor with color syntax hilighting
In-Reply-To: <l9di+kKkXgKL092yn@gnosis.cx>
References: <slrnb8hvka.pki.ramen@lackingtalent.com> <mailman.1049169002.28920.python-list@python.org> <slrnb8j93q.421.ramen@lackingtalent.com> <l9di+kKkXgKL092yn@gnosis.cx>
Message-ID: <3E89E5C9.8050508@movilogic.com>

Lulu of the Lotus-Eaters wrote:

>I *have*
>been using vim on one host recently, and have kinda-sorta learned basic
>navigation, but it won't stick in my head after a month (I know from
>experience).  Likewise with emacs.
>
>Jed has this wonderful feature of having topbar menus that actually let
>you visually select the most common actions (without needing to remember
>a cryptic key-sequence).  This fabulous design had only been in DOS
>text editors since 1982, so it's nice that Linux is starting to catch up
>with that.
>  
>
i dont consider adding text menus to vim to catch up. Vim is very good 
at what it is.

Rant aside, i use like vi because:
- I like the command/edit modes, very helful for fast editing
- i like the portability, i use it on windows and linux
- it has gvim, which i use when runing X or windows. there you have 
menus and all the stuff.

if you learn to use vim, you can take it everywhere with you. The only 
yhing missing (i think) is code autocompletion (it has command 
autocompletion).

Lucio, in sunny vi paradise.





From peter at engcorp.com  Sat Apr 26 15:30:03 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sat, 26 Apr 2003 15:30:03 -0400
Subject: Some Python Code
References: <b8zqa.5100$Zj2.1085711@news20.bellglobal.com>
Message-ID: <3EAADE3B.6ABB4FE6@engcorp.com>

Sean Ross wrote:
> 
> I've implemented a general purpose genetic algorithm
> package and a little circuit description programming language, plus a few
> other AI related items of differing quality. I don't really want to do any
> maintenance on these projects, so I'll not be bothering to put them up in
> some place like sourceforge or pypi. 

Maybe there's something findable via http://uselesspython.com/sourcelinks.html
which would give you a place to dump the code.

-Peter



From creedy at mitretek.org  Tue Apr  8 11:13:17 2003
From: creedy at mitretek.org (Chris Reedy)
Date: Tue, 08 Apr 2003 11:13:17 -0400
Subject: How to protect python code ?
In-Reply-To: <3E921A4C.1020908@hotmail.com>
References: <3E8BE455.4000904@novagrid.com> <mailman.1049389032.4904.python-list@python.org> <u8bka.651$001.591065@monger.newsread.com> <mailman.1049722991.18682.python-list@python.org> <3E921A4C.1020908@hotmail.com>
Message-ID: <3e92e4ba$1_2@corp-news.newsgroups.com>

msw wrote:
> [snip] 
> 
> The software I write falls under the category of those that need to have 
> the source code hidden from customers - for many reasons, not least of 
> which is "because the marketing department says so".  Our customers are 
> unlikely to go after "movl $0x0007, %ax" with a hex editor, but are 
> almost certain to mess with code that looks like "licenseCount = 5".
> 
Given that no source code obfuscation system can provide absolute 
protection, and given that your objective is to keep the user from 
messing with the code, I suggest the following program:

1. Change the name of all local variables to _L1, _L2, etc.

2. Change the name of all parameters, including keywords (that may get a 
little tricky) for functions and methods that are not part of the 
external interface to _P1, _P2, etc.

3. Change the name of all functions and methods that are not part of the 
  external interface to _F1, _F2, etc.

4. Change the name of all classes that are not part of the external 
interface to _C1, etc.

5. Change the name of all module globals that are not part of the 
external interface and not already covered to _G1, etc.

6. Change the name of all packages and modules that are not part of the 
external interface to M1, M2, etc.

7. Remove all comments.

8. Remove all testing code.

9. Remove documentation strings for objects that aren't part of the 
external interface.

10. And anything else I forgot ...

At this point your code is sufficiently obfuscated that no one in their 
right mind would want to touch it. More to the point, anyone that is 
willing to devote that time and effort to understanding this code is 
probably going to penetrate any obfuscation scheme you care to design. 
That is, this will achieve your goal, you will keep the casual user from 
attempting to modify the code.

   Chris

-- 
This is informal and not an official Mitretek Systems position.
Dr. Christopher L. Reedy, Senior Principal Software Engineer
Mitretek Systems, 3150 Fairview Park Drive South, Falls Church, VA 22042
Email: creedy at mitretek.org  Phone: (703) 610-1615  FAX: (703) 610-2203




From martin at v.loewis.de  Fri Apr 25 08:04:09 2003
From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=)
Date: Fri, 25 Apr 2003 14:04:09 +0200
Subject: PyXML - setContentHandler??
In-Reply-To: <3EA91EA8.2050505@ncsa.uiuc.edu>
References: <3EA91EA8.2050505@ncsa.uiuc.edu>
Message-ID: <b8b87m$rs4$01$1@news.t-online.com>

Randy Heiland wrote:

> I'm using a chunk of code that does:
> 
>  >>> parser = xml.sax.make_parser()
> <define a handler>
>  >>> parser.setContentHandler(handler)
> 
> however, using PyXML, I get:
> AttributeError: SAX_expat instance has no attribute 'setContentHandler'

Where do these come from? I.e. what is xml.sax.__file__, and
parser.__class__.__module__?

Regards,
Martin





From db3l at fitlinxx.com  Thu Apr  3 10:37:18 2003
From: db3l at fitlinxx.com (David Bolen)
Date: 03 Apr 2003 10:37:18 -0500
Subject: else: w/o except: - why not?
References: <mailman.1049211247.3694.python-list@python.org>
Message-ID: <uvfxvtwwh.fsf@fitlinxx.com>

Michael Chermside <mcherm at mcherm.com> writes:

> The other situation is running "untrusted plug-in" code of some 
> sort. Something like this:
> 
>     try:
>         randomCodeSomebodyUploaded()
>     except:
>         logProblem()
> 
> But really, that hardly EVER comes up. I wish that guido had made
> it harder to spell "bare except" because these situations seem so
> unusual. The next best solution is to just avoid it unless you're
> SURE you need it.

I think you need to be careful about blanket statements like "hardly
EVER" - it really depends on the sort of target environment being
worked on.  In our case for example, a large amount of our Python code
deals with distributed systems management and communications, and
there are quite a few cases where we have bare excepts in order to
catch any exceptions to be packaged and either reported (as part of a
plugin or dynamic module system as you show above), propagated to some
client, or just logged and execution resumed.  In such cases I
consider it a perfectly good programming practice and in fact one of
the conveniences of Python to ensure an application stays operational.

Now, in other cases with applications that are more standalone or
non-infrastructure related, it is in fact much rarer to need such a
construct, although even there I'll often have at least one such case
per production application near the uppermost code levels as a general
purpose trap to log unexpected failures but permit the application to
resume operation at some initialized point.

To be honest, the only thing we've had to be careful of in such cases
is with respect to the prior comments about possibly not wanting to
catch KeyboardInterrupt or SystemExit.  But we've found it easier to
check and re-raise those in the bare except clause, than to try to
come up with a qualified except clause that includes all other
exceptions.  We did try using "except Exception" instead but that
won't catch older string exceptions which are still used by some
modules we use.

--
-- David
-- 
/-----------------------------------------------------------------------\
 \               David Bolen            \   E-mail: db3l at fitlinxx.com  /
  |             FitLinxx, Inc.            \  Phone: (203) 708-5192    |
 /  860 Canal Street, Stamford, CT  06902   \  Fax: (203) 316-5150     \
\-----------------------------------------------------------------------/



From max at alcyone.com  Sun Apr 27 01:26:44 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sat, 26 Apr 2003 22:26:44 -0700
Subject: Forgetting "()" when calling methods
References: <b8cadp$1fj9$1@ns.felk.cvut.cz> <698f09f8.0304262115.386c2571@posting.google.com>
Message-ID: <3EAB6A14.707AD826@alcyone.com>

Jeremy Fincher wrote:

> Perhaps the answer is to remove the __nonzero__ method on
> functions/methods.  Logically, what purpose does it serve?

	# f holds a callable object or None
	if f:
	    f(...)

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Death to all fanatics!
\__/ Malaclypse the Younger
    Kepler's laws / http://www.alcyone.com/max/physics/kepler/
 A proof of Kepler's laws.



From tim.one at comcast.net  Wed Apr  9 17:57:09 2003
From: tim.one at comcast.net (Tim Peters)
Date: Wed, 09 Apr 2003 17:57:09 -0400
Subject: Odd behaviour of dir(object) with metaclasses.
In-Reply-To: <b71toc$k7m$1@main.gmane.org>
Message-ID: <BIEJKCLHCIOIHAGOKOLHGELAFEAA.tim.one@comcast.net>

[Dirk Gerrits]
> I am a bit puzzled by the behaviour of the dir builtin.
> ...
>
> Perhaps I just don't fully understand what dir does. The first bit in
> the docs says:
>
>      Return an alphabetized list of names comprising (some of) the
>      attributes of the given object, and of attributes reachable from it:
>
> This would seem to correspond with my feeling of what its supposed to
> do, but doesn't. However, reading further, we see that the current
> behaviour is indeed as it is documented:
>
>      No argument:  the names in the current scope.
>      Module object:  the module attributes.
>      Type or class object:  its attributes, and recursively the
>          attributes of its bases.
>      Otherwise:  its attributes, its class's attributes, and recursively
>          the attributes of its class's base classes.
>
> Why is there an unnatural (IMHO) distinction between type/class objects
> and other non-module objects? I assume that a class's class is not
> considered become dir comes from the pre-metaclass era?

No, dir() was rewritten from scratch for 2.2.  The comments in the C code
say

	/* Elif some form of type or class, grab its dict and its bases.
	   We deliberately don't suck up its __class__, as methods belonging
	   to the metaclass would probably be more confusing than helpful. */

That's what Guido believed at the time, and I agreed at the time, so that's
what the code does.  Across at least seven pre-releases of 2.2, nobody
disagreed <wink>.

> Shouldn't the docs really read
> ...

No, becuase the suggested rewording doesn't describe the actual behavior --
as you noted above, the current wording does.

> ...
> I hope you'll be able to clarify these things for me a bit.

dir() was meant to be a convenience in interactive mode, and that's all.
Because of that, no two people will agree on what it "should do" in all
respects in all cases.  Changing it always makes people scream too, so I
doubt it will be changed again.  If you don't like what it does, write a dir
that does what you like -- doing so in Python is easy enough.





From aleax at aleax.it  Fri Apr 18 09:53:30 2003
From: aleax at aleax.it (Alex Martelli)
Date: Fri, 18 Apr 2003 13:53:30 GMT
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr> <20030417155900.A10011@tibia.amotlpaa.bogus> <mailman.1050663512.4067.python-list@python.org>
Message-ID: <urTna.32583$LB6.738590@news1.tin.it>

Donnal Walter wrote:

> Steven Taschuk wrote:
>> Quoth Jean-Paul Roy:
>>>- Python is bad : tail recursion is not iterative (also astonished at
>>>that, I can understand with an interpreter, but compiler ?)
>> 
>> 
>> For teaching purposes, imho tail recursion is a Bad Thing.
>> 
>> Writing an iterative process using tail recursion has always
>> struck me as a clever but obfuscated technique.  Write what you
>> mean: if you want an iterative process, write a loop; if you want
>> a recursive process, make recursive calls.
> 
> What is the difference between "recursive calls" and "tail recursion"?
> Sometimes I have written methods that call the same method on a parent
> or child object, with the depth of recursion being limited by a method
> of the same name NOT calling itself. In the past I have thought of this
> as tail recursion. Is this incorrect?

A function-call is "tail" if it's the last thing the caller does,
i.e., basically, if the caller executes "return somefunction(args)"
or a semantically equivalent sequence.

A function-call is "recursive" if it's a call of the function to
itself (there's also "mutual recursion", where two functions call
each other).

A function-call is "tail recursive" if it's tail and recursive.
Such calls can be optimized by avoiding the allocation of a new
stack frame -- the same frame object that's already on the
stack to handle this specific call can be re-used (if you're
willing to lose traceback information -- Python currently does
not allow that).

Your example is not recursive: the fact that the various functions
being called have the same name is accidental, as they're different
functions anyway.


Alex




From gh at ghaering.de  Wed Apr 30 03:30:48 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Wed, 30 Apr 2003 09:30:48 +0200
Subject: plugin system
In-Reply-To: <WHIra.27759$su3.3307778@twister.nyc.rr.com>
References: <WHIra.27759$su3.3307778@twister.nyc.rr.com>
Message-ID: <3EAF7BA8.70605@ghaering.de>

xam wrote:
> I am interested in writing a plugin system where the main app is python but
> it is capable of accepting helper "plugins" which can be either python or
> C++. I guess I want some sort of an interface spec for the plugins that can
> be implemented in either language. Which technologies should I research? Are
> there other projects that I can refer to as a model?

There could be a plugins/ directory where each module is imported from. 
  Python modules can be written in Python (obviously) and C++. Problem 
solved, or not?

Python itself only has a C API for writing extension modules, but there 
are several C++ wrappers that can make your life much easier, such as 
Boost::Python.

-- Gerhard




From uselpa at myrealbox.com  Sat Apr 26 05:40:12 2003
From: uselpa at myrealbox.com (Patrick Useldinger)
Date: Sat, 26 Apr 2003 11:40:12 +0200
Subject: zipfile with umlaut in filename
In-Reply-To: <b8caue$g65$05$3@news.t-online.com>
References: <3ea9a4bf$1_1@news.vo.lu> <b8caue$g65$05$3@news.t-online.com>
Message-ID: <3eaa52b1_1@news.vo.lu>

Martin v. L?wis wrote:
>> Hello everybody,
>> I have a problem when creating an archive using zipfile. If the name 
>> of the file to be added contains a special character like an umlaut, 
>> this character gets replaced by another character.
>> What can I do against this?
> 
> I'm not sure you can do anything. Zipfiles only support ASCII file 
> names, to my knowledge.

Well, doing the same thing with WinZip works ok and keeps the name as it 
should.
Also, reading *that* archive back in Python yields, as a result of 
namelist():
['Der h\x94fische Charakter der Liebe.doc']
for a file called
'Der h?fische Character der Liebe'.

-PU




From max at alcyone.com  Wed Apr 23 17:00:37 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 23 Apr 2003 14:00:37 -0700
Subject: python language: infimum and supremum of integers
References: <mailman.1051093416.23728.python-list@python.org>
Message-ID: <3EA6FEF5.7B764ECB@alcyone.com>

Gr?goire Dooms wrote:

> I need a lower and higher integer for initializing lower and higher
> bounds in an algorithm.  I know integers have no limits(maximum and
> minimum) in python and infimum and supremum would do the job too.

Well, to start with, there's sys.maxint.  If you know you're dealing
with 2's complement integers, then minint would be -sys.maxint - 1.

> I empirically discovered (Python 2.2.2) that None could be considered
> an
> infimum and any string could be considered a supremum.
> e.g "a">int(x) is always true and None<int(x) is always false
> 
> But this fact is not stated in the language reference.

The language says that the ordering between objects of (some) different
types will be consistent in an implementation, but will be no be
guaranteed from implementation to implementation.  So in your
implementation, None might be less than all integers, but in another
implementation it might not be.

The best solution to your problem is probably what Alex suggested; make
a set of custom types that always compares less (or greater) than any
other object, and use that as your sentinel.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ If I had known, I would have become a watchmaker.
\__/ Albert Einstein
    7 Sisters Productions / http://www.7sisters.com/
 Web design for the future.



From sismex01 at hebmex.com  Mon Apr 14 16:20:59 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Mon, 14 Apr 2003 15:20:59 -0500
Subject: Displaying Unicode on the console (Windows)
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F0623304597725@mail.hebmex.com>

> 
> Paul Moore <gustav at morpheus.demon.co.uk> writes:
> 
> > PS Is there any way of forcing the Windows 2000/XP console 
> > to display Unicode in any form (UTF-8, maybe) rather than
> > having to deal with code pages???
> 
> Yes, through WriteConsoleW.
> 
> Regards,
> Martin
>

It accepts UTF-8?  Or only 16bit chars?

-gca




From a.schmolck at gmx.net  Sun Apr 13 13:01:38 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 13 Apr 2003 18:01:38 +0100
Subject: intern'ed strings and deepcopy()
References: <ef08387c.0304111042.73a16e68@posting.google.com> <3E972D34.6B829FCC@alcyone.com> <yfsel48fvs6.fsf@black132.ex.ac.uk> <3E97584A.5E7FD112@alcyone.com> <yfs8yugfsl0.fsf@black132.ex.ac.uk> <3E977928.B3A6DC7E@engcorp.com> <yfsznmueec1.fsf@black132.ex.ac.uk> <m3he92pef3.fsf@mira.informatik.hu-berlin.de>
Message-ID: <yfsr886e3gt.fsf@black132.ex.ac.uk>

martin at v.loewis.de (Martin v. L?wis) writes:

> If you meant that 
> 
> type(a) == type(b) == str and b == a 'implies' intern(b) is intern(a)
> 
> then yes: Python does guarantee that, for arbitrary a,b.

Yes, this is of course what I meant (I apparently got tripped over a bit in
the process of translating my mentalese into python :)

> 
> Notice that Python does *not* (anymore) guarantee
> 
> type(a) == type(b) == str and b == a 'implies' id(intern(b)) == id(intern(a))

Huh? Really?

I thought  ``id(EXP1) == id(EXP2)`` implies ``EXP1 is EXP2`` and that's the
only difference I can see between the two lines.

So does this no longer hold true in python 2.3: "The `is' operator compares
the identity of two objects; the id() function returns an integer representing
its identity"? (the 2.3a2 reference openly admits its own inaccuracy in
general, but doesn't really seem to shed further light on this issue in particular).

> The intention is that intern gives you a unique distinct
> representative from the set of equal strings.

Yep, that's what I thought it would be :)

'as



From gh at ghaering.de  Tue Apr 15 20:25:32 2003
From: gh at ghaering.de (Gerhard =?iso-8859-1?Q?H=E4ring?=)
Date: Wed, 16 Apr 2003 02:25:32 +0200
Subject: do python's nifty indentation rules spell the death of
	one-liners?
In-Reply-To: <87he90b8rp.fsf@jidanni.org>
References: <87he90b8rp.fsf@jidanni.org>
Message-ID: <20030416002531.GA22395@mephisto.ghaering.test>

* Dan Jacobson <jidanni at dman.ddts.net> [2003-04-15 13:59 +0800]:
> I need to use one liners every day, especially in Makefiles.  I thus
> need to know how to make any kind of python program into a big long one
> liner. However, right out of the starting gates I run into
> $ python -c 'print 2;for i in (1,4):print i'
>   File "<string>", line 1
>     print 2;for i in (1,4):print i
>               ^
> SyntaxError: invalid syntax

Try:

python -c """
print 2
for i in (1, 4):
    print i
"""

It seems to work under a bash script at least. I once did something with
<<EOF but have since forgotten the little knowledge I had a about shell
scripting ;-)

> [...] If I found a bug, report it for me.

If you really find a bug *you* can report it with the Python bug tracker
at Sourceforge.

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/




From andymac at bullseye.apana.org.au  Fri Apr 25 18:58:41 2003
From: andymac at bullseye.apana.org.au (Andrew MacIntyre)
Date: Sat, 26 Apr 2003 09:58:41 +1100 (edt)
Subject: FreeBSD: Python exits on signal 11
In-Reply-To: <b8bpst$43q$03$1@news.t-online.com>
Message-ID: <Pine.OS2.4.44.0304260951350.28674-100000@tenring.andymac.org>

On Fri, 25 Apr 2003, Harald Schneider wrote:

> I get this error on FreeBSD5 with Python 2.2.2, when I run an application
> that does its job without errors on NT and Linux platform. It uses threads
> and sockets as well (xqmlrpclib). It just stops after a while, displays this
> error and does a core dump.

Keep in mind that FreeBSD 5.0 is effectively a pre-release, especially as
the threads support has been re-written to use the new KSE facilities,
and is thus still relatively immature.  If your code runs on FreeBSD 4.8
Ok, I suggest that you need to take this to a FreeBSD forum.

Last I heard, the FreeBSD team don't expect 5.x to become the stable
branch until 5.2; until then the 4.x branch continues as the stable
branch.

--
Andrew I MacIntyre                     "These thoughts are mine alone..."
E-mail: andymac at bullseye.apana.org.au  | Snail: PO Box 370
        andymac at pcug.org.au            |        Belconnen  ACT  2616
Web:    http://www.andymac.org/        |        Australia





From vf at interna.ru  Wed Apr 30 09:26:40 2003
From: vf at interna.ru (vadim)
Date: 30 Apr 2003 06:26:40 -0700
Subject: How to call built in function 'open' from inside object that has 'open' method?
Message-ID: <db81c492.0304300526.7b8d3c86@posting.google.com>

I am new to Python and do not know the answer to very
simple how to 'open' from inside another object that
has 'open' method. I know that in C++ I should use
following syntax ::open. 

More details on reasons. I have met this problem trying to 
write a simple script use ActiveScripting  inside IE.
I am trying "file = open(...) but because context window
object has 'open' method that just opens window it is called
first. Sorry for simple question. I was trying to find 
the answer but either it is not on the surface or too simple.



From dan at telent.net  Thu Apr 17 08:39:58 2003
From: dan at telent.net (Daniel Barlow)
Date: Thu, 17 Apr 2003 13:39:58 +0100
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr>
 <m27k9vfcz3.fsf@mycroft.actrix.gen.nz>
 <3E9DE024.3020807@removethis.free.fr> <yfsu1cy9luw.fsf@black132.ex.ac.uk>
 <3E9EB3E9.1040208@removethis.free.fr>
Message-ID: <87r8818fhd.fsf@noetbook.telent.net>

laotseu <bdesth at removethis.free.fr> writes:

> Ok, cross-posted and fu2 positionned.

Followup reset.  Lisp users already know the answer to this one; if
Python users also want to, it should go to comp.lang.python.

> Lispers friends, let's have you're opinion on that point : is Common
> Lisp a functionnal language ?

Common Lisp is a multiparadigm language.  If the question is "Does
Common Lisp support a functional style of programming", the answer is
"yes".  If the question is "Does Common Lisp _require_ a functional
style of programming", no, absolutely not.

And to say that CLOS is built "on top of" CL, though historically
accurate, is no longer really correct.  CLOS is _part_ of ANSI Common
Lisp.


-dan

-- 

   http://www.cliki.net/ - Link farm for free CL-on-Unix resources 



From Gareth.McCaughan at pobox.com  Mon Apr 28 04:52:43 2003
From: Gareth.McCaughan at pobox.com (Gareth McCaughan)
Date: Mon, 28 Apr 2003 09:52:43 +0100
Subject: Forgetting "()" when calling methods
References: <mailman.1051497342.21537.python-list@python.org> <ge5ra.13872$3M4.378491@news1.tin.it>
Message-ID: <slrnbapqur.on.Gareth.McCaughan@g.local>

Alex Martelli wrote:

[Tim Peters:]
> > OTOH, how many would have scored 100% if also asked what the result of
> > sorting would be?  I'm not sure anyone could have.  For years, ints came
> > out "less than" lists, which in turn came out "less than" longs, so that
> > 
> >     [10, [9], 7L, 8]
> > 
> > was considered to be "in sorted order" as-is.  It was something like 7
> > years before someone finally noticed how nuts that was.

[Alex:]
> *Blink* -- I don't get it.  Even if int < list < long, how CAN the
> two int's here be ``considered in sorted order''???

a0 <= a1 <= ... <= a(n-1). The old ordering wasn't
transitive. That was indeed nuts, of course.

> Here is a typical use case: prettyprinting a dictionary.
...
> Oops!  We can't prettyprint it any more -- kaBOOM.
...
> However, it seems totally arbitrary to me that the standard
> library's pprint.pprint should raise an exception just because 
> the user's trying to prettyprint a dictionary with some complex
> numbers as keys.  I guess one could class this as a bug in
> pprint, but from my POV what it does is highlight a typical
> case in which the inability to ask "just sort this stuff so
> it displays in an order that doesn't appear totally random"
> *hurts*, and hurts pretty badly too.
> 
> What's the ADVANTAGE gained, worth breaking prettyprinting for?
> 
> It seems to me it's basically one of "purity" -- that 1.5.2
> was far more pragmatic in this regard.  I know I'm not gonna get
> the pragmatism back, sigh, but is it SO hard to understand why
> I would want it?!

Hear, hear. The decision to make comparing complex numbers
throw an exception never made sense to me. (And I'm a
mathematician; believe me, I understand the appeal of
purity.)

Here's another example: regression testing. For that,
you want a sort order that's guaranteed not to change
across runs either, though it's probably OK if it
changes with new versions of Python.

The key question, to which I've never seen a convincing
answer, is threefold:

  1 What can you do more easily in Python as a result
    of complex comparisons throwing exceptions?

  2 What actual bugs are prevented by having complex
    comparisons throw exceptions?

  3 How is Python made easier to learn or understand
    by having complex comparisons throw exceptions?

It seems to me that the answers are "nothing", "none"
and "it isn't", respectively. I'm open to persuasion.

-- 
Gareth McCaughan  Gareth.McCaughan at pobox.com
.sig under construc



From exarkun at intarweb.us  Thu Apr 17 09:43:47 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 17 Apr 2003 09:43:47 -0400
Subject: How many of you are Extreme Programmers?
In-Reply-To: <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk>
References: <mailman.1050512881.28736.python-list@python.org> <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk>
Message-ID: <20030417134347.GC25373@meson.dyndns.org>

On Thu, Apr 17, 2003 at 02:37:20PM +0100, Robin Becker wrote:
> In article <3E9EA707.4DB53582 at engcorp.com>, Peter Hansen
> <peter at engcorp.com> writes
> >> This sounds remarkably similar to what we commonly call genetic
> >> optimisation or evolution strategy selection. There a fitness criterion
> >> is applied to randomly selected elements of a pool of solutions, with
> >
> >The key is that XP is _not_ applying "randomly selected elements".
> >The experience and abilities of the programmer are heavily involved
> >in the process.
>
> interestingly genetic programming et al seem to perform worse when
> attempts are made to 'direct' the evolution. 
>

  Only if you direct the evolution in the wrong way >:)

  Jp

> -- 
> Robin Becker
> -- 
> http://mail.python.org/mailman/listinfo/python-list

-- 
It is practically impossible to teach good programming style to
students that have had prior exposure to BASIC: as potential
programmers they are mentally mutilated beyond hope of
regeneration.        -- Dijkstra
-- 
 up 28 days, 9:02, 5 users, load average: 0.04, 0.03, 0.00




From john.abel at pa.press.net  Tue Apr 29 08:50:05 2003
From: john.abel at pa.press.net (John Abel)
Date: Tue, 29 Apr 2003 13:50:05 +0100
Subject: Passing Named Paramater To A C Func?
In-Reply-To: <b8lm5d$4tl$1@main.gmane.org>
References: <3EAE57E5.5010202@pa.press.net> <b8lm5d$4tl$1@main.gmane.org>
Message-ID: <3EAE74FD.2000208@pa.press.net>

Hi,

Thanks for that, its pointed me in the right direction.

Regards

John

Fredrik Lundh wrote:

>John Abel wrote:
>
>  
>
>>I'm writing a C function, which will have a Python wrapper, and I wish
>>to be able to pass it an optional, named parameter.  For example:
>>
>>def funcName( dirName, fileName=None ):
>>
>>   if fileName is None:
>>       Some Other Stuff
>>
>>   cCall( dirName, fileName )
>>
>>Is the best to handle this in the Python wrapper, or is there a way of
>>handling these via C.  None of the extending documents seem to mention
>>anything about this.
>>    
>>
>
>to handle named arguments on the C level, use ParseTupleAndKeywords.
>here's an example, taken from the regular expression engine:
>
>static PyObject*
>pattern_match(PatternObject* self, PyObject* args, PyObject* kw)
>{
>    ...
>
>    PyObject* string;
>    int start = 0;
>    int end = INT_MAX;
>    static char* kwlist[] = { "pattern", "pos", "endpos", NULL };
>    if (!PyArg_ParseTupleAndKeywords(args, kw, "O|ii:match", kwlist,
>                                     &string, &start, &end))
>        return NULL;
>
>you also have to set the METH_KEYWORDS flag in the PyMethodDef
>table:
>
>static PyMethodDef pattern_methods[] = {
>    {"match", (PyCFunction) pattern_match, METH_VARARGS|METH_KEYWORDS},
>    ...
>
>:::
>
>in your case, I'm not sure it's really worth the effort, since you
>need to look at the argument at the Python level anyway.
>
></F>
>
>
>
>
>  
>





From abpillai at lycos.com  Thu Apr 17 09:34:11 2003
From: abpillai at lycos.com (Anand B Pillai)
Date: 17 Apr 2003 06:34:11 -0700
Subject: Python Thread Question
Message-ID: <bd993a2f.0304170534.245aa2d5@posting.google.com>

Hi Pythonistas

 I have written this application which is a kind of intranet
 web-spider. It crawls a given url and retrives the files in
 the url and saves it to the disk. 

 Now when I do this using multiple threads(python threads), 
 assigning each url to a thread I find that the download gets
 completed faster than if it were in a single thread. I assume
 that the reason for this must be simple, that when you use 
 a single thread idiom, the app has to wait till a file is 
 downloaded. Whereas if you use a thread for each download, 
 the app can spawn other threads for other downloads, so no 
 wait is needed. I am firing off a group of threads (limited
 by a maxthread count) and pooling them in a threadgroup. 
 Once the threads are fired for download, the app does not
 try to control them until they finish/killed or a network 
 time-out occurs.

 Ideally speaking, multithreading need not improve the speed
 of an application but in examples like this which involve
 bottlenecks like network traffic, it does. My questions about
 this are:

 1. Does python threads work only if the native platform supports
    threading ?  i.e, is python firing 'C' threads which in turn
    fire the platform API threads (Win32 for windows/ pthreads for
    linux etc)?
 2. Can a software API (Win32/pthreads) do multithreading even if
    the CPU does not support multithreading ? (might seem like a
    superfluous question when almost all cpus does in this age, but
    the question is still valid). Or is multithreading ultimately
    related to how the CPU handles threads ?
 3. Is the apparent increase in speed in my program using multiple
    threads attributable to the CPU or the platform API or python ?

 4. Can I safely say that multithreading will improve my application
    performance if it has similar work to do on many resources at the
    same time ? (egs: a web parser/ spider/ a disk-to-disk file copier/
    directory synchronizer) Or does it depend upon the nature of the
    task at hand ?

Well, that is quite a lot.

Thanks for your help,

Regards

Anand Pillai



From h_schneider at marketmix.com  Fri Apr 18 10:22:27 2003
From: h_schneider at marketmix.com (Harald Schneider)
Date: Fri, 18 Apr 2003 16:22:27 +0200
Subject: SOLVED: py2exe Registry Access from NT Service ...
References: <b7lr3g$b3q$06$1@news.t-online.com> <y929gt00.fsf@python.net> <b7mbnb$pp2$06$1@news.t-online.com> <8yu9gpnb.fsf@python.net> <b7mf9u$2jb$03$1@news.t-online.com> <b7mjlc$6tm$03$1@news.t-online.com>
Message-ID: <b7p1mq$i2t$02$1@news.t-online.com>

Strange ... I changed the order of the import statements from

import win32serviceutil,win32service,win32event,pywintypes
import os,string,rpcserver,xmlrpclib,confighandler
from agent import *
from setuptool import RegistryTool

to

from agent import *
from setuptool import RegistryTool
import win32serviceutil,win32service,win32event,pywintypes
import os,string,rpcserver,xmlrpclib,confighandler

and not it works!

The registry access is done in a class contained in the agent module. Under
Python 2.1 the order played no role, under 2.2 it does!

All the best,
Harald


"Harald Schneider" <h_schneider at marketmix.com> schrieb im Newsbeitrag
news:b7mjlc$6tm$03$1 at news.t-online.com...
> Running the service framework with
>
> if __name__ == '__main__':
>     win32serviceutil.HandleCommandLine(Service)
>
> also works without problem. When I freeze the whole thing, it ceases to
> function ... :-/
>
> --Harald
>
>
> "Harald Schneider" <h_schneider at marketmix.com> schrieb im Newsbeitrag
> news:b7mf9u$2jb$03$1 at news.t-online.com...
> > It's in HKEY_LOCAL_MACHINE.
> >
> > The registry calls are in another module, which is imported. The
registry
> > calls work, when I move them to the Service framework. So this seems to
be
> > an import problem which only happens on 2.2 ...
> >
> > > What happens when you run the service as not-py2exe'd service (via
> > > PythonService.exe)?
> >
> > It just says
> > Could not start the service - error 203
> > Registering the Python Service Manager..
> >
> > Very strange ..
> >
> > --Harald
> >
> >
> >
> > "Thomas Heller" <theller at python.net> schrieb im Newsbeitrag
> > news:8yu9gpnb.fsf at python.net...
> > > "Harald Schneider" <h_schneider at marketmix.com> writes:
> > >
> > > > No permission probs, logged in as Admin. Platform is Win2K, not
> updated
> > from
> > > > NT.
> > > > There are no exceptions. Reading a registry key just claims it would
> not
> > > > exist. When I run the same code outside the service wrapper, I can
> read
> > the
> > > > key.
> > > >
> > >
> > > Strange.
> > >
> > > Do you mean *you* are logged in as Admin, or does the *service* run
> under
> > > the Admin account?
> > >
> > > Per default a service runs under the system account. Maybe the
registry
> > > keys are really not there then? HKEY_CURRENT_USER, probably?
> > >
> > > What happens when you run the service as not-py2exe'd service (via
> > > PythonService.exe)?
> > >
> > > Thomas
> > >
> > > PS: As far as i can tell, there are no differences between the 2.2 and
> 2.1
> > > release of py2exe. Not yet ;-)
> >
> >
>
>





From rmunn at pobox.com  Thu Apr  3 04:57:42 2003
From: rmunn at pobox.com (Robin Munn)
Date: Thu, 03 Apr 2003 09:57:42 GMT
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org> <slrnb8lre3.15s.do-not-spam-ben.hutchings@tin.bwsint.com> <b6esse$pge$1@heckle.cc.uregina.ca> <3E8B0030.E3B0DE00@engcorp.com> <b6f3o3$t1n$1@panix3.panix.com> <3E8B4183.3DF3F57E@engcorp.com>
Message-ID: <slrnb8ncct.eaq.rmunn@localhost.localdomain>

Peter Hansen <peter at engcorp.com> wrote:
> Aahz wrote:
>> 
>> In article <3E8B0030.E3B0DE00 at engcorp.com>,
>> Peter Hansen  <peter at engcorp.com> wrote:
>> >
>> >I believe Google could provide more detail with a search using
>> >"stainless roman numerals" or something like that.
>> 
>> You're using Roman numerals for underwear?
> 
> _What_ underwear?

Semper ubi sub ubi.

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From andrew at acooke.org  Tue Apr 29 12:47:51 2003
From: andrew at acooke.org (andrew cooke)
Date: Tue, 29 Apr 2003 12:47:51 -0400 (CLT)
Subject: Where did you learn to unit test?  How did you learn?
In-Reply-To: <200304290836.42804.shalehperry@attbi.com>
References: <1147e466.0304290638.287687f6@posting.google.com> 
     <200304290836.42804.shalehperry@attbi.com>
Message-ID: <3180.10.2.0.2.1051634871.squirrel@127.0.0.1>

as someone who uses java and tests (not sure if they're "unit tests", but
i use junit and test selected classes), i want to second shaleh's point. 
the java type system catches a lot of errors (and generic java, while not
perfect, is a big advance on ordinary java in this respect (generic java
include c++ template like syntax that allows generic containers to be
specialised)).  so there's no (very little) need to test for type
correctness.

that leaves testing for algorithm correctness in the more general sense of
results being what are expected, which is of course useful (and why i do
unit testing in java).

however, when refactoring, it is the type system that i rely on, not the
tests.  the type system is a huge help in refactoring - if i extend an
interface, for example, the compiler will automatically warn me of every
class that (incorrectly) uses the modified interface.  tests are some help
in refactoring, but once the code compiles most of the hard works has been
done.

imho help with refactoring is by far the biggest bonus of static typing
(in a language like java; in more functional languages refactoring tends
to be more about adding higher order functions which, again, are helped by
static typing, but the experience is somewhat different).  this may depend
on your development environment, of course - at the moment i'm working
with a somewhat ugly codebase that needs a lot of refactoring.

returning to the original question - to introduce tests just start using
them.  add them to the main build process in ant (eg by searching for all
classes that end in "Test").  then wait until someone breaks your code. 
at that point they'll trip a test and realise that it's useful.

another example - someone who works for me wrote some code and asked me to
review it.  i wrote some tests and showed him which cases failed.  that
showed him how tests are useful.

(on the flip side, it's a sad fact of life that you sometimes end up
working with crap programmers.  some aren't going to see the advantages of
tests, just as they fail to see the advantage of almost anything else. 
all i can suggest is getting them moved to someone else's project (or
getting moved to someone else's project yourself, depending on whether
they're above or below you on the ladder ;o)).

a key point to making tests work in java is abstracting out test
specifications.  most of the work of test classes is adding frameworks
that allow cases to be specified in a compact manner.  once that is done,
adding test cases is very simple (for example, i verify trees from a
parser against xml specifications - the framework is fairly complex, but
once done it's easy to test many different cases).  this is particularly
important in java because the language itself (the syntax) is fairly
inflexible and verbose.

this approach (lead by example) seems to me much better than forcing
people to use tests.  after all, they're not always appropriate.  and
there's no surer way of making someone dislike somethig than forcing them
to do it.  also, generally, leading by example seems to get you a "better"
team.

andrew

ps xml spec for javacc asts at
http://www.acooke.org/jara/org-acooke/org/acooke/javacc/NodeVerifier.html

Sean 'Shaleh' Perry said:

> On Tuesday 29 April 2003 07:38, Christopher Blunck wrote:
>> Hi all-
>>
>> This is my second generic question posted to c.l.p (the first was 'How
>> many of you are Extreme Programmers?') but it seems like a somewhat
>> appropriate discussion for the group.
>>
>> I've noticed in my development that Python programmers (in general)
>> write more unit tests than their java counterparts.  If I download a
>> project off of jakarta.apache.org, it very rarely has unit tests.
>> However, when I download a python module, it frequently has unit
>> tests.
>>
>
> I believe this is due in part to Python being an interpreted language.
> With
> Java you have a compiler there to catch many problems but with Python most
> bugs, misfeatures, etc. can only be found by exercising the code.  Python
> coding is all about handling the runtime which makes for rapid writing of
> code but almost requires stronger testing.
>
> As Skip said, I got into unit testing by writing test cases to prove bugs
> were
> fixed.  When more bugs were found either new tests were written or old
> ones
> improved.
>
> After that I got into test writing as a way to actually write my
> applications.
> One thing I find hard for me is interface design.  So now I write the code
> that uses my class or function before I write the function.  This helps me
> identify the ways I want the code to work and also provides me with a test
> infrastructure.
>
> --
> http://mail.python.org/mailman/listinfo/python-list
>
>


-- 
http://www.acooke.org/andrew




From belred1 at yahoo.com  Tue Apr 22 13:04:50 2003
From: belred1 at yahoo.com (Charlene and Bryan)
Date: Tue, 22 Apr 2003 10:04:50 -0700 (PDT)
Subject: xmlrpc introspection
In-Reply-To: <16037.28589.172440.897944@montanaro.dyndns.org>
Message-ID: <20030422170450.33256.qmail@web13907.mail.yahoo.com>

thanks skip,

but that site didn't have what i'm after.  the
introspection API link is dead.  i went to frontier
since i read that they updated their docs, but i can't
find the updates with introspection anywhere.  i did
google around first, but no luck anywhere.

bryan

--- Skip Montanaro <skip at pobox.com> wrote:
> 
>     Bryan> i'm interested in implementing the xmlrpc
> introspecton methods
>     Bryan> ...
> 
> Google is your friend.  Searching for "xmlrpc
> introspection" yields this
> promising item.
> 
>    
>
http://xmlrpc-c.sourceforge.net/xmlrpc-howto/xmlrpc-howto-api-introspection.html
> 
> Skip


__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com




From mwh at python.net  Wed Apr  9 12:47:01 2003
From: mwh at python.net (Michael Hudson)
Date: Wed, 9 Apr 2003 16:47:01 GMT
Subject: Adding item in front of a list
References: <pan.2003.04.09.14.42.36.965880.811@thomas-guettler.de> <mailman.1049892916.24803.python-list@python.org> <pan.2003.04.09.16.59.35.949871.811@thomas-guettler.de>
Message-ID: <7h3brzf1uuz.fsf@pc150.maths.bris.ac.uk>

Thomas Guettler <pan-newsreader at thomas-guettler.de> writes:

> insert(0, 1) behaves like insert(-1, 1). It is better, but
> insert(1) would be best. 
> 
> This is not about the functionality, but about the beauty of python.
> 
> Would be nice if handling lists like queues and stacks would be a bit
> better. 

Ah, but Python is telling you something here: if you want to use a
list as a stack, you should push/pop from the *end* to reduce the
number of allocations...

Cheers,
M.

-- 
  ... with these conditions cam the realisation that ... nothing
  turned a perfectly normal healthy individual into a great political
  or military leader better than irreversible brain damage.
                   -- The Hitch-Hikers Guide to the Galaxy, Episode 11



From peter at engcorp.com  Mon Apr  7 07:06:08 2003
From: peter at engcorp.com (Peter Hansen)
Date: Mon, 07 Apr 2003 07:06:08 -0400
Subject: wxPython update problem
References: <pan.2003.04.07.10.05.55.749576@idi.ntnu.no>
Message-ID: <3E915BA0.A4B5286@engcorp.com>

Nils Grimsmo wrote:
> 
> class MainWindow(wxFrame):
>     label = ["foo", "bar"]
>     button = None
>     def __init__(self, parent, id, title):
>         wxFrame.__init__(self, parent, -4, title, size = (100, 100))
>         self.button = wxButton(self, 13, self.label[0])
>         EVT_BUTTON(self, 13, self.run)
>         self.Show(true)
> 
>     def run(self, event):
>         for c in range(2):
>             for x in [0, 1]:
>                 self.button.SetLabel(self.label[x])
>                 # please update. pretty, please!
>                 time.sleep(1.0)

When do you think/hope run() will be called?  I don't think that
method is looked for by wxPython...  Did you base this on 
some sample code?

-Peter



From matthew at newsgroups.com  Tue Apr  8 05:53:37 2003
From: matthew at newsgroups.com (matthew)
Date: Tue, 08 Apr 2003 21:53:37 +1200
Subject: passing no arguments to a callback function question
In-Reply-To: <3E9294D4.A9F95112@engcorp.com>
References: <b6tqig$93k$1@lust.ihug.co.nz> <3E9294D4.A9F95112@engcorp.com>
Message-ID: <b6u60i$i6p$1@lust.ihug.co.nz>

Peter Hansen wrote:
> matthew wrote:
> 
>>I'm pretty new to python and most of my stumblings get me somewhere but
>>I'm stumped on this. Given the following code/exception how do I avoid
>>the exception generated when notify_listeners is called?
>>
>>     def print_params():
>>        print 'PARAMS: ', params
>>
>>gives:-
>>
>>   File "D:\Python22\Lib\site-packages\TeePee\notifier.py", line 39, in
>>__call__
>>     return self.fn(*(self.args + args), **d)
>>TypeError: print_params() takes no arguments (1 given)
<snip>
Peter
Thanks for the hint. However, functions like print_params are outside of 
the class scope sitting in another module. The idea is that users can 
write their own functions which they can add as listeners for certain 
event types being scanned from a file. I 'pinched' the curry class from 
the py.cookbook and I think I'm probably being tripped up by my only 
vague understanding of argument passing. Here is the listener code 
(which I'm sure must look pretty ugly to you guys):-. Thanks for further 
assistance. matthew.

     class curry(object): <<< copied from cookbook - have no idea how it 
works!

         def __init__(self, fn, *args, **kw):
             self.fn, self.args, self.kw = (fn, args, kw)

         def __call__(self, *args, **kw):
             if self.kw:
                 d = self.kw.copy()
                 d.update(kw)
             else:
                 d = kw
             return self.fn(*(self.args + args), **d)

     def add_listener(self, event, function, args=None):
         if debug=='on': print 'Event & Func: ', event,function,args
         try:
             entry = (function, self.curry(function,args))
             self.event_list[event].append(entry)
         except KeyError:
             if debug=='on': print 'Notifier.add_listener: \'', event, 
'\' not in event_list'
             else: raise # to caller

     def call_listeners(self, event):
         try:
             for listener in self.event_list[event]:
                 listener[1]() #called
         except KeyError:
             if debug=='on': print 'Notifier.call_listeners: \'', event, 
'\' not in event_list'
             else: raise # to caller




From a.schmolck at gmx.net  Mon Apr  7 12:26:05 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 07 Apr 2003 17:26:05 +0100
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk>
Message-ID: <yfsof3ijmua.fsf@black132.ex.ac.uk>

Kim Petersen <kp at kyborg.dk> writes:

> > Remember that the scope of elisp variables for example are not lexical.
> > This is something very hard to emulate in python.
> 
> Why would the variable scheme that elisp uses be tougher to emulate in python
> than a lexical version ? From my pov it should make it a heck of a lot easier
> (think of emacsen as having 1 global dictionary and 1 buffer specific
> dictionary for global/buffer local variables) then every time you call a
> function in elisp - do a shallow copy of local variable space (dictionary) and
> send it as the local vars - you can safely throw away the dict after function
> returns).

Ahem, there are several thousand variables kicking about in global in emacs,
you want to copy them all for each function call?

How do your propose to handle the following efficiently:

(defun foo()
  (message "when foo is called folding is: %s" case-fold-search))
(defun bar()
  (message "when bar is called folding is: %s" case-fold-search)
  (let ((case-fold-search t))
    (foo))
  (message "now folding is: %s again" case-fold-search))
(bar)

when bar is called folding is: nil
when foo is called folding is: t
now folding is: nil again

alex



From skip at pobox.com  Fri Apr 11 15:04:09 2003
From: skip at pobox.com (Skip Montanaro)
Date: Fri, 11 Apr 2003 14:04:09 -0500
Subject: intern'ed strings and deepcopy()
In-Reply-To: <ef08387c.0304111042.73a16e68@posting.google.com>
References: <ef08387c.0304111042.73a16e68@posting.google.com>
Message-ID: <16023.4521.891951.354551@montanaro.dyndns.org>

    fp> I can't seem to find the exact description between these two in the
    fp> Library Reference - so am I correct in assuming that deepcopy() will
    fp> not actually duplicate an intern'ed string?

Strings (interned or not) are immutable, so there's never a need to copy
them:

    >>> s = "abc def ghi jkl"
    >>> id(s)
    9251080
    >>> id(copy.deepcopy(s))
    9251080

    >>> t = intern("abc")
    >>> id(t)
    11982528
    >>> id(copy.deepcopy(t))
    11982528

Skip




From usenet_spam at janc.invalid  Wed Apr 23 19:01:06 2003
From: usenet_spam at janc.invalid (JanC)
Date: Wed, 23 Apr 2003 23:01:06 GMT
Subject: Off topic: What usenet news reader to use?
References: <mailman.1051122829.6158.python-list@python.org>
Message-ID: <Xns9367A58C27FJanC@213.118.75.206>

Stan Baptista <sbaptista at yahoo.com> schreef:

> I also want to switch over to a newsgroup for the same
> reason so I downloaded xnews and tried to access
> these:
> 
> comp.lang.python
> and
> comp.lang.python.announce
> 
> I get error: "10110 unknown error while looking up
> comp.lang.python[.announce]"

Sounds like a Winsock error during DNS lookup?
Did you enter your newsserver's hostname (often news.provider.com or 
something like that) correctly?

-- 
JanC

"Be strict when sending and tolerant when receiving."
RFC 1958 - Architectural Principles of the Internet - section 3.9



From max at alcyone.com  Wed Apr 16 21:31:43 2003
From: max at alcyone.com (Erik Max Francis)
Date: Wed, 16 Apr 2003 18:31:43 -0700
Subject: How many of you are Extreme Programmers?
References: <1147e466.0304160630.5b510724@posting.google.com> <fc30bff.0304161639.2969734a@posting.google.com>
Message-ID: <3E9E03FF.E8E4FFAB@alcyone.com>

John Taylor wrote:

> Can someone please describe to me XP in a few sentences?  I have heard
> the terminology before, but never an explanation.  Also, any good
> links or books would be greatly appreciated.  Python is such a joy to
> program in/with, I'd love to hear about making the process even
> better!

Google sez:

	http://www.extremeprogramming.org/
	http://www.xprogramming.com/

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ It was involuntary.  They sank my boat.
\__/ John F. Kennedy (on how he became a war hero)
    Bosskey.net: Return to Wolfenstein / http://www.bosskey.net/rtcw/
 A personal guide to Return to Castle Wolfenstein.



From staschuk at telusplanet.net  Sun Apr 27 20:38:47 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 27 Apr 2003 18:38:47 -0600
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <b8cgst$20ho$1@nntp6.u.washington.edu>; from donn@u.washington.edu on Fri, Apr 25, 2003 at 11:38:05PM +0000
References: <b7km3s$238m$1@nntp6.u.washington.edu> <LNBBLJKPBEHFEDALKOLCGEFEEDAB.tim.one@comcast.net> <mailman.1050607638.7466.python-list@python.org> <b7n4hk$2g1u$1@nntp6.u.washington.edu> <mailman.1050619770.15662.python-list@python.org> <b7nh0t$33li$1@nntp6.u.washington.edu> <20030417205405.A10672@tibia.amotlpaa.bogus> <mailman.1051310385.1948.python-list@python.org> <b8cgst$20ho$1@nntp6.u.washington.edu>
Message-ID: <20030427183847.C28461@tibia.amotlpaa.bogus>

Quoth Donn Cave:
> Quoth Steven Taschuk <staschuk at telusplanet.net>:
> ...
> | I've submitted a documentation patch for the __str__ and __repr__
> | docs, following ideas discussed in this thread:
> |     <http://www.python.org/sf/727789>
  [...]
> I still think that the normal usage of __str__ can be directly
> described as a data conversion to type string.  I know you don't
> find that completely illuminating on its own, but I think it does
> convey an essential idea that is otherwise likely to be missed.
> See for example the __str__ and __repr__ definitions in Cookie.py.

Hm... "conversion to type string".  So under this concept str()
converts an object to a string, while repr() represents the object
in a string, right?  Is this what you're saying when you
distinguish "the data as a string" from "the object as a string"?
If so, I quite like it.

(For some reason convert/represent sits better in my brain than
data/object.  Don't ask me; I just think here.)

> Since this apparently isn't as obvious to you as it is to me, you
> may be in a better position to explain it if I can convince you
> that there's something to it.  [...]

*chuckle*  True.  (Not that I'm in any special position vis a vis
changing the documentation.  Quite the contrary.)

> [...] If we can decide on some central
> notion behind normal __str__ usage, the examples will make more
> sense and it won't be necessary to describe __str__ as "not __repr__".

I entirely agree.  I'll have to think about the notion of
conversion to type string and see if I can devise a more
satisfactory explanation of __str__ in those terms.  (Further
bulletins as events warrant.)

-- 
Steven Taschuk                                                   w_w
staschuk at telusplanet.net                                      ,-= U
                                                               1 1




From theller at python.net  Sat Apr  5 11:38:51 2003
From: theller at python.net (Thomas Heller)
Date: 05 Apr 2003 18:38:51 +0200
Subject: Input from a file as a command line argument
References: <3E8EDBDA.7020701@pbezone.net>
Message-ID: <istsc31g.fsf@python.net>

Teemu Luojola <teemu.luojola at pbezone.net> writes:

> I have read through Python documentation and tried several searches,
> but I can't find a simple answer to this question.
> 
> 
> If I make an executable script in linux environment, how can I pass
> input to that script from a file on command line? Like the following:
> 
> 
> $ script.py input.txt
> 
> or perhaps
> 
> $ script.py < input.txt
> 

See module fileinput.

import fileinput
for line in fileinput.input():
    process(line)

Thomas



From ramen at lackingtalent.com  Tue Apr  1 10:46:42 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 15:46:42 -0000
Subject: Questions on linked lists
References: <kl8ia.27315$OV.127379@rwcrnsc54> <5vonl-vm3.ln1@beastie.ix.netcom.com>
Message-ID: <slrnb8ja7a.421.ramen@lackingtalent.com>

In article <5vonl-vm3.ln1 at beastie.ix.netcom.com>, Dennis Lee Bieber wrote:
> 
>         I suspect that, in the case of Python, it would be reinventing the 
> wheel. Python has lists, and list operations, as a native data type; 
> one doesn't have to program list manipulation from scratch. What Python 
> does not have is user control of dynamic allocation and access to 
> addresses.
> 

That said, what would be an example of a legitimate need for a custom linked
list data structure in Python? I can't say I've come across one yet. Do
linked lists outperform native Python lists at certain operations?

Dave



From steven at twitch.net  Mon Apr 28 02:41:47 2003
From: steven at twitch.net (Steven Hodgen)
Date: Sun, 27 Apr 2003 23:41:47 -0700
Subject: problems with compiling and finding modules
Message-ID: <oprobt7xfxx3zh5c@news.mindspring.com>

Hello,

I'm currently learning Python and am having a couple of problems.

I use a programmer's editor called "Visual SlickEdit".  I'm trying to set 
it up to work with Python.  So, how do I get it to compile a python file.  
I can't seem to find a command line method for either compiling a .py file 
to a .pyc file or even just do a syntax check for finding errors in the 
editor.  I found a file called py_compile, but this just seems to add a 
function called compile which I can execute from the interpreter but not 
the command line.  The file compileall, will compile everything in 
directory or entire tree, but not just a single file.  I would asume there 
must be a way short of writing my own single file compile script.  I 
understand I'm going to have to do something like:

  python compile_command myfile.py

but what do I put in for compile_command?

Also, how do I change the working directory from within IDLE?  If I open my 
.py file from the Windows Explorer using a right click and choose 'Edit 
with IDLE', everything works fine, but if I 'Open...' a .py file from 
within IDLE, it can't find modules I import from.  I assume this is because 
it isn't in the current directory.

Thanks!

--Steven
steven at twitch.net



From gry at ll.mit.edu  Wed Apr 23 14:07:54 2003
From: gry at ll.mit.edu (george young)
Date: Wed, 23 Apr 2003 14:07:54 -0400
Subject: Exception.__init__ needed in derived class?
Message-ID: <20030423140754.1946afea.gry@ll.mit.edu>

[python-2.3a1]
Does class Exception have a constructor __init__ that needs
to be called?  I couldn't find this in the docs.  I see many
uses of inheriting from Exception without invoking
Exception.__init__(self), but nobody ever *said* that this 
was safe to leave out...

Pychecker complains about this, which is what brought it to
mind.  I can certainly invoke it just to make pychecker 
happy, but I wondered...

-- 
 I cannot think why the whole bed of the ocean is
 not one solid mass of oysters, so prolific they seem. Ah,
 I am wandering! Strange how the brain controls the brain!
	-- Sherlock Holmes in "The Dying Detective"



From jeffbernstein at yahoo.com  Wed Apr  2 22:37:54 2003
From: jeffbernstein at yahoo.com (Jeff Bernstein)
Date: 2 Apr 2003 19:37:54 -0800
Subject: Mac OS X, X11, Python and IDLE: keyboard shortcuts?
Message-ID: <f1e1cae2.0304021937.2c09db7a@posting.google.com>

Greetings:

I have IDLE running in X11 find and dandy. What I'm trying to figure
out is how to get into the menus on IDLE using the keyboard. I figured
out that the Apple/splat key is used in place of Alt and CTRL really
is CTRL. Anyone know how to get to the menus in IDLE via the keyboard?

I'm using an iBook.

Thanks!

Jeff

-----BEGIN PGP SIGNATURE-----
Version: PGP 8.0 - not licensed for commercial use: www.pgp.com

iQA/AwUBPoutQQzQ3D7IrjzcEQIGEQCg0EYB67vRsiCRr2d6nyByXwhlGX8Anj38
HjSk/tqXw9KH16YrpjewVMmn
=kT9v
-----END PGP SIGNATURE-----



From llothar at web.de  Sun Apr  6 15:44:38 2003
From: llothar at web.de (Lothar Scholz)
Date: 6 Apr 2003 12:44:38 -0700
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain>
Message-ID: <6ee58e07.0304061144.35c5ab92@posting.google.com>

pobrien at orbtech.com (Patrick K. O'Brien) wrote in message news:<m2u1dcw22s.fsf at localhost.localdomain>...
> If you were crazy enough to think that you could create a program
> along the lines of Emacs, but written in Python, how would you go
> about it?  How would you design the domain model for files, buffers,
> windows, and frames?  How would you allow the same level of
> customizability?  How would you map functions (or methods or whatever)
> to keys?  Any thoughts?

Look at the basics, that is the 200.000 LOC of C Code in Eiffel.
It's good documented. And then skip it.

> Here is the background.  I'm working on a file editor (called
> PyAlaMode) using wxPython and Scintilla.  It's an extension of the
> PyCrust code base.  I've got a basic version working, and it supports
> editing multiple files and all the usual, basic features.  But I want
> it to be as customizable and extensible as Emacs.  In fact, I want to
> model it after Emacs, so that elisp code could be rewritten in Python
> and work with PyAlaMode.  To do that, I've got to support the same
> primitives as Emacs; expose files, buffers, regions, and windows the
> same way Emacs does; handle the Emacs keybindings and ability to remap
> keys, the same way Emacs does, etc.

Then you have to present the same old and ugly user interface with a
mini  buffer as input line and so on. I don't see any reason at the
moment for cloning EMACS. You should ask people why they like emacs as
an editor.
There are only a handful of features that are normally the reason to
keep
this as your primary editor. For me its the window handling C-X 1, the
way to prefix commands with a repeat command, the macro recording and
a the autoindentation with TAB.

To make clear, i'm using XEMACS 8 hours a day for years now. But it's
absolutely losing ground compared to modern editors. And i don't think
that
extensibility in the emacs sense is really necessary today or a plus,
because
people want out of the box solutions that are specialized to their
domain.

> 
> Sound like fun?

No, sounds like real work. Drinking beer and making love with your
girl friend is real fun. Everything else should be seen as a nonpaid
job.



From oktaysafak at ixir.com  Wed Apr 16 06:28:38 2003
From: oktaysafak at ixir.com (Oktay Safak)
Date: Wed, 16 Apr 2003 13:28:38 +0300
Subject: AOP in Python (was: Post-OOP: not just an "unfortunate acronym"?)
References: <pan.2003.04.15.21.27.33.796814@mn.rr.com> <slrnb9p5bk.vg8.ramen@lackingtalent.com>
Message-ID: <001a01c30402$f4759c60$0200a8c0@b>

"Dave Benjamin" <ramen at lackingtalent.com>, wrote:
> Does anyone have an example of this newfangled "Aspect Oriented"
paradigm's
> techniques implemented in Python? The only thing I've seen in any language
> (and I haven't looked at AspectJ or anything) was a casual reference in a
C#
> text that seemed something like function-call pre/post event-handlers.
What
> might Python stand to gain from this methodology?

See http://peak.telecommunity.com/. It has been some time since I read some
docs about it but as far as I can remember, the previous version of
PEAK(Python Enterprise Application Kit), whose name I can't remember at the
moment, aspired to bring AOP principles to Python enterprise development. Is
PEAK still on that track? That, I don't know. Need to comb the docs but
unfortunately I feel too lazy to do it at the moment ;) Maybe you can do it
and share your findings with us here <wink> ;)

Regards,

Oktay





From jack at performancedrivers.com  Sun Apr 27 11:51:34 2003
From: jack at performancedrivers.com (Jack Diederich)
Date: Sun, 27 Apr 2003 11:51:34 -0400
Subject: metaclasses, how you might actually use 'em
Message-ID: <20030427115134.M15881@localhost.localdomain>

__metaclass__ has been given a bad name.  They aren't as hard as you think and
they're a good deal more handy in your day to day life.  This is just a brief
exposition on metaclasses.  I just finished refactoring a bunch of code and
dropped the linecount by 15% while making the source more concise.  But it is
only more concise if people understand metaclasses, hence this evangelism.
Standard disclaimer, this is how I've come to know and love metaclasses.
Corrections encouraged, pedantry especially.

Brief overview of metaclasses

metaclasses just allow you to tweak a class definition on the fly.  Mainly 
adding or removing attributes just before the class becomes 'real'.  You could
get the same benefits from using eval() with a lot of pain and suffering. I 
use them for the Register pattern, and the Named Prototype pattern[1].

1) The Register pattern.
  Not included in GoF's 'Design Patterns' but one I've used quite frequently.
  Every time you define a subclass of X, store a pointer to it in a central 
  location, usually a factory but handy for generating any list of subclasses
  of X.

class Register(type):
  """A tiny metaclass to help classes register themselves automagically"""

  def __init__(cls, name, bases, dict):
    if ('register' in dict): # initial parent class
      setattr(cls, 'register', staticmethod(dict['register']))
    elif (getattr(object, 'DO_NOT_REGISTER', None)): # don't register this guy
      delattr(cls, 'DO_NOT_REGISTER') # this attribute won't be inherited
    else: # the usual case, register the class
      cls.register(cls)
    return

Here is an example of using the generic Register metaclass

class Factory(object):
  """A relatively useless factory"""
  string_name_to_class = {}
  def make_one(string_classname):
    """if string_classname == 'Foo', return Foo()"""
    return Factory.string_name_to_class[string_classname]()
  make_one = staticmethod(make_one) # make this method static

class Base(object):
  __metaclass__ = Register
  def register(klass):
    """tell Factory about klass"""
    Factory.string_name_to_class[klass.__name__] = klass
    return

class Foo(Base): pass
class Bar(Base): pass

Now let's create a Foo object by calling factory with the string name

>>> print Factory.string_name_to_class
{'Foo': <class 'meta.Foo'>, 'Bar': <class 'meta.Bar'>}
>>> foo = Factory.make_one('Foo')
>>> foo
<Foo object at 0x8153bdc>
>>> 

I used to have a list of lines at the bottom of a module that looked like
register_REST('my_URI_name1', SomeClass1)
register_REST('my_URI_name2', SomeClass2)
.. [lots more lines here]

These registered classes to do handle particular request when my mod_python 
program got a URI like 'http://REST/my_URI_name1'

I've also used it to count instances of subclasses in a graphing program,
there could be multiple lines are bars on a graph.  If there was N bars on
the graph the width of the bar is divided by N and offset from the starting
point.  I didn't want to know how many primitives there were because only
bars have this property. It counts the number of items of a class (lines or
bars) and tells each class you are the first, second, etc out of N for your
class.

[Speculation, the reason this wasn't included in the GoF is because C++
 makes no guarantees about order of class initialization.  So it isn't used.]

2) Named Prototype pattern.
  This one is in 'Design Patterns' by the GoF.  I'm from a C++ background so
  when I started with python I still did this the C++ way.  The prototype
  pattern is when you get a copy of an instance and then call some methods
  on the copy to change its behavior.  Specialization, but at runtime instead
  of by inheritance.

  This pattern is driven by laziness.  I can type quickly but verbosity hurts.
  It also allows for some compile time checking.

  My runtime definition of specialized prototypes looked like

  newob = ob.copy()
  newob.define(param1 = None,
               param2 = 'Sally',
               ..
               paramN = 2 * 8,
              )
  otherob = ob.copy()
  otherob.define(another_long_list_of_parameters)

  The parameters that define the behavior of the new object look like line
  noise.  The define() method would check the arguments, but this didn't happen
  until the code actually ran. Enter the Named Prototype.

class NamedPrototype(type):
  def __init__(cls, name, bases, dict):
    params = {}
    for (param) in my_list_of_important_params:
      params[param] = getattr(cls, param) # might raise an error, that is OK
      delattr(cls, param)
    validate_parameters(params) # might raise an error, that is OK
    
    setattr(cls, 'final_params', params)

class Base(object):
  """Virtual base class for our classes"""
  __metaclass__ = NamedPrototype

class A(Base):
  param1 = None
  param2 = 'Sally'
  ..
  paramN = 2 * 8

All the param* values will be removed from class A, validated, and stored in
the final class definition in the 'final_params' hash

>>> a = A()
>>> a.final_params
{ 'param1' : None, 'param2' : 'Sally' ... }

All this is only done once at the time the class is defined.
If you have every used PLY or SPARK (both lex/yacc rewritten in python)
they use the names of functions and even the doc strings to define the
behavior of the resulting classes.  A lexer in a modified PLY[2] looks like[3]

class MyLexer(BaseLexer):
  t_ignore = ' \t'
  
  def t_newline(self, t):
    r'[\n\r]'
    self.code += "\n"
    return t
  
  t_opencurly = r'\{'
  t_closecurly = r'\}'
  
  def t_leading_white(self, t):
    r'^\s+'
    if (not self.in_block):
      self.whitespace = ''
      self.whitespace = t.value
    return t

all the names starting with t_ are special and define rules for tokens.
Plain tokens are just regular expression strings, fancier tokens will call
a function and are matched by the function's doc string.  The BaseLexer class
has a metaclass that will check all the attributes, delattr() the members and
store them in a more regular structure and validate the results.

All the benefits of the Named Prototype pattern could be achieved through
straight inheritance in many instances, and regular Prototyping in the rest.
All except for easier finger typing.  The coder/user has to know that these 
methods will be treated specially, but he doesn't have to know they would be 
any more special because the metaclass cares or the plain inheritance super
class cares.  Either way they have to have read the documentation.


Ending disclaimer.  I don't read the snippets on ASPN, just c.l.py.  If this 
is known stuff or there is a better standard way just mark this thread as
ignored and enjoy your Sunday.  If people show interest I can provide
more deatils of my actual usage, but this post is already too long.

-jackdied

[1] Register pattern and Named Prototype pattern are TM me, right now, just 
    thought em up, couldn't think of anything better, suggestions welcome.

[2] For my own purposes I bastardized PLY.  An out of date version can
be found as lexer.py at http://jackdied.com/cgi-bin/viewcvs.cgi/pymud/
This code is completely unsupported. Use PLY or SPARK, they have 
good documentation.

[3] This is part of a psp (python server pages) lexer I posted to the
    mod_python development list a couple weeks ago.  Unfortunately, I don't
    think it is archived.




From creedy at mitretek.org  Tue Apr  8 16:28:40 2003
From: creedy at mitretek.org (Chris Reedy)
Date: Tue, 08 Apr 2003 16:28:40 -0400
Subject: Import 'filename'? (was Re: Cross-reference 'import' in a class hierarchy)
In-Reply-To: <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <3e932e9f$1_2@corp-news.newsgroups.com>

Ben Finney wrote:
> On Tue, 08 Apr 2003 15:44:10 +1200, Greg Ewing (using news.cis.dfn.de) wrote:
> 
>>It's been suggested before (and I happen to agree) that Python should
>>provide a general way of making *relative* references to modules or
>>packages.
> 
> 
> I continue to be amazed that there's simply no way to say
> 
>   import "../foo/ook.py"
> 
> as should be possible to implement any well-architected class hierarchy.
> 
> Anyone who can show me otherwise gets a jelly baby.

I had to think about this one for a while. Here's the issue I see:

What exactly does import '../foo/ook.py' actual do?

If you're still interested, read on.

<longdiscussion>

There are three ways (no wait -- make that four ways) to refer to a 
module/package:

(1) It's name as defined by searching the file system using PYTHONPATH:

   This is what you get when you say import MyApp.foo.ook. It looks for 
a file named ook.py in a directory <dir>/MyApp/foo where <dir> is in 
PYTHONPATH.

(2) It's name as defined in sys.modules.

   This is what you get when you say import MyApp.foo.ook and 
sys.modules['MyApp.foo.ook'] already exists.

(3) It's name as defined via variable lookup in the local or global 
namespace.

   That is, when you write x = MyApp.foo.ook.f(), you expect that MyApp 
refers to a module object that has an attribute foo that is a module 
object that .... Remember that if you "import MyApp.foo.ook as eek" you 
can now refer to eek.f().

(4) It's actual location in the file system, i.e. 
<somedir>/MyApp/foo/ook.py.

   Note that this one doesn't correspond to the others in any well 
defined way: ook.py could be loaded as module 'ook', 'foo.ook', ... or, 
as '__main__'. In the former cases, the module would have a __file__ 
attribute giving the correct location. In the latter case it doesn't, 
but sys.path[0] would give you the directory?!

Based on some reading in the email archives, there is clearly a desire 
(maybe a mandate by the BDFL) that some sort of correspondence be 
guaranteed to exist between the above four definitions (especially 1, 2, 
and 3).

</longdiscussion>

So my question is what does import '../foo/ook.py' mean in this context. 
What it's supposed to mean relative to (4) above seems obvious.

However, what it's supposed to mean relative to (2) and (3) is not at 
all clear. In particular, what name(s) would be added to the local 
dictionary and what name would be added to sys.modules?

Also, relative to (1), what would it mean if there is an ook.py that can 
be found via PYTHONPATH and it isn't the one you just loaded.

   Pant, pant,
   Chris

-- 
This is informal and not an official Mitretek Systems position.
Dr. Christopher L. Reedy, Senior Principal Software Engineer
Mitretek Systems, 3150 Fairview Park Drive South, Falls Church, VA 
22042-4519
Email: creedy at mitretek.org  Phone: (703) 610-1615  FAX: (703) 610-2203




From ramen at lackingtalent.com  Wed Apr  2 15:38:25 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Wed, 02 Apr 2003 20:38:25 -0000
Subject: intro to OOP using Python, tutorial?:newbie
References: <3E8A5AE4.1080902@netscape.net> <b6el8v$7d7$2@nntp0.reith.bbc.co.uk>
Message-ID: <slrnb8mfmb.tpj.ramen@lackingtalent.com>

In article <b6el8v$7d7$2 at nntp0.reith.bbc.co.uk>, Michael Sparks wrote:
> vector wrote:
> 
>> is there an introduction to OOp using python tutorial doc,file,pdf etc
>> anywhere on the net. 
> 
> How to Think Like a Computer Scientist: Learning with Python 

I just wanted to mention that this book, along with Learning Python, was one
of the most valuable resources when I was first learning Python. I think
it's a great read and highly recommend it as a tutorial.

My 2c. =)
Dave



From aleax at aleax.it  Wed Apr 30 04:01:05 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 10:01:05 +0200
Subject: Question about accessing class-attributes.
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DE169@admin56.narex.com>
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE169@admin56.narex.com>
Message-ID: <200304301001.05209.aleax@aleax.it>

On Wednesday 30 April 2003 02:15 am, Bjorn Pettersen wrote:
   ...
> > is probably some sound implementation reason for the current
> > behavior, but if so it should be better documented, I think.
>
> SF 729913. I have not been able to find an exact description of the
> lookup rules for new-style classes anywhere.

Good point.  I guess the language reference DOES need to provide
"language lawyers" with those!


> > Yes, of course not -- super(MyClass, self) is NOT an instance
> > of MyClass (isn't THAT why you're calling super...?).
>
> I'm also missing an exact description of what super() returns. It looks
> like it's an object that when you call method 'm', invokes the proper
> super class' method 'm', and if that 'm' invokes any otherm method,
> 'm2', the original objects definiton 'm2' is used. So it looks like it

Again, this needs to go somewhere in the docs, but the effect of

super(aclass, aninst).amethod()

SHOULD be the same as anotherclass.amethod(aninst) where
anotherclass is the first class following aclass in the mro of aninst's
class that defines amethod.  Note that the *lookup* for amethod
CANNOT use aclass itself, BUT the argument PASSED to amethod
MUST be aninst -- or else how would super enable "cooperative
superclass method calling"?

Given
    sup = super(aclass, aninst)
if sup were an instance of aclass, then the lookup for sup.amethod
would start in aclass, thus the typical use case such as:
    def amethod(self):
        print 'before'
        super(thisclass, self).amethod()
        print 'after'
would just recurse endlessly (well, until that raises an exception;-)
making it pretty useless.  But also, unless aninst were the instance
passed to anotherclass.amethod, this too would inhibit the intended
effect -- i.e. in the typical diamond inheritance structure:

class base(object):
    def amethod(self):
        print 'base.amethod'

class der1(base):
    def amethod(self):
        print 'der1.amethod'
        super(der1, self).amethod()

class der2(base):
    def amethod(self):
        print 'der2.amethod'
        super(der2, self).amethod()

class der12(der1, der2): pass

class der21(der2, der1): pass

der12().amethod()
der21().amethod()

how would we obtain the required sequence of prints in both
cases?

I'm NOT claiming this is perfectly stabilized yet nor that all
subtle bugs have yet been shaken off -- I'm just trying to
explain why _I_ don't find this behavior at all surprising.


> should be substitutable, or at least a first class object, but as you

Well it IS a first-class object, surely?  You can pass it as an argument,
return it as a function's result, use it as a value in a list or dict -- 
aren't these the characteristics DEFINING "first-class object"?

> said: "of course not". Am I missing some documentation, or is it
> unreasonable of me to think it should work this way (what way should it
> work... exactly)?

I wouldn't be surprised to find out that the "language lawyer" levels
of docs haven't been written down anywhere yet -- i.e. that there
is nothing much beyond descrintro or the short paragraph under
http://www.python.org/dev/doc/devel/lib/built-in-funcs.html that
describes super.  But as for 'reasonableness', I dunno -- I do not
understand how you would expect 'super' to work to enable its basic
use case of cooperative superclass method calling and yet NOT have

super(aclass, aninst).amethod()

behave just the same as

anotherclass.amethod(aninst)


There may be a good design idea hiding in your expectations -- as
long as it doesn't break super's fundamental use case -- and so I
am quite keen on understanding them!  Indeed if we were able to
classify the current behavior as "a bug" in some respect, AND give
a patch that breaks no currently designed-in behavior, yet "fixes the
bug" and makes super more useful to you, then that's the kind of
fix that COULD go in 2.3 beta 2 (new functionality won't, but in
some sense it might be considered "a bug" if super violates the
normal expectations of clearly-advanced users such as you).

Maybe we should break this thread in two -- one about metaclasses
and one about super -- with suitable subjects... I do not understand
where the 'bugs' of the two may be related nor actually what they
really have to do with "accessing class-attributes" except in the
vaguest, broadest sense.


> > > Perhaps a more realistic exmaple?:
> > >
> > >   class Secure(object):
> > >     def __init__(self, name):
> > >       self.name = name
> > >
> > >     def __getattr__(self, attr):
> > >       return getSecureMethod(self.name, attr, getUserPrivileges())
> > >
> > >   class LogFoo(Secure):
> > >     def __init__(self):
> > >       Secure.__init__(self, 'SalaryDB')
> > >
> > >     def foo(self):
> > >       print 'calling foo'
> > >       return super(LogFoo, self).foo()
> > >
> > >   lf = LogFoo()
> > >   print lf.foo()
> >
> > Sorry, I don't get it.  How is this related to metaclasses...?
>
> Not terribly obvious was it... :-)  And it's a bad example too. The
> thinking was that if Super's metaclass could intercept special methods
> through it's __getattr__, the object returned from super() could look
> more like a regular object.  That Super.__getattr__ doesn't check for a
> __getattr__ in the __mro__ chain is a problem, but a separate one (sorry
> for the confustion).

"the object returned from super" DOES look a lot "like a regular
object", except it has a custom metaclass that ensures its fundamental
behavior -- attribute lookups starts AFTER the given class, but, if and
when any bound-methods are built, the instance they're bound to is
the given one.  We can't break THAT, I think.

I think that, currently, if you want the functionality you seem to be
after in this example, you must ensure that superclass Secure has,
in its class dict at the time type.new is called to build it, explicit
entries for all specialmethods of your interest -- that way, the slots
will be filled in (perhaps with special-purpose descriptors that do
the just-in-time lookup you may want).  But, for methods that you
look up explicitly, a metaclass's __getattr__ (NOT of course the
__getattr__ of the CLASS on which you're looking up -- that only
applies to lookups on INSTANCES of that class) should work fine.

Yes, maybe we should separate super-issues and metaclass-issues
and find a sharp example in each case of where reasonably expected
behavior from docs, and observed behavior in 2.3 beta 1, differ.  I
think it does not need to be a "reasonable use case": that would be
needed if we were arguing for EXTRA functionality of course -- but
then it would go in 2.4 so there's no hurry, to say the least.  If we're
arguing there are BUGS, so the fixes could go in 2.3, then just a
way to show up each bug clearly should suffice (hopefully).


Alex





From xxdigitalhellxx at hotmail.com  Tue Apr  8 13:23:48 2003
From: xxdigitalhellxx at hotmail.com (sik0fewl)
Date: Tue, 08 Apr 2003 11:23:48 -0600
Subject: fdopen() question
In-Reply-To: <slrnb95u9d.r1m.news@cipw08.physik.uni-bonn.de>
References: <slrnb95s3m.u9b.news@cipw08.physik.uni-bonn.de> <slrnb95u9d.r1m.news@cipw08.physik.uni-bonn.de>
Message-ID: <b6v0la$mhl$1@heckle.cc.uregina.ca>

Carsten Gaebler wrote:
> Carsten Gaebler wrote:
> 
>>Traceback (most recent call last):
>>  File "./listener.py", line 27, in ?
>>    f = os.fdopen(fd, 'r')
>>OSError: [Errno 9] Bad file descriptor
>>
>>What am I missing?
>>
> 
> 
> Seems I have to open the file in 'w+' mode to make it work.
> 
> cg.

Make sure the first fdopen() parameter is a valid file descriptor, like 
the ones returned by fileobj.fileno(). The exception that was raised is 
raised if the file descriptor you are trying to open has been closed or 
never existed in the first place.

-- 
Hope this helps,
Ryan




From frithiof.jensen at remove_this.jensen.tdcadsl.dk  Tue Apr 15 10:18:54 2003
From: frithiof.jensen at remove_this.jensen.tdcadsl.dk (Frithiof Andreas Jensen)
Date: Tue, 15 Apr 2003 16:18:54 +0200
Subject: Sv: Python compared toJava
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com> <3e9bdf88$0$24726$edfadb0f@dread14.news.tele.dk> <ExTma.98936$It5.95989@news2.central.cox.net>
Message-ID: <3e9c142a$0$24713$edfadb0f@dread14.news.tele.dk>

Steve Holden <sholden at holdenweb.com> skrev i en
nyhedsmeddelelse:ExTma.98936$It5.95989 at news2.central.cox.net...

> This is a little ingenuous: some modules are only available on some
> platforms, and others have different functionality on each different
> platform....

At least it is documented what is available where - mostly, that is - with
Tix f.ex. this is not so as I recently discovered.

With java, I feel that SUN wished to beat Microsoft soo badly that they *had
to have* the same OS-version/application dependencies as Windows provides -
i.e:

User need both App_A & App_B; however App_A requires JVM X and App_B demands
JVM Y but user can only install One JVM!  Dependencies may change with
platforms too.






From bokr at oz.net  Mon Apr 28 16:37:15 2003
From: bokr at oz.net (Bengt Richter)
Date: 28 Apr 2003 20:37:15 GMT
Subject: How to write RTF from Python?
References: <3EA97D61.3CC9FE74@easystreet.com> <878ytuo3uj.fsf@andreasen.org> <3ead7a4f$0$49103$e4fe514c@news.xs4all.nl>
Message-ID: <b8k3dr$jgf$0@216.39.172.122>

On Mon, 28 Apr 2003 21:00:31 +0200, Irmen de Jong <irmen at -NOSPAM-REMOVETHIS-xs4all.nl> wrote:

>Sorry if this has already been mentioned or not appropriate
>but I'm hopping late into this thread and I missed the
>original messages...
>
>What we did quite successfully was typing a letter in Word, with special
>placeholders such as "%%%DATE%%%", "%%%AGE%%%" on the locations where
>you want your data items to appear.
>Save this as RTF.
>Load it as a 'template' in your program and replace the
>placeholders with the actual data values. Presto.
>
I suggested using (to translate your example) %(DATE)s %(AGE)s
in the locations, so you can use a dict (or something with a __getitem__ method)
to do template % aDict where aDict might be {'DATE':'2003-04-28', 'AGE':29} or whatever.

Also, you are likely to get less cruft in the .rtf if you use wordpad (write.exe) instead
of word, I think. See working example with code prev post.

Regards,
Bengt Richter



From nhodgson at bigpond.net.au  Tue Apr 15 04:19:14 2003
From: nhodgson at bigpond.net.au (Neil Hodgson)
Date: Tue, 15 Apr 2003 08:19:14 GMT
Subject: syntax highlighting file python 2.1
References: <b7gd6g$200$04$1@news.t-online.com>
Message-ID: <6gPma.6270$hF.51358@news-server.bigpond.net.au>

Luigi Monaco:

> Basically I require a full list of methods, modules or a way how to
> extract this from a python release. Keywords etc. should
> suffice from e.g.: /usr/share/vim/vim61/syntax/python.vim.

   This may help, although it can take a long time:
http://www.scintilla.org/gen_python_api.zip

   Neil





From sde at recombinant.demon.co.uk  Wed Apr 30 16:50:16 2003
From: sde at recombinant.demon.co.uk (Stephen D Evans)
Date: Wed, 30 Apr 2003 21:50:16 +0100
Subject: Python 2.3b1 and _tkinter.Tcl_Obj
References: <b8p30s$be4$1$8302bc10@news.demon.co.uk> <m3wuhbg4ys.fsf@mira.informatik.hu-berlin.de>
Message-ID: <b8pcua$m4h$1$8300dec7@news.demon.co.uk>

"Martin v. L?wis" <martin at v.loewis.de> writes:

> - set Tkinter.wantobjects to False before creating the first
>   tkapp object. This will give you the Python 2.2 behaviour.
> - convert Tcl_Obj objects to strings using str()
> - convert Tcl_Obj objects to Unicode using unicode()
> - convert Tcl_Obj objects using
>   ASCII-if-possible-otherwise-Unicode-if-possible-otherwise-bytes
>   algorithm of Python 2.2 using Tcl_Obj.string
>
> If you feel you need to take one of these approaches, please let us
> know, so we can see whether backwards compatibility needs to be
> improved.


My application runs Ok with Tkinter.wantobjects=False

I shall convert the front-end to wxPython with its bigger,
better looking (on windows) widget selection.

Tkinter/Pmw does the job, but  the world has moved on since Python 1.5.2


thanks,
Stephen D Evans





From guy at obstruction-no-spam.com  Fri Apr 18 16:41:11 2003
From: guy at obstruction-no-spam.com (Guy Middleton)
Date: Fri, 18 Apr 2003 20:41:11 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <7x4r4vy2gl.fsf@ruckus.brouhaha.com>
Message-ID: <HpZna.64494$BQi.43307@news04.bloor.is.net.cable.rogers.com>

In article <7x4r4vy2gl.fsf at ruckus.brouhaha.com>,
Paul Rubin  <http://phr.cx at NOSPAM.invalid> wrote:
> guy at obstruction-no-spam.com (Guy Middleton) writes:
> > What is the most Pythonic way to sum the n-th element of a list of lists (or
> > of tuples)?
> > 
> > The obvious (to me) way is to use reduce, a list comprehension, and a lambda:
> > 	... 
> > 	>>> reduce(lambda a, b: a+b, [y[1] for y in x])
> > 	6
> > 	>>>
> > 	$
> > 
> > But I don't really like lambda, is there a better way to do this?
> 
> import operator
> reduce (operator.add, [y[1] for y in x])


Aha!  That's what I was looking for.  Thanks...

 -Guy



From news at saarpfalz.de  Wed Apr  2 01:51:46 2003
From: news at saarpfalz.de (Susanne Ziemerle)
Date: Wed, 2 Apr 2003 08:51:46 +0200
Subject: Problem: 1040, too many connections
References: <3e899a43$0$21745$9b622d9e@news.freenet.de> <Xns9350CD870CCD9cpl19ghumspamgourmet@62.153.159.134>
Message-ID: <3e8a8e35$0$14308$9b622d9e@news.freenet.de>

> Maybe the default number of connections with 4.0.12 is smaller? Look in
> mysql.conf
It is the same default (100).
Also a value of 1000 didn't help to solve the problem.

Greetings,
 Susanne
--






From spam at mattbergin.co.uk  Mon Apr 14 13:14:57 2003
From: spam at mattbergin.co.uk (Matt Bergin)
Date: Mon, 14 Apr 2003 18:14:57 +0100
Subject: Reference Tracking
In-Reply-To: <Z8xma.11688$LB6.322565@news1.tin.it>
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it>
Message-ID: <b7eqag$bl2$1@wisteria.csv.warwick.ac.uk>

Alex Martelli wrote:
...
> In your case, you were talking about tracking objects of a specific
> different type -- not ones that Python will implicitly cache and
> reuse, such as small integers.  Thus, you do not necessarily need
> to work with _changes_ in refcounts rather than with refcounts
> directly -- I think using the refcounts directly may be simpler.

Thanks very much for putting up with my questions! Hopefully I'll be 
answering on comp.lang.python someday...

-- 
Matt Bergin




From jason at tishler.net  Fri Apr  4 09:51:57 2003
From: jason at tishler.net (Jason Tishler)
Date: Fri, 04 Apr 2003 09:51:57 -0500
Subject: Cygwin Python regression testers wanted
In-Reply-To: <v8pun8oringf6@corp.supernews.com>
References: <mailman.1049390947.15214.python-list@python.org>
 <v8pun8oringf6@corp.supernews.com>
Message-ID: <20030404145157.GC1336@tishler.net>

Tim,

On Thu, Apr 03, 2003 at 08:24:22PM -0700, Tim H wrote:
> I know you didn't ask for Win2000, but my box gives different results:
> 
> 169 tests OK.
> 1 test failed:
>     test_dircache

What do you get when you run in verbose mode?

    python -E -tt /usr/lib/python2.2/test/regrtest.py -l -v test_dircache

Are you using NFTS or FAT?

> test_commands does not fail,

The README has the following:

    4. Due to a recent Python change, the test_commands regression test
    will fail if the user and/or group names of the root directory
    (i.e., "/") contain spaces.  Due to a recent Cygwin change, the
    test_commands regression test will fail if the root directory (i.e.,
    "/") has certain ACLs applied.  Both of these issues have been
    corrected in Python CVS.

> plus other differences.

What are the other differences?

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From achrist at easystreet.com  Mon Apr 28 22:54:03 2003
From: achrist at easystreet.com (achrist at easystreet.com)
Date: Mon, 28 Apr 2003 19:54:03 -0700
Subject: Installer Problem WinNT to Win 98
References: <3EADD5C2.ABC48D5E@easystreet.com> <Xns936C27AC56317JanC@213.224.134.81>
Message-ID: <3EADE94B.379C63A0@easystreet.com>

JanC wrote:
> 
> achrist at easystreet.com schreef:
> 
> Did you use the Unicode version of wxPython?  

Thanks, but no.  I've got wxPythonWIN32-2.4.0.7-Py22.exe, not
wxPythonWIN32-2.4.0.7u-Py22.exe.  


I've just spent some time looking inside the various pyd files that
the McMillan installer did package with my app to see what else I might
need.  I found:

In almost all of them:  MSVCRT.DLL.  Is there any way that a pure
vanilla Win 98 machine (very little installed software, nostly used
to run what came with the computer) might be missing this file?  Is it
legit for me to redistribute it with a python app (I do have an MSVC
license)?  

Also:
	
	ADVAPI32.dll
	Shell32.DLL
	OLEAUT32.DLL
	Kernel32.DLL
	User32.DLL

Most of these are used by the win32api.pyd file, which I guess my
app is using to handle registry settings.  Any way that any of these
might be missing from said Win 98 machine?  


Al



From gh at ghaering.de  Tue Apr 29 17:13:41 2003
From: gh at ghaering.de (=?ISO-8859-1?Q?Gerhard_H=E4ring?=)
Date: Tue, 29 Apr 2003 23:13:41 +0200
Subject: apache modules in python ?
In-Reply-To: <pan.2003.04.29.22.09.59.56666@sdf-eu.org>
References: <pan.2003.04.29.22.09.59.56666@sdf-eu.org>
Message-ID: <3EAEEB05.4030808@ghaering.de>

Geirgat Jonas wrote:
> Is it possible to write modules for the apache http server in python ?

Did you check out mod_python, yet? If not, do so :-)

The other solutions (mod_snake, pyapache) seem to be unmaintened.

-- Gerhard




From tim_one at email.msn.com  Fri Apr 11 01:10:17 2003
From: tim_one at email.msn.com (Tim Peters)
Date: Fri, 11 Apr 2003 01:10:17 -0400
Subject: Adding item in front of a list
In-Reply-To: <Xns93596C6F1EA0Eduncanrcpcouk@127.0.0.1>
Message-ID: <LNBBLJKPBEHFEDALKOLCMEJFEGAB.tim_one@email.msn.com>

[Tim]
> Guido says he also wishes list.insert() had been defined with the
> arguments in the opposite order, so that list.insert(object) could
> have a natural default index argument of 0.  I'd like to change that
> too, but it's clearly too late for that one.

[Duncan Booth]
> I don't actually see the problem with this.

"this" doesn't make sense except as referring to Guido's desire that the
arguments had been defined in the other order.  That's what it's too late
for.

> list.insert() currently takes exactly two arguments, therefore there
> would be no backwards compatibility issues with allowing it to take
> one or two arguments where a single argument acts as though you had
> given two but the first defaulted to 0.

That's a different scheme than the one mentioned.  It violates the "natural"
in "have a natural default index argument of 0".

> Just like 'range' does in fact!

That's right, and Guido regrets range's unnatural definition.  He won't add
another like that.

    def insert(self, object, index=0):
        # insert object at self[index:index]

would have been natural.

    # Since the second argument to insert is optional, its default
    # must be specified as an object a user can't possibly pass,
    # else the "but I didn't want a second argument" case can be
    # confused with the two-argument case.
    _unique = []
    def insert(self,
               usually_index_but_object_if_2nd_arg_missing,
               object_if_specified_at_all=_unique):
        if object_if_specified_at_all is _unique:
            object = usually_index_but_object_if_2nd_arg_missing
            index = 0
        else:
            object = object_if_specified_at_all
            index = usually_index_but_object_if_2nd_arg_missing
        # insert object at self[index:index]

is a conceptual mess.

While it's a popular pastime on c.l.py <wink>, Guido rarely goes for
spelling ideas whose only virtue is that they assign a strained meaning to a
previously forbidden or undefined construct.  For that reason, only
rightmost arguments ever become defaultable, and new defaultable arguments
always come after previously accepted arguments.

living-proof-that-people-can-learn-to-type-0-when-they-want-0-ly y'rs
    - tim





From koenig at v-i-t.de  Sun Apr 27 17:11:39 2003
From: koenig at v-i-t.de (Mirko Koenig)
Date: Sun, 27 Apr 2003 23:11:39 +0200
Subject: python shell as standalone
Message-ID: <b8hh2b$9vc$1@news.tu-darmstadt.de>

Hi

I wrote a modul that gives me some extensions to the python shell.
E.g. displaying complex number in different ways.
Now i want to make the shell, with this modul loaded, a stand alone
programm, so that i can run it unter win/dos without having python
installed.

Is that possible? If yes how?
I can't find documentation on creating standalone code with python.

Thanx 

Mirko Koenig



From sdieselil at yahoo.com  Fri Apr  4 11:21:16 2003
From: sdieselil at yahoo.com (sdieselil)
Date: 4 Apr 2003 08:21:16 -0800
Subject: What are the differences between SWIG and Boost.Python?
Message-ID: <169a7283.0304040821.9ad0976@posting.google.com>

Hi

I want to write C extension for Python. I know that there are 2 most
popular libraris for this purpose: SWIG and Boost.Python. What are the
differences between them? Which one is better for little extensions
and which is better for large-scale extensions?



From Thomas.Sunshine at web.de  Thu Apr 10 04:44:04 2003
From: Thomas.Sunshine at web.de (Thomas M.)
Date: Thu, 10 Apr 2003 10:44:04 +0200
Subject: design patterns in german (object-oriented)
Message-ID: <b73asl$m64$1@newsreader2.netcologne.de>

hi,
are there any good german sides or books with design patterns for 
object-oriented programming?



From staschuk at telusplanet.net  Sun Apr  6 16:02:17 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Sun, 6 Apr 2003 14:02:17 -0600
Subject: Python & Music: Performance question
In-Reply-To: <1759581e.0304060218.164d4cd4@posting.google.com>; from forodejazz@yahoo.es on Sun, Apr 06, 2003 at 03:18:50AM -0700
References: <1759581e.0304060218.164d4cd4@posting.google.com>
Message-ID: <20030406140217.C769@tibia.amotlpaa.bogus>

Quoth David Asorey ?lvarez:
  [...]
>    I want define a class, Note, which represents a note. It has two
> attributes, pitch and duration (quarter note, half note, etc). There
> will be a lot of instances of this object, maybe hundreds of notes,
> and the program needs access to all of this objects and read its
> attributes.
>    Which implementation will be more efficient?
  [...]

I don't know about speed (you should profile when you find it
matters), but for memory efficiency, this sounds like a job for
the Flyweight design pattern.

-- 
Steven Taschuk                           staschuk at telusplanet.net
"I'm always serious, never more so than when I'm being flippant."
                            -- _Look to Windward_, Iain M. Banks




From gansevle at ewi.utwente.nl  Tue Apr  8 16:28:31 2003
From: gansevle at ewi.utwente.nl (Fred Gansevles)
Date: Tue, 08 Apr 2003 22:28:31 +0200
Subject: Is there any way to make repr(aString) use double quotes?
References: <698f09f8.0304081057.1ba7041f@posting.google.com>
Message-ID: <b6vbd2$528$1@netlx020.civ.utwente.nl>

Jeremy Fincher wrote:

> The topic really says it all -- I'd like to *force* repr(aString) to
> use double quotes.  Is that possible?
> 
> If it's not, then is it possible to achieve the same thing (everything
> escaped properly, etc.) through other means?
> 
> Thanks,
> Jeremy

the following function does the trick:

def qrepr(s):
     "repr with double-quotes"
     # make one symbol special that isn't treated special in a string, in
     # this case '@'.
     # now modify the string in such a manner that quoting-characters
     # become non-quoting characters, get the repr and reverse the whole
     # operation.
     # replace all '@'-signs with '@@'
     s = s.replace('@', '@@')
     # replace all '"' with '@+@'
     s = s.replace('"', '+ at +')
     # replace all "'" with '@:@')
     s = s.replace("'", '- at -')
     # get the repr
     s = repr(s)[1:-1]
     # and reverse the changes
     s = s.replace('- at -', "'")
     s = s.replace('+ at +', r'\"')
     s = s.replace('@@', '@')
     return '"%s"' % s

if __name__ == '__main__':
    # test
    for s in ('xy', 'x at y', 'x"y', "x'y", 'x@@y', 'x+ at +y', 'x- at -y'):
        print repr(s), qrepr(s)





From sholden at holdenweb.com  Mon Apr 28 08:03:58 2003
From: sholden at holdenweb.com (Steve Holden)
Date: Mon, 28 Apr 2003 12:03:58 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <mailman.1050698431.29009.python-list@python.org> <C8_na.646$dP1.2863@newsc.telia.net> <eppstein-653307.16014718042003@news.service.uci.edu> <Fy7oa.53694$T34.1483263@news2.tin.it> <mailman.1050772891.13211.python-list@python.org> <Iuhoa.58588$T34.1568946@news2.tin.it> <KIioa.58844$T34.1577116@news2.tin.it> <9tSoa.142846$It5.87082@news2.central.cox.net> <nCUoa.41676$LB6.941312@news1.tin.it> <pan.2003.04.22.03.08.58.526973@mn.rr.com> <3japa.44301$LB6.997057@news1.tin.it>
Message-ID: <OM8ra.21447$WQ4.16996@news2.central.cox.net>

"Alex Martelli" <aleax at aleax.it> wrote in message
news:3japa.44301$LB6.997057 at news1.tin.it...

[why Tim and Guido are dead good ...]

> To be properly thankful that we're not using an Alex-designed
> language, we might want to look at a simple example from the
> domain-specific language BBL (Bridge-Bidding Language) which I
> designed and implemented in the '80s [well before I met Python --
> but I was quite familiar with Icon, REXX and other highly
> readable very-high-level languages, so that's no excuse;-)].
>
> Here's the BBL expression to determine whether the current hand
> is suitable for a weak-notrump opener...:
>
> & (                      # conditions for a weak-notrump opener
>     s ( 1 g 4  3  3  3   # shape 4333 (any), or...
>         2 g 4  4  3  2   #       4432 (any), or...
>         3   3- 3- 3- 5   #       5332 w 5 clubs, or...
>         4   3- 3- 5  3-  #       5332 w 5 diamonds;
>       )
>     < ( 11                    # moreover, a 12-14 range for
>         \+ SHDC c( 4 3 2 1 0) # normal Milton-Work point-count
>         15
>       )
>   )
>
> There -- aren't you glad I don't do language design any more?-)
>

On this evidence it's arguable that you never did! This language was
*designed*?

but-anyway-let's-all-be-grateful-for-python-ly y'rs  - steve
--
Steve Holden                                  http://www.holdenweb.com/
Python Web Programming                 http://pydish.holdenweb.com/pwp/







From eric.brunel at pragmadev.com  Fri Apr 25 03:41:33 2003
From: eric.brunel at pragmadev.com (Eric Brunel)
Date: Fri, 25 Apr 2003 09:41:33 +0200
Subject: Tkinter - tkFileDialog - need help
References: <oprn5mygrymu3s0q@news.cis.dfn.de>
Message-ID: <b8aofk$iqf$1@news-reader14.wanadoo.fr>

Klaus Meyer wrote:
> Hi,
> 
> my script uses a directory-select box.
> But it's not possible to find an select a directory that has attribut 
> "hidden" inside the dialog-box (askdirectory).
> 
> It's possible to make "hidden" directory viewable? How? Thanks!
> 
> Win2000:
> 
> import tkFileDialog
> tkFileDialog.askdirectory(initialdir="C:/", title="Test")

Tk uses the regular "open file" dialog from Windows , so the only known way (at 
least to me) is to configure Windows to display the hidden files. This is 
somewhere in the folder options, can't remember exactly where (I don't have a 
Windows box at hand for now...).

HTH
-- 
- Eric Brunel <eric.brunel at pragmadev.com> -
PragmaDev : Real Time Software Development Tools - http://www.pragmadev.com




From ramen at lackingtalent.com  Tue Apr  1 15:18:05 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 20:18:05 -0000
Subject: Terminal-based text editor with color syntax hilighting
References: <slrnb8hvka.pki.ramen@lackingtalent.com> <mailman.1049169002.28920.python-list@python.org> <slrnb8j93q.421.ramen@lackingtalent.com> <mailman.1049223987.18476.python-list@python.org>
Message-ID: <slrnb8jq45.5u1.ramen@lackingtalent.com>

In article <mailman.1049223987.18476.python-list at python.org>,
Lulu of the Lotus-Eaters wrote:
> Dave Benjamin <ramen at lackingtalent.com> wrote previously:
>|Are there any others? What about Fte, I wonder... is that editor/IDE still
>|being actively maintained? According to their website it supports Python
>|color hilighting, but I'm not sure if the s-lang version does. I'll check
>|into it.
> 
> Unfortunately, AFAIK FTE only works in Unix-like systems over XTerm.  A
> console session doesn't do FTE (under Linux, OS/2 has both console and
> PM versions of FTE).  As with the OP, this rules out using FTE on most
> of my connections--my web host doesn't run an X client, my OS/2, Windows
> and MacOSX machines don't run an X server).

Actually, believe it or not, I just looked into this and there are three
versions of FTE, one for X, one for a terminal (like SSH) and one for
console (I imagine they do direct screen-writes, since it seems to require
root privileges). I was able to get color hilighting and pulldown menus
working with a fresh install. So, this is a viable alternative.

> IMO, the best option for the rather common scenario of remote SSH
> connection is Jed. But I would welcome learning of others.  I *have*
> been using vim on one host recently, and have kinda-sorta learned basic
> navigation, but it won't stick in my head after a month (I know from
> experience).  Likewise with emacs.

That's been my experience also, more or less. I really like the speed and
familiarity of Jed. It gets a little screwy about Python triple-quotes
sometimes, depending on what version you have, but overall its
auto-indentation works great for Python.

> Jed has this wonderful feature of having topbar menus that actually let
> you visually select the most common actions (without needing to remember
> a cryptic key-sequence).  This fabulous design had only been in DOS
> text editors since 1982, so it's nice that Linux is starting to catch up
> with that.

Hehe... if you like that, I'd say FTE is even more in that direction. It
reminds me of editors like Boxer and the TurboC++ environment. It's a bit
more of an IDE than a plain-vanilla text editor though.

<ot>
DOS had the advantage of being relatively predictable, at least as far as
text mode was concerned. Running full-screen text applications remotely was
another story entirely (anyone remember Doorway?) because of the direct
screen writes... anyhow, I'd imagine developing full-screen text apps for
*nixes (and being compatible with the various kinds of terminals) is a more
difficult endeavor, even with the help of ncurses/s-lang/etc.
</ot>

Thanks,
Dave



From altis at semi-retired.com  Thu Apr 10 23:34:35 2003
From: altis at semi-retired.com (Kevin Altis)
Date: Thu, 10 Apr 2003 20:34:35 -0700
Subject: Promoting Python
Message-ID: <qGqla.61$Tj7.91655@news.uswest.net>

One of the hot topics at PyCon was Promoting Python. We had two Open Space
sessions at PyCon that were relevant:

  http://www.python.org/cgi-bin/moinmoin/MarketingPythonOpenSpace

  http://www.python.org/cgi-bin/moinmoin/PythonOrgWebsiteRedesignOpenSpace

I would like to invite you to join the marketing-python list if you're
interested in contributing ideas or would like to help implement plans to
increase the user-base of Python and make sure that you never have to answer
the question "What is Python?" ever again. If you have marketing, PR, or
business experience that you can share that is a big plus!

  http://pythonology.org/mailman/listinfo/marketing-python

One of the top priorities is to improve the python.org website. A mailing
list has been created to coordinate activities. We will need volunteers to
help make python.org a better site. If you have experience building large
web sites, are an experienced web designer, or can help with the nuts and
bolts of the web site maintenance your contributions to making a better
python.org will be quite welcome.

  http://mail.python.org/mailman/listinfo/pydotorg-redesign

I recommend that you take time to read the marketing-python and
pydotorg-redesign archives, at least the posts from the last few weeks to
catch-up with the discussions.

Welcome to the revolution!

ka
---
Kevin Altis
altis at semi-retired.com
http://radio.weblogs.com/0102677/





From mstenner at phy.duke.edu  Wed Apr  9 13:55:07 2003
From: mstenner at phy.duke.edu (Michael Stenner)
Date: Wed, 9 Apr 2003 13:55:07 -0400
Subject: a size-intelligent progress bar?
In-Reply-To: <1PYka.1140$RX1.361@newsfep3-gui.server.ntli.net>; from tomchance@gmx.net on Wed, Apr 09, 2003 at 06:20:02PM +0100
References: <1PYka.1140$RX1.361@newsfep3-gui.server.ntli.net>
Message-ID: <20030409135507.A30646@phy.duke.edu>

On Wed, Apr 09, 2003 at 06:20:02PM +0100, Tom Chance wrote:
> Hullo,
> 
> I'm making a shell-based program that includes a progress bar (like wget's
> one). At the moment, I've got it printing an equals sign every 2% so I get
> a progress bar 50 "="s wide, which fits my Konsole perfectly. This isn't
> very handy for other shell sizes though :)
> 
> Does anyone have, or can anyone point me in the direction of a way of making
> a decent progress bar that will automagically detect the width of the shell
> open, even resizing it mid-"progress"?

I don't know how portable it is, but the environment variable COLUMNS
appears to tell me my terminal width and follows resizes.  If you
check that on each update, it should do what you want.

				-Michael
-- 
  Michael Stenner                       Office Phone: 919-660-2513
  Duke University, Dept. of Physics       mstenner at phy.duke.edu
  Box 90305, Durham N.C. 27708-0305




From mhbaek at korea.com  Fri Apr 18 00:08:26 2003
From: mhbaek at korea.com (mhbaek)
Date: 17 Apr 2003 21:08:26 -0700
Subject: How do confirm to Email receive?
Message-ID: <fd5e8b3d.0304172008.1e48e7fc@posting.google.com>

Hello ^^

I will make "confirm to Email receive System" for python

system platfom is MS Windows 2000 server.

I have problem so, how confirmation receive user IP and Time. 

Please, Tell me you did it.

Thank you.



From nhv at cape.com  Sat Apr 12 16:43:58 2003
From: nhv at cape.com (Norman Vine)
Date: Sat, 12 Apr 2003 16:43:58 -0400
Subject: PCA principal component analysis
References: <b1ab591.0304090601.302cc4b6@posting.google.com> <Z7nla.45$W4.8475@news.uchicago.edu> <7oZla.5396$LB6.150919@news1.tin.it>
Message-ID: <b79tj4$jhv$1@main.gmane.org>

Alex Martelli wrote
> Suchandra Thapa wrote:
> 
> > sebastien <s.thuriez at laposte.net> wrote:
> >> Is there any PCA analysis tools for python ?
> >> If it does, do you have any idea on how well it would scale ?
> >> 
> >> I have already seen PyClimate (but it is not available for Windows
> >> which will be one of the target). Is there some LAPACK like packages ?
> > 
> > Have you considered using R (www.r-project.com) and RS/Python
> 
> Hmmmm -- this site's most succint homepage's in English, but all
> links from it seem to be trying to show me lots and lots of Japanese
> text.  What's the right way to get R docs & other downloads these
> days for the non-Japanese-speaker...?

try
http://www.r-project.org/

Norman





From staschuk at telusplanet.net  Fri Apr 18 13:25:04 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 18 Apr 2003 11:25:04 -0600
Subject: Classes in Python
In-Reply-To: <roy-B8E873.12412018042003@reader1.panix.com>; from roy@panix.com on Fri, Apr 18, 2003 at 12:41:20PM -0400
References: <3EA01A9B.4030209@vnet.net> <roy-B8E873.12412018042003@reader1.panix.com>
Message-ID: <20030418112504.A931@tibia.amotlpaa.bogus>

Quoth Roy Smith:
  [...]
> Extra credit for deranged minds: figure out a way to make:
> 
> a = c()
> b = c()
> 
> try:
>    print b.x
> except AttributeError:
>    print "wugga wugga"
> 
> a.x = "foo"
> 
> try:
>    print b.x
> except AttributeError:
>    print "wugga wugga"
> 
> print "wugga wugga" and then print "foo".  [...]

class c:
    def __setattr__(self, attribute, value):
        setattr(self.__class__, attribute, value)

Did you have something cleverer in mind?

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From cbuffer at removethis.rosecott.ukfsn.org  Tue Apr 29 19:24:09 2003
From: cbuffer at removethis.rosecott.ukfsn.org (Ken Parkes)
Date: Wed, 30 Apr 2003 00:24:09 +0100
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <80dac0f4.0304281047.9143cd0@posting.google.com>
Message-ID: <1051658594.16051.0@demeter.uk.clara.net>

On Mon, 28 Apr 2003 11:47:24 +0000, Ionutz Borcoman wrote:

> Hi,
> 
> Why don't you try "How to Think Like a Computer Scientist: Learning with
> Python"
> 
> http://www.ibiblio.org/obp/thinkCSpy/
> 
> It's free and it's easy reading for people with no programming experience.
> 
> Ionutz

I hate reading on a screen and when I try to print this book it cuts off
the last two characters on each line, and the last two lines on each page.
Does it work for other people?  I've never had this problem with anything
else.  ( using cups on Mandrake 9 )

Ken.




From lekishvili at python.qartu.com  Mon Apr  7 09:17:26 2003
From: lekishvili at python.qartu.com (Giorgi)
Date: 7 Apr 2003 06:17:26 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>	 <6ee58e07.0304061758.35070e7c@posting.google.com> <mailman.1049681412.942.python-list@python.org>
Message-ID: <a3b5b608.0304070517.280b93bf@posting.google.com>

Ian Bicking <ianb at colorstudy.com> wrote in message news:<mailman.1049681412.942.python-list at python.org>...
> On Sun, 2003-04-06 at 20:58, Lothar Scholz wrote:
> > No its not a naturally tailored web-programming language, because it
> > is very hard to integrate in Server Pages. You have to change the
> > indentation syntax/semantic of python to embedd it in a real useable
> > way.  This is done by Webware but it looks very strange to normal python
> > programmers.
> 
> OTOH, it's rather easy to integrate HTML into Python code -- easier than
> many other languages (maybe Perl is better, PHP isn't).

exactly!
> 
> > Of course you can use template systems. But this is not always a good
> > economic decision.
> 
> If you want a quick solution, then I'd recommend using no template at
> all.  If you want real templating, Server Pages aren't a particularly
> good implementation.
> 
>   Ian



From alessandro.bottoni at infinito.it  Tue Apr  8 18:05:19 2003
From: alessandro.bottoni at infinito.it (Alessandro Bottoni)
Date: Tue, 8 Apr 2003 22:05:19 +0000
Subject: How about Python + Mozilla for large apps?
Message-ID: <200304082200.10255.alessandro.bottoni@infinito.it>

Is there any project aimed to embed the Python interpreter inside 
Mozilla OR to make available the XUL/XBL GUI engine of Mozilla to the Python 
interpreter OR to create a full-capability bridge between Mozilla and Python 
(maybe via XPCOM)?

I'm thinking that coupling the sophisticated UI features of Mozilla (XUL+XBL) 
to a powerful interpreter like Python (or Java, Ruby, etc...) would allow the 
development of full-blown, well-engineered, cross-platform applications: 
unlimited access to a lot of different RDBMS, unlimited access to the hosting 
platform (Windows, MacOS, and Linux/Unix) and all of its services, unlimited 
access to a lot of different plug-ins (PyNum, PyGame, py*) and so on.

It seems that this idea was somehow already presented by Ian Oescher in this 
article: http://www.mozilla.org/docs/xul/xulnotes/xulnote_oven.html
But I'm unable to find any more information on this topic.

Thanks for your attention.

BTW: As long as I can understand PyXPCOM is not what I'm looking for. It is a 
way to create XPCOM components in Python and use them in Mozilla. Instead, 
I'm looking for a (simple) way to have access to the _whole_ Python 
interpreter from within a Mozilla javascript OR to have access to the whole 
XUL/XBL technology of Mozilla from within a Python script (like it was 
WxPython). Probably I'm wrong, given that ActiveState created Komodo with 
PyXPCOM, but... I have this feeling.

The nsIProcess or nsILocalFile interfaces of Mozilla could probably do this 
job by invoking the interpreter but how about setting/getting parameters, 
communicating with the running interpreter and using a single copy of the 
interpreter for many different task? I'm rather thinking to something like 
mod_python for Apache.
 
---------------------
Alessandro Bottoni






From peter at engcorp.com  Wed Apr 16 08:48:13 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 16 Apr 2003 08:48:13 -0400
Subject: do python's nifty indentation rules spell the death of   one-liners?
References: <87he90b8rp.fsf@jidanni.org> <mailman.1050452786.8192.python-list@python.org> <b7ik4p$1d2t1$1@ID-169208.news.dfncis.de> <b7iksv$1ba93$1@ID-169208.news.dfncis.de>
Message-ID: <3E9D510D.41ECC5E@engcorp.com>

"Greg Ewing (using news.cis.dfn.de)" wrote:
> 
> Oops, looks like I spoke too soon, because this
> *doesn't* work:
> 
> all:
>         python -c '\
>         print "Hi"\
>         for i in range(10):\
>           print i'
> 
> It seems that make puts the whole command on one line,
> no matter how you lay it out in the Makefile. Bugger.

Changing the enclosing quotes to triple-quotes, or explicitly
specifying newlines via \n is required... in this case and
any other with strings split across multiple lines but where
you need to preserve the newlines.

-Peter



From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr 24 10:47:47 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 24 Apr 2003 14:47:47 GMT
Subject: A new and very robust method for doing file locking over NFS?
References: <lcu1cw525a.fsf@gaffa.mit.edu>
Message-ID: <slrnbafu8h.ek.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <lcu1cw525a.fsf at gaffa.mit.edu>, Douglas Alan wrote:
> I'd like to do file locking over NFS without using lockd.  The reason
> I want to avoid using lockd is because many lockd implementations are
> too buggy.
> 
> It is fairly easy to avoid using lockd -- just avoid using lockf() to
> lock a file.  Instead of using lockf(), lock a file by creating a lock
> file that you open with the O_CREAT | O_EXCL flags.
<snip>

O_EXCL doesn't work with NFS prior to version 3, and maybe not even
then.

The usual solution for creating a lock file in an NFS-safe way is:

1. Create a file in the target directory with a unique name (usually
   involving hostname and pid).
2. Attempt to link the lock file name to this file (using the temporary
   file name).  Ignore whether this succeeds or fails, as the result is
   not reliable.
3. stat() the target file name and record the number of links.
4. Unlink the original file name.
5. If the number of links recorded in step 3 was 2, continue, else fail.
   (Or try again, up to some maximum number of times.)

A cleaner solution would be to use a client-server protocol instead of
synchronising through the file system, but that's not so easy to do.

In Python:

    hostname = socket.gethostname().split('.')[0]
    unique_name = '%s.%s-%d' % (lock_name, hostname, os.getpid())
    for i in range(5):
        os.close(os.open(unique_name, os.O_CREAT|os.O_WRONLY))
        try:
            try:
                os.link(lock_name, unique_name)
            except OSError:
                pass
            if os.stat(lock_name).st_nlink == 2:
                break
        finally:
            os.unlink(unique_name)
    else:
        raise OSError((errno.EEXIST, os.strerror(errno.EEXIST), lock_name))



From claird at lairds.com  Mon Apr 28 15:39:53 2003
From: claird at lairds.com (Cameron Laird)
Date: Mon, 28 Apr 2003 19:39:53 -0000
Subject: Textbooks?
References: <oproar1sjtx3zh5c@news.mindspring.com> <a5415304.0304271335.2f569879@posting.google.com>
Message-ID: <var0s9a81psp83@corp.supernews.com>

In article <a5415304.0304271335.2f569879 at posting.google.com>,
Ron Stephens <rstephens at vectron.com> wrote:
			.
			.
			.
>The only actual "Textbook" I am aware of is Deitel and Deitel's Python
>book, which I found pretty good. However, Learning Python by Lutz and
			.
			.
			.
I found the Deitel and Deitel book unsatisfying;
others here "dinged" it harder.

*Making Use of Python* <URL: http://
www.unixreview.com/documents/s=7822/ur0302m/ > is
aimed at college classes, so perhaps it qualifies
as an "actual textbook".  I do NOT recommend it,
though.
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From eppstein at ics.uci.edu  Thu Apr  3 16:09:50 2003
From: eppstein at ics.uci.edu (David Eppstein)
Date: Thu, 03 Apr 2003 13:09:50 -0800
Subject: computer science with python
References: <m3istvtodl.fsf@moroni.pp.asu.edu>
Message-ID: <eppstein-88592A.13095003042003@news.service.uci.edu>

In article <m3istvtodl.fsf at moroni.pp.asu.edu>,
 David Bear <david.bear at asu.edu> wrote:

> I purchased the OReilly book "Computer Science with Perl" (or
> something similar and found it very fun -- even though I really hate
> Perl.  It made me think that python would be a terrific language to
> demonstrate some key computer science concepts.
> 
> Anyone know if a book has already been done?  I find that many CS
> language concepts are debated/discussed here and have to be careful
> when I look at this group because I become too interested in reading
> all the conceptual stuff rather than getting a specific question
> answered.  Seems this list would provide a good starting point for
> such a book.  I image it could be many volumes...

I don't know about textbooks, but I have been using Python (or at 
least, pseudocode syntax very very similar to Python) in my algorithms 
lectures for the last couple of years -- it turns out to be similar in 
a lot of ways to the pseudocode already used by standard algorithms 
texts such as Cormen et al or Goodrich&Tamassia (specifically in its 
lack of variable declarations and indentation-based blocking).  As part 
of this project I've put a random set of implementations of some of the 
algorithms on a web site http://www.ics.uci.edu/~eppstein/161/python/
This isn't a programming course, I don't expect the students to know or 
even learn Python, but the readability of the language gives me a 
convenient notation for describing algorithms in general.

-- 
David Eppstein                      http://www.ics.uci.edu/~eppstein/
Univ. of California, Irvine, School of Information & Computer Science



From bourassa-m at rmc.ca  Tue Apr  8 11:28:32 2003
From: bourassa-m at rmc.ca (bourassa)
Date: Tue, 08 Apr 2003 11:28:32 -0400
Subject: Newbie: Python + C Question
Message-ID: <3e92ebe8$1@rmcnt03a.rmc.ca>

Hi there.  I have a question about Python and the interaction with C. 
 From Python I can use C extensions.  But does doing so speed up 
execution speed?  When Python uses the C extension does it just process 
it in an interpreted fashion or does the code written in C run like a 
compiled C program?  Or have I really misunderstood the whole concept, 
in which case do you have a good reference :-)  TIA

Majb




From ckea25d02 at sneakemail.com  Tue Apr 15 18:50:21 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Wed, 16 Apr 2003 10:50:21 +1200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <1050390062.742825@yasure>
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org> <b7di0n$dd6gj$1@ID-169208.news.dfncis.de> <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net> <1050390062.742825@yasure>
Message-ID: <b7i1mm$17cpu$1@ID-169208.news.dfncis.de>

Donn Cave wrote:
> The idea that repr() must marshal the object is a sorely misguided
> extension of this.  It is not consistently supported enough to be
> clearly useful, and it has done some harm, e.g., it apparently
> contributed to a move to make the repr for floats to display full
> precision.

No, I think the reason for that is to provide something
which can be relied upon to tell you the "whole truth"
about the value of a float; str(aFloat) lies to you slightly,
exacerbating people's confusion about the behaviour of
float arithmetic.

Of course, it's still not *quite* the whole truth --
to get that you'd have to display it in binary.
But it's better than it used to be.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From dave at boost-consulting.com  Wed Apr 23 05:13:37 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Wed, 23 Apr 2003 05:13:37 -0400
Subject: Using __repr__ or __str__ for own printable class?
References: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net> <f6a31cb8.0304220810.547f5acd@posting.google.com>
 <oRKdnYDxiYyb5TijXTWcpA@comcast.com> <1051075309.87191@yasure>
Message-ID: <84d6jdbmpq.fsf@boost-consulting.com>

"Donn Cave" <donn at drizzle.com> writes:

> Quoth "Terry Reedy" <tjreedy at udel.edu>:
> | "Ted Drain" <teddrain at earthlink.net> wrote in message
> | news:f6a31cb8.0304220810.547f5acd at posting.google.com...
> |> My $.02:
> |>
> |> We use Python as a scripting interface to an underlying C++ system.
> |> The users write Python code to do what they want.  We ended up writing
> |> __repr__ and __str__ in our classes to be the same thing because our
> |> users were very annoyed when they typed:
> |>
> |> >>> s
> |> >>> print s
> |>
> |> and got different results.
> |
> | If you think of repr(o) as the codewriter view of an object o and
> | str(o) as the (passive) user view of o, then printing repr(o) in
> | interactive (codewriter) mode makes perfect sense.
>
> Coincidentally, today I was looking into a little Mailman problem.
> Its subclass of Message defines __repr__ to return __str__().
> Unfortunately, __str__ is a fairly involved function that not only
> returns information that's completely inappropriate for situations
> like the interpreter, it's also prone to failure with Unicode troubles.
>
> So here's this object.  You can't repr() it, including implicitly
> from the interpreter prompt - you get a traceback instead.  Of course
> that also goes for the list object that contains this object.  To
> me, this is not right, and maybe it ought to be another point to
> consider if we're reconsidering the documentation for __repr__.
> Wish I could specify this better, but I believe there's a minimalist
> way to think about implementing __repr__ that will tend to make it
> less interesting but more reliable.

One idea might be:

    # pseudocode
    class object:
         def __repr__(self):
            if hasattr(self, '__getinitargs__'):
                return (
                    repr(self.__class__)
                  + repr(self.__getinitargs__())
                )
            else:
                return # whatever we do today

I know that would only handle a small class of types, since most
aren't picklable, but it might encourage people to implement
__getinitargs__ to get the right __repr__ behavior.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com



From ckea25d02 at sneakemail.com  Mon Apr 14 23:33:51 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 15 Apr 2003 15:33:51 +1200
Subject: Reference Tracking
In-Reply-To: <Z8xma.11688$LB6.322565@news1.tin.it>
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk> <z7lma.10014$LB6.267626@news1.tin.it> <b7e60e$re$1@wisteria.csv.warwick.ac.uk> <Z8xma.11688$LB6.322565@news1.tin.it>
Message-ID: <b7ftuh$fu1e$1@ID-169208.news.dfncis.de>



  -- any number of system-maintained
 > data structures may have attributes and items that happen to
 > be worth 1 (or 0, or other small, popular numbers).
 >
 > import gc
 > gc.get_referrers(1)
 >
 > will at any time give you a list of all containers which have
 > a reference to the value 1, if you're curious.  Of course, having
 > such a container in hand will not necessarily make it easy to
 > understand WHAT the container in fact IS -- but many will be
 > dicts with string keys, which lets you do some detective work;-).
 >
 >
 > Note that any Python function call always passes VALUES -- i.e.
 > there is no difference between your call to sys.getrefcount in
 > the above snippet and one that directly does sys.getrefcount(1).
 > In either case, it's the VALUE (1) that you're asking info about;
 > it makes no difference HOW you obtain the value to pass (I guess
 > this is probably already clear to you, but, just in case...).
 >
 >
 > In your case, you were talking about tracking objects of a specific
 > different type -- not ones that Python will implicitly cache and
 > reuse, such as small integers.  Thus, you do not necessarily need
 > to work with _changes_ in refcounts rather than with refcounts
 > directly -- I think using the refcounts directly may be simpler.
 >
 >
 > Alex
 >


Alex Martelli wrote:
> In practice, Python does such caching for small integers, and 1
> is clearly a popular value

Just for fun, I did a quick poll of the popularities
of integers from 0-99:

 >>> for i in range(100): print i, sys.getrefcount(i)
...
0 98
1 45
2 21
3 9
4 14
5 10
6 11
7 10
8 10
9 9
10 14
11 8
12 8
13 10
14 7
15 7
16 6
17 9
18 7
19 7
20 7
21 7
22 7
23 8
24 7
25 7
26 7
27 7
28 7
29 9
30 8
31 7
32 9
33 5
34 5
35 5
....
99 5

It seems that the most popular number is 0, and the most
popular popularity is 5.

0-for-prime-minister-ly,

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From m.faassen at vet.uu.nl  Tue Apr 29 06:10:34 2003
From: m.faassen at vet.uu.nl (Martijn Faassen)
Date: 29 Apr 2003 10:10:34 GMT
Subject: Web programming and a different "type" problem
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it> <vaqvrb3ipqblf2@corp.supernews.com> <slrnbarhfb.67t.jon+usenet@snowy.squish.net>
Message-ID: <b8lj2q$l14$1@newshost.accu.uu.nl>

Jon Ribbens <jon+usenet at unequivocal.co.uk> wrote:
> In article <vaqvrb3ipqblf2 at corp.supernews.com>, Cameron Laird wrote:
>> 'Doesn't work, though.  Here's where the special case
>> enters:  if only one datum with the indicated name is
>> present as input, Zope and the others return it as a
>> simple string, rather than a list with that string as 
>> its sole member.
> 
> jonpy (jonpy.sf.net) avoids this problem. You name your multiple field
> with a '*' suffix (e.g. <input type="checkbox" name="foo*">). All
> fields with a '*' suffix are *always* lists when examined in your
> Python program. All fields without a '*' suffix are *always* strings
> (if there is more than one field with the same name but no '*' suffix
> then all but one are discarded). e.g.:
> 
>  for foo in req.params.get("foo*", ()):
>    process(foo)
> 
> It seems to me that allowing the Evil User to manipulate the type of
> Python variables, as everything except jonpy seems to allow, is
> fundamentally a bad idea.

Doesn't jonpy allow the Evil User to manipulate the type of your Python
variables too? The Evil User could get rid of the '*' in the form and
submit, right? Though I grant it looks like the code will always fail
in that case, which is good. It's a neat idea.

Formulator tries quite hard to avoid the Evil User manipulating HTML forms
problem as well, by the way. Validation information is not sent to the
client but kept on the server, and requests are interpreted by it.

Regards,

Martijn
-- 
History of the 20th Century: WW1, WW2, WW3?
No, WWW -- Could we be going in the right direction?



From aahz at pythoncraft.com  Fri Apr 25 12:04:42 2003
From: aahz at pythoncraft.com (Aahz)
Date: 25 Apr 2003 12:04:42 -0400
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com> <SXcqa.6797$K35.186991@news2.tin.it>
Message-ID: <b8bmaq$mhl$1@panix3.panix.com>

In article <SXcqa.6797$K35.186991 at news2.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
>
>"rm throwaway.py" has clearly miniscule benefits in most cases, and
>the potential cost is obvious -- if you're wrong, and next week
>need that same functionality again, you're going to have to write
>and possibly debug it again.  Leaving your hard-disk strewn with
>dusty old code, however, is quite sub-optimal too -- any single
>given script matters little, but cumulatively they may make things
>untidy indeed.  Keeping your "probably throwaway scripts" in their
>own folder[s] is a good move -- and so is writing SOME docs about
>what it IS that they do, even just a couple lines worth of docstring.

I've got one throwaway script, called x.py.  Any time I need to write a
throwaway script, I edit that file, deleting the current contents first.
(I'll admit that I do usually have multiple instances of that file in
different directories, but the ones in directories get used strictly for
the projects in those directories -- my standard is the one in my home
directory.)
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

"In many ways, it's a dull language, borrowing solid old concepts from
many other languages & styles:  boring syntax, unsurprising semantics,
few automatic coercions, etc etc.  But that's one of the things I like
about it."  --Tim Peters on Python, 16 Sep 93



From llothar at web.de  Sun Apr  6 21:58:24 2003
From: llothar at web.de (Lothar Scholz)
Date: 6 Apr 2003 18:58:24 -0700
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
Message-ID: <6ee58e07.0304061758.35070e7c@posting.google.com>

lekishvili at python.qartu.com (Giorgi) wrote in message news:<a3b5b608.0304040904.146dfff5 at posting.google.com>...
> - Is Python a real web-programming language at all? I mean, has
> somebody ever used it for real, commercial tasks? Please, do not
> consider my question too trivial. There is no language equally
> suitable for all tasks, however, Python seems to me a language
> naturally tailored for web-programming; this or that way, I have more

No its not a naturally tailored web-programming language, because it
is
very hard to integrate in Server Pages. You have to change the
indentation syntax/semantic of python to embedd it in a real useable
way.
This is done by Webware but it looks very strange to normal python
programmers.

Of course you can use template systems. But this is not always a good
economic decision.



From ramen at lackingtalent.com  Tue Apr  1 13:42:08 2003
From: ramen at lackingtalent.com (Dave Benjamin)
Date: Tue, 01 Apr 2003 18:42:08 -0000
Subject: When Python outruns C++ ...
References: <mailman.1048973984.17441.python-list@python.org> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <x7of3qjh4g.fsf@guru.mired.org> <b6ckjr$903$1@panix2.panix.com>
Message-ID: <slrnb8jkg8.5u1.ramen@lackingtalent.com>

In article <b6ckjr$903$1 at panix2.panix.com>, Roy Smith wrote:
> Mike Meyer  <mwm at mired.org> wrote:
>> The trick to getting good performance out of interpreted languages is
>> doing things to take advantage of those libraries.
> 
> This is a very true statement, and one which I think is
> under-appreciated by a lot of people.
> 
> ...
> 
> The bottom line is exactly what Mike said: in interpreted languages,
> the more you can push down into the libraries and interpreter core,
> the better off you are most of the time.  And profiling helps too :-)

This has been my experience as well. It is a very gratifying discovery,
IMHO, because it encourages a less-is-more style of programming where you
say what you mean, and little else.

In general, when concise, declarative code tends to execute faster than
verbose imperative code, that is a good sign for a high-level language.

The presence of well-optimized core libraries with simple interfaces (like
Python's lists and dictionaries) encourages reuse, readability, and
efficiency, all at once.

It does require a shift in thinking for those accustomed to C++ or Java
programming. For instance, iterating through strings character-by-character
is often very efficient in these languages, but in Python you're usually
better off thinking in terms of splits, joins, replacements, regexes, etc.
Profiling is the only surefire way to know.

Dave



From claird at lairds.com  Thu Apr  3 07:26:00 2003
From: claird at lairds.com (Cameron Laird)
Date: Thu, 03 Apr 2003 12:26:00 -0000
Subject: Zope style (was: tuning our site (was: help! advocacy resources needed fast))
References: <mailman.1046819836.25704.python-list@python.org> <irdbl-1ds.ln1@news.lairds.org> <v8mjffpuph1oc2@corp.supernews.com> <4ufsl-ki3.ln1@beastie.ix.netcom.com>
Message-ID: <v8oa2oi9094a0c@corp.supernews.com>

In article <4ufsl-ki3.ln1 at beastie.ix.netcom.com>,
Dennis Lee Bieber  <wlfraed at ix.netcom.com> wrote:
>Cameron Laird fed this fish to the penguins on Wednesday 02 April 2003 
>12:54 pm:
>
>> comes from the supposition that they want stuff to look like
>> HTML, so they can use their "HTML editors".  Is *anyone* doing
>> that with Zope, though?  It would surprise me.  OK, I'm regu-
>
>        If I were actually maintaining a ZOPE site (rather than playing with 
>one on my Windows OS) I would likely have taken the time to try to 
>define tag controls for HomeSite (I'm still running v4).
			.
			.
			.
Interesting!  I want to be clear on the work-flow; you'd edit
and preview in HomeSite, then switch back to the Zope manage-
ment screens, and use the "Upload File" selection to update 
the site itself--do I have that right?  Have I just demon-
strated my ignorance of some far slicker interface?  To me,
that prospect seems frustratingly clumsy.  I like learning
new techniques, though ...
-- 

Cameron Laird <Cameron at Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html



From bbfoto2002 at netscape.net  Wed Apr  2 16:28:04 2003
From: bbfoto2002 at netscape.net (BBFoto)
Date: Wed, 02 Apr 2003 21:28:04 GMT
Subject: need help with game
References: <3e88b3d3.30570106@news-server> <mailman.1049232185.25252.python-list@python.org> <3e8a586c.56142743@news-server> <mailman.1049277905.31117.python-list@python.org> <3e8afeb9.1125578@news-server> <3E8B0060.D9C449A3@engcorp.com> <3e8b1b0f.8380222@news-server> <slrnb8lq1h.d7q.rmunn@localhost.localdomain>
Message-ID: <3e8b5589.10546880@news-server>

I copied your line and pasted it.  

File "<stdin>", line 1
    C:\Python22\Python.exe "c:/program
files/pysol/pysol-4.80/data/PySol_22.pyc"

     ^
SyntaxError: invalid syntax



Double clicking on pysol_22.pyc  or any of the other .pyc files does
nothing.

Thanks,

BB
>>>


On Wed, 02 Apr 2003 19:38:18 GMT, Robin Munn <rmunn at pobox.com> wrote:

>BBFoto <bbfoto2002 at netscape.net> wrote:
>> I apologize for top posting.  I clicked on "Followup Usenet Message".
>> My mistake was unintentional.  I will try to find out more about it as
>> was suggested.  And as to why I tried to execute a .pyc file:  If I
>> knew what I was doing, then I would do it right.  I do appreciate all
>> attempts to help.
>> 
>> I did a search on the PySol directory.  Nowhere is there a file called
>> "PySol.py".  There is a series of files in sequence named
>> "PySol_15.pyc"  "PySol_16.pyc"  "PySol_20..pyc"  "PySol_21.pyc" and
>> "PySol_22.pyc".
>> 
>> 
>> I tried to execute this line,
>>  
>> execfile('c:/program files/pysol/pysol-4.80/data/PySol_22.pyc') 
>> 
>> The path is correct.  After it failed I copied this error message from
>> the Dos screen:
>> 
>>   File "c:/program files/pysol/pysol-4.80/data/PySol_22.pyc", line 1
>>     -_
>>      ^
>> SyntaxError: invalid syntax
>
>OK, try this:
>
>C:\Python22\Python.exe "c:/program files/pysol/pysol-4.80/data/PySol_22.pyc"
>
>(Substitute your path to Python as appropriate for the Python.exe).
>
>If you get an error message about "global name 'win32_gethomedir' is not
>defined", that's because you somehow got the Linux version of PySol
>instead of the Windows version. If that's the case, go here:
>
>    http://avitous.net/software/pysol-windows/index.shtml
>
>and download the PySol 4.81 for Windows installer. I just tried that one
>five minutes ago on my Windows box and it worked just fine; that should
>get you up and running.
>
>-- 
>Robin Munn <rmunn at pobox.com>
>http://www.rmunn.com/
>PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838




From duncan at NOSPAMrcp.co.uk  Tue Apr  1 03:43:58 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Tue, 1 Apr 2003 08:43:58 +0000 (UTC)
Subject: Use Pyrex instead (Re: newbie at SWIG, help needed with typemaps)
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <3E87E64C.F7D2C11E@easystreet.com>
Message-ID: <Xns935061D008C57duncanrcpcouk@127.0.0.1>

achrist at easystreet.com wrote in news:3E87E64C.F7D2C11E at easystreet.com:

> "Greg Ewing (using news.cis.dfn.de)" wrote:
>> 
>> Instead of SWIG, you might prefer to use Pyrex to wrap your C code:
>> 
>>     http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/
>> 
>> This is exactly the sort of problem I designed Pyrex to solve.
> 
> Looks almost too nice to be true.  Any reliability or performance
> issues?
> 
> Any reason to think this won't work with Windows and MSVC?

I'm giving a talk at the UK Python conference about integrating Python with 
C/C++, so I've been playing about with Pyrex a bit recently. I can highly 
recommend it.

Remember it is still a work in progress. This means that if you really push 
the envelope, you may find some areas where it becomes less friendly[1], 
probably because nobody else has been there before you. However, most of 
the time it will really amaze you. When you get an extension written using 
the C Api or SWIG wrong, you spend an age in the debugger. When you get 
your Pyrex extension wrong, you get a Python exception and stack backtrace.

If you use Distutils to do the compilation then you won't even have to 
worry about what compiler options to use: your extension will compile on 
anything that has a compiler. If you want to write your own makefiles you 
may have to think a bit harder, but I can post a simple Pyrex makefile for 
msvc if anyone wants it (actually I'll make a point of putting my talk 
slides and associated examples on the web after the conference).

My only other complaint would be that the documentation is perhaps a little 
bit sparse. Probably it just needs some more examples.

[1] An example where it becomes less friendly. The documentation says that 
to do class methods you have to write the function outside the class:

    	def Spam_method(cls):
    	    	pass

     class Spam:
    	    	method = classmethod(Spam)

I, of course, misread this and wrote:

     cdef class Spam:
    	    	method = classmethod(Spam)

This gives an error from the C compiler: 'method' undeclared identifier.
Fair enough that it doesn't work (the documentation sort of says it won't), 
but not so good that it generated bad C code. (For those who don't know 
Pyrex, 'class Spam' generates a Python class 'Spam', 'cdef class Spam' 
generates a new built-in type 'Spam'.)

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From cpl.19.ghum at spamgourmet.com  Tue Apr  8 15:54:03 2003
From: cpl.19.ghum at spamgourmet.com (Harald Massa)
Date: Tue, 8 Apr 2003 21:54:03 +0200
Subject: cx_Oracle and Number(11)
References: <3e918724@news.mt.net.mk> <Xns9356EDF45F8E9cpl19ghumspamgourmet@62.153.159.134> <3e929143@news.mt.net.mk>
Message-ID: <Xns9357DEDCAF717cpl19ghumspamgourmet@62.153.159.134>

>> Isn't that by definition? Number turns to float, INT to INT?
> hmm, what definition?

there is allways a mapping sql - python And I'm quite sure that the mapping 
from number is allways to float.

So it's "by design" rather than "by definition", sorry.



From jlh at cox.net  Wed Apr 16 00:32:26 2003
From: jlh at cox.net (Jeff Hinrichs)
Date: Wed, 16 Apr 2003 04:32:26 GMT
Subject: [Q] can PySerial work with a USB to serial adapter ?
References: <xZYma.26921$c42.17977@fe01.atl2.webusenet.com>
Message-ID: <u15na.146641$S14.139802@news1.central.cox.net>

Are you sure it's locate on com1? I've used such a device on win2k and the
drivers for the adapter set it up as com6 by default. I used pyserial to
access it just fine.
-jeff
"Rikko" <riknessa at hotmail.com> wrote in message
news:xZYma.26921$c42.17977 at fe01.atl2.webusenet.com...
> Hello python minded,
> I'm trying to use PySerial with Python 2.2.2 on an XP system
> with a USB to RS232 adapter.
> Upon opening the port (making an instance of the Serial class)
> the system return the error message "unable to open file",
> thereby referring to the COM1 port used.
>
> Is it ME who's goofing thing up or what?
> With kind regards, Rik.
>
>
>





From db3l at fitlinxx.com  Wed Apr  9 17:54:51 2003
From: db3l at fitlinxx.com (David Bolen)
Date: 09 Apr 2003 17:54:51 -0400
Subject: wxPython - Can you explain why I crash
References: <33803989.0304090652.174ac9de@posting.google.com> <3E946842.5D897F5@engcorp.com>
Message-ID: <uadezibf8.fsf@fitlinxx.com>

Peter Hansen <peter at engcorp.com> writes:

> Miki Tebeka wrote:
> > 
> > I've written this little client. It used to work fine but now no
> > matter which button I press the application crash.
> > I'm running Python2.2 and wxPython 2.4.0.7
> > Any ideas?
> 
> [snip 480 lines of code]
> 
> I have an idea: spend a little of your own time cutting down the
> program to a *much* smaller example which still crashes.  Then,
> and only then, post it to this newsgroup and ask for others' help.

Agreed - I know for my part that besides the length, the fact that it
would need various network operations and make remote connections that
I wasn't sure would work in my environment anyway was enough to keep
me from answering.

It would also help if the OP could include a better definition than
"crash" - are we talking about a Python traceback, or a Win32
application crash?  If the former, then definitely include the actual
traceback information.  If the latter, then more information about the
platform (e.g., Windows OS version at a miminum) would be helpful.

-- David



From jjl at pobox.com  Tue Apr 22 16:36:17 2003
From: jjl at pobox.com (John J. Lee)
Date: 22 Apr 2003 21:36:17 +0100
Subject: *Newbie Question* --> Is it viable to distribute Python apps over the Internet?
References: <25a672ee.0304210914.1cc2049c@posting.google.com> <mailman.1050946774.25112.python-list@python.org> <uPYoa.240$Fb.4879581@newssvr15.news.prodigy.com> <mailman.1050978337.8048.python-list@python.org>
Message-ID: <87wuhm2rsu.fsf@pobox.com>

Jp Calderone <exarkun at intarweb.us> writes:

> On Mon, Apr 21, 2003 at 08:49:30PM +0000, Dennis Reinhardt wrote:
> >
> > On Windows, which is where Visual Basic runs after all,  a complete Python
> > hello world is 459 KB.   3-4 MB is *not* required.  See the online builder
> > for constructing Python hello-world class executables at
> > http://www.dair.com/autorun/arpyexe.cgi
[...]
> > The python DLL is bundled as are all required PY files in both of the
> > executables above.
> 
>   Is the app that generates these available (commercially or for free)?

Probably py2exe or McMillan Installer, I imagine.  Both know about
dependencies, so you don't end up with the whole Python distribution
in your .exe.


John



From rmunn at pobox.com  Sat Apr  5 12:59:42 2003
From: rmunn at pobox.com (Robin Munn)
Date: Sat, 05 Apr 2003 17:59:42 GMT
Subject: Invalid format character in string
References: <s5vr8vs4hsfkatous02cvfkbv1quvgd9ev@4ax.com> <mailman.1049493668.26508.python-list@python.org> <ns0s8v4nv240k62mvvenit8ble307r0lda@4ax.com> <mailman.1049496189.612.python-list@python.org>
Message-ID: <slrnb8thcf.ldu.rmunn@localhost.localdomain>

Adam Hupp <hupp at cs.wisc.edu> wrote:
> On Sat, Apr 05, 2003 at 12:17:52AM +0200, Nikolai Kirsebom wrote:
>> Hi,
>> The problems arise where the HTML editor used (to produce the file in
>> the first place) has put various expressions of %; and %>.  These are
>> HTML codings the creator of the HTML file does not really know about.
>> As an example, I've just told them (the users creating the HTML file)
>> that they should write the text "%(name)s" where they want the name
>> generated and "%(age)s" where they want the age generated.  Then my
>> script is run for a set of person objects producing a set of HTML
>> files with the correct names and corresponding ages. 
> 
> You could apply this regexp to the string before you do the insertion:
> 
>>> re.sub('%(?!\(\w+\)s)','%%','%foo %(bar)s')
><< '%%foo %(bar)s'
> 
> The '(?!...)' is a negative lookahead assertion.

I'd suggest stripping the "s" off the end of that regexp, myself: you
never know if someone might have decided they wanted to print a decimal
number or whatever with "%(age)d". Yes, he told the users to use s, but
one of them might know a little C programming and decide "Oh, this looks
like printf() format strings, in which case that 's' should be a 'd'."

A little constructive paranoia often comes in handy, I find...

-- 
Robin Munn <rmunn at pobox.com>
http://www.rmunn.com/
PGP key ID: 0x6AFB6838    50FF 2478 CFFB 081A 8338  54F7 845D ACFD 6AFB 6838



From ckea25d02 at sneakemail.com  Wed Apr 16 21:39:04 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 17 Apr 2003 13:39:04 +1200
Subject: Why functional Python matters
In-Reply-To: <ccc7084.0304160450.680bf7ef@posting.google.com>
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com>
Message-ID: <b7kvuh$20aem$1@ID-169208.news.dfncis.de>

Tj wrote:
> Lambda could also stand renaming;
> Pythonistas probably think it's an arrogant name.  (Or exotic?  I
> can't tell.)

Actually, we think it's kind of cute.

   Mary had a little lambda,
   Its syntax white as snow,
   And every program Mary wrote,
   She wrote in Lisp, you know.

Although nowadays she's probably using Python instead.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ianb at colorstudy.com  Tue Apr 22 04:54:41 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 22 Apr 2003 03:54:41 -0500
Subject: Should I prefer an external database
In-Reply-To: <u3ckbymri.fsf@mail.afm.dtu.dk>
References: <u3ckbymri.fsf@mail.afm.dtu.dk>
Message-ID: <1051001680.701.124.camel@lothlorien>

On Tue, 2003-04-22 at 03:11, Brian Elmegaard wrote:
> I am working on a script for organizing a program for a conference. I
> started out using a connection to mySQL, but then I thought that it
> would be easier to keep it within python. So I implemented a class:
> 
> class Paper:
>         def __init__(title,author,session):
>                 self.title=title
>                 self.author=author
>                 self.session=session
> I then have a list to which each new paper instance is appended. I can
> sort by any of the paper instances variables by using the
> decorate-sort-undecorate method.
> 
> I now have it working, but I probably could do better, so:
> 
> Is this a more pythonic way to implement this?

More Pythonic?  Eh... I would say that's fine, from a purely "What's
Pythonic" view.  Is there a *better* way?  Well, there's *always* a
better way... ;)

> To do a select I think I need to use for loops specifying the names of
> the variables, i.e:
> 
> for paper in paperlist:
>         if paper.session=='foo':
>                 print paper.title
> 
> Probably this might be done more elegantly, but how do I know which
> variables (database columns) the instance has?

You could try a number of techniques for making searches easier to use. 
Like:

def search(**kw):
    result = []
    for paper in paperlist:
        matched = True
        for key, value in kw.items():
            if getattr(paper, key) != value:
                matched = False
                break
        if matched: result.append(paper)
    return result

Then you can do search(title="foo", author="bar"), for instance.

You can do lots just with list comprehension, like:

result = [p for p in paperlist if p.session == 'foo']
print '\n'.join(result)

For particular searches you can do all sorts of indexing, mostly a
matter of creating dictionaries of commonly searched-for keys.

> Would I be better of with a list of list than with a list of class
> instances? 

No, I think instances are better.

> What about a class of classes? Could such be made and inherit from the
> list class to support index and sort?

You could, say, store all the instances in a class variable, then use
class methods to search them.  Then you could index the objects as you
find it necessary, hiding the indexing in the functions.



Now, that all said, you might want an external database with an
object-relational mapper.  Using SQLObject (plug: http://sqlobject.org),
you'd do it like:

from SQLObject import *

__connection__ = MySQLConnection(user=...)
# or use Postgres, SQLite, or a DBM backend

class Paper(SQLObject):
    _columns = [StringCol('title'), 
                StringCol('author'), 
                KeyCol('session', foreignKey='Session')]

(this will look a little cleaner in the next version)

Then use it like:

for paper in Paper.select(AND(Paper.q.title == 'foo',
                              Paper.q.author.startswith('Bob'))):
    print paper

or:

for paper in Paper.selectBy(title='foo', author='bar'):...

Plus you get persistence.  A database like SQLite adds code
dependencies, but no environment dependencies, so maybe that's what you
are looking for (SQLite just stores its database in a single file and
isn't client/server).

-- 
Ian Bicking   ianb at colorstudy.com  http://colorstudy.com
PGP:  gpg --keyserver pgp.mit.edu --recv-keys 0x9B9E28B7
4869 N. Talman Ave., Chicago, IL 60625      773-275-7241





From mwh at python.net  Fri Apr 25 06:48:52 2003
From: mwh at python.net (Michael Hudson)
Date: Fri, 25 Apr 2003 10:48:52 GMT
Subject: Dict handling of floats
References: <98imn-ust.ln1@grendel.myth> <slrnbahk9n.ero.grante@localhost.localdomain>
Message-ID: <7h3fzo6luno.fsf@pc150.maths.bris.ac.uk>

Grant Edwards <grante at visi.com> writes:

> In article <98imn-ust.ln1 at grendel.myth>, Jim Richardson wrote:
> 
> > I have come across a weirdness. I have a number of floats, (lat long
> > pairs) which I stuff into a dict, but when I take them out, my nice %2f4
> > number, (say, 47.5894) is something long and *almost* right, like
> > 47.589333333333333339. What gives?
> 
> That's floats.  The best you can hope for is "almost right".

Well, in some circumstances you can hope for a bit more than that.
But unless you're very careful, this is probably a good rule of thumb.

Cheers,
M.

-- 
  > It might get my attention if you'd spin around in your chair,
  > spoke in tongues, and puked jets of green goblin goo.
  I can arrange for this.  ;-)            -- Barry Warsaw & Fred Drake



From ckea25d02 at sneakemail.com  Wed Apr 16 22:10:12 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 17 Apr 2003 14:10:12 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m24r4yh1l8.fsf@orbtech.com>
References: <mailman.1050394215.20671.python-list@python.org> <b7i26g$17t5m$1@ID-169208.news.dfncis.de> <m24r4yh1l8.fsf@orbtech.com>
Message-ID: <b7l1ov$22o4q$1@ID-169208.news.dfncis.de>

Patrick K. O'Brien wrote:
> I suppose one could bind attributes (current frame, current buffer,
> point) to a module and then import that module everywhere that needed
> access to these attributes

That's no better, because they're still global -- there's
still only *one* of each of them, whereas there really
needs to be different ones in different contexts.

A function which needs to operate on a frame should be
passed a frame. Some buffer will be displayed in that
frame: frame.buffer. That buffer will have a point:
frame.buffer.point. Etcetera.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From mike at bindkey.com  Wed Apr 23 13:53:29 2003
From: mike at bindkey.com (Mike Rovner)
Date: Wed, 23 Apr 2003 10:53:29 -0700
Subject: SGMLParser questions
References: <igmbavg5ughls6ksiden2rgnael1s0ub8q@4ax.com> <mailman.1051061136.16221.python-list@python.org> <3EA60335.E3A047C6@alcyone.com>
Message-ID: <b86kks$o9f$1@main.gmane.org>

"Erik Max Francis" <max at alcyone.com> wrote in message
news:3EA60335.E3A047C6 at alcyone.com...
> Mike Rovner wrote:
>
> > dict(attributes).get['src']
>
> <nitpick value="I think you meant
>
> dict(attributes).get('src')
>

Yeah, sure. Thank you for pointing that out.

Mike

PS. It's interesting that even with habit to check my every post BEFORE
sending (including my first reply)
I sometimes still miss some. ;) Hope that adding something to the nature of
human errors that some P...
languages take into account but others don't. I was bitten by them and
switched to Python as way more
error prone.







From exarkun at intarweb.us  Wed Apr  2 16:55:03 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Wed, 2 Apr 2003 16:55:03 -0500
Subject: How to compile a python script to *.pyo
In-Reply-To: <3045.143.166.82.240.1049319577.squirrel@iamlearning.com>
References: <mailman.1049309105.22947.python-list@python.org> <b6fj57$626$1@news-reader11.wanadoo.fr> <3E8B53BA.DBB0C84C@engcorp.com> <3045.143.166.82.240.1049319577.squirrel@iamlearning.com>
Message-ID: <20030402215503.GB15307@meson.dyndns.org>

On Wed, Apr 02, 2003 at 03:39:37PM -0600, danny at iamlearning.com wrote:
> [snip]
> 
> 
> What are other techniques I can use?
> I want my users run *.pyo or *.pyc files instead of *.py.
> 

  Why?

  Jp

-- 
Lowery's Law:
        If it jams -- force it.  If it breaks, it needed replacing anyway.
-- 
 up 13 days, 18:00, 7 users, load average: 0.10, 0.09, 0.08




From BPettersen at NAREX.com  Wed Apr 16 17:35:53 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Wed, 16 Apr 2003 15:35:53 -0600
Subject: Big-O notation
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFA6@admin56.narex.com>

> From: sik0fewl [mailto:xxdigitalhellxx at hotmail.com] 
> 
> Peter van Kampen <news at datatailors.xs4all.nl> wrote in message 
> news:<slrnb9q90a.mhg.news at datatailors.com>...
> >I think I almost 'get it'. Except who or what decides what Big-O a
> >certain algorithm has. Is that by 'hunch' or by running benchmarks or
> >are there still other ways to recognize 'Big-O's (except for the
> >formal computations that I understand are not very common)? 
> 
> I'll try to give a simple explanation. Big-O is the upper 
> bound of the time it will take to complete the function.
[...]

google(algorithm complexity) == lecture notes
(http://www.geog.buffalo.edu/classes/geo554/week4.pdf)

[also "time complexity", "space complexity"].

 O() ::= The big-O notation gives an indication of the growth 
         rate of a function. It identifies the dominant term 
         of the function.

And, for the practically minded...

Rules of Thumb
- A loop that processes each data item in turn is O(n).
- A loop within a loop
    Do n
      Do m
  Is O(m*n)
- If each data set is halved each time through a loop 
  then there will be O(log2n) iterations. (log2n is 
  the number of times you repeatedly half n to reach 1).
- An algorithm that processes every subset of n data 
  items has complexity O(2**n).
- An algorithm that processes every permutation of n 
  data items has complexity O(n!).
- If the number of operations is the same whatever 
  the number of data items then the complexity function 
  does not depend on n. We say the operation has complexity 
  O(1).

-- bjorn




From amk at amk.ca  Wed Apr 16 07:51:23 2003
From: amk at amk.ca (A.M. Kuchling)
Date: Wed, 16 Apr 2003 06:51:23 -0500
Subject: Pythonic way of web-programming
References: <cQ_la.5597$LB6.158861@news1.tin.it> <mailman.1050210449.24337.python-list@python.org> <23891c90.0304140221.20e9e7fb@posting.google.com> <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com> <B8Pma.20554$T34.656139@news2.tin.it> <mailman.1050471334.22000.python-list@python.org>
Message-ID: <0gqdnfw_XuYm3gCjXTWcow@speakeasy.net>

On 16 Apr 2003 00:35:05 -0500, 
	Ian Bicking <ianb at colorstudy.com> wrote:
> I don't think FastCGI has much relation to CGI.  I don't entirely
> understand FastCGI, but from what I can tell it just sucks -- it's
> easier to do the same thing with a simpler socket connection ...

The FastCGI protocol is really complicated because it supports multiplexing 
several different transactions over one socket.  Data is broken up into
multiple packets, each with a unique transaction ID, and the process on the
other end can then reassemble them and hand them out to different threads.
However, I'm not sure the Apache mod_fastcgi supports this and few client
modules seem to, so all of that complexity buys nothing in most cases.

--amk                                                             (www.amk.ca)
FLUTE: Nay, faith, let me not play a woman; I have a beard coming.
      -- _A Midsummer Night's Dream_, I, ii



From oktaysafak at ixir.com  Thu Apr  3 15:22:57 2003
From: oktaysafak at ixir.com (oktaysafak at ixir.com)
Date: Thu, 3 Apr 2003 22:22:57 +0200
Subject: unicode bug in turkish characters?
Message-ID: <20030403202257.PYOK19162.nairobi@[212.252.122.82]>

Martin, thanks for your attention.

Here is the problematic output on my machine (turkish windows98, IDLE):

Python 2.3a2 (#39, Feb 19 2003, 17:58:58) [MSC v.1200 32 bit (Intel)] on win32
Type "copyright", "credits" or "license" for more information.
IDLE 0.8 -- press F1 for help

>>> import sys
>>> sys.getdefaultencoding()
'iso-8859-9'
>>> import locale
>>> locale.getdefaultlocale()
('tr_TR', 'cp1254')
>>> u"i".upper()
u'I'
>>> "i".upper()
'I'
>>> "i".encode("windows-1254").upper()
'I'
>>> "i".encode("iso-8859-9").upper()
'I'
>>> u"i".encode("iso-8859-9").upper()
'I'
>>> "i".upper().decode("iso-8859-9")
u'I'
>>> import unicodedata
>>> unicodedata.name(_)
'LATIN CAPITAL LETTER I'
>>> hex(ord("i".upper()))
'0x49'    # necessary to see the hex value in my IDLE

You see, the values we get are different. However, if I do

>>> hex(ord("*"))  # * stands for capial I with dot
'0xdd'



I get the correct value. To see whether things would improve if the locale sys.defaultencoding are both windows1254, since I'm on a windows machine, here is another run:



Python 2.3a2 (#39, Feb 19 2003, 17:58:58) [MSC v.1200 32 bit (Intel)] on win32
Type "copyright", "credits" or "license" for more information.
IDLE 0.8 -- press F1 for help
>>> import sys, unicodedata, locale
>>> sys.getdefaultencoding()
'windows-1254'
>>> locale.getdefaultlocale()
('tr_TR', 'cp1254')
>>> locale.setlocale(locale.LC_ALL, "tr_tr")
Traceback (most recent call last):
  File "<pyshell#3>", line 1, in ?
    locale.setlocale(locale.LC_ALL, "tr_tr")
  File "C:\PYTHON23\lib\locale.py", line 381, in setlocale
    return _setlocale(category, locale)
Error: locale setting not supported
>>> locale.setlocale(locale.LC_ALL, "turkish")
'Turkish_Turkey.1254'
>>> locale.getdefaultlocale()
('tr_TR', 'cp1254')
>>> "i".upper()
'I'
>>> u"i".upper()
u'I'
>>> locale._print_locale()
Locale defaults as determined by getdefaultlocale():
------------------------------------------------------------------------
Language:  tr_TR
Encoding:  cp1254

Locale settings on startup:
------------------------------------------------------------------------
LC_NUMERIC ...
   Language:  Turkish_Turkey
   Encoding:  1254

LC_MONETARY ...
   Language:  Turkish_Turkey
   Encoding:  1254

LC_TIME ...
   Language:  Turkish_Turkey
   Encoding:  1254

LC_COLLATE ...
   Language:  Turkish_Turkey
   Encoding:  1254

LC_CTYPE ...
   Language:  Turkish_Turkey
   Encoding:  1254


Locale settings after calling resetlocale():
------------------------------------------------------------------------
Traceback (most recent call last):
  File "<pyshell#8>", line 1, in ?
    locale._print_locale()
  File "C:\PYTHON23\lib\locale.py", line 737, in _print_locale
    resetlocale()
  File "C:\PYTHON23\lib\locale.py", line 391, in resetlocale
    _setlocale(category, _build_localename(getdefaultlocale()))
Error: locale setting not supported
>>> hex(ord("i".upper()))
'0x49'


No change. Also, when I try to set the locale to "tr_TR" which you do painlessly, I get an error instead. The error I get with _print_locale is also interesting. 

I hope these outputs make it clear for you to see what's going on.

Thanks again.

Regards,
Oktay





From ianb at colorstudy.com  Wed Apr 16 01:35:05 2003
From: ianb at colorstudy.com (Ian Bicking)
Date: 16 Apr 2003 00:35:05 -0500
Subject: Pythonic way of web-programming
In-Reply-To: <B8Pma.20554$T34.656139@news2.tin.it>
References: <cQ_la.5597$LB6.158861@news1.tin.it>
	 <mailman.1050210449.24337.python-list@python.org>
	 <23891c90.0304140221.20e9e7fb@posting.google.com>
	 <d5mm9v8em9j41kus2tvj2raipd7dagv18k@4ax.com>
	 <B8Pma.20554$T34.656139@news2.tin.it>
Message-ID: <1050471304.30719.131.camel@lothlorien>

On Tue, 2003-04-15 at 03:11, Alex Martelli wrote:
> However -- look at the XML situation for contrast.  There IS just such
> a standard (and parts of it ARE in the standard libraries, while others,
> PyXML stuff, integrate as seamlessly as they can with those via some
> import tricks).  AND YET -- you guys have developed and maintain PyRXP,
> which uses substantially different interfaces -- *AND WELL IT DOES*,
> because thanks to the different interfaces (trees of bare tuples, or
> lists as the case may be, rather than intricate objects as in DOM) it
> can save a LOT of memory, and generally zip along much, much faster.
> 
> Aren't there analogies here...?  

There may be compelling differences in some areas.  For instance, a
template that is DOM-like (i.e., externally controlled) vs. a template
that is self-evaluating, like Cheetah.  But among self-evaluating
templates, the interface seems like it should be possible to unify.

There are deep differences and shallow ones.  It would be really nice to
get rid of the shallow ones.

[...]
> > Finally, there's a transport mechanism.  Here one could use
> > CGI, Twisted, Zope, BaseHTTPServer or whatever else.
> 
> Not sure I'd qualify Zope as a "transport" mechanism!-)  Rather
> the choice here is:
> 
>   -- a custom server program coded in Python (BaseHTTPServer and
>      the like, Medusa, Twisted)
> 
>   -- CGI, the only standard ensuring operability with different
>      webservers (maybe fast-CGI might live here too...?)

I don't think FastCGI has much relation to CGI.  I don't entirely
understand FastCGI, but from what I can tell it just sucks -- it's
easier to do the same thing with a simpler socket connection to a
persistent process (as done with SCGI, WebKit adapters, mod_skunkweb,
and no doubt others).  I think FastCGI manages starting that persistent
process on its own, which seems like a nice feature.

>   -- webserver-specific ways to cooperate with existing webservers:
>      LRWP for Xitami, mod_python or mod_webkit or PyApache for
>      Apache, ASP or lower-level interfaces to MS servers, ...

mod_python is another concept entirely (mod_webkit is just one
implementation of a simple protocol that's web-server-neutral)...
mod_python really is something separate, though maybe it could be made
to look like any other forking request.

> Can we find enough commonality among such disparate needs to
> extract out some common interfaces?  In the DB-API case (and in
> the XML one) there were some existing external standards (not
> necessarily well-observed, vide the SQL standard, but still maybe
> better than nothing) as a reference point.  Besides CGI, and
> HTTP itself, what would we build upon here?  Perhaps the Request-
> Response model you mention would suffice for reasonably high
> level interactions, but much of what one can do e.g. with mod_python
> ISN'T "reasonably high level" (or else one would just use
> PyApache for CGI-substition-level tasks...!).

There is a common thread here: HTTP.  It's not that complicated, and the
interface you'd create for an HTTP transaction seems pretty self-evident
-- request and response objects don't have a lot of novelty to them.

There may be funny Apache-specific things you could do with mod_python,
but I think the benefit of those features is easily outweighed by the
potential benefits of some more unification of web programming for
Python... and besides mod_python, most frameworks share a pretty similar
foundation, even though each one phrases it slightly differently.

  Ian






From mads at troest.NEVERMORE.dk  Sat Apr 12 08:41:46 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Sat, 12 Apr 2003 14:41:46 +0200
Subject: Using __repr__ or __str__ for own printable class?
Message-ID: <uw1i4nqpl33h.1l3wpteil4jb0.dlg@40tude.net>

Hi Pythoners;

If I create a class which I want to be able to print like this ...
	tst = myclass()
	exp = r"test: %(intepol)s";
	print exp % { 'intepol':tst }

... it seems I have two choices, either of which appear to work. I can 
overload the __str__ or the __repr__ method.

Which is the most correct to use; or should I even overload both? The 
documentation says __str__ is for the, quote, informal representation; 
whereas __repr__ is for the, quote, official representation. Furthermore, 
__reptr__ should, quote, look like an expresseion. I'm not quite sure what 
is meant by these distinctions. Is it that __repr__ is used by pickle and 
the like, and - thus - should represent the entire object state? Given that 
I simply want to determine what should happen when one tries to print an 
instance, do I assume correctly when I gather that __str__ would be the 
more appropriate of the two - or is there some other way I should achieve 
this?

Thanks in advance,
   /\/\\ads Orbesen Troest



From max at alcyone.com  Thu Apr 24 17:40:29 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 24 Apr 2003 14:40:29 -0700
Subject: OT: losing memories (was Re: How many of you are Extreme 
 Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com> <oafgavg8e0pf512712g4aenldqktvmb0dr@4ax.com>
Message-ID: <3EA859CD.8F8AFEFA@alcyone.com>

Stephen Horne wrote:

> Experts are only expert in particular fields. Just because Einstein
> was a genius in the field of physics, that did not qualify him to
> state that only 10% of the brain is unused. Assuming he claimed it at
> all - I never checked, but it does get attributed to him now and again
> - it was probably never intended as a statement of strict biological
> fact.

It's an urban legend:

	http://www.urbanlegends.com/science/10_percent_of_brain.html

In fact, I've never even heard it attributed to Einstein.

In some attempt to contribute some positive value to this already
off-topic thread, I discovered something which I thought was insightful
when I was watching a Nova (or some other such high-quality PBS program)
concerning patients with severe brain injury.  Many -- and evidently
this is not at all uncommon -- those who suffer severe brain injury that
leads to profound impairment of physical abilities cannot remember
things being any different.  So not only has the injury caused the
damage, but it also tends to remove the ability of the victim to
remember what things were like before the injury.

One particular example was of a young woman who suffered severe injury
(I cannot remember whether it was due to physical trauma, coma, or
hypoxia ... in any case).  She was unable to walk and was now confined
to a wheelchair, and as she was being given some basic assessment tests
(which she was doing very poorly at, as you might expect), the doctor
asked her if she remembered things being any different, like being able
to remember things or walk.  She thought about it, and said, "Nope."

I'm not really sure whether I find that comforting or not.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Insight -- the titillating knack for hurting!
\__/ Colette
    Blackgirl International / http://www.blackgirl.org/
 The Internet resource for black women.



From maxm at mxm.dk  Sun Apr  6 10:43:44 2003
From: maxm at mxm.dk (Max M)
Date: Sun, 06 Apr 2003 16:43:44 +0200
Subject: Python & Music: Performance question
In-Reply-To: <1759581e.0304060218.164d4cd4@posting.google.com>
References: <1759581e.0304060218.164d4cd4@posting.google.com>
Message-ID: <V%Wja.5267$y3.728616@news010.worldonline.dk>

David Asorey ?lvarez wrote:
> Hello, group.
> 
> I'm developing a small music program, and I have some doubts.
> 
>    I want define a class, Note, which represents a note. It has two
> attributes, pitch and duration (quarter note, half note, etc). There
> will be a lot of instances of this object, maybe hundreds of notes,
> and the program needs access to all of this objects and read its
> attributes.
>    Which implementation will be more efficient?


It is premature optimisation. First you should just get it to work, and 
then you should optimize if there are any problems.

No matter how many notes you have, it will make no problems for you on a 
reasonably moderne computer.


hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From skip at pobox.com  Tue Apr  1 20:12:03 2003
From: skip at pobox.com (Skip Montanaro)
Date: Tue, 1 Apr 2003 19:12:03 -0600
Subject: The Whitespace Language
In-Reply-To: <eGWdnSuSpbdBqBejXTWc3w@comcast.com>
References: <283adf56.0304011143.4f4a6934@posting.google.com>
        <3E8A12CD.6194A9A8@alcyone.com>
        <eGWdnSuSpbdBqBejXTWc3w@comcast.com>
Message-ID: <16010.14563.997079.236784@montanaro.dyndns.org>

    Terry> Do you mean this WhiteSpace (XML-centered Java extension)?
    Terry> http://www.bluering.nl/whitespace/faq.html

No, I think they mean:

    http://developers.slashdot.org/article.pl?sid=03/04/01/0332202&mode=thread&tid=156

Skip




From alloydflanagan at attbi.com  Wed Apr 16 17:14:23 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 16 Apr 2003 14:14:23 -0700
Subject: do python's nifty indentation rules spell the death of one-liners?
References: <87he90b8rp.fsf@jidanni.org> <a8b7f07a.0304160946.73760d3a@posting.google.com> <3e9d9bb9$0$151$a1866201@newsreader.visi.com>
Message-ID: <a8b7f07a.0304161314.361a80fa@posting.google.com>

grante at visi.com (Grant Edwards) wrote in message news:<3e9d9bb9$0$151$a1866201 at newsreader.visi.com>...
> In article <a8b7f07a.0304160946.73760d3a at posting.google.com>, A. Lloyd Flanagan wrote:
> > Dan Jacobson <jidanni at dman.ddts.net> wrote in message news:<87he90b8rp.fsf at jidanni.org>...
> > 
> > python << DONE
> > print 2
> > for i in (1,4):
> >      print i
> > DONE
> 
> But how do you do it in a Makefile?

Err.  Put it in a script file, and call the script file from make?  I
tried sticking it in a Makefile, can't get it to work.  So far.



From andrew-pythonlist at puzzling.org  Tue Apr 22 08:18:36 2003
From: andrew-pythonlist at puzzling.org (Andrew Bennetts)
Date: Tue, 22 Apr 2003 22:18:36 +1000
Subject: Twisted 1.0.4
In-Reply-To: <341410000.1051007646@leeloo.intern.geht.de>
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local> <341410000.1051007646@leeloo.intern.geht.de>
Message-ID: <20030422121836.GA8567@frobozz.local>

On Tue, Apr 22, 2003 at 12:34:06PM +0200, Marc Recht wrote:
> >to Twisted and Twisted's users.  Similarly, Twisted has some C extensions
> >which are typically built with gcc, but gcc's licence (GPL) also has no
> >relevance to Twisted.
> IIRC then gcc libs are LGPL'ed, so it doesn't affect Twisted. I wouldn't be 
> so sure if they were GPL'ed.

No, gcc is GPLed.  I'm not sure what you mean by "gcc libs"?  I'm pretty
sure libgcc1 is part of gcc and thus GPLed also.

The point is that a C program or library compiled with gcc is not a derived
work of gcc, anymore than a thesis written in and printed with MS Word is a
derived work of MS Word.  Python isn't GPLed, despite often being compiled
with gcc.

Twisted is not a derived work of gcc, python, Buildbot, emacs, vim, the
linux kernel, or any other software incidentally involved in its
development.

-Andrew.





From hamifrancy at zipmail.com  Mon Apr  7 16:32:07 2003
From: hamifrancy at zipmail.com (Hamifrancy Meneses)
Date: Mon, 07 Apr 2003 16:32:07 -0400
Subject: help modules Python
Message-ID: <web-44136975@zipmail.com>

HI, 

I am a transportation engineer. 
I am a beginer in Python and I'd some troubles with
modules.
I'd like to know if somebody could give me a copy of this
modules, or a thick (WEB site) how I can get them:

        win32ui   
	pywin.mfc 
        object
	dde
	os.path

Thanks a lot,

Hamifrancy B. Meneses
Tansportation Engineer
Fortaleza, Cear?, Brazil
________________________________________________
Don't E-Mail, ZipMail! http://www.zipmail.com/




From mlsj at earthlink.net  Thu Apr 17 22:09:04 2003
From: mlsj at earthlink.net (Josh)
Date: 18 Apr 2003 02:09:04 GMT
Subject: Downloading TV listings with urllib
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4> <mailman.1050602375.2763.python-list@python.org> <E196EHo-0003Gb-00@swan.mail.pas.earthlink.net> <mailman.1050608946.14831.python-list@python.org> <pLHna.782$_N.471@twister.nyroc.rr.com>
Message-ID: <Xns9360D72F594DjslmAAearthlinknet@130.133.1.4>

darrell <dgallion1 at yahoo.com> wrote in
news:pLHna.782$_N.471 at twister.nyroc.rr.com: 


> 
> This worked.
> 
>         print gzip.GzipFile(fileobj=cStringIO.StringIO(tvg().read())
>         ).read() 
> 
> --Darrell
> 

Yes I tried that too and it worked for me too. Thanks anyway.

Josh



From fperez528 at yahoo.com  Tue Apr 15 12:21:50 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Tue, 15 Apr 2003 10:21:50 -0600
Subject: more informative stack traces
References: <e0bf36b5.0304132100.4324c3b8@posting.google.com> <3e9ab723$0$66212$edfadb0f@dread11.news.tele.dk> <200304141403.08895.scott_list@mischko.com> <mailman.1050376043.21564.python-list@python.org>
Message-ID: <b7hbiv$3m4$1@peabody.colorado.edu>

Skip Montanaro wrote:


> Everyone involved should take a look at the cgitb module in the standard
> distribution.  I submitted a patch for cgitb a long time ago which allows
> plain text traceback generation:
> 
>     http://python.org/sf/569574
> 
> Note Ping's comment appended to that patch that the core code belongs in
> the traceback module.
> 
> If you can get this sorted out quickly so that the core code is in the
> traceback module I suspect you can still squeeze it into 2.3.

Well, if you want to take a look at ipython (http://ipython.scipy.org), it
already has a text (with optional coloring where supported) version of
this.  The code was originally written by Nathan Grey, and he marked is as
being lifted off cgitb.py.  I've later modified it quite a bit.  After
becoming used to ipython, I find it nearly impossible to work with regular
tracebacks.

Here is an example comparing the ipython traceback (and on myscreen it is
color highlighted):

In [2]: run error
hi
Ramp time: 0.0
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)

/usr/local/home/fperez/test/error.py
    109
    110     print 'speedup:',Rtime/RNtime
    111
    112
--> 113 if __name__ == '__main__': main()
        global __name__ = '__main__', main = <function main>

/usr/local/home/fperez/test/error.py in main()
    104     array_num = zeros(size,'d')
    105     for i in xrange(reps):
--> 106         RampNum(array_num, size, 0.0, 1.0)
        global RampNum = <function RampNum>, array_num = array([ 0.,  0., 
0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.]), size = 10
    107     RNtime = time.clock()-t0
    108     print 'RampNum time:', RNtime

/usr/local/home/fperez/test/error.py in RampNum(result=array([ 0.,  0.,  0., 
0.,  0.,  0.,  0.,  0.,  0.,  0.]), size=10, start=0.0, end=1.0)
     88     tmp = zeros(1e2)
     89     step = (end-start)/(size-1-tmp)
---> 90     result[:] = arange(size)*step + start
        result = array([ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.]),
global arange = <built-in function arange>, size = 10, step = array([
0.11111111,  0.11111111,  0.11111111,  0...11111111,  0.11111111, 
0.11111111,  0.11111111]), start = 0.0
     91
     92 def main():

ValueError: frames are not aligned
WARNING: Failure executing file: <error.py>


to the same traceback from plain python:

[~/test]> python error.py
hi
Ramp time: 0.0
Traceback (most recent call last):
  File "error.py", line 113, in ?
    if __name__ == '__main__': main()
  File "error.py", line 106, in main
    RampNum(array_num, size, 0.0, 1.0)
  File "error.py", line 90, in RampNum
    result[:] = arange(size)*step + start
ValueError: frames are not aligned

This code is ready, it works, and it has been in use by ipython's users for
about 1 1/2 years.  Feel free to take a look at it.

Best,

f.



From donnal at donnal.net  Sat Apr 12 22:32:19 2003
From: donnal at donnal.net (Donnal Walter)
Date: Sat, 12 Apr 2003 21:32:19 -0500
Subject: an ugly file-reading pattern
In-Reply-To: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
References: <OJ3ma.462340$sf5.784124@rwcrnsc52.ops.asp.att.net>
Message-ID: <b7ai52$f1k$1@main.gmane.org>

Istvan Albert wrote:
> now there is the readlines() method
> that slurps everything up (no good for me
> since I don't like to build into my systems
> implicit assumptions about filesizes)
> 
> if I choose to process the file line by line
> the "Learning Python" book advises me to use the
> following atrocity:
> 
> while 1:
>     line = file.readline()
>     if not line: break
> 
> Maybe I'm picky here but having to start an infinite loop
> then needing a conditional to  break out of it is
> anything but an elegant pattern.
> 
> This has left me wondering about python,

In Python 2.2 or later, I believe the following will work:

for line in file:
     print line


Donnal Walter
Arkansas Children's Hospital






From max at alcyone.com  Thu Apr  3 18:01:35 2003
From: max at alcyone.com (Erik Max Francis)
Date: Thu, 03 Apr 2003 15:01:35 -0800
Subject: Overloadable Assignment PEP
References: <9383481e.0304012011.78f05548@posting.google.com>
	 <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk>
	 <9383481e.0304030705.524cf0fa@posting.google.com> <mailman.1049398267.20730.python-list@python.org>
Message-ID: <3E8CBD4F.23406A07@alcyone.com>

Lulu of the Lotus-Eaters wrote:

> Then again, I wouldn't mind adding the new operator ":=" which was
> "generic augmentation"... the operator could do -absolutely nothing-
> for
> standard types, but could be available for custom meaning in custom
> types.

That strikes me as a better approach; having a special "augmentation"
operator which has no predefined meaning but can be overridden as
desired, and thereby leaving usual assignment alone.

The only reservation I might have is whether or not := is a good choice,
simply because of the :=/= and =/== language differences out there.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ The decree is the Sultan's, the mountains are ours.
\__/ Dadaloglu
    Official Omega page / http://www.alcyone.com/max/projects/omega/
 The official distribution page for the popular Roguelike, Omega.



From marcus at infa.abo.fi  Mon Apr 14 14:37:19 2003
From: marcus at infa.abo.fi (Marcus Alanen)
Date: 14 Apr 2003 20:37:19 +0200
Subject: Complexity of standard Python data structures (Re: Speed of string += string)
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <slrnb9l0dm.96d.marcus@infa.abo.fi> <Oqwma.11524$LB6.319489@news1.tin.it> <slrnb9lbbq.v4f.marcus@infa.abo.fi> <Qnzma.12335$LB6.335182@news1.tin.it>
Message-ID: <slrnb9lsbe.4pa.marcus@infa.abo.fi>

On Mon, 14 Apr 2003 14:15:12 GMT, Alex Martelli <aleax at aleax.it> wrote:
>Maybe I should clarify that I'm not _attacking_ your statement (nor
>you for making it, Marcus)

Good, it is far too easy to misinterpret each other. Apologies
if I have sounded a bit harsh.

>While performing optimizations may be premature, _understanding_
>what's going on can hardly be

Exactly.

>However, I still don't understand why this would be a problem
>specific to Python rather than any other language -- including
>C++ for those operations that are not part of that subset of C++'s
>standard library for which O()-behavior is specified in the standard.

Actually, I don't think this is specific to Python---I frequently
ask myself about the complexity of the code I'm programming, even in
C/C++. You asked this several times in your answer, so I'll just
answer here once: the memory and speed complexity of common operations
is in my opinion interesting in any language, and in the long
run, an absolute necessity to understand.

Also consider that Python is a relatively new language, at least it is
a language which evolves very fast. Thus, various issues are probably
unfinished, under development and it is uncertain how they act. Common
sense suggests to tread carefully, and learn what the language
actually does. Plus, I'm personally interested in computer languages.

As an example of "unfinished" interfaces: there is no list.clear()
function, even though dict.clear() exists. (Actually I'll send
python-dev@ a mail about this soon, when I get the time to dive into
the python source code).


Don't mistake any of this for python-bashing, it's _great_ that the
language developers are keen on discussing and implementing new things,
and as far as I see, there are a lot of very bright people there.
Perhaps I just was more interested in "how does it do it?" than
"what does it do?" :-)

>What's the O()-behavior of C+'s dynamic_cast<>, for example, and
>on what parameters does that performance behavior depend?  I do

Ah, but here's an interesting part: it probably doesn't matter, you're
probably going to use it only a few times. You do use data structures
and various algorithms much more, and that's why they matter. If I were
to continuously use dynamic_cast<> in lot's of loops, I'd probably check
its behaviour closer.

In a way, it's the same thing as said by higher-level language people:
optimize where it matters. I just draw the line of optimization to
include data structures and algorithms, and leave dynamic_cast<> out.

>not know, and I don't think I *CAN* know -- every implementation is
>free to do its best (or its worst...) to optimize dynamic_cast's
>performance, time-vs-space tradeoffs, whatever.

True, but sometimes there are only so few sane choices. People have
come to expect a certain behaviour from e.g. C/C++ compilers. I
_expect_ good performance from the C++ and the STL algorithms and
their corresponding implementation, since that's what it was designed
for. When seeing e.g. python's list.__getitem__() and list.insert(), I
got confused: either Python is doing something very very odd to get
both methods fast, or then a list is really just a vector and
list.insert() is O(N) --- something which I would've expected a
"warning" about.

But naturally, sometimes expectations bite you, no matter what the
language: http://dforce.sh.cvut.cz/~seli/en/linking2/

>So why did you perceive Python as being different in this respect?
>C doesn't specify the O()-behavior of, e.g., qsort (indeed it's
>often pretty horrible...!) -- C++ doesn't specify performance for

Exactly, but how did you know about qsort? Legacy. Previous
experience. Things that I, quite naturally, lacked regarding Python.

>such things as RTTI, exception handling, or even for that matter
>for the dispatch of virtual methods

And for the record, C++ is by far not clear to me in all respects:
e.g. I've recently become interested in exception handling in C++, and
wonder how fast the actual stack unwinding is, and how much "extra"
code it creates. (Not done any measurements) They are not however the
common path of execution, and so they are less interesting than
e.g. virtual method dispatching.

>But -- DO Python beginners typically start learning the language with
>a *REFERENCE* manual?  Probably not -- yet you explain that the time
>when you were left wondering was just when you STARTED learning.  So,
>perhaps, the document to retouch with a few small mentions of typical
>performance behavior (perhaps in a footnote or two) could be the
>tutorial, rather than either of the reference manuals (language and
>library).  Clearly, the tutorial is being descriptive (of one specific
>implementation it comes with) and not prescriptive (it does not give
>any constraints to what future implementations might choose), which
>may make it more acceptable to give such indications there.
>
>So what do you think -- is it the tutorial that such changes should
>be targeting...?

Um, no, perhaps people new to programming don't care about performance
at all and would be confused.

On the other hand, I've always wished for a document "language X for
people know language Y / programming in general / think they know". We
could skip all that "let's hold hands while I take you though the
mysteries of a dictionary" nonsense, and just concentrate on the issue
at hand: how does our knowledge from language X map to this new
language? What pitfalls do we have in the language at the moment
(e.g. python default arguments are in my humble opinion still
counter-intuitive) ?

>In "Python in a Nutshell", I took a more traditional approach --
>*assuming* that the target reader would want to understand the
>*functionality* of operations first, the *performance
>characteristics* of opeations only quite a bit later.  So, I
>have descriptions of fundamental functionality quite early, in
>chapter 4, with almost no mention of performance issues (except
>a very few ones, such as the fact that pow(a,b,c) returns the
>same result as (a**b)%c "but faster" -- or the fact that
>passing a comparison function to the sort method of a list

All this does sound very good for learning to program.

>"slows down the sort").  Only much later (in Chapter 17,
>"Testing, Debugging, and Optimizing") do I cover the big-O
>behavior of operations on lists, strings and dicts -- including
>the key facts, from which almost all others follow, that lists
>are internally implemented with vectors, dicts with hash tables.

Hmm. "Only much later...the key facts...from which almost all other
[big-O facts] follow". Is there a specific reason to hide the
key facts until much later? Why would they be confusing?

If somebody would say "python's list is a really a vector/array",
that's more or less the only documentation one will ever need,
help(list) tells the rest. The key characteristics are obvious from
that one single sentence, right?

>Do you think that this traditional approach (teach functionality
>first, performance issues only later) is fundamentally flawed?

No, people usually want to get things done, it doesn't matter how
fast or elegant it is. *However*, I don't consider *anybody*
to actually have learnt programming if they don't know that
- programming is about meeting requirements
- those requirements usually involve correctness to a certain
degree, _but_ also memory usage, speed, user interface functionality,
error handling, modularity, documentation, extendability.

So it's not just a functionality-performance relationship.

>That's basically what I see the "first thing I wondered about"
>comment as potentially implying, and what I'm trying to get a
>deeper understanding of, particularly when you seem to single
>out Python wrt other languages in this respect.  Of course, in

This was not my intention: I would have the same concern for
any language :-)

Regarding the implication, I really couldn't say.

>the Nutshell I was constrained by other considerations too (it's
>primarily a *reference* work, only secondarily intended to be
>read from the start onwards -- having decided to devote a part
>of a chapter to optimization, it seemed to me that optimization
>related information was best grouped there, rather than spread
>around -- and _repeating_ information would go against the grain
>of what's meant above all to be a *concise* reference...).  But,
>that book may well not be the last time I have occasion to write
>about Python (or anyway teach it...) -- so, I think I need to
>understand these issues more deeply.

Right, that can't be easy. As you said, repeating is almost out of the
question, but on the other hand, sometimes you have to see problems
from different aspects, and you'd still like to repeat information...

>in the cache.  (Incidentally, cache behavior and other such issues of
>memory hierarchies are by far the most complicated performance aspects
>to understand, analyze or predict -- and yet far too often they can at
>the same time be _dominating_ the performance behavior of your program...
>and no language spec is going to be very helpful there;-).

Indeed.

"Almost all programming can be viewed as an exercise in caching"
  --Terje Mathisen


Hopefully this hasn't turned into incoherent rambling, cheers,
Marcus

-- 
Marcus Alanen
maalanen at abo.fi



From max at alcyone.com  Wed Apr 23 00:53:09 2003
From: max at alcyone.com (Erik Max Francis)
Date: Tue, 22 Apr 2003 21:53:09 -0700
Subject: Twisted 1.0.4
References: <20030421110438.4991f713.itamar@itamarst.org> <006901c3081c$8240d7e0$0200a8c0@b> <20030422013947.GB7666@frobozz.local> <341410000.1051007646@leeloo.intern.geht.de> <20030422121836.GA8567@frobozz.local> <1051031635.18957.24.camel@software1.logiplex.internal> <mailman.1051064256.4034.python-list@python.org> <3EA60532.6ADC21C@alcyone.com> <mailman.1051070318.3249.python-list@python.org>
Message-ID: <3EA61C35.BEAD61A4@alcyone.com>

Andrew Bennetts wrote:

> Yep, I'm aware of this (but I don't mind you pointing it out, because
> it
> might not be obvious to some).

Good, it was taken in the manner intended.

> I believe gcc can also be used on proprietary unix systems where the
> standard libc is *not* glibc, but the vendor's own implementation. 
> For that
> matter, don't the various BSD flavours have their own, BSD-licenced,
> libc?

I don't know, probably.  (I don't fiddle with BSD guts much.)  You can
run gcc -v to build simple programs and see what default linking steps
it takes.  Those can really be changed to do whatever you like.  By
default, though, they will be implicitly linking against LGPL stuff, but
strictly speaking they need not; they're separate processes, but by
default gcc does does the ugly stuff for you.

> Again, my point is that gcc isn't relevant to the licence of C code
> compiled
> with it; it's just a compiler, and compiled code isn't a derived work
> of the
> compiler, or whatever the magic legalese in the GPL is.  The libraries
> compiled into or linked with your code are a seperate matter, and can
> have
> licences relevant to your code.

Indeed.  As another example, emacs is GPL'd software, but any text file
that I create with it is most certainly not affected by that license. 
As for a particular example on my side of the fence, my templating
system EmPy is released under GPL, but obviously any material you
process with it will be unaffected by that license.  It would only be if
you incorporated EmPy has a module in a larger application that GPL's
licensing requirements would apply.

Even Microsoft EULAs fully acknowledge this sort of thing; Microsoft
Word would be a good example.  Even beasts such as Microsoft Comic Chat
can be used to create monstrosities such as [HORRIBLE WEB COMIC
REFERENCE EXPURGATED], and Microsoft can't do anything about it, as much
as they'd invariably prefer to distance themselves from such creations.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ 'Tis man's to fight, but Heaven's to give success.
\__/ Homer
    Maths reference / http://www.alcyone.com/max/reference/maths/
 A mathematics reference.



From adalke at mindspring.com  Wed Apr  2 12:25:20 2003
From: adalke at mindspring.com (Andrew Dalke)
Date: Wed, 2 Apr 2003 10:25:20 -0700
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
Message-ID: <b6f66v$28k$1@slb3.atl.mindspring.net>

robin:
> If you want the biggest, boldest approach and don't care about
> overhead, use CORBA.

What overhead would this be?  From what I see of omniORB,
there isn't really that much.  Also, CORBA is the most complete,
eg, it allows callbacks and passing around object references
the others don't have.

I've wanted to do something in CORBA for years.  I've
never gotten there.  One problem is that I'm used to Python,
where I don't need to describe the interface beforhand.
CORBA wants that IDL, and a change in the object's interface
 must be reflected in the IDL.  That just seems tedious to
me now.

I also do nearly everything in Python, so don't need the
ability for different langauges to interoperate.  I just pass
around Python objects.

I keep hoping that one of the component systems (like
in GNOME or KDE) takes off, so that "scripting" a la
COM takes off in unix systems, but I've been hoping
for that for the last 5 years.

> If you want a simpler approach but with performance and
> implementation difficulties, use SOAP or XML-RPC.

I wouldn't put SOAP as simple, and I've had problems with
interoperability between various packages.  If I needed to pass
simple data around, I would use XML-RPC.

> If you want the ultimate in simplicity and are willing to foresake
> multi-language support, use Dopy, Pyro, or Twisted Spread.

One reason I'm looking at Twisted is because it handles other
interfaces as well.  I need to talk to SQL databases, SOAP and
XML-RPC servers, straight HTTP, and spawned off external
processes.

There is also older interfaces, like PVM and MPI from the
high-performance computing world, and Linda, and more,
but I haven't looked into them for years.

                    Andrew
                    dalke at dalkescientific.com





From anton at vredegoor.doge.nl  Fri Apr 25 07:43:10 2003
From: anton at vredegoor.doge.nl (Anton Vredegoor)
Date: Fri, 25 Apr 2003 13:43:10 +0200
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com> <b86bqd$iu4$1@news.hananet.net> <b86e9f$c15$1@panix1.panix.com> <3EA74EC0.4BAB336@engcorp.com> <vPLpa.51482$LB6.1161039@news1.tin.it> <3EA80534.4B3FFA43@engcorp.com> <QsYpa.3418$K35.76394@news2.tin.it> <3EA87475.F6A19BE@engcorp.com>
Message-ID: <b8b7kj$g74$1@news.hccnet.nl>

Peter Hansen <peter at engcorp.com> wrote:

<throwaway scripts>

>I can't see any point in continuing the discussion, can you?

This discussion about whether throwaway scripts tend to get thrown
away or not is too heavily based on the wrong Dutch quotes. A more
fitting quote would be "ieder voordeel heb zijn nadeel" from a famous
Dutch soccer player (and later soccer trainer in Italy). It means as
much as "every advantage has its disadvantage".

In modern Dutch society this quote and the reverse standard form of it
are several times more popular than all of Dijkstras quotes taken
together. Since the man also went to Italy later in his career there's
probably also an Italian version of this quote and of several other of
his remarks.

Now to get to the point, I would like to suggest that the advantage of
writing throwaway scripts is that there's no Dijkstra or employer
looking over ones shoulder saying; "Are you serious? Is this the way
you are spending my time and money?".

On the other hand, if a programmer has become any good in the art
there's no way this could have been accomplished without having been
playing with code a lot. To a certain extent playing *is* learning.

The problem that comes up is that often, while playing, pieces of code
are produced that could become important later. After all this is the
source of almost all coding creativity isn't it?

So now we have a respectable programmer with all kinds of crazy code
snippets, some of which could become of considerable value to himself,
to the employer, and to other coders. However, there's no way to share
them without losing credibility among the other "professional" coders.
After all, if making money is one of the main objectives, it becomes
contraproductive to share ones less successful attempts[1].

This is where pair coding, sprints, inofficial meetings, answers to
usenet postings of beginners, trivial conversations about quotes etc.
come in handy. It enables the more experienced programmers to share
their knowledge without being made responsible for the crazy stuff.
The downside  of this -yes we're still following Johan Kruif's trail-
is that there's sometimes a lot of boasting necessary to compensate
for engaging in such activities.

It would be better to forget about money -and Dijkstra- all together
in open source communities and just think in terms of the time that
one is willing to contribute to a project.

If a lot of that time is spent playing with the code, the community as
a whole would profit more than from having payed coders for a project.
(Unless someone is willing to hire me for a project of course, I'm
cheap :-) 

So, concluding, write as much throwaway scripts as you can, just don't
throw them away :-)  

Someone else, maybe even yourself will "profit" from them later.

Anton

[1] This problem is older than coding, because it was the reason for
the creation of a scientific community. I consider it a big social
problem that nowadays companies have sufficient amounts of money to
destroy this community by buying away the best scientists. 

There's new hope in using the internet as a medium for sharing
knowledge. I can't see any scientist or professional programmer
retaining the status associated with formerly held positions when
using the new medium however, since such positions are generally
acquired by restricting access for most people, while the internet has
a tendency to work around such barriers. While this is generally a
good thing, the transformation of the old values and the accompanying
social unrest could cause problems.






From ckea25d02 at sneakemail.com  Mon Apr  7 22:41:33 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 08 Apr 2003 14:41:33 +1200
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <mailman.1049687898.18892.python-list@python.org>
References: <m2u1dcw22s.fsf@localhost.localdomain>	 <6ee58e07.0304061144.35c5ab92@posting.google.com>	 <m2k7e78j5s.fsf@localhost.localdomain>	 <mailman.1049662570.29853.python-list@python.org>	 <yfsr88fkppr.fsf@black132.ex.ac.uk> <mailman.1049687898.18892.python-list@python.org>
Message-ID: <b6tcb5$8p1s0$1@ID-169208.news.dfncis.de>

Ian Bicking wrote:
> Given a novel form of input device, perhaps there's something to what
> he's trying to do.  I'm skeptical, though.

Foot pedals. Hold down the Search pedal and type away.

Maybe one of those car-racing-game accessories could
be used to prototype this...?

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From nhodgson at eb2.net.au  Wed Apr 16 20:21:56 2003
From: nhodgson at eb2.net.au (Neil Hodgson)
Date: Thu, 17 Apr 2003 00:21:56 GMT
Subject: Using property in classic class may not work
In-Reply-To: <b7hies$ol8$0@216.39.172.122>
References: <cYHma.5800$hF.46142@news-server.bigpond.net.au> <ifPma.15344$LB6.409776@news1.tin.it> <a8b7f07a.0304150527.5b0b1187@posting.google.com> <hRUma.23848$T34.709378@news2.tin.it> <b7hies$ol8$0@216.39.172.122>
Message-ID: <Esmna.9752$hF.75996@news-server.bigpond.net.au>

Bengt Richter wrote:

> I may be fooling myself here, but a quick experiment suggests being
> able to define a new class that inherits from both a classic class
> that you can't touch and a new-style one that you can give a property, e.g.:
> 
>  Python 2.2.2 (#37, Oct 14 2002, 17:02:34) [MSC 32 bit (Intel)] on win32
>  Type "help", "copyright", "credits" or "license" for more information.
>  >>> class Old3rdParty:
>  ...     def oldMethod(self): print 'oldMethod was called'
>  ...
>  >>> class NewPropertyCarrier(object):
>  ...     def __init__(self, v): self.prop = v
>  ...     def _get_p(self): return self._v
>  ...     def _set_p(self, v): self._v = v
>  ...     prop = property(_get_p, _set_p)
>  ...
>  >>> class Derived(Old3rdParty, NewPropertyCarrier):
>  ...     pass
>  ...
>  >>> d = Derived(123)

    Thanks. Good idea. I implemented this but it still upsets wxPython.

    Neil




From a_abdi406 at yahoo.com  Thu Apr  3 23:19:19 2003
From: a_abdi406 at yahoo.com (abdi)
Date: 3 Apr 2003 20:19:19 -0800
Subject: about NLTK in Python
Message-ID: <97e89849.0304032019.28a612ad@posting.google.com>

hi,

I have question about the modules, I downloaded NLTK module for win32
and I followed the instruction for installation from the web, having
dowloaded I the wizardinstallation guided me to complete the
installation and it installed the same directory as Python22,  I
opened the python interactive window to check whether the installation
was successfull I called the module as follows:

>>> from nltk.token import *
Traceback (most recent call last):
  File "<interactive input>", line 1, in ?
ImportError: No module named nltk.token
>>> 

I consulted the FAQ before posting the question, so what is going on ?

folder structure is as follows:

C:\\Python22\\nltk1.0

                   

thanks in advance



From danra at walla.co.il  Tue Apr 15 17:22:46 2003
From: danra at walla.co.il (Danra)
Date: 15 Apr 2003 14:22:46 -0700
Subject: super-problem
References: <J_Xma.4475$Up3.2766033@twister.nyc.rr.com>
Message-ID: <fd1c664e.0304151322.bb3a98d@posting.google.com>

I posted a question with (almost) exactly the same subject about a
week ago. Check it out.



From robin at jessikat.fsnet.co.uk  Thu Apr 17 19:42:33 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Fri, 18 Apr 2003 00:42:33 +0100
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org>
 <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk>
 <Xns9360989556563duncanrcpcouk@127.0.0.1>
Message-ID: <3XAcSgApvzn+Ewvl@jessikat.fsnet.co.uk>

In article <Xns9360989556563duncanrcpcouk at 127.0.0.1>, Duncan Booth
<duncan at NOSPAMrcp.co.uk> writes
>Robin Becker <robin at jessikat.fsnet.co.uk> wrote in 
>news:XiCE$IAjFmn+Ew43 at jessikat.fsnet.co.uk:
>
>> I get the impression that XP is intended as a design none/test all
>> process where the code some grows organically into the requirement
>> through bursts of coding and testing. It even seems as though the
>> testing is primary.
>
>XP ranks design very highly. Its just that it recognises that coding is an 
>aspect of design. 'Big design up front' thinks design is something you do 
>once before you start coding, XP thinks that design is continuous 
>throughout the process.
.... interesting
>
.........
>What makes test driven development so powerful is that you write the 
>specification while thinking about WHAT the code has to achieve. If you 
>write the test after the code you are inevitably thinking about HOW the 
>code achieves its results. Writing the test first means your tests relate 
>to the specification rather than the implementation.
>
>You are correct to some extent, if the programmer didn't think of 
>particular cases when writing the tests, then XP almost demands that the 
>code doesn't support those cases. In particular, deliberately writing code 
>that fails on all data other than the tested data is a common way to do 
>TDD: Write a test, make sure it fails for the expected reason, then write 
>minimal code to make it pass (which might involve hardwiring the result), 
>make sure the test passes, refactor all the while making sure all tests 
>still pass.

so we have terms in the utility which penalize rather than benefit.

>
>The refactoring stage is the one that (hopefully) renders the code as short 
>and clean as possible, which includes removing special case results and 
>making the code support the general case.
>

this is like a secondary utility or cost function ie amongst candidates,
which are equivalent in primary terms, we prefer candidates with better
secondary costs.

>If later in the process you discover a case that you overlooked, then it is 
>a simple matter to add a test for the new case and modify the code to 
>support it while being sure you aren't breaking other cases.
>
Each test covers a subset of the possibilities. I guess the difficulty
is in ensuring that the tests cover a sufficient volume of them (ideally
all) and that the tests give the right answers for each of those
covered.
-- 
Robin Becker



From lx1978 at 263.net  Sun Apr  6 15:56:06 2003
From: lx1978 at 263.net (Xin Liu)
Date: Mon, 7 Apr 2003 03:56:06 +0800
Subject: Is there swap function in Python?
Message-ID: <b6q0n6$1c55$1@mail.cn99.com>

I know it is a too sample question. I can't find anything in my python
documents about the swap funtion. Of cause, I can write a swap use boost or
an other way. But the language has grown for 13 years. Why not its library
include a so basic funtion?

Excuse me, My English is poor. It isn't my mather tongue.





From boogiemorg at aol.com  Sun Apr  6 20:49:56 2003
From: boogiemorg at aol.com (David Morgenthaler)
Date: 6 Apr 2003 17:49:56 -0700
Subject: pb py32exe and PIL
References: <20030406074208.01628.00000159@mb-cb.aol.com>
Message-ID: <f9b3eae9.0304061649.1d7df6d8@posting.google.com>

ryomanga at aol.com (Ryomanga) wrote in message news:<20030406074208.01628.00000159 at mb-cb.aol.com>...
> As in the title I got a pb with py2exe and the PIL library.
> When I use the exe i've got the following error message :
> 
>   File "Image.pyc", line 960, in open
> IOError: cannot identify image file
> 
> It seems it can't read my pictures. I've try to change their format and
> directory but nothing changes. Is there something special to do with PIL like
> with PMW??
> 
> Thanks for your help
>      Guilh?m

This did it for me:
>python setup.py py2exe -pPIL



From nomad*** at ***absamail.co.za  Sun Apr 13 17:08:35 2003
From: nomad*** at ***absamail.co.za (Nomad)
Date: Sun, 13 Apr 2003 23:08:35 +0200
Subject: OT: Python application deployment?
References: <14bd1399.0304091426.1b08e005@posting.google.com> <b766fi$p3b$1@newstree.wise.edt.ericsson.se> <16411664.REEpLcBOk5@lothar.cengija.com> <amfd9vsugg1hd0cd8qrjdm033opugi5i6e@4ax.com> <1218043.6iNpJ2gbPL@lothar.cengija.com>
Message-ID: <p86j9vsnoh2eaircth84sbp6ub7v4e6pek@4ax.com>

On Sat, 12 Apr 2003 21:44:26 +0200, Davor Cengija
<dcengija_remove_ at inet.hr> wrote:
<snip>
>> Of course not -- it's got to be bloated, but pretty...
>
>Python+wxWindows combination looks pretty. It would help (in such a 
>situation) to have the dropdown menus with shadows etc (a la Windows XP), 
>but then I'd be required to sell my software. I mean, you cannot have a 
>good program (read: pretty looking program) which is distributed for free. 
>"Something fishy is going on. Don't install it, it's a virus!"
>
> :-)

That's why all the best programs are command-line, but the popular
ones have GUI's with cute shadows and large fisher-price style buttons
on the toolbar.

;-)

-- 
Nomad

Wondering of the vast emptyness of the 'net
in search of something cool.



From jepler at unpythonic.net  Tue Apr 15 12:19:37 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Tue, 15 Apr 2003 11:19:37 -0500
Subject: pread/pwrite (was Re: files: direct access - NOT sequential)
In-Reply-To: <S_Sma.22945$T34.690529@news2.tin.it>
References: <mailman.1050398560.8881.python-list@python.org> <S_Sma.22945$T34.690529@news2.tin.it>
Message-ID: <20030415161936.GA30739@unpythonic.net>

On Tue, Apr 15, 2003 at 12:33:54PM +0000, Alex Martelli wrote:
> Patrick Carabin wrote:
> 
> > i search a means to get " direct access " inside files, eg mixing
> > read & write operaions for the same file
> > example:
> > given a file with contents : "qwerty" , positionning on the "w", and
> > writing a "N" , so the file becomes "qNerty"
> > How can i achieve this in Python ?
> 
> thefile = open('thefile', 'r+')    # must be already-existing
> thefile.seek(1)
> thefile.write('N')
> thefile.close()
> 
> Here, in fact, we're NOT mixing read and write operations -- the
> only _operation_ is a write -- but, if we wanted to, we might (that
> is what mode 'r+' means).

Speaking of which, (if a patch were supplied,) is it remotely likely
that the os module would get pread/pwrite functions?  These make
lseek+read or lseek+write atomic, which is useful if two threads share
the same file descriptor and want to use random-access to it in a
threadsafe way.

(nicest would be if file.read and file.write took an optional pos=
argument, but pread/pwrite are for file-descriptors, not FILE*s so stdio
buffering would louse things up)

Jeff




From tismer at tismer.com  Fri Apr 25 15:19:26 2003
From: tismer at tismer.com (Christian Tismer)
Date: Fri, 25 Apr 2003 21:19:26 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <27cc4d4a.0304232113.57bfc2e7@posting.google.com>
References: <mailman.1050569128.32513.python-list@python.org> <27cc4d4a.0304222142.199ba7c9@posting.google.com> <mailman.1051105417.17659.python-list@python.org> <27cc4d4a.0304232113.57bfc2e7@posting.google.com>
Message-ID: <3EA98A3E.9080609@tismer.com>

Patrick Logan wrote:
>>Can you give me some pointers to timings nd maybe comparisons?
>>I'm not familiar with those langauge's web sites.
> 
> 
> Not exactly an apples-to-apples with your numbers, but it gives you an idea...
> 
> http://www.erlang.org/ml-archive/erlang-questions/200212/msg00058.html

Very good, many thanks!

I will create soft channel switching and see how far
I get. My bet is about 8 Mega messages per sec, for
sending messages with soft-switching between tasklets
built upon a CFunction.

Will report on that, ASAP.

ciao - chris

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/





From mertz at gnosis.cx  Tue Apr 29 11:39:31 2003
From: mertz at gnosis.cx (David Mertz)
Date: Tue, 29 Apr 2003 11:39:31 -0400
Subject: POD vs. reST for standalone writing?
References: <7gd6j6kpq0.fsf@pikespeak.metacarta.com>
 <mailman.1051584944.30800.python-list@python.org>
 <b8kuet$ftn$1@panix1.panix.com> <20030429043815.GB13813@frobozz.local>
 <3EAE801D.4010005@python.org> <mailman.1051625745.20516.python-list@python.org>
Message-ID: <zypr+kKkXE2Q092yn@gnosis.cx>                       

|Yep.  For Twisted's documentation (now over 200 pages as a PDF), I think
|Lore was the right choice (even if we did have to invent it first).  Other
|users will have to make their own decisions about what's right for
|them.

FWIW, I wrote my whole 500+ page book (http://gnosis.cx/TPiP/) using my
own "smart ASCII" format, which is even lighter than reST.  Actually, I
believe it is quite a lot "lighter" (take a look).  The final
book--which I believe is quite attractive--is generated 100% from the
smart ASCII source... including use of fonts, headers, cross-references,
decorative elements, and so on.  The chain is (smart ASCII -> LaTeX ->
PDF -> Publisher puts it on paper; a somewhat more crude HTML outputer
is included in an appendix of the book itself, FWIW).

Then again, there is a lot of special casing involved in this format.
The conversion applies to *this specific* book, rather than to every
documentation project.  For example, all cross references are introduced
by the phrase "SEE ALSO," at the start of a paragraph, and the stuff
that comes after has to be a comma separated list of linkable targets.
But given how I want to book to be, this is no more and no less than
what is needed.  Something like the following is extremely easy to
write... and to read without any special training in the format:

  SEE ALSO, [float], `int`, `long`, `sys.maxint`, [operator]

It's true that there is something slightly "funny" about the square
brackets and backticks.  You would not necessarily know at a glance that
the brackets surround module names, and backticks surround standard
function names.  But not knowing it does not distract much from reading
(and the same convention is used in all contexts). reST has a similar
property in readability, but with a larger "vocabulary" of special
names (which is both more expressive and more confusing).

Similarly, starting a header with the words "CHAPTER" or "SECTION" or
"TOPIC" each mean something a bit different to the final output.  But
deciding among those is *exactly* what I need to do while writing, it's
not at all a matter of *remembering* an odd syntax variant.  Even
the horizontal lines of dashes I occassionally use in the source are
only there to make the source look nicer in 'less', they don't affect
the processing.  And for a reader of the ASCII format, this is a
transparent, IMO, as anything can possibly be.  But again, the
organization and "markup" is specific to a book of a fairly similar sort
to mine.

As well...  I lied slightly in the above.  I also did the compositing of
my book.  I can generate a *pretty good* book with no annoying special
punctuation/markup.  But for a *final* book, there are a lot of things
to worry about:  whitespacing between words cannot be too loose,
consecutive lines should not begin with the same words, facing pages
must balance in length, successive lines should not both end with
hyphen, and so on.  At the end of the process (and not at the prior
URL), I used a number of really picky and obscure extra punctuation to
adjust for these camera-ready requirements.  Mostly, I chose a bunch of
hi-bit ASCII characters or obscure punctuation sequences to force layout
issues on a page-by-page basis.  If I were to change the page size by
1/4" or change any fonts, these odd markups would need to go in entirely
different places.  But frankly, I don't think that reST, DocBook, Lore,
or anything else, would manage to avoid such manual final touches for
book-quality output.

Yours, David...

--
 mertz@  _/_/_/_/ THIS MESSAGE WAS BROUGHT TO YOU BY: \_\_\_\_    n o
gnosis  _/_/             Postmodern Enterprises            \_\_
.cx    _/_/                                                 \_\_  d o
      _/_/_/ IN A WORLD W/O WALLS, THERE WOULD BE NO GATES \_\_\_ z e





From robin at jessikat.fsnet.co.uk  Fri Apr 11 15:08:03 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Fri, 11 Apr 2003 20:08:03 +0100
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com>
 <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
 <d116fbae.0304100704.68ef4833@posting.google.com>
 <c48$3e960655$516049d2$3498@nf1.news-service.com>
 <3e96efac$0$49107$e4fe514c@news.xs4all.nl>
 <Yr7Xt6APGvl+EwYQ@jessikat.demon.co.uk>
 <3e96f481$0$49099$e4fe514c@news.xs4all.nl>
 <PEEiWdBwIwl+EwpZ@jessikat.demon.co.uk>
 <3e970abc$0$49101$e4fe514c@news.xs4all.nl>
Message-ID: <71M$aHCTKxl+Ew8l@jessikat.demon.co.uk>

In article <3e970abc$0$49101$e4fe514c at news.xs4all.nl>, Irmen de Jong
<irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> writes
>Robin Becker wrote:
>> In article <3e96f481$0$49099$e4fe514c at news.xs4all.nl>, Irmen de Jong
>> <irmen at -NOSPAM-REMOVE-THIS-xs4all.nl> writes
>> 
>>>Sending lots of ints to the server
>>>Time to send 5000 ints, 4.586772 (0.000917 per call)
>>>
>> 
>> is that 5000 connections with new sockets? Seems really fast to me.
>
>No, Pyro reuses socket connections for method calls.
>
>--Irmen
>
sigh :(
for real distributed computing I wouldn't want to hold the sockets open.

I just tried the original server.py/time-client.py pair again this time
on a freeBSD system

freeBSD-->localhost 500 secs
freeBSD-->192.168.0.3 50 secs (a local win32 machine).


I guess I'm just not doing what I think I'm doing.
-- 
Robin Becker



From do-not-spam-ben.hutchings at businesswebsoftware.com  Thu Apr 24 12:16:01 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 24 Apr 2003 16:16:01 GMT
Subject: Degree symbol (UTF-8 > ASCII)
References: <f7199550.0304160930.6cfeca65@posting.google.com>
Message-ID: <slrnbag3e0.ek.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <f7199550.0304160930.6cfeca65 at posting.google.com>, Peter Clark wrote:
> I'm working with a xml document which doesn't include an encoding, so
> it defaults to UTF-8. Of course, all of the text is ASCII, and likely
> to remain so. I would like to insert the degree symbol (chr(176)), but
> because this is outside the bounds (chr(128) is the limit), Python
> raises an XML error. What's the simplest way of getting including an
> unadorned degree symbol?

Since you mentioned an XML error I'm assuming you're using some XML API 
to read and write the file, so you want a Unicode string and the answer
is:
    unichr(176)

chr(176) returns an ordinary string which should be interpreted in the
local encoding, which is ASCII by default; however, ASCII doesn't define
a character 176.  Python doesn't interpret strings as anything other
than arbitrary bytes until you try to print or convert them, so that's
why the error only shows up later.

> Again, it's not necessary to preserve the UTF-8 encoding, but I'm not
> quite certain as to how to tell Python that the XML document is plain
> ASCII (or I guess in the case of the degree symbol, latin-1).

An XML document in some 8-bit or multibyte encoding other than UTF-8
*must* have the encoding specified in its XML declaration.  You should
not try to override this.



From Peter.Havens at Level3.com  Fri Apr 11 12:28:42 2003
From: Peter.Havens at Level3.com (Havens, Peter)
Date: Fri, 11 Apr 2003 10:28:42 -0600
Subject: Confusion: assignments create object references.
Message-ID: <0C642BE4EE085149B52839125D4FC9D422BD9F@idc1exc0002.corp.global.level3.com>

I am confused about when assignments create references to objects as opposed to copying the value of those objects. I realize this is probably a very common bit of confusion, but I haven't had much luck finding anything in the Python FAQ, from googling the comp.lang.python news group (voluminous thread tangents), or from the regular documentation I've read which has only fueled my confusion. To wit, I read the following "rule" in the O'Reilly book "Learning Python": 

"...Python assignment stores references to objects in names or data structure slots. It always creates references to objects, instead of copying objects."

That seems fairly straight forward. So, I ran this test:

>>> L = [0,1,2]
>>> M = L
>>> L
[0, 1, 2]
>>> M[0] = 9
>>> L[0]
9
>>>

So far so good. This works as I would expect given the rule above.

>>> L = 0
>>> M = L
>>> L
0
>>> M = 1
>>> L
1
>>>

However, I can't wrap my head around this. It seems to violate the rule above. The value in L seems to have been copied. What did I miss? What's the real rule about assignments? Are exceptions to the rule somehow made for certain types of assignments? Also, how do I explicitly create a reference to these types of objects since they appear to be copies? In Perl I would do something like this:

my $M;
my $M_reference = \$M;

Is there an equivalent in Python? I couldn't find a helper function.

Any help would be appreciated. Thanks,
Pete




From edreamleo at charter.net  Sat Apr 19 09:09:07 2003
From: edreamleo at charter.net (Edward K. Ream)
Date: Sat, 19 Apr 2003 08:09:07 -0500
Subject: ANN: Leo's new web site.
Message-ID: <va2ijnhrq7mj05@corp.supernews.com>

Please note my new email address and Leo's new URL below.

Edward
--------------------------------------------------------------------
Edward K. Ream   email:  edreamleo at charter.net
Leo: Literate Editor with Outlines
Leo: http://webpages.charter.net/edreamleo/front.html
--------------------------------------------------------------------





From BPettersen at NAREX.com  Fri Apr 18 10:15:37 2003
From: BPettersen at NAREX.com (Bjorn Pettersen)
Date: Fri, 18 Apr 2003 08:15:37 -0600
Subject: Problem in Accessing C++ Class Instance Global Variable (near success, help please!!)
Message-ID: <60FB8BB7F0EFC7409B75EEEC13E20192022DDFD9@admin56.narex.com>

> From: Alan Pong [mailto:alanpong at hkstar.com] 

Plase be considerate to people who've been up all night and has to go
through your entire reply to get the context. The expected form almost
everywhere is quote what you're talking about, type your
question/answer, repeat until done, and delete everything else. Like I'm
doing here :-)

> > First, PyImport_AddModule( "myext1" ), doesn't load the module [...]
> >
> 1. i did used Py_InitModule.

Please study the manual (e.g. look up what it says the two functions
above do). This topic has too many details to teach introductions here
without rewriting the manual. FWIW, you want something close to:

  Py_Initialize();

  PyObject* mainmod = PyImport_AddModule("__main__");
  m_namespace = PyModule_GetDict(mainmod);
  Py_INCREF(m_namespace);

  PyObject* nrx = PyRun_String("import nrx\n", Py_file_input,
m_namespace, m_namespace);

  if (nrx == NULL) {
    throw new NException(getTraceback().c_str(), 25, __FILE__, 0);
  } else {
    Py_DECREF(nrx);
  }

  FILE* fp = fopen("test.py", "r"); // you seem to have missed this
step?
  PyObject* result = PyRun_File(fp, "test.py", m_namespace,
m_namespace);
  etc., etc.

> > The symptoms imply that your "library" (wrapper.cpp) is most likely
> > compiled statically against both your program (test.exe) and your
dll
> > (myext1.cpp). [...explanation why that's two different 'varables' at
> > runtime...]
> >
> 2. correction: wrapper.cpp compiled to wrapper.obj which is linked to
> form myext1.dll.   test.exe doesn't contain wrapper.obj

Then there is something horribly wrong with what you sent us. If
wrapper.cpp and myext1.c are compiled to object files, there is an
external reference to gtest in wrapper.obj. The linker must find it
somewhere to be able to create a dll. Either you're also linking in
test.cpp with the dll (the only file with a defintion of gtest),
resulting in what I described above, or your code and the code in the
email are different.

> > fix1: "Don't do that" <wink>.
> > fix2: pass the object as an argument.
> >
> 3.  sorry i don't understand fix2 and fix3 mentioned below.

#1 Global variables are considered a bad solution to most problems,
usually when you think you _have_ to have it, there is a better and
simpler way.

#2 One way of not referencing a global G in function f is to pass G as
an argument to f.

> 4. i re-think about: can i access a global variable like this???? ...
> 
> any sample code please?

These are really very fundamental C and dll issues you're dealing with
(by now you should have checked the address of gtest when you're in main
and when you're in SetString etc.). I wouldn't know where to start
helping you without rewriting your program in some form, and I doubt
that will help very much...?

-- bjorn





From cain72 at hanafos.com  Wed Apr 23 22:45:38 2003
From: cain72 at hanafos.com (¹Ú±Ù¹ü)
Date: Thu, 24 Apr 2003 11:45:38 +0900
Subject: [Question] I want to know what script more speedy....  VB script and Python Script...
Message-ID: <b87h29$bju$1@news1.kornet.net>

Plz... answer me...





From peter at engcorp.com  Thu Apr 10 12:58:15 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 10 Apr 2003 12:58:15 -0400
Subject: Python compared toJava
References: <c9e6ac92.0304100824.eec0f5a@posting.google.com>
Message-ID: <3E95A2A7.238BFA6C@engcorp.com>

Blaktyger wrote:
> 
> I recently read that the Python interpreter is like the Java Virtual Machine.
> Is that true?

Yes!  In fact, it's exactly the same thing!  We just use a different name
for it to avoid the wrath of Sun Microsoftsystems, and because we kind of like
snakes.

Note: see http://www.catb.org/~esr/faqs/smart-questions.html before posting
again please. 

-Peter



From kp at kyborg.dk  Wed Apr  9 12:06:00 2003
From: kp at kyborg.dk (Kim Petersen)
Date: Wed, 09 Apr 2003 18:06:00 +0200
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <3E9040CF.2040606@removethis.free.fr> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E9267F6.9020605@kyborg.dk> <yfsu1d9hqw6.fsf@black132.ex.ac.uk>
Message-ID: <3E9444E8.4070805@kyborg.dk>

Alexander Schmolck wrote:
> Kim Petersen <kp at kyborg.dk> writes:
> 
> 
>>Alexander Schmolck wrote:
>>
>>>Kim Petersen <kp at kyborg.dk> writes:
>>
>> >> (snip)
>>
>>>>function in elisp - do a shallow copy of local variable space (dictionary) and
>>>>send it as the local vars - you can safely throw away the dict after function
>>>>returns).
>>>
>>>Ahem, there are several thousand variables kicking about in global in emacs,
>>>you want to copy them all for each function call?
>>
>>look above - *Local* variable space ... globals stay globals.
> 
> 
> Maybe we have a misunderstanding. 
> 
> I thought in your scheme every call to a lisp function foo (from python *or*
> another lisp function) is to be implemented by making a copy of the locals and
> passing it to foo and this copy is discarded after return. If so, this won't
> work (because foo can mess with its callers *locals*; also globals *don't*
> stay globals -- a let binding of a global name means it can no longer be
> globally modified further up the call tree).

No misunderstanding... and possibly not efficient - thats why i said 
that you could do push/pop of bindings instead... since then i 
remembered that elisp (or any other lisp for that matter *doesn't* have 
global variables - so it would be inefficient). But to make it efficient 
and usable you can do it with a class instead (that simulates a dictionary):

class Borg: pass

class LispNameSpace:
    def __init__(self):
       self.nstack=[]
       self.vars={}
       self.undefined=Borg()
    def getcontext(self): return len(self.nstack)
    def restorecontext(self,n):
       for i in xrange(len(self.nstack),n,-1):
          if self.nstack[i][1]==self.undefined:
             del self.vars[self.nstack[i][0]]
          else:
             self.vars[self.nstack[i][0]]=self.nstack[i][1]
    def localvar(self,name,value=Null):
       if self.vars.has_key(name):
          self.nstack.append((name,self.vars[name]))
       else:
          self.nstack.append((name,self.undefined))
      self.vars[name]=value

only thing missing now - is overwriting __getitem__,__setitem__ to do 
lookup and set of variables - possibly on failure to find a name in 
namespace default down to python namespace - making it really usefull.

implementation of let:

def let(cons,namespace):
    # of course a scheme to reasonably map lisp functions to python
    # implementations is needed - i just assume that (f x y z) is called
    # here as f(lisplist(x,y,z),namespace)
    #
    context=namespace.getcontext()
    defs=cons.car()
    body=cons.cdr()
    if iscons(defs.car()):
       namespace.localvar(defs.car().car(),defs.car().cdr().car())
    else:
       namespace.localvar(defs.car())
    lispeval(body,namespace)

> 
> A better example to clarify:
> 
> (defun foo()
>   (message "when foo is called folding is: %s" case-fold-search)
>   (setq case-fold-search :evil))
> (defun bar()
>   (message "when bar is called folding is: %s" case-fold-search)
>   (let ((case-fold-search t))
>     (foo)
>     (message "what did foo do to folding? %s!" case-fold-search))
>   (message "bar's last folding is: %s again" case-fold-search))
> (defun quux()
>   (message "before bar:%s" case-fold-search)
>   (bar)
>   (message "after bar:%s" case-fold-search))
> (quux)
> 

and all this would be handled gracefully.

> 
> before bar:nil
> when bar is called folding is: nil
> when foo is called folding is: t
> what did foo do to folding? :evil! ;; !!!
> bar's last folding is: t again 
> after bar:nil
> 
> Tell me if you think this would work, too and I'll just reread your message.

Yep it will - and no need to look back - i think i clarified?

> 
> 
> 'as





From andrew at acooke.org  Mon Apr 28 15:40:07 2003
From: andrew at acooke.org (andrew cooke)
Date: Mon, 28 Apr 2003 15:40:07 -0400 (CLT)
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
In-Reply-To: <vaqv0fg30kdc85@corp.supernews.com>
References: <mailman.1051532982.25352.python-list@python.org> 
     <UXara.22428$K35.745468@news2.tin.it> 
     <yu99sms2k35f.fsf@europa.research.att.com> 
     <mailman.1051548461.3290.python-list@python.org> 
     <vaqv0fg30kdc85@corp.supernews.com>
Message-ID: <1609.10.2.0.2.1051558807.squirrel@127.0.0.1>

claird at lairds.com said:

> In article <mailman.1051548461.3290.python-list at python.org>,
> andrew cooke <andrew at acooke.org> wrote:
[...]
>>[1,4)+[4,7) = [1,7)
>>
>>isn't there some (fairly old) popular math book by a famousish person
>>(knuth?  hofstadter?) that develops maths (at least arithmetic) based on
>>defining sets of points from the "number line"?  i wonder what that says
>>about this?  (i should admit that i picked up the above from one of
>>celko's db books).
[...]
> So:  there are several relatively popular mathematically-oriented
> books that illustrate, for example, Peano arithmetic.

i'm pretty sure that it specifically mentioned "the number line".  it
could have been a pamphlet rather than a book (i have a curiously clear
memory of reading it in a library at college (dusty and with shelves made
from that meccano-like metal stuff) - possibly the library at the dept of
app maths & theoretical physics, although i have no idea why i'd have been
there (i was a physicist (and not a very theoretical one), not a
mathematician)).  it was pretty old then (15 years ago).

(another example of doing this kind of thing is modelling (to use the
correct term - thank you!) natural numbers as functions in lambda calculus
which must be the most repeated and most boring example in functional
programming textbooks...)

now i need to go and google for peano arithmetic :o)

cheers,
andrew

-- 
http://www.acooke.org/andrew




From sismex01 at hebmex.com  Wed Apr 30 09:57:01 2003
From: sismex01 at hebmex.com (sismex01 at hebmex.com)
Date: Wed, 30 Apr 2003 08:57:01 -0500
Subject: Three questions about Tkinter
Message-ID: <F7DB8D13DB61D511B6FF00B0D0F06233048D6837@mail.hebmex.com>

> From: Arnal [mailto:arnaud at tribu.ch]
> Sent: Wednesday, April 30, 2003 8:43 AM
> 
> Hi,
> 

howdy!

> I'm new at python programming (as well as english writing!), 
> and I'm trying to make a simple text editor.

You too?  ;-)

> 
> I'm using Tkinter, but I've the following problems:
> 
> -I can't figure out how to use a printer to print under Windows.
> 

That's a problem with Tk/Tkinter.  I don't know if Tk has a printer
output mode; I *do* know that the Canvas widget ("wee-dgit" ;-)
can generate postcript, probably describing it's current contents,
but I've never used it.

Printer output is a whole can of worms of it's own.

OTOH... if you want to simply output the contents of a text
control to the printer, minding not any fancyness, you can
send it simply as text:

...<python code>
# Open printer as output file.
prn = open("PRN:", "wt")

# Send contents of text widget to printer file.
prn.write(text_widget.get())

# And send a form feed for good measure.
prn.write(chr(12))
prn.close()
...<end python code>

> -I use a "text" control (I dislike the "widget" term, sorry), 
>  but I can't resize it on the fly (the width and height properties
>  are measured in characters and lines, not in pixels). My current
>  workaround is to make the window not resizable but it would be
>  nice to do it.
> 

Yes indeed.

When you .pack() your text widget, be sure to add the following
arguments to pack():

    fill="x|y|both"   -> tells your text widget that you want it
                         to fill the current space in x, y or both
                         directions.

    expand="yes"      -> tells the text widget to grow along with
                         it's container.

> -I've a sub menu for recents files. When I read in my pref 
>  file to list every file paths stored in it, I would like to
>  add them in the recents submenu. I can't find how to do it.
>  I tried, inside a "while" with the variable i:
> 
> MRecents.insert_command(0, label=GetFileName(file),
>                         accelerator=acc,
>                         command=OpenRecent(i))
> 

OK, you've a problem here.  "Openrecent(i)" is a function call,
so unless it returns a function reference (er, "function pointer")
as a result, then it's surely not doing what you want.

> MRecents.insert_command(0, label=GetFileName(file),
>                         accelerator=acc,
>                         command=OpenRecent i)

er... this ain't Lispy nor Perly ;-)

> 
> (MRecents is a global variable representing my submenu, 
> GetFileName is a function, no special meaning in the problem,
> acc is defined). 
> In the first example, the OpenRecent function is called directly,
> that's the normal behavior, and there is a syntax error in the
> second example.
> How to define an array of menu items with all calling the 
> same function with a parameter?
> 

Most probably, a function reference, dynamically created,
which upon evaluation calls your given function with the
arguments you defined for it.  You Want A Lambda (run! hide!).

Does this work?

MRecents.insert_command(0, label=GetFileName(file),
                        accelerator=acc,
                        command=lambda : OpenRecent(i))

> 
> Has someone any information on one of these problems? It 
> would be great.
>
> Thanks in advance.
>

You're welcome, happy pythoning :-)

-gca




From staschuk at telusplanet.net  Tue Apr 29 12:21:07 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Tue, 29 Apr 2003 10:21:07 -0600
Subject: Python code -> to C code
In-Reply-To: <b8lpof$q0f$1@pegasus.csx.cam.ac.uk>; from hygt2@cam.ac.uk on Tue, Apr 29, 2003 at 01:07:02PM +0100
References: <b8kk2i$p78$1@pegasus.csx.cam.ac.uk> <3EADEEDD.7B740C6C@engcorp.com> <b8lpof$q0f$1@pegasus.csx.cam.ac.uk>
Message-ID: <20030429102107.B338@tibia.amotlpaa.bogus>

Quoth Garrison Tsang:
  [...]
> There is a program written in Python by a friend.  It takes in a file and
> decrypts it , which I would like to integrate into a C program I am writing.

Integrating Python code into a C program is called "embedding",
that is, embedding the Python interpreter in your C program so the
interpreter can run the Python code for you at request.

> I want to see how the python code was implemented and whether I can
> implement it is C.

Now this seems like a completely different and unrelated
objective.  For this, it seems to me you need to learn enough
Python to read your friend's program, then translate the algorithm
into C.

In principle such conversions could be done automatically, but
Python's dynamic typing (etc.) would probably entail very verbose
output.  Better to translate it by hand.

(Of course, you could just write *your* program in Python.  You
say you don't know any Python but are very familiar with C.  I
will echo others and say that you'll be able to get things done in
Python in a matter of hours, you'll be effective with it in a
matter of days, and familiar with it in a matter of weeks.)

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From jeremy at zope.com  Mon Apr 28 12:52:24 2003
From: jeremy at zope.com (Jeremy Hylton)
Date: 28 Apr 2003 12:52:24 -0400
Subject: __base__ and __bases__[0]
In-Reply-To: <2259b0e2.0304280813.5a94c41f@posting.google.com>
References: <2259b0e2.0304280813.5a94c41f@posting.google.com>
Message-ID: <1051548744.16915.32.camel@slothrop.zope.com>

On Mon, 2003-04-28 at 12:13, Michele Simionato wrote:
> I thought __base__ and __bases__[0] where just equivalent, but it
> seems it is not so:
> 
> >>> class C(object): pass
> ...
> >>> class M(C,type): pass
> ...
> >>> M.__bases__[0]
> <class '__main__.C'>
> >>> M.__base__
> <type 'type'>
> 
> Question: how __base__ is computed and what's the rationale about it ?

__base__ corresponds to the tp_base slot in a type object.  At the C
level, Python has a single inheritance model that determines the memory
layout of instances.  There is a chain involving base classes that
contribute to the instance layout.  __base__ is the base class that is
involved in that chain.

It's possible that there are multiple base classes and none of them
contribute to the layout.  This would occur if each of the base classes
is a plain old Python class.  In that case, I think the choice is of
__base__ is arbitrary.

Here's a concrete example:

class A(object): pass
class B(dict): pass
class C(A, B): pass

C.__base__ is B

because B's base class defines a custom instance layout.

BTW, I don't think this is documented anywhere.  I have to ask Guido
;-).

Jeremy






From john.abel at pa.press.net  Fri Apr 11 12:39:15 2003
From: john.abel at pa.press.net (John Abel)
Date: Fri, 11 Apr 2003 17:39:15 +0100
Subject: string.split() documentation incorrect?
References: <20030411095232.A8283@ash.int.ca.pason.com>
Message-ID: <003501c30049$1af74090$66656565@hallows>

This looks like a bug with 2.2.2.  I can reproduce it with 2.2.2 (Win32).
If I import the string,py from 2.3, it works fine.

Regards

John
----- Original Message -----
From: "Jerry Seutter" <jerry.seutter at pason.com>
To: <python-list at python.org>
Sent: Friday, April 11, 2003 4:52 PM
Subject: string.split() documentation incorrect?


> I'm trying to use string.strip to remove the comma off the end of
> a string, eg:
>
> string.strip("id INT NOT NULL,", ",")
>
> The current documentation says I can do this.  Python 2.2.2 gives
> the following error:
>
> >>> strip('  hello  ', ',')
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> TypeError: strip() takes exactly 1 argument (2 given)
>
> So, I think either the documentation is incorrect or there is
> a bug.  Thoughts?
>
> Jerry Seutter
>
>
> --
> http://mail.python.org/mailman/listinfo/python-list
>





From mis6 at pitt.edu  Mon Apr 28 17:29:33 2003
From: mis6 at pitt.edu (Michele Simionato)
Date: 28 Apr 2003 14:29:33 -0700
Subject: Class attribute or instance attribute?
References: <vaqtlsqbi52j93@corp.supernews.com>
Message-ID: <2259b0e2.0304281329.12ac184b@posting.google.com>

"Paul Watson" <pwatson at redlinec.com> wrote in message news:<vaqtlsqbi52j93 at corp.supernews.com>...
> When is an attribute a class attribute or an instance attribute?

I give you a subtle example:

class C(object):
      x=1 # class attribute
      def __init__(self):
          self.x=self.x # the class attribute becomes an instance attribute


c=C()

C.x=2

print c.x,C.x 

This code prints 1,2. You see that the line self.x=self.x is far
from being trivial: Python looks first for the self on the right,
looking for the instance attribute self.x: since it does not find anything,
it looks for the class attribute; then the class attribute
is assigned to the instance attribute on the left hand side.

If you forget the apparently useless self.x=self.x line, you get a 
different result:

class C(object):
      x=1 # class attribute
      def __init__(self):
          pass

c=C()

C.x=2

print c.x,C.x 

This prints 2,2, since the class attribute has been changed to 2
and c has no instance attribute.

Hoping-having-not-increased-your-confusion-ly,


                    Michele



From a.schmolck at gmx.net  Fri Apr 11 11:36:49 2003
From: a.schmolck at gmx.net (Alexander Schmolck)
Date: 11 Apr 2003 16:36:49 +0100
Subject: Could Emacs be rewritten in Python?
References: <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain> <6ee58e07.0304061722.2472b3d0@posting.google.com> <3E9175A5.1060607@kyborg.dk> <yfsof3ijmua.fsf@black132.ex.ac.uk> <3E92E00B.6000708@kyborg.dk> <mailman.1049814375.24609.python-list@python.org> <b6vq8r$9qev7$1@ID-169208.news.dfncis.de> <mailman.1049911336.19017.python-list@python.org> <yfs3ckrihgi.fsf@black132.ex.ac.uk> <mailman.1049943676.18025.python-list@python.org> <yfsbrzeh3kt.fsf@black132.ex.ac.uk> <9Qhla.7802$d8.3212@nwrdny02.gnilink.net>
Message-ID: <yfswui1f3la.fsf@black132.ex.ac.uk>

Carl Banks <imbosol-1049990205 at aerojockey.com> writes:

> What you're missing is that Globals are Evil(tm), and furthermore, that if

Oh, so *that's* what I was missing -- silly me.

> your program depends on changing a global in a dynamic scope, it badly needs
> to be redesigned. (To wit, that functions that refer to the global you want
> to change need to be changed to accept it as an argument.)

You did notice, of course, that you don't only need to change that particular
function foo, but also everything (directly or indirectly) calling foo? As
well as everything that is called (directly or indirectly) by foo and that
uses this global state information? Including 3rd party code and builtin
constructs of the language?
 
> I think dynamic scpoing is a wholly bad idea, and simplifying a very
> misguided programming practice doesn't make it any less so.

How much time did you actually spend thinking about global variables and
dynamic scope or what I wrote before you posted?

Why are global variables evil?

- because they can lead to poor encapsulation and code that is impossible to
  understand and dangerous to modify because of hidden dependencies
  (parameterizing e.g. logging as an argument for every single method and
  function doesn't exactly help either, however)

- because they make it hard to avoid inadvertent changes to other parts of the
  program (this problem can be much reduced by dynamic scoping which ensures
  that localized and temporary state changes will be automatically cleaned up,
  *even* if there are errors and *even* if you use multiple thread ( e.g. I
  want to only log a particularly critical code section with a high logging
  level -- how do you do that in a multithreaded python application?))

Ever wondered why python (and about any non-functional language) still has
global variables (and makes use of them in core modules such as sys)?

Because certain aspects of a program are orthogonal in a way that is best
captured by state that is not explicitly passed around (even if by chance it
would be possible to do it that way, because you can anticipate and correctly
parameterize all state and have full control over all the involved code (see
above)).

Examples that come to mind are:

* pervasive concerns that affect all or large portions of code such as:
  - logging (e.g. I should be able to change the logging level for a
    particular, critical operation)

  - localization and printing options e.g. output precision, pretty printing,
    maximal print length etc.

  - state information to ease development and debugging (e.g. controlling
    assertions, debugging and hooking into functions for tracing or
    input/output verification etc.). 

* fine tuning aspects of behavior of a class conceptually related functions
  that is are however very widely used

* interfacing to the system or another application and the tracking of changes
  and transactions for error recovery as well as adjustment of details of
  error recovery and handling

All this should *not* require rewriting each and every function and method
call! How do you propose to address these issues (bearing in mind error
handling and multithreading)?

Maybe there is a superior way to handle all or some of those issues (AOP,
Monads, god knows what), but you haven't mentioned it.

Similarly, you have given no indication why you think lexically scoped global
state better address these problems than dynamically scoped state (or why its
downsides such as new sources for bugs and complicating the language would
outweigh the benefits). Rather you just pretended the issue didn't exist and I
was just being stupid for not realizing this to start with.

'as



From tyler at cg1.org  Mon Apr 28 21:15:28 2003
From: tyler at cg1.org (Tyler Eaves)
Date: Tue, 29 Apr 2003 01:15:28 GMT
Subject: saving output to file
References: <MUera.658292$L1.187296@sccrnsc02>
Message-ID: <Qmkra.713$5C2.307@tornadotest1.news.pas.earthlink.net>

Eric Anechiarico wrote:

> I am still not as proficient as I would like to be in Python, so this may
> be a trivial question to most.
> 
> I want to have the output I get in my IDE window (using MacPython 2.2.2
> primarily in Carbon) to be written to a file instead of having to copy the
> output and then paste it to a text file myself.  How would I go about
> doing this?
> 
> Mainly I would be using the code to do this with programs that calculate
> outputs of large numbers (primes, Fibonacci, etc.) and they sometimes
> overflow the buffer and I cannot get back very far to view the beginning.
> Make sense?  Basically since I am doing this in the IDE window anyway.
> 
> Any suggestions?
> 
> Thanks!

import sys
outf = open('somefile.txt','w')
sys.stdout = outf

Leave the rest of your program unchagned.

Has the advantage that you can easily change the behavior back and forth by
commenting out the sys.stdout = line.






From skip at pobox.com  Mon Apr 28 17:14:14 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 28 Apr 2003 16:14:14 -0500
Subject: Linux SUID
In-Reply-To: <16041.21654.333496.299591@montanaro.dyndns.org>
References: <3ea94eae$0$23154$ba620e4c@reader0.news.skynet.be>
        <3EA951F2.10207@ghaering.de> <3EA951F2.10207@ghaering.de>
Message-ID: <16045.39334.107783.63109@montanaro.dyndns.org>

    >> There one can turn on the SUID-exec permission bit which is honoured
    >> by Perl.
    ...
    >> Is there something similar in Python?

    Gerhard> Short answer: No.

    Gerhard> There's a demo suid wrapper in the source distribution ...

    Skip> I think the Mailman distribution has/does something similar, using
    Skip> a compiled C program to do the suid bit.  If my recollection is
    Skip> correct it's likely to be better tested than the demo wrapper in
    Skip> the Python distribution.

I was just checking up on possible software updates for some machines I just
took over at Northwestern and noticed this item on dev.perl.org about the
suidperl capability:

    A potential security vulnerability in the optional suidperl
    component of Perl was identified in August 2000.  suidperl is
    neither built nor installed by default.  As of November 2001 the
    only known vulnerable platform is Linux, most likely all Linux
    distributions.  CERT and various vendors and distributors have
    been alerted about the vulnerability. See
    http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
    for more information.

    The problem was caused by Perl trying to report a suspected
    security exploit attempt using an external program, /bin/mail.  On
    Linux platforms the /bin/mail program had an undocumented feature
    which when combined with suidperl gave access to a root shell,
    resulting in a serious compromise instead of reporting the exploit
    attempt.  If you don't have /bin/mail, or if you have 'safe setuid
    scripts', or if suidperl is not installed, you are safe.

    The exploit attempt reporting feature has been completely removed
    from Perl 5.8.0 (and the maintenance release 5.6.1, and it was
    removed also from all the Perl 5.7 releases), so that particular
    vulnerability isn't there anymore.  However, further security
    vulnerabilities are, unfortunately, always possible.  The suidperl
    functionality is most probably going to be removed in Perl 5.10.
    In any case, suidperl should only be used by security experts who
    know exactly what they are doing and why they are using suidperl
    instead of some other solution such as sudo ( see
    http://www.courtesan.com/sudo/ ).

So, if you count on it in Perl, prepare for it to disappear.  Be extra
careful if you use anything similar with Python such as the suid
capabilities Gerhard and I referred to.

Skip




From cnetzer at mail.arc.nasa.gov  Tue Apr 29 21:23:49 2003
From: cnetzer at mail.arc.nasa.gov (Chad Netzer)
Date: 29 Apr 2003 18:23:49 -0700
Subject: Problem in scalling............
In-Reply-To: <20030430001533.3975.qmail@web41708.mail.yahoo.com>
References: <20030430001533.3975.qmail@web41708.mail.yahoo.com>
Message-ID: <1051665828.8110.55.camel@sayge.arc.nasa.gov>

On Tue, 2003-04-29 at 17:15, perl lover wrote:

> When i scallled
> the canvas stil i got the same value (762*477) if the
> mouse is placed at bottom right of the canvas. Since i
> used 2 as scalling factor for both x,y, I expect
> bottom right value is as half as normal canvas (ie,
> 381,248). How can i get it?

When you scale things in the canvas, you are scaling the objects, not
the canvas itself.  So the mouse coordinates in the canvas do not change
with scaling.  From the Tk documentation for the canvas object:

"Canvases do not support scaling or rotation of the canvas coordinate
system relative to the window coordinate system."

If you want to treat the canvas a a zoomable window, such that you can
transform the coordinate system of the canvas itself, I suppose you
could intercept the canvas scale method, and keep track of all the
scaling for 'ALL' objects, and use that to help compute the zoom factor
for converting mouse coordinates.

Lots of people, inclusing myself, have learned that the Tk Canvas sucks
for doing things like you want to (anything requiring lots of scaling,
rotating, etc.).  In particular, the kinds of transformations you can do
to objects is limited, and those the canvas almost non-existant
(translations only).

You can try using other toolkit canvases (Gtk, Qt, WxWindows?).  You can
also try using OpenGL (which can very easily handle transformations of
it's space, but is a heavyweight rendering library).  The Sketch
software project, written mostly in Python, needed a very powerful and
flexible canvas device, and uses an interface to X-Windows drawing (they
call it Pax) and implements its own primitives; you could take a look at
it, although it is daunting code for a beginner.  It all depends on what
you want to do.

On the other hand, doing things with the Tk canvas, and learning its
limitations and how to deal with them, can be very instructive (and
sometimes very frustrating).

-- 

Chad Netzer
(any opinion expressed is my own and not NASA's or my employer's)





From andersjm at dancontrol.dk  Fri Apr 11 11:22:19 2003
From: andersjm at dancontrol.dk (Anders J. Munch)
Date: Fri, 11 Apr 2003 17:22:19 +0200
Subject: Could Emacs be rewritten in Python?
References: <mailman.1050048561.24768.python-list@python.org> <mailman.1050068661.5348.python-list@python.org>
Message-ID: <3e96dd8e$0$66251$edfadb0f@dread11.news.tele.dk>

"Beni Cherniavsky" <cben at techunix.technion.ac.il> wrote in message 
> 
> So I argue that dynamic scoping done right, unlike simple globals, is
> not always evil.  It's completely equivallent to passing all your
> environment, except that you don't have to write that explicitly,
> which can make all the difference.  Frequently, what you really want
> is to parametrize a *proccess*, not just a specific call.
> 
> For example, in Unix, I/O redirection, the current working directory
> and most other process parameters are dynamically scoped: it's
> implemnented by copying from each process to its children but to you,
> it's just magically inherited.  

And this is a common source of frustration.  Scripts may work fine
from the command line but fail mysteriously when run from init or a
cron job, because somehow the environment is different.  The CWD ought
to have been local to the shell as the notational device for writing
abbreviated filenames that it is; inheriting it just causes problems.
Ever tried replacing "cd somedir; make" with "make -f
somedir/Makefile"?  This will invariably fail because an unexpected
CWD is inherited.

What works well for customising your interactive shell experience does
not work well for programming.

speaking-as-the-author-of-a-lexically-scoped-make-utility-ly y'rs,
Anders

-- 
Anders Munch. Software Engineer, Dancontrol A/S, Haderslev, Denmark
Capon - a software build framework - http://www.dancontrol.net/share/capon/





From staschuk at telusplanet.net  Fri Apr 11 23:20:12 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Fri, 11 Apr 2003 21:20:12 -0600
Subject: instance creatiion as an "event"
In-Reply-To: <pan.2003.04.11.23.20.04.65410@webone.com.au>; from simonb@webone.com.au on Sat, Apr 12, 2003 at 09:20:05AM +1000
References: <mailman.1050095779.7490.python-list@python.org> <pan.2003.04.11.23.20.04.65410@webone.com.au>
Message-ID: <20030411212012.A3063@tibia.amotlpaa.bogus>

Quoth Simon Burton:
> On Fri, 11 Apr 2003 17:15:32 -0400, Arthur wrote:
  [...]
> > The only thing I seem to be missing is some kind of "hook" to trigger a 
> > general update cycle of all existing class instances upon the creation 
> > of a new class instance at the interactive prompt
> 
> class Member:
> 	members = []
> 	def __init__(self):
> 		for m in Member.members:
> 			m.notify(self)
> 		Member.members.append(self)
> 	def notify(self,other):
> 		print self,"a new member:",other

Though you will probably want to use weak references in
Member.members, unless you intend that all instances of the class
live as long as the class does.

-- 
Steven Taschuk                             staschuk at telusplanet.net
"I may be wrong but I'm positive."  -- _Friday_, Robert A. Heinlein




From titanrebel at comcast.net  Wed Apr  9 18:26:09 2003
From: titanrebel at comcast.net (kkennedy)
Date: 9 Apr 2003 15:26:09 -0700
Subject: Python application deployment?
Message-ID: <14bd1399.0304091426.1b08e005@posting.google.com>

Is there a way to deploy the Python "runtime" with your Python
scripts?  I am only concerned with Windows (since most Linux distros
would already have it installed).  I would like it to be smaller than
the 7 MB download available from python.org, and be able to be
installed transparent to the user.  In other words, it would be nice
to be able to simply unzip the files needed to run my scripts without
the user having to download and install the full Python windows setup,
and then download and install my Python scripts seperately.

Are there any tools to make this doable by the common programmer?



From jepler at unpythonic.net  Fri Apr 25 15:40:27 2003
From: jepler at unpythonic.net (Jeff Epler)
Date: Fri, 25 Apr 2003 14:40:27 -0500
Subject: pythong referencing/aliasing
In-Reply-To: <vaioqs7e92v14d@corp.supernews.com>
References: <ylcqa.9116$h%2.963225@read1.cgocable.net> <mailman.1051284819.25791.python-list@python.org> <vaioqs7e92v14d@corp.supernews.com>
Message-ID: <20030425194024.GG30690@unpythonic.net>

On Fri, Apr 25, 2003 at 04:33:32PM -0000, Cameron Laird wrote:
> I'll make a positive assertion of this:  in C and
> C++, stylish programmers frequently use references
> to good effect.  Python rarely, very rarely, needs
> to do the same.  It might be instructive to choose
> a typical C++ reference idiom, and illustrate how
> its typical use case is better served in Python be
> a derived return, or dictionary construction, or ...

Well, some of the places you use a pointer/reference *and* a return
value, you use multiple return values in Python.  For instance:

// two outputs from function, one through a pointer
double modf(double x, double *iptr)

# In Python, return the two values as a tuple
>>> math.modf(3.14)
(0.14000000000000012, 3.0)

Sometimes it's instead of true exception handling (less so in C++):
// return whether an error happened, otherwise stick the result in the
// preallocated buffer and return the length
int read(int fd, void *buf, size_t count)

# In Python, raise an exception if the read fails.  Otherwise,
# return a freshly allocated buffer (which implicitly carries its
# length)
>>> sys.stdin.read(16)
No one expects the spanish inquisition # (typed in)
'No one expects t'

Somtimes it's instead of some stateful information (less so in C++):
// The state needed to continue iterating through tokens is in the
// third arg
char *strtok_r(char *s, const char *delim, char **ptrptr);

# In Python, you'd have an object.  This may not quite match C's strtok,
# but it gives the flavor of the implementation
class strtok:
    def __init__(self, s, delim="\t\r\n "):
        self.re = re.compile('[%s]+|$' % delim) # approximately right
        self.s = s
        m = self.re.match(self.s)
        if m:
            self.idx = m.end()
        else:
            self.idx = 0
    def __iter__(self): return self
    def next(self):
        old_idx = self.idx
        if old_idx >= len(self.s): raise StopIteration
        m = self.re.search(self.s, old_idx)
        self.idx = m.end()
        return self.s[old_idx:m.start()]

Jeff




From exarkun at intarweb.us  Sun Apr 20 20:55:20 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Sun, 20 Apr 2003 20:55:20 -0400
Subject: Mail and text extraction
In-Reply-To: <20030420234554.GA1084@mephisto.ghaering.test>
References: <pan.2003.04.21.01.19.48.976609.3145@mindlessSPAM.com> <20030420234554.GA1084@mephisto.ghaering.test>
Message-ID: <20030421005520.GA9490@meson.dyndns.org>

On Mon, Apr 21, 2003 at 01:45:55AM +0200, Gerhard H?ring wrote:
> [snip]
> 
> Perhaps you could use IMAP instead? I believe that with IMAP it is
> possible to download parts of email messages (MIME parts), but even here
> I doubt it's easily possible to cleanly download only a few lines of a
> MIME part.

  IMAP4 lets you download byte ranges of specific MIME parts.  So if the
structure is *very* well known, it may be possible.

  Jp

-- 
Examinations are formidable even to the best prepared, for
even the greatest fool may ask more the the wisest man can answer.
                -- C.C. Colton
-- 
 up 31 days, 21:02, 6 users, load average: 0.22, 0.15, 0.10




From zephyr01 at alltel.net  Thu Apr 24 23:43:08 2003
From: zephyr01 at alltel.net (Hans Nowak)
Date: Thu, 24 Apr 2003 23:43:08 -0400
Subject: Vim Newbie =?windows-1252?Q?=96_Python_Edit/Compile/Run_?=
 =?windows-1252?Q?cycle_and_integrating_py_help_questions?=
In-Reply-To: <3EA7EBD3.3010605@ghaering.de>
References: <a791ab19.0304240532.5a025b81@posting.google.com> <3EA7EBD3.3010605@ghaering.de>
Message-ID: <b8aad5$o9b$1@main.gmane.org>

Gerhard Haering wrote:

> a) http://vim.org/ - "Scripts" and "Tips" - you'll also find 
> Python-related stuff there.
> 
> b) The vim mailing list (also available via the news.gmane.org newsserver).
> 
>>  I'm really enjoying the power of Vim!!
> 
> I like vim a lot, too :-)

A not entirely unrelated followup...  (I searched the vim site in vain, and 
before I ask on the vim mailing list or elsewhere, I figured I'd try c.l.py 
first, since I'm more comfortable here. :)

I have been trying to edit Python source code with vim on Mac OS X.  Most of 
the settings in my .vimrc file (copied from Windows) work just fine, except for 
one.  I use Tab to indent and Shift-Tab to dedent (a line, not a block).  Tab 
behaves as usual, but Shift-Tab doesn't seem to work; it just behaves like I 
pressed Tab.

The line I use in my Windows _vimrc to achieve this looks like this:

   imap <S-Tab> ^D

As said, this doesn't work in the vim for Mac OS X.  Using Ctrl-V to get a 
representation of the key doesn't work at all (you know, you press Ctrl-V and 
then F1 or something and you get '<F1>' back).  Typing <S-Tab> verbatim doesn't 
work either.  I also used other variants of *map, to no avail.

Anybody knows how to solve this?  I am using the vim for Jaguar that can be 
found here: http://macvim.swdev.org/OSX/.

TIA,






From duncan at NOSPAMrcp.co.uk  Wed Apr  9 04:38:40 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Wed, 9 Apr 2003 08:38:40 +0000 (UTC)
Subject: String splitting question
References: <9afd3f36.0304082245.33670341@posting.google.com> <b70icc$9n0ff$1@ID-99293.news.dfncis.de> <LqQka.8341$y3.1005660@news010.worldonline.dk>
Message-ID: <Xns93586089ECB1Fduncanrcpcouk@127.0.0.1>

Max M <maxm at mxm.dk> wrote in 
news:LqQka.8341$y3.1005660 at news010.worldonline.dk:

>> out = []
>> for i in 'abcd/df/a/iiwk/abcdefghijkl/b/c'.split('/'):
>>     out.append( (len(i), i) )
>> print max(out)[1]
> 
> Or the ultra terse version, which basically does the same thing:
> 
> s = 'abcd/df/a/iiwk/abcdefghijkl/b/c'
> print max([(len(sub), sub) for sub in s.split('/')])[-1]

It might be worth noting that if there are multiple strings of the same 
longest length, these solutions both give you the last one.
The variant below will give you a list of all the longest strings from 
which you can select first, last or any other combination.

>>> s = 'abcd/df/a/iiwk/ijkl/b/c'
>>> print max([(len(sub), sub) for sub in s.split('/')])[-1]
ijkl
>>> elements = s.split('/')
>>> length = max([len(sub) for sub in elements])
>>> print [ sub for sub in elements if len(sub) == length ]
['abcd', 'iiwk', 'ijkl']
>>> 

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From maxm at mxm.dk  Fri Apr  4 00:56:12 2003
From: maxm at mxm.dk (Max M)
Date: Fri, 04 Apr 2003 07:56:12 +0200
Subject: mp3 or wave  ---> midi , in python , how?
In-Reply-To: <b6isqq$23hs$1@mail.cn99.com>
References: <b6isqq$23hs$1@mail.cn99.com>
Message-ID: <s59ja.2658$y3.447971@news010.worldonline.dk>

Frank Zheng wrote:
> Hi, All:
> 
>    Who can tell me how to do this in python ,are there any modules existing
> can help me?

This is far from trivial, and there are no modules in Python which does 
this.

You should check out "Melodyne" which can do it fairly well on 
monophonic instruments. Unfortunately it is rather expensive!

http://www.celemony.com/melodyne/

On polyphonic instruments it cannot be done yet.

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From andrew at acooke.org  Wed Apr 23 10:39:45 2003
From: andrew at acooke.org (andrew cooke)
Date: Wed, 23 Apr 2003 10:39:45 -0400 (CLT)
Subject: Data-driven testing
In-Reply-To: <b867d7$f53$1@panix2.panix.com>
References: <b867d7$f53$1@panix2.panix.com>
Message-ID: <56223.127.0.0.1.1051108785.squirrel@127.0.0.1>

this isn't an exact solution, but maybe it will help you.  i had a similar
problem while writing a parser.  in that case i wanted to test parse
trees, but didn't want to specify the whole tree by hand for each test
case.  what i did was develop a simple specification "language" that would
check against only the sub-tree that i was interested in, allowing other
sub-trees to be excluded.

for example:

    compare("p1|p2",
            "<ASTPatternNT>"+
            "<ASTLocationPathPatternNT>...</ASTLocationPathPatternNT>"+
            "<ASTLocationPathPatternNT>...</ASTLocationPathPatternNT>"+
            "</ASTPatternNT>",
            COMPLETE);

this specifies that the description is for the whole tree (COMPLETE), but
that sub-nodes of ASTLocationPathPatternNT are not checked ("p1|p2" is the
xpath statement that needs to be parsed).

or:

    compare("id(' a'  )",
            "<ASTIdT getToken=' a'/>",
            FRAGMENT);

which specifies only that an Id token exists with the correct name
somewhere in the tree (i even checked my whitespace handling ;o).

andrew

ps this code is all in java, i'm afraid - it can be downloaded from my web
site (see sig)


aahz at pythoncraft.com said:

> As I just mentioned in another post, I'm writing an ASCII database for
> my address book.  Actually, the main reason for writing it was because
> I've been maintaining multiple lists for various kinds of party invites,
> and the update problem was killing me.
>
> While writing it, it occurred to me that I *ought* to be using unit
> tests to make sure that it was working correctly, but two things stopped
> me.  Secondly, I was in a hurry.
>
> But firstly, I ran into the same problem I had when writing the BCD
> module: how the heck does one write tests when the results are
> hand-calculated?  There were several times when I wrote what I thought
> were correct tests for the BCD module, and my code was right and my test
> was wrong.
>
> In the case of my ASCII database, I'm going to need a big matrix to
> properly test the query language.  How do I build that matrix?


-- 
http://www.acooke.org/andrew




From maxume at yahoo.com  Thu Apr  3 15:20:41 2003
From: maxume at yahoo.com (Max Erickson)
Date: Thu, 03 Apr 2003 20:20:41 -0000
Subject: wxSplitterWindow to fill up its avalible space
References: <6I%ia.3871$Du.12407@newsc.telia.net>
Message-ID: <Xns93529C2F73807notrealmail@216.168.3.44>

"Tim Gahnstr?m /Bladerman" <tim at bladerman.com> wrote in
news:6I%ia.3871$Du.12407 at newsc.telia.net: 

> Hi
> I have put a wxSplitterWindow in a wxBoxSizer together with a panel.
> My problem is that I must set the splitterWindowsize by hand when I
> create it
> 
>     wxSplitterWindow(self, -1, size=(300,300))
> 
> Otherwise it gets no size at all. If I just put two panels there



Are you adding any children to the splitter? It will take sizing cues from 
them...



From fperez528 at yahoo.com  Fri Apr 18 17:40:07 2003
From: fperez528 at yahoo.com (Fernando Perez)
Date: Fri, 18 Apr 2003 15:40:07 -0600
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com>
Message-ID: <b7prbn$fbi$1@peabody.colorado.edu>

Guy Middleton wrote:

> What is the most Pythonic way to sum the n-th element of a list of lists
> (or of tuples)?
> 
> The obvious (to me) way is to use reduce, a list comprehension, and a
> lambda:
> 
> $ python
> Python 2.2.1 (#1, Jul  4 2002, 15:23:46)
> [GCC 2.95.3 20010315 (release) [FreeBSD]] on freebsd4
> Type "help", "copyright", "credits" or "license" for more information.
> >>> x = (("a", 1), ("b", 2), ("c", 3))
> >>> # want to sum 1+2+3 = 6
> ...
> >>> reduce(lambda a, b: a+b, [y[1] for y in x])
> 6

People often don't know (or forget) that Numeric is not only useful for
numerical work, but also because it has the cleanest possible syntax for
handling multi-dimensional arrays (even heterogeneous ones):

In [16]: x=array((("a", 1), ("b", 2), ("c", 3)),typecode='O')

In [17]: sum(x[:,1])
Out[17]: 6


Best,

f.



From max at alcyone.com  Tue Apr 15 00:38:26 2003
From: max at alcyone.com (Erik Max Francis)
Date: Mon, 14 Apr 2003 21:38:26 -0700
Subject: Built-ins vs. modules
References: <mailman.1050378626.14581.python-list@python.org>
Message-ID: <3E9B8CC2.5E775AB1@alcyone.com>

Paul Berkowitz wrote:

> Why import the string module if you can do the same things anyway? The
> page
> of built in String Methods bears an uncanny resemblance to the page on
> the
> string module. Is there a historical reason? Are the built in type
> methods
> quite recent?

The functions in the string module used to be the only way to do this. 
Starting with 2.0, string methods meant that for fiddling with strings,
you need not import the string module.

There are still a few things in the string module that are unavailable
via string methods; e.g., string.lowercase and the like.  Though there
was recent discussion about incorporating these into the string class in
some manner, which seems to be receive favorable responses (though I
don't know what its status is).

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ They have rights who dare defend them.
\__/ Roger Baldwin
    EmPy / http://www.alcyone.com/pyos/empy/
 A templating system for Python.



From theller at python.net  Wed Apr 16 12:05:53 2003
From: theller at python.net (Thomas Heller)
Date: 16 Apr 2003 18:05:53 +0200
Subject: A better unittest
References: <3ckier9a.fsf@python.net> <v9qot54nmk87cb@news.supernews.com>
Message-ID: <u1cyqvfi.fsf@python.net>

"John Roth" <johnroth at ameritech.net> writes:

> "Thomas Heller" <theller at python.net> wrote in message
> news:3ckier9a.fsf at python.net...
> > I'm moving more and more to TDD (test driven development [1]), which
> > makes tests an integral part of the development process.  Failing >
> > tests are ocurring very often with this development style, because
> > you write the tests first and then implement the functionality.
> [snip]
> 
> > What do people think? Is this a useful change, and should I submit a
> > patch?
> 
> I don't think I'd find it useful for two reasons. One is that it doesn't
> get rid of enough clutter; removing the stack trace cuts the output
> down by maybe 50%, and leaves in a good deal of poorly formatted
> visual clutter.
> 
Hm, the latest version of the patch which I posted typically prints this:

======================================================================
FAIL: test_failUnlessEqual (__main__.FailingTests)
----------------------------------------------------------------------
TestFailed: 0 != 1
  File "xunit.py", line 12, in test_failUnlessEqual
    self.failUnlessEqual(self.a, self.b)

Is this poorly formatted? I won't argue about this.
Anyway, this is IMO the information I want to have:
filename, line number, source code line, and the actual#
values which made the test fail.

> The other reason is that it's really hard to figure out when the error
> is expected so the stack trace is redundant, and when it's not expected
> so that the stack trace includes useful information.

It doesn't look too much to a traceback anymore.

> 
> In any case, I really only expect *one* test to fail at any point
> in time, and I suppose that I'm of the "more information is better"
> school of thought. If more than one test fails, I suspect that I've
> done something really stupid on the last change.

This is irrelevant. I did post an example with multiple failures/errors
to show the output for different cases.

Concerning the 'more is better': If the test crashes (raises an
unexpected exception), the full traceback is printed.
If the test fails because a condition is not fulfilled, I'm *not*
interested to see how unittest raises this error.

Thomas



From max at alcyone.com  Sun Apr  6 22:47:52 2003
From: max at alcyone.com (Erik Max Francis)
Date: Sun, 06 Apr 2003 19:47:52 -0700
Subject: Cross-reference 'import' in a class hierarchy
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake>
Message-ID: <3E90E6D8.34A507BA@alcyone.com>

Ben Finney wrote:

> It's not.  Even if I make it so on my computer, it's not going to be
> so
> on someone else's computer when they download my package tree.

If it's important that the project be a real Python package, then make
that an installation requirement.  Provide instructions in the form of a
README or make a distutils setup.py script that people can use to
install your software.

Make sure you put __init__.py files (even if blank) at the root of your
subdirectories to make them true packages.

-- 
 Erik Max Francis / max at alcyone.com / http://www.alcyone.com/max/
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ Sit loosely in the saddle of life.
\__/ Robert Louis Stevenson
    Maths reference / http://www.alcyone.com/max/reference/maths/
 A mathematics reference.



From roy at unice.fr  Fri Apr 18 04:14:57 2003
From: roy at unice.fr (Jean-Paul Roy)
Date: Fri, 18 Apr 2003 10:14:57 +0200
Subject: Teaching : Python, Scheme, Java...
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <3E9C9A91.8090209@removethis.free.fr> <m27k9vfcz3.fsf@mycroft.actrix.gen.nz> <3E9DE024.3020807@removethis.free.fr> <roy-8E37D1.09554617042003@malibu.unice.fr> <mailman.1050609281.18019.python-list@python.org>
Message-ID: <roy-08BA75.10145718042003@malibu.unice.fr>

In article <mailman.1050609281.18019.python-list at python.org>,
 Ian Bicking <ianb at colorstudy.com> wrote:

> On Thu, 2003-04-17 at 02:55, Jean-Paul Roy wrote:
> > - Python is bad : tail recursion is not iterative (also astonished at 
> > that, I can understand with an interpreter, but compiler ?)
> 
> Some would argue that tail recursion is not good for teaching and
> understanding, because you loose the call stack.  In the likely event of
> a traceback, it can be very confusing if you don't fully indicate how
> you got there, or if points upon the path were lost in an optimization.

That's a point, but in that case it would suffice to enveloppe the call 
by an identity function...
Let me cite the Normalization Report on Scheme:

"Programming languages should be designed not by piling feature on top 
of feature, but by removing the weaknesses and restrictions that make 
additional features appear necessary."

And so the wile or for... loops could be defined in Python as in Scheme 
with the tail recursion optimzation.
In the same spirit as Paul Graham's invited talk at PyCon'2003.

My teaching experience is that students knowing loops through the while 
construct actually *understand* the concept of iteration when they see 
that it is modelized by a tail recursive function which transforms a 
"vector of loop variables" (a state vector). Then, it is only a GOTO 
which passes arguments. I'm really happy to teach that way.

     -jpr



From news at snakefarm.org  Wed Apr  9 18:36:47 2003
From: news at snakefarm.org (Carsten Gaebler)
Date: 9 Apr 2003 22:36:47 GMT
Subject: Wraparound problems with time.clock() on Linux
References: <b722d0$bau$1@peabody.colorado.edu> <b722km$bau$2@peabody.colorado.edu>
Message-ID: <slrnb9983v.vss.news@cipw08.physik.uni-bonn.de>

In article <b722km$bau$2 at peabody.colorado.edu>, Fernando Perez wrote:
> Sorry, scratch that.  I went looking further in the docs, and it seems that
> time.clock() calls the system's clock(). From man clock:
> 
>        Note  that  the  time  can  wrap  around.   On  a  32bit  system 
> where
>        CLOCKS_PER_SEC  equals 1000000 this function will return the same
> value
>        approximately every 72 minutes.
> 
> 

> Anyone know of a reliable way to time long-running codes?

Without having tested it for more than 72 minutes, I'd try resource.getrusage().

cg.



From ckea25d02 at sneakemail.com  Fri Apr 11 02:11:58 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 11 Apr 2003 18:11:58 +1200
Subject: Import 'filename'? (was Re: Cross-reference 'import' in a classhierarchy)
In-Reply-To: <slrnb9aopc.ln1.marcus@infa.abo.fi>
References: <slrnb91k1d.og.bignose-hates-spam@rose.localdomain.fake> <3E90D2DB.1A799612@alcyone.com> <slrnb91lge.og.bignose-hates-spam@rose.localdomain.fake> <7h3llymio0q.fsf@pc150.maths.bris.ac.uk> <slrnb94509.p4.bignose-hates-spam@rose.localdomain.fake> <b6tg0i$8t3kr$1@ID-169208.news.dfncis.de> <slrnb94lqs.p4.bignose-hates-spam@rose.localdomain.fake> <3e932e9f$1_2@corp-news.newsgroups.com> <b6vri8$9ob33$1@ID-169208.news.dfncis.de> <3E9428A8.9090005@mitretek.org> <b72kcp$aejp6$1@ID-169208.news.dfncis.de> <slrnb9aopc.ln1.marcus@infa.abo.fi>
Message-ID: <b75log$bcvga$1@ID-169208.news.dfncis.de>

Marcus Alanen wrote:
> I'm not sure what your "good enough" algorithm actually does.
> Consider also the fact that e.g. on Linux you can mount the same
> filesystem several times, in several places.

Sure it's *possible* to do that, but I think if anyone
sets their system up in such a screwy way and it causes
them problems with importing their Python modules, they
deserve what they get.

> Personally I'd just like a simple stat() for each (non-symlink) file
> and remembering their st_dev and st_ino values. I don't know how this
> would be emulated on other operating systems, though.

That's the rub -- st_dev and st_ino are extremely Unix-
specific notions.

On classic MacOS you could use (vRefNum, fileId), I
suppose. Windows, I have no idea.

But at least they all have something akin to a symbolic
link.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From juneaftn at REMOVETHIShanmail.net  Wed Apr 23 11:35:10 2003
From: juneaftn at REMOVETHIShanmail.net (Changjune Kim)
Date: Thu, 24 Apr 2003 00:35:10 +0900
Subject: Data-driven testing
References: <b867d7$f53$1@panix2.panix.com>
Message-ID: <b86bqd$iu4$1@news.hananet.net>

Hi, Aahz.

"Aahz" <aahz at pythoncraft.com> wrote in message
news:b867d7$f53$1 at panix2.panix.com...
> As I just mentioned in another post, I'm writing an ASCII database for
> my address book.  Actually, the main reason for writing it was because
> I've been maintaining multiple lists for various kinds of party invites,
> and the update problem was killing me.
>
> While writing it, it occurred to me that I *ought* to be using unit
> tests to make sure that it was working correctly, but two things stopped
> me.  Secondly, I was in a hurry.

On the second:

KentBeck : "If you write automated tests, you'll get more done every day."
Johnny : "But we have too much to do. We don't have time to write tests."
KentBeck : "No, no. You didn't understand. You'll get _more_ done, not
less."
Johnny : "Yes, but we don't have time for testing."

>
> But firstly, I ran into the same problem I had when writing the BCD
> module: how the heck does one write tests when the results are
> hand-calculated?  There were several times when I wrote what I thought
> were correct tests for the BCD module, and my code was right and my test
> was wrong.
>
> In the case of my ASCII database, I'm going to need a big matrix to
> properly test the query language.  How do I build that matrix?

You make tests in order to increase confidence. If you lose any confidence
writing tests, find another way. One option is to divide and conquer. Don't
try to compare the (initial input, final output) pairs at the first step.
Start from the parts; it's called _unit_ testing.

For example, if your program caculates a polynomial expression like
x^2+2x-1 and you are afraid your hand-calculated value is wrong, test
squaring and adding, substracting first, for instace. Now you have one step
further. Try unittesting step by step. If you have some time :) have a look
at TDDBE by Kent Beck.


With warmest regards,

June Kim





From sam at insidegabriel.com  Thu Apr 24 18:07:33 2003
From: sam at insidegabriel.com (Sam)
Date: 24 Apr 2003 15:07:33 -0700
Subject: boost.python - Virtual Functions with Default Implementations
Message-ID: <3c539fdc.0304241407.54dac3dc@posting.google.com>

I'm trying to get virtual functions with default implementations to be
exposed to Python.  When I follow the examples given on the boost.org
site...
http://www.boost.org/libs/python/doc/tutorial/doc/virtual_functions_with_default_implementations.html

...I get the following error:
c:\projects\publiclibs\boost\boost_1_29_0\boost\python\detail\def_helper.hpp(131)
: error C2440: 'return' : cannot convert from 'int (__thiscall
giPython::BaseWrap::*const )(void)' to 'const char *'
        There is no context in which this conversion is possible
        c:\projects\publiclibs\boost\boost_1_29_0\boost\python\detail\def_helper.hpp(130)
: while compiling class-template member function 'const char
*__thiscall boost::python::detail::def_helper<int (__thiscall
giPython::BaseWrap::*)(void),struct
boost::python::detail::not_specified,struct
boost::python::detail::not_specified>::doc(void) const'

Anyone have any ideas?  Here is my code:

using namespace boost::python;
namespace giPython
{
    struct Base
    {
        virtual int f() { return 0; }
    };

    struct BaseWrap : Base
    {
        BaseWrap(PyObject* self_)
            : self(self_) {}
        int f() { return call_method<int>(self, "f"); }
        int default_f() { return Base::f(); }
        PyObject* self;
    };

void SetupPythonWrappers (scope& this_module)
{
    class_<Base, BaseWrap>("Base")
    .def("f", &Base::f, &BaseWrap::default_f);
}

} // namespace giPython



From peter at engcorp.com  Sun Apr 20 21:52:22 2003
From: peter at engcorp.com (Peter Hansen)
Date: Sun, 20 Apr 2003 21:52:22 -0400
Subject: importing modules with the same name but from different directories
References: <b7vhad$1t6$1@news.tu-darmstadt.de>
Message-ID: <3EA34ED6.F0361E7D@engcorp.com>

Mirko Koenig wrote:
> 
> In every modul directory is a diretory for the language files. These
> files are used to display the modules label in different languages.
> eg.
> 
> .../modules/module1/language/german.py
> .../modules/module1/language/english.py
> .../modules/module2/language/german.py
> .../modules/module2/language/english.py
> 
> My problem is:
> If is start modul1 from within its modul directory ( .../modules/modul1 )
> And modul1 calls modul2, then modul2 doesn't load its own language file.
> Instead it loads the langugae file of modul1.

Although it's not 100% clear to me what you are trying to do, I'd
like to suggest a couple of alternatives to what you seem to be
trying, which might be simpler.

Option 1 is to just use the execfile() builtin function to execute
the necessary files, rather than trying to use "import".

Option 2 is to rename the files to something other than .py and
load them as regular files, then decode them with exec().

Option 3 is to find one of the standard tools for supporting
internationalization and use a standard approach instead of 
rolling your own like this.  Check the "gettext" standard
library module for details.

-Peter



From pan-newsreader at thomas-guettler.de  Mon Apr  7 05:53:58 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Mon, 07 Apr 2003 11:53:58 +0200
Subject: COM and MS Word
References: <2eb85cfb.0304070115.6aecb644@posting.google.com>
Message-ID: <pan.2003.04.07.11.53.57.522248.788@thomas-guettler.de>

On Mon, 07 Apr 2003 11:15:31 +0200, CTony wrote:

> It appears that some data files I need are in WordPerfect (.dat) format,
> and I need to get them into Python for processing and later entry into a
> database. I haven't been able to find a way of interfacing to
> WordPerfect directly, but MS Word 2000 can read these files...

Maybe you can solve your problem by avoiding COM. You can create HTML
from Word-Documents with wvWare. I think you can handle this easier than
COM. Our save the word document as XML and parse this.

Sorry, this was no answer to your question, but maybe it helps, too.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From news at titanic.co.uk  Sat Apr  5 14:13:45 2003
From: news at titanic.co.uk (Simon Faulkner)
Date: Sat, 05 Apr 2003 20:13:45 +0100
Subject: List of Numbers
Message-ID: <1lau8vcctfn4g2pi9k5df3km7qgg5r72sv@4ax.com>

I have a list of about 5000 numbers in a text file - up to 14 digits
each.

I need to check for duplicates.

What would people suggest as a good method?

Simon



From alloydflanagan at attbi.com  Mon Apr 21 09:21:49 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 21 Apr 2003 06:21:49 -0700
Subject: Why functional Python matters
References: <slrnb9on0q.sho.ramen@lackingtalent.com> <yfs7k9vcu2p.fsf@black132.ex.ac.uk> <ccc7084.0304160450.680bf7ef@posting.google.com> <RoacnSAeGeS6CACjXTWcoQ@comcast.com> <a8b7f07a.0304181341.5e4b69f3@posting.google.com> <ilLoa.1695$Qg2.126929@news20.bellglobal.com>
Message-ID: <a8b7f07a.0304210521.2d87cab8@posting.google.com>

Warren Postma <warren.postma at sympatico.ca> wrote in message news:<ilLoa.1695$Qg2.126929 at news20.bellglobal.com>...
> > I'll second that.  Maybe we could get func added as a synonym for lambda, at least.
> 
> 
> Maybe we should translate all language keywords into some other popular 
> languages than english.  In which case, Lambda can remain as is, for our 
> greek friends, and the rest of the keywords can change to greek words.
> 
> My-big-fat-greek-python-ly-yr's
> 
> Warren

What's the greek word for "python"?



From imbosol-1049309043 at aerojockey.com  Wed Apr  2 14:03:26 2003
From: imbosol-1049309043 at aerojockey.com (Carl Banks)
Date: Wed, 02 Apr 2003 19:03:26 GMT
Subject: Best practice for object attributes?
References: <b6ekqi$7d7$1@nntp0.reith.bbc.co.uk> <3E8AEC44.2090909@removethis.free.fr> <b6eusn$avn$1@nntp0.reith.bbc.co.uk>
Message-ID: <2uGia.77$nG6.43@nwrdny01.gnilink.net>

Michael Sparks wrote:
> laotseu wrote:
> 
>> One outstanding advantage of the second 'idiom' is that it's the right
>> way to define *instance* attributes !-) (which is probably what your
>> looking for)
> 
> Given the first idiom seems to create instance attributes by cloning the
> current state of the class attributes, I'm finding it useful is where
> certain instance attributes are unlikely to differ much between
> instances, or changes infrequently related to object creation. 

It doesn't.


> For example: if you're creating large numbers of objects which need a
> ballpark timestamp, you could timestamp the class periodically, and
> reduce your number of calls to gettimeofday.

Are you sure?  Look:

>>> class A:
...     time = 'today'
... 
>>> B = A()
>>> B.time
'today'
>>> A.time = 'tomorrow'
>>> C = A()
>>> C.time
'tomorrow'

So far so good.  But look at what happened to B:

>>> B.time
'tomorrow'



> Another example is in a custom logging class - it has default logging
> levels for most objects defined by the class attributes, but can be
> raised/lowered on a per instance basis, or globally for all new objects
> by changing the class attributes.
> 
>> With the first form you get *class* attributes.  Make 2 instances of
>> Shrubbery1, change the name of first one to 'Brian', print the name of
>> the second one, and then try to understand what is the purpose of the
>> 'self' argument in methods.
> 
> I'm aware of the meaning of self :) and know that the first form gives
> you class attributes. The reason I was asking is because I'd noticed
> the following interesting behaviour, which if _intentional_ is useful.
> 
>>>> a=shrubbery1()
>>>> b=shrubbery1()
>>>> a.name,b.name,shrubbery1.name
> (None, None, None)                  # a.name, b.name exist - presumed
>                                    # references to shrubbery1.name
>>>> a.name="hello"
>>>> a.name,b.name,shrubbery1.name
> ('hello', None, None)               # Bzzzt! a.name seems to be
>                                    # independent
>>>> b.name="brian"
>>>> a.name,b.name,shrubbery1.name
> ('hello', 'brian', None)            # Hmm. So is b.name
>>>> shrubbery1.name="Surprise"
>>>> a.name,b.name,shrubbery1.name
> ('hello', 'brian', 'Surprise')      # Changing class leaves existing
>                                    # instances unaffected.

Try changing the class *before* changing the instances.  You will see
that it does affect the instances.


>>>> a=shrubbery1()                  # Get 2 fresh copies.
>>>> b=shrubbery1()
>>>> a.name,b.name,shrubbery1.name
> ('Surprise', 'Surprise', 'Surprise') # a.name, b.name cloned from
>                                     # shrubbery1.name
> 
> As far as I can tell the behaviour seems to be that you automatically
> gain instance attributes that clone the state of the class attributes
> at object creation time, which are then independent and can be modified
> via the usual self.<attr> syntax later on. (They cannot however be
> modified via the self.__dict__[<attr>] syntax since they're not stored
> there)

You're missing an important point however: when you "modify" a class
attribute using self, what you're really doing it adding an item to
the instance dict.  Before you modified the attribute, there was no
such attribute in the instance dict



> For reference, this behaviour holds with the following python versions:
>   * python 2.3 alpha 2
>   * python 2.2.1
>   * python 2.2.2 stackless
> 
> Given it seems consistent across versions, and seems useful (changing
> default values for later objects just by updating a class attribute -
> eg an external timestamper perhaps) I was wondering if there's a good
> reason not to use this.


I highly recommend you initialize all instance variables in __init__,
and never use class variables as "default instance varibles."  It can
lead to all kinds of subtle bugs.  Even if you fully understood the
dynamics (and it seems from this post you don't), another person
working on the code might not.  It's also possible that you might want
to do something later that depends on instance variables being in the
instance's dict.  (Maybe you want to pickle it, and expect it to have
the same value when unpickled.)

Generally, you should only use class variables when you want *all*
(not most, all) instances to have the same attribute, and you want
*all* instances to reflect any changes to it.


-- 
CARL BANKS



From rainer.klueting at epost.de  Thu Apr 10 11:40:45 2003
From: rainer.klueting at epost.de (Rainer Klueting)
Date: Thu, 10 Apr 2003 17:40:45 +0200
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com> <urnka.4658$Du.16154@newsc.telia.net> <b68f71a7.0304080323.75b5ed1a@posting.google.com> <2me59vcigrdkoeapbtb267i46jf07kr9gr@4ax.com> <nqzka.4693$Du.16451@newsc.telia.net>
Message-ID: <op2b9v0lbtjubvo0itmo2iq53kqrk3kg19@4ax.com>

Hi Tim and all of you helpful people,

I will have another look on tkTable. I got an information by Bob Terek
by mail --I don't find this message here --about a Python-wrapper for
the latest version of tkTable. According to Bob it can be found at
http://tktable.sourceforge.net/. (I have not yet checked it.)

My search has come to an interesting result: I found a book from
German author Michael Lauer of university of Frankfurt, Germany,
called "Python und GUI-Toolkits". (I don't know whether a translation
to English is available.) This Guy has done a remarkable job: He
presents Tkinter, wxWindow, Qt, GTK+ and, in short form, several other
GUI-toolkits by completely writing a non trivial application with each
of these toolkits, explaining almost every line of code.

And, you won't believe it: The application, a contact manager, uses a
table. For his Tkinter implementation he had to write an appropriate
table-class in pure Tkinter.

I'm just deep down in this book, not only because of the table, of
course! 

See you on re-surfacing! ;-)

Rainer Klueting, Stuttgart, Germany



From robin at jessikat.fsnet.co.uk  Thu Apr 17 04:10:11 2003
From: robin at jessikat.fsnet.co.uk (Robin Becker)
Date: Thu, 17 Apr 2003 09:10:11 +0100
Subject: How many of you are Extreme Programmers?
References: <mailman.1050512881.28736.python-list@python.org>
Message-ID: <XiCE$IAjFmn+Ew43@jessikat.fsnet.co.uk>

I get the impression that XP is intended as a design none/test all
process where the code some grows organically into the requirement
through bursts of coding and testing. It even seems as though the
testing is primary.

This sounds remarkably similar to what we commonly call genetic
optimisation or evolution strategy selection. There a fitness criterion
is applied to randomly selected elements of a pool of solutions, with
the fittest being combined/mutated etc and flung back into the next
testing round. In GP the expertise is almost always in the fitness
criterion and in the representation of solution candidates.

Can XP be reduced to choosing candidates from the pool of programs that
best satisfy a set of tests? If so experience from GP suggests that the
tests require care to avoid common pitfalls. As an example a program
might complete all the tests perfectly, but fail on all other instances
of data (this commonly happens to overspecialised solutions).

Do XPers combine differing solutions etc etc? 
-- 
Robin Becker



From grisha at ispol.com  Tue Apr  1 14:57:36 2003
From: grisha at ispol.com (Gregory (Grisha) Trubetskoy)
Date: Tue, 01 Apr 2003 19:57:36 GMT
Subject: Anything like Mailman's LockFile.py in std lib?
Message-ID: <20030401144641.L28887-100000@eden.ispol.com>

Is there anything being planned for the standard lib that provides locking
a la LockFile.py in Mailman?

I find it extremely useful, but just recently came across a project where
it can't be used becuase it's GNU licensed - thus the question.

Grisha




From do-not-spam-ben.hutchings at businesswebsoftware.com  Tue Apr  8 12:07:56 2003
From: do-not-spam-ben.hutchings at businesswebsoftware.com (Ben Hutchings)
Date: 8 Apr 2003 16:07:56 GMT
Subject: executable won't run on another machine
References: <v90mh6ppf9bc09@corp.supernews.com>
Message-ID: <slrnb95sur.t8.do-not-spam-ben.hutchings@tin.bwsint.com>

In article <v90mh6ppf9bc09 at corp.supernews.com>, Lance wrote:
> Hi All,
> 
> I created an executable using McMillan's Installer. It runs great on the
> development machine. I copied it to another machine and it doesn't run. I
> type the executable name from a command prompt, hit return, and get another
> command prompt.
<snip>

Did you put './' in front of the executable name?  Most Unix systems are
configured not to include the current directory in the search path.  If
you forgot to do that then you could be running some other program that is
in the search path.



From pacquets at newsguy.com  Tue Apr 29 13:17:12 2003
From: pacquets at newsguy.com (pacquets at newsguy.com)
Date: 29 Apr 2003 10:17:12 -0700
Subject: Tiny (with respect to memory) Python
Message-ID: <b8mc2o0c8m@drn.newsguy.com>

I can't find any recent postings about this, so I hope someone can help me out
here:

I ported 2.0 to our weird little real-time OS awhile ago, and I found the memory
consumption pretty unacceptable for an embedded device (for our low-end models,
anyway).  Some stats:

 * enter the interpreter, add two numbers, exit.  High-water mark:  153K

 * enter again, do the following & exit:

>>> import os
>>> os.listdir('.') 

high-water mark:  850K

This is not to throw stones;  I think Python's a wonderful language and I'd like
all our users to have it.  But I really need to bring that number down
drastically for the low-end products.  I accept that that might require *really*
draconian measures.  What's the best way to do that?

Some thoughts:

1) there is Pippy, which is now on SourceForge and seems pretty well abandoned. 
I haven't figured out yet what they did to bring the footprint down.

2) Minimal Python looks promising, but it doesn't seem like they're very far
along yet.

3) Question:  Is there some way to get an idea where the memory is going?  I
found this:

http://www.nightmare.com/medusa/memory-leaks.html

the function at the end of that looks sort of promising;  is that the best way?




From staschuk at telusplanet.net  Wed Apr 16 23:01:53 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Wed, 16 Apr 2003 21:01:53 -0600
Subject: Degree symbol (UTF-8 > ASCII)
In-Reply-To: <f7199550.0304161732.48c75c42@posting.google.com>; from pc451@yahoo.com on Wed, Apr 16, 2003 at 06:32:06PM -0700
References: <f7199550.0304160930.6cfeca65@posting.google.com> <3e9daad2$0$49115$e4fe514c@news.xs4all.nl> <f7199550.0304161732.48c75c42@posting.google.com>
Message-ID: <20030416210153.D6846@tibia.amotlpaa.bogus>

Quoth Peter Clark:
> [...] However, when I put it into the code, I get this error:
> 
>     w += [deg.encode('UTF-8') + scale.strip()]
> UnicodeError: ASCII decoding error: ordinal not in range(128)

Aha!  Yes, it's a *decoding* error.  Your 'deg' variable is a
normal string, right?

    >>> '\xb0'.encode('utf-8')
    Traceback (most recent call last):
      File "<stdin>", line 1, in ?
    UnicodeError: ASCII decoding error: ordinal not in range(128)
    >>> '\xb0'.decode('latin-1').encode('utf-8')
    '\xc2\xb0'
    >>> '\xb0'.decode('latin-1') # produces Unicode string
    u'\xb0'
    >>> u'\xb0'.encode('utf-8') # uses Unicode string
    '\xc2\xb0'

See?  For .encode to work, it has to know what characters are in
the string being encoded.  If it's a normal string with bytes
outside of range(128), they're not ASCII, so it quite properly
refuses to guess what they are.  But if you tell it the correct
encoding, it can construct the right sequence of characters (as a
Unicode string), and then encode that as instructed.

>     Since the output is meant to be read to be displayed by a font
> which is in essentially latin-1 encoding, I need to restrict the
> manner in which the degree symbol is displayed to one byte. [...]

So just do something like
    print >>fileobject, chr(176)
    print >>fileobject, u'\N{DEGREE SIGN}'.encode('latin-1')

(Note that if you do want your file to be in ISO-8859-1, the XML
declaration should say that's what it is.)

-- 
Steven Taschuk                                                   w_w
staschuk at telusplanet.net                                      ,-= U
                                                               1 1




From mwm at mired.org  Wed Apr  2 13:18:36 2003
From: mwm at mired.org (Mike Meyer)
Date: Wed, 02 Apr 2003 18:18:36 GMT
Subject: PyQt QTableItem inheritance...
References: <x77kaejf9b.fsf@guru.mired.org> <mailman.1049223123.616.python-list@python.org>
Message-ID: <x7istwiwzm.fsf@guru.mired.org>

Phil Thompson <phil at river-bank.demon.co.uk> writes:

> On Tuesday 01 April 2003 5:31 pm, Mike Meyer wrote:
> > I've already tried posting this to the pyKDE mail list, and gotten no
> > response.
> >
> > I'm trying to create an item in a QTable that uses the
> > QDoubleValidator on the input edit field. The only way I can see to do
> > this is subclass QTableItem, and override createEditor with a method
> > that sets the validator, like so:
[code elided]
> > While I get the output from __init__, I never see the output from
> > createEditor. Any help would be appreciated.
> 
> A small but complete example would help.

The rest of it is biolerplate, but here it is.

        <mike

import sys
from qt import *
from qttable import *

class Float_Field(QTableItem):

    def __init__(self, table, data):
        print "Creating float field."
        QTableItem.__init__(self, table, QTableItem.OnTyping, str(data))

    def createEditor(self):
        print "createEditor"
        editor = QLineEdit(self.text(), self.table().viewport())
        self.connect(editor, SIGNAL("textChanged(QString &)"),
                     self.table(), SLOT("doValueChanged"))
        editor.setValidator(QDoubleValidator(self))
        return editor


app = QApplication(sys.argv)
table = QTable(1, 1)
table.setItem(1, 1, Float_Field(table, "12.34"))
app.setMainWidget(table)
table.show()
app.exec_loop()


-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



From duncan at NOSPAMrcp.co.uk  Mon Apr  7 11:24:08 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Mon, 7 Apr 2003 15:24:08 +0000 (UTC)
Subject: select a class in execution time
References: <mailman.1049724192.1994.python-list@python.org>
Message-ID: <Xns9356A6A539CDduncanrcpcouk@127.0.0.1>

cybersamurai at mac.com wrote in news:mailman.1049724192.1994.python-
list at python.org:

> have some like that on Python?
> 
> Object x = Class.forName( "mypackage.TheClass" ).newInstace();
> (this is a java code)

    def find_class(module, name):
        __import__(module)
        mod = sys.modules[module]
        klass = getattr(mod, name)
        return klass


-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From skip at pobox.com  Mon Apr 14 09:53:39 2003
From: skip at pobox.com (Skip Montanaro)
Date: Mon, 14 Apr 2003 08:53:39 -0500
Subject: an ugly file-reading pattern
In-Reply-To: <E194y4Q-0002W1-00@tswings.swing.cluster>
References: <16024.52407.405254.424039@montanaro.dyndns.org>
        <E194y4Q-0002W1-00@tswings.swing.cluster>
Message-ID: <16026.48483.929843.87511@montanaro.dyndns.org>

    Christian> Maybe you should have written it as :

    Christian>     open = file("somefile")
    Christian>     for line in open:
    Christian>         process(line)

    Christian> <2.2 wink>

Yeah, stupid mistake.  I had part of the OP's original solution (vrbl named
file) and part of mine...

Skip




From peter at engcorp.com  Thu Apr 10 10:02:39 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 10 Apr 2003 10:02:39 -0400
Subject: class too big for one file! How can I split it?
References: <e956b5f1.0304100512.597654d3@posting.google.com>
Message-ID: <3E95797F.3A24944F@engcorp.com>

HHaegele wrote:
> 
> Now, OTLib.py has more then 32000 lines. And that doesn?t work. Python
> 1.5 can?t execute then. 

I presume you've already eliminated the alternative of *not* using
the as-good-as-obsolete Python 1.5?



From none at none.com  Sat Apr 26 05:59:23 2003
From: none at none.com (Iwan van der Kleyn)
Date: Sat, 26 Apr 2003 09:59:23 +0000
Subject: GUI libs
In-Reply-To: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it>
Message-ID: <3EAA587B.5090905@none.com>

Alessio Pace wrote:
> Hi, I'd like to make a GUI for a program written in python, meant to run on
> mac, win and linux: which is the preferable solution? I saw around tkinter,
> wxpython and pyqt....

For me it came down to a choice between Tkinter and WxPython. In the end 
I settled for WxPython.  A rather biased comparrison of Tkinter and 
WxPython you can find at 
http://wiki.wxpython.org/index.cgi/Choosing_20wxPython_20over_20Tkinter

I like the quote "Yes, with wxPython you pay the price up front, in the 
learning curve. With Tk you pay it later, when you're developing 
applications with real interface requirements."

Part of the learning curve of WxPython is the fact that WxPython has no 
documentation worth mentioning. The C++ library on which it runs, 
WxWindows, has terrific documentation but it's near to useless if you do 
not have any C++ experience. So if you have litle experience and want to 
have results fast, then tkinter is a good choice. However, the 
usercommunity around WxPython is very supportive. And Boa Contructor is 
a terrific IDE/Form builder which uses WxPtyhon. With this you can 
create complex forms fast and use the generated code to learn WxPython 
quikcly. http://boa-constructor.sourceforge.net/

Regards,

Iwan




From jmeyer at pdi.com  Tue Apr  8 14:16:46 2003
From: jmeyer at pdi.com (Jim Meyer)
Date: 08 Apr 2003 11:16:46 -0700
Subject: exceptions == errors?
In-Reply-To: <b6v0nn$1h8$1@panix2.panix.com>
References: <P6oka.3208$Bg2.2310@newssvr19.news.prodigy.com>
	 <16018.5597.933725.584314@montanaro.dyndns.org>
	 <mailman.1049817314.14810.python-list@python.org>
	 <b6v0nn$1h8$1@panix2.panix.com>
Message-ID: <1049825806.3175.995.camel@utonium.pdi.com>

On Tue, 2003-04-08 at 10:26, Roy Smith wrote:
> Jim Meyer  <jmeyer at pdi.com> wrote:
> >It's always been my impression that exceptions impose considerable
> >performance overhead and therefore should be used to handle only the
> >truly exceptional cases; for any predictable case, one should perform
> >validation and/or error-handling outside the exception mechanism. 
> >
> >[BTW, I picked up this habit after reading "Effective C++" by Scott
> >Meyers, a book I cannot recommend highly enough. =]
> 
> "Effective C++" is indeed an excellent book for improving your C++
> code.  But why would you think it would teach you anything about
> improving your Python code?  Most of the book is very C++ specific.

I wouldn't think such (and don't =). The two weren't intended to
directly relate; I apologize for causing you to infer otherwise. =]

The first paragraph you quoted relates to benchmarking I'd done some
time ago (mentioned in a portion of the same posting not included in
your quotation of it). The second bit you quoted relates to the bit of
Mark Harrison's original posting (also quoted in my original post)
regarding there being "the same feeling as in C++ that exceptions can be
used for non-error situations".

Sorry for the confusion!

--j
-- 
Jim Meyer, Geek at Large                                  jmeyer at pdi.com





From lucio at movilogic.com  Tue Apr  8 17:15:28 2003
From: lucio at movilogic.com (lucio)
Date: Tue, 08 Apr 2003 18:15:28 -0300
Subject: Tiobe index algorithm now published
In-Reply-To: <20030408194829.GA4589@meson.dyndns.org>
References: <b6v8id$qli$1@reader1.tiscali.nl> <20030408194829.GA4589@meson.dyndns.org>
Message-ID: <3E933BF0.8070204@movilogic.com>

Jp Calderone wrote:

>On Tue, Apr 08, 2003 at 09:37:31PM +0200, GerritM wrote:
>  
>
>>We have had discussions about the "TIOBE Programming Community Index" at
>>http://www.tiobe.com/tpci.htm
>>One of our consistent critics was that the algorithm was not published nor
>>verifiable. Tiobe has now added the (simple) algorithm on this page. They
>>also explain which "programming languages" are monitored.
>>    
>>
>
>  Ooh, a Google search.  How scientific. ;)
>
Comparing the popularity of XML, C and SQL is like... lets see, some 
example without fruits... yes, comparing cars, planets and diseases.

btw: as far as i know HTML does not manipulate data
/Q: What definition of programming languages has been used?/

A: The adopted definition is "any computer language which is either 
interpreted or compiled and is capable of manipulating data". As a 
consequence, HTML and XML are considered programming languages as well. 
ASP is /not/ because it is regarded a technique that makes use of other 
languages such as JavaScript and VBScript.

back to python, now.

Lucio.





From carroll at tjc.com  Sat Apr 12 21:12:06 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Sat, 12 Apr 2003 18:12:06 -0700
Subject: nntplib, huge xover object
References: <ZCidndFDJdNGwBejXTWc3g@giganews.com> <bmqn8vonc4qkmnr2kcl573k4l58c2t70fq@4ax.com> <slrnb8opkc.174.do-not-spam-ben.hutchings@tin.bwsint.com> <443s8v8mehlrhapp0d6af7k0ps22k2tvii@4ax.com> <09is8votcf3116a94euite6vhdfffgefko@4ax.com> <slrnb8ti36.ldu.rmunn@localhost.localdomain> <6j8u8v0hjh3n9dm1jjflu3a970ada6qho4@4ax.com> <slrnb933p7.46e.rmunn@localhost.localdomain> <0au59v0n82d016vs3mai634d4088uf9f0k@4ax.com> <slrnb98hor.8ee.rmunn@localhost.localdomain>
Message-ID: <68eh9v0dl9k9llqjb322li4rv5nq6afur2@4ax.com>

On Wed, 09 Apr 2003 22:08:40 GMT, Robin Munn <rmunn at pobox.com> wrote:

> I hope I've been able to explain properly
>why that's not the Python way.

Yeah, thanks, I actually found this to be a pretty interesting
discussion.

I've submitted my patch, by the way: 720468.  We'll see how it goes.




From mykbourassa at cogeco.ca  Wed Apr  9 13:10:38 2003
From: mykbourassa at cogeco.ca (myk)
Date: Wed, 09 Apr 2003 13:10:38 -0400
Subject: When Python outruns C++ ...
In-Reply-To: <3e912546$0$10390$edfadb0f@dread11.news.tele.dk>
References: <mailman.1048973984.17441.python-list@python.org> <omqha.31133$Jg1.677484@news1.tin.it> <tyfr88nwbx6.fsf_-_@pcepsft001.cern.ch> <db5ia.43$DT2.151963@twister.nyc.rr.com> <yfsvfxzm2n2.fsf@black132.ex.ac.uk> <na6ia.126$n95.214909@twister.nyc.rr.com> <yfs65pxn89e.fsf@black132.ex.ac.uk> <x7u1dgiz1k.fsf@guru.mired.org> <3e8be5d0$0$10380$edfadb0f@dread11.news.tele.dk> <x7el4jibzl.fsf@guru.mired.org> <Kp2ja.2589$y3.426276@news010.worldonline.dk> <x7znn6gkfn.fsf@guru.mired.org> <3e912546$0$10390$edfadb0f@dread11.news.tele.dk>
Message-ID: <swYka.972$Jl.399908@read2.cgocable.net>

Hire more programmers?!?!?!  Mr. Brooks ("The Mythical Man-month" - 
Addison-Wesley) would have you for breakfast :)

Anders J. Munch wrote:

>"Mike Meyer" <mwm at mired.org> wrote:
>  
>
>>>No limit on programmer time?
>>>      
>>>
>>You just have to hire more programmers.
>>    
>>
>
>That only works if your project isn't late <0.7 wink>
>
>- Anders
>
>
>  
>




From aleax at aleax.it  Wed Apr 30 18:47:43 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 30 Apr 2003 22:47:43 GMT
Subject: Question about accessing class-attributes.
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE169@admin56.narex.com> <mailman.1051689777.10470.python-list@python.org> <2259b0e2.0304301148.36cdd823@posting.google.com>
Message-ID: <joYra.44343$K35.1283700@news2.tin.it>

Michele Simionato wrote:
   ...
> class B(object): pass
> class C(B): pass
> c=C()
> super(C,c).__thisclass__() #=> C
> super(C,c).__self_class__() #=> C
> 
> I would like to have an additional attribute (not a method)
> returning the superclass of C, in this case B:

*blink* and if "class D(E, F, G, H, I):", what would "the
superclass" *BE*?

Is what you're looking for something like you could obtain
by, e.g.:

class metasc(type):
    def __new__(meta, name, bases, clasdict):
        if bases: clasdict['__superclass__'] = bases[0]
        else: clasdict['__superclass__'] = None
        return type.__new__(meta, name, bases, clasdict)

???  If so, I fail to see the usefulness.  Otherwise,
could you please explain better?


> Certainly the documentation in
> http://www.python.org/dev/doc/devel/lib/built-in-funcs.html
> 
> has to be changed. It says:
> 
> """
>  super(type[object-or-type])
> 
>       Return the superclass of type. If the second argument is omitted
> the super object returned is unbound. If the second argument is an
> object, isinstance(obj, type) must be true. If the second argument is
> a type, issubclass(type2, type)
> must be true. """
> 
> But it DOES not return the superclass! I admit I myself at the
> beginning was confused.

The documentation is surely imperfect (e.g. by making no reference to
the mro, which is crucial in the case in which the 2nd argument is
there and is an instance of the first one).  But doesn't the second
sentence say explicitly that e.g. super(C) returns a "super object"
that is unbound?  Maybe by "superclass" you mean "base", a common
usage, but then the singular "the superclass" would be absurd.

I'm not sure what the practical use case for the 2nd argument being 
a type is:

>>> class A(object): a = 23
...
>>> class B(object): a = 45
...
>>> class C(A, B): a = 67
...
>>> C.a
67
>>> super(A, C).a
45

but I don't know how one would get, from C, some object whose
attribute named 'a' is worth 23.  So, I guess I just don't 
understand this part of super's behavior.


Alex




From jason at tishler.net  Wed Apr  9 15:04:05 2003
From: jason at tishler.net (Jason Tishler)
Date: Wed, 09 Apr 2003 15:04:05 -0400
Subject: Cygwin Python regression testers wanted
In-Reply-To: <20030403160205.GA2896@tishler.net>
References: <20030403160205.GA2896@tishler.net>
Message-ID: <20030409190404.GC2444@tishler.net>

On Thu, Apr 03, 2003 at 11:02:05AM -0500, Jason Tishler wrote:
> I'm in the process of updating the Cygwin section of the official
> Python README.  Since I only have access to Windows NT 4.0 and 2000, I
> am seeking Cygwin Python regression testers for the following Windows
> platforms:
> 
>     95
>     95 OSR2
>     98
>     98 Second Edition
>     Me
>     XP Home
>     XP Professional

I would like to thank the regression testers that have submitted test
reports for the following platforms:

    95 OSR2
    98
    98 Second Edition
    XP Home
    XP Professional

However, I still need a 95 (not OSR2) and a Me user willing to help.  If
interested, then please see the following:

    http://cygwin.com/ml/cygwin/2003-04/msg00314.html

Additionally, is anyone willing to debug the non-catastrophic failures
on any 9x/Me platform:

    test_dircache
    test_hotshot
    test_mmap

Even running the regression tests with a "-v" option would be helpful.

Of course, if your are willing to debug the catastrophic ones:

    test_fork1
    test_popen2
    test_pty
    test_socket

then I won't try to stop you. :,)

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From dave at boost-consulting.com  Mon Apr 21 12:10:22 2003
From: dave at boost-consulting.com (David Abrahams)
Date: Mon, 21 Apr 2003 12:10:22 -0400
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <60FB8BB7F0EFC7409B75EEEC13E20192022DE025@admin56.narex.com> (Bjorn
 Pettersen's message of "Mon, 21 Apr 2003 09:38:50 -0600")
References: <60FB8BB7F0EFC7409B75EEEC13E20192022DE025@admin56.narex.com>
Message-ID: <844r4rrfv5.fsf@boost-consulting.com>

"Bjorn Pettersen" <BPettersen at NAREX.com> writes:

>> From: David Abrahams [mailto:dave at boost-consulting.com] 
>
>> Also, "stacklessness" is really an implementation detail AFAICT - it
>> just says that all execution context is stored in
>> dynamically-allocated structures instead of using local variables on
>> the 'C' stack (via a recursive-calling implementation).
>> 
>> Have I misunderstood this whole thing?
>
> Which property exactly makes the C stack frames not dynamically
> allocated? (the sequential property of the allocator, or the fixed size
> of the stack itself?)

Sorry; I guess they're dynamically allocated... on the stack.  IOW,
not on the heap and with lifetime directly tied to the location of the
program counter.  That's a more accurate way to say what I meant.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com





From pan-newsreader at thomas-guettler.de  Mon Apr  7 09:10:02 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Mon, 07 Apr 2003 15:10:02 +0200
Subject: how to convert from unicode to another codepage
References: <mailman.1049717651.6046.python-list@python.org>
Message-ID: <pan.2003.04.07.15.10.02.330094.788@thomas-guettler.de>

On Mon, 07 Apr 2003 14:16:26 +0200, Ruslan Spivak wrote:

> Hello, python users!
> 
> I need to convert string from utf-8 in 'cp1251' codepage. How to achieve
> that?

cyrillic_utf8='....'
u=unicode(cyrillic, 'utf8')
cyrillic_cp1251=u.encode('cp1251')

 thomas


-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From staschuk at telusplanet.net  Thu Apr 17 18:14:36 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 17 Apr 2003 16:14:36 -0600
Subject: Memory leak while looping
In-Reply-To: <n8t1n-ms4.ln1@isomorphism.org>; from bbeck@NOSPAM.austin.rr.com on Thu, Apr 17, 2003 at 06:29:29AM +0000
References: <1020cbf8.0304162100.6fc49e6f@posting.google.com> <n8t1n-ms4.ln1@isomorphism.org>
Message-ID: <20030417161436.C10011@tibia.amotlpaa.bogus>

Quoth Brandon Beck:
> [...] Consider using the generator (or maybe it's just generator like)
> version of range, xrange.  That should run in constant memory.

To be precise, xrange creates an immutable sequence object -- an
iterable, not an iterator (and in particular not a generator).

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From python at dylanreinhardt.com  Sat Apr 19 13:00:56 2003
From: python at dylanreinhardt.com (Dylan Reinhardt)
Date: 19 Apr 2003 10:00:56 -0700
Subject: Parent references: hot or not?
In-Reply-To: <20030419145756.GA25916@meson.dyndns.org>
References: <1050763980.14721.41.camel@ida.dylanreinhardt.com>
	 <20030419145756.GA25916@meson.dyndns.org>
Message-ID: <1050771656.14717.48.camel@ida.dylanreinhardt.com>

On Sat, 2003-04-19 at 07:57, Jp Calderone wrote:
> > 4. It will be impossible to fully delete a non-empty Container.
> 
>   I'm not sure what "fully delete" means here.  The container will certainly
> be able to do "del self.items".  If "self.items" was the only reference to
> the items in the container, then the will all be collected by Python and
> de-allocated.

By that, I meant that if you deleted a non-empty Container, its Items
would still carry a reference to it.  You pretty much answered this
question in your response to my question about the potential for
recursive issues.

Sounds like I should be sure to write a __del__ method for Container
that ensures things work correctly.

Thanks for your comments,

Dylan






From tim at frontier.net.nospam  Thu Apr  3 22:24:22 2003
From: tim at frontier.net.nospam (Tim H)
Date: Thu, 3 Apr 2003 20:24:22 -0700
Subject: Cygwin Python regression testers wanted
References: <mailman.1049390947.15214.python-list@python.org>
Message-ID: <v8pun8oringf6@corp.supernews.com>

>>> Python 2.2.2 (#1, Mar  9 2003, 08:18:26)
>>> [GCC 3.2 20020927 (prerelease)] on cygwin
>>>
>>> The output is attached.
>>>
>>> ...
>>> 172 tests OK.
>>> 1 test failed:
>>>    test_commands
>>> ...
>>
>> XP Pro is the same as NT 4.0 and 2000 (as expected).
>>

Hmm,
    I know you didn't ask for Win2000, but my box gives different results:

169 tests OK.
1 test failed:
    test_dircache
23 tests skipped:
    test_al test_cd test_cl test_curses test_dbm test_dl test_email
    test_email_codecs test_gl test_imgfile test_largefile
    test_linuxaudiodev test_locale test_minidom test_nis test_pyexpat
    test_sax test_socket_ssl test_socketserver test_sunaudiodev
    test_unicode_file test_winreg test_winsound
Ask someone to teach regrtest.py about which tests are
expected to get skipped on cygwin.

test_commands does not fail, plus other differences.

System:
Python 2.2.2 (#1, Mar  9 2003, 08:18:26)
[GCC 3.2 20020927 (prerelease)] on cygwin

Win2000 sp3.  I let cygwin setup update all packages when installing rebase
today.

Tim H





From blunck at gst.com  Sun Apr 27 23:06:48 2003
From: blunck at gst.com (Christopher Blunck)
Date: 27 Apr 2003 20:06:48 -0700
Subject: error in
References: <mailman.1050622890.4267.python-list@python.org>
Message-ID: <1147e466.0304271906.113457ad@posting.google.com>

Mod_Python executes your modules using restricted execution mode. 
Under restricted execution mode, a module cannot execute arbitrary
code (which your handler module qualifies as).  PyXML tries to load
encodings (usually UTF-8) when processing your xml message.  The
loading of the encodings is quantified as executing arbitrary code. 
It is promptly stopped with the message you displayed.

I'm looking into ways to either Bastion PyXML or to get mod_python to
support PyXML.  Will post more when I learn more.

-c

Vince Buonassisi <vjb at vuit.com> wrote in message news:<mailman.1050622890.4267.python-list at python.org>...
> hi,
> 
> i am getting the following error when i load my test script in my web  
> server:
> 
> Mod_python error: "PythonHandler reqhandler"
> 
> Traceback (most recent call last):
> 
>    File "/usr/lib/python2.2/site-packages/mod_python/apache.py", line  
> 181, in Dispatch
>      module = import_module(module_name, _req)
> 
>    File "/usr/lib/python2.2/site-packages/mod_python/apache.py", line  
> 335, in import_module
>      module = imp.load_module(mname, f, p, d)
> 
>    File "/usr/local/apache/htdocs/reqhandler.py", line 9, in ?
>      from xml.dom.ext.reader.Sax2 import FromXmlStream, FromXml
> 
>    File  
> "/usr/lib/python2.2/site-packages/_xmlplus/dom/ext/reader/__init__.py",  
> line 21, in ?
>      encoder = codecs.lookup("utf-8")[0] # encode,decode,reader,writer
> 
>    File  
> "/usr/src/build/143041-i386/install/usr/lib/python2.2/encodings/ 
> __init__.py", line 51, in search_function
> 
> RuntimeError: cannot unmarshal code objects in restricted execution mode
> 
> the script i am running is very simple.  i just send a text message to  
> the browser.  i am using the following versions of software:
> 
> apache 1.3.26
> python 2.2.1
> mod python 2.7.8
> PyXML 0.7.1
> linux redhat 8
> 
> any help would be appreciated.
> 
> thanks
> 
> vince
> --



From bokr at oz.net  Fri Apr 11 14:40:23 2003
From: bokr at oz.net (Bengt Richter)
Date: 11 Apr 2003 18:40:23 GMT
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <b7726n$8pq$0@216.39.172.122>

On Thu, 10 Apr 2003 15:13:11 +0200, Alex Martelli <aleaxit at yahoo.com> wrote:

>I _know_ there's a decent algorithm to solve the following problem,
>but I don't recall its name and thus can't rapidly google for the 
>details... can somebody help?
>
>
>The problem: I need to design a "decision criterion" to classify
>"observations".  For each observation I measure the values of a
>number N of features, x1, x2, ... xN; the desired design criterion
>is a set of weights w1, w2, ... wN such that for any observation I 
>will then just compute a weighted sum
>  S = w1*x1 + w2*x2 + ... + wN*xN
>and classify the observation as Black if S<=1, White if S>1.  To
>train my classifier I have a large corpus of observations already
>made and annotated with "ground-truth" data about whether each
>given observation should have been classified as B or W, and an
>error-cost value for each kind of classification (Ebw is the cost
>of erroneously classifying a feature as W when it should be B,
>Ewb is that of classifying it as B when it should be W).  So,
>what's the algorithm to estimate the weights given the corpus of
>observation and ground-truth data, and the error-costs?
>
My first question(s) would be about the "observations": E.g., are
they orthogonal/independent? Second, what kinds of distributions do
they have? What kinds of joint distributions do they have?

Is a simple linear combination really appropriate (which is like
a projection of a multi-dimensional point onto a line which you bisect
to make the decision)? What if your multi-dimensional points are really
in a bunch of clusters that should be labeled with ground-truth? I.e, if
there is no projection that doesn't overlap clusters with conflicting
ground-truth, so that you have to evaluate new data by figuring out what
cluster new data belongs to and then look up your result in a dictionary
of clusters identified and labeled according to your training data.

If there is no knowledge other than what is implicit in your training set,
what implicit assumptions are being made? Are there measurement error
distributions as well as distributions of perturbations of the measured
things themselves (e.g., digital roundoff plus shaky hands plus measurement
of water depth in a pool with some waves) and is that worth thinking about
vs just making a gaussian or boxcar assumption?

Are the assumptions explicit? Is there knowledge about constraints on relationships
between observations (i.e., if they are not totally independent) that could help?
E.g., absolute mutual exclusions of concurrent existence of valid measurements
(which would suggest partitioning the problem algorithmically before applying
perhaps different continuous math models) as opposed to cross correlations etc.,
or coding relationships such as between bits of an error correcting code?

Well, that was my reaction, FWIW. I'm sure I'm not the only one curious
as to what you are measuring and classifying ;-)

Regards,
Bengt Richter



From uche at ogbuji.net  Thu Apr 10 11:04:40 2003
From: uche at ogbuji.net (Uche Ogbuji)
Date: 10 Apr 2003 08:04:40 -0700
Subject: distributed computing implementations
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <m3r88kk24p.fsf@mira.informatik.hu-berlin.de>
Message-ID: <d116fbae.0304100704.68ef4833@posting.google.com>

martin at v.loewis.de (Martin v. L?wis) wrote in message news:<m3r88kk24p.fsf at mira.informatik.hu-berlin.de>...
> escalation746 at yahoo.com (robin) writes:
> 
> > If you want the biggest, boldest approach and don't care about
> > overhead, use CORBA.
> 
> I can't agree with that analysis: CORBA has, of all your alternatives,
> the least network bandwidth requirements.

True, and Mike Olson did some pretty thorough analysis on this.  See:

http://www-106.ibm.com/developerworks/library/ws-pyth9/


--Uche
http://uche.ogbuji.net



From bokr at oz.net  Tue Apr 29 13:56:11 2003
From: bokr at oz.net (Bengt Richter)
Date: 29 Apr 2003 17:56:11 GMT
Subject: Determining the volume ID for a file/drive?
References: <uv7savkms6lsgchcv6jh8ahtccu9lon3oi@4ax.com> <JZtra.19431$3M4.563286@news1.tin.it> <zlvra.19900$3M4.571480@news1.tin.it>
Message-ID: <b8mebr$9i2$0@216.39.172.122>

On Tue, 29 Apr 2003 13:45:03 GMT, Alex Martelli <aleax at aleax.it> wrote:

>Alex Martelli wrote:
>   ...
>> Quick and dirty...:
>> 
>> import os
>> x = os.popen('C:\\')
>
>OOPS -- TOO quick -- sorry, I meant os.popen('DIR C:\\')!!!
>
Or
 >>> import os
 >>> os.popen('vol C:').readlines()
 [' Volume in drive C is System\n', ' Volume Serial Number is 14CF-C4B9\n']

 >>> def drive2volser(drive):
 ...     import os
 ...     return os.popen('vol %s'%drive).readlines()[1].split()[-1]
 ...
 >>> drive2volser('C:')
 '14CF-C4B9'

(Note that volume labels can have spaces, so you might want to .split(' is ')
and .strip() if you want that).

Regards,
Bengt Richter



From aleax at aleax.it  Tue Apr 29 17:59:43 2003
From: aleax at aleax.it (Alex Martelli)
Date: Tue, 29 Apr 2003 21:59:43 GMT
Subject: Question about accessing class-attributes.
References: <mailman.1051637056.20383.python-list@python.org>
Message-ID: <jBCra.33970$K35.1085354@news2.tin.it>

Bjorn Pettersen wrote:

>> From: Alex Martelli [mailto:aleax at aleax.it]
> [...]
>> > however I haven't been able to find a good description of attribute
>> > lookup for new-style classes... (references?) E.g. why is
>> > the metaclass' __getattr__ never called?
   ...
> Here's the case I'm confused about:

[special-methods problem snipped]

Ah yes, I see now!  Yes, functions such as len() rely on slots in
the type object, e.g. as you've noticed:

> finding the function they're needing, e.g. for len, it looks like it
> uses:
> 
>   m = o->ob_type->tp_as_sequence;
>   if (m && m->sq_length)
>     return m->sq_length(o);
> 
>   return PyMapping_Size(o);

and the "incredibly complex thinking" (quoting from typeobject.c)
in update_one_slot doesn't seem to work except for operations the
which "the class overrides in its dict" (again from a comment in
typeobject.c, this one for fixup_slot_dispatchers).

The issue may be with _PyType_Lookup (again in the same ,c file),
which just gives up if it can't find a name somewhere along the mro
(it doesn't "look upwards" to the metaclass) while type_getattro
DOES work upwards on the metaclass too.  Hmmmm.  I'm not sure I
really understand all that's going on here - it IS a rather hairy
zone of the code.  Maybe you can post this as a bug in 2.3 beta 1
on sourgeforge (ideally showing where in the docs it defines the
semantics that it then doesn't respect) so we can get this looked
at by the few people who really DO grasp these parts...;-).  There
is probably some sound implementation reason for the current
behavior, but if so it should be better documented, I think.


> which would never call the metaclass' __getattr__ (since tp_as_sequence
> would presumably be NULL?).
> 
> The relevance this has to super, is that you couldn't do a (contrived)
> example like:
> 
>   for item in super(MyClass, self):
>       ...
> 
> (in general, you can't pass super(MyClass, self) to a function expecting
> a MyClass instance, even if MyClass defines no new behavior...)

Yes, of course not -- super(MyClass, self) is NOT an instance of MyClass
(isn't THAT why you're calling super...?).  

> 
> Perhaps a more realistic exmaple?:
> 
>   class Secure(object):
>     def __init__(self, name):
>       self.name = name
>   
>     def __getattr__(self, attr):
>       return getSecureMethod(self.name, attr, getUserPrivileges())
>         
>   class LogFoo(Secure):
>     def __init__(self):
>       Secure.__init__(self, 'SalaryDB')
>       
>     def foo(self):
>       print 'calling foo'
>       return super(LogFoo, self).foo()
>       
>   lf = LogFoo()
>   print lf.foo()

Sorry, I don't get it.  How is this related to metaclasses...?


Alex




From intentionally at blank.co.uk  Thu Apr 24 16:49:48 2003
From: intentionally at blank.co.uk (Stephen Horne)
Date: Thu, 24 Apr 2003 21:49:48 +0100
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com> <3EA80827.6277D27C@engcorp.com>
Message-ID: <oafgavg8e0pf512712g4aenldqktvmb0dr@4ax.com>

On Thu, 24 Apr 2003 11:52:07 -0400, Peter Hansen <peter at engcorp.com>
wrote:

>I think if pathways in the brain can be built up, through repetition of
>stimuli, dreaming, etc., then they can likewise be dissolved, whether
>through drug use, pyschological problems or chemical imbalances caused 
>by lack of sleep, and so on.  Perhaps simple disuse leads only to an
>asymptotic decay which never quite reaches zero, as your argument might
>suggest, but I think other processes could act in different ways.

There are those who claim that the human brain contains a record of
everything that ever happened to it, and that an inability to remember
results from the inability to bring that record back to conscious
recollection. Complete garbage. The brains capacity may be immense,
but it isn't that immense - and if it was, wasting resources in
remembering numerous memories that cannot be recalled or applied in
any useful way would be clearly against the principal of survival of
the fittest. Especially considering that the brain uses much more
energy per unit of weight than any other part of the body.

Some memories may still be present but difficult or impossible to
recall - due to the loss of the associative links between memories -
but the resources dedicated to such memories are prone to being
re-used. Memories don't stay in one part of the brain - at least, not
for several years at least (when they may finally make it into
permanant memory). They get shuffled back and forward between two
different parts of the brain during resorting processes. There are
plenty of chances to reallocate those resources to other memories if
they are unused.

Cognitive psychology and neurology have determined some very
interesting facts about human memory. One such fact is that all human
memory is subjected to extreme lossy compression. Basically, only
aspects of the memory which are considered particularly relevant and
different to expectations are stored. Everything else is factored out.
On recall, the full memory is reconstructed by filling in the gaps
with current knowledge, expectations and prejudices.

If the general state of knowledge, expectations and prejudices has
radically changed in the mean time, the reconstructed memory may not
resemble the original event very closely at all.

A surprising fact that courts love to ignore - a persons confidence in
the accuracy of their recall is no guide to the true accuracy of what
they recall. Confident people may speak confidently and may convince
people very effectively, but they also may not question the accuracy
of their own memory. A less confident person may report memories more
accurately, and with a more accurate sense of the limitations of those
memories, yet will tend to be disregarded and disbelieved due to that
lack of confidence.

There have been many cases where people convicted due to the evidence
given by victims or witnesses, who were absolutely confident in their
memories, have later been proved innocent. Fancy spending a few
decades in prison for rape, only to later be proven innocent by DNA
evidence, because of a victims over-confidence in an intense but
inaccurate memory? - It *has* happened.

Experts are only expert in particular fields. Just because Einstein
was a genius in the field of physics, that did not qualify him to
state that only 10% of the brain is unused. Assuming he claimed it at
all - I never checked, but it does get attributed to him now and again
- it was probably never intended as a statement of strict biological
fact. Benjamins view of Dijkstra's point can probably be seen in a
similar light.

Of course it is a basic fact of physics that any observer is affected
by everything it observes (as well as affecting the thing it
observes), and chaos theory states that the impact of small affects is
often unexpectedly large, but that isn't the same as a memory. You
can't reconstruct a recollection from those affects even if you could
accurately quantify them (which you can't). Its an information theory
thing.


Sorry - I have a broken brain (Asperger syndrome) myself, so I have an
obsessive interest in certain aspects of the brains inner workings.




From camel at oasis.com  Tue Apr  8 12:30:50 2003
From: camel at oasis.com (dromedary)
Date: Tue, 08 Apr 2003 12:30:50 -0400
Subject: Palindrome finder: help find ways to speed up?
Message-ID: <080420031230508991%camel@oasis.com>

I've written a palidrome finder two ways (see below), both of which
work, but verrrrrry slooooooooooowly. Would anybody out there care to
look over the code and offer an idea or two about how to speed it up?

The first example uses recursion (handy for combining any number of
words); the second example uses for loops (not fast but faster, though
clunkier).

But?if you suggest a C/C++ wrapper or extension, please please give me
guidance on how to implement it. I've never translated a Python widget
into C/C++.

___________________________________________________

CODE 1 (uses recursion to combine words, then looks for palindromes)

##### Combine 'em
def combinations(alist, numb, blist=[]):
   if not numb: 
      return rPal(''.join(blist))
   for i in range(len(alist)):
      blist.append(alist.pop(i))
      combinations(alist, numb-1, blist)
      alist.insert(i, blist.pop())

##### Look for symmetry
def rPal(w):
   x=0
   y=1
   L=len(w)
   h=L/2
   if L % 2 > 0:
      while x < h:
         if w[h-y]==w[h+y]:
            x=x+1
            y=y+1
            if x == h:
               print w
               break
         else:
            break          
   else:
      while x < h:
         if w[h-y]==w[h+x]:
            x=x+1
            y=y+1
            if x == h:
               print w
               break 
         else:
            break

##### Get a word list
words = open(SOME_WORD_FILE).read().splitlines()

##### Set number of words to combine
n=SOME_NUMBER_OF_WORDS

########################################
##### Call the recursive function to print the palindromes
if __name__ == '__main__': 
   
   combinations(words, n)


^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

CODE 2 (for loops; the example uses three loops, but you can see how
complex it can get for each additional iteration)

##### Get a word list
words = open(SOME_WORD_FILE).read().splitlines()

##### Combine pairs of words into keys and values
for i in words:
   for j in words:
      for k in words:
         if i != j and i != k and j !=k:
            w=i+j+k
            x=0
            y=1
            L=len(w)
            h=L/2
            if L % 2 > 0:
               while x < h:
                  if w[h-y]==w[h+y]:
                     x=x+1
                     y=y+1
                     if x == h:
                        print w
                        break
                  else:
                     break
                     
            else:
               while x < h:
                  if w[h-y]==w[h+x]:
                     x=x+1
                     y=y+1
                     if x == h:
                        print w
                        break 
                  else:
                     break



From alloydflanagan at attbi.com  Fri Apr 11 19:46:50 2003
From: alloydflanagan at attbi.com (A. Lloyd Flanagan)
Date: 11 Apr 2003 16:46:50 -0700
Subject: Macros in Python? (PEP 310)
References: <b71qde$q71$07$1@news.t-online.com> <mailman.1050064218.12167.python-list@python.org> <b770t8$2b8$00$1@news.t-online.com> <b771jb$ac6$03$1@news.t-online.com>
Message-ID: <a8b7f07a.0304111546.3e5a5636@posting.google.com>

Dominic <oblivious at web.de> wrote in message news:<b771jb$ac6$03$1 at news.t-online.com>...
> Dominic wrote:
> > Beni Cherniavsky wrote:
> > 
> >> Dominic wrote on 2003-04-09:
> >>
> >>
> >>> (defmacro with-root-privs (() &body body)
> >>>   (let ((oldidsym (gensym)))
> >>>     `(let ((,oldidsym (geteuid)))
> >>>        (seteuid 0)
> >>>        (unwind-protect
> >>>        (progn , at body)
> >>>      (seteuid ,oldidsym)))))
> >>>
> >>> Is there an easy way to produce similar code
> >>> in Python which gurantees that certain things
> >>> happen around some code? (Maybe by using
> >>> generators?)
> >>>
> >>
> ...

I think you are making this a lot harder than it needs to be. 
Remember that function definitions are first class objects in python:

def y(a, b):
    " a simple function to do nothing significant
    return a + b

_y = y  #save defn. of y

def x(a, b):
    # function to wrap stuff around y
    print "hi"
    c = _y(a, b)
    print "bye"
    return c

y = x  #rebind y so it looks identical to caller

>>> y(2, 3)
hi
bye
5

If this isn't what you are trying to do, could you clarify what's
missing?



From aleax at aleax.it  Wed Apr 16 06:30:02 2003
From: aleax at aleax.it (Alex Martelli)
Date: Wed, 16 Apr 2003 10:30:02 GMT
Subject: instance creatiion as an "event"
References: <mailman.1050095779.7490.python-list@python.org> <pan.2003.04.11.23.20.04.65410@webone.com.au> <mailman.1050117319.5102.python-list@python.org> <b7gt8r$em0$1@panix1.panix.com> <P4Tma.22981$T34.691702@news2.tin.it> <b7i63s$1869h$1@ID-169208.news.dfncis.de>
Message-ID: <Kgana.21251$LB6.523284@news1.tin.it>

Greg Ewing (using news.cis.dfn.de) wrote:

> Alex Martelli wrote:
>> If Member.members is coded
>> as above, then you have circular reference loops involving the
>> class and each instance so just about everything becomes sort of
>> "immortal" -- with weak references in Member.members, no circular
>> reference loops, no problems (or, am I missing something...?).
> 
> Circular references aren't a disaster these days, with
> the cyclic garbage collector.

...unless something in the loop has a __del__, in which case
you're skirting disaster-area again.

But that's not really what we were talking about.  It's unlikely
the class object will go away anytime soon -- most of the time
it's a top-level name in a module, or referenced by one, and you
know how persistent those are.  Now, if the class object has
NON-weak references to all of the instances, those instances
become deucedly persistent too.  So why should the references
from class to instances be weak, since they're NOT intended
to keep the instances alive, merely to be able to reach those
that are STILL alive for other reasons when something happens?


> But if you want to be sure things are reclaimed as
> soon as possible, weak refs are still a good idea.

More generally, weak references are useful when you want to
be able to refer to something if that something still exists,
but NOT keep that something in existence otherwise -- that's
their design intent, after all.


Alex




From zyziu at interia.pl  Thu Apr 17 05:51:15 2003
From: zyziu at interia.pl (zyziu)
Date: Thu, 17 Apr 2003 11:51:15 +0200
Subject: scrollbar
References: <b7kh3g$51u$1@nemesis.news.tpi.pl> <MAlna.4040$ej5.2777@news01.roc.ny.frontiernet.net>
Message-ID: <b7ltfi$49c$1@atlantis.news.tpi.pl>

----- Original Message -----
From: "Lee Harr" <missive at frontiernet.net>
Newsgroups: comp.lang.python
Sent: Thursday, April 17, 2003 1:22 AM
Subject: Re: scrollbar


> In article <b7kh3g$51u$1 at nemesis.news.tpi.pl>, zyziu wrote:
> > Hello. I made lanChat in Python. I have problem with Text widget. I want
to
> > text from Entry goes to Text widget and scrollbar goes down, because I
can't
> > see what write another user. Now I must do it manually. How do it to
> > scrollbar goes down automatically? Thanks for answer. Zyziu from Poland.
> >
> >
>
> You should probably be a little bit more specific, and maybe include
> some of the code around where you are having the problem.
>
> That said, if this is Tkinter, maybe something here will help:
>
>
>       self.wout = Tkinter.Tk()
>       self.frame = Tkinter.Frame(self.wout, bg='brown')
>
>
>       self.frame.sc = Tkinter.Scrollbar(self.frame,
orient=Tkinter.VERTICAL)
>
>       self.frame.tout = Tkinter.Text(self.frame,
yscroll=self.frame.sc.set,
> bg='black', fg='white')
>
>       self.frame.sc['command'] = self.frame.tout.yview
>       self.frame.sc.pack(side='right', fill='y')
>
>
>          t = self.c.read_eager().replace('\r','')
>
>          if t:
>             self.w.frame.tout.config(state='normal')
>             self.w.frame.tout.insert(Tkinter.END, t)
>             self.w.frame.tout.config(state='disabled')
>             self.w.frame.tout.yview(Tkinter.END)
>
>
>
>
> This is just bits and pieces of a file I wrote a couple of years ago...
> I remember having trouble with the same thing.  Hope this helps.
>
>
Oops! Sorry. I use Python 2.2.2 with Idle 0.8. You wrote:
   t = self.c.read_eager().replace('\r','')
What is 'c'? Is it scrollbar? Is 't' read position of scrollbar? Look at my
code and tell me where I must put it into my code. Please correct my code
and send me back if you can, please. Thanks. Zyziu from Poland.

# Tkinter
from Tkinter import *
from time import *
import string

class App(Frame):

    def __init__(self, master=None):
         Frame.__init__(self,master)
         self.master.title('Zyziu123 wzywa')
         self.grid(row=2, column=2)
         self.konstrukcja() #wywolanie funkcji

    def konstrukcja(self):
         self.scrollbar = Scrollbar(self, orient=VERTICAL)
         self.okno = Text(self, yscrollcommand=self.scrollbar.set)
         self.scrollbar.config(command=self.okno.yview)
         self.scrollbar.grid(row=0, column=2,sticky=S+N+W+E)
         self.okno.grid(row=0,column=0,columnspan=2)

         self.tekst=Entry(self)
         self.tekst.grid(row=1,column=0,columnspan=3,sticky=W+E)

         self.guzik=Button(self, text='send', command=self.Pisz)
         self.guzik.grid(row=1,column=3,sticky=E)

    def Pisz(self):
         self.dane = self.tekst.get()
         self.tekst.delete(0,'end')
         if (self.dane<>''):
             self.okno.insert('end',strftime("%H:%M:%S")+'\n')
             self.okno.insert('end',self.dane+'\n'+'\n')


ap = App()
ap = mainloop()






From tim.one at comcast.net  Mon Apr 21 13:40:21 2003
From: tim.one at comcast.net (Tim Peters)
Date: Mon, 21 Apr 2003 13:40:21 -0400
Subject: negative lookahead question
In-Reply-To: <16036.6978.270321.861442@montanaro.dyndns.org>
Message-ID: <LNBBLJKPBEHFEDALKOLCKELMEDAB.tim.one@comcast.net>

[Skip Montanaro]
> This re.sub call lives in Lib/stmplib.py as a way to make line endings
> canonical:
>
>     re.sub(r'(?:\r\n|\n|\r(?!\n))', CRLF, data)
>
> This certainly seems to do what's desired, however, it looks
> overly complex to me.  First, the non-grouping parens are unnecessary.

Yup.

> Second, I don't think the negative lookahead assertion is required.

That's also so.  Because alternatives are tried left to right, if we get to
the third alternative then we know we can't be looking at \r\n (that was the
first alternative, but the match attempt didn't stop there).

> This simpler function call seems to do the trick:
>
>     re.sub(r'\r\n|\n|\r', CRLF, data)

It should.  Other simpler equivalents are the regexps

    \r\n|[\r\n]

and

    \r\n?|\n

An interesting question is whether it's really necessary to replace \r\n
with itself (CRLF == '\r\n').  IOW, what's wrong with

    \r(?!\n)|\n

?  The problem is that it doesn't actually leave \r\n alone.  There's no
match when looking at the \r in \r\n, but then re.sub moves forward a
character and tries again, looking at the \n in \r\n.  That one would match,
and \r\n would get changed to \r\r\n.





From ckea25d02 at sneakemail.com  Wed Apr  2 19:34:46 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Thu, 03 Apr 2003 12:34:46 +1200
Subject: a pyrex-inspired for-i-from-1-to-n construct backported to Python.
   Good idea? Bad Idea? PEP?
In-Reply-To: <RbCia.6782$9I.555309@news20.bellglobal.com>
References: <Xns934E6D8A81E5BjslmAAearthlinknet@130.133.1.4> <b68jti$2mj66$1@ID-169208.news.dfncis.de> <RbCia.6782$9I.555309@news20.bellglobal.com>
Message-ID: <b6fv3c$5218t$1@ID-169208.news.dfncis.de>

WP wrote:
> Greg Ewing wrote the following Pyrex snippet:
> 
> >     for i from 0 <= i < n:
> >       result.append(a[i])
 >
> does anyone else thin[k] that for the extremely common ... case 
 > where you need to do a for loop over a range ... of integers, that a
 > language level keyword construct ... being expressed in a bytecode
 > optimized fashion would be good?


The idea for the above syntax came out of an earlier discussion
about this. In a Python context, the main motivation for it wasn't
speed -- as has been pointed out, there are ways that integer
for-loops could be optimised that don't require language changes.

Rather, the motivation was readability and expressiveness. The
things I like about it are that it gives very clear and unambiguous
control over the inclusion or exclusion of both endpoints and
the direction of iteration.

Contrast this with range(m, n, s), which only gives you one
option inclusion of each endpoint, and if s is negative, there
are a number of different and equally plausible things it could
mean.

The range() function with one argument is very useful if you
want to generate indices for a sequence, where it does just
what you want; but if you're using the indices for something
else that has nothing to do with sequences, it's just as
likely to do a rather poor job.

J. P. Calderone (exarkun at intarweb.us):
 >  Hopefully something better than that Pyrex-style loop you've got above.

Actually, I'd like to make it just

    for 0 <= i < n:
      ...

This is what I wanted to do originally; the "i from" was added
to make it easier to parse. Having had some experience of using
it now, I'm feeling that my original instinct was right. I'm
going to have a go at getting rid of the "from" some day.

 > (How do you do steps with that, anyway?)

You don't -- you map a sequence with step 1 to whatever
other sequence you want.

I'm not inclined to try to shoehorn a step into it, because
it would spoil the symmetry. The way it is, the two endpoints
are treated identically. With a step, one of them would have
to be singled out as the "starting" value, which would always
be hit, with the other one possibly being hit or not depending
on the circumstances. Then you're well on the way towards
getting back all the ugliness of range(m, n, s).

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From ckea25d02 at sneakemail.com  Sun Apr  6 23:21:56 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 07 Apr 2003 15:21:56 +1200
Subject: How to convert certain typedefs to Pyrex?
In-Reply-To: <b6lq3p$hls$1@bob.news.rcn.net>
References: <b6lq3p$hls$1@bob.news.rcn.net>
Message-ID: <b6qqb6$82pho$2@ID-169208.news.dfncis.de>

Edward C. Jones wrote:
> How do I treat the following typedefs in Pyrex?
>
> typedef struct _IplTileInfo IplTileInfo;
> typedef struct _CvContourScanner* CvContourScanner;
> typedef struct CvPOSITObject CvPOSITObject;

ctypedef struct IplTileInfo
cdef struct _CvContourScanner
ctypedef _CvContourScanner *CvContourScanner
ctypedef struct CvPOSITObject

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From gh at ghaering.de  Thu Apr 24 07:23:14 2003
From: gh at ghaering.de (Gerhard Haering)
Date: Thu, 24 Apr 2003 13:23:14 +0200
Subject: Problems at python.org??
In-Reply-To: <Pine.LNX.4.44.0304240713090.12990-100000@bagend.odc.vzwcorp.com>
References: <Pine.LNX.4.44.0304240713090.12990-100000@bagend.odc.vzwcorp.com>
Message-ID: <3EA7C922.5050708@ghaering.de>

Scherer, Bill wrote:
> Pointing my browser at http://www.python.org, I receive:
> 
> """
> No more connections can be made to this remote computer at this 
> time because there are already as many connections as the 
> computer can accept.
> """
> 
> Has there been a sudden Python Popularity Explosion?

That's strange, because a Google search suggests that you're really 
getting this message from a Windows 2000 Server, not from the python.org 
webserver.

-- Gerhard






From whisper at oz.net  Sun Apr  6 18:51:14 2003
From: whisper at oz.net (David LeBlanc)
Date: Sun, 6 Apr 2003 15:51:14 -0700
Subject: Could Emacs be rewritten in Python?
In-Reply-To: <m2k7e78j5s.fsf@localhost.localdomain>
Message-ID: <GCEDKONBLEFPPADDJCOEKEINJFAA.whisper@oz.net>

> The main reason I started PyAlaMode is that I wanted the same
> autocompletion and calltips that I've got in the PyCrust shell, but
> while I'm editing a file.  Those features are working now, and are
> pretty sweet, if I do say so myself, since they are based on dynamic
> namespaces updated whenever you want, one per buffer, with no
> conflicts no matter how many buffers you have open, with automatic
> reloading of imported modules.

Scite (based on Scintilla) has calltips and auto-completetion and is
friendly towards Python. It's multi-platform too. I've often thought a
python-scriptable Scite would be a great thing. I'd bet that wrapping (boost
or swig) the main scite (C++) class would give you a lot of the same
primatives as emacs. Graft on user keystroke assignment and you're well on
the way...

Dave LeBlanc
Seattle, WA USA





From tismer at tismer.com  Sat Apr 19 00:05:33 2003
From: tismer at tismer.com (Christian Tismer)
Date: Sat, 19 Apr 2003 06:05:33 +0200
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <b7pchr$ig0$1@main.gmane.org>
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <b7pchr$ig0$1@main.gmane.org>
Message-ID: <3EA0CB0D.5060105@tismer.com>

Mike Rovner wrote:
> "Martin v. L?wis" <martin at v.loewis.de> wrote in message
> 
> 
>>>Is there a short way to say why Guido didn't want continuations?
>>
>>He didn't like the implementation strategy; I believe he was neutral
>>about the feature per se.
> 
> 
> Second that.
> AFAIK one of the points was/is the ability to implement it in Jython,
> which is/was impossible while you are mangling with C stack frames.

Stackless 1.0 did continuations without mangling C stack frames.
I still don't see how this is unimplementable in Jython, especially
with the existing generators, which are dong something limited but
similar.

-- 
Christian Tismer             :^)   <mailto:tismer at tismer.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  pager +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/






From ckea25d02 at sneakemail.com  Fri Apr  4 01:08:44 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Fri, 04 Apr 2003 18:08:44 +1200
Subject: Overloadable Assignment PEP
In-Reply-To: <9383481e.0304031917.1d0100bf@posting.google.com>
References: <9383481e.0304012011.78f05548@posting.google.com> <3e8c083b$0$10386$edfadb0f@dread11.news.tele.dk> <9383481e.0304030705.524cf0fa@posting.google.com> <mailman.1049388384.29614.python-list@python.org> <9383481e.0304031917.1d0100bf@posting.google.com>
Message-ID: <b6j711$5m5dh$2@ID-169208.news.dfncis.de>

Drew Moore wrote:
> # modify the voltages, in a very simple, readable syntax.
> vdduc = -3.8 # try a reduced supply
> vreset = -3.2 # and increased reset drive...
> vbias = 2.3 # and decreas the bias a little..

Maybe our brains work differently, but I'd find it
FAR less confusing if the dac were an object and
one did

   dac.vdduc = -3.8
   dac.vreset = -3.2
   dac.vbias = 2.3

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From aleax at aleax.it  Sun Apr 13 18:02:07 2003
From: aleax at aleax.it (Alex Martelli)
Date: Sun, 13 Apr 2003 22:02:07 GMT
Subject: Reference Tracking
References: <b79km7$iur$1@wisteria.csv.warwick.ac.uk> <KUYla.7409$T34.201025@news2.tin.it> <b7c46c$cvo$1@wisteria.csv.warwick.ac.uk>
Message-ID: <z7lma.10014$LB6.267626@news1.tin.it>

Matt Bergin wrote:
   ...
>> You can try to avoid the periodic scanning overhead by using a
>> tricky __del__, or a non-tricky __del__ or weak-reference callback
>> for some object that's "packed together with" the object you give
>> out (keeping the latter alive in a per-class container anyway).  A
>> radically different and far cleaner approach would be to specify
>> a way (e.g. a method) for the using-code to say "I'm finished with
>> this object, reuse it freely from now on" -- no black magic at all
>> (of course, perhaps marginally smaller convenience for the using
>> code, but it's not all that bad to ensure a finalization method
>> gets explicitly called, IMHO).
> 
> I'm just worried that if an exception occurs in a thread then the cursor
> will never be returned, using the 'no black magic' method which appeals
> to me the most. Could I prevent this by careful use of try: finally?

Yes.  And you can *in addition* keep the cursors you've handed out
in a list and ALSO periodically do the scan checking refcounts just
to help you debug user threads that aren't returning cursors properly
(or use other tricks as above but for debugging-help rather than
functional purposes).


Alex




From user at domain.invalid  Mon Apr  7 21:49:18 2003
From: user at domain.invalid (nihilo)
Date: Mon, 07 Apr 2003 18:49:18 -0700
Subject: doing hundreds of re.subs efficiently on large strings
In-Reply-To: <b6s9a5$525$0@216.39.172.122>
References: <3E80CE1D.1000200@NOmyrealCAPSbox.com> <b5qu69$t98$0@216.39.172.122> <3E8370F2.2030806@NOmyrealCAPSbox.com> <b65cq3$mi5$0@216.39.172.122> <3E90C634.1030900@NOmyrealCAPSbox.com> <b6s9a5$525$0@216.39.172.122>
Message-ID: <3E922A9E.5000705@domain.invalid>

Bengt Richter wrote:
> On Mon, 07 Apr 2003 00:28:13 GMT, nihilo <exnihilo at NOmyrealCAPSbox.com> wrote:
> 
> 
>>Bengt Richter wrote:
>>
>>>Thanks for testing, that is interesting. It would be interesting to see
>>>the test code you ran, so we could see the reasons for .25 vs .43 (and
>>>I could revise my reality-model as necessary ;-)
>>>
>>>Regards,
>>>Bengt Richter
>>
>>Hi Bengt,
>>
>>Here is some code that I created to test the two versions.
>>
>>Original:
>>
> 
> [...]
> 
>>print time.clock() - start
>>
>>And the modified code is:
>>
>>import time, re
>>start = time.clock()
>>data = '' # same data
>>dict = { 'US and Britain' : 'Oceania', 'US and British' : 'Oceanian', 
> 
> [...~ 135 items ...]
> 
> 
> 
>>pat = re.compile('(Canada|the U\.S\.A\.|air force|International Herald 
> 
> [... ditto ...]
> 
>>strings = pat.split(data)
>>for i in xrange(1,len(strings),2):
>>    strings[i] = dict[strings[i]]
>>data = ''.join(strings)
>>
>>I just checked these again, and the results were .052 on average for the 
>>first version, and .080 on average for the second. The second would 
>>obviously be a lot faster if I pre-compiled the pattern and cPickled it, 
> 
> I had assumed pre-compiled regex, but I learned something anyway:
> The split is a _lot_ slower than I would have thought. I had guessed that
> it would have compiled to a fast state machine like flex might generate
> for or-ed static string matches, but something else must be happening.
> Maybe it was not worth special casing static ors.
> 
> The loop to substitute into the list plus joining the result is quite fast,
> but the split is 2+ orders of magnitude slower than the latter, and the
> regex compilation is somewhat faster than the latter, but not much, apparently.
> Something like 175:150:1 for split:regexcomp:subst-in-list-using-dict-and-join
> in a test I ran on a 63.5k html file with 100 matching substrings in the data.
> 
> 
>>but I wanted to include compilation time in the test.  The fastest 
> 
> Why? Just curious -- would the patterns be changing as often as the data?

Hmmm, I was thinking in terms of equivalent operations, or starting from 
a blank slate in each case, but that is silly. What is important is the 
time, and if one solution allows you to do some of the work beforehand, 
so much the better.

> 
>>version that I have tested is using Alex's replacement suggestion in 
> 
> What version of python are you running? Mine wouldn't compile the regex
> in the form Alex suggested (with every '|' term separately parenthesized,
> like '(x)|(y)' vs '(x|y)' ). Maybe I typoed when I typed the change, but
> I got
> 
>       File "D:\python22\lib\sre_compile.py", line 442, in compile
>         assert p.pattern.groups <= 100,\
>     AssertionError: sorry, but this version only supports 100 named groups
> 
> I was hoping to compare regex speed using the two forms. Can't pursue it now though...

I'm running 2.3a2, and got the same error. I compared without the 
parentheses, figuring that it would give me a rough estimate of how 
quick it was. I am going to break the pattern up into sets of 100 named 
groups when I have a little time. I'll post final times when I get 
around to it if you're curious, including your version with the patterns 
precompiled.

>>combination with pre-compilation and cPickle, but the naive data = 
>>data.replace is surprisingly fast.
>>
> 
> Yes.
> BTW, neither of the regex-based methods are equivalent in semantics
> to the naive sequence of data.replace statements, because each data.replace
> has the possibility (in general, even if maybe not in the actual example)
> of matching something created by a previous data.replace. 

Yes, I realize that. What I actually want is the semantics of the 
regexes, since I don't want to substitute for any of the words that have 
already been substituted. The data.replace method was just an easy way 
of getting the same result, since none of the data.replace output words 
appear as inputs to later data.replaces.

> Thanks for posting the code.
> 
> Regards,
> Bengt Richter

No problem. Thanks for all your help.

regards,

-nihilo




From 'tert at pncen.hf'.decode  Mon Apr 21 01:44:59 2003
From: 'tert at pncen.hf'.decode (Greg Krohn ('rot-13'))
Date: Mon, 21 Apr 2003 00:44:59 -0500
Subject: python dynamic scoping question
References: <79ad5955.0304201513.6ffd4f83@posting.google.com> <b7vg9t01mdo@enews2.newsguy.com> <mailman.1050900693.8645.python-list@python.org>
Message-ID: <b800h901rgl@enews3.newsguy.com>

"Jp Calderone" <exarkun at intarweb.us> wrote in message
news:mailman.1050900693.8645.python-list at python.org...
> On Sun, Apr 20, 2003 at 07:49:55PM -0500, Greg Krohn wrote:
> > [snip]
> >
> > From the docs:
> >
> > globals() - Return a dictionary representing the current global symbol
> > table. This is always the dictionary of the current module (inside a
> > function or method, this is the module where it is defined, not the
module
> > from which it is called).
> >
> > >>> def f():
> > ...   def g():
> > ...     x = 5
> > ...     globals()['x'] = 5
> > ...   g()
> > ...   print x
> > ...
> > >>> f()
> > 5
> >
> >
> > Cool?
> >
>
>   No.  The right way to do this is:
>
>     def f():
>         def g():
>             global x
>             x = 5
>         g()
>         print x
>

Hm. I guess you can tell I never use the global statement. I though it did
something else. Lesson learned.

>   But isn't the answer to the original poster's question, afaict.

Really? The only differnece I see is 'global x'.

>
>   Jp
>
> --
>         "I quite agree with you," said the Duchess; "and the moral of
> that is -- Be what you would seem to be' -- or, if you'd like it put
> more simply -- Never imagine yourself not to be otherwise than what it
> might appear to others that what you were or might have been was not
> otherwise than what you had been would have appeared to them to be
> otherwise.'"       -- Lewis Carrol, "Alice in Wonderland"
> --
>  up 32 days, 1:02, 3 users, load average: 0.21, 0.21, 0.25
>





From bokr at oz.net  Thu Apr  3 18:22:27 2003
From: bokr at oz.net (Bengt Richter)
Date: 3 Apr 2003 23:22:27 GMT
Subject: OT: top-posting and the blind (was Re: need help with game)
References: <3e88b3d3.30570106@news-server> <3E88C1D8.3070305@motorola.com> <3e8a57d0.55986091@news-server> <3E8B1602.72BB1A37@juno.com> <3E8B3F3E.ADA82191@engcorp.com>
Message-ID: <b6ifnj$elu$0@216.39.172.122>

On Wed, 02 Apr 2003 14:51:26 -0500, Peter Hansen <peter at engcorp.com> wrote:

>Jeff Sandys wrote:
>> 
>> Re: top-posting: Blind people with text to speech converters
>> prefer top-posting so they don't have to listen to the same
>> stuff over and over again.
>
>Presumably blind people are not reading their own postings that
>often, so unless you are replying to a blind person with a text
>to speech converter (and one who doesn't match the profile 
>posted by Robin on the issue) then that doesn't make it a good
>idea for you to do it.
>
>If *you* are blind and find it easier for us to top-post in our
>replies to you, then we would likely readily accommodate that 
>when talking to you.
>
But wonder if there isn't a smarter reader, that can skip to
the first unquoted text, and navigate relative context, etc.
Are they just brute force top-to-bottom text-to-speech things?

Maybe Python can help ;-)

Regards,
Bengt Richter



From jmeyer at pdi.com  Wed Apr  9 11:21:21 2003
From: jmeyer at pdi.com (Jim Meyer)
Date: 09 Apr 2003 08:21:21 -0700
Subject: String splitting question
In-Reply-To: <9afd3f36.0304082245.33670341@posting.google.com>
References: <9afd3f36.0304082245.33670341@posting.google.com>
Message-ID: <1049901680.14487.26.camel@utonium.pdi.com>

On Tue, 2003-04-08 at 23:45, Jim Shady wrote:
> Hello,
> 
> I am a beginner at Python and I just joined this group.
> 
> I have a string:
> 
> abcd/df/a/iiwk/abcdefghijkl/b/c
> 
> I need to get the longest string between the /s of the string. For
> example, longest_str() for the above line should return
> 'abcdefghijkl'. How do I go about doing this?

Toss this one onto the heap; it feels a bit more pythonic and
straightforward than some of the solutions offered:

  byLen = lambda a,b : cmp(len(a), len(b))
  words = 'abcd/df/a/iiwk/abcdefghijkl/b/c'.split('/')
  words.sort(byLen)
  print words[-1]

Certainly not as clever as the solves authored by Peter Abel, William
Park, and Max M, but much more easily understood and maintained --
important if you ever want to leave a codebase behind without having to
leave the company. =]

BTW, if you're splitting on "/" because it's a path separator, you might
consider using os.sep instead; it makes your code a bit more portable.

Cheers!

--j
-- 
Jim Meyer, Geek at Large                                  jmeyer at pdi.com





From aahz at pythoncraft.com  Sun Apr  6 14:27:18 2003
From: aahz at pythoncraft.com (Aahz)
Date: 6 Apr 2003 14:27:18 -0400
Subject: Could Emacs be rewritten in Python?
References: <m2u1dcw22s.fsf@localhost.localdomain> <m2k7e7addl.fsf@localhost.localdomain> <b6pgcm$1t3$1@panix3.panix.com> <m27ka7aae0.fsf@localhost.localdomain>
Message-ID: <b6pri6$drf$1@panix2.panix.com>

In article <m27ka7aae0.fsf at localhost.localdomain>,
Patrick K. O'Brien <pobrien at orbtech.com> wrote:
>
>And since Emacs has a Vi mode, maybe we'll even get you using
>PyAlaMode.  ;-)

Blech.  The reason I stick with vi is because I can walk up to any
system (well, other than Windoze) and start working.  Bare emacs is
unusable.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz, c.l.py, 2/4/2002



From duncan at NOSPAMrcp.co.uk  Thu Apr 10 09:54:20 2003
From: duncan at NOSPAMrcp.co.uk (Duncan Booth)
Date: Thu, 10 Apr 2003 13:54:20 +0000 (UTC)
Subject: class too big for one file! How can I split it?
References: <e956b5f1.0304100512.597654d3@posting.google.com>
Message-ID: <Xns9359977DE4BCduncanrcpcouk@127.0.0.1>

meinmailfilter at gmx.de (HHaegele) wrote in 
news:e956b5f1.0304100512.597654d3 at posting.google.com:

> Now, OTLib.py has more then 32000 lines. And that doesn?t work. Python
> 1.5 can?t execute then. So, how can I add the method Btt()? How can I
> make another file, which knows the global XX and who can be handled
> from the ot_starter, like it is in one file? Good would also be, if
> the Btt() can access XX the same way like Att() can. Is that possible?
> Or how must I change the structure?

I strongly recommend you split your class into several smaller classes. 
However, if that really isn't possible, then you could use the execfile 
function. The example below executes each of the included files as though 
it were indented in the class definition even though it is a separate file.

You will get a performance hit on startup though as the included files will 
be compiled every time you run the program instead of being cached in the 
.pyc file.

OTLib.py:
---------------
XX = None

class Ot:
    def __init__(self, a = None, e = None):
        global XX
        XX = e    
    execfile("OTLib_Att.py")
    execfile("OTLib_Btt.py")
---------------

OTLib_Att.py
---------------
def Att(self):
    global XX
    print XX
---------------
OTLib_Btt.py
---------------
def Btt(self):
    global XX
    print XX
---------------

Also though, if I were you I'd drop the global and just make XX a class 
variable you can refer to as Ot.XX, or self.XX if not changing the value. 
If you do that then you can simply define the extra methods in another 
module and assign them into the class. This gets round the compilation hit 
that I described above.

-- 
Duncan Booth                                             duncan at rcp.co.uk
int month(char *p){return(124864/((p[0]+p[1]-p[2]&0x1f)+1)%12)["\5\x8\3"
"\6\7\xb\1\x9\xa\2\0\4"];} // Who said my code was obscure?



From exarkun at intarweb.us  Thu Apr  3 16:56:23 2003
From: exarkun at intarweb.us (Jp Calderone)
Date: Thu, 3 Apr 2003 16:56:23 -0500
Subject: distributed computing implementations
In-Reply-To: <a626bbd7.0304031149.7abe84c0@posting.google.com>
References: <a626bbd7.0304020829.1e5cac2f@posting.google.com> <b6f66v$28k$1@slb3.atl.mindspring.net> <v8m9a4nk4dpjf7@corp.supernews.com> <a626bbd7.0304031149.7abe84c0@posting.google.com>
Message-ID: <20030403215623.GA7416@meson.dyndns.org>

On Thu, Apr 03, 2003 at 11:49:05AM -0800, robin wrote:
> claird at lairds.com (Cameron Laird) wrote:
> 
> > SOAP and such are just concessions to commercial misunder-
> > standings about what business needs.
> 
> I would like to know more about what you mean by this.
> 
> Jp Calderone <exarkun at intarweb.us> wrote:
> 
> > Personally, I think SOAP is worth ignoring.  ;)  XML-RPC is what I would
> > choose if I were going for that sort of solution, too.
> 
> I'd like more details here too. Why ignore SOAP?

  It's XML-based.  This is a major design flaw for a wire protocol.  The
only reason XML-RPC gets off the hook here is that it was first (afaik.  And
if not, it's at least the most popular), and you're allowed to make a
mistake once >:)

  It's HTTP-based.  HTTP is almost sufficient for web pages (not if you want
really dynamic content though, then it pretty much sucks).  As an RPC
transport, it's a flop.  Luckily, I don't -think- SOAP is completely tied to
HTTP, it's just a too-commonly used transport.

  In addition to these two bad design choices, which probably could have
been worked around, it has a generally high per-message overhead (-even- for
XML).  Here's an example of a message, directly from the standards document:

    POST /StockQuote HTTP/1.1                                                                                                            
    Host: www.stockquoteserver.com                                                                                                       
    Content-Type: text/xml; charset="utf-8"                                                                                             
    Content-Length: nnnn                                                                                                                
    SOAPAction: "Some-URI"                                                                                                              
                                                                                                                                
    <SOAP-ENV:Envelope                                                                                                                
      xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"                                                                       
      SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">                                                               
       <SOAP-ENV:Body>                                                                                                       
           <m:GetLastTradePrice xmlns:m="Some-URI">                                                                                      
               <symbol>DIS</symbol>                                                                                                    
           </m:GetLastTradePrice>                                                                                                        
       </SOAP-ENV:Body>                                                                                                                 
    </SOAP-ENV:Envelope>                                                                                                              


  This message is asking for the last price "DIS" traded at.  It is almost
2kB long.

  That's the short of it.

  Jp

-- 
"There is no reason for any individual to have a computer in their
home."
                -- Ken Olson, President of DEC, World Future Society
                   Convention, 1977
-- 
 up 14 days, 18:00, 5 users, load average: 0.56, 0.34, 0.22




From carroll at tjc.com  Tue Apr 15 17:52:39 2003
From: carroll at tjc.com (carroll at tjc.com)
Date: Tue, 15 Apr 2003 14:52:39 -0700
Subject: O'Reilly Python Books
References: <12240-3E9C5F83-187@storefull-2371.public.lawson.webtv.net>
Message-ID: <30uo9vo46k53v4to6d65koh72nqf0p8nvn@4ax.com>

On Tue, 15 Apr 2003 15:37:39 -0400 (EDT), wilfred_s at webtv.net (W S)
wrote:

>Hello,
>
>I find the material, as presented, in the book Programming Python
>frustrating and extremely difficult to understand.
>
>Are the nutshell and cookbook O'Reilly Python books easier to read?

I really dislike the "Programming Python" book.[1]  After playing
around unstructuredly with Python for a few days.  I bought it and the
New Riders "Python Essential Reference."  I found the PET valuable
(although it's a reference, not a tutorial), and didn't find
"Programming Python" worthwhile at all.

"Learning Python," (which is co-authored by author of "Programming
Python,") on the other hand, is, in my opinion, excellent.

My recommendation is to start with two books, one as a tutorial, and
one as a reference.  I think "Learning Python" is outstanding as a
tutorial.  Since you have that already, I'd work through that.  I
don't offhand know the contents of the other two books you have ("Core
Python" and "Python Bible"), although I've browsed them in bookstores.
>From memory, I think they would suffice for good references.  The
Nutshell also would.   For me, it's the Python Essential Reference.

When you're comfortable with basic Python, maybe you want further
books that better explain a particular area that you find of interest
(say, XML or Tkinter for those features, or the Cookbook for general
advanced topics).

I've been playing with Python for a few months now (which you can take
either as a warning that I might not know what I'm talking about, or
as data that I'm close to the same audience as you are), and just
ordered the Nutshell and Cookbook a couple days ago.

[1] My major objection to Programming Python is that 1) part of it's
too advanced to be a good tutorial; 2) it's not properly organized to
be a good reference; and 3) too many sections of it depend on having
read all the prior sections, which is somewhat acceptable for a
tutorial, but really falls short for a book where you're going to skip
around -- the examples aren't very self-contained.  It seems you have
to feign interest in the packer application that is introduced at the
beginning of the book, and follow through how it's developed, or else
you won't understand what the book is talking about when it starts
talking about adding menus.  Too much of a time investment reading all
that unnecessary information, just to follow parts of the example that
have nothing to do with menus.  It seems the whole book operates more
as a case study, requiring an inordinate amount of time to learn the
case (which I don't care about) in order to learn about the features
tried to be explained.

In fairness, here's a review that just *loves* "Programming Python,"
calling it "a work of art": http://www.awaretek.com/programpython.html
So opinions can vary.  But when I gave my views of "Programming
Python" on the Python Tutor list, I got a lot of "Me, too" posts.  I
suspect that, to someone familiar with Python, it looks like it would
be a useful book to a novice; the actual novices seem to disagree,
though.

(The "case study" issue I describe above, which I consider a "bug,"
the above review considers to be a feature: "Many of the example
scripts and programs build on one another from chapter to chapter,
which helps to give the book a narrative flavor and hold the reader's
attention."  I would say "many of the example scripts and programs
build on one another from chapter to chapter, making it difficult to
isolate and understand a feature being introduced, without going back
and reading or rereading prior material in which the reader has no
interest.")

I was really expecting "Programming Python" to be to Python what
"Programming Perl" is to Perl: a pretty good reference, with lots of
(independent!) examples and occasional descents into tutorial mode.





From peter at engcorp.com  Wed Apr 23 22:46:40 2003
From: peter at engcorp.com (Peter Hansen)
Date: Wed, 23 Apr 2003 22:46:40 -0400
Subject: [Question] I want to know what script more speedy....  VB script and 
 Python Script...
References: <b87h29$bju$1@news1.kornet.net>
Message-ID: <3EA75010.D9F9C857@engcorp.com>

"??????" wrote:
> 
> Plz... answer me...

It sort of depends on the type of application.

Please read http://www.catb.org/~esr/faqs/smart-questions.html
and try posting again...

-Peter



From ckea25d02 at sneakemail.com  Mon Apr 14 22:53:00 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Tue, 15 Apr 2003 14:53:00 +1200
Subject: Using __repr__ or __str__ for own printable class?
In-Reply-To: <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net>
References: <mailman.1050200792.6949.python-list@python.org> <mailman.1050253975.31483.python-list@python.org> <b7di0n$dd6gj$1@ID-169208.news.dfncis.de> <h6o2e6ubks0j.1uimq2uxys0rx.dlg@40tude.net>
Message-ID: <b7frhv$gaol$1@ID-169208.news.dfncis.de>

Mads Orbesen Troest wrote:
> But then one wouldn't be able to do a standard "print" of the class, would 
> one?

No, you'd have to say

   print myStringRep(someInstanceOfMyClass)

But this is only an issue if you're printing a built-in
container such as a list. You could just as well make
the __str__ of MyClass do whatever myStringRep does.
However, if any of the instance variables of MyClass
are built-in containers, you would have to not simply
str() them, but treat them more carefully.

 > I would like to be able to  do this ...
 >
> 	print someInstanceOfMyClass
> 
> This works with __str__ but 
> I was/am afraid if this overriding of __str__ will conflict with other uses 
> of __str__ and/or __repr__.

If it works for you, then that's fine. The only other
built-in use of __str__ that I can think of is the
str() function. If you're happy for anyone who calls
str() on your class to get the same result, then
you don't need to worry.

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From dale at riverhall.NOTHANKS.co.uk  Tue Apr 22 15:55:13 2003
From: dale at riverhall.NOTHANKS.co.uk (Dale Strickland-Clark)
Date: Tue, 22 Apr 2003 20:55:13 +0100
Subject: Off topic: What usenet news reader to use?
References: <mailman.1051036295.31460.python-list@python.org>
Message-ID: <d97bavkeagbpfkvm3649725tptqhkskds9@4ax.com>

"Wai-Yip Tung \(wtung\)" <wtung at cisco.com> wrote:

>I haven't used Usenet news for a long time. Now that I have subscribed
>to python-list I find that email does not scale well to the high volumne
>of discussion. Instead of deleting emails all day I think I should go
>back to Usenet news. Is there any good Windows news reader you would
>recommend. I'm happily using Opera as the browser. But it's
>revolutionary news and email reader eludes me. What are you using?
>
>By the way I'm also looking for a free or low cost email client. My
>company is using Outlook but I want to find one for personal use. Thank
>you for you suggestions.

You don't say whether you're interested in an offline or online reader
so here's both. In no particular order, have you tried:

* Outlook Express
* Free Agent
* Mozilla
* XNews
* And what's wrong with the one in Opera 7.10?

-- 
Dale Strickland-Clark
Riverhall Systems Ltd



From db3l at fitlinxx.com  Tue Apr 29 13:00:41 2003
From: db3l at fitlinxx.com (David Bolen)
Date: 29 Apr 2003 13:00:41 -0400
Subject: Visual Basic - Python compare
References: <mailman.1051603275.12037.python-list@python.org> <1bf4033f.0304290627.40f7d831@posting.google.com>
Message-ID: <uptn56xxi.fsf@fitlinxx.com>

fcslka at yahoo.com (Kent) writes:

> 7.  In the interest of being honest, Python support in development
> tools is weak.  There are many IDEs for Python, most of which are
> immature, or are Python-only.  What needs to happen is for full
> support of Python to be built into existing IDEs.  There's Visual
> Studio .NET support, but boy, it's expensive.  NetBeans and Eclipse
> could really help us out here if we could get more than just cursory
> support for Python in those otherwise great environments.

At the risk of over-emphasizing this aspect, I do think that the OP
needs to consider this when it comes to replacing a VB environment
with Python, particularly when it sounds like the application is
already fairly well separated into back-end logic (the C++ COM
objects) and front end UI (the VB code).

I'd completely agree that in a pure language comparison, there's very
little to recommend VB over Python.  However, to the extent that you
are really comparing the RAD for VB development and/or its ability to
design and implement the UI (and only the UI) for a system, the
comparison is a lot closer.

We're at the point of making some similar decisions for our upcoming
projects.  In our case we've got existing VB licenses, and budget
constraints, so fair or otherwise, a switch to Python is difficult if
it involves too much cost.  I can break budgets if I need to, but I'm
probably hardest on myself in terms of justifying it.

We're definitely moving to Python as our core language for the new
projects, and Python will be used to implement all core code and
business logic.  However, there's still the question about the pure-UI
presentation layer.  I've kept abreast of the IDE and UI development
environments for Python over the past few years, and to be honest, at
the moment I've come to the conclusion that today it's still difficult
to come up with a GUI layer RAD for Python that is comparable to VB.

Even without trying to move to VB.NET, the existing VB 6 environment
within Visual Studio is darn nice for screen layout and basic forms
construction.  Our developers are already experienced with it, and to
be honest, I'm not sure a Python solution would improve our
productivity or end result that dramatically for the UI layer.

When it comes to Python UI, I have yet to find an obvious solution
over VB, if you want a RAD and a screen designer.  There just aren't
that many RAD tools (even including commercial), particularly at
reasonable cost.

I happen to be in the wxPython camp and that's what we've used for all
Python UI projects internally to date (generally manually designing
screens in code, which I'm fine with but which I don't think scales to
larger developments with our development staff).  I expect I'd
seriously consider Qt instead with its excellent QtDesigner, and
particularly with its portability to embedded systems, if it didn't
cost an arm and a leg for commercial development.  Heck, I can buy
multiple MSDN licenses covering multiple languages and tools for the
cost of one Qt license on one OS.  That's a hard pill to swallow.

In terms of Python IDEs, most are code/debug only, no GUI.  In the
wxPython world, BOA Constructor is getting there but still too much of
a moving target to my mind, and you have to like the Delphi model.
PythonCard may be promising but is also early on.  wxDesigner is nice,
but not integrated into a more general IDE.  For Qt, I was initially
excited about BlackAdder on the commercial front because of it's
inclusion of a Python Qt and designer at reasonable cost, but it feels
like a dead project to me at this point (still beta after how long?).
For Tk, Pythonworks is interesting, but I just don't prefer the look
of Tk of under Windows.

My leading contender for a VB replacement was to use wxDesigner
(reasonably priced) for screen design for wxPython.  But to be honest,
it's still almost a wash if I keep the UI code strictly presentation,
so it's unclear that it's worth moving away from VB at this point.
It is, however, my general plan should I want to move in that direction.

That's not to say that it would be a bad idea to replace VB with
Python in a project.  But to the extent that VB is being used solely
as a presentation layer and screen management, I think the tradeoffs
are much closer than might be suggested by an overall language
comparison, since the environment plays a substantially larger part in
the productivity of the system.

--
-- David
-- 
/-----------------------------------------------------------------------\
 \               David Bolen            \   E-mail: db3l at fitlinxx.com  /
  |             FitLinxx, Inc.            \  Phone: (203) 708-5192    |
 /  860 Canal Street, Stamford, CT  06902   \  Fax: (203) 316-5150     \
\-----------------------------------------------------------------------/



From fredrik at pythonware.com  Fri Apr 25 16:01:26 2003
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 25 Apr 2003 22:01:26 +0200
Subject: GUI libs
References: <MB7qa.25958$DN.734132@tornado.fastwebnet.it> <b8b1k3$t5f$00$1@news.t-online.com> <Dl9qa.26067$DN.734905@tornado.fastwebnet.it> <vai9f3ug54144@corp.supernews.com> <3ea947e1$0$198$a1866201@newsreader.visi.com>
Message-ID: <b8c44q$jni$1@main.gmane.org>

Grant Edwards wrote:

> Tkinter is indeed a pretty good place to start -- just try not
> to think about the fact that you're pulling in a Tcl interpreter
> also.  ;)

I do find it a bit worrying that every time I start a Python program,
it pulls in a C runtime library.  Guess I have to do something about
that...

</F>







From paddy3118 at netscape.net  Mon Apr 28 17:02:19 2003
From: paddy3118 at netscape.net (Paddy McCarthy)
Date: 28 Apr 2003 14:02:19 -0700
Subject: Slice confusion : a[n:p] is a list exclude the last  element p
References: <mailman.1051532982.25352.python-list@python.org> <UXara.22428$K35.745468@news2.tin.it>
Message-ID: <2ae25c6b.0304281302.46cac07f@posting.google.com>

Alex Martelli <aleax at aleax.it> wrote in message news:<UXara.22428$K35.745468 at news2.tin.it>...
<<SNIP>>
> For example, for all XX, x, y, z, XX[x:y]+XX[y:z] == XX[x:z].  This
> kind of regularity is exactly what facilitates reasoning about
> programs, and in particular modifying programs.
> 

For x<=y<=z surely.

<<SNIP>>
> Alex

<example>

>>> XX=range(10)
>>> XX
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> x,y,z = 1,4,7
>>> print XX[x:y] +XX[y:z], XX[x:z]
[1, 2, 3, 4, 5, 6] [1, 2, 3, 4, 5, 6]
>>> y
4
>>> y = -2
>>> print XX[x:y] +XX[y:z], XX[x:z]
[1, 2, 3, 4, 5, 6, 7] [1, 2, 3, 4, 5, 6]
>>> XX[y:z]
[]
>>> 

</example>

But I agree with Alex, it turned out to be less error prone for me if
I adopted the inclusive lower bound, non-inclusive upper bound way of
doing things.
It may look odd at first but you get used to it and the array bounds
maths I find easier.

Cheers, Paddy.



From tjreedy at udel.edu  Tue Apr 29 12:48:48 2003
From: tjreedy at udel.edu (Terry Reedy)
Date: Tue, 29 Apr 2003 12:48:48 -0400
Subject: Problem subclassing tuple
References: <mailman.1051631814.18940.python-list@python.org>
Message-ID: <2MmdnZmgXowvMTOjXTWcpA@comcast.com>

"John Wilson" <tug at wilson.co.uk> wrote in message
news:mailman.1051631814.18940.python-list at python.org...
> Given this class:
>
> class Holder(tuple):
>     def __init__(self, *values):
>         tuple.__init__(self, values)
>
>
> When I try to create an instance:
>
> Holder(1, 2)
>
> I get:
>
> TypeError: tuple() takes at most 1 argument (2 given)
>
> It looks like the constructor for the superclass is being called
rather than
> the __init__ method on Holder.
>
> I'm obviously missing something here. Do I have to supply a __new__
method?
> Any help would be appreciated.


Because tuples are immutable, one cannot mutate them in an init
method, even to set the initial values.  So, as you apparently
suspected, they are initialized in the __new__ method from a sequence.
This is the source of the exception (see below). tuple.__init__ should
be a no op.

>>> class Holder(tuple):
...      def __init__(self, *values):
...          print "Holder init entered"
...
>>> Holder(1,2)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
TypeError: tuple() takes at most 1 argument (2 given)
# init not entered
>>> Holder((1,2))
Holder init entered
(1, 2)

While Holder.__init__() cannot modify the value of self, it *can* add
attributes (which surprised me a bit) so it can have a use.

>>> class Holder(tuple):
...     def __init__(self, seq):
...         self.y = 1
...
>>> a=Holder((1,2))
>>> a.y
1

I don't know what more you could do in a __new__ method.

Terry J. Reedy







From stephen.boulet at motorola.com  Tue Apr  1 09:58:07 2003
From: stephen.boulet at motorola.com (Stephen Boulet)
Date: Tue, 01 Apr 2003 08:58:07 -0600
Subject: Regular expression help
References: <2FGdnaUSmIhkghSjXTWc-g@speakeasy.net>
Message-ID: <3E89A8FF.4020007@motorola.com>

Stephen Boulet wrote:
> How do I use the sub function to replace my match with my match 
> prepended by a newline?
> 
> Thanks.
> 
> -- Stephen
> 

Thanks to all who answered. I learned stuff.

Here's the final script that formats text on the windows clipboard 
(cleanDates.pyw):

import win32clipboard,sys,re

# Get current string from clipboard. If string empty, error out.
win32clipboard.OpenClipboard(0)
if win32clipboard.IsClipboardFormatAvailable(win32clipboard.CF_TEXT):
	s = win32clipboard.GetClipboardData(win32clipboard.CF_TEXT)
else:
	from tkMessageBox import showerror
	showerror('Error!','No text on clipboard')
	sys.exit()

# Fix current string
p = re.compile(r'\d+/\d+/\d+')

def mysub(x):
	return '\n\n' + x.group(0)

s = p.sub(mysub, s)

# Paste string to clipboard
win32clipboard.OpenClipboard(0)
win32clipboard.EmptyClipboard()
win32clipboard.SetClipboardText(s.strip())
win32clipboard.CloseClipboard()




From cybersamurai at mac.com  Mon Apr  7 13:09:24 2003
From: cybersamurai at mac.com (cybersamurai at mac.com)
Date: Mon, 7 Apr 2003 14:09:24 -0300
Subject: selecting a class in execution time
Message-ID: <AE59EE7E-691B-11D7-8CA9-000393B10A78@mac.com>

have some way to do this on Python?

Object x = Class.forName( "mypackage.TheClass" ).newInstace();
(this is a java code)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 224 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20030407/d55836f6/attachment.bin>

From tebeka at cs.bgu.ac.il  Wed Apr  9 10:52:45 2003
From: tebeka at cs.bgu.ac.il (Miki Tebeka)
Date: 9 Apr 2003 07:52:45 -0700
Subject: wxPython - Can you explain why I crash
Message-ID: <33803989.0304090652.174ac9de@posting.google.com>

Hello All,

I've written this little client. It used to work fine but now no
matter which button I press the application crash.
I'm running Python2.2 and wxPython 2.4.0.7
Any ideas?

10x.
Miki

--- ui.py ---
#!/usr/bin/env python
'''Compas client emulator UI'''

__author__ = 'Miki Tebeka <miki_tebeka at amat.com>'
__id__ = '$Id: ui.py,v 1.3 2003/03/06 14:20:15 MTebeka90603 Exp $'

# $Source: /u/CVSROOT/compass_client/ui.py,v $

from wxPython.wx import *
from os.path import isfile
from ConfigParser import ConfigParser
import client

# Ids
ID_GET_RECIPES = wxNewId()
ID_SELECT_RECIPE = wxNewId()
ID_GET_WAFERS = wxNewId()
ID_LOAD = wxNewId()
ID_UNLOAD = wxNewId()
ID_RUN = wxNewId()
ID_STOP = wxNewId()
ID_DISCONNECT = wxNewId()
ID_CONNECT = wxNewId()
ID_LAST_ERROR = wxNewId()
ID_STATUSBAR = wxNewId()
ID_SELECT_ALIGN = wxNewId()

# Connection messages
CONNECTED = 'Connected'
DISONNECTED = 'Disconnected'

class CompassClient(wxFrame):
    '''Client frame'''
    def __init__(self):
        # No resizable border
        wxFrame.__init__(self, None, -1, 'CompassSim',
                         style=wxDEFAULT_FRAME_STYLE &
(~wxRESIZE_BORDER))
        self.CreateStatusBar(1, 0, ID_STATUSBAR)
        self.GetStatusBar().SetFieldsCount(2)
        # Get mouse click on task bar
        EVT_LEFT_DOWN(self.GetStatusBar(), self.last_error)

        sizer = wxBoxSizer(wxVERTICAL)
        # Recipe list
        t = wxStaticText(self, -1, 'Recipes:')
        sizer.Add(t)
        self.recipes = wxListBox(self, -1)
        sizer.Add(self.recipes, 0, wxEXPAND)
        hsizer = wxBoxSizer(wxHORIZONTAL)
        self.add_button(ID_GET_RECIPES, 'Read Data', self.get_recipes,
hsizer)
        self.add_button(ID_SELECT_RECIPE, 'Select',
self.select_recipe, \
                        hsizer)
        sizer.Add(hsizer)
        self.add_sep(sizer)
        # Wafer List
        t = wxStaticText(self, -1, 'Wafers:')
        sizer.Add(t)
        self.wafers = wxListBox(self, -1, style=wxLB_EXTENDED)
        sizer.Add(self.wafers, 0, wxEXPAND)

        # Load / Unload
        hsizer = wxBoxSizer(wxHORIZONTAL)
        self.add_button(ID_LOAD, 'Load', self.load, hsizer)
        self.add_button(ID_UNLOAD, 'Unload', self.unload, hsizer)
        sizer.Add(hsizer)
        
        # Run / Stop
        self.add_sep(sizer)
        hsizer = wxBoxSizer(wxHORIZONTAL)
        self.add_button(ID_RUN, 'Run', self.run, hsizer)
        self.add_button(ID_STOP, 'Stop', self.stop, hsizer)
        sizer.Add(hsizer)
        self.times = wxSpinCtrl(self, -1, size=(75,-1))
        self.times.SetRange(1, 999999)
        self.times.SetValue(1)
        sizer.Add(self.times)
        self.add_sep(sizer)

        # Connect / Quit
        hsizer = wxBoxSizer(wxHORIZONTAL)
        self.add_button(ID_CONNECT, 'Connect', self.connect, hsizer)
        self.add_button(-1, 'Quit', lambda e: self.Close(1), hsizer)
        sizer.Add(hsizer)
        
        # Auto resize      
        self.SetSizer(sizer)
        self.SetAutoLayout(1)
        sizer.Fit(self)
        self.connect(None)


    def status(self, success):
        '''Status of last commands'''
        if success:
            self.SetStatusText('OK', 1)
        else:
            self.SetStatusText('Error', 1)

    def busy(self):
        '''Busy state'''
        self.SetCursor(wxHOURGLASS_CURSOR)

    def idle(self):
        '''Idle state'''
        self.SetCursor(wxSTANDARD_CURSOR)

    def add_button(self, id, title, func, sizer, center = 0):
        '''Add a button to sizer'''
        b = wxButton(self, id, title)
        EVT_BUTTON(self, id, func)
        if center:
            sizer.Add(b, 0, wxALIGN_CENTER)
        else:
            sizer.Add(b)
    
    def add_sep(self, sizer):
        '''Add a seperator to sizer'''
        sizer.Add(0, 5, 0, wxEXPAND)

    def error(self, msg):
        '''Show error message'''
        dlg = wxMessageDialog(self, msg, 'Error', \
                              wxOK | wxICON_EXCLAMATION)
        dlg.ShowModal()
        dlg.Destroy()

    def last_error(self, e):
        '''Show last error'''
        err = client.LAST_ERROR
        if not err:
            err = 'Unknown Error'
        dlg = wxMessageDialog(self, err, 'Last Error', \
                              wxOK | wxICON_INFORMATION)
        dlg.ShowModal()
        dlg.Destroy()
    
    def get_recipes(self, e):
        '''Get recipe list'''
        print 'get_recipes'
        try:
            self.busy()
            if not client.get_lot_data():
                self.status(0)
                return
            recipes = client.get_recipe_list()
            if not recipes:
                self.status(0)
                return
            # Update listbox
            self.recipes.Clear()
            for recipe in recipes:
                self.recipes.Append(recipe)
            self.status(1)
        finally:
            self.idle()
        
    def select_recipe(self, e):
        '''Select a recipe'''
        try:
            self.busy()
            if self.recipes.GetCount() == 0:
                self.error('No Recipes')
                return
            if not client.select_recipe(self.recipes.GetStringSelection()):
                self.status(0)
                return
            wafers = client.get_wafer_list()
            if not wafers:
                self.status(0)
                return
            # Update listbox
            self.wafers.Clear()
            for wafer in wafers:
                self.wafers.Append(wafer)
            self.status(1)
        finally:
            self.idle()
    
    def _load(self, wafer):
        # set_alignment_parameters must bee called before load
        # Image must have /work/wf prefix (see stub_functions.cpp
AlignData)
        params = ['/work/wf/coarse.img', 'TIFF', # Name and format
                  '1.634888', '1.660389', # Pixel size
                  '-264.066777', '-9704.485647', # Center to Coarse
                  '-24.486192', '-21.873070', # Coarse to fine
                  '-21.873070', '5.359366', # Corner to Fine
                  '1'] # Unpatterned
        if not client.set_alignment_parameters(*params):
            client.LAST_ERROR = "Can't set alignment data"
            return 0
        return client.load(wafer)

    def load(self, e):
        '''Load wafer'''
        try:
            self.busy()
            if self.wafers.GetCount() == 0:
                self.error('No Wafers')
                return
            if len(self.wafers.GetSelections()) != 1:
                self.error('Please select only one wafer')
                return
            wafer = self.wafers.GetString(self.wafers.GetSelections()[0])
            self.status(self._load(wafer))
        finally:
            self.idle()
        
    def unload(self, e):
        '''Unload wafer'''
        try:
            self.busy()
            self.status(client.unload())
        finally:
            self.idle()

    def run(self, e):
        '''Run process for given number of iterations'''
        try:
            self.busy()
            wafers = map(lambda i: self.wafers.GetString(i), \
                                   self.wafers.GetSelections())
            for t in range(self.times.GetValue()):
                for wafer in wafers:
                    if not self._load(wafer):
                        self.status(0)
                        return
                    if not client.start_process():
                        self.status(0)
                        return
                    if not client.unload():
                        self.status(0)
                        return
            self.status(1)
        finally:
            self.idle()

    def stop(self, e):
        '''Stop process'''
        try:
            self.busy()
            self.status(client.stop())
        finally:
            self.idle()

    def connect(self, e):
        '''Connect to tcp server'''
        try:
            self.busy()
            if not client.connect():
                self.connected = 0
                self.SetStatusText(DISONNECTED)
                return
            self.SetStatusText(CONNECTED)
        finally:
            self.idle()


def main():
    app = wxPySimpleApp()
    frame = CompassClient()
    frame.Show(1)
    app.MainLoop()

# Main
if __name__ == '__main__':
    main()

--- ui.py ---

--- client.py ---

'''Compas client emulator'''

__author__ = 'Miki Tebeka <miki_tebeka at amat.com>'
__id__ = '$Id: client.py,v 1.5 2003/03/06 14:20:15 MTebeka90603 Exp $'

# $Source: /u/CVSROOT/compass_client/client.py,v $

from socket import error as SocketError, socket, AF_INET, SOCK_STREAM
import re
from xml.dom.minidom import parseString
from sys import exit

SOCKET = None # Global socket
HOST = 'localhost' # Host name
PORT = 3333 # Port
LAST_ERROR = '' # Last error

def header(name):
    '''Create header xml'''
    return '<Header To="BEE" From="WF" Type="Request" Name="%s"/>' %
name

def connect():
    '''Connect to tcp server'''
    global SOCKET, LAST_ERROR
    SOCKET = socket(AF_INET, SOCK_STREAM)
    try:
        SOCKET.connect((HOST, PORT))
        return 1
    except SocketError, e:
        LAST_ERROR = str(e)
        SOCKET = None
        return 0

def disconnect():
    global SOCKET
    SOCKET = None

def is_connected():
    global SOCKET
    return SOCKET != None

def _send(msg):
    '''send "len msg"'''
    SOCKET.sendall('%d %s' % (len(msg), msg))

# Regular expressions
NUM_RE = re.compile('\d')
WS_RE = re.compile('\s')

def _read():
    '''Read message. Return error + extra'''
    size = ''
    while 1:
        d = SOCKET.recv(1)
        if not NUM_RE.match(d):
            break
        size += d
    return _parse_respone(SOCKET.recv(int(size)))

class Error:
    '''Error class'''
    def __init__(self, value = 0, message = ''):
        self.value = value # Error value
        self.message = message # Error message

    def __repr__(self):
        return 'Error (%d): %s' % (self.value, self.message)

    def ok(self):
        '''True not error occured'''
        return self.value == 0

def _parse_respone(msg):
    '''Parse a respone. Return error + extra'''
    global LAST_ERROR
    doc = parseString(msg)
    datanode = doc.getElementsByTagName('Data')[0]
    assert(len(datanode.childNodes) in [1,2])
    errnode = None
    for node in datanode.childNodes:
        if node.nodeName == 'Error':
            errnode = node
            datanode.removeChild(node)
            break
    assert(errnode)
    attrs = errnode.attributes
    err = Error(int(attrs['ErrNum'].value), attrs['ErrText'].value)
    LAST_ERROR = err.message
    if len(datanode.childNodes):
        extra = datanode.childNodes[0]
    else:
        extra = None
    return err, extra

def _simple_cmd(msg):
    '''Simple command'''
    if not SOCKET:
        return 0
    _send(msg)
    err, extra = _read()
    return err.ok()

def _list_cmd(msg):
    '''Command that returns a list'''
    if not SOCKET:
        return 0
    _send(msg)
    err, extra = _read()
    if not err.ok():
        return []
    else:
        return map(lambda n: n.nodeValue, extra.attributes.values())

def _state_cmd(msg):
    '''Command that returns a state'''
    if not SOCKET:
        return 0
    _send(msg)
    err, extra = _read()
    if not err.ok():
        return -1
    return int(extra.attributes['Value'].nodeValue)
    
# Interface functions    
def get_lot_data():
    '''Get lot data'''
    return _simple_cmd(header('GetLotData'))

def get_recipe_list():
    '''Get recipe list'''
    print 'get_recipe_list'
    return _list_cmd(header('GetRecipeList'))

def select_recipe(recipe):
    '''Select a recipe'''
    msg = header('SelectRecipe')
    msg += '<Data><Recipe Name="%s"/></Data>' % recipe
    return _simple_cmd(msg)

def get_wafer_list():
    '''Get wafer list'''
    return _list_cmd(header('GetWafersList'))

def set_alignment_parameters(*params):
    '''Set alignment parameters'''
    assert(len(params) == 11)
    msg = header('AlignData') + '''
<Data>
    <CoarseImage Name="%s" Format="%s"/>
    <CoarsePixelSize psX="%s" psY="%s"/>
    <CenterToCoarse CntrToCrsX="%s" CntrToCrsY="%s"/>
    <CoarseToFine CrsToFineX="%s" CrsToFineY="%s"/>
    <CornerToFine CrnToFineX="%s" CrnToFineY="%s"/>
    <WaferType UnPatternedWafer="%s"/>
</Data>''' % params
    return _simple_cmd(msg)

def prepare_to_load(slot):
    '''Prepare to load'''
    msg = header('PrepareToLoad')
    msg += '<Data><Slot SlotID="%s"/></Data>' % slot
    return _simple_cmd(msg)

def load_ended():
    '''Load ended'''
    return _simple_cmd(header('LoadEnded'))

def load(slot):
    '''Load sequence'''
    if not prepare_to_load(slot):
        return 0
    return load_ended()

def start_process():
    '''Start process'''
    return _simple_cmd(header('StartProcess'))

def prepare_to_unload():
    '''Prepare to unload'''
    return _simple_cmd(header('PrepareToUnload'))

def unload_ended():
    '''Unload ended'''
    return _simple_cmd(header('UnloadEnded'))
    
def unload():
    '''Unload sequence'''
    if not prepare_to_unload():
        return 0
    return unload_ended()

def get_state():
    '''Get state'''
    return _state_cmd(header('GetState'))

def get_vacuum_state():
    '''Get vacuum state'''
    return _state_cmd(header('GetVacuumState'))

def stop():
    '''Stop'''
    return _simple_cmd(header('bee_stop_process'))

--- client.py ---



From linux42 at freemail.c3.hu  Fri Apr 11 07:57:15 2003
From: linux42 at freemail.c3.hu (Gabor Nagy)
Date: Fri, 11 Apr 2003 13:57:15 +0200
Subject: win32 COM test question
Message-ID: <20030411115715.GB1858@swordfish.capgemini.hu>

Hi!

I have downloaded the win32all extension from
http://starship.python.net/crew/mhammond/win32/Downloads.html

I have followed the description in
http://www.python.org/windows/win32com/QuickStartServerCom.html

and created a python script, which when ran, registered my test server.

So far, so good.

Now I'd like to test my class from anything but python. On the above
mentioned page the "Testing our Class" paragraph says this:

For the purposes of this demonstration, we will test the class using
Visual Basic. This code should run under any version of Visual Basic,
including VBA found in Microsoft Office. Any COM compliant package could
be used alternatively.
This is not a tutorial in VB. The code is just presented! Run it, and it
will work!

The problem is, there is no code on the page, and I don't know how to call
the exported functions of my new class

So could someone write me the missing VB example? (I don't know VB or COM,
so I couldn't figure it out myself, but I simply want to put this call in
an existing VB program)

Regards,
Gabor




From gregryman at taekwondotimes.com  Mon Apr  7 15:05:55 2003
From: gregryman at taekwondotimes.com (Greg Ryman)
Date: Mon, 7 Apr 2003 14:05:55 -0500
Subject: TaeKwonDo Times Magazine Offer
Message-ID: <B0007837728@mail.prairieinet.net>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20030407/5cbd2832/attachment.html>

From maxm at mxm.dk  Fri Apr  4 01:54:30 2003
From: maxm at mxm.dk (Max M)
Date: Fri, 04 Apr 2003 08:54:30 +0200
Subject: mp3 or wave  ---> midi , in python , how?
In-Reply-To: <b6j6sl$5m5dh$1@ID-169208.news.dfncis.de>
References: <b6isqq$23hs$1@mail.cn99.com> <b6j6sl$5m5dh$1@ID-169208.news.dfncis.de>
Message-ID: <7Y9ja.2675$y3.451371@news010.worldonline.dk>

Greg Ewing (using news.cis.dfn.de) wrote:
> Frank Zheng wrote:

> My first thought was that this sounds almost impossible,
> Python or no. 


You should try and download the demo of melodyne, you will be amazed and 
amused for the better part of a day ;-)

-- 

hilsen/regards Max M Rasmussen, Denmark

http://www.futureport.dk/
Fremtiden, videnskab, skeptiscisme og transhumanisme




From paul at boddie.net  Thu Apr 10 07:52:36 2003
From: paul at boddie.net (Paul Boddie)
Date: 10 Apr 2003 04:52:36 -0700
Subject: Nice quote
References: <just-60BB53.18302728032003@news1.news.xs4all.nl> <pan.2003.03.28.20.18.26.483935@jerf.org> <3E8597A2.13FE92DE@engcorp.com> <pan.2003.03.30.10.38.29.454614@jerf.org>
Message-ID: <23891c90.0304100352.33bc3658@posting.google.com>

"Jeremy Bowers" <jerf at jerf.org> wrote in message news:<pan.2003.03.30.10.38.29.454614 at jerf.org>...
> On Sat, 29 Mar 2003 07:54:58 -0500, Peter wrote:
> > 

[Tim Bray's comments]

> > Basically, he was conveying that Python people don't find processing XML
> > to be a huge problem in any particular way, and I think that's a fair
> > assessment.  IMHO.
> 
> I didn't take his statement that way because it doesn't make sense some to
> me. I've done SAX and DOM in both Perl and Python and they are equal pains
> in the ass.

My interpretation of the "Python people's" presumed statement of
contentment, via Mr Bray's interpretation of that statement, is that
there exists a collection of Python libraries/modules which implement
contemporary XML processing standards to a satisfactory level such
that real work can be done with them. While I remain unsettled by
various versioning issues with PyXML and 4Suite (the main packages for
XML in Python), it is true that they provide decent functionality and,
due to their adherence to standards, are accessible to people like Mr
Bray who come from the wider XML community.

As to whether SAX or DOM are good APIs or not, I personally only
"touch down" on the DOM API when I've made use of other techniques
(notably XPath). Having not read the DOM specification cover-to-cover,
I can't claim to be an authority on DOM, but it always surprises me
that people bring out the "DOM means loading everything at once" when
it is conceivable that implementations could actively avoid such
strategies. That's a side issue, however.

[...]

> The gnosis.xml packages *are* easier then standard DOM processing, and at
> least reduce XML to relatively straight-forward tree processing, though as
> the author of those utilities mentioned, I doubt Tim Bray was thinking of
> those since Tim doesn't know Python. However it's possible that the
> "Python people" mailing Tim may have been thinking of that.

I'm not familiar with gnosis.xml, but it seems to me that non-standard
approaches (particularly the "Pythonic" APIs) would not be of
significant interest to someone who wants to remain close to the
standards. Moreover, from what I've seen of the "Pythonic" element
tree APIs, I don't see anything radically better than the DOM - yes,
the notation may be nicer or cleaner, but there isn't anything obvious
with any of them that provides the benefit that something like XPath
can give, as far as I have seen.

[...]

> I'm still probably going to noodle around more later with a new library
> for Python that will actually construct an object model out of objects you
> provide for it, but it would be a long time before it would be publicly
> available because I'd want to use it in a lot of situations before I could
> be confident it was powerful enough.

My personal belief is that the XML and RDF technologies provide
sufficiently interesting standardised means of querying and
manipulating object hierarchies that one would be best suited by
staying as close to the standardised object models as much as
possible.

Paul



From frobozz_electric at hotmail.com  Thu Apr 10 10:53:03 2003
From: frobozz_electric at hotmail.com (Sean Ross)
Date: Thu, 10 Apr 2003 10:53:03 -0400
Subject: looking for "optimal weighting" algorithm
References: <b73ql8011te@enews4.newsguy.com>
Message-ID: <lzfla.8029$N94.727741@news20.bellglobal.com>

Hi.
It sounds like you're doing a 'Decision Tree' problem, from machine
learning. If this is correct, you can find a Python implementation here:
    http://www.norvig.com/python/learning.html
This contains code for the Decision Tree as well as a few Learner classes.
It looks like your problem will not fit exactly into this framework, but
this code may still provide a basis for your solution. Of course, if this
isn't a Decision Tree problem, then the code won't serve as much of anything
besides light reading ;)
I hope this will have been helpful, and good luck with your project,
Sean Ross


"Alex Martelli" <aleaxit at yahoo.com> wrote in message
news:b73ql8011te at enews4.newsguy.com...
> I _know_ there's a decent algorithm to solve the following problem,
> but I don't recall its name and thus can't rapidly google for the
> details... can somebody help?
>
>
> The problem: I need to design a "decision criterion" to classify
> "observations".  For each observation I measure the values of a
> number N of features, x1, x2, ... xN; the desired design criterion
> is a set of weights w1, w2, ... wN such that for any observation I
> will then just compute a weighted sum
>   S = w1*x1 + w2*x2 + ... + wN*xN
> and classify the observation as Black if S<=1, White if S>1.  To
> train my classifier I have a large corpus of observations already
> made and annotated with "ground-truth" data about whether each
> given observation should have been classified as B or W, and an
> error-cost value for each kind of classification (Ebw is the cost
> of erroneously classifying a feature as W when it should be B,
> Ewb is that of classifying it as B when it should be W).  So,
> what's the algorithm to estimate the weights given the corpus of
> observation and ground-truth data, and the error-costs?
>
> Responses welcome either here or directly to my mailbox -- if I
> do get responses to my mailbox, I'll eventually summarize here --
> as a bonus, I promise to make available a Python implementation of
> whatever algorithm I end up with, if one's not already around...;-).
>
>
> TIA,
>
> Alex
>





From pan-newsreader at thomas-guettler.de  Thu Apr  3 04:09:29 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Thu, 03 Apr 2003 11:09:29 +0200
Subject: Persistent server with GUI interactive functions ?
References: <mgqn8v01utuhhioq5vq32rm2r2oadtrm8i@4ax.com>
Message-ID: <pan.2003.04.03.11.09.29.238820.809@thomas-guettler.de>

On Thu, 03 Apr 2003 10:29:28 +0200, Frank wrote:

> I need to write a GUI app that will serve as a peer to peer
> client/server.  At the highest level it would be similar to a chat
> client, which has a server component (waiting for any incoming
> connections) as well as the client part (initiating a connection,
> executing user commands).

How are your client and server connected? 

Can both create tcp sockets to each other? 
You can't do this if a client is behind a masquarding (NAT)
router. If you can use TCP/IP, I think Pyro is the best choice. I have
not used it yet.

Can the client make a tcp connection to the server? Some networks only
allow http through a proxy. If some clients are behind a http proxy
I would recommend xml-rpc.
 

> Each of those function appears to involve entering an "endless loop" as
> the last logical step to initiate processing (i.e. "mainloop()" for the
> user interface and "serve_forever()" for the server).  I haven't yet
> grasped the logical structure that will allow me to have two
> non-returning calls and maintain control of the program.

I would do it with threads. Maybe there it is possible to do it without 
threads, too.

In the subject you used the word "Persistent". I would use pickles for
development. If you have too much data you can switch to ZODB later.

 thomas

-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de



From staschuk at telusplanet.net  Thu Apr 24 14:00:59 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Thu, 24 Apr 2003 12:00:59 -0600
Subject: Overloading and?  was <RE: Should I prefer an external database>
In-Reply-To: <b87il2$3rf$1@slb6.atl.mindspring.net>; from adalke@mindspring.com on Wed, Apr 23, 2003 at 08:40:49PM -0600
References: <mailman.1051023275.29338.python-list@python.org> <b8420j$629$1@slb3.atl.mindspring.net> <mailman.1051043856.10118.python-list@python.org> <b86gim$60r$1@slb5.atl.mindspring.net> <mailman.1051127503.25664.python-list@python.org> <b87il2$3rf$1@slb6.atl.mindspring.net>
Message-ID: <20030424120059.B252@tibia.amotlpaa.bogus>

Quoth Andrew Dalke:
  [...overloadable, short-circuiting operators...]
> Ahhh, a nit-picker, eh?  :)
> 
> Let's see.  I said:
> ] There's no way to implement that behaviour in Python, so
> ] it cannot be overridden.
> 
> I submit that a language which worked in your suggested manner
> would not be Python.
> 
> The nit is now in your court!

Hm.  I'm not sure I understand what would make such a language not
Python.

I agree, of course, that no known Python has such special __and__
and __or__ methods.  But that no *known* Python works as I
suggested does not mean that no *possible* Python works thus.
(Consider that not long ago the yield statement was not part of
any known Python.  And imho generators are a much larger change to
the language than two new special methods with unusual arguments.)

-- 
Steven Taschuk              Aral: "Confusion to the enemy, boy."
staschuk at telusplanet.net    Mark: "Turn-about is fair play, sir."
                             -- _Mirror Dance_, Lois McMaster Bujold




From tim at bladerman.com  Tue Apr  8 05:38:59 2003
From: tim at bladerman.com (Tim Gahnström /Bladerman)
Date: Tue, 08 Apr 2003 09:38:59 GMT
Subject: Datagrid-Widget for Python?
References: <ibc39v0igd6chg8egcauum1m0b68b2t9hi@4ax.com> <urnka.4658$Du.16154@newsc.telia.net> <i6059vgtvavjp99ilkfeo8rtc003mha57l@4ax.com>
Message-ID: <TMwka.4685$Du.16488@newsc.telia.net>

"Rainer Klueting" <rainer.klueting at epost.de> wrote in

> I'm just learning Python; I dropped onto it about a week ago and had
> the strong impression that it might be what I was looking for. :-)

You were probably right there :-)

> I had already seen that there is another GUI toolkit called wxWindow,
> and I even had downloaded it. But my impression was, that Tkinter is
> kind of a GUI-standard for Python, and I tried to stick to the
> standard.

> Is that true? Do you have any information about how widespread the use
> of wxWindows is? I don't want to steer into a dead-end-road.

My feeling is that tkinter is a little more widspread since it is in the
distribution but I am not sure.
Have a look at this thread ,or search google groups for tkinter wxPython and
you will find a lengthy discussion about wich one is the best.
http://groups.google.com/groups?hl=sv&lr=&ie=UTF-8&oe=UTF-8&threadm=mailman.
980465131.18622.python-list%40python.org&rnum=2&prev=/groups%3Fhl%3Dsv%26lr%
3D%26ie%3DUTF-8%26oe%3DUTF-8%26q%3Dtkinter

My personal feelig without having used anyone much is that tkinter is more
"scriptic"  and not as good looking as wxPython wich seems more robust and
always uses native OS widgets.
None is "native python" (whatever that is) but relies on underlying librarys
(tcl/tk and wxWindows). tkinter is probably mor easy to make crossplattform
since it is already in all distributions but now I am in way over my head
and will stop rambling :-)

wxPython is most likely not a dead end anyway.

Tim





From danny at iamlearning.com  Thu Apr  3 12:47:32 2003
From: danny at iamlearning.com (danny at iamlearning.com)
Date: Thu, 3 Apr 2003 11:47:32 -0600 (CST)
Subject: How to compile a python script to *.pyo
In-Reply-To: <3E8B67AD.7AD74C0C@engcorp.com>
References: <mailman.1049309105.22947.python-list@python.org>
        <b6fj57$626$1@news-reader11.wanadoo.fr>
        <3E8B53BA.DBB0C84C@engcorp.com>
        <mailman.1049319846.7316.python-list@python.org>
        <3E8B67AD.7AD74C0C@engcorp.com>
Message-ID: <32609.143.166.224.133.1049392052.squirrel@iamlearning.com>

> danny at iamlearning.com wrote:
>>
>> I want my users run *.pyo or *.pyc files instead of *.py.
>
> To expand on Jp's response a little: please explain why you think
> you want this, so we can better understand the need that's driving it.
> Maybe you are trying to do something you shouldn't be trying
> to do, or maybe you're just going about it the wrong way.  What
> makes you think doing what you are asking is a good thing?
>
> -Peter

I am working in test environment, and I didn't like testers alter my
script (this happened to me). Basicly, I want my scripts are under control
with my knowledge. If the problem come up, I know where to start
troubleshooting.

Thanks,






From staschuk at telusplanet.net  Mon Apr 21 14:18:43 2003
From: staschuk at telusplanet.net (Steven Taschuk)
Date: Mon, 21 Apr 2003 12:18:43 -0600
Subject: Stackless 3.0 alpha 1 at blinding speed
In-Reply-To: <3EA2C43F.2080802@tismer.com>; from tismer@tismer.com on Sun, Apr 20, 2003 at 06:01:03PM +0200
References: <3E9E3040.5040409@tismer.com> <mailman.1050607597.6802.python-list@python.org> <m3he8w9oqo.fsf@mira.informatik.hu-berlin.de> <mailman.1050725072.21332.python-list@python.org> <fcr2avoef04bbrqscqb61fqovduudr0sk8@4ax.com> <b7ti62$ibg$1@news.hccnet.nl> <ruadnZIgGvclOj-jXTWcpA@speakeasy.net> <b7ufg4$o1d$1@news.hccnet.nl> <3EA2C43F.2080802@tismer.com>
Message-ID: <20030421121843.A596@tibia.amotlpaa.bogus>

Quoth Christian Tismer:
  [...]
> Quoting Bjorn Pettersen about continuations:
> 
> """
> They also seem to have the property that before you understand them, no
> explanation will improve your understanding; and after you understand
> them you lose the ability to explain it <wink>.
> """

Hm...

Before studying continuations, function calls are function calls
and loops are loops.  While studying continuations, function calls
are not function calls and loops are not loops.  After studying
continuations, once again function calls are function calls and
loops are loops.

-- 
Steven Taschuk                               staschuk at telusplanet.net
"What I find most baffling about that song is that it was not a hit."
                                          -- Tony Dylan Davis (CKUA)




From nirmalkannan at hotmail.com  Tue Apr 22 07:51:20 2003
From: nirmalkannan at hotmail.com (N.K)
Date: 22 Apr 2003 04:51:20 -0700
Subject: Python Thread Question
References: <bd993a2f.0304170534.245aa2d5@posting.google.com> <mailman.1050587609.6119.python-list@python.org> <b7meoo$mf6$1@panix2.panix.com>
Message-ID: <5b04353b.0304220351.2034a26@posting.google.com>

aahz at pythoncraft.com (Aahz) wrote in message news:<b7meoo$mf6$1 at panix2.panix.com>...
> 
> This isn't quite true on a multi-CPU box.  Because I/O releases the GIL,
> the OS can schedule threads to run on all available CPUs.  So it's
> certainly possible for a threaded app to outrun non-blocking I/O.


  Heard that new Intel Processors support Hyper Threading ( HT)
Architecture from which multi-threaded apps can benefit from. Is that
true?



From mlsj at earthlink.net  Thu Apr 17 22:07:54 2003
From: mlsj at earthlink.net (Josh)
Date: 18 Apr 2003 02:07:54 GMT
Subject: Downloading TV listings with urllib
References: <Xns93607E136EA66jslmAAearthlinknet@130.133.1.4> <mailman.1050602375.2763.python-list@python.org> <E196EHo-0003Gb-00@swan.mail.pas.earthlink.net> <mailman.1050608946.14831.python-list@python.org>
Message-ID: <Xns9360D6FCC737FjslmAAearthlinknet@130.133.1.4>

Gerhard H?ring <gh at ghaering.de> wrote in
news:mailman.1050608946.14831.python-list at python.org: 

> Josh wrote:
>> On 17 Apr 2003, you wrote in comp.lang.python:
>>>Josh wrote: [...]
>>>>        h.putheader('Accept-Encoding:',
>>>>        'gzip,deflate,compress;q=0.9') [...]
>>>>when I use err.read() to read the object returned I get everything
>>>>in Hex. 
>>>
>>>My guess is that you get what you're asking for - gzip-compressed
>>>data. 
>> 
>>  Well I don't want gzipped data. I want HTML so I can parse it and
>>  capture 
>> the listings table into a list or something. Any ideas?
> 
> Don't tell the webserver that you'll accept gzip-compressed data?
> 
> I. e. remove the line I quoted above.
> 
> -- Gerhard
> 

 Thanks Gerhard, it worked. I just removed the 'Accept-Encoding' part and 
it worked.

Josh



From peter at engcorp.com  Thu Apr 24 11:52:07 2003
From: peter at engcorp.com (Peter Hansen)
Date: Thu, 24 Apr 2003 11:52:07 -0400
Subject: OT: losing memories (was Re: How many of you are Extreme Programmers?)
References: <mailman.1050512881.28736.python-list@python.org> <3E9EA707.4DB53582@engcorp.com> <ge+JwjAQ4qn+Ewcd@jessikat.demon.co.uk> <mailman.1050587068.2733.python-list@python.org> <b7o3rn$3hc$1@panix1.panix.com> <3E9FFCB5.79AEA22E@engcorp.com> <slrnbaf7bi.gb9.ramen@lackingtalent.com>
Message-ID: <3EA80827.6277D27C@engcorp.com>

Dave Benjamin wrote:
> 
> Well, I think Dijkstra's basic point was that you can't really "unlearn"
> anything. You can ignore what you've learned, but it's still there. Once
> you've learned some BASIC, it's a part of your brain, and that imprint will
> remain there FOR THE REST OF YOUR LIFE. But it's okay, it hasn't stopped me
> either. Maybe someday we'll start a therapy group.

After having spent the last few months getting on average something like 
five hours of sleep (my need is for more like 7.5 or 8), I'd have to say
I'm a living rebuttal to your argument.  I have completely forgotten a
variety of interesting things which I used to know.  More than that, I've
forgotten *that* I used to know these things, and am going on the say-so
of others whose memories are more intact.  And I strongly doubt I will 
ever recover those memories, or even the memory of the memories.

I think if pathways in the brain can be built up, through repetition of
stimuli, dreaming, etc., then they can likewise be dissolved, whether
through drug use, pyschological problems or chemical imbalances caused 
by lack of sleep, and so on.  Perhaps simple disuse leads only to an
asymptotic decay which never quite reaches zero, as your argument might
suggest, but I think other processes could act in different ways.

-Peter



From cliechti at gmx.net  Sun Apr  6 14:15:57 2003
From: cliechti at gmx.net (Chris Liechti)
Date: 6 Apr 2003 20:15:57 +0200
Subject: Python commands
References: <3e8bb9c5$1_1@news.iprimus.com.au> <Xns9352D1FE6C1E0cliechtigmxnet@62.2.16.82> <3e8e5d85$1_1@news.iprimus.com.au>
Message-ID: <Xns9355CE635AD66cliechtigmxnet@62.2.16.82>

"Paul Dexter" <dextech at terrigal.net.au> wrote in
news:3e8e5d85$1_1 at news.iprimus.com.au: 

>     Hi all,
>                 Thanks for all the advice and links to the libraries.
>     I sometimes build hardware to connect to the computer via the
>     serial or 
> parrallel ports, so a peripheral driver is not usually available.
>     I now have a lot of reading to do, to get Python to transfer data
>     thro'  the port to my device.

>     Things aren't as straight forward as the DOS days with all the
> multitasking multithreading features we seem to demand.         LOL

1. do you realy need threads for your app?
1a. if so consider using a Queue and exactly one thread for reading and 
writing (or if your protocol allows this, one thread for reading and an 
other for writing)

2. learn asynchronous programming. Twisted provides a good framework to do 
so. it already has serial port support (for win32 too, in the CVS version)
it does not yet have parallel port support, but depending on your 
application, it may be easy to add.

both, the queue or async programming, can make your life a lot easier.

chris
-- 
Chris <cliechti at gmx.net>




From martin at v.loewis.de  Wed Apr  2 16:40:16 2003
From: martin at v.loewis.de (Martin v. =?iso-8859-15?q?L=F6wis?=)
Date: 02 Apr 2003 23:40:16 +0200
Subject: unicode bug in turkish characters?
References: <mailman.1049269265.16795.python-list@python.org>
Message-ID: <m3znn8k27z.fsf@mira.informatik.hu-berlin.de>

"Oktay Safak" <oktaysafak at ixir.com> writes:

> When I try to convert the character "i" to uppercase what comes
> out is "I" where it should have a dot on top of it instead. 

How do you do convert to upper-case? If you use

>>> u"i".upper()
u'I'

then this uses the Unicode character database, which is
language-independent, and the upper-case variant of "i" is "I".

However, if you do

>>> locale.setlocale(locale.LC_ALL,"tr_TR")
'tr_TR'
>>> "i".upper()
'\xdd'
>>> "i".upper().decode("iso-8859-9")
u'\u0130'
>>> unicodedata.name(_)
'LATIN CAPITAL LETTER I WITH DOT ABOVE'

then this generates the upper-case version according to the
locale. Notice that this works for byte strings only.

> Also, when I try to convert the uppercase i with dot to lowercase,
> it comes out as itself where "i" should be the character produced.

I can't reproduce this:

>>> u'\u0130'.lower()
u'i'

seems to work fine.

> I might be missing something but I have a feeling that this is
> a bug since the case toggle works perfectly with turkish
> characters that do not exist in ascii. With i and I though, which
> do exist in ascii, it's all messed up.

I'm sure you are missing something. Since you did not say what you did
*exactly* (i.e. which functions you've used with what parameters), it
is hard to tell what your problem is.

However, I'm almost certain that Python has no bug here.

Regards,
Martin



From jason at tishler.net  Thu Apr  3 12:20:28 2003
From: jason at tishler.net (Jason Tishler)
Date: Thu, 03 Apr 2003 12:20:28 -0500
Subject: Cygwin Python regression testers wanted
In-Reply-To: <A7B5C22F20EE3C4E95744E2F03F1887F0AF50F@MAYOR.retzlaff.com>
References: <A7B5C22F20EE3C4E95744E2F03F1887F0AF50F@MAYOR.retzlaff.com>
Message-ID: <20030403172028.GD2896@tishler.net>

Jimmy,

Please post instead of sending private email.

On Thu, Apr 03, 2003 at 08:48:24AM -0800, Jimmy Retzlaff wrote:
> I'm not at all experienced with Cygwin Python (I use win32 Python and
> have cygwin around for various utilities), but I have an XP Pro system
> and thought I'd give it a shot. I updated my cygwin installation just
> before doing this. Python reports itself as:
> 
> Python 2.2.2 (#1, Mar  9 2003, 08:18:26)
> [GCC 3.2 20020927 (prerelease)] on cygwin
> 
> The output is attached.
>
> ...
> 172 tests OK.
> 1 test failed:
>    test_commands
> ...

XP Pro is the same as NT 4.0 and 2000 (as expected).

> Let me know if you'd like me to change anything and try again.

Did you rebase your system before running the regression test?

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6




From terin at alpha-designs.net  Mon Apr  7 17:35:26 2003
From: terin at alpha-designs.net (Terin)
Date: 7 Apr 2003 14:35:26 -0700
Subject: Got a kick out of this
Message-ID: <237247fb.0304071335.619a468b@posting.google.com>

I was doing random Google searches, and came upon this nugget of
humor. Someone else suggested I post it, so here's the link, hopefully
it will work...

http://vil.nai.com/vil/content/v_994.htm



From bokr at oz.net  Thu Apr 10 22:24:27 2003
From: bokr at oz.net (Bengt Richter)
Date: 11 Apr 2003 02:24:27 GMT
Subject: optimizing memory usage of large in memory structures
References: <aebfcc5a.0304090817.2c683c9e@posting.google.com>
Message-ID: <b7590r$u8h$0@216.39.172.122>

On 9 Apr 2003 09:17:40 -0700, theory79 at yahoo.com (Hugo Liu) wrote:

>hi,
>
>i'm hoping that someone out there has a solution to hack this problem.
>
>working with corpus analysis and large knowledge bases, i often have
>to load large arrays into memory.  each array element is just a text
>string.  although the text is only 5mb, python consumes 40+ mb.  i
>would suspect that the overhead is from storing type info for each
>element.  if the strings were all fixed length, i'd just store them
>ordered and fixed width into a single string and use bisect, but they
>aren't fixed width.
>
>does anyone know if there is a way to fake static typing for this
>situation or any other way to make this more efficient (without
>sacrificing operating system interoperability or having to write
>external C code?)  any suggestions or pointers would be much
>appreciated!
>
If your strings were put in an efficient-storage magic black box,
how would you want to access the strings? One sequential sweep
is a lot different from zillions of random accesses, and if
you do random access, what are the selection criteria?

You could write a class that held your 5mb as a single string
and used the array module to hold a an array of 'l'-integers
for start positions (assuming fixed-length delimiters in the big string
-- if not, another array of 'B'-integers for string lengths
(assuming 0 to 255) could be used. Or look for the delimiter each time).

That would define storage that should take little more than 5mb+4*ns
(+ns if separate length array) bytes, where ns is the number of strings.

But then you need access methods. E.g., you could define __getitem__ and
access (read) your variable length strings by an index to get the nth string.
This would allow random or sequential read access of your variable-length strings
as if it were an ordinary list of ordinary Python strings.

If it is a linguistic corpus, the chances are you have many repetitions of words,
and you could represent the original 5mb as a sequence of integer word identifiers
and use a dictionary so as not to store word-strings more than once each.

And you could pack unique word-strings into a homogeneous array end to end together with
an offset array and use indices into the latter as keys for lookup, so you could throw away
the dictionary you used for speed while building the compact lookup data.

There may well be better ways to store and manipulate your data, depending on what you
really have to do with it.

Then again, if space is your only worry, why not just buy some more memory? ;-)

Regards,
Bengt Richter



From ckea25d02 at sneakemail.com  Mon Apr 14 02:21:01 2003
From: ckea25d02 at sneakemail.com (Greg Ewing (using news.cis.dfn.de))
Date: Mon, 14 Apr 2003 18:21:01 +1200
Subject: Speed of string += string
In-Reply-To: <J5gma.9328$Gb1.6409@nwrdny01.gnilink.net>
References: <18s6dy4mcsgjg$.y4vahh6w5lp3.dlg@40tude.net> <a9Zla.7472$T34.203357@news2.tin.it> <roy-06B012.20145112042003@reader1.panix.com> <bJ2ma.5447$883.4511@nwrdny02.gnilink.net> <iz86drbfq4va$.1r0k6kjvji1h5.dlg@40tude.net> <J5gma.9328$Gb1.6409@nwrdny01.gnilink.net>
Message-ID: <b7djc8$deo4s$2@ID-169208.news.dfncis.de>

Carl Banks wrote:
>>But is that /guaranteed/ (as it is with STL), or does it just happen to be 
>>that way with the current implementation (however unlikely it is to 
>>change)?
> 
> It's not guaranteed, but if it ever changes, I'll eat my hat.

If it ever did change (for the worse), I think there would be a
LOT of VERY unhappy Python users!

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg




From guy at obstruction-no-spam.com  Fri Apr 18 17:49:43 2003
From: guy at obstruction-no-spam.com (Guy Middleton)
Date: Fri, 18 Apr 2003 21:49:43 GMT
Subject: Pythonic way to sum n-th list element?
References: <U6Zna.64376$BQi.35175@news04.bloor.is.net.cable.rogers.com> <Ab_na.34355$LB6.777939@news1.tin.it>
Message-ID: <Xp_na.64908$BQi.2699@news04.bloor.is.net.cable.rogers.com>

In article <Ab_na.34355$LB6.777939 at news1.tin.it>,
Alex Martelli  <aleax at aleax.it> wrote:
> Guy Middleton wrote:
> 
> > What is the most Pythonic way to sum the n-th element of a list of lists
> > (or of tuples)?
> 
> _most Pythonic_ is probably:
> 
> total = 0
> for sublist in listoflists:
>     total = total + sublist[n]
> 
> I think that's what Guido would do if asked -- it's simple, it's entirely
> obvious, it's totally maintainable and clear.

It turns out this way is almost twice as fast as the list comprehension:

listoflists = (
    (1, 1),
    (2, 2),
    ...
    (100, 100)
    )

print time.clock()
for i in xrange(10000):
    total1 = 0
    for sublist in listoflists:
	total1 = total1 + sublist[1]
print time.clock()
for i in xrange(10000):
    total2 = reduce(operator.add, [x[1] for x in listoflists])
print time.clock()
print total1, total2


$ ./tt.py 
0.046875
2.5703125
7.234375
5050 5050
$



From mwh at python.net  Tue Apr  8 13:27:07 2003
From: mwh at python.net (Michael Hudson)
Date: Tue, 8 Apr 2003 17:27:07 GMT
Subject: [Slightly OT]: More on ints and floats
References: <fjss6b.1021.ln@boundary.tundraware.com> <mailman.1049768113.16134.python-list@python.org> <kauu6b.f2d1.ln@boundary.tundraware.com>
Message-ID: <7h3wui4528k.fsf@pc150.maths.bris.ac.uk>

Tim Daneliuk <tundra at tundraware.com> writes:

> I am unclear, though, on whether a mathematician understands 1.0 to
> mean 1.0000000... (a single number) or the _neighborhood_ around 1
> with one digit precision.

I'm a mathematician, albeit not a numerical analyst, and I don't think
considering a float as the interval of real numbers to which it is the
closest floating point approximation is a particularly helpful point
of view.

Floats are perfectly normal rational numbers; it's the *operations*
which are weird.

Cheers,
M.

-- 
 The rapid establishment of social ties, even of a fleeting nature,
 advance not only that goal but its standing in the uberconscious
 mesh of communal psychic, subjective, and algorithmic interbeing.
 But I fear I'm restating the obvious.  -- Will Ware, comp.lang.python



From knight at baldmt.com  Tue Apr 15 12:54:41 2003
From: knight at baldmt.com (Steven Knight)
Date: Tue, 15 Apr 2003 11:54:41 -0500 (CDT)
Subject: ANN: Capon 0.1 - a software build framework
Message-ID: <Pine.LNX.4.44.0304151046120.16820-100000@theopera.home.baldmt.com>

Hi--

> > So how does Capon differ from SCons?

I originally replied to this question directly to Aahz with some
information about SCons.  He suggested this was of general interest and
that I share the info with the list.

Since I'm the project lead for SCons, please keep in mind the usual
caveats about my own biases, the fact that I don't know Capon as well as
I know SCons, etc...

> SCons has lots of features.  The main thrust of Capon is on providing
> the means to add features.

SCons has a lot of features partly because we've put a lot of emphasis
on clean APIs for adding features.  For example, SCons does have a lot
of canned, out-of-the-box objects for building programs, object files,
libraries, tar and zip files, pre-compiled headers, MSVC++ resource
files, and various document formats.  However, SCons creates these using
the same API that you'd use yourself if you were defining your own
Builder object.  User feedback indicates that this makes it very easy to
extend SCons.

>                             Capon is a language by itself, and Capon
> scripts are not Python scripts.  That sounds complex but isn't, the
> language is just Python in disguise.

Anders is alluding to the fact that the canonical SCons interface is
that build files are Python scripts that call the build-engine API.

This isn't the only way to use SCons functionality, though.  The SCons
build engine has been designed from the ground up as a separate set of
modules from the Python-script external interface.  You can thus embed
the SCons build engine in any other interface you want.

For example, Stephen Kennedy has already integrated the SCons build
engine with Python FAM (http://sourceforge.net/projects/python-fam) to
create a prototype continuous-integration build GUI, with very fast
rebuild times.  Asko Kauppi is even working at putting a Lua interface
around it...!

> It's not inconceivable that I could find a straightforward way to
> bridge Capon and SCons, providing "Caponic" style access to SCons
> features.  It's not inconceivable that someone else could, either,
> given the extensible nature of Capon.

Actually, this would be pretty easy to do, given the separability of the
SCons build engine that I described above.  (And we'd love to support
more interfaces... :-)

> Take a look at this implementation of 'clean', a function to delete
> intermediary files.
> 
> | import os  # import a Python module
> | .clean():  # lazy notation; think of this as "def clean(self):"
> |     # dependency_tree() is a generator that lists dependencies recursively
> |     for dep in .dependency_tree():
> |         dep.clean_this()
> |
> | .clean_this():
> |     pass
> |
> | .file.gen/clean_this():
> |     if .exists():
> |         if .verbose:
> |             print "clean() deleting %s" % (.filename,)
> |         os.unlink(.filename)
> 
> 'clean' is a simple function that walks a generator and calls
> 'clean_this' on each node.  The nodes typically represent files but
> don't have to.  Except for implicit self argument lists and function 
> bodies are straight Python.

The SCons build engine has a Node-based internal structure that
sounds very similar to this.  It looks to me like in Capon, you add
capabilities by writing Node methods directly, whereas in SCons you go
through an API.  (Anders, please correct me if I'm mistaken.)

The SCons Node infrastructure has pretty robust subclassing for the file
system Nodes, with clean separation between the dependency management
parts (in the base class) and the file-system-specific stuff in the
subclass.  File system Nodes are further subclassed so you can actually
have a build depend on a *directory*.  A directory's "build" consists of
building all of its contents (including subdirectories).  This comes in
really handy for packaging trees into tar and zip files.

If anyone has any follow-on questions about SCons, please let me know.
Thanks!

        --SK





From nonone at nowhere.com  Sun Apr 20 03:20:38 2003
From: nonone at nowhere.com (Mark Gibson)
Date: 20 Apr 2003 01:20:38 -0600
Subject: xmlrpclib, connection refused? revisited
References: <ARi5a.9343$Ua.656611@news2.west.cox.net> <mailman.1045817162.16116.python-list@python.org> <3E562CCF.749B6D18@engcorp.com>
Message-ID: <3ea24a46$1_3@omega.dimensional.com>

Peter Hansen <peter at engcorp.com> wrote:
: John Abel wrote:
:> 
:> I'm guessing your using SimpleXMLRPCServer?  What interface is the
:> server bound too?  It needs to be bound to the IP, rather than
:> localhost.  When specifiying the host/port, try using:
:> 
:> (socket.gethostbyname( socket.gethostname() ), 8888 )

: Better yet, use ('', 8888) as your address to listen on *all*
: interfaces, including the loopback interface (localhost).

: -Peter

I had the same problem, and using '' fixed it.  Can someone explain it what's
happening?

Mark



From ajsiegel at optonline.net  Tue Apr 22 10:54:17 2003
From: ajsiegel at optonline.net (Arthur)
Date: Tue, 22 Apr 2003 10:54:17 -0400
Subject: instance creatiion as an "event"
Message-ID: <000501c308df$0d12c910$0c02a8c0@Arts>

I had written -

>Standalone, yes. But, unless I'm missing something, the effects in an
>inheritance chain is different. If I inherit from TestPoint, and need to
>call TestPoint.__init__ in the Point_Child.__init__ to pick up
initilization
>routines common to the chain, I will now unavoidably be calling TestPoint's
>redraw method before I call Point_Child's.

So I'm thinking I should perhaps go back to a fundamentally different
scheme.  Instead of factory functions calling classes based on argument
composition - and therefore lots of classes and attendant inheritance
issues, there should be master classes with alternative redraw methods that
are selected in __init__ on the basis of the argument composition:

In effect:

class Point:
def __init__(self,*args):
   if args are 3 planes:
      self.redraw = self.redraw_from_planes
       elif  args are 2 lines:
            self.redraw = self.redraw_from_lines
        ...
       self.redraw()

I had experimented with this at one point and found it workable.  But went
to a factory function scheme, probably for no better reason then I thought
it was a more standard approach to the general issue I faced.  Serves me
right (if I am now right).

Art







From unendliche at hanmail.net  Tue Apr  1 21:18:54 2003
From: unendliche at hanmail.net (Seo Sanghyeon)
Date: 1 Apr 2003 18:18:54 -0800
Subject: PEP 313 - Roman numeral literals
References: <x7y92uhzpw.fsf@guru.mired.org>
Message-ID: <45e6545c.0304011818.474e324c@posting.google.com>

Mike Meyer wrote:
> PEP: 313
> Title: Adding Roman Numeral Literals to Python

A similar thing already has been implemented by Jean-Paul Calderone of
Twisted fame. Check out intrinsics.c, intrinsics_example.py and example.py at:
http://intarweb.us:8080/evil/
(Last two .py are basically identical. The latter seems better.)

Although this does not resolve _literal_ issue, I think this implementation
is worth a look for anyone interested in PEP 313.



From sschwarzer at sschwarzer.net  Mon Apr 21 12:39:57 2003
From: sschwarzer at sschwarzer.net (Stefan Schwarzer)
Date: Mon, 21 Apr 2003 18:39:57 +0200
Subject: FTP
In-Reply-To: <e62b26d3.0304151727.6de2e2e4@posting.google.com>
References: <e62b26d3.0304151727.6de2e2e4@posting.google.com>
Message-ID: <3EA41EDD.90801@sschwarzer.net>

Matt Singer wrote:
> I have an application with a PYQT interface that uploads large files. 
> Is there an async method to get how much data has been sent with
> storlines to update a progress bar?  Or is there an alternate ftp
> interface?

Some time ago, I wrote ftputil, which may help you; see
http://www.sschwarzer.net/python/python_software.html .

Stefan




From saveenr at online.microsoft.com  Sun Apr 13 22:54:06 2003
From: saveenr at online.microsoft.com (Saveen Reddy [MSFT])
Date: Sun, 13 Apr 2003 19:54:06 -0700
Subject: properties & exceptions
References: <3e9a0144$1@news.microsoft.com> <mailman.1050282441.10913.python-list@python.org>
Message-ID: <3e9a39c7@news.microsoft.com>

You got me.

Ugh, I edited and copied and pasted the code so many times, I completely
lost track of the need (and the code fragment) to do this.

Thanks very much, this solved my problem.
-s





From mads at troest.NEVERMORE.dk  Tue Apr  8 15:05:03 2003
From: mads at troest.NEVERMORE.dk (Mads Orbesen Troest)
Date: Tue, 8 Apr 2003 21:05:03 +0200
Subject: Loading a class the name of which is only known runtime
References: <13n0q1sc6o6oh$.hfa48a72df4n.dlg@40tude.net> <pan.2003.04.08.17.35.15.762651@webone.com.au>
Message-ID: <14ix3p3bd2ze4$.1nhibartngu4y$.dlg@40tude.net>

On Wed, 09 Apr 2003 03:35:19 +1000, Simon Burton wrote:

Hi Simon, tbanks for your response.

>>>> instance = locals()[name]()

Interesting suggestion (I learned something new here :), but the problem is 
that the class(es) must be external, residing in separate .py files 
(possible in a package) outside the executing script, so the solution 
proposed here won't do.

Regards,
   /\/\\ads Orbesen Troest



From pan-newsreader at thomas-guettler.de  Sat Apr  5 02:49:46 2003
From: pan-newsreader at thomas-guettler.de (Thomas Guettler)
Date: Sat, 05 Apr 2003 09:49:46 +0200
Subject: Pythonic way of web-programming
References: <a3b5b608.0304040904.146dfff5@posting.google.com>
Message-ID: <pan.2003.04.05.09.49.45.708544.878@thomas-guettler.de>

On Fri, 04 Apr 2003 19:04:37 +0200, Giorgi wrote:

> Hi all!
> 
> Let me start with the statement that I am not a professional
> web-programmer. My tasks are rather linked with the scientific
> calculations and algorithm prototyping.
> 
> Recently I have started to make a Georgian language based web-site for
> Python. It was a natural feeling not to use, say, php or perl for
> web-content management. I have gone through different approaches
> available in Python; namely:
> - CGI (with or without mod-python and / or cpy ) ; - ZOPE ; - Jython /
> J2EE ;
> - ASP .NET ;
> - Webware PSP.

It depends on what you want to do.

Zope is bloated. Nevertheless I use it because the ZCatalog lets you
index python objects easily. The downside: Most providers don't let you
use it.

If performance is not very important, I would use cgi. 

I was told, that scgi is better than mod_python. But never used one of
both: http://www.mems-exchange.org/software/scgi/

If you want it to be 100% Python you can use medusa: A webserver written
in python.

 thomas



-- 
Thomas Guettler <guettli at thomas-guettler.de>
http://www.thomas-guettler.de